Progress in Brain Research Volume 129 Nervous System Plasticity and Chronic Pain

List of Contributors

D. Alvares, Department of Anatomy and Developmental Biology, University College London, London WClE 6BT, UK L. Arendt-Nielsen, Center for Sensory-Motor Interaction, Laboratory for Experimental Pain Research, Aalborg University, Fredrik Bajers Vej 7D-3, DK-9100 Aalborg, Denmark H. Baba, Neural Plasticity Research Group, Department of Anesthesia and Critical Care, Massachusetts General Hospital, 149 Thirteenth Street, Room 4309, Charleston, MA 02 129-2000, USA B. Beland, Department of Anatomy and Developmental Biology, University College London, London WCIE 6BT, UK J. Benrath, Institut fur Physiologie und Pathophysiologie, Universitlt Heidelberg, Im Neuenheimer Feld 326 D-69 120 Heidelberg, Germany A. Berthele, Department of Neurology, Technical University Munich, Moehlstr. 28, 81675 Munich, Germany J.A. Black, Department of Neurology and PVA/EPVA Neuroscience Research Center, Yale University School of Medicine, 333 Cedar Street, New Haven, CT 06510, USA S.G. Black, Department of Neurology and PVA/EPVA Neuroscience Research Center, Yale University School of Medicine, 333 Cedar Street, New Haven, CT 06510, USA C. Brechtel, Institut ftir Physiologie und Pathophysiologie, Universitat Heidelberg, Im Neuenheimer Feld 326, D-69120 Heidelberg, Germany B. Bromm, Institute of Physiology, University Hospital Eppendorf, D-20246 Hamburg, Germany A. Calignano, Department of Pharmacology, University of Naples, Naples 80131, Italy G. Carli, Instituto di Fisiologia Umana, Universidad degli Studi di Siena, Via Aldo Moro, I-53 100 Siena, Italy E. Carstens, Section of Neurobiology, Physiology and Behavior, University of California, 1 Shields Avenue, Davis, CA 95616, USA K.L. Casey, University of Michigan, Neurology Service, VA. Medical Center, Ann Arbor, MI 48105, USA J.M. Castro-Lopes, Institute of Histology and Embryology and IBMC, Faculty of Medicine of Oporto, Al Hemani Monteiro, 4200 Porto, Portugal D.R. Clohisy, Department of Orthopaedic Surgery and Cancer Center, University of Minnesota, Minneapolis, MN 55455, USA B. Conrad, Department of Neurology, Technical University Munich, Moehlstr. 28, 81675 Munich, Germany S.P Cook, Vollum Institute, Oregon Health Sciences University, 3181 SW Sam Jackson Park Road, Portland, OR 97201, USA S.V. Coutinho, UCLA/CURE Neuroenteric Disease Program, WLA VA Medical Center, Building 115, Room 223, 11301 Wilshire Boulevard, Los Angeles, CA 90073, USA

vi

A.D. Craig, Division of Neurosurgery, Barrow Neurological Institute, 350 West Thomas Road, Phoenix, AZ 85013, USA T.R. Cummins, Department of Neurology and PVA/EPVA Neuroscience Research Center, Yale University School of Medicine, 333 Cedar Street, New Haven, CT 06516, USA S.D. Dib-Hajj, Department of Neurology and PVA/EPVA Neuroscience Research Center, Yale University School of Medicine, 333 Cedar Street, New Haven, CT 06510, USA J.O. Dostrovsky, Department of Physiology, Medical Sciences Building, Room 3305, 1 King’s College Circle, University of Toronto, Toronto, ON M5S lA8, Canada P.M. Dougherty, Department of Neurosurgery, Johns Hopkins University, Meyer Building 7-113, 600 North Wolfe Street, Baltimore, MD 21287-7713, USA A.M. Drewes, Department of Medical Gastroenterology, Aalborg Hospital, DK-9100 Aalborg, Denmark A. Ebersberger, Institut fur Physiologie Friedrich-Schiller-Universitat Jena, Teichgraben 8, D-07740 Jena, Germany S.P. Sutherland, Vollum Institute, Oregon Health Sciences University, 3181 SW Sam Jackson Park Road, Portland, OR 9720 I, USA M.P. Finke, Neurosystems Center and Departments of Preventive Sciences, Psychiatry, Neuroscience and Cancer Center, University of Minnesota, Minneapolis, MN 55455, USA and VA Medical Center, Minneapolis, MN 55417, USA M. Fitzgerald, Department of Anatomy and Developmental Biology, University College London, London WClE 6BT, UK H. Flor, Central Institute of Mental Health, J5, Neuropsychology and Clinical Psychology Unit 68342 Mannheim, Germany C. Forster, Institut fur Physiologie und Experimentelle Pathophysiologie, Universitat Erlangen/Nurnberg, Universitltsstr. 17, 91054 Erlangen, Germany I.M. Garonzik, Department of Neurosurgery, Johns Hopkins University, Meyer Building 7-113,600 North Wolfe Street, Baltimore, MD 21287-7713, USA G.F. Gebhart, Department of Pharmacology, Bowen Science Building, University of Iowa College of Medicine, Iowa City, IA 52242, USA G. Gerber, Department of Anatomy, Histology and Embryology, Semmelweis University of Medicine, Tuzolto utca 58, 1094 Budapest, Hungary H.-J. Habler, Physiologisches Institut, Christian-Albrechts-Universitat zu Kiel, Olshausenstr. 40, 24098 Kiel, Germany H.O. Handwerker, Institut fur Physiologie und Experimentelle Pathophysiologie, Universitat Erlangen/Ntimberg, Universitatsstr. 17, 91054 Erlangen, Germany M. Hasenbring, Abteilung Medizinische Psychologie, Ruhr Universitat Bochum, Gebaude MA-O/145, Universitatsstr 150,44780 Bochum, Germany B. Heinke, Institut fur Physiologie und Pathophysiologie, Universitat Heidelberg, Im Neuenheimer Feld 326, D-69120 Heidelberg, Germany U. Hoheisel, Institut fur Anatomie und Zellbiologie, Im Neuenheimer Feld 307, D-69120 Heidelberg, Germany K. Hole, Department of Physiology, University of Bergen, Arstadveien 19, N-5009 Bergen, Norway P. Honore, Neurosystems Center, Departments of Preventive Sciences, Psychiatry and Neuroscience, University of Minnesota, 18-208 Moos Tower, 515 Delaware Street, Minneapolis, MN 55455, USA and VA Medical Center, Minneapolis, MN 55417, USA C.H. Hsu, Department of Biomedical Sciences, Iowa State University, Ames, IA 500111250, USA

vii SE. Hua, Department of Neurosurgery, Johns Hopkins University, Meyer Building 7-113, 600 North Wolfe Street, Baltimore, MD 21287-7713, USA T. Imanishi, Department of Anesthesiology, Kansai Medical University, lo-15 Fumizono, Moriguchi Osaka 570, Japan S.L. Ingram, Oregon Health Sciences University, 3 181 SW Sam Jackson Park Road, Portland, OR 97201, USA D. Isaev, Department of Biomedical Sciences, Iowa State University, Ames, IA 5001 I1250, USA S. Ito, Department of Medical Chemistry, Kansai Medical University, lo-15 Fumizono, Moriguchi Osaka 570, Japan W. Jlnig, Physiologisches Institut, Christian-Albrechts-Universitat zu Kiel, Olshausenstr. 40,24098 Kiel, Germany T.S. Jensen, Department of Neurology, Aarhus University Hospital, DK-8000 Aarhus C, Denmark B.J. Kerr, Neuroscience Research Centre, Division of Physiology, St. Thomas’ Hospital Campus, Lambeth Palace Road, London SE1 7EH, UK G. La Rana, Department of Pharmacology, University of Naples, Naples 8013 1, Italy R.J. Laursen, Laboratory for Experimental Pain Research, Center for Sensory-Motor Interaction, Aalborg University, Fredrik Bajers Vej 7D-3, DK-920 Aalborg, Denmark J.-I. Lee, Department ef Neurosurgery, Samsung Medical Center, Sungkyunkwan University School of Medicine, Seoul, Korea EA. Lenz, Department of Neurosurgery, Meyer Building 7-113, Johns Hopkins Hospital, 600 North Wolfe Street, Baltimore, MD 21287-7713, USA A.R. Light, Department of Cellular and Molecular Physiology, University of North Carolina, School of Medicine, Chapel Hill, NC 27599, USA P. Loubet-Lescoulie, Department of Pharmacology, University of California, 360 Med Surge II, Irvine, CA 92697-4625, USA W. Magerl, Institute of Physiology and Pathophysiology, Johannes Gutenberg-University, Saarstr. 21, D-55099 Mainz, Germany A.B. Malmberg, Roche Bioscience, Neurobiology Unit, 3401 Hillview Avenue, Palo Alto, CA 94304, USA I?W. Mantyh, Neurosystems Center, University of Minnesota, 18-208 Moos Tower, 515 Delaware Street, Minneapolis, MN 55455, USA E.W. McCleskey, Vellum Institute, Orgeon Health Sciences University, 3181 SW Sam Jackson Park Road, Portland, OR 97201, USA S.B. McMahon, Neuroscience Research Centre, Guy’s, King’s and St. Thomas’ School of Biomedical Sciences, King’s College London, London SE1 7EH, UK P.M. Menning, Neurosystems Center, Departments of Preventive Sciences, Psychiatry and Neuroscience, University of Minnesota, 18-208 Moos Tower, 5 15 Delaware Street, Minneapolis, MN 55455, USA and VA Medical Center, Minneapolis, MN 55455, USA S. Mense, Institut fiir Anatomie und Zellbiologie, Im Neuenheimer Feld 307, D-69120 Heidelberg, Germany T. Minami, Department of Medical Chemistry, Kansai Medical University, lo-15 Fumizono, Moriguchi Osaka 570, Japan K.A. Moore, Massachusetts General Hospital-East, Neural Plasticity Research Group, 149 Thirteenth Street, Room 4309, Charleston, MA 02129-2000, USA J. Nebe, Rheinische Landes- und Hochschulklinik Essen, Virchowstrassse 174, D-45147 Essen, Germany

...

Vlll

V. Neugebauer, Department of Neuroanatomy and Neurosciences and Marine Biomedical Institute, University of Texas Medical Branch, Galveston, TX 77555-1069, USA M.L. Nichols, Neurosystems Center, Departments of Preventive Sciences, Psychiatry and Neuroscience, University of Minnesota, 18-208 Moos Tower, 515 Delaware Street, Minneapolis, MN 55455, USA and VA Medical Center, Minneapolis, MN 55417, USA L. Nikolajsen, Departments of Anaesthesiology and Danish Pain Research Center, Aarhus University Hospital, DK-8000 Aarhus C, Denmark E. Okuda-Ashitaka, Department of Medical Chemistry, Kansai University, lo-15 Fumizono, Moriguchi Osaka 570, Japan A. Pertovaara, Department of Physiology, Institute of Biomedicine, University of Turku, Kiinamyllynkatu 10, FIN-20520 Turku, Finland G. PethS, Institute of Pharmacology and Pharmacotherapy, Faculty of Medicine, University of Pets, Szigeti u. 12, H-7624 Pets, Hungary D. Piomelli, Department of Pharmacology, University of California, 360 Med Surge II, Irvine, CA 92697-4625, USA M.L. Ramnaraine, Department of Orthopaedic Surgery and Cancer Center, University of Minnesota, Minneapolis, MN 55455, USA M. Randic, Department of Biomedical Sciences, Iowa State University, Ames, IA 5001 l1250, USA P.W. Reeh, Institute of Physiology and Experimental Pathophysiology, Universitltsstr. 17, D-91054 Erlangen, Germany M. Reynolds, Department of Anatomy and Developmental Biology, University College London, London WClE 6BT, UK R. Ringler, Institut ftir Physiologie und Experimentelle Pathophysiologie, Universitat Erlangen/Numberg, Universitatsstr. 17, 91054 Erlangen, Germany S.D. Rogers, Neurosystems Center and Departments of Preventive Sciences, Psychiatry, Neuroscience and Cancer Center, University of Minnesota, Minneapolis, MN 55455, USA and VA Medical Center, Minneapolis, MN 55417, USA L.H. Rowland, Department of Neurosurgery, Johns Hopkins University, Meyer Building 7-l 13,600 North Wolfe Street, Baltimore, MD 21287-7713, USA R. Ruscheweyh, Institut fur Physiologie und Pathophysiologie, Universitat Heidelberg, Im Neuenheimer Feld 326, D-69120 Heidelberg, Germany J.L. Salak-Johnson, Neurosystems Center and Departments of Preventive Sciences, Psychiatry, Neuroscience and Cancer Center, University of Minnesota, Minneapolis, MN 55455, USA and VA Medical Center, Minneapolis, MN 55417, USA J. Sandktihler, Institute of Physiology and Pathophysiology, University of Heidelberg, Im Neuenheimer Feld 326, D-69 120 Heidelberg, Germany J. Schadrack, Max-Planck Institute of Psychiatry, Clinical Neuropharmacology, Kraepelinstr. 2, 80804 Munich, Germany H.-G. Schaible, Institut fur Physiologie, Friedrich-Schiller-Universitat Jena, Teichgraben 8, D-07740 Jena, Germany E. Scharein, Institute of Physiology, University Hospital Eppendorf, D-20246 Hamburg, Germany M.J. Schwei, Neurosystems Center and Departments of Preventive Sciences, Psychiatry, Neuroscience and Cancer Center, University of Minnesota, Minneapolis, MN 55455, USA and VA Medical Center, Minneapolis, MN 55417, USA

ix J.N. Sengupta, Gastrointestinal Pharmacology, Preclinical R&D, AstraZeneca, S-43 1 83 Molndal, Sweden X. Su, Department of Pharmacology, College of Medicine, The University of Iowa, Bowen Science Building, Iowa City, IA 52242, USA F. Svendsen, Department of Physiology, University of Bergen, Arstadveien 19, N-5009 Bergen, Norway T.R. Tolle, Department of Neurology, Technical University Munich, Moehlstr 28, 81675 Munich, Germany S.W.N. Thompson, Neuroscience Research Centre, Guy’s, King’s and St. Thomas’ School of Biomedical Sciences, King’s College London, London SE1 7EH, UK A. Tjolsen, Department of Physiology, University of Bergen, Arstadveien 19, N-5009 Bergen, Norway C. Torsney, Department of Anatomy and Developmental Biology, University College London, London WClE 6BT, UK R.-D. Treede, Institute of Physiology and Pathophysiology, Johannes Gutenberg-University, Saarstr. 21, D-55099 Maim, Germany C. Vahle-Hinz, Institute of Physiology, University Hospital Eppendorf, D-20246 Hamburg, Germany H. Vanegas, Instituto Venezolano de Investigaciones Cientificas (IVIC), Apartado 21827, Caracas 1020-A, Venezuela C.J. Vierck, Jr., Department of Neuroscience, University of Florida, College of Medicine, Brain Research Institute, Gainesville, FL 32610-0244, USA S.G. Waxman, Department of Neurology and LCI 707 Neuroscience Research Center, Yale University School of Medicine, 333 Cedar Street, New Haven, CT 06510, USA O.H.G. Wilder-Smith, Nociception Research Group Berne University, Bubenbergplatz 11, CH-3011 Berne, Switzerland C. J. Woolf, Neural Plasticity Research Group, Department of Anesthesia and Critical Care, Massachusetts General Hospital, 149 Thirteenth Street, Room 4309, Charleston, MA 02129-2000, USA D. Wynick, Department of Medicine, University of Bristol, Marlboro Street, Bristol B52 8HW, UK R.P. Yezierski, Department of Neurological Surgery and the Miami Project, University of Miami, 1600 N.W. 10th Avenue, R-48, Miami, FL 33136, USA D.-H. Youn, Department of Biomedical Sciences, Iowa State University, Ames, IA 5001 l1250, USA M. Zhuo, Departments of Anesthesiology, Anatomy and Neurobiology, Washington University School of Medicine, Campus Box 8054, 660 S. Euclid Avenue St. Louis, MO 63110, USA W. Zieglgansberger, Max-Planck Institute of Psychiatry, Clinical Neuropharmacology, Kraepelinstr. 2, 80804 Munich, Germany

Xl

Foreword The end of the Decade of the Brain and the beginning of a new century provide an excellent opportunity to review the advances that have been made in the neurobiology of pain. The evidence is indisputable that the nervous system does not always reliably encode and convey unaltered the consequence of nociceptor activation. Rather, nervous system plasticity may contribute to or become the sole cause for the amplification and chronicity of pain. During the last decade the number of research groups and clinics that have devoted their efforts to study the neurobiology of pain and pain management has increased worldwide and impressive progress has been made in at least three major fields: . The multiple disciplines involved in pain research have developed a common language to express questions and interpret results. This advance is both prerequisite to and proof of the importance of an interdisciplinary approach to enhanced understanding of the neurobiology of pain and improved pain management. . Vertical integration in pain research has reached a level that now allows correlation of well defined clinical symptoms with new information about the molecular and cellular mechanisms of pain. . The clinical relevance of pain research performed in basic science laboratories has increased remarkably, and this is highly dependent on the above two items. The present volume on Nervous System Plasticity and Chronic Pain reflects recent progress that has been made in all of these areas. The chapters cover a broad spectrum of latest achievements in pain research from the molecular to the perceptual level. When reading the contributions, it becomes clear that new concepts and ideas developed in one arena of pain research have had impact on concepts and hypotheses important to other fields of pain research. As in all of the life sciences, the degree of reductionism is dictated by the mutual interdependence of available technologies and the scientific question asked. While the attention paid by the scientific community to the various hierarchical levels of nociception waxes and wanes, admittedly with some fashionable exaggerations, future pain research must unquestionably continue to address all organizational levels of the nervous system from the molecular biology and synaptic impulse transmission to systems and higher brain functions. Much of the foundation on which future pain research will rest is described in this volume. However, neither acute nor chronic pain are nosological entities and one should therefore not expect that pain research will eventually develop a unifying theory. Rather, further progress is to be expected by defining better models and mechanisms for specific, well-characterized clinical pain syndromes. This will be the key to rational, evidence- and mechanism-based management and prevention of chronic pain. The progress in pain research documented in this book justifies optimism at the beginning of this century that this goal is attainable in the near future. J. Sandktihler B. Bromm G.F. Gebhart (Editors)

J. Sandkiihler, B. Bmmm and G.F. Gebhart (Eds.) Progressin Brain Research,Vol. 129 0 2000 Elsevier Science B.V. All rights reserved

CHAPTER 1

Sodium channels and the molecular pathophysiology of pain Theodore R. Cummins, SulaymanD. Dib-Hajj, Joel A. Black and StephenG. Waxman * Department of Neurology and PVAIEPVA Neuroscience Research Centel; Yale Universig School of Medicine, New Haven, CT 06510; and Rehabilitation Research Center: VA Medical Center; West Haven, CT 06516, USA

Introduction Nociceptive pathways begin with dorsal root ganglion (DRG) and trigeminal neurons, which constitute the first links in the chain of neurons making up the somatosensory system. These cells encode information in the form of series of action potentials whose depolarizing upstrokes are produced by sodium channels. Under normal circumstances DRG and trigeminal neurons are relatively quiescent unless they are stimulated, and when stimulated they produce highly modulated series of action potentials that convey, to the brain, information about the external world. Following some nerve injuries and when there is inflammation of peripheral target tissues, however, these primary sensory neurons can become hyperexcitable, and can give rise to unprovoked spontaneous action potential activity or pathological bursting which can contribute to chronic pain (Ochoa and Torebjork, 1980; Wall and Devor, 1981; Nordin et al., 1984; Zhang et al., 1997). Classical neurophysiological doctrine referred, until the last decade, to the sodium channel, but modem molecular neurobiology has taught us that at least eight distinct voltage-gated sodium channels are present within the nervous system. The different sodium channels

* Corresponding author: S.G. Waxman, Department of Neurology LCI 707, Yale Medical School, New Haven, CT 06510, USA. Tel.: +1 (203) 7856351; Fax: +l (203) 785-7826; E-mail: [email protected]

share a common motif but are encoded by different genes which endow them with different amino acid sequences, and they are expressed in a regionallyand temporally-specific manner within the nervous system. DRG neurons have provided an especially tractable cell-type in which to dissect sodium channel expression. Over the past decade, molecular and electrophysiological methods have begun to identify a complex ensemble of sodium channels in DRG neurons. It has recently become clear that there are several sensory neuron-specific sodium channels that are preferentially expressed in DRG and trigeminal neurons. Moreover, it has become apparent that hyperexcitability of injured DRG neurons is due, at least in part, to injury-induced changes in the expression of sodium channels, including the downregulation of transcription of several sodium channel genes and the up-regulation of transcription of at least one previously silent sodium channel gene. In this chapter we review recent advances in the understanding of sodium channel expression in DRG neurons, and of the changes in expression of sodium channel genes that occur in these sensory neurons following injury. Multiple sodium channels in dorsal root ganglion neurons It has been known since the earliest patch clamp studies that DRG neurons exhibit multiple, distinct sodium currents, which can be differentiated on the basis of their different voltage-dependence

4

1.OL-o

q

0.8 -

8.

2m -0

.

F

0 (\ ‘\

0.6

L

IOnA 2 ms

0.4 -

: ,I !

b

0.2 0.0

i. \\

, -100

L-ii&q-D -80 -60

I

!

1, -40

-20

, 0

1.0 0.8

=: $ rr.

0.6

2

0.4

ti cr:

0.2

5 $ a

0.0

voltage (mV)

0.0 b

I -100

I-Be-4 -80 -60

-40

b-., -20

I (-Jo 0

voltage (mV)

1.0 0.8

3 f$ -.

0.6

2

voltage (mV) Fig. 1. Different types of voltage-gated sodium currents in different types of DRG neurons, (A) Patch clamp recording showing fast, TTX-sensitive sodium current (left) and the corresponding steady-state activation and inactivation curves, from a muscle afferent (right). (B) Slow, TTX-resistant sodium current from a small lB4+ DRG neuron. (C) Sodium currents from an lB4small DRG neuron exhibit both fast, TTX-sensitive and slow, TTX-resistant components. This can be appreciated both in the recordings and in the steady-state activation and inactivation curves. The different currents are produced by distinct sodium channels,

and kinetics; these sodium currents also show pharmacological differences, including varying degrees of sensitivity to the neurotoxin tetrodotoxin (TTX)

(Kostyuk et al., 1981; Caffrey et al., 1992; Roy and Narahashi, 1992; Elliott and Elliott, 1993). Fig. 1 shows sodium currents from three different DRG

5

neurons and illustrates their diversity. Electrophysiological evidence indicates that different functional classes of DRG neurons, (e.g., cutaneous vs. muscle afferents) express functionally different types of sodium channels (Honmou et al., 1994). Moreover, patch clamp recordings demonstrate the presence of multiple sodium currents in some types of DRG neurons, implying that they can co-express several types of sodium channels (Cummins and Waxman, 1997; Cummins et al., 1999; Dib-Hajj et al., 1999~). Study of the mRNAs encoding the various channels, by RT-PCR or in situ hybridization, provides precise information about which sodium channel genes are transcriptionally active. As might be predicted from the multiple sodium currents that are present in DRG neurons, these cells express at least six mRNAs encoding different sodium channels as seen in Fig. 2 (Black et al., 1996). Sodium channels a1 and Na6 (which are also expressed at high levels by other neuronal cell types within the CNS) produce ITX-sensitive sodium currents, and are expressed at high levels in large and medium-size DRG neurons and at lower levels in small DRG neurons. Notably, DRG and trigeminal neurons express at least three sodium channel transcripts which are not present at significant levels in other neuronal cell types. (1) PN 1/hNE, which is present in virtually all DRG neurons, encodes a TTX-sensitive channel. In in vitro model systems where it has been studied, PNl/hNE channels are preferentially located near the terminals of DRG neurons (Toledo-Aral et al., 1997) and,

as described below, they have physiological properties which poise them to amplify sensory generator potentials (Cummins et al., 1998). (2) SNS/PN3,

B M123456789M

600 bp Fig. 2. (A) DRG neurons express multiple sodium channels. Sodium channel o-subunit mRNAs visualized in adult rat DRG by in situ hybridization with subtype-specific antisense riboprobes. mRNAs for six different sodium channels (a-l, Na6, hNE/PNl, SNS, NaN and NaG) are present at moderate-to-high levels in DRG neurons. Hybridization with sense riboprobes, e.g., for NaG(s), does not result in signal. Bar indicates 100 pm. (Reproduced with permission from Black et al. (1996)) (B) Restriction mapping of Na channel domain 1 shows a similar distribution of sodium channel mRNAs in DRG. ‘M’ lanes contain lOO-bp ladder marker. Lane 1 contains the RT-PCR amplification product from DRG cDNA. Lanes 2-9 show the result of cutting this DNA with EcoRV, EcoNl, Aval, Sphl, BamHl, AfIIII, X&l, and EcoRI, which are specific to subunits a-1, -II, -III, Na6, PNl, SNS, NaG, and NaN, respectively. Reproduced with permission from Dib-Hajj et al. (1998b).

200

100

6 which is expressed preferentially in small DRG and trigeminal neurons, produces a sodium current that is relatively resistant to TTX (Akopian et al., 1996; Sangameswaran et al., 1996). (3) NaN, initially cloned and sequenced by Dib-Hajj et al. (1998b) and subsequently by Tate et al. (1998) who call it SNS-2, is also expressed preferentially in small DRG and trigeminal neurons, especially IB4-positive neurons which are responsive to GDNF (Fjell et al., 1999b). The human gene for NaN has been mapped to locus 3~21-24 on chromosome 3 (Dib-Hajj et al., 1999a). Based on the presence of a serine at position 355, Dib-Hajj et al. (1998b) predicted that NaN encodes a TTX-resistant sodium channel; consistent with this, a persistent voltage-dependent TTX-resistant sodium current is present in DRG neurons from SNS-null transgenic mice (Cummins et al., 1999). The preferential expression of two TTX-resistant sodium channels, SNS/PN3 and NaN, within small DRG neurons, provides a molecular basis for the electrophysiological observation (Kostyuk et al., 1981; Caffrey et al., 1992; Roy and Narahashi, 1992; Rizzo et al., 1994; Rush et al., 1998; Scholz et al., 1998) that these DRG neurons express prominent TTX-resistant sodium currents as well as the more conventional TTX-sensitive sodium currents that are present in most other neuronal cells. SNS/PN3 and NaN are both preferentially expressed in small-diameter DRG neurons, which include nociceptive cells, although NaN tends to be preferentially expressed in IB4-positive cells in which there are ret and GFR receptors, while SNS is not (Fjell et al., 1999b). There is, in fact, evidence that TTX-resistant sodium channels participate in the conduction of action potentials within nociceptive sensory neurons and their axons (Jeftinija, 1994; Quasthoff et al., 1995; Brock et al., 1998). Different sodium channels contribute different electrical properties to DRG neurons The different types of sodium channels possess different physiological characteristics including voltage-dependence, kinetics, recovery properties, etc. The presence of multiple types of sodium channels within DRG neurons raises the question of the contribution of each to the overall function of the cell. One approach to this problem is provided by a

‘bottom-up’ analysis. Here, the gene for a single type of sodium channel is transfected into a host cell that normally does not express high levels of sodium channels. As a result of this, the channel of interest can be studied in relative isolation. An example of the ‘bottom-up’ approach is shown in Fig. 3B,C,D, which shows the result of this type of analysis (Cummins et al., 1998) for the PNl/hNE channel. For comparison, in Fig. 3A, patch clamp recordings are shown for SkMl (muscle) sodium channels, also transfected into human embryonic kidney (HEK293) cells so that they are expressed in relative isolation. Like most sodium channels, the SkMl channels require sudden, relatively large, depolarizations in order to be activated; as a result, in response to slow depolarization close to resting potential these channels do not open. This can be demonstrated by recording the response to a slow ramp-like stimulus (0.23 mV/ms); in response to this type of stimulus the SkMl channels do not generate a current (Fig. 3A). In contrast to this, the PNl/hNE channels, expressed in HEK293 cells, activate and show a distinct response to the slow ramp-like stimuli (Fig. 3B). Moreover, the activity of these channels shows a unique pharmacological signature, since the currents they produce are blocked by TTX (Fig. 3C) and enhanced by cadmium (Fig. 3D). Given that the PNl/hNE channel exhibits these properties when studied by ‘bottom-up’ analysis, the next question is: Does the PNl/hNE channel play a similar functional role in its native environment, within DRG neurons? To answer this question we used the information gleaned from the ‘bottom-up’ analysis, to electrophysiologically dissect DRG neurons using a ‘top-down’ approach (Cummins et al., 1998). Having learned that small, gradual depolarizations constitute an effective stimulus for the PNl/hNE sodium channel within the HEK expression system, we applied ramp stimuli to intact DRG neurons, and found that they evoke a depolarizing response within these cells (Fig. 3E) that is similar to that of the isolated PNl/hNE channels (Fig. 3B). This ramp current within intact DRG neurons exhibits a pharmacological profile of responses to TTX and cadmium (Fig. 3F) that is identical to that of the isolated PNl/hNE channels. The ‘bottom-up’ and ‘top-down’ analyses thus converge, and demonstrate that in intact DRG neurons, the PNl/hNE chan-

-120 0

-80

40

(O.*SmVlmseOc)

f%msec

peak scaled -o-peak 100%

Fig. 3. ‘Bottom-up’ and ‘top-down’ analyses of the PNl/hNE sodium channel. In HEK 293 cells and in DRG neurons. (A) SkMl sodium channels, transfected into HEK293 cells, do not activate in response to slow ramp-like (0.23 mV/ms) depolarizations. (B) In contrast, ramp stimuli activate PNl/hNE channels transfected into HEK293 cells, generating a distinct inward current. These ramp currents are elicited at potentials close to resting potential. (C) PNl/hNE currents are blocked by TTX. (D) PNl/hNE currents are enhanced by Cd 2+. (E) In a ‘top-down’ analysis, similar stimuli are applied to DRG neurons, yielding a similar inward current. F: The ramp currents in intact DRG neurons exhibit a pharmacologic profile similar to these of isolated PNl/hNE currents, being blocked by TI’X and enhanced by Cd2+. (G) Within intact DRG neurons, the threshold for activation of ramp currents is similar to that for isolated PNI/hNE currents. (H) PNl/hNE currents in HEK293 cells, and ramp currents in intact DRG neurons, plotted together. Modified from Cummins et al. (1998).

nel responds to small, slow depolarizations close to resting potential, activating so as to produce inward (depolarizing) currents. Ramp currents have been recorded in a number of types of neurons, where they serve to amplify small depolarizing inputs, thereby acting as boosters (see, e.g., Schwindt and Crill, 1995; Jung et al., 1997; Pennartz et al., 1997; Parri and Crunelli, 1998) As a result of its targeting to nerve terminals of DRG neurons (Toledo-Aral et al., 1997), the PNl/hNE channel appears to be deployed to regions where it can amplify excitatory inputs such as sensory generator potentials (Cummins et al., 1998). Transgenic (‘knock-out’) technology provides another approach to the analysis of the physiology contribution of each channel subtype. Using patch clamp to study SNS knockout mutants (Akopian et al., 1999), Cummins et al. (1999) have observed a TTX-resistant (Ki = 39 f 9 PM) persistent sodium current (presumably produced by NaN channels) in mouse DRG neurons in which functional SNS channels are not expressed (Fig. 4). A similar TTX-resis-

tant persistent current can be recorded from rat DRG neurons (Cummins et al., 1999) and human DRG neurons obtained at autopsy (Dib-Hajj et al., 1999~) using appropriate stimulus protocols. The persistent current exhibits a hyperpolarized voltage dependence of activation (threshold - -70 mV, midpoint of activation = -41 mV in mouse DRG neurons) and steady-state inactivation (midpoint = -44 mV), with substantial overlap between activation and steadystate inactivation curves (Fig. 4B) which suggests that persistent sodium currents should be active near resting membrane potential (Cummins et al., 1999). Persistent sodium currents have been implicated in setting membrane resting potential (Stys et al., 1993), in subthreshold membrane potential oscillations (Kapoor et al., 1997), in amplification of depolarizing inputs (Schwindt and Crill, 1995) and in impulse initiation (Stafstrom et al., 1982). The low voltage-activated TTX-resistant persistent sodium channel in DRG neurons appears to participate in setting resting potential and regulating excitability in these cells.

I

Pn

120ms

voltage (mV)

C

Fig. 4. Persistent TTX-resistant sodium currents are expressed in small DRG neurons. (A) Representative TTX- resistant sodium currents recorded from a SNS-null neuron with 100 ms test pulses. (B) Activation (unfilled squares) and steady-state inactivation (filled squares) curves exhibit significant overlap for the TTX- resistant current in SNS-null neurons. Steady-state inactivation was measured with 500 ms prepulses. (C) ‘ITX- resistant persistent currents from a SNS-null neuron elicited with 2 s step depolarizations to the voltage indicated. All recordings were made with 250 nM TTX, 100 FM cadmium (to block calcium currents) and Vh,,td= -120 mV. Modified from Cummins et al. (1999).

Neuronal hyperexcitability axonal injury

can develop following

Early microelectrode studies (Eccles et al., 1958; Kuno and Llinas, 1970) demonstrated long-term alterations in the excitability of motor neurons following axonal transection, suggesting that there was an increase in sodium channel expression over the cell body and the dendrites as a consequence of axonal injury. Similar changes in excitability were subsequently observed following axonal injury in sensory neurons (Gurtu and Smith, 1988). Ion substitution experiments and pharmacological studies have confirmed a role of sodium channels in this hyperexcitability (Sernagor et al., 1986; Titmus and Faber, 1986). Immunocytochemical observations have more recently demonstrated the accumulation of abnormal aggregations of sodium channels, including the SNS/PN3 and type III channels, at the distal tips of injured axons (Devor et al., 1989; England et al., 1994, 1996a; Black et al., 1999). Taken together with electrophysiological studies which have suggested that increased sodium conductance can bias electro-

genesis so that DRG neurons produce inappropriate, repetitive action potential activity following axonal injury (Matzner and Devor, 1992, 1994; Zhang et al., 1997), and with observations on partial efficacy of sodium channel-blocking agents in experimental models of neuropathic pain and in humans with chronic neuropathic pain (see, e.g., Chabal et al., 1989; Devor et al., 1992; Omana-Zapata et al., 1997; Rizzo, 1997), these findings tend to support the conclusion that activation of sodium channels within injured DRG neurons can contribute to sensory neuron hyperexcitability associated with chronic pain. This conclusion leads to another question: which types of sodium channels are involved in abnormal hyperexcitability of DRG neurons after injury to these cells? Abnormal expression of sodium channel genes in injured DRG neurons Early studies (Waxman et al., 1994) demonstrated a significant up-regulation of expression of the previously silent a-III sodium channel gene in DRG

Control 13 M C-A C-A

5 C-A

7 C-A

Axotomy

14 C-A

a-SNS

123456 666 600

a-NaN -

400

Fig. 5. Expression of mRNA for sodium channel a-111 (top) is up-regulated, and mRNA for SNS (middle) and NaN (bottom) is down-regulated, in DRG neurons following transection of their axons within the sciatic nerve. The micrographs (right side) show in situ hybridizations in control DRG, and at 5-7 days post-axotomy. RT-PCR (left side) shows products of co-amplification of a-111 (top) and SNS (middle) together with l%actin transcripts in control and axotomized DRG (days post-axotomy indicated above gels), with computer enhanced images of amplification products shown below gels. Co-amplification of NaN (392 bp) and GAPDH (606 bp) (bottom) shows decreased expression of NaN mRNA at 7 days post-axotomy (lanes 2,4, 6) compared to controls (lanes 1, 3, 5). Top and middle modified from Dib-Hajj et al. (1996); bottom modified from Dib-Hajj et al. (1998b).

neurons following axonal transection within the sciatic nerve. More recent studies have demonstrated, in addition, a down-regulation of the SNS/PN3 gene expression which can persist at least 210 days following axotomy (Dib-Hajj et al., 1996), and downregulation of the NaN gene expression (Dib-Hajj et al., 1998b) in DRG neurons following axonal transection. Fig. 5 illustrates these changes in sodium channel gene expression in axotomized DRG neurons. Because SNS/PN3 and NaN encode TTX-resistant channels in DRG neurons, and since there is re-

duced expression of the SNS/PN3 and NaN sodium channel genes in DRG neurons following axonal transection, a reduction in TTX-resistant sodium currents would be expected in these cells following axotomy. Patch clamp studies have shown that there is, in fact, a significant attenuation of TTX-resistant sodium currents in DRG neurons following axonal transection within the sciatic nerve (Rizzo et al., 1995); this down-regulation persists in small DRG neurons for at least 60 days (Cummins and Waxman, 1997), consistent with the long-lasting changes in sodium channel gene expression that have been de-

control

C 4

1.00

1000

z 5

0.75

‘s,”

0.25 0.50

Density

500 250

0.00 -100

-50

0

0

mV

6

Current

750 g

ii t

TTX-R

days post axotomy

DPA6

D

Fig. 6. TTX-resistant sodium currents in small DRG neurons are down-regulated following axonal transection within the sciatic nerve. (A and B, left side) Whole-cell patch clamp recordings from representative control (A) and axotomized (B, 6 days post-axotomy) DRG neurons. The slowly-inactivating TTX-resistant component of sodium current is attenuated following axotomy. Steady-state inactivation curves (A and B, right side) show loss of a component characteristic of TTX-resistant currents. (C) Attenuation of TTX-resistant current persists for at least 60 days post-axotomy. (D) Attenuation of TTX-resistant persistent sodium currents, measured at 40 ms, in axotomized DRG neurons (6 days post-axotomy). Modified from Cummins and Waxman (1997).

scribed (Dib-Hajj et al., 1996, 1998b) in these cells (Fig. 6A-C). As shown in Fig. 6D, the attenuation of TTX-resistant sodium current in axotomized DRG neurons includes a significant reduction in TTXresistant persistent currents (Cummins and Waxman, 1997). A switch in the properties of the TTX-sensitive sodium currents is also observed in DRG neurons following axotomy (Fig. 7), with the emergence of a rapidly-repriming current (i.e., a current that recovers rapidly from inactivation) (Cummins and Waxman, 1997). Following axotomy, the time constant for recovery of TTX-sensitive sodium currents from inactivation is accelerated about four-fold in axotomized DRG neurons. It has been suggested that the expression of cl-111sodium channel underlies the emergence of the rapidly-repriming sodium current (Cummins and Waxman, 1997). The suggestion that type III channels produce a rapidly-repriming TTX-sensitive current is supported by several observations: First, rapidly-repriming TTX-sensitive current and expression of type III sodium channel protein show parallel

patterns of up-regulation following axotomy of the peripherally-directed (sciatic nerve) axons of DRG

-0 2

P8

1 .oo

0.75

i?! c 0.50 .g t2

0.25

l

-

0.00

recovery

time

DPA6 + DPA22

control

TTX-S

TTX-S

(ms)

Fig. 7. A rapidly repriming TTX-sensitive sodium current emerges in axotomized DRG neurons. The graph displays repriming (recovery of the TTX-sensitive sodium current from inactivation) in DRG neurons following axonal transection (6 and 22 days post-axotomy, results pooled). Repriming in uninjured controls is shown for comparison. The leftward shift in the recovery curve indicates that recovery from inactivation is accelerated, in DRG neurons studied at 6 and 22 days post-axotomy (DPA). Modified from Cummins and Waxman (1997).

11

neurons but not following axotomy of the centrallydirected (dorsal root) axons of these cells (Black et al., 1999). Second, type III sodium channels display rapid-repriming when expressed in a mammalian expression system (HEK 293 cells; Cummins, Dib-Hajj and Waxman, unpublished observations). Notably, abnormal accumulations of type III sodium channel protein can be detected close to the tips of injured axons within the experimental neuromas (Black et al., 1999), a site where aberrant hyperexcitability has been demonstrated (Scadding, 1981; Burchiel, 1988; Matzner and Devor, 1994). There are several mechanisms by which these electrophysiological changes may poise DRG neurons to fire spontaneously, or at inappropriately high frequencies, following injury to their axons. First, increased sodium conductance due to increased numbers of channels, per se, would be expected to lower

the threshold for action potential generation (Waxman and Brill, 1978; Matzner and Devor, 1992). Second, because of overlap between steady-state activation and inactivation curves, together with the relatively weak voltage-dependence of TTX-resistant sodium channels, co-expression of abnormal mixtures of channels may permit subthreshold potential oscillations, supported by TTX-resistant channels, to cross-activate TTX-sensitive sodium channels, thereby producing abnormal action potential activity (Rizzo et al., 1996). Third, as a result of the rapid repriming of the TTX-sensitive sodium current in DRG neurons following axotomy, refractory period would be expected to be shorter in injured DRG neurons, so that they can sustain higher firing frequencies (Cummins and Waxman, 1997). Fourth, some sodium channel subtypes in DRG neurons appear to be inactivated close to their resting potential

A

B Control

TTX-R current density

CCI

IOnA L-25 ms

Control

TTX-S current recovery at -80 mV ,P=e-. A’ /

D

TTX-S current repriming

kinetics

0.6.

/-,m’

d+ -L-

/ i / I,. ,./I 9-A k -.-.’ I

1

,,

0.01

recovery

Control Axotomy CCI

duration

Fig. 8. (A) Representative TTX-resistant Tl’X-resistant current density is reduced from Dib-Hajj et al. (1999b).

/

-140

0.1

(set)’

-120

recovery

-100

/

-80

potential

-60

-40

(mV)

sodium currents in control DRG neurons and after chronic constriction after CCI. (C and D) Recovery from inactivation is more rapid after CU.

injury (CCI). (B) A and B modified

12

(see, e.g. Kostyuk et al., 1981; Caffrey et al., 1992); if persistent currents contribute to the setting of resting potential in DRG neurons or their axons as demonstrated in axons of the optic nerve (Stys et al., 1993), reduction of TTX-resistant currents following axotomy could lead to a hyperpolarizing shift in resting potential in DRG neurons which, by relieving resting inactivation, might increase the availability of TTX-sensitive sodium current; this would, in turn, lower the threshold for action potential generation (Cummins and Waxman, 1997). Similar, though less extensive, changes occur in sodium channel gene expression in the chronic constriction injury model of neuropathic pain (Dib-Hajj et al., 1999b). As seen in Fig. 8, as would be expected in view of the changes in sodium channel mRNA, TTX-resistant sodium currents are attenuated and there is more rapid repriming of TTXsensitive currents in DRG neurons in the chronic constrictive injury model (Dib-Hajj et al., 1999b). These results are consistent with the observations of >80% loss of myelinated fibers and 60-80% loss of unmyelinated fibers in this model (Carlton et al., 1991). Neurotrophin modulation of sodium channel expression in DRG neurons Neurotrophins modulate sodium channel expression in DRG neurons. Early in vitro demonstrations that nerve growth factor (NGF) can affect sodium channel expression in DRG neurons (Aguayo and White, 1992; Zur et al., 1995), suggested that the effects of peripheral axotomy on sodium channel expression in DRG neurons could be due, at least in part, to loss of access to a peripheral supply of NGF. Black et al. (1997) studied an in vitro model that mimics axotomy, and demonstrated that NGF, delivered directly to DRG cell bodies in vitro, down-regulates u-111 mRNA and up-regulates SNS/PN3 mRNA expression in small DRG neurons. Dib-Hajj et al. (1998a) studied DRG neurons in vivo following axotomy, and showed that delivery of exogenous NGF to the nerve stump results in a partial rescue of SNS/PN3 mRNA levels and of TTX-resistant sodium current in small DRG neurons (Fig. 9). Fjell et al. (1999a) showed that NGF is required for SNS expression in uninjured DRG neurons of

adult rats, indicating that NGF participates in the maintenance of steady-state SNS/PN3 levels in DRG neurons in the adult nervous system in vivo. Changes in o-111 and SNS/PNS sodium channel expression in DRG neurons following axonal transection within the sciatic nerve thus may be, at least in part, a result of loss of access to peripheral pools of NGF. NGF is not, however, the only trophic factor that modulates the expression of sodium channels in DRG neurons. NaN expression in DRG neurons appears to be, at least in part, regulated by a pathway involving a different growth factor, glial-derived growth factor (GDNF) (Fjell et al., 1999b). NaN tends to be expressed in DRG neurons that bind with the isolectin IB4, a marker for cells that express the ‘ret’ and GFR receptors necessary to respond to GDNF. Studying an in vitro model of axotomy where NaN mRNA levels are reduced, Fjell et al. (1999b) showed that exposure to GDNF restores expression of NaN to levels at or above normal in IB4-positive DRG neurons (Fig. 10A). Moreover, administration of GDNF leads to a large up-regulation of TTX-resistant sodium current in these neurons (Fig. 10B). These findings indicate that GDNF modulates the expression of NaN. Consistent with this conclusion, intrathecal administration of GDNF ameliorates the reduction in conduction velocity of C-fibers that follows axotomy (Bennett et al., 1998). Brain-derived growth factor (BDNF) has been found not to alter sodium currents in DRG neurons, although it has significant effects on GABA receptor expression in these cells (Oyelese et al., 1997). As shown above, the effects of neurotrophins on sodium channel expression in DRG neurons appear to be specific (up-regulation of SNS/PN3, together with down-regulation of cl-111by NGF; up-regulation of NaN by GDNF). It is possible that neurotrophins may, in fact, have combinatorial effects on channel expression in cells that express multiple neurotrophin receptors. Abnormal sodium channel expression in inflammatory pain models Abnormal expression of sodium channels may also contribute to the pathophysiology of inflammatory pain. It is now clear that inflammatory molecules, e.g., prostaglandins and serotonin, can modulate

13

12345678

M

800 600 400

0

Ringer

NGF

4.0 i5 3.0 9% 2.0 2 - 1.0

A

0.0 IO-12dpa

Control

Axot + NGF

Axot + Ringer’s

piJ\

E17, -120 -80 -40 0 test potential

!!I\ -120

-so 40

0

-120

-80

-40 0

(mV)

Fig. 9. Delivery of NGF to the proximal nerve stump results in partial rescue of SNS mRNA and TTX-resistant sodium currents in axotomized DRG neurons following (A) RT-PCR showing co-amplification of SNS (979 bp) and GAPDH (666 bp) products in Ringer-treated axotomized DRG (Lanes 1, 2, 5, 6) and NGF-treated axotomized DRG (lanes 3, 4, 7 and 8). Bar graph shows the magnitude of the increase in SNS amplification product in NGF-treated DRG. (B) In situ hybridization showing down-regulation of SNS mRNA in DRG following axotomy (axotomy + Ringers), and the partial rescue of SNS mRNA by NGF. (C, D, E) Representative patch clamp recordings demonstrate partial rescue of slowly-inactivating TTX-resistant sodium currents in axotomized DRG neurons following exposure to NGF. Corresponding steady-state inactivation curves are shown below the recordings. Modified from Dib-Hajj et al. (1998a).

TTX-resistant sodium currents in DRG neurons (Gold et al., 1996), possibly acting through a cyclic AMP-protein kinase A cascade (England et al., 1996b; see also Malmberg, 2000, this volume). In addition, there are changes in sodium channel gene

expression in inflammatory models of pain. On the basis of our previous observations of peak changes in SNS/PN3 mRNA in DRG neurons five days following axotomy (Dib-Hajj et al., 1996), we examined the expression of sodium channels in DRG neurons

B 164+

NaN

1 DIV

Control

NGF

GDNF +GN”G”F’

t OJd

1 DIV

Control

NGF

184

D Control

GDNF

+GDNF/ NGF

GDNF +G$‘;

I

I 7 DIV

Fig. 10. Exposure to GDNF increases NaN mRNA levels, and TTX-resistant sodium currents, to near-normal levels in axotomized IB4+ DRG neurons. (A) NaN mRNA levels in IB4+ neurons are decreased 7 days following dissociation and culturing (7 DIV). NGF does not restore the level of NaN mRNA. However, GDNF has a strongly positive effect. (B) Micrographs showing NaN in situ hybridization in IB4+ (left column) and IB4- (right column) DRG neurons 7 days following dissociation and plating. NGF does not have a significant effect, compared to controls. GDNF, in contrast, produces a strong up-regulation of NaN in IB4+ cells. (C) TTX-resistant sodium current amplitude in DRG neurons, shown graphically to illustrate the effect of GDNF. (D) Patch clamp recordings showing TTX-resistant current traces from representative DRG neurons, 7 days following dissociation and culturing. Currents from control, NGF-treated, GDNF-treated, and GDNF+ NGF-treated neurons are shown. There is an up-regulation of TTX-resistant current in response to GDNF. Modified from Fjell et al. (1999b).

in rats four days following injection of the inflammatory agent carrageenan into the hind paw (Tanaka et al., 1998). SNS/PN3 mRNA expression was significantly increased in DRG neurons projecting to the inflamed limb, compared with DRG neurons from the contralateral side or naive (uninjected) controls (Fig. 11). Consistent with this, there was a significant increase in TTX-resistant sodium current amplitude in small DRG neurons projecting to the inflamed limb (31.7 f 3.3 nA) compared to the contralateral side (20.1 f 2.1 nA) four days post-injection (Tanaka

et al., 1998). The TTX-resistant current density was also significantly increased in the carrageenan-challenged DRG neurons. Other studies have demonstrated an increase in NaN mRNA seven days following injection of complete Freund’s adjuvant (Tate et al., 1998), and an increase in sodium channel immunoreactivity in DRG neurons following injection of complete Freund’s adjuvant into their projection field, which persists for at least 2 months (Gould et al., 1998). While details of the molecular pathway(s) underlying to

15

Carr (4d)

D

Contra

E

naive

F

Fig. 11. SNS mRNA levels and TTX-resistant sodium currents are increased within small DRG neurons 4 days following injection of the inflammatory agent carrageenan into the ion projection fields. Top: In situ hybridization showing SNS mRNA in carrageenar-injected (A), contralateral control (B), and naive (C) DRG. Patch clamp recordings (D, E, F) do not reveal any change in voltage- dependence of activation or steady-state inactivation of TTX-resistant sodium currents following carrageenan injection, but indicate that TTX-resistant current density and amplitude are increased (D). Modified from Tanaka et al. (1998).

these changes in sodium channel gene expression in inflammatory pain are not known, altered neurotrophin levels may be involved. Cell types that include fibroblasts, Schwamr cells, and keratinocytes normally produce NGF within peripheral target tissues; inflammatory agents stimulate NGF production in immune cells, and increased NGF concentrations have been observed within tissues exposed to inflammatory agents such as carrageenan and Freund’s adjuvant (Weskamp and Otten, 1978; Woolf et al., 1994). Sodium channels and the molecular pathophysiology of pain There seems to be little question that sodium channels contribute, in at least several ways, to the molecular pathophysiology of primary sensory neuron hyperexcitability. The molecular organization of axotomized DRG neurons shows profound changes, in-

cluding the down-regulation of some sodium channel subtypes, and the up-regulation of other, previously undetectable, sodium channel subtypes. As a result of these transcriptional changes, there are changes in the types of sodium channels that are deployed, and in the characteristics of the sodium currents that are produced in DRG neurons. The membranes of DRG neurons and their processes thus appear to be re-tuned after axonal injury, and this re-tuning may poise the cell to fire inappropriately. This may be clinically relevant because, in some cases, DRG neuron hyperexcitability can be associated with pain. Many important questions still require study. We need, for example, to learn more about the regulatory mechanisms that turn on, and off, the expression of various sodium channel genes; perhaps this process can be controlled. We also need to know whether secondary sensory neurons in the nociceptive pathway, located post-synaptic to primary sensory cells, exhibit altered expression of sodium channel genes

16

in chronic pain states. A growing body of evidence suggests that sodium channel expression of central neurons, even uninjured central neurons, is a dynamic process. For example, deafferentation of the olfactory bulb of developing rats, via transection of the olfactory nerve, results in a down-regulation of cl-11sodium channel mRNA expression of tufted and mitral cells (Sashihara et al., 1996). This change in sodium channel expression appears to be due to a change in the level of incoming synaptic activity since similar changes occur following deprivation of olfactory stimuli without denervation of the olfactory bulb (Sashihara et al., 1997). Within the hypothalamus of normal adult rats, supraoptic magnocellular neurons display changes in expression of sodium channel genes in response to physiological changes in the osmotic milieu; these transcriptional changes produce alterations in the deployment of specific sodium channels and in the levels of the sodium currents produced by these channels, which in turn alter the excitability of these cells (Tanaka et al., 1999). Given the dynamic nature of sodium channel expression in intact neurons where it has been studied, it would not be surprising if there were changes in sodium channel expression in uninjured neurons, located centrally along nociceptive pathways following nerve injury. The precise role(s) of each sodium channel subtype in the physiology of intact DRG neurons and in the pathophysiology of DRG neurons along pain-signaling pathways following nerve injury will hopefully soon be understood. Assessment of the effects of channel knock-out or selective blockade of each channel subtype will undoubtedly provide important information. The preferential expression of several sodium channel genes (SNS/PN3, NaN, PNl) in DRG and trigeminal neurons, at levels much higher than in other neuronal cell types, and the up-regulation of other sodium channel genes (type-III) in these cells following axonal injury, may provide an opportunity for selective pharmacological manipulation of sodium channels within primary sensory neurons in general, or nociceptive neurons in particular. The complexity of primary sensory neurons and the multiplicity of sodium channels within them present a challenge as we attempt to understand, at the molecular level, the basis for their behavior in normal and pathological states. It also, however,

presents an opportunity for the exploration of new therapeutic targets that may be helpful in the control of abnormal sensory phenomena including chronic pain. Acknowledgements Work in the authors’ laboratory has been supported in part by grants from the National Multiple Sclerosis Society and the Paralyzed Veterans of America/Eastern Paralyzed Veterans Association, and by the Medical Research Service and Rehabilitation Research Service, Department of Veterans Affairs. References Aguayo, L.G. and White, G. (1992) Effects of nerve growth factor on TTX- and capsaicin-sensitivity in adult rat sensory neurons. Brain Rex, 510: 61-67. Akopian, A.N., Sivilotti, L. and Wood, J.N. (1996) A tetrodotoxin-resistant voltage-gated sodium channel expressed by sensory neurons. Nature, 379: 257-262. Akopian, A.N., So&ova, V., England, S., Okuse, K., Ogata, N., Ure, J., Smith, A., Kerr, B.J., McMahon, S.B., Boyce, S., Hill, R., Stanfa, L.C., Dickenson, A.H. and Wood, J.N. (1999) The tetrodotoxin-resistant sodium channel SNS has a specialized function in pain pathways. Nat. Neurosci., 2: 541-548. Bennett, D.L., Michael, G.J., Ramachandran, N., Munson, J.B., Averill, S., Yan, Q.. McMahon, S.B. and Priestly, J.V. (1998) A distinct subgroup of small DRG cells express GDNF receptor components and GDNF is protective for these neurons after nerve injury. J. Neurosci., 18: 3059-3072. Black, J.A., Dib-Hajj, S., McNabola, K., Jeste, S., Rizzo, M.A., Kocsis, J.D. and Waxman, S.G. (1996) Spinal sensory neurons express multiple sodium channel a-subunit mRNAs. Mol. Brain

Res., 43: 117-132.

Black, J.A., Langworthy, K., Hinson, A.W., Dib-Hajj, S.D. and Waxman, S.G. (1997) NGF has opposing effects on Na+ channel III and SNS gene expression in spinal sensory neurons. NeuroRepoti,

8: 2331-2335.

Black, J.A., Cummins, T.R., Plumpton, C., Chen, Y., Clare, J. and Waxman, S.G. (1999) Upregulation of a previously silent sodium channel in axotomized DRG neurons. J. Neumphysiol., 82: 2776-2785.

Brock, J.A., McLachlan, E.M. and Belmonte, C. (1998) Tetrodotoxin-resistant impulses in single nociceptor nerve terminals in guinea-pig cornea. J. Physiol., 512: 211-217. Burchiel, K.J. (1988) Carbamazepine inhibits spontaneous activity in experimental neuromas. Exp. Neurol., 102: 249-253. Caffrey, J.M., Eng, D.L., Black, J.A., Waxman, S.G. and Kocsis, J.D. (1992) Three types of sodium channels in adult rat dorsal root ganglion neurons. Brain Res., 592: 283-297. Carlton, S.M., Dougherty, P.M., Pover, C.M. and Coggeshall,

17 R.E. (1991) Neuroma formation and numbers of axons in a rat model of experimental peripheral neuropathy. Neurosci Len., 131: 88-92. Chabal, C., Russell, L.C. and Burchiel, K.J. (1989) The effect of intravenous lidocaine, tocainide, and mexiletine on spontaneously active fibers originating in rat sciatic neuromas. Pain, 38: 333-338. Cummins, T.R. and Waxman, S.G. (1997) Downregulation of tetrodotoxin-resistant sodium currents and up-regulation of a rapidly repriming tetrodotoxin -sensitive sodium current in small spinal sensory neurons after nerve injury. J. Neurosci., 17: 3503-3514. Cummins, T.R., Howe, J.R. and Waxman, S.G. (1998) Slow closed-state inactivation: A novel mechanism underlying ramp currents in cells expressing the hNE/PNl sodium channel. J. Neurosci., 18: 9607-9619. Cummins, T.R., Dib-Hajj, S.D., Black, J.A., Akopian, A.N., Wood, J.N. and Waxman, S.G. (1999) A novel persistent tetrodotoxin-resistant sodium current in small primary sensory neurons, J. Neurosci., 19: RC43. Devor, M., Keller, C.H., Deerinck, T.J. and Ellisman, M.H. (1989) Na+ channel accumulation on axolemma of afferent endings in nerve end neuromas in Apteronotus. Neurosci. Lea., 102: 149-154. Devor, M., Wall, PD. and Catalan, N. (1992) Systemic lidocaine silences ectopic neuroma and DRG discharge without blocking nerve conduction. Pain, 48: 261-268. Dib-Hajj, S., Black, J.A., Felts, P and Waxman, S.G. (1996) Down-regulation of transcripts for Na channel o-SNS in spinal sensory neurons following axotomy. Proc. Null. Acad. Sci. U.S.A., 93: 14950-14954. Dib-Hajj, SD., Black, J.A., Cummins, T.R., Kenney, A.M., Kocsis, J.D. and Waxman, S.G. (1998a) Rescue of a-SNS sodium channel expression in small dorsal root ganglion neurons after axotomy by nerve growth factor in vivo. J. Neurophysiol., 79: 2668-2678. Dib-Hajj, S.D., Tyrrell, L., Black, J.A. and Waxman, S.G. (199813) NaN, a novel voltage-gated Na channel, is expressed preferentially in peripheral sensory neurons and down-regulated after axotomy. Proc. N&l. Acud. Sci. U.S.A., 95: 89638969. Dib-Hajj, S.D., Tyrrell, L., Escayg, A., Wood, P.M., Meisler, M.H. and Waxman, S.G. (1999a) Coding sequence, genomic organization, and conserved chromosomal localization of the mouse gene scnlla encoding the sodium channel NaN. Genomics,

59: 309-3

18.

Dib-Hajj, S.D., Fjell, J., Cummins, T.R., Zheng, Z., Fried, K., LaMotte, R., Black, J.A. and Waxman, S.G. (1999b) Plasticity of sodium channel expression in DRG neurons in the chronic constriction injury model of neuropathic pain. Pain, 83: 591600. Dib-Hajj, S.D., Tyrell, L., Cummins, T.R., Black, J.A., Wood, PM. and Waxman, S.G. (1999~) Two tetrodotoxin-resistant sodium channels in human dorsal root ganglion neurons. FEBS Lat., 462: 117-120. Eccles, J.C., Libet, B. and Young, R.R. (1958) The behavior of

chromatolysed motoneurons studied by intracellular recording. Lond., 143: 1l-40. Elliott, A.A. and Elliott, J.R. (1993) Characterization of TI’Xsensitive and ‘FIXresistant sodium currents in small cells from adult rat dorsal root ganglia. 1. Physiol. Land., 463: 3956. England, J.D., Gamboni, F., Ferguson, M.A. and Levinson, S.R. (1994) Sodium channels accumulate at the tips of injured axons. Muscle Nerve, 17: 593-598. England, J.D., Happel, L.T., Kline, D.G., Gamboni, F., Thouron, C.L., Lui, Z.P. and Levinson, S.R. (1996a) Sodium channel accumulation in humans with painful neuromas. Neurology, 47: 272-276. England, S., Bevan, S. and Docherty, R.J. (1996b) PGEz modulates the tetrodotoxin-resistant sodium current in neonatal rat dorsal root ganglion neurones via the cyclic AMP-protein kinase A cascade. J. Physiol. Land., 495: 429-440. Fjell, J., Cummins, T.R., Fried, K., Black, J.A. and Waxman, S.G. (1999a) In vivo NGF deprivation reduces SNS/PN3 expression and TTX-R sodium currents in IBCnegative DRG neurons. J. Neurophysiol., 81: 803-810. Fjell, J., Cummins, T.R., Dib-Hajj, S.D., Fried, K., Black, J.A. and Waxman, S.G. (1999b) Differential role of GDNF and NGF in the maintenance of two lTX-resistant sodium charnels in adult DRG neurons. Mol. Bruin Rex, 67: 267-282. Gold, MS., Reichling, D.B., Shuster, M.J. and Levine, J.D. (1996) Hyperalgesic agents increase a tetrodotoxin-resistant Na+ current in nociceptors. Proc. Nutl. Acad. Sci U.S.A., 93: 1108-1112. Gould III, H.J., England, J.D., Liu, Z.P. and Levinson, S.R. (1998) Rapid sodium channel augmentation in response to inflammation induced by complete Freund’s adjuvant. Bruin J. Physiol.

Res., 802: 69-74.

Gurtu, S. and Smith, PA. (1988) Electrophysiological characteristics of hamster dorsal root ganglion cells and their response to axotomy. J. Neumphysiol., 59: 408-423. Honmou, O., Utzschneider, D.A., Rizzo, M.A., Bowe, CM., Waxman, S.G. and Kocsis, J.D. (1994) Delayed depolarization and slow sodium currents in cutaneous afferents. L Neurophysiol., 71: 1627-1641. Jeftinija, S. (1994) The role of tetrodotoxin-resistant sodium channels of small primary afferent fibers. Bruin Rex, 639: 125-134. Jung, H.-Y., Mickus, T. and Spruston, N. (1997) Prolonged sodium channel inactivation contributes to dendritic action potential attenuation in hippocampal pyramidal neurons. J. Neurosci., 17: 6639-6646. Kapoor, R., Li, Y-G. and Smith, K.J. (1997) Slow sodiumdependent potential oscillations contribute to ectopic firing in mammalian demyelinated axons. Bruin Res., 120: 647-652. Kostyuk, PG., Veselovsky, N.S. and Tsyandryenko, A.Y. (1981) Ionic currents in the somatic membrane of rat dorsal root ganglion neurons. I. Sodium currents. Neuroscience, 6: 24232430. Kuno, M. and Llinas, R. (1970) Enhancement of synaptic transmission by dendritic potentials in chromatolysed motoneurons of the cat. J. Physiol. Lund., 210: 807-821.

18

Malmberg, A. (2000) Protein kinase subtypes involved in injury-induced nociception. In J. Sandktihler, B. Bromm and G.F. Gebhart (Eds.), Nervous System Plasticity and Chronic Pain, Progress in Brain Research, Vol. 129. Elsevier Amsterdam, pp. 5 l-59. Matzner, 0. and Devor, M. (1992) Na+ conductance and the threshold for repetitive neuronal firing. Bruin Res., 597: 9298. Matzner, 0. and Devor, M. (1994) Hyperexcitability at sites of nerve injury depends on voltage-sensitive Na+ channels. J. Neurophysiol., 12: 349-359. Nordin, M., Nystrom, B., Wallin, U. and Hagbarth, K.-E. (1984) Ectopic sensory discharges and paresthesiae in patients with disorders of peripheral nerves, dorsal roots and dorsal columns. Pain, 20: 231-245. Ochoa, J. and Torebjork, H.E. (1980) Paresthesiae from ectopic impulse generation in human sensory nerves. Brain, 103: 835854. Omana-Zapata, I., Khabbaz, M.A., Hunter, J.C. and Bley, K.R. (1997) QX-314 inhibits ectopic nerve activity associated with neuropathic pain. Brain Res., 771: 228-237. Oyelese, A.A., Rizzo, M.A., Waxman, S.G. and Kocsis, J.D. (1997) Differential effects of NGF and BDNF on axotomy-induced changes in GABAA-receptor-mediated conductance and sodium currents in cutaneous afferent neurons. J. Neurophysiol., 78: 31-42. Parri, H.R. and Cmnelli, V. (1998) Sodium current in rat and cat thalamocortical neurons: role of a non-inactivating component in tonic and burst firing. J. Neurosci., 18: 854-867. Pennartz, C.M.A., Bierlaagh, M.A. and Guersten, A.M.S. (1997) Cellular mechanisms underlying spontaneous firing in rat suprachiasmatic nucleus: involvement of a slowly inactivating component of sodium current. J. Neurophysiol., 78: 181 l1825. Quasthoff, S., Grosskreutz, J., Schroder, J.M., Schneider, U. and Grafe, P. (1995) Calcium potentials and tetrodotoxin-resistant sodium potentials in unmyelinated C fibres of biopsied human sural nerve. Neumscience, 69: 955-965. Rizzo, M.A. (1997) Successful treatment of painful traumatic mononeuropathy with carbamazepine: insights into a possible molecular pain mechanism. J. Nemo1 Sci., 152: 103-106. Rizzo, M.A., Kocsis, J.D. and Waxman, S.G. (1994) Slow sodium conductances of dorsal root ganglion neurons: Intraneuronal homogeneity and intemeuronal heterogeneity. J. Neurophysiol., 72: 2796-2816. Rizzo, M.A., Kocsis, J.D. and Waxman, S.G. (1995) Selective loss of slow and enhancement of fast Na+ currents in cutaneous afferent dorsal root ganglion neurones following axotomy. Neurobiol. Dis., 2: 87-96. Rizzo, M.A., Kocsis, J.D. and Waxman, S.G. (1996) Mechanisms of paresthesiae, dysesthesiae, and hyperesthesiae: Role of Na+ channel heterogeneity. EM,: Neural., 36: 3-12. Roy, M.L. and Narahashi, T. (1992) Differential properties of tetrodotoxin-sensitive and tetrodotoxin-resistant sodium channels in rat dorsal root ganglion neurons. J. Neurosci., 12: 2104-2111. Rush, A.M., Brau, M.E., Elliott, A.A. and Elliott, J.R. (1998)

Electrophysiological properties of sodium current subtypes in small cells from adult rat dorsal root ganglia. J. Physiot, 5 1 1: 771-789. Sangameswaran, L., Delgado, S.G., Fish, L.M., Koch, B.D., Jakeman, L.B., Stewart, G.R., Sze, I?, Hunter, J.C., Eglen, R.M. and Herman, R.C. (1996) Structure and function of a novel voltage-gated Tetrotodoxin-resistant sodium channel specific to sensory neurons. .I. Biol. Chem., 271: 5953-5956. Sashihara, S., Greer, CA., Oh, Y. and Waxman, S.G. (1996) Cell-specific differential expression of Na+-channel beta Isubunit mRNA in the olfactory system during postnatal development and after denervation. J. Neurosci., 16: 702-7 13. Sashihara, S., Waxman, S.G. and Greer, C.A. (1997) Down-regulation of Na+ channel mRNA following sensory deprivation of tufted cells in the neonatal rat olfactory bulb. NemoReport, 8: 1289-1293. Scadding, J.W. (1981) Development of ongoing activity, mechanosensitivity, and adrenalin sensitivity in severed peripheral nerve axons. Exp. Neurol., 73: 345-364. Scholz, A., Appel, N. and Vogel, W. (1998) Two types of TTXresistant and one ITX-sensitive Na+ channel in rat dorsal root ganglion neurons and their blockade by halothane. Eur: J. Neurosci., 10: 2547-2556. Schwindt, P.C. and Crill, W.E. (1995) Amplification of synaptic current by persistent sodium conductance in apical dendrite of neocortical neurons. J. Neurophysiol., 74: 2220-2224. Semagor, E., Yarom, Y. and Werman, R. (1986) Sodium dependent regenerative Responses in dendrites of axotomized motoneurons in the cat. Proc. Natl. Acad. Sci. U.S.A., 83: 966-7970. Stafstrom, C.E., Schwindt, P.C. and Crill, W.E. (1982) Negative slope conductance due to a persistent subthreshold sodium current in cat neocortical neurons in vitro. Brain Res., 236: 221-226. Stys, PK., Ransom, B.R. and Waxman, S.G. (1993) Non-inactivating, TTX-sensitive Na+ Conductance in rat optic nerve axons. Proc. Natl. Acad. Sci. U.S.A., 90: 6976-6980. Tanaka, M., Cummins, T.R., Ishikawa, K., Dib-Hajj, S.D., Black, J.A. and Waxman, S.G. (1998) SNS Na+ channel expression increases in dorsal root ganglion neurons in the carrageenan inflammatory pain model. NeuroReport, 9: 967-972. Tanaka, M., Cummins, T.R., Ishikawa, K., Black, J.A., Ibata, Y. and Waxman, S.G. (1999) Molecular and functional remodeling of electrogenic membrane of hypothalamic neurons in response to changes in their input. Proc. Natl. Acad. Sci. U.S.A., 96: 1088-1093. Tate, S., Benn, S., Hick, C., Trezise, D., John, V., Mannion, R.J., Costigan, M., Plumpton, C., Grose, D., Gladwell, Z., Kendall, G., Dale, K., Bountra, C. and Woolf, C.J. (1998) Two sodium channels contribute to the TTX-R sodium current in primary sensory neurons. Nat Neurosci., 1: 653-655. Titmus, M.J. and Faber, D.S. (1986) Altered excitability of goldfish mauthner cell following axotomy. II. localization and ionic basis. J. Neurophysiol., 55: 1440-1454. Toledo-Aral, J.J., Moss, B.L., He, Z.-J., Koszowski, A.G., Wbisenand, T., Levinson, S.R., Wolff, J.J., Silos-Santiago, I., Halegoua, S. and Mandel, G. (1997) Identification of PNl,

19 a predominant voltage-dependent sodium channel expressed principally in peripheral neurons. Proc. Natl. Acad. Sci. U.S.A., 94: 1527-1532.

Wall, RD. and Devon M. (1981) The effect of peripheral nerve injury on dorsal root potentials and on transmission of afferent signals into the spinal cord. Brain Rex, 209: 95-111. Waxman, S.G. and Brill, M.H. (1978) Conduction through demyelinated plaques in multiple sclerosis: computer simulations of facilitation by short internodes. J. Neurol. Neurosurg. Psychiatry. 41: 408417. Waxman, S.G., Kocsis, J.K. and Black, J.A. (1994) Type III sodium channel mRNA is expressed in embryonic but not adult spinal sensory neurons, and is reexpressed following axotomy. J. Neurophysiol., 72: 466-471. Woolf, C.J., Safieh-Garabedian, B., Ma, Q.-l?, Crilly, F?and Win-

ters, J. (1994) Nerve growth factor contributes to the generation of inflammatory sensory hypersensitivity. Neuroscience, 62: 327-331.

Weskamp, G. and Otten, U. (1978) An enzyme-linked immunoassay for nerve growth factor (NGF) A tool for studying regulatory mechanisms involved in NGF production in brain and in peripheral tissues. L Neurochenz., 48: 1779-1786. Zhang, J.-M., Donnelly, D.F., Song, X.-J. and LaMotte, R.H. (1997) Axotomy increases the excitability of dorsal root ganglion cells with unmyelinated axons. J. Neurophysiol., 78: 2790-2794.

Zur, K.B., Oh, Y., Waxman, S.G. and Black, J.A. (1995) Different up-regulation of sodium channel o- and 8-subunit mRNAs in cultured embryonic DRG neurons following exposure to NGF. Mol. Brain Rex, 30: 97-103.

J. Sandkiihler, B. Bromm and GE Gebhart (Ed%) Progress in Brain Research, Vol. 129 0 2000 Else&-r Science B.V. All rights reserved

CHAPTER 2

Chemical mediators of pain due to tissue damage and ischemia StephaniP Sutherland*, SeanP. Cook and Edwin W. McCleskey Vellum

Institute,

Oregon

Health

Sciences

Universiv,

3181 SW Sam Jackson

Park Rd., Portland,

OR 97201,

USA

Introduction

Rapid detection of tissue damage

Physical and chemical stimuli are transduced into action potentials at the peripheral endings of sensory neurons. Our understanding of the molecular mechanisms for transduction of somatic sensations is not as advanced as that for vision, audition and olfaction. However, there has been important progress on several fronts, notably the sensation of noxious heat (see Reeh and Petho, 2000, this volume). The present chapter discusses transduction of two other painful stimuli: tissue damage and cardTHiac ischemia. We review evidence suggesting that the initial sensation of tissue damage is mediated by extracellular adenosine triphosphate (ATP) acting on an ATP-gated ion channel called the P2X receptor. We also review evidence that ischemic heart pain - chronic angina and the pain of heart attack - are triggered when protons open an acid-sensing ion channel (ASIC). Both hypotheses should be considered debatable at this point. The hypotheses rely heavily on evidence of high expression of P2X and ASIC receptors in cells specialized to detect tissue damage and cardiac ischemia, respectively. However, increasing evidence from whole animal studies lends support to these ideas.

To describe the mechanisms of pain caused by tissue injury one must first identify active compounds that are released near peripheral sensory endings. The list of potential active compounds is long. Therefore, we have limited the scope of this section by focusing our discussion on compounds believed to act immediately in response to tissue injury and not on those involved in pain from inflammation or other chronic pain states (Levine and Taiwo, 1994; Millan, 1999). To put it more simply, which compounds act quickly enough to account for the initial pain felt by stepping on a tack? Clearly, such compounds should be immediately available for rapid release. One idea is that noxious injury disrupts the plasma membrane of cells near nociceptors and releases excitatory compounds that act directly on sensory endings. This requires that the compounds be at low external concentrations relative to the internal concentrations of a damaged cell. Furthermore, the compounds must have targets on the nerve ending at sufficient density to increase action potential firing. What compounds, then, meet these requirements?

* Corresponding author: S.P.Sutherland, Vollum Institute, Oregon Health SciencesUniversity, 3 181 SW Sam Jackson Park Rd., Portland, OR 97201, USA. Tel.: +l (503) 494-4321; Fax: +1 (503) 494-6972; E-mail: [email protected]

Immediate pain due to K+ Probably the most common agent implicated in the immediate response to tissue damage is potassium. Potassium levels within cytosol are between 100 and 150 mM, whereas plasma concentration is 5 r&l. Elevated extracellular potassium caused by its release from cytosol depolarizes neurons by chang-

22

ing the K+ electrochemical driving force for current through potassium channels open at the resting potential. Because potassium channels provide the major resting permeability in all sensory neurons, an increase in potassium will excite neurons that mediate non-painful sensations as well as nociceptors. Regardless of its non-selective actions, when potassium is injected into the skin, it causes pain (Keele and Armstrong, 1964). It is unknown whether tissue injury can increase extracellular K+ to the concentrations necessary to elicit pain. At least 100 mM Kf is needed to produce even moderate pain when injected into human skin (Lindahl, 1961). The weak potency of K+ requires that cytosolic potassium not be diluted more than IO-fold before it loses its effectiveness and should severely restrict the actions of K+ to an area immediately around the damaged cell. As a direct test of the potassium hypothesis, Bleehen et al. (1976) applied human erythrocyte lysates to a skin blister. The lysates represent a soluble fraction of cell cytosol and presumably mimic compounds, including K+, that are released by cell damage. As expected, the lysates caused pain, but the pain was greater than that produced by an equivalent amount of K+ . This was the first indication that K+ was not the only component in erythrocyte lysates that can cause pain. Immediate pain due to adenosine triphosphate (ATP)

To identify other algogenic components within erythrocyte lysates, Keele and Armstrong (1964) switched from human to cat erythrocytes. Cat erythrocytes contain K+ at plasma levels. These lysates still produced pain. After fractionation and biochemical analysis they identified three compounds - ATP, ADP, and AMP - that were present at similar concentrations in lysates and caused pain when injected into blisters (Fig. 1) (Bleehen et al., 1976). In erythrocytes, as in most tissues, ATP concentration is 0.8-I mM. ADP and AMP, on the other hand, are present at lo- and IOO-fold lower concentrations. They recognized that the unusually high concentrations of ADP and AMP in the fractionated lysates resulted from ATP degradation during dialysis and fractionation. Because all three adenine nucleotides had similar actions in their pain assay and ATP was by far the most abundant nucleotide in fresh tis-

sue preparations, they concluded that ATP was the lysate’s major pain-producing substance. A target for ATP

In the early eighties, ATP was shown to open nonspecific cation channels (P2X receptors) on sensory neurons, thus forming a direct connection between ATP release during tissue damage and excitation of sensory neurons (Jahr and Jessell, 1983; Krishtal et al., 1983). The work from several laboratories showed that >90% of sensory neurons express P2X receptors (Chen et al., 1995; Robertson et al., 1996; Cook et al., 1997). This current averages several nA in neurons dissociated from various sensory ganglia and, along with the acid-induced current and vanilloid-induced current, forms the largest current for any algogen in these tissues (Fig. 2) (Benson et al., 1999). A comparison of the kinetic and pharmacological properties of sensory neuron currents to those of the cloned P2X receptors gives clues to the identity of P2X receptors that serve to detect tissue damage. At least seven P2X receptor subtypes have been cloned to date (North and Barnard, 1997). These receptors share little sequence identity with any other receptor family, but they have a transmembrane topology similar to the channel family that includes the degenerins (DEG), acid-sensing ion channels (ASICs) and epithelial sodium channels (ENaCs) (North, 1996). P2X receptors have two transmembrane regions, short amino- and carboxy-termini that are presumed to be intracellular, and an extracellular domain of approximately 200 amino acids. P2X receptors are multimers formed from single or multiple types of P2X subunits. When expressed in heterologous systems, P2X currents can be divided into two broad classes based on kinetics of desensitization. The properties of the cloned channels match those found in native tissues (Fig. 3). The P2Xi and P2X3 subtypes desensitize to ATP within several hundred milliseconds. All other P2X receptors produce currents that desensitize over loo-fold more slowly (MacKenzie et al., 1999). Sensory-specific P2X receptors

Sensory neurons express both slowly and rapidly desensitizing kinetic classes. This suggests that they

23

Fig. 1. Algogenic action on the human blister base preparation of ATP, ADP, AMP, adenosine and ACh. Applications of agents in blister Ringer (concentrations in FM) are indicated by bars. The tick marks are at 30-s intervals. Pain intensity from 0 (no pain) to 3 (severe pain). From Bleehen et al. (1976).

A=

pH 5.0

Cap

8 ‘TL cu QL tr 4 FL -

2 set

B.

ATP

5HT

-

-

* 2 set

-

---

* 2 set

- 10 set

100 ii 80 7J 5 60 3 40 Cm

”

$g 10 3;f3 6 g4

pH

ATP

Ach I 0 0

BK

Aden

DRG heart DRG unlabeled Nodose heart

5 2 is 0 ATP 5HT Cap Ach BK Aden PH l! Fig. 2. Currents evoked by the indicated agent in a representative cardiac sympathetic afferent (A). Adenosine and bradykinin (BK) evoke only small currents (400 PA) in fewer than 50% of dissociated DRG neurons and labeled cardiac sympathetic afferents (B,C). Application of pH 5.0 evokes large (>8 nA) currents in over 90% of cultured cardiac sympathetic afferents. Capsaicin evokes very small currents in a small minority of cardiac afferents. Open bars describe randomly selected DRG neurons; black bars describe cardiac afferents with cell bodies in DRG (cardiac sympathetic afferents); gray bars describe cardiac afferents with cell bodies in the nodose ganglia (cardiac parasympathetic afferents). From Benson et al. (1999).

24

cloned receptor

native receptor

P2Xl ATP

a&meATP

P2X2 ATP

a@meATP

muscle

PC12

-ATP

V

uPa3

ATP

ATP

@-meATP

f’%Y3

DRG neuron

a&eATP

flodose

I/-

IJ

neuronv-

ATP

akmeATP

ATP

V

gland

ATP

ATp

apmeATP

-I

0.5 nA

4s

up-meATP

aj3-meATP

V

Fig. 3. Two distinct classes of desensitization kinetics are seen for cloned and native P2X receptors. Inward currents in response to brief applications of ATP or cc&methylene ATP (as indicated by bars above traces) recorded from HEK293 cells transfected with P2X1, P2X2, P2X3, P2x4, or P2X2,3 receptors (left-hand traces) or from native tissues (right-hand traces). From MacKenzie et al. (1999).

contain at least two different P2X isoforms. The relative proportion of the two classes depends on the type of sensory ganglion studied. In rat DRG, most P2X current desensitizes rapidly, as expected for P2Xi or P2Xs receptors (Robertson et al., 1996). In rat nodose ganglia, whose neurons innervate visceral

tissues, P2X current desensitizes slowly (Khakh et al., 1995). Rat trigeminal sensory neurons have both rapidly and slowly desensitizing currents (Cook et al., 1997). The physiological significance of these two kinetic classes and their differential distribution within various sensory ganglia is unknown. It is

25

P2X3 in Tooth Pulp

b

30 pM ATP +/- prepuff

30 PM agme

1 pMATP +I- suramin -c

Fig. 4. (a) Antisera generated against PZXs receptor labeled the peripheral endings of nociceptors in tooth pulp. (b-d) Pharmacological comparison of ATP-activated currents in nociceptors and muscle-stretch receptors. Nociceptors express two types of currents that are consistent with the expression of P2Xs homomers (desensitizing) and P2X 2,s heteromers (non-desensitizing). Stretch receptors express a different subtype of P2X (probably homomeric P2Xs or P2X2). (b) Desensitizing current in nociceptors peaked and declined during ATP application (30 FM, 0.3 s) and was greatly diminished by prior (15 s) ATP (prepuff, asterisk). This current was activated by a,@-methylene ATP (30 CM), and blocked by suramin (asterisk), a P2X antagonist (10 FM suramin, 1 J.LMATP). (c,d) Non-desensitizing nociceptor current of nociceptors was sensitive to a$-methylene ATP (c), while stretch receptor current was not (d). Scale bar in a: 25 urn. From Cook et al. (1997).

26

possible that both classes are important for detecting tissue damage since both are seen in a preparation of nociceptors from the rat trigeminal ganglia (Fig. 4). P2X.Y

An important clue to the identity of the P2X receptor in pain-sensing neurons came when the P2X3 subtype was cloned. The messenger RNA for P2X3 was found exclusively in sensory neurons and their projections in the spinal cord (Chen et al., 1995; Lewis et al., 1995). The absence of P2X3 message elsewhere in the body suggested that the P2Xs receptor is specialized for sensory transduction. In the dorsal root ganglia, small diameter sensory neurons had the highest levels of P2X3 message. This subclass is enriched in unmyelinated C fibers, many of which are nociceptors. Treatment of neonatal rats with the algogen capsaicin results in a loss of small diameter neurons and, as expected, a loss of P2X3 message (Chen et al., 1995). Further evidence for the specialization of P2X3 in detection of tissue damage came when antibodies were used to find P2X3 protein in nociceptive sensory endings within the tooth pulp (Fig. 4a), and large ATP-evoked currents were seen in nociceptors in primary culture (Cook et al., 1997). When expressed in heterologous cell systems, homomerit P2X3 forms a channel with pharmacology and kinetics identical to the rapidly desensitizing current of nociceptors (Fig. 4b) (Cook et al., 1997). Which P2X receptor forms the slowly desensitizing current in sensory neurons? A puzzle emerged when it was found that the slowly desensitizing P2X current found in nodose ganglia was sensitive to a$-methylene ATP, an agonist at the rapidly desensitizing P2X3 and P2Xr subtypes (Lewis et al., 1995). This pharmacology did not match any of the cloned P2X receptors when subtypes are expressed individually; all the cl&methylene ATP-sensitive channels have rapidly desensitizing kinetics. However, when P2X3 and P2X2 receptors are expressed together, they produce channels with pharmacology and kinetics that matched the nodose current (Lewis et al., 1995). This suggests that P2X2,3 heteromers form in vivo. Heteromeric P2X2,3 receptors are also likely in tooth pulp nociceptors where the slowly desensitizing P2X current also is sensitive to c&methylene ATP (Fig. 4c) (Cook et al., 1997).

Selectivity in nociceptors While P2Xs receptors, either as homomers or heteromers with P2X2, form the major P2X current in nociceptors, they do not mediate current from neurons that sense muscle stretch. Muscle spindle afferents have P2X current that is smaller on average than that found in nociceptors and is insensitive to cl&methylene ATP (Fig. 4c) (Cook et al., 1997). The large quantity of P2X current in nociceptors and restricted expression of P2X3 to nociceptors, but not muscle spindle afferents, suggest that ATP plays an important role in the initiation of pain. Whole animal studies With the use of pharmacological probes for P2X receptors, the role of ATP in producing pain has been recently examined in intact animal models. When the P2X3- and P2Xi -selective agonist a$-methylene ATP was injected into the hind paw of the rat, there was an increase in pain behaviors that was similar in magnitude and duration to that produced by the inflammatory mediator bradykinin (Fig. 5) (BlandWard and Humphrey, 1997). The effectiveness of c&methylene ATP in eliciting pain behaviors is consistent with activation of P2X3 receptors. ATP was far less effective than c&methylene ATP. Although it is no more potent than ATP in activating P2X3 of cultured neurons, c&methylene ATP is resistant to breakdown by ectonucleotidases, a common activity in intact tissue. The increased potency of non-hydrolysable ATP analogues suggests that ATP breakdown may provide an important mechanism for limiting the initial pain of tissue damage. Pain mediated by released ATP may be heightened during inflammation. When the rat hind paw is inflamed by various agents (e.g., UV light, carrageenan, or prostaglandin), the nociceptive response and sensitivity to a&methylene ATP increases dramatically (Hamilton et al., 1999). An increased sensitivity to ATP after inflammation may also occur in humans (Bleehen and Keele, 1977; Coutts et al., 198 1). The mechanism whereby inflammatory mediators increase the sensitivity to ATP is not known. Possibly P2X receptors are upregulated in response to inflammation. Alternatively, decreased ATP degradation during inflammation may increase the sensitiv-

27

l

a&Methylene

ATP

0 Bradykinin

A a&Methylene

ATP

0 0

20

10 Time

after

subplantar

injection

30 (min)

0.1

1

10 Dose

100

1000

(nmol)

Fig. 5. Activation of P2X receptors causes pain behaviors in whole animals. (a) Time course of hindpaw lifting behavior in rats injected subplantar with either a$-methylene ATP (200 nmol), bradykinin (10 nmol), or formalin (O.S%, 0.1 ml). Extent and duration of the pain behavior was similar for bradykinin or the P2X agonist, cr,b-methylene ATP. (b) The effect of various subplantar doses of bradykinin, cl&methylene ATP, and ATP on hindpaw lifting behavior. The ectonucleotidase-resistant ATP analogue a$-methylene ATP is > loo-fold more potent than ATP. This analogue is a potent agonist of the P2Xs receptor subtype. From Bland-Ward and Humphrey (1997).

ity to exogenously applied ATP Consistent with this latter possibility, ATP degradation is slowed in the synovial fluid of arthritic joints (Park et al., 1996). Adenosine: the fate of extracellular ATP A ubiquitous ectonucleotidase action on extracellular ATP quickly produces adenosine that may further propagate the pain of tissue injury. Like ATP, adenosine causes pain in the human blister base (Bleehen and Keele, 1977). However, the response to adenosine can be distinguished from that of ATP. When adenosine or ATP was injected arterially, close to the rat knee joint, primary afferents were excited (Dowd et al., 1998). This adenosine response took 10 s to develop and, consequently, would not be expected to mediate sharp pain to tissue injury. Ln contrast, the response to ATP occurred in two kinetic phases. The first phase of excitation was immediate and was followed by a second delayed phase, which could be due to breakdown of ATP to adenosine. An indirect action of adenosine on nociceptor excitation could account for the delayed primary afferent excitation, and could explain the lack of significant current evoked by adenosine in cultured primary sensory (see Fig. 2).

While it is likely that the action of ATP on nociceptor P2X receptors mediates pain caused by ATP, more convincing evidence is needed. A block of P2Xs activity in whole-animal models of tissue damage, either by genetic or pharmacological methods, has not yet been demonstrated. Unfortunately, such approaches have been hindered by a lack of specific and potent P2X antagonists for use in whole tissue. Although recently P2X antagonists have had some success in antagonizing ATP-mediated hyperalgesia when administered centrally (Tsuda et al., 1999) and in blocking the ATP-evoked firing of mesenteric afferent nerves (Kirkup et al., 1999), they have not yet been used to block pain caused by tissue injury. Immediate pain due to other compounds Several compounds, such as serotonin (5HT) bradykinin, prostaglandin, and histamine, cause pain when injected into the skin (Keele and Armstrong, 1964) (many stinging insects take advantage of this). These compounds are present in normal somatic tissue only at a low concentration. Therefore, these compounds must first be synthesized or cells that contain them be recruited before they can accumulate at sites of tissue damage - a process that is

28

too slow to mediate the acute pain accompanying injury. For example, bradykinin is made from circulating plasma kininogen at sites of injury by the actions of activated kallikreins. Serotonin is released by activated platelets and mast cells. For this reason, the role of these compounds in nociception is

ciceptive reflexes in sensory afferents that innervate the rat tail (Ault and Hildebrand, 1993). The nonNMDA glutamate receptor agonist kainate evokes currents in 63% of freshly dissociated sensory neurons from rat DRG (Huettner, 1990). Kainate-evoked currents average around 100 pA and are > IO-fold

probablymost importantduring inflammationand

smallerthanthe averagecurrentsinducedby ATP or

in chronic pain states (Wood and Docherty, 1997; Millan, 1999). However, there are other compounds that may act in a more acute manner.

In some tissues, acetylcholine (ACh) might play a role in the initial response to tissue injury. ACh causes pain when injected into blistered skin, although relatively high concentrations are needed

protons. An enhancement of pain from released glutamate may occur by several mechanisms. In sensory neurons, non-NMDA glutamate receptor channels may be enriched in central rather than peripheral processes in vivo where they may serve a presynaptic modulatory role (Huettner, 1990). Thus the direct gating of sensory neuron channels by glutamate may not occur at sites of peripheral injury. In the periphery, glutamate receptors may contribute to inflammatory pain. Glutamate concentrations increase

(KeeleandArmstrong,1964).ACh concentrationis

in theplantarof rat hindpawsafterformalin-induced

high in cornea1 epithelium, but its packaging and potential for release during injury are unknown (Pesin and Candia, 1982). Nicotinic ACh-receptor agonists activate a subset of sensory afferent fibers that innervate cornea1 epithelium (Tanelian, 1991). However, direct activation of nicotinic channels on dissociated

inflammation (Omote et al., 1998), and glutamate receptor antagonists lessen formalin-induced pain behavior (Davidson et al., 1997). Glutamate receptor agonists promote mechanical allodynia and hyperalgesia when injected into the rat hindpaw (Zhouet al., 1996). These studies are consistent with the involve-

sensoryneuronshas not beendemonstratedby all

ment of peripheralglutamatereceptorsin pain, but

laboratories (Fig. 2). Where they have been detected, nicotinic ACh-receptor agonists activate small inward currents in 23-51% of sensory neurons de-

also suggest that peripheral glutamate contributes more to chronic inflammatory pain rather than to pain from acute injury.

Acetylcholine

pendingon the type of ganglia studied.In DRG, the current ranges between -5 and -40 pA and is therefore approximately loo-fold smaller than the average current evoked by ATP (Sucher et al., 1990). Sensitivity is mainly restricted to neurons with larger cell diameters, and is not enriched in small-diameter capsaicin-sensitive neurons (Liu et al., 1993; Liu and

Simon, 1996).The presenceof ACh sensitivityin this cell population may indicate that ACh sensitivity is enriched in myelinated neurons that mediate sharp pain. However, this cell population is also enriched in neurons that mediate non-nociceptive sensations. Glutamate

Glutamate may also mediate pain to tissue injury (Wood and Docherty, 1997). Glutamate concentration within the cytosol of somatic tissue can be several mM, so tissue damage could result in high levels of extracellular glutamate. Glutamate produces no-

Chemical mediators of ischemic pain In addition to sensing pain caused by tissue damage, chemosensitive neurons detect ischemia. Ischemia, from the Greek ‘suppression of blood’, is defined as a “deficiency of blood in a part, due to functional constriction or actual obstruction of a blood vessel” (Taylor, 1988). The condition includes hypoxia, an insufficiency in the oxygen supply relative to demand, and also includes a reduction in the nutrients carried by the plasma. When ischemia occurs chronically in the heart, it produces pain called angina pectoris; acute ischemic pain is experienced during myocardial infarction (heart attack). While many of the earliest experiments aimed at understanding ischemia were performed in skeletal muscle, the following discussion will focus on myocardial ischemia. Ischemic pain in skeletal muscle and in myocardium are thought to be mediated by

29 the same process, as Lewis (1932) suggested. The experiments presented here shape our current understanding of the processes involved in the generation of ischemic pain. The review by Keele and Armstrong (1964) provides a good overview of the following early investigations aimed at understanding the nature of ischemic pain. Occlusion of a blood vessel at rest does not produce pain; therefore, pain is dependent not only on hypoxia or arrested circulation, but also on continued muscle contraction, which results in an increase in oxygen demand. MacWilliam and Webster (1923) were the first to recognize this and attributed such pain to “want of oxygen and its consequences, with excessive accumulation of metabolic products, acids and other bodies”. They suggested that the pain was “protective in character, tending to limitation of effort and shielding the muscle from being spurred on to further and injurious activity”. Pickering and Wayne (1933) showed in anemic patients that ischemic pain depends on the oxygenation state of the blood and not the flow itself; thus they postulated that the algogenic substance that accumulates in tissue spaces can be removed by reoxygenation. Thus, while ischemia consists of more than simply hypoxia, the lack of oxygen plays the major part in producing pain rather than depletion of some other substance in blood. Finally, Lewis (1932) concluded that the algogenic substance arises from muscle contraction and acts directly at nerve endings in the muscle tissue. The search has since ensued for a metabolite that (1) is produced in contracting muscle (perhaps only in hypoxic conditions), (2) is stable enough to accumulate during ischemia, and (3) the effects of which would be alleviated by reoxygenation. Lewis termed the substance ‘factor P’, and stressed that factor P may be present before and after the sensation of pain, but that only during the ischemic condition does it accumulate to a level sufficient to produce pain (Lewis, 1932). While potassium ions were considered a possible mediator early on, the evidence does not point to a major specific role of K+ in mediating ischemic pain. Three more likely candidates considered in this review are adenosine, lactic acid, and bradykinin. Since this discussion will focus on myocardial ischemia, it will be helpful at this time to review the sensory innervation of the heart.

Sensory anatomy of heart attack and angina Since the time of these early studies, it has been recognized that a potential mediator of ischemic pain must stimulate the afferent pathway from the muscle to the CNS. The anatomy of cardiac innervation was largely determined through clinical experience (White, 1957). Three populations of afferent neurons innervate the heart, as illustrated in Fig. 6 (Benson et al., 1999). Neurons that innervate the pericardial membrane, which encapsulates the heart, have axons that follow the phrenic nerve to their cell bodies in the upper cervical dorsal root ganglia (Cs-Cs). There is no evidence that these pericardial nerve endings sense cardiac ischemia. Two populations of sensory neurons innervate the cardiac muscle (myocardium) itself: vagal afferents and sympathetic afferents. Vagal afferent axons follow the vagus nerve to their cell bodies located in the nodose ganglia. Sympathetic afferent axons follow the sympathetic nerve to their cell bodies in the upper thoracic dorsal root ganglia (C&T& Sympathetic afferents may be the primary mediators of myocardial pain because sympathectomy relieves pain in most patients suffering from chronic angina (Cutler, 1927). Moreover, Brown (1967) found that coronary occlusion causes pain in cats that is accompanied by afferent nerve activity in the sympathetic nerve tract. However, the vagal afferents should not be discounted as a source of ischemic sensation from the heart. In a comprehensive review of the experimental literature, Meller and Gebhart (1992) present an overview of cardiac afferent innervation. Sympathectomy is ineffective in relieving angina in some patients, and transection of vagal fibers has been found to relieve cardiac pain in some cases. Meller and Gebhart propose spatial differences in cardiac innervation by sympathetic and vagal afferents. Perhaps patients suffering from angina in an area innervated by vagal afferents were relieved of pain by vagotomy, whereas sympathectomy would have been ineffective. There certainly is evidence that vagal afferents transduce some form of cardiac pain, but the vast majority of literature on cardiac pain has studied the sympathetic afferents. Therefore the following discussion will only address pain mediated by cardiac sympathetic afferents. It is generally accepted that pain is the only conscious sensation arising from the heart, and that

-Pel*icardium

Fig. 6. Illustration of sensory innervation of the heart. Axons of afferent neurons that innervate the pericardium follow the phrenic nerve to their cell bodies in the upper cervical DRG (C&s). The myocardium is innervated by sympathetic afferents, the axons of which follow the sympathetic nerve to cell bodies in the upper thoracic DRG (Cs-Ts), and by vagal afferents, with axons that follow the vagus nerve to their cell bodies in the nodose ganglia. From Benson et al. (1999).

myocardial ischemia is the source of cardiac pain. Cervero (1994) has compiled an extensive review of literature examining the modalities of cardiac afferents and their possible functions. Cardiac sympathetic afferents consist of myelinated and unmyelinated fibers. The myelinated fibers display both mechanosensitive and chemosensitive behavior, and the unmyelinated fibers are predominantly chemosensitive, but may display various degrees of mechanosensitivity. Uchida and Murao (1975) found

that all A-6 and C fibers responded to coronary occlusion and displayed chemosensitivity regardless of mechanosensitivity. Mechanosensitive neurons fire rhythmically, but only once per cardiac cycle under normal conditions, so they are not well suited to provide information about the cycle. However, they respond robustly to coronary occlusion and to application of lactic acid (Uchida and Murao, 1975) or bradykinin (Baker et al., 1980) with a firing pattern that is dictated by the cardiac cycle. This suggests

31

that these neurons may provide information about the cycle under ischemic conditions by integrating chemical and mechanical signals. In contrast, the unmyelinated, exclusively chemosensitive neurons fired with an irregular burst pattern. Despite the fact that a significant portion of sympathetic afferents is mechanosensitive, the entire population appears able to sense ischemic conditions, and might all transmit pain sensation. Both the mechanically and chemically sensitive neurons may be important to elicit the sympathetic reflex response that arises with coronary occlusion or myocardial ischemia. Malliani et al. (1969) found that the sympathetic, rather than the vagal, afferents are instrumental in exciting the sympathetic efferent neurons, resulting in a cardiocardiac sympathetic reflex arc. This curious reflex causes the heart to contract more forcefully at a time when it is already oxygen-deficient. Clearly counterproductive, this can cause an increase in myocardial injury in the setting of a heart attack or chronic ischemia (Leenan, 1999). Thus it appears that the cardiac sympathetic afferents respond to myocardial ischemia with two consequences: the sensation of pain, and the sympathetic reflex response. Both of these functions are probably mediated by some of the following chemicals produced during ischemia. Adenosine Adenosine has been considered a candidate for a chemical mediator of ischemic pain for many years, but the evidence for and against a major role of adenosine are in conflict. Intracoronary administration of adenosine to patients with angina caused chest discomfort that was similar to their typical angina1 pain (Crea et al., 1990). Thames et al. (1993) found that cardiac sympathetic afferent firing in response to a coronary occlusion was inhibited by i.v. administration of the adenosine receptor antagonist aminophylline, and the firing was augmented by dipyradimole, an adenosine uptake inhibitor. Gnecchi-Ruscone et al. (1995) found that adenosine caused sympathetic afferent firing in a dose-dependent manner that was blocked by administration of aminophylline; however, afferent firing in response to coronary occlusion was not reduced by aminophylline, indicating that adenosine is not the

sole mediator of cardiac pain. These studies support a role for adenosine in cardiac sensation. Other studies, however, do not. In a recent study by Pan and Longhurst (1995), intracardial injection or epicardial application of adenosine failed to cause cardiac afferents to fire. Application of the Al receptor agonist N6-cycloadenopentasine did not cause firing, and dipyradimole did not potentiate the response of ischemically sensitive afferents to 5 min of myocardial ischemia. Furthermore, aminophylline did not attenuate firing during ischemia. Therefore, the authors concluded that adenosine does not play a substantial role in activation of cardiac sympathetic afferents during myocardial ischemia. Abe et al. (1998) found that only one of 12 fibers tested was sensitive to epicardial application of adenosine, and this fiber did not fire in response to coronary occlusion. Others have shown that adenosine failed to cause a sympathetic reflex response when applied epicardially (Veelken et al., 1996) and failed to activate afferents in skeletal muscle (Rotto and Kaufman, 1988). The part played by adenosine in cardiac sensation remains unclear due to these conflicting results. Bradykinin Bradykinin (BK) has also been shown to play a role in activating afferent neurons in the heart. Baker et al. (1980) found with single unit recordings of cardiac afferent neurons that both mechanically and chemically sensitive neurons responded to intrapericardial application of BK. They examined many mechanosensitive neurons (140) and found that BK either directly stimulated the nerve endings or sensitized them to the mechanical stimulus of the beating heart. The small number (18) of purely chemosensitive neurons they tested all showed a robust response to BK. Abe et al. (1998) found that 13 of 16 mechanosensitive sympathetic afferents responded to epicardial application of BK. Stebbins and Longhurst (1986) examined the role of BK in eliciting the reflex cardiovascular response to contracting skeletal muscle. They showed with pharmacological manipulation of bradykinin metabolism that BK contributes to stimulation of this reflex response. Likewise, Veelken et al. (1996) found that epicardial application of BK induced a sympathoexcitatory response through B2

32

-5

4

-3

-2

-1

0

12

3

4

5

6

7

Time (min) Fig. 7. Frequency of action potentials recorded in vivo from an umyelinated cardiac afferent during control, ischaemia, and reperfusion before (A) and after (B) treatment with isotonic neutral phosphate buffer. From Pan et al. (1999).

receptors, which probably reside on cardiac afferent neurons. Thus it is clear that BK activates cardiac sympathetic afferents, plays a role in eliciting a sympathetic response to contracting muscle, and may evoke the sensation of pain. Lactic acid Lindahl (1974) suggested that protons are the sole mediator of nociception. This was an overestimate of the role of acid, but it was an early indicator of its importance in the nociceptive system, Ischemic tissue acidification results from generation of lactic acid. Lactate and ATP are both products of anaerobic glycolysis, and breakdown of ATP produces protons. Lactate and protons produced during myocardial ischemia rise first in tissue and later in blood (Opie et al., 1973; Cobbe et al., 1982). This suggests that the acid metabolites accumulated during ischemia are washed into the bloodstream by reperfusion. Lactic acid solutions stimulate sympathetic cardiac afferents, and prior application of H+ buffer blocks afferent firing in response to coronary occlusion (Fig. 7) (Uchida and Murao, 1975;

Pan et al., 1999). Uchida and Murao (1975) found that while the chemosensitive afferent firing was markedly inhibited by buffer during occlusion, the mechanosensitive afferents were only slightly (O25%) inhibited. This suggests that lactic acid may not be the critical stimulus for all types of cardiac afferents during ischemia. Benson et al. (1999) found that 93% of labeled, dissociated cardiac sympathetic afferents responded to a drop in pH with very large depolarizing currents (Fig. 2), while BK and adenosine evoked very small currents in a much smaller fraction of the cells. Interestingly, the effects of pH seen in cardiac afferents are somehow modulated by lactate. Pan et al. (1999) showed that cardiac afferents were stimulated better by lactic acid than by an acidic phosphate buffer solution at the same pH. Furthermore hypercapnia, which decreased epicardial pH to the same level as lactic acid, had no effect on afferent firing. Thimm et al. (1984) found that perfusion of the rat hindleg with lactic acid caused changes in heart rate that increased with lactate concentration. Hong et al. (1997) also found that while injection of lactic acid to the right atrium caused a decrease in arterial blood

33

pH and caused pulmonary C fibers to fire, HCI at the same pH had no effect on blood pH or afferent firing. Formic acid (with an identical pK,) caused a similar fall in blood pH, but the afferent response was not as robust. Although the lactate ion does not directly cause firing of sensory afferents, it may potentiate the response to protons. These experiments are consistent with the hypothesis that a decrease in pH due to production of lactic acid during anaerobic metabolism is the primary mechanism for activation of cardiac sympathetic afferents. Lactic acid is produced during ischemia, but is the level of acidosis reached in ischemic tissue sufficient to stimulate sensory nerve endings? While early measurements of pH in ischemic tissue were attempted, they were not accurate prior to 1978 (PooleWilson, 1978). Technological advances since then allow much faster and more accurate measurements of tissue pH. Cobbe and Poole-Wilson (1980) used a pH-sensitive needle electrode in the epicardium to measure changes in pH during myocardial ischemia. The accumulation of protons caused pH to fall by 0.15 units after 5 min of ischemia, and by 0.66 units after 15 min. Recently, Pan et al. (1999) measured epicardial pH using the needle electrode during myocardial ischemia and found that tissue pH fell from 7.35 to 6.98 within 5 min, and this decrease was prevented by neutral phosphate buffer in the pericardial space. Furthermore, they made single unit recordings from cardiac sympathetic C-fiber afferents and found that the afferent firing matched the fall in tissue pH and was also diminished by buffering (Fig. 7).

-ia) ;I-

10 s -10

4x10 A

M

Proton-sensing molecules

Krishtal and Pidoplichko (1980) were the first to propose a molecular sensor of pH at sensory neuron terminals. They found that isolated cultured sensory neurons of the rat responded to a rapid decrease in pH from 7.4 to 6.9 and lower with an inward, depolarizing current carried by sodium. They showed that 74% of neurons smaller than 26 urn were proton-sensitive, while 75% of neurons larger than 26 urn were proton-insensitive (Krishtal and Pidoplichko, 198 l), suggesting that this small-diameter, nociceptor-enriched population of sensory neurons is specialized to sense protons. They described three currents that desensitized at different rates (Fig. 8).

Fig. 8. Examples of the three different types of desensitization kinetics of acid-gated currents in sensory neurons, obtained from different trigeminal ganglia neurons. From Krishtal and Pidoplichko (1981).

The slowly activating, non-inactivating type of current has been the cause of much debate over sensory neuron proton-gated currents. Bevan et al. (1993) has contended that the slowly or non-inactivating current is present only in cells that display a current in response to capsaicin, and that this slow acid current is carried by the same ion channels that are opened by capsaicin. In contrast, Zeilhofer et

34

al. (1997) found that while the slow, non-inactivating acid current and the capsaicin-sensitive current were both carried somewhat by Ca*+ ions, the proportional contribution of Ca2+ to each current was significantly different, indicating two populations of channels. However, they did find that a drop in extracellular protons augmented the responses to capsaicin. Others have found currents in sensory neurons that could be activated by either capsaicin or protons without complete overlap in cell distribution (Steen et al., 1992; Liu and Simon, 1994). These reports indicate that sensory neurons express a diverse population of channels, some of which respond only to capsaicin OYprotons, and some that can respond to either ligand. The recently cloned vanilloid receptor VR- 1, also referred to as the capsaicin receptor, can integrate multiple stimuli (Caterina et al., 1997). VR-1 is opened by noxious heat and is augmented by protons. The channel can be opened in the absence of capsaicin by an external pH of 5.9 or lower at room temperature (Tominaga et al., 1998). These channels may exist in sub-populations of nociceptors specialized to detect noxious heat. The investigations of dissociated neurons discussed thus far have been undertaken in sensory neurons, many in a nociceptive-enriched population. However, the specific sensory modality or site of innervation of these neurons was not known. To examine cardiac sensory neurons and their responses to chemical mediators of ischemia, Benson et al. (1999) devised a method to label the sensory neurons innervating the myocardium in vivo for later identification in vitro. Surprisingly, these cardiac afferent neurons are remarkably insensitive to capsaicin as a population, and those that did respond to capsaicin had very small currents (~200 PA). Conversely, 93% responded to pH 5.0 with large currents (average 8.6 nA) (Fig. 2). This suggests that VR-1 is not the major proton sensor in cardiac afferent neurons. The acidsensitive current in sympathetic afferents resembles the fastest of the three kinetic currents described by Krishtal (Fig. 8), is activated by pH 7.0 (PHI,, 6.7), and is blocked by amiloride (an ASIC antagonist, see below) (Benson et al., 1999). Cardiac afferents display a biphasic current. In addition to the large, transient current evoked at pH 7 and below, there is a smaller sustained current

evoked at very low pH levels (~5.0). This current is the same as that described by Bevan and Yeats (1991). Its sustained nature seems important because cardiac pain is sustained. However, the extreme pH required for activation of this current is not achieved even after 60 min of total myocardial ischemia (Cobbe and Poole-Wilson, 1980), which would be a fatal insult. Thus, there are two parts of the ischemia-sensing current, neither of which is obviously suited to signal persistent pain. One possibility is that the sustained component is activated by less extreme pH levels in vivo than seen with electrophysiology. The discrepancy may be due to experimental conditions that differ during measurements of cardiac physiology and cellular electrophysiology: cellular experiments are performed at room temperature and in the absence of cardiac metabolites (other than low pH). Alternatively, the transient current might generate persistent firing because it does not fully desensitize at the pH achieved during myocardial ischemia (Benson et al., 1999). Cloned acid-sensing channels

In recent years, a group of channels has been cloned called ASKS, or acid-sensing ionic channels (Waldmann et al., 1999). These channels belong to a diverse family that includes epithelial sodium channels (ENaCs) and the degenerin channels (DEG). Members of this family share a membrane topology consisting of two transmembrane domains, short intracellular N- and C-termini, and a large extracellular loop. They all preferentially pass Na+ ions, are blocked by amiloride, and are insensitive to voltage. The family displays great diversity as well: some channels are constitutively open (ENaC), others are gated by ligands including protons (ASIC) or peptides (FMRF-amide gated channel, FaNaC), and still others become active after a mutation and may be mechanically sensitive (DEG). These channels are found in a variety of species; they include mammals as well as Drosophila, the snail Helix aspersa, and the nematode Caenorhabditis elegans (Waldmann and Lazdunski, 1998). Members of this family probably form tetramers of subunits, as the FaNaC channel does (Coscoy et al., 1998). Incidentally, while the capsaicin-sensitive channel VR-1 is sensitive to protons, ASICs are insensitive to cap-

35 saicin. ASICs and vanilloid receptors are members of entirely different families with distinct properties and membrane topologies. The chronology of the cloning of the protongated channels has resulted in some discrepancy in the nomenclature. Most of the proton-gated channels have been discovered by Lazdunski’s group, the first of which is now referred to as ASICla (Waldmann et al., 1997a). This channel is activated by a rapid decrease in pH to 6.9 or lower. The current displays single exponential activation and desensitization kinetics, which appear similar to Krishtal’s intermediate-kinetic current (Fig. 8). A splice variant, called ASIClb, has been cloned that displays the same pH-dependency and kinetics as ASICla. The current properties of the two channels differ in permeability to Ca2+ and inhibition by extracellular Ca2+ ions (Chen et al., 1998). While ASICla is expressed throughout the central nervous system and the dorsal root ganglia, ASICI b is absent from CNS neurons and is expressed in a subset of DRG neurons. Before ASICl, the mammalian homologue to the degenerin channels was cloned and named MDEG (Waldmann et al., 1996). Shortly after ASICla was cloned, MDEGl was found to be gated by protons, and a splice variant, MDEG2, was cloned (Lingueglia et al., 1997). These channels are now referred to as ASIC2a and ASIC2b. ASIC2a differs from ASICl in that it has slower kinetics and opens at a much lower pH, with a half-maximal activation at pH 4.35. ASIC2b does not form a functional channel when expressed as a homomer, but it can associate with other ASIC subunits to change their permeation properties. ASIC2a and ASIC2b are expressed in the same areas of CNS, but, of the two, only ASIC2b is present in sensory neurons (Lingueglia et al., 1997). ASIC3 may be the most important cloned proton-gated channel in terms of sensory physiology. ASIC3 mRNA is abundant in DRG, so this channel was initially called DRASIC (Waldmann et al., 1997b). ASIC3 is absent from brain and all other areas examined, suggesting that this channel is specialized to detect sensory signals. Like the acid currents reported in some sensory neurons (Bevan and Yeats, 1991; Benson et al., 1999), this current displays biphasic kinetics: a fast transient compo-

nent that is very sensitive to protons (pHi,, 6.5), followed by a smaller sustained component that is far less sensitive (pHtjz 3.5) (Waldmann et al., 1997b). However unlike the sensory neuron currents, in which the sustained component is a non-selective cation conductance, both the transient and sustained components of ASIC3 are Na+-selective. How could the ASIC3 channel achieve the differential selectivity that is seen in neurons, but not in the heterologous system? One possibility is that the channel undergoes some post-translational modification in neurons, such as phosphorylation. Another way that ASIC3 could account for the properties seen in native neurons is by heteromerization with another ASIC channel subunit. Although ASIC2b does not form a functional channel as a homomer, it can alter the properties of other ASIC channels when coexpressed as heteromers. In fact, ASIC3 and ASIC2b heteromers display the kinetic and selectivity characteristics seen in sensory neurons (Lingueglia et al., 1997). It is evident that members of the new ASIC family play an important part in sensory physiology. The diversity of currents seen in sensory neurons (Krishtal and Pidoplichko, 1981; Bevan and Yeats, 1991; Benson et al., 1999) are almost certainly carried by ASIC channels, a family that will perhaps continue to grow. Cardiac afferent neurons are particularly well suited to use an ASIC channel, an exquisitely sensitive proton sensor, as a way to detect myocardial ischemia and elicit cardiac pain. List of abbreviations BK ACh ATP ADP AMP DRG

bradykinin acetylcholine adenosine n-i-phosphate adenosine di-phosphate adenosine mono-phosphate dorsal root ganglion

References Abe, T., Morgan, D., Sengupta, J.N., Gebhart, G.F. and Gutterman, D.D. (1998) Attenuation of ischemia-induced activation of cardiac sympathetic afferents following brief myocardial ischemia in cats. L Auton. Nov. Syst., 71: 28-36. Auk, B. and Hildebrand, L.M. (1993) L-Glutamate activates peripheral nociceptors. Agems Actions, 39: C142-Cl44

36 Baker, D.G., Coleridge, H.M., Coleridge, J.C.G. and Nerdrum, T. (1980) Search for a cardiac nociceptor: stimulation by bradykinin of sympathetic afferent nerve endings in the heart of the cat. .I. Physiol., 306: 519-536. Benson, C.J., Eckert, S.P. and McCleskey, E.W. (1999) Acidevoked currents in cardiac sensory neurons: a possible mediator of myocardial ischemic sensation. Circ. Res., 84(8): 921928. Bevan, S., Forbes, C.A. and Winter, J. (1993) Protons and capsaicin activate the same ion channels in rat isolated dorsal root ganglion neurones. J. Physiol., 459: 4OlP. Bevan, S. and Yeats, J. (1991) Protons activate a cation conductance in a sub-population of rat dorsal root ganglion neurones. J. Physiol., 433: 145-161. Bland-Ward, PA. and Humphrey, P.P. (1997) Acute nociception mediated by hindpaw P2X receptor activation in the rat. BE J. Pharmacol., 122: 365-371. Bleehen, T. and Keele, CA. (1977) Observations on the algogenic actions of adenosine compounds on the human blister base preparation. Pain, 3: 367-317. Bleehen, T., Hobbiger, F. and Keele, C.A. (1976) Identification of algogenic substances in human erythrocytes. J. Physiol., 262: 131-149. Brown, A.M. (1967) Excitation of afferent cardiac sympathetic nerve fibers during myocardial ischaemia. J. Physiol., 190: 3.5-53. Caterina, M.J., Schumacher, M.A., Tominaga, M., Rosen, T., Levine, J. and Julius, D. (1997) The capsaicin receptor: a heat-activated ion channel in the pain pathway. Nature, 389: 818-824. Cervero, F. (1994) Sensory innervation of the viscera: peripheral basis of visceral pain. Physiol. Rev., 74: 95-138. Chen, CC., Akopian, A.N., Sivilotti, L., Colquhoun, D., Bumstock, G. and Wood, J.N. (1995) A P2X pminoceptor expressed by a subset of sensory neurons. Nature, 377: 42% 431. Chen, C.-C., England, S., Akopian, A.N. and Wood, J.N. (1998) A sensory neuron-specific, proton-gated ion channel. Proc. Nutl. Acad. Sci. USA, 95: 10240-10245. Cobbe, S.M. and Poole-Wilson, P.A. (1980) The time of onset and severity of acidosis in myocardial ischaemia. J. Mol. Cell. Card.,

12: 745-760.

Cobbe, S.M., Parker, D.J. and Poole-Wilson, P.A. (1982) Tissue and coronary venous pH in ischemic canine myocardium. C&n. Card.,

5: 153-156.

Cook, S.P., Vulchanova, L., Hargreaves, K.M., Elde, R. and McCleskey, E.W. (1997) Distinct ATP receptors on pain-sensing and stretch-sensing neurons. Nature, 387: 505-508. Coscoy, S., Lingueglia, E., Lazdunski, M. and Barbry, P. (1998) The Phe-Met-Arg-Phe-amide-activated sodium channel is a tetramer. J. Biol. Gem., 273(14): 8317-8322. Coutts, A.A., Jorizzo, J.L., Eady, R.A., &eaves, M.W. and Bumstock, G. (1981) Adenosine triphosphate-evoked vascular changes in human skin: mechanism of action. Em J. Phanacol., 76: 391-401. Crea, F., Pupita, G., Galassi, A.R., El-Tamini, H., Kaski, J.C.,

Davies, G. and Maseri, A. (1990) Role of adenosine in pathogenesis of angina1 pain. Circulation, 81: 164-172. Cutler, EC. (1927) Summary of experiences up to date in the surgical treatment of angina pectoris. Am. J. Med. Sci., 173: 613-624. Davidson, E.M., Coggeshall, R.E. and Carlton, S.M. (1997) Peripheral NMDA and non-NMDA glutamate receptors contribute to nociceptive behaviors in the rat formalin test. NeuroReport,

8: 941-946.

Dowd, E., McQueen, D.S., Chessell, I.P. and Humphrey, P.P. (1998) Adenosine Al receptor-mediated excitation of nociceptive afferents innervating the normal and arthritic rat knee joint. BI: J. Pharmacol., 125: 1267-1271. Gnecchi-Ruscone, T., Montano, N., Contini, M., Guazzi, M., Lombardi, F. and Malliani, A. (1995) Adenosine activates cardiac sympathetic afferent fibers and potentiates the excitation induced by coronary occlusion. J. Auton. Nerv. Syst., 53: 175184. Hamilton, S.G., Wade, A. and McMahon, S.B. (1999) The effects of inflammation and inflammatory mediators on nociceptive behaviour induced by ATP analogues in the rat. BI: J. Pharmacol.,

126: 326-332.

Hong, J.L., Kwong, K. and Lee, L.-Y. (1997) Stimulation of pulmonary C fibres by lactic acid in rats: contributions of H and lactate ions. J. Physiol., 500: 3 19-329. Huettner, J.E. (1990) Glutamate receptor channels in rat DRG neurons: activation by kainate and quisqualate and blockade of desensitization by Con A. Neuron, 5: 255-266. Jahr, C.E. and Jessell, T.M. (1983) ATP excites a subpopulation of rat dorsal horn neurones. Nature, 304: 730-733. Keele, C.A. and Armstrong, D. (1964) Substances Producing Pain and Itch. Edward Arnold, London. Khakh, B.S., Humphrey, I?P. and Surprenant, A. (1995) Electrophysiological properties of P2X-purinoceptors in rat superior cervical, nodose and guinea-pig coeliac neurones. J. Physiol., 484: 385-395. Kirkup, A.J., Booth, C.E., Chessell, I.P., Humphrey, P.P. and Grundy, D. (1999) Excitatory effect of P2X receptor activation on mesenteric afferent nerves in the anaesthetised rat. J. Physiol., 520: 551-563. Krishtal, O.A. and Pidoplichko, V.I. (1980) A receptor for protons in the nerve cell membrane. Neuroscience, 5: 2325-2327. Krishtal, O.A. and Pidoplichko, V.I. (1981) A receptor for protons in the membrane of sensory neurons may participate in nociception. Neuroscience, 6: 2599-2601. Krishtal, O.A., Marchenko, S.M. and Pidoplichko, V.I. (1983) Receptor for ATP in the membrane of mammalian sensory neurones. Neurosci. Len., 35: 41-45. Leenan, F.H. (1999) Cardiovascular consequences of sympathetic hyperactivity. Can. J. Cardiol., lS(Supp1 A): 2A-7A. Levine, J. and Taiwo, Y. (1994) Inflammatory Pain. In: P.D. Wall and R. Melzack (Eds.j, Textbook of Pain, 3rd edn., Churchill Livingstone, Edinburgh, pp. 45-54. Lewis, C., Neidhart, S., Holy, C., North, R.A., Buell, G. and Surprenant, A. (1995) Coexpression of P2X2 and P2X3 receptor subunits can account for ATP- gated currents in sensory neurons. Nature, 371: 432-435.

37 Lewis, T. (1932) Pain in muscular ischemia. Arch. Inf. Med., 49(5): 713-727. Lindahl, 0. (1961) Experimental skin pain induced by injection of water-soluble substances in humans. Actu Physiol. Stand., 5 l(Supp1. 179): 32-239. Lindahl, 0. (1974) Pain - a general explanation. Adv. Neural., 4: 45-47. Lingueglia, E., de Weille, J.R., Bassilana, F., Heurteaux, C., Sakai, H., Waldmann, R. and Lazdunski, M. (1997) A modulatory subunit of acid sensing ion channels in brain and dorsal root ganglion cells. J. Biol. Chem., 272: 29778-29783. Liu, L. and Simon, S.A. (1994) A rapid capsaicin-activated current in rat trigeminal ganglion neurons. Neurobiology, 91: 738-741. Liu, L. and Simon, S.A. (1996) Capsaicin and nicotine both activate a subset of rat trigeminal ganglion neurons. Am. J. Physiol., 270: C1807-C1814. Liu, L., Pugh, W., Ma, H. and Simon, S.A. (1993) Identification of acetylcholine receptors in adult rat trigeminal ganglion neurons. Brain Res., 617: 37-42. MacKenzie, A.B., Surprenant, A. and North, R.A. (1999) Functional and molecular diversity of purinergic ion channel receptors. Ann. New York Acad. Sci., 868: 716-729. MacWilliam, J.A. and Webster, W.J. (1923) Some applications of physiology to medicine. Br Med. J., i: 51-53. Malliani, A., Schwartz, l?J. and Zanchetti, A. (1969) A sympathetic reflex elicited by experimental coronary occlusion. Am. J. Physiol., 217: 703-709. Meller, S.T. and Gebhart, G.F. (1992) A critical review of the afferent pathways and the potential chemical mediators involved in cardiac pain. Neuroscience, 48: 501-524. Millan, M.J. (1999) The induction of pain: an integrative review. Prog.

Neurobiol.,

57: l-164.

North, R.A. (1996) Families of ion channels with two hydrophobic segments. Curr: Opin. Cell Biol., 8: 474-483. North, R.A. and Barnard, E.A. (1997) Nucleotide receptors. Curr: Opin.

Neurobiol.,

7: 346-357.

Omote, K., Kawamata, T., Kawamata, M. and Nan&i, A. (1998) Formalin-induced release of excitatory amino acids in the skin of the rat hindpaw. Bruin Rex, 787: 161-164. Opie, L.H., Owen. P., Thomas, M. and Samson, R. (1973) Coronary sinus lactate measurements in assessment of myocardial ischemia. Am. J. Cardiol., 32: 295-305. Pan, H.-L. and Longhurst, J.C. (1995) Lack of a role of adenosine in activation of ischemically sensitive cardiac sympathetic afferents. Heart Circ. Physiol., 38: H106-H113. Pan, H.-L., Longhurst, J.C., Eisenach, J.C. and Chen, S.-R. (1999) Role of protons in activation of cardiac sympathetic C-fibre afferents during ischaemia in cats. J. Physiol.. 518: 857-866. Park, W., Masuda, I., Cardenal-Escarcena, A., Palmer, D.L. and McCarty, D.J. (1996) Inorganic pyrophosphate generation from adenosine triphosphate by cell-free human synovial fluid. J. Rheumatol., 23: 665-671. Pesin, S.R. and Candia, O.A. (1982) Acetylcholine concentration and its role in ionic transport by the cornea1 epithelium. Invest. Ophthalmol.

Es. Sri., 22: 651-659.

Pickering, G.W. and Wayne, E.J. (1933) Observations on angina pectoris and intermittent claudication in anemia. Clin. Sci., I: 305-325. Poole-Wilson, P.A. (1978) Measurement of myocardial intracellular pH in pathological states. J. Mol. Cell. Curdiol., 10: 51 l-526. Reeh, P. and Petho, G. (2000) Nociceptor excitation by thermal sensitization - a hypothesis. In .I. Sandkiihler, B. Bromm and G.F. Gebhart (Eds.), Nervous System Plasticity and Chronic Pain, Progress in Brain Research, Vol. 129, Elsevier Science, Amsterdam, pp. 39-50. Robertson, S.J., Rae, M.G., Rowan, E.G. and Kennedy, C. (1996) Characterization of a P2X-purinoceptor in cultured neurones of the rat dorsal root ganglia. Br J. Pharmacol., 118: 951956. Rotto, D.M. and Kaufman, Ml? (1988) Effect of metabolic products of muscular contraction on discharge of group III and IV afferents. J. Appl. Physiol., 64: 2306-2313. Stebbins, C.L. and Longhurst. J.C. (1986) Bradykinin in reflex cardiovascular responses to static muscular contraction. J. Appl. Physiol., 6 1: 27 l-279. Steen, K.H., Reeh, PW., Anton, E and Handwerker, H.O. (1992) Protons selectively induce lasting excitation and sensitization to mechanical stimulation of nociceptors in rat skin, in vitro. .I. Neurosci., 12: 86-95. Sucher, N.J., Cheng, T.P and Lipton, S.A. (1990) Neural nicotinic acetylcholine responses in sensory neurons from postnatal rat. Brain Rex, 533: 248-254. Tanelian, D.L. (1991) Cholinergic activation of a population of comeal afferent nerves. Enp. Brain Rex, 86: 414-420. Taylor, E.J. (Ed.), Dorland’s Illustrated Medical Dictionary, W.B. Saunders, Philadelphia, PA, 1988. Thames, M.D., Kinugawa, T. and Dibner-Dunlap, M.E. (1993) Reflex sympathoexcitation by cardiac sympathetic a&rents during myocardial ischemia: role of adenosine. Circulation, 87: 1698-1704. Thimm, F., Carvalho, M., Babka, M. and Meier zu Verl, E. (1984) Reflex increases in heart-rate induced by perfusing the hind leg of the rat with solutions containing lactic acid. P&g. Arch.,

400: 286-293.

Tominaga, M., Caterina, M.J., Mamberg, A.B., Rosen, T.A., Gilbert, H., Skinner, K., Raumann, B.E., Basbaum, AI. and Julius, D. (1998) The cloned cap&in receptor integrates multiple pain-producing stimuli. Neuron, 21: 531-543. Tsuda, M., Ueno, S. and Inoue, K. (1999) In vivo pathway of thermal hyperalgesia by intrathecal administration of alpha,beta-methylene ATP in mouse spinal cord: involvement of the glutamate-NMDA receptor system. BK J. Pharmacol., 127: 449-456. Uchida, Y. and Murao, S. (1975) Acid-induced excitation of afferent cardiac sympathetic nerve fibers. Am. J. Physiol., 228: 27-33. Veelken, R., Glabasnia, A., Stetter, A., Higers, K.F., Mann, J.F.E. and Schmieder, R.E. (1996) Epicardial bradykinin B2 receptors elicit a sympathoexcitatory reflex in rats. Hypertension, 28: 615-621. Waldmann, R. and Lazdunski, M. (1998) H-gated cation chan-

38 nels: neuronal acid sensors in the NaC/DEG family of ion channels. CUUKOpin. Neurobiol., 8: 418. Waldmann, R., Champigny, G., Voilley, N., Lauritzen, I. and Lazdunski, M. (1996) The mammalian degenerin MDEG, an amiloride-sensitive cation channel activated by mutations causing neurodegeneration in C. elegans. J. Biol. Chem., 271(18): 10433-10436. Waldmann, R., Bassilana, F., de Weille, J., Champigny, G., Heurteaux, C. and Lazdunski, M. (1997b) Molecular cloning of a non-inactivating proton-gated Na+ channel specific for sensory neurons. J. Biol. C/rem., 272: 20975-20978. Waldmann, R., Champigny, G., Bassilana, F., Heurteaux, C. and Lazdunski, M. (1997a) A proton-gated cation channel involved in acid-sensing. Nature, 386(6621): 173-177.

Waldmann, R., Champigny, G., Lingueglia, E., De Weille, J.R., Heurteaux, C. and Lazdunski, M. (1999) H(+)-gated cation channels. Ann. New York Acad. Sci., 868: 67-76. White, J.C. (1957) Cardiac pain: anatomic pathways and physiologic mechanisms. Circulation, 16: 644-655. Wood, 3.N. and Docherty, R. (1997) Chemical activators of sensory neurons. Annu. Rev. Physiol., 59: 457-482. Zeilhofer, H.U., Kress, M. and Swandulla, D. (1997) Fractional Ca currents through capsakin- and proton-activated ion channels in rat dorsal root ganglion neurones. J. Physiol., 503: 67-78.

Zhou, S., Bonasera, L. and Carlton, SM. (1996) Peripheral administration of NMDA, AMPA or KA results in pain behaviors in rats. NeuroReport, 7: 895-900.

J. Sandkiihler, B. Bromm and GE Gebhart (Ed%) Progress in Brain Research, Vol. 129 0 2000 Elsevier Science B.V. All rights reserved

CHAPTER

3

Nociceptor excitation by thermal sensitization a hypothesis Peter W. Reeh L* and GciborPethij 2 2 Institute

’ Institute of Physiology and Experimental of Pharmacology and Pharmacotherapy,

Pathophysiolog): UniversitZitsstl: Faculty of Medicine, University

Introduction Sensitivity to noxious heat of the skin and oral cavity appears to be a useful protective mechanism. However, many deep tissues including skeletal muscle, dura, testis and colon, are also reported to be innervated by polymodal, mechano-heat-sensitive primary afferents (Kumazawa et al., 1991; Bove and Moskowitz, 1997; Kumazawa and Mizumura, 1997; Su and Gebhart, 1998). Fig. 1 originates from a single-fiber recording in the rabbit phrenic nerve, in vitro, and the recorded unit formed a mechanoreceptive field in the epineurium of the nerve and the joined mediastinal pleura. Radiant heat was very effective in inducing a sharp rise of the discharge frequency from close to 40°C on, and a very steep stimulus-response function precisely encoded the tissue temperature up to the peak of 46°C. Another surprising deep tissue with graded responses to noxious heat is the peripheral nerve (Fig. 2). Axons of the desheathed sciatic nerve have recently been shown to release calcitonin gene-related peptide (CGRP) in response to heat. The threshold and slope of the stimulus-response function exhibits typical nociceptive magnitudes (Sauer et al., 1999). Thus,

*Corresponding author: P.W. Reeh, Institute of Physiology and Experimental Pathophysiology, Universit%str. 17, D-91054 Erlangen, Germany. Tel.: +49 (9131) 8522228; Fax: +49 (913 1) 852-2497; E-mail: [email protected]

17, O-91054

Erlangen,

Germany Pets,

of Pets, Szigeti ~4.12, H-7624

Hungary

sensitivity to noxious heat is not restricted to sensory nerve endings. What may be the biological purpose of a sensory capacity to signal core temperatures that are more than sufficient to kill immediately? We believe that heat sensitivity and the transduction mechanisms involved subserve a second function, namely to sense inflammation and attacks by naturally occurring irritants. This is achieved by intracellular mechanisms in the nociceptive terminal that lower the heat threshold rapidly and profoundly, so that body or tissue temperature can become a gating force of nociceptor excitation and pain. Of course, this is a hypothesis meant to stimulate experimental investigation, but, at first glance, it explains the immediate pain relief that results from cooling injured, inflamed or chemically irritated tissue. From routine clinical practice, it is well known that pain associated with acute inflammatory states, such as appendicitis or dental pulpitis and with injury such as surgical wounds or distorsions, can efficiently be alleviated by local cooling. In the human skin, pain induced by infiltration with a buffer solution of low pH or by topical capsaicin application was abolished by cooling of the skin (Steen and Reeh, 1993; Kilo et al., 1995). The frequency of action potentials recorded from the rat saphenous nerve in response to intradermahy injected capsaicin was rapidly reduced by ice cooling of the innervated skin area (Szolcsanyi, 1977). Also, in the rat saphenous nerve, excitation of C-mechano-heat-sensitive (C-MH) nociceptors by topical mustard oil was

40

.

II 5 8 c 1’1 * 8 II s 8 - ‘I 1 c (I”‘b

‘IS b’Im 8 7 b m I’#

‘I

30 40s 20 10 Fig. 1. A mechano-heat-sensitive C-fiber in the rabbit pleuraphrenic nerve preparation, in vitro. Radiant heat stimulation evoked a log-linear increase in discharge rate (each dot representing one spike) in many C-fibers of this preparation, mechanical (von Frey) thresholds were widely scattered between 1 and 64 mN. The regression line was calculated through seconds 17-25 during the active heating phase and extrapolated to cut the l/s-line; by that, a heat threshold (7’) could be estimated (Sauer. Weidner and Reeh, unpublished observation). 0

blocked by placing ice on the receptive fields (Reeh et al., 1986). Room temperature

as heat stimulus

Discovery of heat-evoked ionic currents and of the heat-activated capsaicin receptor (VRl) and its homologue (VRL-1) in sensory neurons of the rat dorsal root ganglion (DRG) has shed important new light on the transduction mechanisms by which heat, inflammatory mediators and chemical irritants may excite nociceptors and contribute to pain (Cesare and McNaughton, 1996; Caterina et al., 1997, 1999). Substances that are able to lower the heat threshold in DRG cells or in VRl-transfected cells are putative candidates, in the light of the above hypothesis, to exert their excitatory effect on nociceptors by a detour through heat transduction pathways.

Indeed, bradykinin (BK) and phorbol ester, which both can activate (see also Malmberg, 2000, this volume) protein kinase C (PKC; Burgess et al., 1989; Dray et al., 1992; McGuirk and Dolphin, 1992), were shown to sensitize DRG cells to heat, decreasing their heat thresholds from above 40 to below 30°C within seconds (Cesare and McNaughton, 1996). In these experiments BK by itself induced a small transient inward current, whereas phorbol ester did not, which was taken as an argument for a PKC- and, thus, sensitization-independent mechanism of BK-induced excitation. However, the first description of BK-induced currents in DRG cells already had established that staurosporine, a blocker of protein kinases, including PKC, prevents this BK effect (Burgess et al., 1989). Thus, the sensitizing and the excitatory mechanism seem to share the same key enzyme, suggesting that they may be identical. Whether a sensitizing agent is strong enough to induce a depolarizing current by itself depends on whether the instantaneous heat threshold is lower or higher, respectively, than the ambient temperature, which is not well defined by ‘room temperature’ as a starting point in most patch-clamp studies. Differences in ambient temperature may explain why BK regularly induced excitatory inward currents in rat DRG cells in some laboratories, but rarely in others (Burgess et al., 1989; Nicol and Cui, 1994; Kress et al., 1997; Vyklicky et al., 1998). When action potentials were recorded in deep tissues (at body temperature), BK excited almost all nociceptive C-fibers, the prevalence of BK sensitivity being 92% in cat knee joint (Kanaka et al., 1985), 93% in dog testis (Kumazawa and Mizumura, 1980), 87% in dog muscle (Kumazawa and Mizumura, 1997), 83% in cat cardiovascular system (Baker et al., 1980), and 73% in cat gastrointestinal organs (Longhurst et al., 1984). In the skin of the limbs (at 32°C and similar temperatures), only about 40-50% of the C-MH fibers are excited by BK in the cat (Beck and Handwerker, 1974), rat (Lang et al., 1990) and monkey (Khan et al., 1992). In the skin, however, the threshold to noxious heat was able to drop from 45 to 30-31°C after sensitization, by topical capsaicin in this case, as revealed in psychophysical experiments (Szolcsanyi, 1977). Thus, the differences in chemical responsiveness between skin and deep tissues may result from their

41

,

5

10

15

20

25

30

min

Fig. 2. Stimulated release of calcitonin-gene related peptide (CGRP, measured by enzyme-immunoassay) from isolated stretches of desheathed rat sciatic nerve (n = 6-16); the noxious temperatures caused significant enhancements of the CGRP content in eluates which were entirely or to 90% (at 52°C) dependent on external calcium. The log-linear stimulus response curve (insert) of the peak release demonstrates a high thermal coefficient Qlo = 15 (Sauer, Bove and Reeh, unpublished observation).

higher temperature rather than from more sensitive nociceptors. In a recent patch-clamp study of rat DRG cells, the ambient temperature was precisely maintained at 22-24°C and capsaicin and BK as well as pH 6.1 were found to lower the threshold of heat-activated currents to the respective bath temperature (Vyklicky et al., 1999). Actually, if heat stimulation was superimposed on capsaicin superfusion at a low concentration, the heat-activated current developed directly out of a smaller chemically induced inward current without any apparent threshold, suggesting that both currents are finally gated by thermodynamic forces and passing through channels which are controlled by temperature in the first place (Fig. 3). In this view “capsaicin is not an agonist per se, but functions as a modulatory agent, lowering the channel’s response threshold to the ubiquitous actions of heat.” This concept has explicitly been proposed, based on a study demonstrating the polymodality of VRl expressed in HEK 293 cells or Xenupus oocytes (Tominaga et al., 1998). In this paper, low pH was introduced as a third activator of VR 1, after capsaicin and heat, and a mutual interaction between these algogenic stimuli was shown: the higher the proton concentration, the lower was the threshold to heat stimulation and vice versa; the higher the ambient temperature, the lower was the proton concentra-

tion needed to evoke a sustained excitatory inward current. This ‘communicating tube’ system only developed in cells transfected with VRl, and heat thresholds close to room temperature were found at pH 6.3. The cellular and molecular findings are sufficient reason to admit low pH, as in inflammatory tissue acidosis, to our hypothesis after having incorporated capsaicin and phorbol esters as exogenous irritants, and BK as an inflammatory mediator. Temperature

coefficients and protein stability

Heat-activated currents as well as nociceptor discharges present with very steep stimulus-response curves with Qlo (temperature coefficient over a 10°C temperature range) values well beyond 10 (see Fig. 1; Vyklicky et al., 1999). The same holds true for noxious heat-induced CGRP release from peripheral axons (Sauer et al., 1999). These temperature coefficients are far higher than those describing transport or enzymatic reactions and most voltagegated ion currents (except for the deactivation kinetics of a special Cl- channel from Torpedo showing a Qlo of 40, Pusch et al., 1997). Vyklicky et al. (1999) recently translated the high Qlo values into ‘net free energies of stabilization’ that act to determine the conformation of a hypothetical thermosensitive

42

Fig. 3. Interaction of heat and capsaicin stimulation in evoking inward currents in cultured rat DRG cells patch-clamped to -60 mV membrane potential. The neuron showed a heat threshold of 42S’C before capsaicin and a heat-activated peak current of 0.7 nA. During capsaicin-induced sensitization the current took off with the onset of heat stimulation at 24°C and peaked with 2.9 nA. From Dittert et al. (1998), with permission.

membrane protein. The result corresponded to activation energy (E,) values of 195-300 kJ M-’ that have to be invested in order to change the conformational structure and, in turn, to open the gates of heatactivated ion channels. Such high E, values may reflect reversible loosening of ionic or covalent bonds, thus, real fragmentation of the protein whose subunits tend to reassemble with high affinity when returning to lower temperature (Vyklicky et al., 1999). Vyklicky et al. (1999) tend to assume a G-protein-like structure that breaks apart upon heating and activates, through one of its fragments, an ionic conductance conveyed in another part of the protein complex. However, since a heat sensor together with an ion conductance could be expressed in HEK 293 cells or Xenopus oocytes by transfecting them with only one, the VRl gene, it appears more likely that both capacities are located in one and the same protein structure (Caterina et al., 1997). If one considers the high energy absorption necessary to induce significant heat-activated currents, it

appears unlikely that the algogenic mediators could provide comparably high and sufficient gating energy when binding to their specific membrane receptors with binding constants in the micromolar concentration range (Kress and Reeh, 1996). The recent patch-clamp study on heat-activated currents in DRG cells provides an intriguing solution of this energy problem which, in addition, supports the central hypothesis of this chapter. Capsaicin, BK and pH 6.1 lowered the heat threshold (by up to 20°C) and increased the heat response, probably by increasing open channel probability and, by that, the conductance of heat-activated channels at any given stimulus temperature, but, at the same time, the steepness of the stimulus-response curves was drastically reduced, corresponding to a drop of the Qio values from an average 18 to a range of 1.9-2.8 (Vyklicky et al., 1999). In our interpretation, this could mean that the mediators induce conformational changes in the heat-sensitive protein(s) which go along with a marked loss of ‘net free energy of stabilization’ (i.e. a marked destabilization or sensitization). Capsaicin and protons may exert their conformational effect through binding directly to the VRl receptorchannel complex and, in addition, by inducing calcium influx followed by secondary reactions (see below), and BK may act through calcium-independent PKC, as recently shown (Zeilhofer et al., 1997; Cesare et al., 1999). Generally, phosphorylation is thought to reduce the stability of proteins. By inducing this, the algogenic mediators may leave it to body or tissue temperature to activate the excitatory currents that finally lead to pain. Heat sensitivity pharmacologically VRl and VRL- 1 are perfect models for the transduction mechanisms and their interactions in nociceptive nerve endings but, of course, they do not fully account for all aspects of heat- and low-pH-induced nociception. This is indicated by some pharmacological inconsistencies. For example, both the competitive capsaicin receptor antagonist capsazepine (10 FM) and the non-competitive antagonist (blocker of the capsaicin-gated cation channel) ruthenium red (10 FM) block the effect not only of capsaicin, but also of noxious heat (by about 80%) on VRl expressed in Xenopus oocytes or HEK 293 cells

43

(Caterina et al., 1997; Tominaga et al., 1998). In the latter system, capsazepine also substantially inhibits the membrane current induced by low pH (Tominaga et al., 1998). On the other hand, only ruthenium red, and not capsazepine has a suppressive effect on the sustained pH-induced current in cultured DRG neurons (Zeilhofer et al., 1996; Vyklicky et al., 1998). In cat cornea1 nociceptors, excitation by carbon dioxide (low pH) is not blocked by capsazepine (Chen et al., 1997). Also, both blockers are only partially effective in reducing (by about 30%) heat-activated currents in DRG cells (Kirschstein et al., 1999). A more extensive comparison of channel properties and pharmacological profiles has recently been published (Kress and Zeilhofer, 1999). The pharmacological discrepancies regarding heat sensitivities have partly been resolved with the identification and cloning of a capsaicin receptor homologue, VRL-1, which is expressed in rat DRG cells, shows a very high threshold (around 52°C) for noxious heat and is only blocked by ruthenium red and not by capsazepine (Caterina et al., 1999). However, VRl- or VRL-1-transfected cells and cultured DRG cells are just models of the nociceptive nerve ending in the periphery. It should be remembered that in initial studies capsazepine (5-20 pM) was found not to inhibit the noxious heat response of dorsal horn neurons or of polymodal nociceptors in the rat skin (Dickenson and Dray, 1991; Seno and Dray, 1993). Also, capsazepine (10 pM) does not inhibit the noxious heat-induced CGRP release in the rat from either cutaneous nociceptors (Petho and Reeh, unpublished) or axons of the desheathed sciatic nerve (Sauer et al., 1999). Similarly, ruthenium red (10 FM) is completely ineffective against heat (Fig. 4) and pH responses (data not shown) of rat cutaneous nociceptors and does not inhibit CGRP release from peripheral axons induced by noxious heat (Sauer et al., 1999). Ruthenium red is not a very selective drug, blocking many different ion channels, and finding it effective would not provide much mechanistic insight. Finding it ineffective, however, may indicate that neither VRI nor WI-1 in their original structure are responsible for nociceptive heat sensitivity. Further homologues or splice variants and even proteins unrelated to VRl need to be searched and insensitivity to ruthenium red block should be a pharmacological search criterion.

1

corium temperature Fig. 4. Heat stimulus-response curves of two polymodal nociceptor populations in rat skin, in vitro, treated with ruthenium red 1 and 10 FM, respectively. The underlying heat stimulus was a linear rise of temperature from 32 to 45°C in 20 s followed by passive cooling; regression lines were calculated up to the peak discharge. The rising phase of the heat response corresponded to Qta = 16 and was not altered by ruthenium red (1 KM) pretreatment of the receptive fields for 5 min (upper panel). Ruthenium red had a weak, though significant and concentration-dependent excitatory effect by itself, and it blocked capsaicin-induced excitation as well as desensitization completely in a slowly reversible manner (data not shown). At 10 pM concentration, ruthenium red pretreatment even showed a small sensitizing effect (lower panel) increasing Qtc to 22 (St. Pierre and Reeh, in preparation; M.D. Thesis, University Medical School, Erlangen, 1993).

Sensitizing second messengers A very good agreement exists between cellular or molecular models and findings from peripheral nociceptive terminals when it comes to characterizing

44

the essential role of calcium influx, which seems to be a sufficient, though not exclusive mechanism for induction of sensitization to heat. In the skin-nerve preparation, in vitro, loading nerve endings with surplus calcium-buffering capacity (using BAPTA-AM) consistently prevented the heat sensitization that followed application of either capsaicin, pH 6.1 or 5.4, ionomycin (a calcium ionophore) or intracellular UV-photolysis of ‘caged calcium’ (NITR-S/AM; Gunther et al., 1999; Kress and Gunther, 1999). In the same studies, three of the sensitizing stimuli, capsaicin, low pH and ionomycin, were applied to patch-clamped DRG cells loaded with the calcium fluorochrome Fura-2, and a marked sensitization of heat-activated currents occurred which paralleled fluorometric changes in calcium content with respect to both time course and magnitude (Gunther et al., 1999; Kress and Gunther, 1999). If calcium was removed from the extracellular solution around the cultured neurons, the capsaicin- and heat-activated currents were essentially unchanged, but calcium influx and sensitization to heat were abolished (Gunther et al., 1999). This is consistent with the results described above using BAPTA-AM and, together, the findings may provide further indication that VRl is not directly responsible for basic heat sensitivity or for sensitization of nociceptors to heat. Moreover, in capsaicin-induced heat sensitization, VRI seems to play the crucial role of a ligand-gated calcium channel that allows calcium to enter the nerve endings and exert the sensitizing effect through other heat-activated ion channels. The BAPTA-AM loading method did not prevent BK-induced sensitization of primary afferents to heat, which lends further support to a calcium-independent, most likely PKC,-mediated mechanism (Cesare et al., 1999; Kessler et al., 1999). Activation of PKC may also be the mechanism by which histamine, through Hi receptors, can induce sensitization to heat of testicular nociceptors (Koda et al., 1996). With respect to the sensitizing role of intracellular calcium, it cannot yet be decided whether it acts through calcium-activated isoenzymes of PKC, through calcium-calmodulin kinases or through activating adenylyl cyclase isoenzymes, increasing cyclic AMP (CAMP) content and stimulating protein kinase A (PKA) (see also Malmberg, 2000, this volume, for the role of protein kinase subtypes).

A contribution of the latter cascade is indicated by studies on cultured DRG neurons and by the fact that high concentrations of prostaglandin Ez (PGE2) can induce sensitization of testicular nociceptors to heat (Mizumura et al., 1993; Cui and Nicol, 1995). In the knee joint preparation, close-arterial injection of PGE:! can even induce weak excitatory effects (Schaible and Schmidt, 1988; Birrell and McQueen, 1993), a finding supported by effects of PGE2 injection into receptive fields of human itch receptors (C-fibers) recorded microneurographically (Schmelz et al., 1998). PGEz is thought to act by elevating CAMP content, and stable membrane-permeant analogs of CAMP (but not cGMP) induce heat sensitization of cutaneous nociceptors (Kress et al., 1996). Thus, the intracellular pathways to sensitization seem to converge in protein phosphorylation by different protein kinases, acting on the putative heatactivated channels and counter-balanced by protein phosphatases whose cellular control mechanisms are largely unknown (Kress and Zeilhofer, 1999). Alternative or complementary theories of inflammatory sensitization do not aim at the process of sensory transduction, but at the conversion of generator potentials into sequences of action potentials (see also Cummins et al., 2000, this volume, for the role of sodium channels). Patch-clamp work done on cultured sensory neurons offered two modulatory mechanisms: shortening of the after-hyperpolarization of the action potentials, owing to BK receptor-mediated prostaglandin formation, and increase of tetrodotoxin-resistant voltage-gated Na+ currents due to simultaneous activation of PKC and PKA, for example by PGE2 (Weinrich et al., 1995; Gold et al., 1998). Both mechanisms could cooperate in increasing the gain of the conversion and, by that, the heat-induced discharge of nociceptive nerve fibers. However, neither mechanism can differentiate between heat- and mechanically evoked generator potentials and, thus, cannot account for the high selectivity of inflammatory sensitization to heat, and not to mechanical stimulation, in polymodal C-fibers (e.g. Kenins, 1984; Kocher et al., 1987; Kessler et al., 1992). Nonetheless, if as in other work PKA was activated in an assumed isolated manner by application of CAMP analogs, the result was sensitization to both heat and mechanical stimulation in cutaneous nociceptors (Kress et al., 1996).

45

Dilemma and cool solution The general concept to which the present chapter is devoted was encouraged by the pioneering results from David Julius’ laboratory (see above, Tominaga et al., 1998), but it was actually born in a dilemma (reviewed in Reeh and Sauer, 1997). In a series of studies using the skin-nerve preparation of the rat,

A

Exp.data

we had repeatedly met with strong tachyphylaxis (i.e. homologous desensitization) to the excitatory effect of BK on polymodal nociceptors, questioning this mediator’s role in sustained inflammatory pain. BK’s sensitizing effect to heat occurs at much lower concentration (Kumazawa et al., 1991), and in our experiments it turned out not to be subject to any obvious tachyphylaxis, but to be repeatable or sustained

CMH 0.38 m/s

40 35 0 30 8 RJ 25 c $ 2-g 20 In 15 10 5 0

be.,Ot t” ;” t I5 t ‘“t ‘t” 3o 35t min

* heat thresholds (exp.data) Fig. 5. Novel theory of BK action. (A) Histograms from a rat cutaneous CMH-fiber responding to BK superfusion (black columns) and to repeated heat stimulation (white columns) demonstrate the apparent paradox of desensitizing BK response and, at the same time, sustained sensitization to heat induced by BK. (B) The theory denies a direct excitatory effect of BK and, instead, suggests a fast sensitizing action of BK that lowers nociceptive heat thresholds far below ambient temperature of the skin-nerve preparation (32”C, dotted line). By that, the actual tissue temperature becomes an effective stimulus inducing nociceptor discharge that is then subject to sensory adaptation; together, both effects determine the initial peak discharge. In addition, desensitization of the BK transduction allows the ‘heat’ threshold to rise again with a tendency to level out. The heat thresholds (asterisks) are derived from the heat responses in A; the question marks denote two responses superimposed on vivid discharge preventing threshold estimation. The connecting lines are hypothetical and drawn to mirror-image the BK-induced discharge in A.

46

for as long as BK was present (Fig. 5A). This was in accordance with previous psychophysical work indicating that the thermal hyperalgesic effects of BK are much less prone to tachyphylaxis and more sustained than the algesic effects (Manning et al., 1991). In our experiments, the sensitizing effect of BK extended to a much larger subpopulation of polymodal afferents than the excitatory effect, and it even recruited a proportion of high-threshold C-mechanoreceptors (C-HTM) that reversibly became heat-sensitive after BK superfusion. The sensitization to heat was much more pronounced in those units that were excited by BK than in the ones not excited. The latter result is in accordance with a previous study that showed a strong correlation between the magnitude of BK-induced excitation and the degree of sensitization to heat in testicular nociceptors (Kumazawa et al., 1991). Both findings are consistent with our theory that excitation by BK is the extreme of sensitization. In a further study using subtype-selective BK receptor antagonists, half of the C-MH and C-HTM fibers not excited by BK were sensitized to heat, and in half of these cases the sensitization could not be prevented by HOE140, the B2 receptor antagonist; the effects of BK were blocked by des-Arg9-[Leu*]-BK, the Bi receptor antagonist. The Bi receptor does not exhibit homologous desensitization, which would fit the lack of tachyphylaxis of the sensitizing BK effect. However, in the vast majority of fibers, the B2 receptor was found responsible for both sensitization and excitation, if present. Here the dilemma arose. If tachyphylaxis to BK is due to desensitization of the established transduction pathway (through PKC) and includes internalization of the B2 receptor (Pizard et al., 1999), how could such a mechanism possibly spare the BK-induced nociceptor sensitization to heat? The only way out was to assume that the real magnitude of the sensitizing effect was much larger, at first, and declined to what we could detect when the excitatory effect of BK was fading (Fig. 5B). The peak of the sensitizing effect and its real time course may be hidden behind the apparent excitatory phase during superfusion with BK. To uncover this period, we had to get rid of the excitatory effect, remembering that moderate cooling had previously been shown to abolish ongoing nociceptor discharge

l

0

5

10

15

30

20 25

BK dM 45-j p

40

0

35

g30

I.

*-o-e-0 / li\

'

l ,*. . /+.- - * - .

. . . . . .

25 kf 25 20

, 0

I

, 5

I

, 10

I

, 15

I

, , 20 25

1

, 30

BK 1CfM

*2oi

I, 0

5

10

15

20 25

30

time (min) Fig. 6. Effect of BK on heat thresholds of four different C-polymodal nociceptors in the cooled rat skin, in vitro. The skinnerve preparation was cooled to 18-20°C before BK superfusion to prevent the apparent excitatory effect; the dotted lines represent the normal bath temperature of 32°C. Radiant heat stimuli were ramp-shaped (19-4S’C, 40 s) but discontinued shortly after exceeding threshold in order to allow for frequent determinations (Petho and Reeh, unpublished observation).

in inflamed rat skin as well as pain in humans after cutaneous capsaicin application or acid infusion (Szolcsanyi, 1977; Reeh et al., 1986; Steen and Reeh, 1993; Kilo et al., 1995). Fig. 6 shows the outcome of this consideration. In this ongoing study, the isolated skin-nerve preparation is cooled to 18-20°C before BK or other

47

algogens are applied, and this has so far prevented all BK-, capsaicin- and most of the pH-induced discharge. Then, ramp-shaped heat stimuli are applied at short intervals and discontinued as soon as a few spikes are recorded to permit response threshold determination (temperature at second spike). The figure gives examples from four individual polymodal C-fibers and shows the fast and dramatic sensitizing effect of a moderate concentration of BK. In several fibers, response threshold dropped by 20°C and more within less than 1 min; the first threshold determined during BK was always the lowest. It was followed by a partial increase in threshold (recovery toward the pre-BK threshold), thus revealing the previously missing decline of the sensitizing effect (‘desensitization’) of BK and attaining a sustained level after a while. Rapid reversibility of the sensitization within a few minutes during wash-out of BK had previously been demonstrated and was not the subject of the present study (Koltzenburg et al., 1992). If one considers (Fig. 6) the temporal profile of the thresholds staying below the normal temperature of the organ bath, it is easy to imagine that this nadir actually reflects the apparent excitatory effect of BK which normally occurs at higher ambient temperatures. Repeated short applications of BK would presumably induce less and less of the peak sensitizing effect due to desensitization of BK transduction and, by that, the mirror image of sensitization, the apparent excitatory effect, would present with the well-known tachyphylaxis. We speculate that the early and fast peak sensitization of nociceptors is mediated mainly by PKC because the sensitization of the heat response of DRG neurons by BK is equally fast, mimicked by the phorbol ester PMA and reversed by the PKC inhibitor staurosporine (Cesare and McNaughton, 1996). In a subsequent study, BK-induced translocation of PKC, into the plasma membrane of DRG neurons was clearly shown (Cesare et al., 1999). The plateau of sensitization indicated in Fig. 6 and, in particular, the short-lasting after-effect following wash-out of BK (which was the subject of the previous investigations into BK-induced nociceptor sensitization) may progressively depend on secondary prostaglandin production. This is supported by an ongoing study showing that the active S(+) but not the inactive R(-) enantiomer of the cyclooxygenase

inhibitor flurbiprofen abolished heat sensitization induced by sustained (5 min) BK exposure, and this effect could partially be overcome by exogenously applied PGE2 (Petho et al., 1999). The proposed role of both PKC and prostaglandins in BK-induced heat sensitization is in accordance with the fact that BK, via B2 as well as B t receptors, can couple not only to phospholipase C/PKC (see above), but also to phospholipase AZ (Pizard et al., 1999), which initiates prostaglandin formation in the nerve ending itself or in the surrounding tissue. Supporting this view, BK-induced release of PGE2 from isolated rat skin has recently been measured (using an ELISA) and found sustained (over 15 min) at a high level after an initial phase of desensitization (Petho, Izydorczyk and Reeh, unpublished observation). During the BK experiments with the cooled skinnerve preparation, we occasionally applied complete heat stimuli, linearly increasing the temperature from 19 to 45°C in 40 s, to determine the pattern of response of sensitized fibers to heat. The result was the usual log-linear increase in the discharge rate, with a considerable number of spikes below the normal organ bath temperature of 32°C. The slopes of the stimulus-response curves were abnormally flat with Qia values around 3, just as predicted by Vyklicky et al. (1999) based on their work on heat-activated currents in DRG cells using capsaicin, BK and low pH as sensitizing agents at controlled room temperature (see above). In ongoing studies, with the cooled skin-nerve preparation, capsaicin had a similar sensitizing effect as BK, but, as expected, with a much stronger decline and desensitization after an initial peak effect. Low pH effects on heat thresholds were slow to develop, less profound, but very well sustained, again as expected from the sustained excitatory effect of this stimulus on nociceptors (Steen et al., 1992). Conclusions There is growing evidence that several inflammatory mediators, including BK (possibly PGE2 and histamine) and low pH as well as the sensitizing model agent capsaicin, act on nociceptors, at least partly, by lowering the threshold of their heat transduction mechanisms so profoundly that body or lower tissue temperatures become a driving force of ex-

48

citation and pain. This unifying theory is attractive for pharmaceutical research and development because diverse and multiple nociceptive mechanisms are converging onto one novel target, heat-activated ion channels, which, however, await final molecular identification. Acknowledgements The authors’ work was supported by Deutsche Forschungsgemeinschaft (SFB 353) European Union and Sander-Stiftung. G.P.‘s fellowship was granted by the Humboldt-Stiftung. Alexandra Derow made valuable contributions to the most recent data and Iwona Izydorczyk provided expert technical help. References Baker, D.G., Coleridge, H.M., Coleridge, J.G.G. and Nerdrum, T. (1980) Search for a cardiac nociceptor: stimulation by bradykinin of sympathetic afferent nerve endings in the heart of the cat. .I. Physiol., 306: 519-536. Beck, PW. and Handwerker, H.O. (1974) Bradykinin and serotonin effects on various types of cutaneous nerve fibers. Pjfiigers

Arch.

EM

J. Physiol.,

314: 209-222.

Birrell, G.J. and McQueen, D.S. (1993) The effects of capsaicin, bradykinin, PGEz and cicaprost on the discharge of articular sensory receptors in vitro. Brain Res., 611: 103-107. Bove, G.M. and Moskowitz, M.A. (1997) Primary afferent neurons innervating guinea pig dura. J. Neurophysiol., 77: 299308. Burgess, G.M., Mullaney, I., McNeil, M., Dunn, P.M. and Rang, H.P. (1989) Second messengers involved in the mechanism of action of bradykinin in sensory neurons in culture. J. Neurosci.,

9: 3314-3325.

Caterina, J.M., Schumacher, M.A., Tominaga, M., Rosen, T.A., Levine, J.D. and Julius, D. (1997) The capsaicin receptor: a heat-activated ion channel in the pain pathway. Nature. 389: 816-824. Caterina, M.J., Rosen, T.A., Tominaga, M., Brake, A.J. and Julius, D. (1999) A caps&in-receptor homologue with a high threshold for noxious heat. Nature,398: 436-441. Chen, X., Belmonte, C. and Rang, H.P. (1997) Capsaicin and carbon dioxide act by distinct mechanisms on sensory nerve terminals in the cat cornea. Pain, 70: 23-29. Cesare, P, Dekker, L.V., Sardini, A., Parker, P.J. and McNaughton, P.A. (1999) Specific involvement of PKC, in sensitization of the neuronal response to painful heat. Neuron, 23: 611-624. Cesare, P. and McNaughton, P. (1996) A novel heat-activated current in nociceptive neurons and its sensitization by bradykinin. Proc. Natl. Acad. Sci. USA, 93: 15435-15439.

Cui, M. and Nicol, G.D. (1995) Cyclic AMP mediates the prostaglandin Ez-induced potentiation of bradykinin excitation in rat sensory neurons. Neuroscience, 2: 459-466. Cummins, T.R., Dib-Haji, S.D., Black, J.A. and Waxman, S.G. (2000) Sodium channels and the molecular pathophysiology of pain. In: J. Sandktihler, B. Bromm and G.F. Gebhart (Eds.), Nervous System Research, Vol.

Plasticity

and Chronic

Pain, Progress

in Brain

129, Elsevier Science, Amsterdam, pp. 3-19. Dickenson, A.H. and Dray, A. (1991) Selective antagonism of capsaicin by capsazepine: evidence for a spinal receptor site in capsaicin-induced antinociception. BE J. Pharmacol., 104: 1045-1049. Dittert, I., Vlachova, V., Knotkova, H., Vitaskova, Z., Vykhcky, L., Kress, M. and Reeh, P.W. (1998) A technique for fast application of heated solutions of different composition to cultured neurons. J. Neurosci. Methods, 82: 195-201. Dray, A., Patel, LA., Perkins, M.N. and Rueff, A. (1992) Bradykinin-induced activation of nociceptors: receptor and mechanistic studies on the neonatal rat spinal cord-tail preparation in vitro. BI: J. Pharmacol., 107: 1129-l 134. Gold, MS., Levine, J.D. and Correa, A.M. (1998) Modulation of TTX-R/N, by PKC and PKA and their role in PGEz-induced sensitization of rat sensory neurons in vitro. J. Neurosci., 18: 10345-10355. Gunther, S., Reeh, P.W. and Kress, M. (1999) Rises in [Ca’+li mediate capsaicin- and proton-induced heat sensitization of rat primary nociceptive neurons. Eur: J. Neumsci., 1 I: 31433150. Kanaka, R., Schaible, H.G. and Schmidt, R.F. (1985) Activation of fine articular afferent units by bradykinin. Brain Res., 327: 81-90. Kenins, P. (1982) Responses of single nerve fibres to capsaicin applied to the skin. Neurosci. Lett., 29: 83-88. Kessler, W., Kirchhoff, C., Reeh, P.W. and Handwerker, H.O. (1992) Excitation of cutaneous afferent nerve endings in vitro by a combination of inflammatory mediators and conditioning effect of substance P. Exp. Bruin Res., 91: 467-476. Kessler, F., Habelt, C., Averbeck, B., Reeh, P.W. and Kress, M. (1999) Heat-induced release of CGRP from isolated rat skin and effects of bradykinin and the protein kinase C activator PMA. Pain, 83: 289-295. Khan, A.A., Raja, S.N., Manning, D.C., Campbell, J.N. and Meyer, R.A. (1992) The effects of bradykinin and sequencerelated analogs on the response properties of cutaneous nociceptors in monkeys. Somatosens. Mot. Res., 9: 97-106. Kilo, S., Forster, C., Geisslinger, G., Brune, K. and Handwerker, H.O. (1995) Inflammatory models of cutaneous hyperalgesia are sensitive to effects of ibuprofen in man. Pain, 62: 187193. Kirschstein, T., Greffrath, W., Bttsselberg, D. and Treede, R.D. (1999) Inhibition of rapid heat responses in nociceptive primary sensory neurons of rats by vanilloid receptor antagonists. J. Neurophysiol.,

82: 2853-2860.

Kocher, L., Anton, F., Reeh, PW. and Handwerker, H.O. (1987) The effect of carrageenan-induced inflammation on the sensitivity of unmyelinated skin nociceptors in the rat. Pain, 29: 363-373.

49

Koda, H., Minagawa, M., S&Hong, L., Mizumura, K. and Kumazawa, T. (1996) Hl receptor mediated excitation and facilitation of the heat response by histamine in canine visceral polymodal receptors studied in vitro. J. Neurophysiol., 76: 1396-1404. Koltzenburg, M., Kress, M. and Reeh, P.W. (1992) The nociceptor sensitization by bradykinin does not depend on sympathetic neurones. Neuroscience, 46: 465-413. Kress, M. and Gunther, S. (1999) Role of [Ca’+]i in the ATPinduced sensitization process of rat nociceptive neurons. J. Neurophysiol., 81: 2612-2619. Kress, M. and Reeh, P.W. (1996) Transduction mechanisms in nociceptors - chemical excitation and sensitization in nociceptors. In: F. Cervero and C. Belmonte (Eds.), Neurobiology of Nociceptors, Oxford University Press, Oxford, pp. 258-297. Kress, M. and Zeilhofer, H.U. (1999) Capsaicin, protons and heat: new excitement about nociceptors. Trends Pharmacol. Sci., 20: 112-118. Kress, M., Rod], J. and Reeh, PW. (1996) Stable analogs of cyclic AMP but not cyclic GMP sensitize unmyelinated primary afferents in rat skin to mechanical and heat stimulation but not to inflammatory mediators, in vitro. Neuroscience, 74: 609-617. Kress, M., Reeh, PW. and Vyklicky, L. (1997) An interaction of inflammatory mediators and protons in small diameter dorsal root ganglion neurons. Neutosci. Lett., 224: 37-40. Kumazawa, T. and Mizumura, K. (1980) Chemical responses of polymodal receptors of the scrotal contents in dogs. J. Physiol., 299: 219-232. Kumazawa, T. and Mizumura, K. (1997) Thin-fibre receptors responding to mechanical, chemical, and thermal stimulation in the skeletal muscle of the dog. J. Physiol., 273: 179-194. Kumazawa, T., Mizumura, K., Minagawa, M. and Tsujii, Y. (1991) Sensitizing effects of bradykinin on the heat responses of the visceral nociceptor. J Neurophysiol., 66: 18 19-1824. Lang, E., Novak, P., Reeh, P.W. and Handwerker, H.O. (1990) Chemosensitivity of fine afferents from rat skin in vitro. J. Neurophysiol., 63: 887-901. Longhurst, J.C., Kaufman, D.P., Ordway, G.A. and Mush, T.I. (1984) Effects of bradykinin and capsaicin on endings of afferent fibres from abdominal visceral organs. Am. J. Physiol. Regul. Integr Comp. Physiol., 247: R552-R559. Malmberg, A. (2000) Protein kinase subtypes involved in injury-induced nociception. In: J. Sandkiihler, B. Bromm and G.F. Gebhart (Eds.), Nervous System Plasticity and Chronic Pain, Progress in Brain Research, Vol. 129. Elsevier Science, Amsterdam, pp. 51-59. Manning, D.C., Raja, S.N., Meyer, R.A. and Campbell, J.N. (1991) Pain and hyperalgesia after intradermal injection of bradykinin in humans. Clin. Pharmacol. Ther, 50: 721-729. MC&irk, S.M. and Dolphin, A.C. (1992) G-protein mediation in nociceptive signal transduction: an investigation into the excitatory action of bradykinin in a subpopulation of cultured rat sensory neurons. Neuroscience, 49: 117-128. Mizumura, K., Minagawa, M., Tsujii, Y. and Kumazawa, T. (1993) Prostaglandin Ez-induced sensitization of the heat re-

sponse of canine visceral polymodal receptors in vitro. Neurosci. Len., 161: 117-119. Nicol, G.D. and Cui, M. (1994) Enhancement by prostaglandin Ez of bradykinin activation of embryonic rat sensory neurones. J Physiol., 480: 485-492. Petho, G., Derow, A. and Reeh, P.W. (1999) Cyclooxygenase products are involved in the bradykinin-induced sensitization to heat of rat cutaneous nociceptors, in vitro. Pfliigers Arch. Eur J. Physiol., 437: R134. Pizard, A., Blaukat, A., Muller-Esterl, W., Alhenc-Gelas, F. and Rajerison, R.M. (1999) Bradykinin-induced internalization of the human B2 receptor requires phosphorylation of three serine and two threonine residues at its carboxyl tail. J. Biol. Chem., 18: 12738-12747. Pusch, M., Ludewig, U. and Jentsch, T.J. (1997) Temperature dependence of fast and slow gating relaxations of ClC-0 chloride channels. J. Gen. Physiol., 109: 105-l 16. Reeh, P.W. and Sauer, S. (1997) Chronic aspects in peripheral nociception. In: Jensen et al. (Eds.), Proceedings of the 8th World Congress on Pain, IASP Press, Seattle, pp. 115-13 1. Reeh, P.W., Kocher, L. and Jung, S. (1986) Does neurogenic inflammation alter the sensitivity of unmyelinated nociceptors in the rat?. Brain Rex, 384: 42-50. Sauer, SK., Bove, G.hl. and Reeh, PW. (1999) Heat-induced CGRP release from rat sciatic nerve: partial block by capsaicin antagonists. Pfliigers Arch. Eur J. Physiol., 437: R54. Schaible, H.G. and Schmidt, R.F. (1988) Excitation and sensitization of fine articular afferents from cat’s knee joint by prostaglandin El, J. Physiol., 403: 9 I- 104. Schmelz, M., Schmidt, R., Weidner, C., Torebjork, H.E. and Handwerker, H.O. (1998) Chemical responsiveness of mechanosensitive and -insensitive C nociceptors in human skin. Sot. Neurosci. Abstr, 24: 383. Seno, N. and Dray, A. (1993) Capsaicin-induced activation of fine afferent fibres from rat skin in vitro. Neuroscience, 55: 563-569. Steen, K.H. and Reeh, P.W. (1993) Sustained graded pain and hyperalgesia from harmless experimental tissue acidosis in human skin. Neutosci. Lett., 154: 113-l 16. Steen, K.H., Reeh, PW., Anton, F. and Handwerker, H.O. (1992) Protons selectively induce lasting excitation and sensitization to mechanical stimulation of nociceptors in rat skin, in vitro. J. Neurosci., 12: 86-95. Su, X. and Gebhart, G.F. (1998) Mechanoseusitive pelvic nerve afferent fibers innervating the colon of the rat are polymodal in character. J. Neurophysiol., 80: 2632-2644. Szolcsanyi, J. (1977) A pharmacological approach to elucidation of the role of different nerve fibres and receptor endings in mediation of pain. J. Physiol. (Paris), 73: 251-259. Tominaga, M., Caterina, M.J., Malmberg, A.B., Rosen, T.A., Gilbert, H., Skinner, K., Raumann, B.E., Basbaum, A.I. and Julius, D. (1998) The cloned capsaicin receptor integrates multiple pain-producing stimuli. Neuron, 2 1: 53 l-543. Vyklicky, L., Knotkova-Urbancova, H., Vitaskova, Z., Vlachova, V., Kress, M. and Reeh, P.W. (1998) Inflammatory mediators at acidic pH activate capsaicin receptors in cultured sensory neurons from newborn rat. J. Neurophysiol., 79: 670-676.

SO

Vyklicky, L., Vlachova, V., Vitaskova, Z., Dittert, I., Kabat, M. and Orkand, R.K. (1999) Temperature coefficient of membrane currents induced by noxious heat in sensory neurones in the rat. .I. Physiol., 517: 181-192. Weinrich, D., Koschorke, GM., Undem, B.J. and Taylor, G.E. (1995) Prevention of the excitatory actions of bradykinin by inhibition of PC& formation in nodose neurones of the guineapig. J. Physiol., 483: 135-746.

Zeilhofer, H.U., Reeh, P.W., Swandulla, D. and Kress, M. (1996) Ca2+ permeability of the sustained proton-induced cation current in adult rat dorsal root ganglion neurons. J. Neurophysiol., 76: 2834-2840.

Zeilhofer, H.U., Kress, M. and Swandulla, D. (1997) Fractional Ca2+ currents through capsaicin- and proton-activated ion channels in rat dorsal root ganglion neurones. J. Physiol., 503: 67-78.

J. Sandkiihler, B. Bromm and GE Gebhart (Eds.) Progress in Brain Research, Vol. 129 0 2000 Elsevier Science B.V. All rights reserved

CHAPTER 4

Protein kinase subtypes involved in injury-induced nociception Annika B. Malmberg * Neurobiology Unit, Roche Bioscience, 3401 Hillview Avenue, Palo Alto, CA 94304, USA

Introduction Peripheral tissue or nerve injury may result in long lasting pain conditions that may be refractory to current treatment strategies. Characteristic for these persistent pain states are an enhanced response to noxious stimuli (hyperalgesia) and an inappropriate pain sensation to innocuous stimuli, such as touch and pressure (allodynia). There is considerable evidence that the underlying mechanisms of persistent pain after injury results from long-term changes in nociceptive processing of peripheral sensory neuron (peripheral sensitization) and/or neurons of the central nervous system (central sensitization). While membrane receptors may contribute to the initiation of peripheral and central sensitization, it is likely that second messenger-activated changes in neuronal activity accounts for the maintenance of persistent pain conditions. Peripheral sensitization may be evoked by a variety of inflammatory mediators, including cyclooxygenase products. There is substantial evidence that the mechanism of prostaglandin-induced sensitization of sensory neurons involves activation of the cyclic AMP transduction cascade (Taiwo et al., 1989; Taiwo and Levine, 1991; Hingtgen et al., 1995; England et al., 1996; Kress et al., 1996). At

the spinal cord level, several studies have demonstrated that the persistent pain states after tissue or nerve injury are sensitive to antagonists selective for the N-methyl-D-aspartate (NMDA) subtype of glutamate receptors (Coderre and Melzack, 1992; Nlsstriim et al., 1992; Chaplan et al., 1997; see also Gerber et al., 2000, this volume; Moore et al., 2000, this volume; Sandkiihler et al., 2000, this volume; Svendsen et al., 2000, this volume). The NMDA receptor is highly permeable to Ca2+ and several second messenger activated pathways have been implicated in NMDA receptor signaling. These include Ca2+/diacyl-glycerol-dependent protein kinase (also known as protein kinase C, PKC), calcium/calmodulin-dependent kinase and cyclic GMP-dependent protein kinase. After stimulation, the activated protein kinase may phosphorylate various substrate proteins, such as ion channels, G-protein-coupled receptors and other enzymes, to enhance neuronal function and pain signaling (Fig. 1). Specifically, PKC and cyclic AMP-dependent protein kinase (PKA) have been implicated in the establishment of changes in excitability of neurons involved in nociceptive transmission (Taiwo et al., 1989; Taiwo and Levine, 1991; Cerne et al., 1992, 1993; Mao et al., 1993; Palecek et al., 1994; Lin et al., 1996). Protein kinase A

* Corresponding author: A.B. Malmberg, Neurobiology Unit, Roche Bioscience, 3401 Hillview Avenue, Palo Alto, CA 94304, USA. Tel.: +l (650) 855-5369; Fax: +l (650) 852-1050; E-mail: [email protected]

PKA has been implicated in persistent pain mechanisms both in the peripheral and central nervous system. As mentioned above, the sensitization of pri-

52 Extracellular Neurotransmitters,

nerve impulses, hormones

Membrane Ion channels,

signals

1

receptors

G-protein

Intracellular CAMP

Cyclic AMPdependent protein kinase (PKA)

coupled receptors

signals

Ca2+

Ca2+/diacylglycerol dependent protein Kinase (PKC)

Substrate Ion channels,

cGMP

Ca2+/calmodulin dependent protein kinase

Cyclic GMPdependent protein kinase

proteins

receptor proteins, enzymes

Fig. 1. Schematic illustration of signal transduction pathways. Extracellular signaling molecules interact with and activation of intracellular second messengers. These messengers may then interact with one or several result in phosphorylation of substrate proteins that may change the excitability of the neuron.

mary afferent neurons that occur in the setting of inflammation have been shown to involve cyclic AMP- and PKA-dependent mechanisms (England et al., 1996; Kress et al., 1996). At the spinal cord

specific protein

membrane proteins kinases, which can

level, injection of cyclic AMP or the catalytic subunit of PKA enhances the response of dorsal horn neurons to glutamate-gated ion channel activation (Ceme et al., 1992, 1993). These studies are consis-

Cyclic AMP-dependent Protein Kinase (PKA) regulatory subunk catslytlc subunits:

Fig. 2. The PKA state, the enzyme to the regulatory catalytic subunits.

Rla, RIP, Rlla, Rllp Ca, Cp

holoenzyme consists of a regulatory dimer (R) and a catalytic (C) subunit bound to each regulatory subunits. In this is inactive. Cyclic AMP (CAMP), which is generated by adenylate cyclase (AC) activation, activates PKA by binding subunits. The CAMP binding results in dissociation of the PKA enzyme into free a regulatory subunit and two free Each activated catalytic subunit can then phosphorylate substrate proteins.

53 tent with other studies suggesting that PKA plays a role in prolonged changes in neuronal synaptic efficacy, which include studies of long-term potentiation in the hippocampus (Frey et al., 1993; Huang et al., 1996). It has been suggested that persistent pain signaling may involve a similar mechanism since after high-frequency electrical stimulation of primary afferent neurons, a type of long-term potentiation has been demonstrated in dorsal horn neurons of the spinal cord (Randic et al., 1993; Lozier and Kendig, 1995; Liu and Sandktihler, 1995; see also Gerber et al., 2000, this volume; Sandktihler et al., 2000, this volume). Pk7A subunits There are several different isoforms of PKA, which have distinct patterns of expression in the brain and are likely to be involved in specific signaling events (Cadd and M&night, 1989). The PKA holoenzyme is a tetramer composed of a regulatory subunit dimer, which contains the CAMP binding sites, and a single catalytic subunit bound to each regulatory subunit (Fig. 2). At least four regulatory (RIa, RIB, RIIa, RIIB) and two catalytic (Ccl, Cg) subunits have been characterized. The cl-isoforms of the PKA subunits are ubiquitously expressed in neural and non-neural tissues, but the fi-isoforms are more restricted and are highly expressed in the nervous system (Cadd and M&night, 1989). While selective inhibitors of the PKA isoforms are not available, mutant mice that lack one of the subunits have been developed to study the function of PKA isoforms. PKA RI/3 mutant mice Mice that selectively lack the gene the encodes the neuronal specific isoform of the type I regulatory subunit (RIB) of PKA are viable and show normal general behaviors (Brandon et al., 1995). The PKA RIB mutant mice showed normal responses in tests of acute nociception, including thermal, mechanical and chemical stimuli, suggesting that activation of nociceptive afferents per se does not require PKA RIB (Malmberg et al., 1997a). In contrast, the RI8 mutant mice showed a marked reduction of the pain behavior in a model of persistent pain after tissue injury (e.g. second phase of the formalin test). The

second phase of the formalin test is known to results from both inflammation-evoked primary afferent activity and from sensitization of dorsal horn neurons secondary to the prolonged afferent input to the spinal cord (Dickenson and Sullivan, 1990; Puig and So&in, 1996). In contrast to attenuated pain behavior in the formalin test, the PKA RIB mutant mice showed normal behavioral and anatomical changes in the spinal cord after peripheral nerve injury, a model of neuropathic pain. The PKA RIB mutant mice also displayed reduced formalin-evoked paw swelling compared to the wildtype mice, suggesting the reduced pain behavior may be a result of attenuated peripheral inflammation in the formalin test (Malmberg et al., 1997a). Since the RI8 subunit of PKA is neuronal specific, it would be assumed that the mechanism of reduced inflammation was related to a reduction to neurogenic inflammation. A common tool to address this issue is to use capsaicin, which selectively activates small diameter, unmyelinated, neuropeptide-containing primary afferent fibers to produce neurogenic inflammation (see Bevan et al., 1987; and Holzer, 1991). In the RIB mutant mice it was shown that capsaicin-evoked plasma extravasation was indeed reduced suggesting that there is a significant peripheral defect in these animals likely to account for the reduction of persistent pain behavior in these mice (Malmberg et al., 1997a). Further studies showed high levels of expression of PKA RIB in dorsal root ganglion (DRG) cells, including many that coexpressed trkA, a marker of nociceptive afferent neurons (Averill et al., 1995). PKA RIB was also coexpressed with CGRP, suggesting that the phenotype of the PKA RIB mutant mice may be related to a defect in the peptide containing primary afferent nociceptor. Mechanisms of PkA-induced sensitization Activators of PKA, including CAMP analogs, have been shown to enhance peptide release from cultured sensory neurons (Hingtgen et al., 1995; Supowit et al., 1995). This raises the possibility that the reduced inflammatory response and pain behavior in the formalin test in PKA RIB mutant mice results from a defect in PKA-mediated regulation of neurotransmitter release from the peripheral terminals of primary afferent C-fibers. In support of this idea, a

54 CAMP-dependent phosphorylation of a Ca2+ channel has been demonstrated and it is well established that Ca2+ channels are an essential component of the mechanisms for neurotransmitter release (Hell et al., 1995). Several studies have demonstrated that the lowering of pain threshold produced by paw injection of prostaglandin E2 (PGE2) can be reduced by inhibiting PKA (Taiwo et al., 1989; Taiwo and Levine, 1991). This underlying mechanism for this effect is unlikely to be attributed to changes in the release of neuropeptides from the primary afferent terminal. It has been demonstrated that the sensitization is manifest as a lowered threshold for firing of the terminal which results in a behavioral allodynia in which non-noxious stimuli can evoke withdrawal reflexes and pain behavior (Martin et al., 1986; Schaible and Schmidt, 1988). It is also known that PGE2 increases the conductance of a TTX-resistant Naf channel that is highly expressed in capsaicin sensitive, small diameter DRG cells and that CAMP mimics these effects (England et al., 1996; Gold et al., 1996). Taken together, PKA activation appears to affect the peripheral nociceptive neurons in two ways; one that involved the neurotransmitter release and one that Ca*+-diacylglycerol/phospholipid-dependent - 12 diflwent

contributes to sensitization processes of the nociceptive neuron. Both these mechanisms may be involved in persistent pain conditions and the mechanisms underlying the phenotype of the PKA RI8 mutant mice. Protein kinase C Protein kinase C (PKC) is a family of intracellular serine/threonine kinases which are critical components of many signal transduction pathways in mammalian cells. Members of the PKC family are activated by intracellular lipid second messengers phosphatidylserine, diacylglycerol (DAG) and to various extent by Ca 2+ (Nishizuka, 1992). In the absence of activators, PKC is found primarily in the cytosol (Fig. 3). Activation is accompanied by dissociation of the pseudosubstrate region from the catalytic domain and involves association with specific membrane proteins. These proteins, termed receptors for activated C kinases (RACK), function as selective scaffolds for activated PKC at discrete subcellular compartments (Mochly-Rosen, 1995). PKC activation terminates with decay of the second-messenger signal followed by relocalization of PKC to the cytoso1. protein

kinase (PKC)

isoforms

- a, PI, PII, y9 6, G 57 r\, @I, h, p

inactive

PKC Fig. 3. PKC consists of a regulatory domain (R) attached by a hinge region to a catalytic domain (C). The regulatory domain contains the autoinhibitory pseudosubstrate (P) sequence. The catalytic domain includes the substrate binding site (S). In the absence of activators, PKC is found primarily in the cytosol. Activation is accompanied by dissociation of the pseudosubstrate region from the catalytic domain and involves association with specific membrane proteins.

55 PKC isoenzymes The mammalian PKC isoenzymes are grouped into three subfamilies based on similarities in primary amino acid sequence, domain structure and sensitivity to different activators (for review and references see Mellor and Parker, 1998). The classical/conventional or cPKC subfamily includes PKCcl (alpha), PKCbI and PKQII (two alternatively spliced forms of the @ (beta) gene), and PKCy (gamma). Members of cPKC subfamily, which and are activated by Ca2+, phosphatidylserine and DAG have four homologous domains (Cl, C2, C3 and C4) interspaced with isoenzyme-unique domains. The cPKCs are also targets of the tumor-promoting phorbol ester PMA (phobol 12-myristate 13-acetate), also known as TPA (tetradecanoic phorbol acetate). Members of the second subfamily, termed novel or nPKC include, PKCE (epsilon), PKCG (delta), PKCn (eta), and PKCB (theta). The nPKC subfamily lacks the C2 domain and does not require Ca2+ for activation. The third subfamily is the atypical or aPKC family, represented by PKC< (zeta) and PKCt (iota). PKCt was first identified in the mouse and termed PKCh (lambda). Members of the aPKC subfamily lack both the C2 and half of the Cl domain and are insensitive to Ca2+, DAG and phorbol esters. The human protein kinase D (PKD) and the mouse homolog PKCp (mu) are phorbol-ester- and DAG-stimulated protein kinases that have been assigned to the atypical subfamily. However, studies of this protein kinase catalytic domain, substrate specificity and sensitivity to inhibitors have indicated that PKD/pPKC may be more related to CaMKII-like protein kinases (Johannes et al., 1994; Valverde et al., 1994). In addition to these groups, a possible fourth group of PKC-related kinases (PRKs) has been described, consisting of at least three members, PRK l-3 (Palmer et al., 1995). Like the aPKCs, the PRKs are insensitive to Ca2+, DAG and phorbol esters. Since the various PKC isoforms display specific expression pattern, cofactor and substrate requirements, it is likely that individual PKC isoforms are involved in specific cellular processes and functions. However, to date, there is limited information about the function of individual PKC isoforms. The PKC family of enzymes has been shown to control signals

for essential processes, such as cell growth, differentiation, transformation and signal transduction of extracellular receptors, including those for neurotransmitters, hormones, antigens, and growth factors. It is thus not surprising that PKC has been implicated in a wide range of physiological functions, such as oncology, immunology, cardiovascular functions and neurobiology, including pain transmission. PKC and nociception Studies using non-subtype selective PKC inhibitors have suggested that this enzyme is critically involved in persistent pain states, including pain behaviors following tissue or nerve injury (Coderre, 1992; Mao et al., 1993; Yashpal et al., 1995). These studies suggested that PKC contributed to an NMDA-receptor mediated hyperexcitability of dorsal horn neurons in the spinal cord. However, PKC isoenzymes are highly expressed in neuronal tissues of both the peripheral nervous system (PNS) and central nervous system (CNS) and may interfere with nociception at the both levels. For better understanding of the involvement of PKC in nociception, focus has to be on specific isoenzymes and the expression pattern of the isoenzymes. PKC appear to interfere with nociception both in the PNS and CNS. This is suggested by studies using mice that selectively lack PKCE (Khasar et al., 1999), which is expressed in the PNS, and in mice that lack PKCy, (Malmberg et al., 1997b), which is selectively expressed in the CNS. PKCE mutant mice It was recently demonstrated that mice which carry a selective mutation in the gene that encodes PKCE displayed decreased sensitivity to epinephrine-induced hyperalgesia and pain behaviors produced by inflammation (Khasar et al., 1999). In contrast, PGE2-evoked hyperalgesia was similar in the mutant and wild-type mice indicating that the reduced pain responses of the mutant mice are not related to an unspecific change in sensitivity of peripheral neurons. PKCE immunoreactivity was demonstrated in the peripheral processes of sensory neurons and over 90% of small-diameter DRG neurons, suggesting that PKCs-dependent changes

56 in nociceptive neurons are important for the phenotype of the mutant mice (Khasar et al., 1999). Furthermore, epinephrine-induced enhancement of tetrodotoxin-resistant sodium current in cultured rat DRG neurons were inhibited by a PKCs-selective inhibitor peptide (Khasar et al., 1999; see also Cummins et al., 2000, this volume). Taken together, it appears that PKCE is important for the regulation of nociceptive function and certain types of pain behaviors via modulation of tetrodotoxin-resistant sodium currents. PKCy mutant mice

In mice with a selective deletion of the gene that encodes for the neuronal specific isoform PKCy, increases in thermal and mechanical sensitivity after nerve injury, which is indicative of neuropathic pain, were attenuated (Malmberg et al., 1997b). In contrast, thermal or mechanical stimulation in absence of injury was normal indicating that transmission of acute ‘pain’ messages was intact in the mutant mice. In addition to deficits in the behavioral responses to peripheral nerve injury, neurochemical changes in the spinal dorsal horn were blunted in the PKCy mutant mice. However, neurochemical alterations in the DRG after nerve injury were similar in the PKCy mutant and wild-type mice, indicating that the spinal cord was the critical site for the phenotype of the PKCy mutant mice in the nerve-injury model. In agreement with this finding, PKCy is expressed only in the central nervous system, and no PKCy-labeled neurons were found in DRG cells of the PNS. Specifically, PKCy is expressed in a subset of interneurons of the inner part of the substantia gelatinosa (lamina II) of the spinal dorsal horn (Mori et al., 1990; Malmberg et al., 1997b; Martin et al., 1999). Interneurons of the inner part of lamina II receive input from a neurochemically distinct population of unmyelinated primary afferent neurons, known as the non-peptidergic population of small-diameter afferent neurons. These primary afferent neurons bind the lectin Bandeiraea simplicifolia, express an ATP-sensitive P2X3 receptor, and contain a fluorideresistant acid phosphatase in their central terminals (see Snider and McMahon, 1998; Sutherland et al., 2000, this volume). Since the interneurons of the inner lamina II respond preferentially to non-noxious

inputs (Light et al., 1979; Bennett et al., 1980; Woolf and Fitzgerald, 1983), PKCy-regulated changes in the processing of non-noxious inputs by dorsal horn neurons may be critical to the development of neuropathic pain after nerve injury. Persistent pain behavior after tissue injury and inflammation, as indicated by reduced paw edema, was also reduced in the PKCy mutant mice (Malmberg et al., 1997b). While the concomitant reduction of the behavioral and anatomical response to nerve injury points to the superficial dorsal horn as the critical locus of the PKCy contribution to neuropathic pain, it is unclear how a deletion of PKCy in lamina II interneurons result in inflammation deficits. Decreased central sensitization may explain the reduction of tissue injury-induced pain behaviors, but the mechanism underlying the reduction of neurogenic inflammation in the PKCy mutant mice is uncertain. One possibility is that a deletion of PKCy in the dorsal horn reduces intemeuron-generated dorsal root reflexes, which has been suggested to modulate the release of peptides from the peripheral terminals of primary afferents and reduce neurogenic inflammation (Sluka and Westlund, 1993; Rees et al., 1996). However, peripheral inflammation may also be influenced by sympathetic and hormonal factors, which may have been affected by the deletion of PKCy at multiple CNS sites. (Coderre et al., 1989; Hargreaves et al., 1990). While these factors cannot be excluded, the demonstration that PKCy immunoreactivity is increased in the spinal dorsal horn after peripheral inflammation suggests that these neurons are actively involved in the physiological consequences of peripheral inflammation (Martin et al., 1999). Furthermore, it was shown that changes PKCy immunoreactivity paralleled the time course of mechanical allodynia suggesting that PKCy contributes to the maintenance of the allodynia produced by peripheral inflammation. Although the exact mechanism remains to be determined, the restricted expression of PKCy in lamina II intemeurons makes these neurons ideally located to regulate both input and output of projection neurons of both the superficial and deep dorsal horn (Martin et al., 1999).

57 Conclusions There is substantial evidence to suggest that intracellular second messenger-activated protein kinases are important for the development of increased neuronal excitability and persistent nociceptive processing. Pharmacological studies using inhibitors or mice with a selective gene deletion suggest that PKA may mainly be involved in tissue injury and inflammation-evoked pain states, but not neuropathic pain, whereas PKC may be involved in all three of these persistent pain states. It appears that targeting PKA and/or PKC may be of potential medical benefit for the treatment of pain. However, most drug candidates that inhibit PKC that has been tested so far display little isoenzyme selectivity and are highly toxic (Goekjian and Jirousek, 1999). This is not surprising because of the involvement of PKC in various normal functions. Targeting specific isoforms that are specifically involved in pain processing may be a more promising approach. It is also possible that targeting the proteins that function as anchoring proteins for activated PKC may provide efficacy with less toxicity (Mochly-Rosen, 1995; MochlyRosen and Kauvar, 1998). Further characterization of specific subtypes of protein kinases and their contribution to nociceptive processing is required for identification of new potential targets directed to signaling pathways for the treatment of persistent pain.

Averill, S., McMahon, S.B., Lary, D.O., Reichardt, L.F. and Priestley, J.V. (1995) lmmunocytochemical localization of trkA receptors in chemically identified subgroups of adult rat sensory neurons. Eur J. Neurosci., 7: 1484-1494. Bennett, G.J., Abdelmoumene, M., Hayashi, H. and Dubner, R. (1980) Physiology and morphology of substantia gelatinosa neurons intracellularly stained with horseradish peroxidase. J. Comp. Neural., 194: 809-827. Bevan, S.J., James, I.F., Rang, HP., Winter, J. and Wood, J.N. (1987) The mechanism of action of capsaicin - a sensory neurotoxin. In: P. Jenner (Ed.), Neurotoxins and their Pharmacological Implications, Raven Press, New York, pp. 261271. Brandon, E.P., Zhuo, M., Huang, Y.Y., Qi, M., Gerhold, K.A., Burton, K.A., Kandel, E.R., M&night, G.S. and ldzerda, R.L. (1995) Hippocampal long-term depression and depotentiation are defective in mice carrying a targeted disruption of the gene encoding the RI beta subunit of CAMP-dependent protein kinase. Proc. Nail. Acad. Sci. USA, 92: 8851-8855.

Cadd, G. and M&night, G.S. (1989) Distinct patterns of CAMPdependent protein kinase gene expression in mouse brain. Neuron, 3: 71-19. Ceme, R., Jiang, M.C. and Randic, M. (1992) Cyclic adenosine 3’5’-monophosphate potentiates excitatory amino acid and synaptic responses of rat spinal dorsal horn neurons. Brain Res., 596: 11 l-123. Ceme, R., Rusin, K.I. and Randic, M. (1993) Enhancement of the N-methyl-D-aspartate response in spinal dorsal horn neurons by CAMP-dependent protein kinase. Neurosci. L&t., 161: 124-128. Chaplan, S.R., Malmberg, A.B. and Yaksh, T.L. (1997) Efficacy of spinal NMDA receptor antagonism in formalin hyperalgesia and nerve injury evoked allodynia in the rat. .I. Phannacol. Exp. Ther, 280: 829-838. Coderre, T.J. (1992) Contribution of protein kinase C to central sensitization and persistent pain following tissue injury. Neurosci. I&t., 140: 181-184. Coderre, T.J. and Melzack, R. (1992) The contribution of excitatory amino acids to central sensitization and persistent nociception after formalin-induced tissue injury. J. Neurosci., 12: 3665-3670. Coderre. T.J., Basbaum, A.I. and Levine, J.D. (1989) Neural control of vascular permeability: interactions between primary afferents, mast cells, and sympathetic efferents. J. Neurophysml., 62: 48-58. Cummins, T.R., Dib-Hajj, S.D., Black, J.A. and Waxman, S.G. (2000) Sodium channels and the molecular pathophysiology of pain. In: J. Sandktihler, B. Bromm and G.F. Gebhart (Eds.), Nervous System Plasticity and Chronic Pain, Progress in Brain Research, Vol. 129. Elsevier Science, Amsterdam, pp. 3-19. Dickenson, A.H. and Sullivan, A.F. (1990) Differential effects of excitatory amino-acid antagonists on dorsal horn nociceptive neurons in the rat. Brain Res., 506: 31-39. England, S., Bevan, S. and Docherty, R.J. (1996) PGEz modulates the tetrodotoxin-resistant sodium current in neonatal rat dorsal root ganglion neurones via the cyclic AMP-protein kinase cascade. J. Physiol., 495: 429-440. Frey, U., Huang, Y.Y. and Kandel, E.R. (1993) Effects of CAMP simulate a late stage of LTP in hippocampal CA1 neurons. Science, 260: 1661-1664. Gerber, G., Youn, D.-H., Hsu, C.H., lsaev, D. and Randic, M. (2000) Spinal dorsal horn synaptic plasticity: involvement of group I metobotropic glutamate receptors. In: J. Sandktihler, B. Bromm and GE Gebhart (Eds.), Nervous System Plasticity and Chronic Pain, Progress in Brain Research, Vol. 129. Elsevier Science, Amsterdam, pp. 115-134. Goekjian, PG. and Jirousek, M.R. (1999) Protein kinase C in the treatment of disease: signal transduction pathways, inhibitors, and agents in development. CurK Med. Chem., 6: 877-903. Gold, M.S., Reichling, D.B., Shuster, M.J. and Levine, J.D. (1996) Hyperalgesic agents increase a tetrodotoxin-resistant Na+ current in nociceptors. Proc. Nail. Acad. Sci. USA, 93: 1108-1112. Hargreaves, K.M., Flores, C.M., Dionne, R.A. and Mueller, G.P. (1990) The role of pituitary beta-endorphin in mediating

58 corticotropin-releasing factor-induced antinociception. Am. J. Physiol., 258: E235-242. Hell, J.W., Yokoyama, C.T., Breeze, L.J., Chavkin, C. and Catterall, W.A. (1995) Phosphorylation of presynaptic and postsynaptic calcium channels by CAMP-dependent protein kinase in hippocampal neurons. EMSO J, 14: 3036-3044. Hingtgen, C.M., Waite, K.J. and Vasko, M.R. (1995) Prostaglandins facilitate peptide release from rat sensory neurons by activating the adenosine 3’5’~cyclic monophosphate transduction cascade. J. Neurosci., 15: 541 l-5419. Holzer, P. (1991) Capsaicin: cellular targets, mechanisms of action, and selectivity for thin sensory neurons. Pharmucol. Rev., 43: 144-201. Huang, Y.-Y., Kandel, E.R., Varshavsky, L., Brandon, E.P., Qi, M., Idzerda, R.L., M&night, G.S. and Bourtchouladze, R. (1996) A genetic test of the effect of mutations in PKA on mossy fiber LTP and its relation to spatial and contextual learning. Cell, 83: 121 l-1222. Johannes, F.J., Prestle, J., Eis, S., Oberhagemann, P. and Pfizenmaier, K. (1994) PKCu is a novel, atypical member of the protein kinase C family. J. Biol. Chem., 269: 6140-6148. Khasar, S.G., Lin, Y.-H., Martin, A., Dadgar, J., McMahon, T., Wang, D., Hundle, B., Aley, K.O., Isenberg, W., McCarter, Cl., Green, P.G., Hodge, C.W., Levine, J.D. and Messing, R.O. (1999) A novel nociceptor signaling pathway revealed in protein kinase C epsilon mutant mice. Neuron, 24: 253-260. Kress, M., Rodl, J. and Reeh, P.W. (1996) Stable analogues of cyclic AMP but not cyclic GMP sensitize unmyelinated primary afferents in rat skin to heat stimulation but not to inflammatory mediators in vitro. Neuroscience, 74: 609-617. Light, A.R., Trevino, D.L. and Per], E.R. (1979) Morphological features of functionally defined neurons in the marginal zone and substantia gelatinosa of the spinal dorsal horn. J. Camp. Neural., 186: 151-171. Lin, Q.. Peng, Y.B. and Willis, W.D. (1996) Possible role of protein kinase C in the sensitization of primate spinothalamic tract neurons. J. Neurosci., 16: 3026-3034. Liu, X.G. and Sandkiihler, J. (1995) Long-term potentiation of C-fiber-evoked potentials in the rat spinal dorsal horn is prevented by spinal N-methyl-D-aspartic acid receptor blockade. Neurosci. Lett., 191: 43-46. Lazier, A.P. and Kendig, J.J. (1995) Long-term potentiation in an isolated peripheral nerve-spinal cord preparation. L Neurophysiol., 74: 1001-1009. Malmberg, A.B., Brandon, E.P, Idzerda, R.L., Liu, H., McKnight, G.S. and Basbaum, Al. (1997a) Diminished inflammation and nociceptive pain with preservation of neuropathic pain in mice with a targeted mutation of the type I regulatory subunit of CAMP-dependent protein kinase. .I. Neurosci., 17: 7462-7470. Malmberg, A.B., Chen, C., Tonegawa, S. and Basbaum, AI. (1997b) Preserved acute pain and reduced neuropathic pain in mice lacking PKCgamma. Science, 278: 279-283. Mao, J., Mayer, D.J., Hayes, R.L. and Price, D.D. (1993) Spatial patterns of increased spinal cord membrane-bound protein kinase C and their relation to increases in 14C-2-deoxyglucose

metabolic activity in rats with painful peripheral mononeuropathy. J. Neurophysiol., 70: 70-8 1. Martin, H.A., Basbaum, AI., Kwiat, G.C., Goetzl, E.J. and Levine, J.D. (1986) Leukotriene and prostaglandin sensitization of cutaneous high-threshold C- and A-delta mechanonociceptors in the hairy skin of rat hindlimbs. Neuroscience, 22: 651-659. Martin, W.J., Liu, H., Wang, H., Malmberg, A.B. and Basbaum, A.I. (1999) Inflammation-induced up-regulation of protein kinase Cgamma immunoreactivity in rat spinal cord correlates with enhanced nociceptive processing. Neuroscience, 88: 1267-1274. Mellor, H. and Parker, P.J. (1998) The extended protein kinase C superfamily. Biochem. J., 332: 281-292. Mochly-Rosen, D. (1995) Localization of protein kinases by anchoring proteins: a theme in signal transduction. Science, 268: 247-251. Mochly-Rosen, D. and Kauvar, L.M. (1998) Modulating protein kinase C signal transduction. Adv. Pharmncol., 44: 91-145. Mori, M., Kose, A., Tsujino, T. and Tanaka, C. (1990) Immunocytochemical localization of protein kinase C subspecies in the rat spinal cord: light and electron microscopic study. L Camp. Neural., 299: 167-177. Moore, K.A., Baba, H. and Woolf, C.J. (2000) Synaptic transmission and plasticity in the superficial dorsal horn. In: J. Sandkiihler, B. Bromm and G.F. Gebhart (Eds.), Nervous System Plasticity and Chronic Pain, Progress in Brain Research, Vol. 129. Elsevier Science, Amsterdam, pp. 63-80. Nlsstrom, J., Karlsson, U. and Post, C. (1992) Antinociceptive actions of different classes of excitatory amino acid receptor antagonists in mice. Eur: J. Pharmacol., 212: 21-29. Nishizuka, Y. (1992) Intracellular signaling by hydrolysis of phospholipids and activation of protein kinase C. Science, 258: 607-614. Palecek, J., Paleckova, V., Dougherty, P.M. and Willis, W.D. ( 1994) The effect of phorbol esters on the responses of primate spinothalamic neurons to mechanical and thermal stimuli. J. Neurophysiol., 71: 529-537. Palmer, R.H., Ridden, J. and Parker, P.J. (1995) Cloning and expression patterns of two members of a novel protein-kinaseC-related kinase family. ELK .I. Biochem., 227: 344-351. Puig, S. and Sorkin, L.S. (1996) Formalin-evoked activity in identified primary afferent fibers: systemic lidocaine suppresses phase-2 activity. Pain, 64: 345-355. Randic, M., Jiang, M.C. and Cerne, R. (1993) Long-term potentiation and long-term depression of primary afferent neuretransmission in the rat spinal cord. J. Neurosci., 13: 522% 5241. Rees, H., Sluka, K.A., Lu, Y., Westlund, K.N. and Willis, W.D. (1996) Dorsal root reflexes in articular afferents occur bilaterally in a chronic model of arthritis in rats. .I Neurophysiol., 76: 4190-4193. Sandktihler, J., Benrath, J., Brechtel, C., Ruscheweyh, R. and Heinke, B. (2000) Synaptic mechanisms of hyperalgesia. In: J. Sandkiihler, B. Bromm and G.F. Gebhart (Eds.), Nervous System Plasticity and Chronic Pain, Progress in Brain Research, Vol. 129. Elsevier Science, Amsterdam, pp. 81-100.

59 Schaible, H.G. and Schmidt, R.F. (1988) Excitation and sensitization of fine articular afferents from cat’s knee joint by prostaglandin Ea. J. Physiol., 403: 91-104. Sluka, K.A. and Westlund, K.N. (1993) Centrally administered non-NMDA but not NMDA receptor antagonists block peripheral knee joint inflammation. Pain, 55: 217-225. Snider, W.D. and McMahon, S.B. (1998) Tackling pain at the source: new ideas about nociceptors. Neuron, 20: 629-632. Supowit, SC., Christensen, M.D., Westlund, K.N., Hallman, D.M. and DiPette, D.J. (1995) Dexamethasone and activators of the protein kinase A and C signal transduction pathway regulate neuronal calcitonin gene-related peptide expression and release. Brain Rex, 686: 77-86. Sutherland, SF!, Cook, S.P. and McCleskey, E.W. (2000) Chemical mediators of pain due to tissue damage and ischemia. In: J. Sandkilhler, B. Bromm and G.F. Gebhart (Eds.), Nervous System Plasticity and Chronic Pain, Progress in Brain Research, Vol. 129. Elsevier Science, Amsterdam, pp. 21-38. Svendsen, F., Hole, K. and Tjolsen, A. (2000) Long-term potentiation in single WDR neurons induced by noxious stimulation in intact and spinalized rats. In: J. Sandktihler, B. Bromm and G.F. Gebhart (Eds.), Nervous System Plastic& and Chronic

Pain, Progress in Brain Research, Vol. 129. Elsevier Science, Amsterdam, pp. 153-161. Taiwo, Y.O. and Levine, J.D. (1991) Further confirmation of the role of adenyl cyclase and of CAMP-dependent protein kinase in primary afferent hyperalgesia. Neuroscience, 44: 13 1-135. Taiwo, Y.O., Bjerknes, L.K., Goetzl, E.J. and Levine, J.D. (1989) Mediation of primary afferent peripheral hyperalgesia by the CAMP second messenger system. Neuroscience, 32: 577-580. Valverde, A.M., Sinnett-Smith, J., Van Lint, J. and Rozengurt, E. (1994) Molecular cloning and characterization of protein kinase D: a target for diacylglycerol and phorbol esters with a distinctive catalytic domain. Proc. N&l. Acad. Sci. USA, 91: 8572-8576. Woolf, C.J. and Fitzgerald, M. (1983) The properties of neurones recorded in the superficial dorsal horn of the rat spinal cord. J. Comp. Neural., 221: 313-328. Yashpal, K., Pitcher, GM., Parent, A., Quirion, R. and Coderre, T.J. (1995) Noxious thermal and chemical stimulation induce increases in 3H-phorbol 12,13-dibutyrate binding in spinal cord dorsal horn as well as persistent pain and hyperalgesia, which is reduced by inhibition of protein kinase C. J. Neurosci., 15: 3263-3272.

I. Sandkiihler, B. Bromm and G.F. Gebhart (Eds.) Progress in Brain Research, Vol. 129 0 2000 Elsevier Science B.V. All rights reserved

CHAPTER 5

Synaptic transmission and plasticity in the superficial dorsal horn Kimberly A. Moore *, Hiroshi Baba and Clifford J. Woolf Neural

Plasticity Research Group, Department of Anesthesia and Critical Care, Massachusetts General Hospital and Harvard Medical School, 149 Thirteenth Street, Room 4309, Charleston MA 02129-2000, USA

Introduction The dorsal horn of the spinal cord is the site of the first synaptic relay in nociceptive pathways, transferring and processing input from primary sensory neurons to intrinsic and projection neurons. The synaptic contact between the central terminals of primary sensory and dorsal horn neurons is highly ordered functionally, chemically and topographically. Each highly specialized primary afferent has a particular arrangement of its central terminal related to the location of its peripheral terminals, as well as to its axon caliber, presence of myelin, transmitter content, growth factor responsiveness, threshold and modality sensitivity. This provides a dense but ordered presynaptic architecture or neuropil that encodes key features of the receptive field properties of the primary sensory neurons. Dorsal horn neurons also display a considerable degree of order with multiple neuronal types exhibiting diverse expression of receptors, ion channels, transmitters, dendritic fields and axonal trajectories which varies systematically with the location of the neurons in the dorsal horn. The transfer of information from primary afferent to

* Corresponding author: K.A. Moore, Neural Plasticity Research Group, Department of Anesthesia and Critical Care, Massachusetts General Hospital, 149 Thirteenth Street, Room 4309, Charleston MA 02129-2000, USA. Tel.: +l (617) 724-3306; Fax: +l (617) 724-3632; E-mail: [email protected]

second order neuron is, however, more than a static synaptic relay dictated simply by the anatomical constraints of the relative interaction between preand postsynaptic structures. Instead, the strength of the synaptic contact is highly modulated. Synaptic plasticity, increasing or decreasing synaptic efficacy, is key to the function of the system. The most superficial laminae of the dorsal horn are of fundamental importance for nociceptive transmission by virtue of the fact that it is here that most small caliber myelinated (As) and unmyelinated (C) fibers terminate. Lamina II is of particular interest as the sensory input to this area is almost entirely C-fiber in nature. Unlike laminae I and V, where many projection neurons are located, almost all lamina II neurons are intrinsic intemeurons. Such a high density of interneurons raises the following questions: What is the precise function of lamina II? What contribution does lamina II make to the transfer, integration and modulation of nociceptive input? What is the role of lamina II in generating synaptic plasticity in the spinal cord? The primary focus of this review will be the current state of knowledge about synaptic transmission in lamina II. Techniques for studying synaptic transmission and plasticity in the superficial dorsal horn A number of electrophysiological recording techniques have been utilized to study synaptic transmission and plasticity in the superficial dorsal horn.

64 A Suction electrode Dorsal

root (15-20 mm)

Computer P-CLAMP

L-l

Fig. 1. Schematic diagram of the in vitro slice preparation and patch clamp recording setup. (A) Thick (500-600 urn) transverse spinal cord slices with an attached dorsal root are cut using a vibrating microslicer and placed in the recording chamber (see below). The substantia gelatinosa (SG, lamina II) is readily discernible under the low power magnification of a dissecting microscope due to its characteristically translucent appearance. Synaptic currents are evoked by dorsal root stimulation using a suction electrode with a wire inside and recorded using the blind whole-cell patch clamp technique. (B) In the recording chamber, the slice is supported by a nylon mesh and held in placed with a titanium electron microscopy grid supported by a wire loop. Slices are perfused at -20 ml/min with Krebs solution saturated with 95% 02/5% CO2 and maintained at 36 f 1°C. Synaptic currents are amplified with a patch clamp amplifier and are monitored using an oscilloscope or chart recorder. Data are digitized with an A-D converter, then collected and analyzed using p-Clamp software (Axon Instruments, CA) on a dedicated PC.

Each of these techniques, extracellular, intracellular, and whole-cell patch clamp recording, is associated with its own unique set of advantages and disadvantages. Extracellular recording is a relatively simple technique that permits recording for long periods of time during which a number of experimental manipulations can be performed. However, this technique does not allow one to study transmission at specific synapses or the cellular mechanisms underlying synaptic transmission and plasticity. Nevertheless, until recently, extracellular recording was the main method used to measure activity in the spinal cord in vivo. Such recordings are difficult, though, in lamina II where neurons are small and densely packed, reducing extracellular field potentials and making single spike analysis problematic. Intracellular and whole-cell patch clamp recording techniques permit recording from single dorsal horn neurons. Intracellular recording provides the advantage of minimal disruption of normal cytosolic contents; allowing easy recording of G-protein mediated responses. Intracellular recording in lamina

II has been performed in vivo, and in isolated or hemisected neonatal spinal cord preparations in vitro for a number of years, to study synaptic transmission in the dorsal horn of the spinal cord. Compared to other in vitro preparations (e.g. slice preparations or dorsal horn cultures) a greater synaptic network is preserved in the hemisected spinal cord. However with such preparations it is difficult to reliably ascertain the recording position during the experiment, making it difficult to record from specific laminae. In addition, impaling neurons with intracellular recording electrodes can produce membrane damage with a major leak current and the high impedance of intracellular recording electrodes makes voltage clamp extremely difficult. Whole-cell patch clamp recording is less damaging than intracellular recording but cytosolic contents are dialyzed with the internal solution of the patch pipette, making it difficult to record responses mediated by metabotropic receptors. The perforated patch technique has been utilized to overcome such obstacles, but intracellular dialysis can be used to

65 the experimenter’s advantage. If desired, specific ion channel antagonists or G-protein inhibitors can be included in the internal solution. The whole-cell recording technique has been utilized to record from both cultured dorsal horn neurons and in vitro spinal cord slice preparations (see also Gerber et al., 2000, this volume). A microisland co-culture technique for embryonic primary afferents (dorsal root ganglion neurons) and dorsal horn neurons has recently been introduced (Gu and MacDermott, 1997). This technique has the distinct advantage that one can easily identify and study monosynaptic connections between primary afferents and dorsal horn neurons. However, the microisland culture has not yet been adapted for postnatal dorsal horn neurons. Therefore, due to developmental differences, this technique may not be appropriate for studying synaptic transmission and plasticity as it relates to nociceptive processing in the mature system. In the early 1980’s, neonatal (Miletic and Randic, 1982) and adult (Yoshimura and Nishi, 1982; Yoshimura and North, 1983) rat transverse in vitro spinal cord slice preparations were developed. In the late 1980’s, the adult preparation was modified to include a long dorsal root thus permitting differential primary afferent stimulation (Yoshimura and Jesse& 1989). Unique features of this in vitro preparation and of lamina II have facilitated blind whole-cell patch clamp recording from substantia gelatinosa (SC) neurons (Fig. 1). The thickness (-600 km) and orientation of the spinal cord slice preserves a large network of synaptic connections. Small neurons with highly branched dendrites (central cells of Cajal), along with unmyelinated or lightly myelinated axon terminals, are densely packed in lamina II (Willis and Coggeshall, 1991). This gives the SG its characteristic translucent appearance under low power magnification, allowing for its identification using a dissecting microscope. Much of the current data regarding synaptic transmission and plasticity in the superficial dorsal horn has been gleaned from experiments utilizing this in vitro preparation and it will be the main focus of this review.

Synaptic transmission in lamina II

Fast excitatory transmission Glutamate. Today, it is well accepted that glutamate released by both primary afferent neurons and interneurons mediates excitatory synaptic transmission in lamina II of the spinal dorsal horn (Yoshimura and Jessell, 1990; Yoshimura and Nishi, 1992; Fig. 2) through activation of all three classes of ionotropic glutamate receptors, AMPA, kainate and NMDA. The 6-cyano-7-nitroquinoxaline (CNQX)-sensitive AMPA receptors are the primary mediators of fast synaptic transmission, while the D-2-amino-5-phospho-novalerate (APV)sensitive NMDA receptors are thought to mediate the slower synaptic effects of glutamate. In cultured dorsal horn neurons, a subset of AMPA receptors with significant calcium permeability have recently been identified (Gu et al., 1996; Engelman et al., 1999). Unlike the highly voltage sensitive NMDA receptors, calcium permeable AMPA receptors can be activated at the resting membrane potential and may therefore play a role in synaptic strengthening. Interestingly, cobalt labeling of calcium permeable AMPA receptors in neonatal rats reveals a pattern consistent with a role for these receptors in nociceptive signaling. The highest density of calcium permeable AMPA receptors is found in lamina I and outer lamina II, regions of the dorsal horn innervated by nociceptive A& and C-fibers (Engelman et al., 1999). Until recently the function of kainate receptors in the spinal cord was largely unknown. However, Li et al. (1999a) have identified kainate receptor-mediated synaptic responses in lamina II neurons of the neonatal spinal cord. Kainate excitatory postsynaptic currents (EPSCs) can be elicited by low frequency, high intensity stimulation of primary afferents, but not by low intensity stimulation. This suggests that kainate receptors may be selectively expressed in dorsal horn neurons that exclusively receive high threshold (nociceptive) afferent inputs. A potential role for kainate receptors in nociceptive transmission is strengthened by the observation that a p-opioid receptor agonist, [D-Ala(2),N-MePhe(4),Gly-ol(5)lenkephalin (DAMGO), blocks kainate-mediated postsynaptic currents.

B

A6 mono

I

Graded intensity

-I

100

pA

10 ms

I

A6 mono + C mono

I I

Strychnike + Blcuculline I 10 ms

50 pA

20 ms

Fig. 2. Primary afferent evoked EPSCs and IPSCs in lamina II neurons. The lop, middle, and bottom panels show EPSCs evoked by graded intensity A@, A& and C-fiber dorsal root stimulation, respectively. Four or five traces are superimposed in each panel. The Afi-fiber evoked EPSCs in this neuron are polysynaptic (note the variable latency), while the A& and C-fiber EPSCs are monosynaptic (note the constant latency). All records were obtained from a single neuron. Polysynaptic IPSCs evoked by graded stimulation (H-40 pA, 0.05 ms). As the intensity increases, the synaptic latency decreases. Strychnine (2 FM, glycine receptor antagonist) eliminates transient IPSCs while bicuculline (20 KM, GABA* receptor antagonist) blocks longer lasting inhibitory responses. (With permission from Baba et al., 1999a).

In addition to its role as a neurotransmitter, kainate may also negatively modulate excitatory synaptic transmission in lamina II of the dorsal horn. Kainate reversibly attenuates both bath applied NMDA-induced currents and dorsal rootevoked NMDA receptor-mediated EPSCs (Sequeira and Nasstrom, 1998). Additionally, a kainate receptor antagonist, NS-102, significantly enhanced glutamate-induced currents in SG neurons (Sequeira and Nasstrom, 1998). The locus of the modulatory kainate effect was not systematically investigated. However, since responses to both exogenously applied NMDA and endogenously released glutamate were depressed, it appears that postsynaptic mechanisms are involved.

Adenosine Triphosphate (ATP). The role of ATP as

a neurotransmitter in the dorsal horn is controversial. However, it does appear that ATP mediates fast EPSCs via P2X receptor activation in a very small population (~5%) of SG neurons in slice preparations (Bardoni et al., 1997) and in -50% of co-cultured dorsal horn (lamina I-III) neurons (Jo and Schlichter, 1999). In contrast, the role of ATP as a neuromodulator in the dorsal horn is less divisive (Table 1). Using microisland co-cultures of dorsal horn and DRG neurons, Gu and MacDermott (1997) observed transient, ATP evoked currents in dorsal horn neurons. ATP-evoked currents were sensitive to the AMPA receptor antagonist, CNQX, suggesting that ATP facilitated

P,

TrkB

Kainate

BDNF

P2X

EP2

(Substance A)

a2

Post

Post

Post

Pre Post

Pre

Pre Post

Pre

5HT2 5HTx

a2

Pre Post

Facilitation

Facilitation

Facilitation Facilitation

Facilitation

Inhibition Inhibition

Facilitation Inhibition

Inhibition Inhibition

Inhibition Inhibition Inhibition

Inhibition

Post

Pre Post Pre

Inhibition

Pre

5HTj

Prostanoids

Tacykinins Neurokinin

ATP

Norepinephrine

Serotonin

6 K Orphanin

I*

Facilitation Inhibition Inhibition

Pre Pre Post

GABAn

Opioids

Inhibition

Effect

Pre

Site of action (pre/post-synaptic)

transmission

GABAA

synaptic

GABA

of excitatory

Receptor

1

Neuromodulator

Modulators

TABLE

neurotransmitter

release

release

release

release

$ NMDA

t NMDA

currents

(1)

t nonselective potentials

Depolarization,

cation

conductance

t Excitatory neurotransmitter release t AMPA currents f NMDA currents, t quisqualate currents

t Glutamate

4 Glutamate release Hyperpolarization, t K+conductance

-1 Neurotransmitter Hyperpohuization 4 NMDA currents t Neurotransmitter

Hyperpolarization & AMPA currents TJ, NMDA currents $ Excitatory neurotransmitter release Hyperpolarization, 4 input resistance

4 Excitatory

t Cl- conductance, release t neurotransmitter release release f Kf conductance

of action

Depolarization, 4 neurotransmitter Depolarization, & Neurotransmitter Hyperpolarization,

Mechanism

observations

and Nasstrom,

et al., 1999 Sequeira

Kerr

Baba et al., 1999

Kangrga and Randid, Rusin et al., 1992 Rusin et al., 1993

Gu and MacDermott,

Travagli and Williams, North and Yoshimura,

1998

1990

1997

1996 1984; Grudt

Hori et al., 1996 Grudt et al., 1995 Murase et al., 1990; Lopez-Garcia, Hori et al., 1996 Khasabov et al., 1999

et al., 1995

1998

Jeftinija, 1988; Hori et al., 1992; Glaum et al., 1994; Kohno et al., 1999 Yoshimura and North, 1983; Jeftinija, 1988 Kolaj and RandiC, 1996 Rusin and Rand%, 1991 Kohno et al., 1999 RandiC et al., 1995 Lai et al., 1997; Liebel et al., 1997

unpub.

et al., 1999

Xi and Akasu, 1996 Iyatomi et al., 2000 Kangrga et al., 1991; Baba and Yoshimura,

MacDermott

References

8

68

glutamate release. In addition, afferent fiber evoked glutamatergic EPSCs in SG neurons were attenuated by a selective P2X receptor antagonist (pyridoxal phosphate-6-azophenyl-2’,4’-disulfonic acid, PPADS) (Li et al., 1998), confirming a modulatory role for ATP. Several lines of evidence support a presynaptic site of ATP action. First, the frequency of spontaneous glutamatergic EPSCs is increased by ATP or the selective P2X receptor agonist, ~$3 methylene ATP (Gu and MacDermott, 1997). Second, antagonism of P2X receptors by PPADS shifts paired-pulse depression of EPSCs to paired pulse facilitation (Li et al., 1998). Finally, when ATP was applied to dorsal horn monocultures in the presence of tetrodotoxin, no glutamate-mediated responses were observed (Gu and MacDermott, 1997). Slow excitatory transmission Glutamate. Focal stimulation of the spinal dorsal horn can also elicit slow excitatory synaptic potentials in the majority (~90%) of lamina II neurons (Yajiri et al., 1997). A significantly smaller fraction (-25%) of SG neurons reveal slow EPSCs upon dorsal root stimulation, suggesting that intemeurons or descending fibers contribute to the generation of these slow potentials. Slow excitatory potentials are mediated in part by glutamate activation of NMDA receptors (Yajiri et al., 1997). These receptors are highly permeable to calcium (MacDermott et al., 1986) and are, therefore, thought to be essential for neuroplasticity in the spinal dorsal horn. Tachykinins. Though tachykinins (substance P and neurokinin A) contribute to slow excitatory responses in deep dorsal horn neurons (Murase and RandiC, 1984; Urban and Randic, 1984; King et al., 1997), whether they contribute significantly to excitatory transmission in the SG is debated. Only ten percent (or less; see Yoshimura et al., 1993) of SG neurons respond to tachykinins or selective agonists (Bleazard et al., 1994), reflecting a lack of NKl receptor expression by lamina II neurons. In contrast, in the deep dorsal horn, -45% of neurons are depolarized by tachykinin agonists (Bleazard et al., 1994). Furthermore, intracellular application of a non-hydrolyzable analog of GDP, GDPps, does not affect slow excitatory potentials

suggesting that G-protein coupled receptors (e.g. neurokinin receptors and metabotropic glutamate receptors) do not participate in slow synaptic transmission in lamina II (Yajiri et al., 1997). Selective NKl and NK2 tachykinin receptor antagonists (RP67580 and MEN10376, respectively) can, however, reduce C-fiber evoked ventral root potentials (Nagy et al., 1994) and the NK2 receptor antagonist can reduce C-fiber evoked postsynaptic responses recorded in lamina II-IV neurons (Nagy et al., 1993). Together, this suggests that tachykinins may contribute to slow synaptic responses in neurons located in laminae other than lamina II. Interestingly, following peripheral inflammation, tachykinins may play a greater role in dorsal horn synaptic transmission. Inflammation elicits the expression of substance P in a subset of Ab-fibers, which are not normally substance P-immunoreactive (Neumann et al., 1996). In addition, NKl receptor mRNA is also upregulated in the superficial dorsal horn almost two-fold over control levels (Schafer et al., 1993). Furthermore, following peripheral inflammation, Ab-fibers acquire the novel capacity to induce an NKl receptor mediated windup-like phenomenon (Thompson et al., 1994; Herrero and Cervero, 1996) and innocuous stimulation can elicit NKl receptor internalization in lamina III and IV, regions where C-fibers do not terminate (Abbadie et al., 1997). Together, these data indicate that after peripheral inflammation, substance P can be released from Ab-fibers, can activate NKl receptors, and can elicit short-term changes in synaptic transmission in the dorsal horn. Tachykinins can also function as modulators of excitatory transmission (Table 1). Nanomolar concentrations of substance P enhance AMPA, NMDA and quisqualate, but not kainate, responses in a majority of acutely isolated dorsal horn (lamina I-III) neurons (Rusin et al., 1992, 1993a,b). Neurokinin A produces similar effects, potentiating both AMPA and NMDA currents (Rusin et al., 1992, 1993b). Substance P-induced modulation of NMDA currents is suppressed by BAPTA, an intracellular calcium buffer, and staurosporine, an antagonist of both protein kinase A and protein kinase C (Rusin et al., 1992), suggesting that calcium-dependent activation of protein kinases may be essential for tachykinin-induced potentiation of excitatory transmission (see also Malm-

69

berg, 2000, this volume). In addition to these postsynaptic effects, tachykinins can also act presynaptitally to enhance basal and evoked glutamate release (Kangrga and RandiC, 1990). Inhibitory transmission The in vitro slice preparation has been utilized extensively to study transmission by inhibitory interneurons. It appears that both GABAergic and glycinergic interneurons can be activated by primary afferents, as two distinct types of polysynaptic inhibitory postsynaptic potentials (IPSPs) can be elicited by A&fiber stimulation (Baba et al., 1994; Yoshimura and Nishi, 1995). Pure bicuculline sensitive GABAergic IPSPs are observed in -40% of SG neurons, while -50% of SG neurons display pure strychnine sensitive glycinergic IPSPs. Mixed GABAergic and glycinergic IPSPs are observed in the remainder (Yoshimura and Nishi, 1995; and see Fig. 2). The kinetic parameters of GABA* and glycine receptor mediated IPSC(P)s differ considerably (Baba et al., 1994; Yoshimura and Nishi, 1995). The time to peak (-4 ms) and half decay time (-11 ms) of glycinergic IPSCs are considerably shorter than that of GABAergic IPSCs (time to peak -13 ms, half decay time -42 ms; Yoshimura and Nishi, 1995). Therefore, GABA* receptor-mediated IPSCs may play a greater role in modulating nociceptive transmission from the spinal dorsal horn to the CNS. Dorsal horn neuron excitability is also decreased by GABAa receptor activation. The majority of dorsal horn neurons (86%) in neonatal slice preparations are directly hyperpolarized by the GABAa receptor agonist, baclofen (Kangrga et al., 1991). Similarly, baclofen elicits outward currents associated with an increased K+ conductance in the majority of adult SG neurons (Baba and Yoshimura, unpublished observations). The reversal potentials of both GABA* and glycinergic IPSCs are - -70 mV indicating that both of these receptors are coupled to a Cl- conductance. In a subset of SG neurons, however, A&fiber stimulation elicits long-lasting, slow IPSC(P)s that reverse polarity at - -90 mV, the Nemst potential for K+ (Baba et al., 1994). The neurotransmitter and receptor type that underlies this long lasting IPSCs has yet to be identified. However, possible candidates include a number of neurotransmitters

that hyperpolarize SG neurons through activation of a K+ conductance (e.g. somatostatin, GABA acting through GABAn receptors, norepinephrine acting via ~12receptors, dopamine, acetylcholine acting through muscarinic receptors, and enkephalin; Yajiri et al., 1997). Recent data suggests that a long-lasting IPSP in the spinal dorsal horn may be mediated by dopamine. Focal stimulation of the deep dorsal horn can evoke long-lasting IPSCs that reverse polarity at -90 mV and are occluded by exogenously applied dopamine (Nakatsuka et al., 1999a). Presynaptic inhibition GABA

GABA may also modulate nociceptive transmission through activation of presynaptic GABA* and GABAn receptors (Table 1). Classic studies in muscle afferents suggest that presynaptic GABAA receptor activation inhibits synaptic transmission (reviewed by MacDermott et al., 1999) via primary afferent depolarization (PAD). GABA* receptor-mediated PAD is associated with an increased Cl- conductance (Eccles et al., 1961; Takeuchi and Takeuchi, 1967). Because the chloride equilibrium potential in these neurons is positive to the resting membrane potential, the driving force for Cl- is outward; therefore, GABA* receptor activation produces presynaptic membrane depolarization. There are number of potential mechanisms through which PAD could inhibit neurotransmitter release and thereby inhibit synaptic transmission. PAD may inactivate calcium channels that are required for neurotransmitter release. Alternatively, the increased Cl- conductance might shunt the calcium conductance required to activate the presynaptic release machinery. Recent work has confirmed the depolarizing action of presynaptic GABA* receptor activation in the spinal cord. However, in contrast to the earlier studies in muscle afferents, GABA* depolarization of primary afferent neurons facilitates neurotransmitter release (reviewed by Xi and Akasu, 1996). Whether presynaptic activation of GABA* receptors in superficial dorsal horn facilitates or inhibits synaptic transmission remains unsettled. In contrast, it is clear that GABAa receptor activation inhibits excitatory synaptic transmission in the

70

superficial dorsal horn. Baclofen, a GABAs receptor agonist, acts presynaptically to decrease the frequency of miniature EPSCs in SG neurons (Iyadomi et al., 2000). GABAn receptors can be positively coupled to a Kf conductance or negatively coupled to a Ca2+ conductance, either or both of these cellular mechanisms may account for the GABAn receptor-mediated reduction of neurotransmitter release. As a result of decreased neurotransmitter release, dorsal root-evoked monosynaptic EPSCs in SG neurons are significantly depressed by baclofen, acting predominantly on C-fiber, rather than A&fiber, terminals (Ataka et al., 1999). Opioids

p-Opioid agonists can also depress synaptic transmission in neonatal (Jeftinija, 1988; Hori et al., 1992; Glaum et al., 1994) and adult (Kohno et al., 1999) dorsal horn slice preparations via pre-synaptic mechanisms (Table 1). The frequency of miniature EPSCs is reduced by both p- and 61-opioid receptor agonists (DAMGO and D-Pen(2,5)-enkephalin (DPDPE), respectively), while the frequency of miniature IPSCs is unaffected (Hori et al., 1992; Kohno et al., 1999). This suggests that neurotransmitter release from excitatory, but not inhibitory terminals, is negatively modulated by u- and 6i-opioid receptor activation. As a result, the amplitude of As-fiber evoked glutamatergic EPSCs in SG neurons is diminished by p- and 8,-opioid agonists (Glaum et al., 1994; Kohno et al., 1999). Similar to the other opioids, nociceptin (also referred to as orphanin FQ) inhibits excitatory synaptic transmission in the superficial dorsal horn through a presynaptic mechanism (Lai et al., 1997; Liebel et al., 1997) (Table 1). However, the inhibitory action of nociceptin in the dorsal horn is naloxone insensitive, and is therefore not mediated by one of the ‘classic’ opioid receptors (p, 6, or K). Inhibition of excitatory Synaptic transmission in the spinal dorsal horn may explain the antinociceptive effects of nociceptin as well as the other, more classical, opioid receptor agonists.

plex fashion (Hori et al., 1996) (Table 1). In -50% of SG neurons, activation of 5-HT, receptors decreases the frequency of miniature AMPA receptor mediated EPSCs. In contrast, activation of 5-HT2 receptors by 5-HT or selective agonists (DO1 or cl-methyl 5-HT) increases the frequency of miniature EPSCs in the remaining 50% of neurons. Changes in the frequency of miniature EPSCs were independent of changes in the amplitude distribution, suggesting a presynaptic locus of action. In keeping with these effects, activation of 5-HTi or 5-HTs receptors also suppresses dorsal root evoked EPSC/EPSPs (Hori et al., 1996; Lopez-Garcia, 1998; Khasabov et al., 1999). In some cases, transient inhibition of evoked EPSCs is followed by 5-HT2 receptor-mediated potentiation (Hori et al., 1996). In a synergistic fashion, 5-HT can also facilitate inhibitory neurotransmission (Table 2). In -20% of trigeminal SG neurons, 5-HT evokes glycinergic IPSPs (Grudt et al., 1995). The TTX-sensitive nature of these effects suggests that 5-HT directly depolarizes glycinergic interneurons thus enhancing inhibitory neurotransmitter release. Norepinephrine

NE also acts presynaptically to diminish excitatory transmission and enhance inhibitory transmission (Tables 1 and 2). NE activation of ~12 receptors decreases the frequency of spontaneous miniature glutamatergic EPSPs (Travagli and Williams, 1996), while NE enhances TTX-sensitive GABAergic IPSCs in the 20% of trigeminal SG neurons (Grudt et al., 1995). Together this suggests that NE reduces the release of excitatory neurotransmitter(s) and enhances the release of GABA from inhibitory interneurons. In the spinal SG, NE facilitates TTXresistant quanta1 release of both GABA and glycine in 90% of neurons studied (Baba et al., 2000a). TTXsensitive GABA* receptor-mediated IPSCs are also enhanced, suggesting that NE activates GABAergic neurons at somatodendritic receptor sites (Baba et al., 2000b).

Serotonin (5HT)

Acetylcholine

Through differential receptor activation, 5-HT regulates presynaptic neurotransmitter release in a com-

Muscarinic agonists can also act presynaptically to increase TTX-sensitive and TTX-resistant GABA*

71 TABLE 2 Modulators of inhibitory synaptic transmission Neuromodulator

Receptor

Site of action (pre/post-synaptic)

Effect

Mechanism of action

References

Serotonin Norepinephrine

S-HTI~ ~(1

Pre Pre

Facilitation Facilitation

Pre Post

Facilitation Inhibition

t Glycine release t GABA release t Glycine release t GABA release J. GABAA currents

Grudt et al., 1995 Grudt et al., 1995 Baba et al., 2OOOa,c Baba et al., 1998 Jo and Schlichter, 1999

a2

Acetylcholine Adenosine

Muscarinic Al

IPSC(P)s in spinal SG neurons (Baba et al., 1998) and TTX-sensitive GABAA IPSC(P)s in trigeminal SG neurons (Travagli, 1996) (Table 2). Postsynaptic inhibition

currents (Chen et al., 1995). In some SG neurons, however, dynorphin depresses EPSPs without affecting passive membrane properties (membrane potential and input resistance), indicating an additional presynaptic site of action.

Opioids

Serotonin (5HT)

Activation of p-opioid receptors directly decreases superficial dorsal horn neuron excitability by hyperpolarizing the resting membrane potential (Yoshimura and North, 1983; Jeftinija, 1988). This and additional postsynaptic targets of p-receptor activation may contribute to opioid modulation of nociceptive transmission (Table 1). In a subset of acutely isolated dorsal horn neurons (kunina I-III), a u-opioid agonist (DAMGO) regulates NMDA currents in a biphasic fashion (Rusin and RandiC, 1991). An initial depression of NMDA currents by DAMGO is followed by a longer lasting (550 min) enhancement of NMDA currents upon removal of the agonist. In contrast, modulation of AMPA currents by k-opioid agonists (DAMGO and PL017) is less complex. In the majority of acutely isolated dorsal horn neurons (lamina I-IV), AMPA currents are reduced by DAMGO and PLO17 (Kolaj and RandiC, 1996). The ic-opioid agonist, dynorphin, modulates monosynaptic AMPA receptor-mediated EPSPs in SG neurons in a concentration-dependent manner (RandiC et al., 1995). Activation of ~1 receptors by nanomolar concentrations of dynorphin decreases the amplitude of monosynaptic EPSPs. In contrast, micromolar concentrations of dynorphin enhance EPSPs in SG neurons. Suggestive of a postsynaptic locus of action, dynorphin can also directly hyperpolarize most SG neurons. Dynorphin also acts postsynaptically through naloxone insensitive opioid pathways to reduce NMDA receptor mediated

5-HT also regulates excitatory synaptic transrnission in the spinal dorsal horn via postsynaptic mechanisms (Table 1). Activation of 5-HTr receptors reduces responses to bath applied NMDA in acutely isolated dorsal horn neurons (Murase et al., 1990) and in the hernisected spinal cord preparation (Lopez-Garcia, 1998). These inhibitory effects may underlie, in part, the analgesic action of this biogenic amine. Norepinephrine (NE)

Excitatory synaptic transmission in the superficial dorsal horn is also inhibited by NE activation of postsynaptic (~2 receptors (Table 1). Postsynaptic ~12receptors are coupled to a K+ conductance that directly hyperpolarizes the majority of SG neurons (North and Yoshimura, 1984; Grudt et al., 1995). Consequently, NE reduces the amplitude of EPSPs in the spinal dorsal horn. Adenosine

Recently, Jo and Schlichter (1999) demonstrated that activation of postsynaptic adenosine receptors on cocultured dorsal horn neurons depresses GABA* receptor-mediated synaptic currents (Table 2). In their system, ATP and GABA are co-released, and adenosine is generated extracellularly by ATP metabolism. Since adenosine is provided by GABAergic intemeu-

72

rons, this may be a model of activity-dependent depression of inhibitory transmission. Synaptic plasticity Use-dependent plasticity Windup Windup is a form of short-term, use-dependent plasticity that, due to similarities in underlying mechanisms, may be involved in initiating longer lasting alterations in synaptic efficacy (e.g. central sensitization or synaptic long term potentiation). It relies on repetitive, low frequency (OS-3 Hz) C-fiber intensity stimulation, during which there is a progressive increase in the number of action potentials (Mendell, 1966). Windup persists for a few minutes after the cessation of the stimulus train, at most, and can not be elicited by A-fiber stimulation under control conditions. However, under pathological conditions, A-fibers can acquire the capacity to elicit action potential windup in dorsal horn neurons (Thompson et al., 1994). Properties of C-fiber evoked postsynaptic responses may account for the unique ability of these unmyelinated fibers to generate windup. C-fiber stimulation elicits slow NMDA or tachykinin receptor mediated synaptic potentials (Urban and RandiC, 1984; Yoshimura and Jessell, 1989) that can summate with low frequency repetitive stimulation to produce a gradual membrane depolarization and action potential windup (Thompson et al., 1990). In addition, calcium entry via NMDA receptors, activation of protein kinase C, and NMDA receptor phosphorylation to relieve the voltage dependence can produce action potential windup and other more prolonged changes in dorsal horn neuron excitability that manifest at a system level as central sensitization (Woolf, 1996). Long-term potentiation (LTP) and long-term depression (LTD) of synaptic transmission Long-term synaptic plasticity (LTP and LTD) in the spinal dorsal horn has been studied extensively using extracellular recording techniques. (LTP: Svendsen et al., 1997; Sandktihler and Liu, 1998; Liu and

Sandktihler, 1998; Rygh et al., 1999 and LTD: Liu et al., 1998; Sandktihler et al., 2000, this volume). Both LTP and LTD of synaptic transmission can also be observed at the cellular level. High frequency afferent stimulation (100 Hz) elicits LTP of both AMPA and NMDA receptor-mediated EPSCs in -40% of superficial dorsal horn neurons (RandiC et al., 1993). NMDA receptor activation is essential for the induction of LTP, however it is not required for maintenance of potentiated transmission (RandiC et al., 1993). Ablation of C-fibers by neonatal capsaicin treatment does not prevent the induction of LTP, suggesting that C-fiber input is not required for the generation of this form of long-term synaptic plasticity in the dorsal horn (RandiC et al., 1993). Interestingly, in a sub-population of neurons, the same high frequency stimulus can elicit either LTP or LTD depending on the membrane potential (RandiC et al., 1993). Under conditions where LTD is favored (a hyperpolarized membrane potential), LTD can also be observed following low-frequency (1 Hz) A&fiber stimulation (Sandktihler et al., 1997). Induction of LTD requires NMDA receptor activation (RandiC et al., 1993; Sandktihler et al., 1997) and an elevation of intracellular calcium, as iontophoresis of BAPTA to buffer intracellular calcium significantly reduced the amplitude of LTD (Sandktihler et al., 1997). In contrast, LTD induction is independent of protein phosphatases (PP 1 and 2A), as well as GABA* and glycine receptor activation (RandiC et al., 1993; Sandktihler et al., 1997). In sum, these types of long-term plasticity may play a functional role in the generation of post-injury hypersensitivity (LTP) or antinociception (LTD). Unmasking of silent synapses Strengthening of ineffective or ‘silent’ synapses may contribute the synaptic plasticity observed in the dorsal horn following peripheral nerve injury or inflammation (Zhuo, 2000, this volume). Silent synapses have been identified in the superficial dorsal horn of neonatal spinal cord slices that can be unmasked by serotonin, acting through 5-HT2 receptors (Li and Zhuo, 1998). Either presynaptic (e.g. increased neurotransmitter content or decreased neurotransmitter breakdown) or postsynaptic (e.g. increased number of functional receptors or changes in membrane

73

excitability) mechanisms, or both, may contribute to serotonin-induced synaptic strengthening. Recent data (Li et al., 1999b) suggests that recruitment of AMPA receptors to the postsynaptic membrane is essential for serotonin-induced unmasking of silent synapses (see Zhuo, 2000, this volume). We, too, have observed silent NMDA-receptor mediated synapses in neonatal SG neurons. However, we do not detect silent synapses in adult lamina II or III neurons (Baba et al., 2000b). Since silent NMDA receptor mediated synapses are thought to be more prevalent at immature synapses (Isaac et al., 1997; Petralia et al., 1999), they might be expected to instead dominate at synapses produced by sprouting Ab-fibers. However, even following sciatic nerve transection to induce Ab-fiber sprouting, the number of SG neurons with silent synapses is not increased (Baba et al., 2000). Therefore, it is unlikely that unmasking of silent synapses contributes to dorsal horn plasticity related to the pathogenesis of chronic pain in the adult. InJEammation-induced synaptic plasticity Recently, we (Baba et al., 1999a) and others (Nakatsuka et al., 1999b) have examined the effects of chronic peripheral inflammation in adult rats on primary afferent evoked synaptic responses in SG neurons. Inflammation was elicited by injection of complete Freund’s adjuvant (CFA) into the rat hindpaw. Synaptic responses were recorded, either two days (Baba et al., 1999a) or seven to ten days (Nakatsuka et al., 1999b) later, from in vitro spinal cord slice preparations of the Ls segment. The frequency and amplitude of spontaneous EPSCs recorded in SG neurons from rats with peripheral inflammation were similar to that of age-matched controls (Baba et al., 1999a; Nakatsuka et al., 1999b), as was the frequency of spontaneous IPSCs (Baba et al., 1999a). We observed no monosynaptic A$-fiber mediated excitatory responses in SG neurons from naive or inflamed rats. However, the percentage of SG neurons with polysynaptic Ab-fiber mediated EPSCs was substantially increased from 25% in naive rats to 63% in rats with peripheral inflammation for two days (Baba et al., 1999a). The percentage of SG neurons with polysynaptic Afi-fiber mediated IPSCs was also significantly increased from 16% to 54% follow-

ing two days of hindpaw inflammation. Consistent with the observation that peripheral inflammation enhances Ab-fiber input to lamina II of the dorsal horn, both the threshold intensity required to evoke A-fiber mediated EPSCs and IPSCs, as well as the response latency of EPSCs, were significantly decreased following peripheral inflammation (Fig. 3). Nakatsuka et al. (1999b) observed a small percentage (7%) of SG neurons with monosynaptic Ab-fiber inputs in slices from naive rats. Following seven to ten days of peripheral inflammation, the percentage of SG neurons with monosynaptic Afi-fiber mediated EPSCs increased to 33%. There was also a small increase in the number of SG neurons with polysynaptic Al+fiber input following seven days of peripheral inflammation (from 2% in naive rat slices to 7% in slices from inflamed rats). Furthermore, following long-term hindpaw inflammation, the percentage of SG neurons with monosynaptic A&fiber EPSCS decreased from 69 to 20%, while polysynaptic A&fiber inputs increased from 22 to 40%. Two general types of mechanisms may underlie peripheral inflammation-induced recruitment of Al5fiber mediated inputs to SG neurons. Pre-existing but ineffective synapses (‘silent’ synapses) may be strengthened by inflammation (see above). Altematively, structural reorganization of afferent terminals may alter synaptic connectivity. Since no monosynaptic Ai3 fiber-mediated EPSCs were observed after 2 days of inflammation, it seems likely that a ‘synaptic strengthening’ mechanism mediates early changes in dorsal horn synaptic transmission. However, longer-term changes (7-10 days) may be mediated by sprouting of low threshold Ab fibers into regions of the dorsal horn usually innervated exclusively by nociceptive afferents, though such new input has not yet been demonstrated morphologically. A number of inflammatory mediators and/or neurotrophic factors may contribute to the hyperreponsiveness of dorsal horn neurons following peripheral inflammation. Here we discuss two examples that may be of particular interest, prostanoids and brain derived neurotrophic factor. Prostanoids Experimentally, intrathecally administered prostaglandins generate hyperalgesia and allodynia (Uda

74

A EPSC stimulation

threshold

AP fiber EPSC latency

Naive Rat

14

10

Naive Rat

6

12 10

6

6 6 cr8 =

4

:

20

4 =u)

0

r . 20

40

A

60

60

100

6

I

0246

I,

,lp,. 6

, 10

12,

14,I!,

16

14

16

rn

t

z Inflamed Rat

147 12.

2-


95000

Peptide m-NST r-NST h-NST b-NST b-NST-BP NST-6P NST-5P

opioid receptor ‘excluding’ FGGFTGARKSARKLANQ minimal core a Concentration of 50% inhibition of [3H]Noc/OFQ binding b Dose of 50% inhibition of Noc/OFQ (50 pg/mouse)-induced

Structure of Noc/OFQ

ID50b

Sequence MPR-(21 MPR-(15 MPR-(10

1 10 a.a.)-AEPGADDAEEVEQKQLQ a.a.)-AEPVADEADEVEQKQLQ a.a.)-PEPGMEEAGEMEQKQLQ TEPGLEEVGEIEQKQLQ EIEQKQLQ EQKQLQ QKQLQ

TEPGLEEVGEIEQKQLQ minimal

(pg)

1.80 10.5 6.60 0.72 0.13 14.2 >500

core

to rat brain homogenates (Dooley and Houghten, 1996). allodynia (Okuda-Ashitaka et al., 1998; Minami et al., 1998).

and NST

Organization of preproNoc/OFQ The preproNoc/OFQ (pPNON) gene has structure and organization quite similar to those encoding precursors of the endogenous peptides preproopiomelanocortin (pPOMC), preproenkephalin (pPENK) and preprodynorphin (pPDYN). As shown in Fig. 1, pPNON generates, upon maturation, at least three bioactive peptides: NST, Noc/OFQ, and

NocII/OFQ2. The bovine, human, rat and mouse pPNONs encode proteins of 176, 176, 181 and 187 amino acid residues, respectively. While amino acid sequences of Noc/OFQ and NocII/OFQ2 are conserved beyond species, that of NST is not well conserved among species. While bovine pPNON contains Lys-Arg, a general cleavage site, at positions 109-110, the human, rat, and mouse pPNONs are devoid of the cleavage site. This may generate processed products of length 30, 35, and 41 amino-acid residues, respectively (Table 1). Al-

207

though amino acid sequences of NST are not well conserved among species and their lengths are variable, the variation in length is essentially due to the repetition of Asp-Ala-Glu-Pro-Gly-Ala, present once in human, twice in rat and three times in mouse (Mollereau et al., 1996). In addition to bovine NST, mature forms of NST with 30, 35, and 41 residues were recently identified in human, rat and mouse brains (Lie et al., 1999). The amount of NST in mouse whole brain was 6.4 pmol/g tissue, comparable to that (10.1 pmol/g tissue) of Noc/OFQ in mouse hypothalamus (Quigley et al., 1998). Brooks et al. (1998) recently reported that Noc/OFQ contents in cerebrospinal fluid (CSF) were 52.49f34.25 p&/ml and 63.39f33.26 pg/ml in patients undergoing Caesarean section and receiving spinal analgesia for established labor, respectively. As the Noc/OFQ concentration was not significantly different between these two groups, they concluded no association between Noc/OFQ concentration in CSF and acute pain of labor. Because human NSTlike immunoreactivity was also detected in CSF obtained from patients with chronic low backache and knee pain (Lie et al., 1999), comprehensive clinical studies with well-defined patients will be necessary to define the role of Noc/OFQ and NST in acute and chronic pain. Structure-activity relationship of Noc/OFQ and NST Noc/OFQ is 17 amino acids long and possesses some structural homology with the endogenous opioid peptide dynorphin A(l-17), but structurally lacks the N-terminal tyrosine essential for peptides to be active at p, K and 6 opioid receptors. Early studies with complete alanine scan analysis and truncated mutagenesis of Noc/OFQ from the N- and C-terminals showed that amino acid residues at positions l-4 together with Arg8 are the most crucial for biological activity in Noc/OFQ (Dooley and Houghten, 1996). In contrast to the complete loss of activity upon N-terminal truncation, relatively unchanged affinities were observed for C-terminal truncation analogues up to position 13, suggesting that the N-terminal half of Noc/OFQ appeared to be more important for receptor binding than the C-terminal part. Extensive structure-activity studies on Noc/OFQ revealed intriguing findings. (1)

While the substitution of Tyr’+Phe’ in dynorphin A eliminated binding activity to the K receptor, the substitution of Phe’-+Tyr’ in Noc/OFQ almost fully retained the binding activity to the Noc/OFQ receptor and in vivo hyperalgesic activity (Shimohigashi et al., 1996). (2) Tyr’-Noc/OFQ(l9) inhibited forskolin-stimulated CAMP accumulation in ic-opioid receptor-expressing CHO cells with an EC50 of 6.89 nM, but not in Noc/OFQ receptorexpressing CHO cells (Reinscheid et al., 1998). And (3) Ala8,9-Noc/OFQ and Ala12~13-Noc/OFQ were without binding activity to the Noc/OFQ receptor. These results demonstrate that amino acids at positions lo-15 as well as Phe’ may contain a domain which prevents Noc/OFQ from activating classical opioid receptors. Civelli’s group proposed the positive and negative regulatory domains in Noc/OFQ (Table 1). As with other neuropeptides, the effects of Noc/OFQ can be regulated by the degradation of the peptide. Sandin et al. (1999) showed that Noc/OFQ was step-wisely metabolized in vivo into Noc/OFQ( 1- 13) and Noc/OFQ(14- 17) followed by Noc/OFQ( l-9) and Noc/OFQ(lO-13). Consistent with structure-activity studies, Noc/OFQ( 1-13) had no effects on spatial learning and motor function. Using homogenates from human cell lines and rat cortical cells in primary culture, Vloskovska et al. (1999) showed Noc/OFQ( 1-9) and Noc/OFQ( 1-13) as two major metabolites. Because the cleavage was inhibited by dynorphin A(2-13), they suggested the involvement of dynorphin convertase, which cleaves peptide bonds positioned before, between or after basic amino acid doublets. These results also suggest that Arg’-Lys’ and Arg12-Lys13 may have dual roles in Noc/OFQ action: receptor activation and peptide degradation. As shown in Table 1, the structure-activity relationship of Noc/OFQ is in sharp contrast to that of NST in which the minimal active core resides in the C-terminal hexapeptide (EQKQLQ). All NSTs blocked Noc/OFQ-induced allodynia with ID50 values of 0.7-10.5 pg/mouse (Minami et al., 1998). Conserved C-terminal hexapeptide blocked the allodynia with an ID50 of 14.2 pg, comparable to that of human NST. The presence of activity in the conserved C-terminal of NST supports its general role in the central nervous system beyond species.

Central functions of Noc/OFQ

and NST

Knockout mice Noc/OFQ is known to have many central effects including nociception, locomotion, anxiety, neuroendocrine secretion, cognitive processes and aversive motivation. The Noc/OFQ receptor is a member of the G-protein-coupled receptor family and manifests high (>60%) sequence identity with the classical u, 6 and K opioid receptors. Although Noc/OFQ shows some structural analogy to opioid peptides, particularly dynorphin A, and acts at the molecular and cellular levels in almost the same way as the classical opioids do, it produces pharmacological effects that oppose, or at least differ from, those of opioids. Recently, knockout mice deficient in pPNON and Noc/OFQ receptor have been generated by gene targeting. The implications and the significance of the Noc/OFQ system in various physiological and pathophysiological processes are being examined and assessed. Because pharmacological roles of Noc/OFQ in central and peripheral actions have been extensively reviewed (Meunier, 1997; Darland et al., 1998), here we summarize the findings obtained with the knockout mice, and discuss the significance of these studies. Noc/OFQ knockout mice do not differ from wildtype litter mates in health, growth and reproduction, and do not have any obvious behavioral or anatomical abnormalities. Noc/OFQ injected into the hippocampus was previously shown to impair spatial learning in rats (Sandin et al., 1997) and Noc/OFQ was suggested to function as an inhibitory modulator regulating synaptic transmission and synaptic plasticity in the hippocampus (Yu et al., 1997). Interestingly, Manabe et al. (1998) showed that mice lacking the Noc/OFQ receptor have better learning ability and memory, and larger long-term potentiation in the hippocampal CA1 region than control mice. Consistent with these findings, the knockout mice showed an enhanced retention of spatial attention in the water-finding test. This gain-of-function mutation in memory and learning was suggested to be produced by an altered intracellular signal transduction system, rather than by synaptic plasticity. Noc/OFQ is also known to be involved in locomotion and nociception. While Noc/OFQ-induced hypoactivity was observed

in wild-type mice, as expected, an obvious effect was not observed by i.c.v. Noc/OFQ in Noc/OFQ receptor knockout mice. Similarly, while i.c.v. Noc/OFQ induced hyperalgesia in wild-type mice, it did not induce hyperalgesia in mutant mice in the tail-flick test. Further, studies with the mutant mice suggested the involvement of the Noc/OFQ system in the development of morphine tolerance (Ueda et al., 1997). However, there was no obvious difference in basal locomotor activity and nociceptive thresholds in four different testing paradigms (Nishi et al., 1997; Mamiya et al., 1998). Due to the absence of any obvious abnormality in nociception and locomotor activity in the mutant mice, Nishi et al. (1997) concluded that the Noc/OFQ system was not essential for the regulation of these activities. However, results with the mutant mice may just imply that fundamental behaviors such as nociception and locomotion are controlled by highly complex neural circuits mediated by many substances, including Noc/OFQ. Noc/OFQ is known to produce anxiolytic-like effects in mice and rats (Jenck et al., 1997). Consistent with pharmacological studies, pPNON knockout mice were shown to display significantly higher levels of anxiety-like behaviors than their wild-type litter mates in the open-field, elevated plus-maze, and light-dark box tests. Conversely, spatial learning was normal in pPNON-deficient mice in the Morris water-maze task (Koster et al., 1999). Because basal and post-stress concentrations of plasma corticosterone were elevated, adaptive responses to repeated stress were significantly impaired in the mutant mice. Thus, Civelli’s group concluded that Noc/OFQ is an integral constituent of the neuronal systems regulating physiological responses to stress. However, Mamiya et al. (1998) reported that the Noc/OFQ receptor was not involved in either anxiety or anti-anxiety on the basis of the findings that the Noc/OFQ receptor knockout mice did not show any behavioral changes in the same elevated plus-maze task. The discrepancy of phenotypes between the ligand and the receptor knockout mice may result from the divergent genetic background by use of different embryonic stem cell lines. Because NST antagonizes the Noc/OFQ-evoked behaviors and NST does not exert its actions through the Noc/OFQ receptor, the action of NST may explain different results between the ligand and the receptor knockout mice.

209

lntracerebroventricular Cerebrum

Inhibition analgesia

Effect

of morphine by NocIOFQ

Hyperalgesia carageenarVkaolin NodOFQ-induced of learning and

antagonism

by

inhibition impairment memory

improvement

Inhibition of sympathetic activity by Noc/OFQ

no effect

nerve

NodOFQ-induced

allodynia

antagonism

NocIOFQ-induced

hyperalgesia

antagonism

Hyperalgesla by 2% formalin NodOFQ-induced analgests

inhibition m em?&

NocIOFQevoked aggravation of pain by 1% formalin

antagonism

PGEz

inhibition

pain

responses

(500 pg/rat) (5-50

antagonism

Scopolamine-induced impairment of learning and memory

-induced

of NST a

pmol/rat)

(0.54

nmogmouse)

(0.55 (lo-100

b

nmollmouse) nmograt)

(10-1000 (&w

d

e

(50 pg/mouse) (500

c

f

pg/mouse)

f

pg/mouse)

g

(10 w/mouse) (10 pg-1

@mouse)

f

Intravenous Antidromic

vasodilation

no effect

(100

nmol/rat)

h

Fig. 2. Biological functions of NST and antagonism with Noc/OFQ. References: aZhao et al. (1999); b Nakagawa et al. (1999); ’ Hiramatsu and Inoue (1999a); d Hiramatsu and Inoue (1999b); e Shirasaka et al., 1999; f Okuda-Ashitaka et al., 1998; Minami et al., 1998; s Nakano et al., 2000; h HIbler et al., 1999.

Central roles of NST Since several papers on NST have been recently reported, we depict the roles of NST in memory and/or learning and nociception in relation to Noc/OFQ (Fig. 2). In line with previous reports (Sandin et al., 1997; Yu et al., 1997; Manabe et al., 1998), Hiramatsu and Inoue (1999a) showed that Noc/OFQ decreased spontaneous alteration in the Y-maze test and shortened the step-down latency in a passive avoidance test, demonstrating that Noc/OFQ impaired learning and/or memory in mice. NST reversed Noc/OFQ-induced alterations in behaviors. Further, NST also alleviated scopolamine-induced learning and memory impairment without affecting the acquisition of memory in normal mice (Hiramatsu and Inoue, 1999b). These results suggest that the Noc/OFQ and NST system may play opposite roles in learning and memory by regulating intracellular signal pathways subtly. Since i.c.v. injection of Noc/OFQ was initially reported to induce hyperalgesia (Meunier et al., 1995;

Reinscheid et al., 1993, a lot of attention has been paid to a possible role of Noc/OFQ in nociception. However, conflicting results have been reported by i.c.v. administration of Noc/OFQ, depending on the testing paradigm, animal species, doses, and route of administration. Although i.c.v. NST induced neither analgesia nor hyperalgesia and had no effect on morphine-induced analgesia by itself, i.c.v. NST was shown to reverse the anti-morphine effect of Noc/OFQ in rats in the tail-flick test (Zhao et al., 1999) and to exhibit analgesic effects on the inflammatory hyperalgesia induced by carrageenan/kaolin (Nakagawa et al., 1999) at doses comparable to that reported by us (Okuda-Ashitaka et al., 1998). Although i.c.v. Noc/OFQ was reported to block stressand opioid-mediated analgesia (Mogil et al., 1996), the pPNON mutant mice developed stress-induced analgesia after forced swimming (Koster et al., 1999). Unless the region specific to Noc/OFQ is targeted, the action of NST must be eliminated simultaneously with that of Noc/OFQ in the mutant mice and the interpretation of the phenotype is not straightforward.

210

On the other hand, i.c.v. administration of Noc/OFQ caused a dose-dependent decrease in blood pressure, heart rate and renal sympathetic nerve activity in rats, possibly through a central action. These hypotensive and bradycardic responses as well as inhibition of sympathetic nerve activity were not antagonized by simultaneous injection or pretreatment with lo-100 nmol of NST (Shirasaka et al., 1999). This dose was extremely high as compared with effective doses in other studies in which the effect of Noc/OFQ was antagonized by NST (Fig. 2). Because pPNON immunoreactivity and Noc/OFQ receptors are widely distributed in the brain, when interpreting these behavioral data, it should be noted that Noc/OFQ participates in many distinct circuits in many different brain regions expressing Noc/OFQ receptors and, possibly through other receptors at high doses. Formalin injection into the hind-paw induces biphasic pain behaviors: the first transient phase (O-5 min) is ascribed to the direct effect of formalin on nociceptors and the second prolonged phase (lo-30 min) is related to the development of inflammation and central sensitization. The formalin-induced licking behavior is sensitive to various classes of analgesic drugs and has been used as a reliable model of inflammatory pain. We systematically examined the effect of Noc/OFQ and NST administered i.t. on formalin-induced pain responses in mice (Nakano et al., 2000). As shown in Fig. 3A, i.t. Noc/OFQ markedly inhibited 2% formalin-induced pain at high doses of 0.3 to 3 pg, but sedation and ataxia were also observed at 3 pg. On the other hand, a low dose (10

A

aline Noc/OFQ 3/J9

s/J9

+ NST

pg) of Noc/OFQ aggravated the second phase of 1% formalin-induced pain behavior (Fig. 3B). These results are generally consistent with previous reports showing that it. administration of Noc/OFQ had inhibitory and facilitatory influence on nociception at high doses (nanogram to microgram doses) and low doses (picogram to nanogram doses), respectively (Meunier, 1997). NST significantly attenuated the second phase of 2% formalin test at 10-1000 pg, but had no effect at high doses of 0.3-3 pg. This analgesic effect was not antagonized by naloxone, demonstrating that classical opioid receptors did not participate the analgesic action of NST in the formalin test. While simultaneous administration of 10 pg NST completely reversed the aggravating effect of 10 pg Noc/OFQ on the 1% formalin-induced pain, inhibition by 3 pg of Noc/OFQ of the 2% formalin-induced pain behavior was not affected by 3 pg of NST in mice (Fig. 3). Recently we found that repeated i.t. administration of antisense oligonucleotide to Noc/OFQ receptors to mice reduced the secondphase pain responses induced by 2% formalin (K. Abe, unpublished observation). These results suggest that Noc/OFQ is involved in the second phase of formalin-induced pain behavior and that, under the protracted pain condition, NST exhibits functional antagonism against Noc/OFQ at the spinal level as well as supraspinal level. However, there seems no antagonistic relationship at high doses (nanogram to microgram doses). Yamamoto and Sakashita (1999) also reported that i.t. injection of 30 pg of NST failed to block the analgesic effect of 30 pg of Noc/OFQ in the rat formalin test.

1601

Saline

Noc/OFQ low

NST 1OW

NocJOFQ + NST

Fig. 3. Effects of Noc/OFQ and NST on formalin-induced pain. Noc/OFQ and NST (3 Kg of each for A and 10 pg of each for B) were separately or concurrently administered 10 min after formalin injection. The total time spent licking and biting the injected paw was measured during lo-30 min after subcutaneous injection of formalin into the hind-paw.

211

Noc/OFQ

Allodynia Hot plate test Formalin

e-m

test Nociception I I fs

I I

P9

I

I

I

w

I44

w

I

I

I

Allodynia Hot plate test Formalin

test

NST

Fig. 4. Dose-response of NOC/OFQ and NST for pain transmission. A range of effective doses of Noc/OFQ and NST on nociception (e) and antinociception ( ) is shown. Data for allodynia and thermal hyperalgesia in the hot-plate test by Okuda-Ashitaka et al. (1998) and the formalin test by Nakano et al. (2000).

Dose-response of Noc/OFQ and NST

Noc/OFQ administered i.t. induced thermal hyperalgesia in the hot plate test and tactile allodynia in mice and NST also blocked the Noc/OFQ-induced allodynia (Minami et al., 1998; Okuda-Ashitaka et al., 1998). NST did not affect nociception by itself. Fig. 4 summarizes the effective dose-response range of Noc/OFQ and the antagonism by NST for nociception: allodynia, hyperalgesia in the hot-plate test and pain behavior in the formalin test. While dose-dependency of Noc/OFQ for hyperalgesia was monophasic over a wide range of doses from 1 fg to 1 rig/mouse, that of Noc/OFQ-induced allodynia showed a bell-shaped pattern from 1 to 100 pg (Hara et al., 1997). NST blocked hyperalgesia and allodynia induced by Noc/OFQ in mice over the same dose range (100 fg-100 pg). NST attenuated pain behaviors in the formalin test and Noc/OFQ-evoked aggravation of formalin-induced pain at the same picogram range. NST also blocked the allodynia induced by prostaglandin E2 at 2.5-750 pg/mouse (Okuda-Ashitaka et al., 1998). These doses are attainable, judging from the amount of Noc/OFQ and NST found in the brain (Brooks et al., 1998; Lie et al., 1999). In fact, i.t. pretreatment of conscious mice with anti-NST antibody decreased the threshold of Noc/OFQ-induced allodynia by 2.5 orders and the

effective dose of Noc/OFQ was shifted from 1 pg to 10 fg, suggesting that NST blocks the induction of allodynia by Noc/OFQ under physiological conditions. On the other hand, the formalin-induced pain behaviors were blocked by 0.3-l pg of Noc/OFQ, close to a dose for inducing a motor deficit. Blockade of formalin-induced pain behaviors by a high dose of Noc/OFQ was not antagonized by lrg of NST (Fig. 3A). Binding of [‘251]Tyr’4-Noc/OFQ to the Noc/OFQ receptor was saturable and of high affinity with a Kd value of 0.1 nM (Reinscheid et al., 1995). Noc/OFQ inhibited forskolin-stimulated CAMP formation with an EC50 value of 0.9-1.0 nM and a maximal effect was observed at 100 nM (Reinscheid et al., 1995; Meunier et al., 1995). Binding of [*251]Lys14(Tyr)-NST to crude membranes of mouse brain and spinal cord was saturable and of high aflinity with a Kd value of 5 nM (Okuda-Ashitaka et al., 1998). While 2 pg = 1 nmol/lO pl (the volume employed for i.t. injection), which corresponds to 100 hM, 2 pg = 1 fmol/lO ~1 corresponds to 0.1 nM. Considering that Noc/OFQ exerts actions through the Noc/OFQ receptor, a high dose of Noc/OFQ may exert actions through a receptor different from the cloned Noc/OFQ receptor (Mathis et al., 1997).

212 Neurochemical

actions of Noc/OFQ and NST

Effect of neurotransmitter NST

release of Noc/OFQ and

At the cellular level, Noc/OFQ exerts several cellular actions common with classical opioids, such as inhibiting CAMP production, increasing inwardly rectifying K+ currents and depressing Ca2+ currents. These actions were expected to reduce neuron excitability and to modulate neurotransmitter release through the Noc/OFQ receptors on presynaptic terminals of nerve endings. Since Nicol et al. (1996) directly demonstrated that Noc/OFQ inhibited high K’-evoked glutamate release from rat cerebrocortical slices, there have been many reports on inhibition of neurotransmitter release by Noc/OFQ (for reviews, see Meunier, 1997). Table 2 summarizes recent studies on modulation of neurotransmitter release by Noc/OFQ and NST. In the brain, Noc/OFQ inhibits acetylcholine release in rat striatum (Itoh et al., 1999), dopamine release in the nucleus accumbens (Murphy and Maidment, 1999), electrically-stimulated release of noradrenaline from mouse brain slices (Werthwein TABLE

Pharmacological studies of NoclOFQ-induced at the spinal level

pain

Compared with the brain, neural circuits for pain transmission seem to be more simple in the spinal cord and pPNON and Noc/OFQ receptors are densely expressed in the superficial dorsal horn where fine primary afferent nerve fibers terminate

2

Modulation

of neurotransmitter

Transmitter peptide

or

NoclOFQ Brain ACh = Dopamine

NAa 5-HTa Periphery ACh Substance

=

release

P

by Noc/OFQ

and NST

Stimulus

Animal

Region

Route

Dose

basal

basal high K basal ESa ESa

rat rat rat rat rat rat mouse rat

striatum Acb a striatum VTAa cortex slice VTA a cortex slice cortex slice

microdialysis microdialysis microdialysis microdialysis in vitro microdialysis in vitro in vitro

10 uM 0.1-l l-10 0.1-I 51 nMb 0.1-l 30nMb 288 nM

ES” ES a ES=

guinea rat rat

trachea trachea trachea

in vitro in vitro in vitro

100 nM 100 nM

1

Pate1 et al. (1997) Helyes et al. (1997) Helyes et al. (1997)

in vitro

100 nM

h

Nicol

basal

GABA a Glutamate

CGRP

et al., 1999) and serotonin release from rat cerebral cortex slices (Siniscalchi et al., 1999). Recently, Noc/OFQ was reported to increase dopamine release from the striatum (Konya et al., 1998) and GABA and glutamate release from the ventral tegmental area (Murphy and Maidment, 1999). The facilitatory effects of Noc/OFQ might be explained by the disinhibition of inhibitory neurotransmitters through Noc/OFQ receptors located on interneurons. Nicol et al. (1998) also showed that NST antagonized the inhibition by Noc/OFQ of high K+-evoked glutamate release. In agreement with wide distribution of pPNON, it is clear that Noc/OFQ and NST are capable of modulating a large variety of neurons and neural circuits in the brain.

pig

Effect

Reference

u IJ 9 fi

Itoh et al. (1999) Murphy and Maidment (1999) Konya et al. (1998) Murphy and Maidment ( 1999) Nicol et al. (1996) Murphy and Maidment (1999) Werthwein et al. (1999) Siniscalchi et al. (1999)

mM uM mM

u 9 IT u

mM b

0.1-3 uM

IJ

NST Glutamate

inhibition Noc/OFQ

a ACh, acetylcholine; electrical stimulation; b IC50 value.

by

rat

GABA, y-aminobutyric Acb, nucleus accumbens;

cortex

slice

acid; NA, noradrenaline; VTA, ventral tegmental

5-HT, area.

serotonin;

CGRP,

calcitonin

et al. (1998)

gene-related

peptide;

ES,

213

Nociceptin + D-AP5 + GAMS + CP 99,994 Saline 10

9

6

-log[i.t. Glutamate

7

6

5

antagonist

4

0

10

Response

(g)]

20

Time(s)

Nociceptin

+ Glycine + Muscimol + Baclofen Saline IO

9

-log[i.t.

a

7

6

Agent (g)]

0

20

10

Response

Time(s)

of neurotransmitters in NOC/OFQ-induced allodynia (A, C) and hyperalgesia (B, D). (A, C) Noc/OFQ (50 pg) was Fig. 5. Involvement injected simultaneously with various doses of agents into the subarachnoid space of conscious mice. Allodynia was assessed 10 min after i.t. injection. (B, D) Noc/OFQ (50 pg) was injected simultaneously with agents into the subarachnoid space of conscious mice. Hyperalgesia was assessedby response on the hot plate 15 min after i.t. injection. Agents and doses employed are as follows: glycine, 1 kg; muscimol, 0.1 bg; baclofen, 10 ng; D-APS, 0.25 kg; GAMS, 1 kg; and CP 99,994, 10 pg.

and incoming somatosensory information is transferred to secondary neurons. Therefore, studies on mechanism of action of Noc/OFQ and NST in the spinal cord may offer an advantage over those in the brain. Activation of primary afferent C-fibers gives rise to spinal release of substance P and glutamate and these mediators facilitate the cascade of nociceptive processing. As shown in Fig. 5A, Noc/OFQ-induced allodynia was dose-dependently blocked by D-AP5 (an NMDA-receptor antagonist) and by GAMS (a non-NMDA-receptor antagonist), but not by L-AP4 (an mGluR antagonist) or CP 99,994 (an NK1 tachykinin receptor antagonist). On the other hand, Noc/OFQ-induced hyperalgesia was blocked by CP 99,994, but not by D-AP5, GAMS, or L-AP4. As shown in Fig. 5C,D, both Noc/OFQ-induced allodynia and hyperalgesia were

blocked by glycine, but not affected by the GABA* receptor agonist muscimol or the GABAa receptor agonist baclofen. Noc/OFQ-induced hyperalgesia and allodynia were abolished by neonatal capsaicin treatment. As depicted in Fig. 6, we postulate that Noc/OFQ may initiate a common biochemical event beginning with disinhibition of the inhibitory glycinergic response, probably inhibition of glycine release from intemeurons. Decrease in glycine may in turn stimulate the release of substance P and glutamate from the nerve endings of C-fibers or increase in excitability of postsynaptic neurons, leading to hyperalgesia and allodynia. Allodynia is considered to result from an expression of neural plasticity. At present, two models are proposed for the mechanism of allodynia accompanied with acute inflammation and injury: the first model focuses on increased postsy-

214

Primary afferent C fibers I (Substance

Pt

~~~

Glutamate Presynaptic

?) terminal

ganglion send fibers to the periphery and the superficial layers of dorsal horn, this conclusion is consistent with our hypothesis for hyperalgesia (Fig. 6). As shown in Table 2, however, Noc/OFQ inhibited the discharge of acetylcholine and the sensory neuropeptides substance P and calcitonin gene-related peptide from the stimulated sensory nerve endings and plasma extravasation (Helyes et al., 1997; Pate1 et al., 1997). The mechanism of the facilitatory effect by Noc/OFQ in the periphery remains unknown. Localization

Hyperalgesia

Allodynia Postsynaptic

neuron

Fig. 6. Possible pathways involved in Noc/OFQ-induced algesia and allodynia in the spinal cord.

hyper-

naptic neuron excitability and responsiveness, known as the sensitization of wide dynamic range neurons in the deeper laminae (Woolf, 1994; see also Moore et al., 2000, this volume), and the second model focuses on presynaptic interactions between lowthreshold mechanoreceptors and C-fibers through secondary interneurons (Cervero and Laird, 1996). Elucidation of mechanism(s) of Noc/OFQ-induced allodynia and a pathway(s) conveying somatosensory information from the periphery needs further investigation. Consistent with our previous results (Hara et al., 1997), it has recently been reported that i.t. injection of picogram doses of Noc/OFQ produced hyperalgesia in the tail-flick test through NKi receptors (Sakurada et al., 1999). NMDA receptor antagonists D-AI-3 and MK-801 and the NO synthase inhibitor L-NAME failed to block the hyperalgesia. In this connection, intraplantar injection of Noc/OFQ also elicited behavioral responses in picogram doses in a flexor-reflex paradigm (Inoue et al., 1998). Interestingly, these pain responses were blocked by NKi receptor antagonists, capsaicin pretreatment or in tachykinin gene knockout mice, suggesting that Noc/OFQ indirectly stimulates nerve endings of nociceptive primary afferent neurons through substance P release. Since small-size neurons in the dorsal root

Recent studies have characterized in detail the localization of pPNON mRNA in mouse brain (Ikeda et al., 1998; Boom et al., 1999), pPNON mRNA and immunoactivity (Neal et al., 1999a) and Noc/OFQ receptor mRNA in mouse brain (Ikeda et al., 1998) and Noc/OFQ receptor mRNA and Noc/OFQ binding in rat brain (Neal et al., 1999b). The distribution of pPNON mRNA is closely correlated with that of Noc/OFQ receptor. Both pPNON and Noc/OFQ receptors are distributed in many brain regions including the telencephalon, diencephalon, midbrain, pons, and medulla oblongata, but not cerebellum. The distribution of NST immunoreactivity is abundant in spinal trigeminal tract, paramedian raphe nucleus and ventromedial nucleus of hypothalamus and appears to be essentially identical to that of Noc/OFQ (Okuda-Ashitaka, E., unpublished data). The wide distribution of pPNON and Noc/OFQ receptors in the brain supports a broad range of biological actions and implies the antagonism of Noc/OFQ and NST in diverse regions in the central nervous system. Since relationship between the distribution and function of the Noc/OFQ system has been extensively described (Neal et al., 1999a), here we only describe the distribution of Noc/OFQ, NST and Noc/OFQ receptors in relation to spinal actions. Ikeda et al. (1998) showed that pPNON mRNA was highly expressed in spinal cord laminae I, II, and IX; the signal was expressed in nearly a half of the neurons in lamina I and in a minor population in lamina II. On the other hand, Neal et al. (1999a) reported that lamina I was devoid of pPNON mRNA-containing neurons, but that lamina II had heavy expression. In the thoracic spinal cord, Noc/OFQ immunoreactivity was modestly detected in fibers and terminals

215

in laminae I and III with moderate numbers of dark neurons in lamina II. NST immunoreactivity was most abundant in the superficial laminae of mouse spinal dorsal horn (Okuda-Ashitaka et al., 1998). Furthermore, while Ikeda et al. (1998) reported that Noc/OFQ receptor mRNA expression was weak, but present in almost all regions of the spinal cord, Neal et al. (1999b) reported that its expression was higher in the ventral horn than in the dorsal horn. Thus, there is marked discrepancy in the distribution of pPNON and Noc/OFQ receptors in the spinal cord. In the latter study (Neal et al., 1999b), while lamina I contained no n-RNA-expressing neurons and laminae II and III contained only scattered mRNAexpressing cells, the dorsal root ganglion was filled with numerous, large mRNA-expressing cells. In contrast to mRNA expression, Noc/OFQ binding sites were dense in laminae II-IV, but not in lamina I or dorsal root ganglion. This mismatch may result from receptor trafficking; Noc/OFQ receptors expressed in the deeper layers of the dorsal horn are probably postsynaptic and those in the superficial laminae probably reside on presynaptic fibers originating from the dorsal root ganglion (Neal et al., 1999b). If Noc/OFQ stimulates nerve endings of nociceptive primary afferent neurons through substance P release, localization of the Noc/OFQ receptor mRNA in large-size ganglion cells is not in agreement with the notion that substance P-containing fibers originate from capsaicin-sensitive, small-size neurons in dorsal root ganglia. In addition, even though pPNON mRNA was not expressed in the dorsal root ganglia of naive animals, peripheral inflammation markedly induced its expression (Andoh et al., 1997). [3H]Noc/OFQ binding also increased in laminae I and II, but not lamina X after inflammation (Jia et al., 1998). The disparity of pPNON and Noc/OFQ receptor localization among researchers and alterations in their expression under pathophysiological conditions have limited our ability to define accurately the relevant circuitry. Interestingly, Monteillet-Agius et al. (1998) demonstrated that u-opioid and Noc/OFQ receptors were expressed predominantly on different fibers in areas involved in pain processing, including the spinal cord. The different cellular expression and organization of the Noc/OFQ and opioid systems support their distinct actions in pain transmission at spinal level.

Conclusions Given the existence of Noc/OFQ and NST, which play opposite roles in central actions, the perceptual processes mediated by the Noc/OFQ system seem to be complicated by the complex and efficacious neuronal circuits by which the control of afferent processing occurs. Noc/OFQ and NST are considered to exert their activity mediated by respective receptors at specific synaptic junctions as do other neuropeptides. Identification of the conditions in which the peptides are released, cloning of the NST receptor and development of selective antagonists for Noc/OFQ and NST receptors will be crucial to understand the Noc/OFQ system completely. This may clarify the circuitry on which they are acting and solve the discrepancy among researchers. The multiple functions of Noc/OFQ and NST produced from the same precursor should provide new insights into brain physiology, such as neural plasticity, and may lead to therapeutic applications to the management of pain and neural disorders. Abbreviations Noc/OFQ NST pPNON CSF

nociceptin/orphanin nocistatin preproNoc/OFQ cerebrospinal fluid

FQ

Acknowledgements This work was supported in part by Grants-in-Aids for Scientific Research on Priority Areas, Scientific Research (B) (11470044, 11470329, 11558093) and (C) (10671450, 11671589) from the Ministry of Education, Science, Sports and Culture of Japan or Japan Society for the Promotion of Science, and by grants from the Science Research Promotion Fund of the Japan Private School Promotion Foundation and Naito Foundation. References Andoh, T., Itoh, M. and Kuraishi, Y. (1997) Nociceptin gene expression in rat dorsal root ganglion induced by peripheral inflammation. NeuroReport, 8: 2793-2796. Bennett, G.J. (1994) Neuropathic Pain. In: P.D. Wall and R.

216

Melzack (Eds.), The Textbook of Pain. Textbook of Pain, 3rd Edn. Churchill Livingstone, Edinburgh, pp. 201-224. Boom, A., Mollereau, C., Meunier, J.-C., Parmentier, M., Vanderhaegen, J.-J. and Schiffmann, S.N. (1999) Distribution of the nociceptin and nocistatin precursor transcript in the mouse central nervous system. Neuroscience, 91: 991-1007. Brooks, H., Elton, CD., Smart, D., Rowbotham, D.J., M&night, A.T. and Lambert, D.G. (1998) Identification of nociceptin in human cerebrospinal fluid comparison of levels in pain and non-pain states. Pain, 78: 71-73. Cervero, E and Laird, J.M.A. (1996) Mechanisms of touchevoked pain (allodynia): a new model. Pain, 68: 12-23. Darland, T., Heinricher, M.M. and Grandy, D.K. (1998) Orphanin FQ/nociceptin: a role in pain and analgesia, but so much more. Trends Neurosci., 21: 215-221. Dooley, C.T. and Houghten, R.A. (1996) Orphanin FQ: receptor binding and analog structure activity relationships in rat brain. Life Sci., 59: 23-29. Dray, A., Urban, L. and Dickenson, A. (1994) Pharmacology of chronic pain. Trends Pharrnacol. Sci., 15: 90-197. Florin, S., Suaudeau, C., Meunier, J.-C. and Costentin, J. (1997) Orphan neuropeptide NocII, putative pronociceptin maturation product, stimulates locomotion in mice. NemoReport, 8: 705707. Habler, H.-J., Timmermann, L., Stegmann, J.-U. and Janig, W. (1999) Effects of nociceptin and nocistatin on antidromic vasodilatation in hairless skin of the rat hindlimb in vivo. BI: J. Pharmacol., 127: 1719-1727. Helyes, Z., Nemeth, Z., Pint&, E. and Szolcsanyi, J. (1997) Inhibition by nociceptin of neurogenic inflammation and the release of SP and CGRP from sensory nerve terminals. Bi: J. Pharmacol., 121: 613-615. Hara, N., Minami, T., Okuda-Ashitaka, E., Sugimoto, T., Sakai, M., Onaka, M., Mori, H., Imanishi, T., Singu, K. and Ito, S. (I 997) Characterization of nociceptin hyperalgesia and allodynia in conscious mice. BI: J. Pharmacol., 121: 401-408. Henderson, G. and M&night, A.T. (1997) The orphan opioid receptor and its endogenous ligand-nociceptin/orphanin FQ. Trends Pharmacol. Sci., 18: 293-300. Hiramatsu, M. and Inoue, K. (1999a) Effects of nocistatin on nociceptin-induced impairment of learning and memory in mice. Eur J. Phamzacol., 367: 151-155. Hiramatsu, M. and moue, K. (1999b) Nociceptin/orphanin FQ and nocistatin on learning and memory impairment induced by scopolamine in mice. BI: J. Pharmacol., 127: 655-660. Ikeda, K., Watanabe, M., Ichikawa, T., Kobayashi, T., Yano, R. and Kumanishi, T. (1998) Distribution of prepro-nociceptin/orphanin FQ mRNA and its receptor mRNA in developing and adult mouse central nervous systems. J. Comp. Neural., 399: 139-151. moue, M., Kobayashi, M., Kozaki, S., Zimmer, A. and Ueda, H. (1998) Nociceptin/orphanin FQ-induced nociceptive responses through substance P release from peripheral nerve endings in mice. Proc. Natl. Acad. Sci. U.S.A., 95: 10949-10953. Itoh, K., Konya, H., Takai, E., Masuda, H. and Nagai, K. (1999) Modification of acetylcholine release by nociceptin in conscious rat striatum. Brain Res., 845: 242-245.

Jenck, F., Moreau, J.-L., Martin, J.R., Kilpatrick, G.J., Reinscheid, R.K., Monsma Jr., F.J., Nothacker, H.-P and Civelli, 0. (1997) Orphanin FQ acts as an anxiolytic to attenuate behavioral responses to stress. Proc. Natl. Acad. Sci. U.S.A., 94: 14854-14858. Jia, Y.-P., Linden, D.R., Serie, J.R. and Seybold, VS. (1998) Nociceptin/orphanin FQ binding increases in superficial laminae of the rat spinal cord during persistent peripheral inflammation. Neurosci. Lett., 250: 21-24. Konya, H., Masuda, H., Itoh, K., Nagai, K., Kakishita, E. and Matsuoka, A. (1998) Modification of dopamine release by nociceptin in conscious rat striatum. Brain Res., 788: 341344. Koster, A., Montkowski, A., Schultz, S., Stube, E.M., Kuaudt, K., Jenck, F., Moreau, J.L., Nothacker, H.P., Civelli, 0. and Reinscheid, R.K. (1999) Targeted disruption of the orphan FQ/nociceptin gene increases stress susceptibility and impairs stress adaptation in mice. Proc. Natl. Acad. Sci. U.S.A., 96: 10444-10449. Lie, T.-L., Fung, F.M.Y., Chen, E-G., Chou, N., Okuda-Ashitaka, E., Ito, S., Nishiuchi, Y., Kimura, T. and Tachibana, S. (1999) Identification of human, rat and mouse nocistatin in brain and human cerebrospinal fluid. NeuroReport, 10: 1537-1541. Mamiya, T., Noda, Y., Nishi, M., Takeshima, H. and Nabeshima, T. (1998) Enhancement of spatial attention in nociceptin/orphanin FQ receptor-knockout mice. Bruin Res., 783: 236-240. Manabe, T., Noda, Y., Mamiya, T., Katagiri, H., Houtani, T., Nishi, M., Noda, T., Takahashi, T., Sugimoto, T., Nabeshima, T. and Takeshima, H. (1998) Facilitation of long-term potentiation and memory in mice lacking nociceptin receptors. Nature, 394: 577-581. Mathis, J.P., RyanMoro, J., Chang, A., Horn, J.S.D., Scheinberg, D.A. and Pastemak, G.W. (1997) Biochemical evidence for orphanin FQ/nociceptin receptor heterogeneity in mouse brain. Biochem. Biophys. Res. Commun., 230: 462-465. Meunier, J.-C., Mollereau, C., Toll, L., Suaudeau, C., Moisand, C., Alvinerie, P., Butour, J.-L., Guillemot, J.-C., Ferrara, P., Monsarrat, B., Mazarquil, H., Vassart, G., Parmentier, M. and Costentin, J. (1995) Isolation and structure of the endogenous agonist for opioid receptor-like ORLi receptor. Nature, 377: 532-535. Meunier, J.-C. (1997) Nociceptin/orphanin FQ and the opioid receptor-like ORLl receptor. Eur: J. Phannacol., 340: l-15. Minami, T., Okuda-Ashitaka, E., Nishiuchi, Y., Kimura, T., Tachibana, S., Mori, H. and Ito, S. (1998) Anti-nociceptive responses produced by human putative counterpart of nocistatin. Br J. Pharmacol., 124: 1016-1018. Mogil, J.S., Grisel, J.E., Reinscheid, R.K., Civelli, O., Belknap, J.K. and Grandy, D.K. (1996) Orphanin FQ is a functional anti-opioid peptide. Neuroscience, 75: 333-337. Mollereau, C., Simons, M.-J., Soularue, P., Liners, F., Vassart, G., Meunier, J.-C. and Palmentier, M. (1996) Structure, tissue distribution, and chromosomal localization of the prepronociceptin gene. Proc. Natl. Acad. Sci. U.S.A., 93: 8666-8670. Monteillet-Agius, G., Fein, J., Anton, B. and Evans, C.J. (1998) ORLl and mu opioid receptor antisera label different fibers

217 in areas involved in pain processing. J. Camp. Neural., 399: 313-383. Moore, K.A., Baba, H. and Woolf, C.J. (2000) Synaptic transmission and plasticity in the superficial dorsal horn. In: J. Sandkiihler, B. Bromm and G.F. Gebhart (Eds.), Nervous System Plasticity and Chronic Pain, Progress in Brain Research, Vol. 129. Elsevier. Amsterdam, pp. 63-80. Murphy, N.P. and Maidment, N.T. (1999) Orphanin FQ/nociceptin modulation of mesolimbic dopamine transmission determined by microdialysis. J. Neurochem., 13: 179186. Nakagawa, T., Kaneko, M., Imamura, S. and Satoh, M. (1999) Intracerebroventricular administration of nocistatin reduces inflammatory hyperalgesia in rats. Neurosci. Lett., 265: 64-66. Nakano, H., Minami, T., Abe, K., Arai, T., Tokumura, M., Ibii, N., Okuda-Ashitaka, E., Mori, H. and Ito, S. (2000) Effect of intrathecal nocistatin on the formalin-induced pain in mice versus that of nociceptin/orphanin FQ. J. Pharmacol. Exp. Thee, 292: 331-336. Neal Jr., C.R., Mansour, A., Reinscheid, R., Nothacker, H.-P., Civelli, 0. and Watson Jr., S.J. (1999a) Localization of orphanin FQ (nociceptin) peptide and messenger RNA in the central nervous system of the rat. J. Comp. Neural., 406: 503541. Neal Jr., C.R., Mansour, A., Reinscheid, R., Nothacker, H.-P., Civelli, O., Akil, H. and Watson Jr., S.J. (1999b) Opioid receptor-like (ORLl) receptor distribution in the rat central nervous system: comparison of ORLl receptor mRNA expression with ‘251-[‘4Tyr]orphanin FQ binding. J. Comp. Neural., 412: 563-605. Neumann, S., Doubell, T.P., Leslie, T. and Woolf, C.J. (1996) Inflammatory pain hypersensitivity mediated by phenotypic switch in myelinated primary sensory neurons. Nature, 384: 360-364. Nicol, B., Lambert, D.G., Rowbothman, D.J., Smart, D. and M&night, A.T. (1996) Nociceptin induced inhibition of K+ evoked glutamate release from rat cerebrocortical slices. BK J. Pharmacol., 119: 1081-1083. Nicol, B., Lambert, D.G., Rowbotbman, D.J., Okuda-Ashitaka, E., Ito, S., Smart, D. and M&night, A.T. (1998) Nocistatin reverses nociceptin inhibition of glutamate release from rat brain slices. Eur: J. Pharmacol., 356: Rl-R3. Nishi, M., Houtani, T., Noda, Y., Mamiya, T., Sato, K., Doi, T., Kuno, J., Takeshima, H., Nukada, T., Nabeshima, T., Yamashita, T., Nosa, T. and Sugimoto, T. (1997) Unstrained nociceptive response and disregulation of hearing ability in mice lacking the nociceptin/orphanin FQ receptor. EMBO J., 16: 1858-1864. Okuda-Ashitaka, E., Tachibana, S., Houtani, T., Minami, T., Masu, Y., Nishi, M., Takeshima, H., Sugimoto, T. and Ito, S. (1996) Identification and characterization of an endogenous ligand for opioid receptor homologue ROR-C: its involvement in allodynic response to innocuous stimulus. Mol. Brain Res., 43: 96-104. Okuda-Ashitaka, E., Minami, T., Tachibana, S., Yoshihara, Y., Nishiuchi, Y., Kimura, T. and Ito, S. (1998) Nocistatin, a

peptide that blocks nociceptin action in pain transmission. Nature, 392: 286-289. Olson, G.A., Olson, R.D. and Kastin, A.J. (1995) Endogenous peptides. Peptides, 16: 1517-1555. Patel, H.J., Giembycz, M.A., Spicuzza, L., Barnes, P.J. and Belvisi, M.G. (1997) Naloxone-insensitive inhibition of acetylcholine release from parasynaptic nerves innervating guineapig trachea by the novel opioid, nociceptin. BK J. Pharmacol., 120: 735-736. Quigley, D.I., McDougall, J., Darland, T., Zhang, G., Ronnekliev, O., Grandy, D.K. and Allen, R.G. (1998) Orphanin FQ is the major OFQl-17-containing peptide produced in the rodent and monkey hypothalamus. Peptides, 19: 133- 139. Reinscheid, R.K., Nothacker, H.-P., Bourson, A., Ardati, A., Henningsen, R.A., Bunzow, J.R., Grandy, D.K., Langen, H., Monsma Jr., F.J. and Civelli, 0. (1995) Orphanin FQ: a neuropeptide that activates an opioidlike G protein-coupled receptor. Science, 210: 792-194. Reinscheid, R.K., Higelin, J., Henningsen, R.A., Monsma Jr., P.J. and Civelli, 0. (1998) Structures that delineate orphanin FQ and dynorphin A pharmacological selectivities. J. Biol. Chem., 273: 1490-1495. Rossi, G.C., Mathis, J.P. and Pasternak, G.W. (1998) Analgesic activity of orphanin FQ2, murine prepro-orphanin FQ141-157 in mice. NeuroReport, 9: 1165-l 168. Sakurada, C., Sakurada, S., Katsuyama, S., Sasaki, J., Tan-No, K. and Sakurada, T. (1999) Involvement of tachykinin NK1 receptors in nociceptin-induced hyperalgesia in mice. Brain Rex, 841: 85-92. Sandin, J., Georgieva, J., Schott, PA., Ggren, S.O. and Terenius, L. (1997) Nociceptin/orphanin FQ microinjected into hippocampus impairs spatial learning in rats. Eur J. Neurosci., 9: 194-197. Sandin, J., Georgieva, J.: Silberring, J. and Terenius, L. (1999) In vivo metabolism of nociceptin/orphanin FQ in rat hippocampus. NemoReport, 10: 11-16. Shimohigashi, Y., Hatano, R., Fujita, T., Nakashima, R., Nose, T., Sujaku, T., Saigo, A., Shinjo, K. and Nagahisa, A. (1996) Sensitivity of opioid receptor ORLl for chemical modification on nociception, a naturally occurring nociceptive peptide. J. Biol. Chem., 211: 23642-23645. Shirasaka, T., Kunitake, T., Kato, K., Takasaki, M. and Kannan, H. (1999) Nociceptin modulates renal sympathetic nerve activity through a central action in conscious rats. Am. J. Physiol., 277: R1025-1032. Siniscalchi, A., Rodi, D., Beani, L. and Bianchi, C. (1999) Inhibitory effect of nociceptin on [3H]-5-HT release from rat cerebral cortex slices. Br J. Pharmacol., 128: 119-123. Ueda, H., Yamaguchi, T., Tokuyama, S., Inoue, M., Nishi, M. and Tekeshima, H. (1997) Partial loss of tolerance liability to morphine alangesia in mice lacking the nociceptin receptor gene. Neurosci. L&t., 231: 36-138. Vloskovska, M., Kasakov, L., Suder, P., Silberring, J. and Terenius, L. (1999) Biotransformation of nociceptin/orphanin FQ by enzyme activity from morphine-naive and morphine-treated cell cultures. Brain Res., 818: 213-220. Werthwein, S., Bauser, U., Nakazi, M., Kathman, M. and

218 Schlicker, E. (1999) Further characterization of the ORLt receptor-mediated inhibition of noradrenaline release in the mouse brain in vitro. BI: J. Pharmacol., 127: 300-308. Woolf, C.J., Shortland, P. and Coggeshall, R.E. (1992) Peripheral nerve injury triggers central sprouting of myelinated afferents. Nature, 355: 75-78. Woolf, C.J. (1994) The dorsal horn: state-dependent sensory processing and the generation of pain. In: PD. Wall and R. Melzack (Eds.), Textbook of Pain, 3rd Edn. Churchill Livingstone, Edinburgh, pp. 101-l 12.

Yamamoto, T. and Sakashita, Y. (1999) Effect of nocistatin and its interaction with nociceptin/orphanin FQ on the rat formalin test. Neurosci. Lett., 262: 179-182. Yu, T.P., Fein, J., Phan, T., Evans, C.J. and Xie, C.W. (1997) Orphanin FQ inhibits synaptic transmission and long-term potentiation in rat hippocampus. Hippocampus, 7: 88-94. Zhao, C.-S., Li, B.-S., Zhao, G-Y., Liu, H.-X., Luo, F., Wang, Y., Tian, J.-H., Chang, J.-K. and Han, J.-S. (1999) Nocistatin reverses the effect of orphanin FQ/nociceptin in antagonizing morphine analgesia. NeuroReport, 10: 297-299.

J. Sandkiihler, B. Bromm and GE Gebhart (Eds.) Progress in Brain Research, Vol. 129 0 2000 Elsevier Science B.V. All rights reserved

CHAPTER 15

The biological role of galanin in normal and neuropathic states Bradley J. Kerr ‘T*,David Wynick 2, StephenW.N. Thompson 1 and StephenB. McMahon ’ ’ Neuroscience

Research

Centre,

2 Department

Guy’s,

of Medicine,

King’s

and St Thomas’ School of Biomedical Sciences, King’s College London SE1 7EH, UK University of Bristol, Marlboro Street, Bristol B52 8HW UK

Introduction Injuries to peripheral nerves in man are associated with a range of debilitating consequences, the most severe of which is the emergence of chronic neuropathic pain. The etiology of such pain is poorly understood and existing clinical treatment is largely ineffective. There have been intense efforts over the last decade to characterize the changes that take place in the somatosensory system, particularly in primary sensory neurons, in a variety of animal models of nerve damage. It is now clear that many aspects of primary sensory neuron function are altered following injury or disease of peripheral axons. One important change is a mobilization of repair mechanisms within the damaged neurons, such as an up-regulation of a number of structural proteins. This regenerative response may be adaptive in the peripheral axon, but it also seems to affect the anatomy of central terminations of damaged sensory afferents in the dorsal horn, and here it may be maladaptive (reviewed in McMahon and Bennett, 1999). There are * Corresponding author: B.J. Kerr, Neuroscience Research Centre, Division of Physiology, St. Thomas’ Hospital Campus, Lambeth Palace Road, London SE1 7EH, UK. Tel.: +44 (171) 928-9292, ext. 2241; Fax: +44 (171) 928-0729; E-mail: [email protected]

London,

also changes that are related to sensory transduction in damaged neurons. For instance, the expression of a number of receptors (e.g. VRl, B2, P2Xs) is reduced. Ion channel expression, notably several types of sodium channels are also up- or down-regulated, and these changes may contribute to abnormal sensory states in the form of altered afferent excitability and the generation of ectopic discharges (reviewed in Devor and Seltzer, 1999). But there is another class of change that affects damaged neurons, an alteration in the expression and central release of neurotransmitters/modulators. Several neuropeptides which are constitutively expressed by sub-groups of sensory neurons (e.g. SP and CGRP) are down-regulated after peripheral injury. Others, however, are up-regulated or induced de novo. This review will focus one of these peptides, gala& the levels of which increase 120-fold in rat DRG after peripheral axotomy (Villar et al., 1989; for review see Hokfelt et al., 1994). Although this fact has been appreciated for more than a decade, there are still many fundamental questions that remain unanswered regarding its functional significance. It has been proposed to be an antinociceptive neuropeptide involved in sensory transmission (Xu and Wiesenfeld-Hallin, 1997) and that its up-regulation following nerve injury acts as a compensatory mechanism to dampen the increased excitability which ensues following nerve damage (Villar et al., 1989).

220

ThermalSensitivity

A 8.0

7.0

assumed, revealing interelated inhibitory and excitatory effects. What is the physiological role for galanin in naive animals?

6.0 5.0 4.0 3.0 2.0 1.0 0

+/+

-I-

MechanicalSensitivity 0.40 0.35 0.30

0.25 0.u) 0.15

The 29-amino acid neuropeptide galanin was first isolated from porcine intestine (Tatemoto et al., 1983) and has since been shown to be widely distributed throughout the central nervous system (Ch’ng et al., 1985; Melander et al., 1986). Galanin is cleaved from a 123-amino acid prepropeptide along with a 59-amino acid sequence known as galanin message associated peptide (GMAP) neither of which are homologous with any other known peptides (Crawley, 1996). Galanin is expressed at high levels in several sensory nervous tissues (trigeminal and dorsal root ganglia) as early as embryonic day 14 continuing until postnatal day 1 when levels in DRG and trigeminal ganglia are down-regulated (Xu et al., 1996a). In the adult, galanin is constitutively expressed by less than 5% of the primary sensory neurons in the adult dorsal root ganglion (DRG). These cells are predom-

0.10 0.05

Fig. I. Sensitivity of wild-type and galanin knock-out mice to threshold noxious stimuli. (A) Galanin knock-out mice (filled bar, n = 15) show a mild, though significant, decrease in noxious thermal withdrawal thresholds compared to wild-type mice (dashed bar, n = 15), (P < 0.05, t-test). (B) Mechanical withdrawal thresholds are also significantly reduced in galanin knockout mice (filled bar, n = 15) compared to wild-type controls (dashed bar, II = 15), (P < 0.05, Mann-Whitney, rank sum test).

Evidence now suggests that galanin may act in a more complex manner after nerve injury than as a purely inhibitory neuropeptide (Ma and Bisby, 1997; Kerr et al., 2000). Here we will focus on galanin’s role in nociceptive sensory function and discuss our recent data using pharmacological approaches as well as a mouse strain carrying a null mutation in the galanin gene. Our work suggests that central galaninergic mechanisms may be more complicated than previously

Fig. 2. The effect of chronic intrathecal galanin (25 rig/h) on thermal and mechanical sensitivity in the naive rat. (A) Neither galanin (filled circles, n = 8) nor saline (open squares, n = 10) had any effect on thermal withdrawal latencies compared to baseline control values (P > 0.05, one-way repeated measures ANOVA on ranks). (B) Vehicle treatment (open squares, n = 10) had no significant effect on mechanical withdrawal threshold compared to baseline control values (P > 0.05, repeated measures ANOVA on ranks). Galanin-treated rats showed a significant drop in mechanical thresholds from test day 2 onwards (P < 0.05, one-way repeated measures ANOVA on ranks, Dunnett’s post hoc test). Asterisks (*) indicate significant differences in mechanical thresholds between groups, (P < 0.05, MannWhitney rank sum test). (C-E) The effect of chronic intrathecal galanin (25 rig/h) on c-fos expression in the dorsal horn of the spinal cord. (C) Photomicrograph showing the absence of c-jbs-immunoreactive nuclei in saline-treated rat and (D) the dorsal horn of a galanin-treated rat with c-@-positive nuclei; note the predominance of c-fos immunoreactivity in deeper lamina. (E) Mean c-@-positive cell counts for saline-treated (dashed bar, n = 8) and galanin-treated (filled bar, n = 8) rats in superficial and deep dorsal horn lamina. Galanin treatment resulted in a significant increase in c-fos immunoreactivity in both superficial and deep lamina, (*P < 0.05, t-test).

221

A

g 2 .-g f ;0

B

160 140 120 100 80 60 40

120

z

100 \

i2 E

80

.g 60 0 saline

20

-j

40

s

20 0

0

1 6 8 12 Post surgeryDay

14

0

1

6

8

12

PostsurgeryDay

I

Superficial Deep Lamina

n

IT-Galanin

14

222

inantly small diameter nociceptors with unmyelinated, slowly conducting axons that co-express the neuropeptide calcitonin gene related peptide (CGRP) (Villar et al., 1989). At the spinal level, a network of galanin immunoreactive afferent fiber terminals can be found in the superficial laminae (I and II) of the dorsal horn (Zhang et al., 1993). Galanin is also expressed in a population of intrinsic interneurons within the superficial lamina of the dorsal horn. A majority of these cells co-express the inhibitory neurotransmitter y-amino butyric acid (GABA) as well as the inhibitory neuropeptide enkephalin (Simmons et al., 1995; Zhang et al., 1995). In response to nerve injury, levels of galanin mRNA and peptide are rapidly up-regulated in the DRG and within the afferent terminals of the spinal cord (H&felt et al., 1987; Villar et al., 1989, 1991; Ma and Bisby, 1997) with little change in expression within the intrinsic intemeuronal population in the spinal cord (Zhang et al., 1995). To date, three receptor subtypes for galanin have been isolated and cloned, (rGALR1, Parker et al., 1995; Burgevin et al., 1996; rGALR2, Howard et al., 1997; Smith et al., 1997; Wang et al., 1997a; Fathi et al., 1998; rGALR3, Wang et al., 1997b) and several studies have recently described the distribution of the GALRl and R2 subtypes within the rat dorsal root ganglion and spinal cord. To date, no data haves been published on the expression of GALR3 in the spinal cord or DRG. Although the GALRl and R2 subtypes share significant sequence homologies, each receptor utilizes distinct second messenger signalling cascades. The GALRl receptor utilizes a Gi/G, coupled protein to inhibit CAMP formation, while the GALR2 subtype signals though a Gq/Gll coupled pathway that results in phosphoinositide turnover and increased

intracellular calcium (Smith et al., 1997; Wang et al., 1998). Thus, galanin appears to have at least two potential effector pathways. Furthermore, distinct subsets of DRG cells are known to express GALRl and R2 receptors. GALRl expression appears restricted to large diameter cells within the DRG, while GALR2 appears restricted to a subset of exclusively small cells (O’Donnell et al., 1999). These receptors are also expressed intrinsically within the spinal cord and in situ hybridization has shown there is a distinct, non-overlapping expression of these receptors in spinal cord neurons (O’Donnell et al., 1999). These receptors have been shown to be regulated independently (Shi et al., 1997; Burazin and Gundlach, 1998): Both are downregulated after peripheral nerve injury (Xu et al., 1996b; Shi et al., 1997), but after inflammation there is an up-regulation of GALR2 (Shi et al., 1997). In axotomized motor neurons in the facial nucleus, there is an up-regulation of GALR2 while GALRl is down-regulated (Burazin and Gundlach, 1998). In addition to its anatomical distribution within structures known to be involved in processing of noxious sensory input, there is significant release of galanin from primary afferents under basal conditions in the spinal dorsal horn (Morton and Hutchisson, 1989; Hope et al., 1994). Taken together, these facts suggest that galanin may serve a tonic modulatory function. In support of this, we have shown recently that galanin knock-out mice display mild, though significant increases in baseline sensitivity to noxious thermal and mechanical stimulation (Fig. 1) (Kerr et al., 2000). This finding implies that galanin may act in an inhibitory fashion to modulate excitability at the spinal level in the absence of injury.

Fig. 3. The effect of chronic intrathecal galanin (25 rig/h) on PKCy immunoreactivity in the dorsal horn of the rat spinal cord. (A,B) Low power photomicrograph of the superficial dorsal horn of galanin-treated (A) or saline-treated (B) rats showing immunoreactivity for PKCy. Scale bar: 100 Km. (C) Graph of mean staining intensity for galanin-treated (filled bar, n = 4) and saline-treated (shaded bar, n = 5) rats. Galanin-treated rats show a significantly greater overall staining intensity compared to saline-treated animals, (P < 0.05, Mann-Whitney rank sum test). (D-F) High power photomicrographs of galanin-treated (D) and saline-treated (E) superficial dorsal horns. Arrows in D indicate cell bodies with typical membrane associated PKCy immunoreactivity. Scale bar: 100 km. (F) Graph showing mean cells per section with membrane associated PKCy for galanin-treated (filled bar, n = 4) and saline-treated (dashed bar, n = 5). Galanin treatment resulted in a significant increase in the number of cells with active PKCy, (* P i 0.05, Mann-Whitney rank sum test).

223

C

.a : E .g .3 ;

0.6 0.5 0.4 0.3 0.2

IT galanin

IT galanin

IT saline

IT saline

224

What is the pharmacological spinal sensory processing?

effect of galanin on

While experiments with galanin knock-out mice have confirmed an inhibitory role for endogenous galanin in naive states, addition of exogenous galanin to non-nerve injured preparations demonstrates that galaninergic modulation of spinal excitability is not uni-modal (Wiesenfeld-Hallin et al., 1988). Indeed, behavioral and electrophysiological evidence suggest that exogenously applied galanin can both facilitate and inhibit nociceptive sensory function in a dose-dependent manner (Cridland and Henry, 1988; Wiesenfeld-Hallin et al., 1989a,b). Using behavioral measures, increased (Cridland and Henry, 1988; Kuraishi et al., 1991) and decreased (Post et al., 1988) sensitivity to noxious stimuli have been reported in response to exogenous galanin administration. With electrophysiological measures, galanin is reportedly capable of producing a pure depression of spinal reflex excitability (Yanagisawa et al., 1986; Nussbaumer et al., 1989) as well as pure facilitation (Wiesenfeld-Hallin et al., 1988). In our own experiments, we have chronically infused a low dose of galanin (600 rig/day) intrathecally to naive rats and examined the thermal and mechanical thresholds over a 2-week testing period. We found a significant increase in mechanical sensitivity in galanin-treated rats compared to saline treated controls (Fig. 2A,B) with no effect on thermal sensitivity. Thus, it would appear that elevating lev-

els of galanin centrally in this circumstance leads to an overall increase in spinal activity. This finding is supported by increased expression of the immediate early gene c-fos (a marker for postsynaptic activity) (Fig. 2C-E) in galanin-treated rats as well as increased staining intensity for the gamma isoform of protein kinase C (PKCy) (Fig. 3), an important factor in the signalling cascade which underlies the development of central sensitization (Coderre, 1992; Yashpal et al., 1995; Malmberg et al., 1997; Martin et al., 1999). What is perhaps more revealing, however, is the body of evidence which demonstrates the pharmacological dependence of galanin’s central modulatory effects (Wiesenfeld-Hallin et al., 1989a). In normal preparations, intrathecal galanin at low doses produces a pure facilitation of spinal reflex excitability. Facilitation is followed by bi-phasic effects at higher doses. At the highest doses tested a pure depression of the spinal flexion reflex was seen (WiesenfeldHallin et al., 1989a). However, galanin has also been shown to have pure facilitatory effects on the responses of wide dynamic range (WDR) neurons in the dorsal horn (Reeve et al., 1999). In these experiments, intrathecally applied galanin potentiated evoked A6 and C-fiber responses as well as increasing postdischarge and wind-up. Intracellular recording of extirpated DRG sensory neurons in vitro show predominantly inward, depolarizing currents in response to galanin application (Puttick et al., 1994). Others however, have reported inward currents in response to galanin application

Fig. 4. Galanin up-regulation and behavioral responses are impaired following partial sciatic nerve injury in galanin knock-out mice. (A,B) Photomicrographs showing galanin immunoreactivity in the dorsal horn on the ipsilateral (right-hand side) and contralateral side of (A) wild-type and (B) galanin knock-out mice 2 weeks after partial sciatic ligation. Scale bar: 100 pm. (C) Group mean latencies for thermal withdrawal thresholds in wild-type (dashed bars, n = 9) and galanin knock-out mice (filled bars, n = 9) following partial sciatic ligation. Comparison between genotypes revealed significant interactions between genotype and test day (P < 0.05, two-way repeated measures ANOVA). Galanin knock-out mice display significantly reduced baseline withdrawal thresholds, but this trend is reversed on day 3, when thermal withdrawal latencies are significantly raised compared to wild-type controls (#I’ < 0.05, Dunnett’s post hoc test). Comparison within genotype revealed that wild-type mice display significant decreases in thermal withdrawal thresholds on the injured side compared to baseline control values from test day 3 onwards (*P < 0.05, one-way repeated measures ANOVA, Dunnett’s post hoc test). However galanin knock-out mice show no significant change from baseline control values (P > 0.05, one-way repeated measures ANOVA). (D) Alterations in mechanical thresholds following partial sciatic ligation in wild-type (dashed bars, n = 7) and galanin knock-out mice (filled bars, n = 7). Both wild-type and knock-out mice have significant decreases in mechanical thresholds on the ipsilateral side following nerve injury compared to baseline control values (P < 0.05, one-way repeated measures ANOVA on ranks, Dunnett’s post hoc test). Wild-type mice show a consistent reduction in mechanical thresholds with significant reductions from day 3 onwards (*P < 0.05, Dunnett’s post hoc test). The degree of mechanical allodynia is less dramatic in galanin knock-out mice which show significant reductions compared to baseline values only on test days 10 and 17 (‘P < 0.05, Dunnett’s post hoc test).

225

Thermal Sensitivity

0

1

3

7

10

14

17

Mechanical Sensitivity 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0

T B t1t a -I-

O

1

3

7 Test Day

10

14

17

226

only after a preconditioning 1997).

axotomy (Xu et al.,

81

What is the physiological role for galanin in neuropathic states? The experiments described above demonstrate the variety of effects of exogenously applied galanin on sensory transmission in the naive state and suggest that complex pharmacological mechanisms may determine the nature of central galaninergic signalling. However, these experiments do not address the functional role of endogenous galanin on central excitability after it is up-regulated in response to nerve injury. All animal models of peripheral nerve injury associated with neuropathic pain behaviors are also associated with an up-regulation of galanin in DRG neurons. These include complete axotomy (H&felt et al., 1987; Villar et al., 1989; Xu et al., 1990), nerve crush, (Villar et al., 1989), chronic constriction injury (Nahin et al., 1994), and partial nerve ligation (Ma and Bisby, 1997). However, peripheral nerve injuries are not equivalent: the present animal models of neuropathic pain are distinct in terms of the degree, modality and duration of neuropathic behaviors observed (see Bennett and Roberts, 1996 for review). There is currently some debate over the relationship between the extent of peripheral nerve injury, the degree of galanin expression, the type of sensory neuron affected and the extent of the associated behavioral hypersensitivity (Ma and Bisby, 1997, 1999; Shi et al., 1999). It has been reported that peripheral nerve injuries which produce partial damage to a nerve trunk are associated with a high degree of galanin expression and a high level of behavioral hypersensitivity (Ma and Bisby, 1997). However, in another model of neuropathic pain it has been suggested that galanin may protect animals from developing behavioral hypersensitivity (Shi et al., 1999). We have used a mouse strain carrying a null mutation in the galanin gene (gal (-/-)) (Wynick et al., 1998) to examine behavioral responses to nerve injury in animals in which galanin cannot be up-regulated as a result of nerve damage. In these mice, we have found that both the thermal hyperalgesia and mechanical allodynia that normally develop after partial sciatic nerve ligation were reduced (Fig. 4).

0

2

4

6

8

10

12

14

16

18

M

Day Post Axotomy Fig. 5. Galanin knock-out mice show greatly reduced autotomy following sciatic axotomy. Following complete transection of the left sciatic nerve, all wild-type mice display characteristic selfmutilation behavior (autotomy) of the affected hindpaw (open squares, n = 9). This response is significantly attenuated in the galanin knock-out mice (filled circles, n = 9).

In addition, the autotomy behaviors that arise after full sciatic transection were completely absent in these animals (Fig. 5) (Kerr et al., 2000). These findings are supported by recent data with Interleukin-6 (IL-6) knock-out mice (Murphy et al., 1999). These mice display significantly compromised neuropathic pain behaviors following chronic constriction injury of the sciatic nerve. These diminished behavioral responses are also associated with significantly reduced galanin expression in the central projection areas of sensory neurons (Murphy et al., 1999; see also Berthele et al., 2000, this volume). At apparent odds with these findings are previous reports using the flexor-reflex paradigm as a measure of central excitability. It is reported that the inhibitory effects of galanin on the flexor-reflex are augmented after nerve injury (Wiesenfeld-Hallin et al., 1989b, 1992). Additionally, nerve-injured rats treated with chimeric-peptide antagonists for galanin receptors show increased autotomy behaviors (Verge et al., 1993). Such experiments have focused attention on the inhibitory aspects of galanin after nerve injury. However, these findings are complicated by recent work examining the signalling characteristics of the various subtypes of galanin receptors (Forray et al., 1996; Smith et al., 1997). The chimeric peptide an-

227

either a pro- or anti-nociceptive peptide after nerve injury may not be feasible. Instead, the effects of this neuropeptide on spinal excitability may be better viewed in the context of a complex pharmacological balance. If, as the evidence suggests, there is a bell-shaped activity curve for galaninergic signalling, this may allow us to resolve apparently conflicting facilitatory/inhibitoty aspects of galanin’s physiology. As previously discussed, low levels of endogenous galanin can act to tonically modulate spinal excitability in an inhibitory manner in the absence of nerve injury (Kerr et al., 2000). If galanin levels are raised, as with exogenous application or after nerve injury, inhibition would switch to excitation (Wiesenfeld-Hallin et al., 1988, 1989a; Ma

tagonists for galanin binding sites that have been used to date (Wiesenfeld-Hallin et al., 1992; Verge et al., 1993; Xu et al., 1998) are now known to have potent agonist activity in vitro at the galanin receptor GALR2 (Smith et al., 1997). Preliminary data now also suggests that one of these antagonists (M40, [galanin-(1-12)-Pros-(Ala-Leu)z-Ala amide]) can act as an agonist at all three galanin receptor subtypes in vivo (Durkin et al., 1999). A bell-shaped dose-response curve for galanin and CNS excitability? In light of these conflicting reports, (see Table 1 for summary) attempts to directly label galanin as TABLE

1

Summary

of several

experiments

using

exogenous

galanin

or galanin

knock-out

Experiment

Species

Treatment

(A) Behavioral measures Post et al. (1988) Cridland and Henry (1988)

mouse (naive) rat (naive)

it-gal, (l-10 it-gal (2-20

rat (naive) rat (ischemic nerve injury) rat (CCI-partial nerve injury) rat (naive) formalin test mouse (partial nerve injury)

it-gal (0.3-3 kg) it-gal (cumulative 30 KS) it-gal (3-20 kg)

Ku&hi et al. (1991) Hao et al. (1999) Yu et al. (1999) Kerr Kerr

et al. (1999) et al. (2000)

(B) Electrophysiological Yanagisawa et al. (1986) Wiesenfeld-Hallin (1988) Wiesenfeld-Hallin (1989a) Wiesenfeld-Hallin (1989b) Puttick

et al. (1994)

Xu et al. (1997)

measures Pl-P4 rat spinal (in vitro) et al. rat (naive)

et al.

rat (naive)

et al.

rat (sciatic

cord

axotomy)

neonatal rat DRG (in vitro) rat DRG (in vitro) (naive and axotomy)

(A) Behavioral measures used to assess galanin function galanin knock-out mice (gal -/-). (B) Electrophysiological bi-phasic facilitation followed by inhibition; it, intrathecal;

it-gal

mice (gal (-/-))

(dose)

illustrating

Effect

I*g) kg)

dose

(10 ng, 1 kg, 10

t latency to noxious heat, TF HP tests f latency in TF test 4 mechanical threshold suspected motor impairments in TF test 4 mechanical threshold (PW) alleviate mechanical/cold allodynia (PW) bi-lateral t to heat, cold and pressure stimuli nerve injured animals (PW) prolong second phase formalin response

wg) gal C--F)

or (+/+I

gal (-/-), decreased

gal (0.1-5

FM)

J monosynaptic

it-gal

(10 ng-10

it-gal

(10 ng-100

it-gal

(10 ng-10

its range of effects

pg)

abolished neuropathic

autotomy behaviors

in

behaviors,

reflex

gal (l-300

nM)

f flexor-reflex excitability at all doses (greater f with mechanical stimuli) t flexor-reflex excitability at low doses (l-10 ug) bi-phasic t$ highest dose, pure JfJ. flexor-reflex excitability seen at lower doses with axotomy (10-100 ng), incidence of pure 4 increases with axotomy at higher doses (I- 10 ug) inward currents

gal (0.01-l

p,M)

inward

pg)

wg)

currents

after axotomy

only

after exogenous administration in naive or nerve-injured rats or mice and using measures using in vivo and in vitro paradigms. t, increase; -1, decrease; TJ,, TF, tail-flick test; HP, hot plate test; PW, paw withdrawal.

228

and Bisby, 1997; Murphy et al., 1999; Kerr et al., 2000). Increasing galanin levels in the CNS further still, for example in cases of nerve injury coupled with exogenous galanin administration (Hao et al., 1999; Yu et al., 1999) may result in a final depression of spinal excitability. Assessing galaninergic function in this framework can now account for the increased depression seen in flexor-reflex excitability after nerve injury or the inhibition of behavioral hyperalgesia (Wiesenfeld-Hallin et al., 1989b; Hao et al., 1999; Yu et al., 1999) as well as the bi-phasic or pro-nociceptive effects reported to date. Conclusions: does galanin belong in the same functional class as other dual acting neuromodulators? The precise mechanisms which could confer this mode of galaninergic signalling are unknown. Recent advances in the understanding of galanin receptor subtypes, their distribution and their signalling pathways may provide useful insights into galanin’s pharmacological profile. As outlined earlier, galanin receptor subtypes 1 and 2 show distinct distribution and signalling properties which suggests that these receptors may serve distinct functional roles (Wang et al., 1998; O’Donnell et al., 1999). It is likely that the availability of galanin at these receptor subtypes and the degree of activation of each receptor class has a profound and possibly diverging effect upon pre- or postsynaptic systems. The ability of the GALR2 receptor to transduce information via distinct second messenger cascades from the GALRl subtype represents a novel mechanism whereby a single neuropeptide may have opposing effects on neuronal excitability. Therefore, galanin may be better understood in the context of other complex acting neurotransmitters/neuromodulators such as serotonin or the endogenous opioid dynorphin. Both exhibit the same bi-phasic, facilitatory/inhibitory characteristics on nociceptive sensory systems (Randic et al., 1995; Zhuo and Gebhart, 1997; see also Moore et al., 2000, this volume). These effects are dependent on dose, site of administration and/or differential receptor subtype activation (Dubner and Ruda, 1992; Hori et al., 1996; Li and Zhuo, 1998). The task at hand now is to better define the conditions and mechanisms in

which galanin exerts its own dual effects on spinal excitability in naive and neuropathic state. Acknowledgements B.J.K. wishes to thank S. Yanow and J. Boucher for helpful comments on the manuscript. This work was supported by the Wellcome Trust (UK) and Astra Pharma Inc. References Berthele et al. (2000) In: J. Sandkuhler, B. Bromm and G.F. Gebhart (Eds.), Nervous System Plasticity and Chronic Pain, Progress in Bruin Research, Vol. 129. Elsevier Science, Amsterdam, pp. 191-203. Bennett, G.J. and Roberts, W.J. (1996) Animal models and their contribution to our understanding of complex regional pain syndromes I and II. Prog. Pain Res. Manage., 6: 102-107. Burazin, T.C.D. and Gundlach, A.L. (1998) Inducible galanin and GalR2 receptor system in motor neuron injury and regeneration. J. Nearochem., 71: 879-882. Burgevin, M.C., Loquet, I., Quartonet, D. and Habert-Ortoli, E. (1996) Cloning, pharmacological characterization, and anatomical distribution of a rat cDNA encoding for a galanin receptor. J. Mol. Neurosci., 6: 33-41. Ch’ng, J.L., Christofides, N.D., Anand, P., Gibson, S.J., Allen, Y.S., Su, H.C., Tatemoto, K., Morrison, J.F.B., Polak, J.M. and Bloom, S.R. (1985) Distribution of galanin immunoreactivity in the central nervous system and the responses of galanin containing neuronal pathways to injury. Neuroscience, 16: 343-3.54. Coderre, T.J. (1992) Contribution of protein kinase C to central sensitization and persistent pain following tissue injury. Neurosci. Lett., 140: 181-184. Crawley, J. (1996) Galanin-acetylcholine interactions: relevance to memory and Alzheimer’s disease. Life Sci., 58(24): 21852199. Cridland, R.A. and Henry, J.L. (1988) Effects of intrathecal administration of neuropeptides on a spinal nociceptive reflex in the rat: VIP, galanin CGRP, TRH, somatostatin, and angiotensin II. Neuropeptides, I 1: 23-32. Devor, M. and Seltzer, Z. (1999) Pathophysiology of dam aged nerves in relation to chronic pain. In: P.D. Wall and R. Melzack (Eds.), The Textbook of Pain, 4th Ed. Churchill Livingstone, New York, pp. 129-164. Dubner, R. and Ruda, M.A. (1992) Activity dependent neuronal plasticity following tissue injury and inflammation. Trends Neurosci., 15(13): 96-103. Durkin, M.M., Walker, M.W., Forray, C. and Branchek, T.A. (1999) Pharmacological properties of [I- 1251-M-40 binding in rat brain sections: GTPgammaS-dependent modulation. Sot. Neurosci. Abstrz, 25: 185. Fathi, Z., Cunningham, A.M., Iben, L.G., Battaglino, PB., Ward, S.A., Nichol, K.A., Pine, K.A., Wang, J.C., Goldstein, M.E.,

229 Iismaa, T.P. and Zimanyi, LA. (1998) Cloning, pharmacological characterization and distribution of a novel galanin receptor. Mol. Brain Rex, 53: 348. Forray, C., Smith, K.E., Gerald, C., Vaysse, P., Weinshank, R. and Branchek, T. (1996) Pharmacological characterization of the recombinant human and rat GALRl galanin receptor. Sot. Neurosci. Abstx, 22: 1304. Hao, J.X., Shi, T.J., Xu, IS., Kaupilla, T., Xu, X.J., H&felt, T., Bartfai, T. and Wiesenfeld-Hallin, Z. (1999) Intrathecal galanin alleviates allodynia-like behaviour in rats after partial peripheral nerve injury. Eul: J. Neurosci., 11: 427-432. H&felt, T., Wiesenfeld-Hallin, Z., Villar, M. and Melander, T. (1987) Increase of galanin-like immunoreactivity in rat dorsal root ganglion cells after peripheral axotomy. Neurosci. L&t., 83: 217-220. H&felt, T., Zhang, X. and Wiesenfeld-Hallin, Z. (1994) Messenger plasiticity in primary sensory neurons following axotomy and its functional implications, Trends Neurosci., 17: 22-30. Hope, P.J., Lang, C.W., Grubb, B.D. and Duggan, A.W. (1994) Release of immunoreactive galanin in the spinal cord of rats with ankle inflammation: studies with antibody microprobes. Neuroscience, 60: 801-807. Hori, Y., Endo, K. and Takahashi, T. (1996) Long lasting synaptic facilitation induced by serotonin in superficial dorsal horn neurons of the rat spinal cord. J. Physiol., 492: 867-876. Howard, A.D., Tan, C., Shiao, L.L., Palyha, O.C., Mckee, K.K., Weinberg, D.H., Feighner, S.D., Cascieri, M.A., Smith, R.G., VanDerPloeg, L.H.T. and Sullivan, K.A. (1997) Molecular cloning and characterization of a new receptor for galanin. FEBS L.&t., 405: 285-290. Kerr, B.J., Gupta, Y., Thompson, S.W.N. and McMahon, S.B. (1999) Intrathecal galanin modulates second phase formalin response in the rat. Sot. Neurosci. Abst,:, 25(2): 1436. Kerr, B.J., Cafferty, W.B.J., Gupta, Y., Bacon, A., Wynick, D., McMahon, S.B. and Thompson, S.W.N. (2000) Galanin knock-out mice reveal nociceptive deficits following peripheral nerve injury. Eul: J. Neurosci., 12: 793-802. Kuraishi, Y., Kawamura, M., Yamaguchi, T., Houtani, T., Kawabata, S., Futaki, S., Fuji, N. and Satoh, M. (1991) Intrathecal injections of galanin and its antiserum affect nociceptive responses of rat to mechanical, but not thermal stimuli. Pain, 44: 321-324. Li, P and Zhuo, M. (1998) Silent glutaminergic synapses and nociception in mammalian spinal cord. Nature, 393: 695-698. Ma, W. and Bisby, M.A. (1997) Differential expression of galanin immunoreactivities in the primary sensory neurons following partial and complete sciatic nerve injuries Neuroscience, 79: 1183-l 195. Ma, W. and Bisby, M.A. (1999) Ultrastructural localization of increased neuropeptide immunoreactivity in the axons and cells of the gracile nucleus following chronic constriction injury of the sciatic nerve. Neuroscience, 83: 335-348. Malmberg, A.B., Chen, C., Tonegawa, S. and Basbaum, AI. (1997) Preserved acute pain and reduced neuropathic pain in mice lacking PKC gamma. Science, 278: 279-283. Martin, W.J., Liu, H., Wang, H., Malmberg, A.B. and Basbaum, AI. (1999) Inflammation induced upregulation of protein ki-

nase C (gamma) immunoreactivity in rat spinal cord correlates with enhanced nociceptive processing. Neuroscience, 88: 1267-1274. McMahon, S.B. and Bennett, D.L.H. (1999) Trophic factors and pain. In: P.D. Wall and R. Melzack (Eds.), The Textbook of Pain, 4th Ed. Churchill Livingstone, New York, pp. 105-128. Melander, T., Hokfelt, T. and Rokaeus, A. (1986) Distribution of galanin-like immunoreactivity in the rat central nervous system. J. Comp. Neurol., 248: 475-517. Moore, K.A., Baba. H. and Woolf, C.J., Synaptic transmission and plasticity in the superficial dorsal horn. In: J. Sandktihler, B. Bromm and G.F. Gebhart (Eds.), Nervous System Plasticity and Chronic Pain, Progress in Brain Research, Vol. 129. Elsevier Science, Amsterdam, pp. 63-80. Morton, C.R. and Hutchisson, W.D. (1989) Release of sensory neuropeptides in the spinal cord: studies with calcitonin gene related peptide and galanin. Neuroscience, 31: 807-815. Murphy, PG., Ramer, MS., Borthwick, L., Gauldie, J., Richardson, PM. and Bisby, M.A. (1999) Endogenous interleukin-6 contributes to hypersensitivity to cutaneous stimuli and changes in neuropeptides associated with chronic nerve constriction in mice. Eur: J. Neurosci., 11: 2243-2253. Nahin, R.L., Marino De Leon, K.R. and Ruda, M. (1994) Primary sensory neurons exhibit altered gene expression in a rat model of neuropathic pain. Pain, 58: 95-108. Nussbaumer, J.C., Yanagisawa, M. and Otsuka, M. (1989) Pharmacological properties of a C-fibre response evoked by saphenous nerve stimulation in an isolated spinal cord-nerve preparation of the newborn rat. BI: J. Pharmacol., 98: 373-382. O’Donnell, D., Ahmad, S., Wahlestedt, C. and Walker, P. (1999) Expression of the novel galanin receptor subtype GALR-2 in the adult rat CNS: Distinct distribution from GALR-1. J. Comp. Neurol., 409: 469-48 1. Parker, E.M., Izzarelli, D.G., Nowak, H.P., Mahle, C.D., Iben, L.G., Wang, J. and Goldstein, M.E. (1995) Cloning and Char acterization of the rat GALRI galanin receptor from Rinl4B insulinoma cells. Mol. Brain Res., 34: 179- 189. Post, C., Alari, L. and Hokfelt, T. (1988) Intrathecal galanin increases the latency in the tailflick and hot-plate tests in the mouse. Acta Physiol. Stand., 132: 583-584. Puttick, R.M., Pinnock, R.D. and Woodruff, G.N. (1994) Galanin-induced membrane depolarization of neonatal rat cultured dorsal root ganglion cells. Eur: J. Pharmacol., 254: 303306. Randic, M., Cheng, G. and Kojic, L. (1995) K-Gpioid receptor agonists modulate excitatory transmission in substantia gelatinosa neurons of the rat spinal cord. J. Neurosci., 15: 68096826. Reeve, A.J., Walker, M.J.K. and Urban, L. (1999) Galanin facilitates the responses of dorsal horn wide-dynamic neurones in the anaesthetized rat. BI: J. Pharmacol., 128: 181. Shi, T.J.S., Zhang, X., Holmberg, K., Xu, Z.Q.D. and H&felt, T. (1997) Expression and regulation of galanin-R2 receptors in rat primary sensory neurons: effect of axotomy and inflammation. Neurosci. L&t., 237: 57-60. Shi, T.J.S., Cui, J.-G., Meyerson, B.A., Linderoth, B. and Hokfelt, T. (1999) Regulation of galanin and neuropeptide-Y in

230

dorsal root ganglia and dorsal horn in rat mononeuropathic models: possible relation to tactile hypersensitivity. Neuroscience, 93: 741-757. Simmons, D.R., Spike, R.C. and Todd, A.J. (1995) Galanin is contained in GABAergic neurons in the rat spinal dorsal horn. Neurosci. Left., 187: 119-122. Smith, K.E., Forray, C., Walker, M.W., Jones, K.A., Tamm, J.A., Bard, J., Branchek, T.A., Linemeyer, D.L. and Gerald, C. (1997) Expression cloning of a rat hypothalamic galanin receptor coupled to phosphoinositide turnover. J. Biol. Chem., 272: 24612-24616. Tatemoto, K., Rokaeus, A., Jomvall, H., McDonald, T.J. and Mutt, V. (1983) Galanin - a novel biologically active peptide from porcine intestine. FEBS Left., 164: 124-128. Verge, V.M.K., Xu, X.J., Langel, U., Hokfelt, T., WiesenfeldHallin, Z. and Bartfai, T. (1993) Evidence for endogenous inhibition of autotomy by galanin in the rat after sciatic nerve section: demonstrated by chronic intrathecal infusion of a high affinity galanin receptor antagonist. Neurosci. Let?., 149: 193197. Villar, M.J., Cortes, R., Theodorsson, E., Wiesenfeld-Hallin, Z., Schalling, M., Fahrenkrug, J., Emson, P.C. and H&felt, T. (1989) Neuropeptide expression in rat dorsal root ganglion cells and spinal cord after peripheral nerve injury with special reference to galanin. Neuroscience, 33: 587-604. Villar, M.J., Wiesenfeld-Hallin, Z., Xu, X.J., Theodorsson, E., Emson, PC. and H&felt, T. (1991) Further studies on galanin, substance P, and CGRP-like immunoreactivities in primary sensory neurons and spinal cord: effects of dorsal rhizotomies and sciatic nerve lesions. Eq. Neural., 112: 29-39. Wang, S., Hashemi, T., He, C.G., Strader, C. and Bayne, M. (1997a) Molecular cloning and pharmacological characterization of a new galanin receptor subtype. Mol. Pharmacol., 52: 337-343. Wang, SK., He, C.G., Hashemi, T. and Bayne, M. (1997b) Cloning and expressional characterization of a novel galanin receptor - identification of different pharmacophores within galanin for the three galanin receptor subtypes. J. BioL Chem., 272: 31949-31952. Wang, S., Hashemi, T., Fried, S., Clemmons, A.L. and Hawes, B.E. (1998) Differential intracellular signaling of the GalRl and GalR2 galanin receptor subtypes. Biochemistry, 37: 671 l6717. Wiesenfeld-Hallin, Z., Villar, M. and H&felt, T. (1988) Intrathecal galanin at low doses increases spinal reflex excitability in rats more to thermal than mechanical stimuli. Exp. Bruin Res., 7 1: 663-666. Wiesenfeld-Hallin, Z., Villar, M.J. and H&felt, T. (1989a) The effects of intrathecal galanin and C-fiber stimulation on the flexor reflex in the rat. Brain Res., 486: 205-213. Wiesenfeld-Hallin, Z., Xu, X.J., Villar, M. and Hokfelt, T. (1989b) The effect of intrathecal galanin on the flexor reflex in rat: increased depression after sciatic nerve section. Neurosci. Len., 105: 149-154. Wiesenfeld-Hallin, Z., Xu, X.J., Langel, U., Bedecs, K., Hokfelt,

T. and Bartfai, T. (1992) Galanin mediated control of pain: enhanced role after nerve injury. Proc. Natl. Acad. Sci. U.S.A., 89: 3334-3337. Wynick, D., Small, C.J., Bacon, A., Holmes, F.E., Norman, M., Ormandy, C.J., Kilic, E., Kerr, N.C., Ghatei, M., Talamantes, E, Bloom, S.R. and Pachnis, V. (1998) Galanin regulates prolactin release and lactotroph proliferation. Proc. Nutl. Acad. Sci. U.S.A., 95: 12671-12676. Xu, I.S., Grass, S., Xu, X.J. and Wiesenfeld-Hallin, Z. (1998) On the role of galanin in mediating spinal flexor reflex excitability in inflammation. Neuroscience, 85: 827-835. Xu, X.J., Wiesenfeld-Hallin, Z., Villar, M.J., Fahrenkrug, J. and H&felt, T. (1990) On the role of galanin substance P and other neuropeptides in primary sensory neurons of the rat: studies on spinal reflex excitability. Eur J. Neurosci., 2: 733743. Xu, Z.Q., Shi, T.J. and H&felt, T. (1996a) Expression of galanin and a galanin receptor in several sensory systems and bone anlage of rat embryos. Proc. Nafl. Acad. Sci. U.S.A., 93: 14901-14905. Xu, Z.Q., Shi, T.J., Landry, M. and H&felt, T. (1996b) Evidence for galanin receptors in primary sensory neurons and effect of axotomy and inflammation. NeuroReport, 8: 237-242. Xu, Z.Q.D.. Zhang, X., Grillner, S. and H&felt, T. (1997) Electrophysiological studies on rat dorsal root ganglion neurons after peripheral axotomy: changes in responses to neuropeptides. Proc. Natl. Acad. Sci. U.S.A., 94: 13262-13266. . Xu, X.J. and Wiesenfeld-Hallin, Z. (1997) Novel modulators of nociception. In: A. Dickenson and J.M. Besson (Eds.), The Pharmacology of Pain. A Handbook of Experimental Pharmacology, Vol. 130. Springer, Heidelberg, pp. 21 l-234. Yanagisawa, M., Yagi, N., Otsuka, M., Yanaihara, C. and Yanaihara, N. (1986) Inhibitory effects of galanin on the isolated spinal cord of the newborn rat. Neurosci. Lea., 70: 278-282. Yashpal, K., Pitcher, G.M., Parent, A., Quirion, R. and Coderre, T.J. (1995) Noxious thermal and chemical stimulation induced increases in 3H-phorbol 12,13-dibutyrate binding in spinal cord dorsal horn as well as persistent pain and hyperalgesia, which is reduced by inhibition of protein kinase C. J. Neurosci., 15: 3263-3272. Yu, L.-C., Lundeberg, S., An, H., Wang, F.-X. and Lundeberg, T. (1999) Effects of intrathecal galanin on nociceptive responses in rats with mononeuropathy. Life Sci., 64: 1145-l 153. Zhang, X., Nicholas, A.P. and H&felt, T. (1993) Ultrastructure studies on peptides in the dorsal horn of the spinal cord-I.Co-existence of galanin with other peptides in primary afferents in normal rats. Neuroscience, 57: 365-384. Zhang, X., Nicholas, A.P. and Hokfelt, T. (1995) Ultrastructure studies on peptides in the dorsal horn of the rat spinal cord-II. Co-existence of galanin with other peptides in local neurons. Neuroscience, 64: 875-891. Zhuo, M. and Gebhart, G.F. (1997) Biphasic modulation of spinal nociceptive transmission from the medullary raphe nuclei in the rat. J. Neurophysiol., 78: 746-758.

J. Samktihler, B. Bromm and G.F. Gebhart (Ed%) Progress in Brain Research, Vol. 129 0 2000 Elsevier Science B.V. All rights reserved

CHAPTER 16

Plasticity in descending pain modulatory systems Antti Pertovaara* Department

of Physiology,

Institute

of Biomedicine,

University

Introduction The development of experimental animal models of neuropathy and other chronic pain conditions has provided the possibility to study cellular mechanisms underlying pathophysiological pain. The studies performed using these models have provided abundant evidence indicating that various pathophysiological conditions may cause dramatic changes in the transmission and modulation of pain-related information. Recently, we performed a series of studies in which we attempted to characterize changes in central transmission of nociceptive information, and, in particular, descending modulation of nociception, in a rat model of experimental neuropathy caused by unilateral ligation of two spinal nerves (Kim and Chung, 1992). This model produces a robust, long-lasting and highly reproducible allodynia and hyperalgesia to mechanical and cooling stimulation (Kim and Chung, 1992; Rijyttg et al., 1999) without causing distress or measurable anxiety to the animals (Kontinen et al., 1999). Also, heat hyperalgesia has been described following spinal nerve ligation in many laboratories (Kim and Chung, 1992). However, in contrast to hypersensitivity to mechanical and cooling stimulation, heat hyperalgesia in this model has been more variable and it has not been observed in all experimental conditions (Bian et al.,

*Corresponding author: A. Pertovaara, Department of Physiology, Institute of Biomedicine, University of Turku, Kiinamyllynkatu 10, FIN-20520 Turku, Finland. Tel.: +358 (40) 760-7123; Fax: +358 (2) 250-2610; E-mail: [email protected]

of Turku,

Kiinamyllynkatu

IO, FIN-20520

Turku,

Finland

1998; Riiyttg et al., 1999). Concomitant with an increased sensitivity to stimuli activating predominantly thick myelinated fibers, spinal nerve ligation causes a marked reduction in the number of myelinated axons innervating the allodynic skin (R6yttti et al., 1999). This paradoxical finding suggests that spinal nerve ligation induces a central amplification, disinhibition of tactile signals, or both. To determine whether findings with the spinal nerve ligation-induced model of neuropathy can be generalized to other pathophysiological conditions, a parallel series of studies on plasticity of central pain transmission and modulation was performed using an acute model of hyperalgesia and allodynia induced by neurogenic inflammation. Role of supraspinal controls It is well-established that brainstem spinal pathways have an important role in the control of spinal nociceptive responses (Fields and Basbaum, 1999). In general, the influence of the brainstem on spinal nociception has been considered to be predominantly inhibitory. However, there is accumulating evidence indicating that brainstem-spinal pathways may not only attenuate pain, but descending controls may also contribute to facilitation or disinhibition of nociceptive signals at the spinal cord level depending on the submodality of test stimulation, pathophysiological condition and the neural pathway (for reviews, see Maier et al., 1992; Lima et al., 1998; Urban and Gebhart, 1999). Studies addressing the role of supraspinal control in the spinal nerve ligation-induced model of neuropathy have shown that a reversible block of the rostroventromedial medulla

232

INTRACEREBRAL

INJECTION

SYSTEMIC

INJECTION

IO' . ctrl paw post 20 /~g RVM 0 oper paw post 20 pg RVM D oper paw post SAL RVM

lzm

TIME [min]

TIME [mini

(Cl

(W

INTRACEREBRAL

TAIL-FLICK

INJECT.

50 1

SAL RVM

2Ql.N

4oW3

20 P!3

RVM

RVM

PAG

TEST

IO,

I

SAL RVM LIDOCAINE

20 I4 RVM

20 /Jg

PAG

DOSE

Fig. 1. Selective attenuation of tactile allodynia following lidocaine-induced block of the RVM or PAG in rats with a spinal nerve ligation-induced neuropathy. (A) Following lidocaine (20 kg) administration in the RVM the hindlimb withdrawal threshold to monofilament stimulation was elevated in the neuropathic side (oper paw), but not in the contralateral side (ctrl paw). Saline (SAL) administration in the RVM had no influence on tactile allodynia. (B) The selective attenuation of tactile allodynia following supraspinal administration of lidocaine was not due to a systemic effect, since systemic administration of lidocaine at the dose of 40 ug had no influence on tactile ahodynia. (C) The antiallodynic effect of lidocaine in the RVM was dose-related. Lidocaine at the dose of 20 ug produced an equipotent antiallodynia following administration in the RVM and the PAG. (D) An antiallodynic dose of lidocaine in the RVM or PAG did not produce any change in the latency of the radiant heat-induced tail-flick in neuropathic animals. The error bars represent S.E.M. (n = 4-7). *P < 0.05 (reference: in A and B, the corresponding pre-injection threshold; in C and D, the saline group). Adapted from Pertovaara et al. (1996).

(RVM) or periaqueductal gray (PAG) by lidocaine selectively attenuate tactile allodynia (Fig. 1; Pertovaara et al., 1996). In line with this, midthoracic spinalization also selectively attenuated tactile allodynia in neuropathic animals (Fig. 2; Bian et al., 1998; Kauppila et al., 1998). These findings, together with similar findings in other models of hypersensitivity (Herrero and Cervero, 1996; Urban et al.,

1996; Kauppila, 1997; Mans&a and Pertovaara, 1997; Pertovaara, 1998; Sung et al., 1998), suggest that a positive feedback loop involving structures rostra1 to the injured segment contributes to mechanical hypersensitivity at the spinal cord level. However, because mechanical hypersensitivity induced by neurogenic inflammation is not completely abolished by spinalization (Pertovaara, 1998), at least

233

(A)

Monofilament-induced

reflex

(W

100

7.5

z e8 eI_

10

s c 5 2 2.5

g

30

3 1

Ctrl

Np-ct

Np-ip

5.0 0.0 bh

Heat-induced

hindlimb

reflex

ClPre I Post

# ?Y*

Ctrl

NP

-

Fig. 2. Effect of spinalization on hindlimb withdrawal reflex response elicited by mechanical stimulation with monofilaments (A) versus radiant heat (B) in rats with no other pathophysiology (CM) and in rats with a spinal nerve ligation-induced neuropathy (Np). Np-ip, neuropathic side; Np-ct, side contralateral to neuropathy; Pre, prior to spinal transection; Post, 10 h after midthoracic transection of the spinal cord. Note a marked elevation of threshold to mechanical stimulation (antiallodynic effect) that is accompanied by a decrease in radiant heat-induced withdrawal latency (hyperalgesic effect) in the neuropathic limb following spinal transection. *P < 0.05 (reference: the corresponding value prior to spinalization; n = 64, #P c 0.05 (reference: the corresponding value in the Ctrl-group). Adapted from Kauppila et al. (1998).

part of the mechanical hypersensitivity observed in spinal dorsal horn neurons is due to spinal segmental mechanisms, which is in line with earlier studies demonstrating central hypersensitivity in spinal preparations (Coderre et al., 1993). In contrast to mechanically evoked responses, heat-evoked responses were enhanced following spinalization (Kauppila et al., 1998), indicating that heat-evoked responses are predominantly under tonic inhibitory controls. The tonic descending inhibition of heat-evoked responses varied depending on the segmental level, being stronger on tail- than limb-evoked responses (Bian et al., 1998; Kauppila et al., 1998), and more distinct in neuropathic than control animals (Fig. 2B; Kauppila et al., 1998). The influence of supraspinal controls on spinal nociception may also vary depending on whether the nociceptive signals originate from the area of primary hyperalgesia (the injured site in which hypersensitivity is predominantly due to peripheral sensitization of nociceptors) or from the area of secondary hyperalgesia (adjacent to the injured site, in which hypersensitivity is mainly due to central mechanisms). That is, brainstem-spinal pathways predominantly facilitate secondary hyperalgesia, but not primary hyperalgesia (Urban et al., 1996, 1999b; Pertovaara, 1998). The complexity of descending modulatory systems is illustrated by the previous findings that at least in some non-neurogenie inflammatory conditions, the descending inhibitory control of ascending nociceptive signals from the inflamed region is enhanced (Schaible et

al., 1991; Ren and Dubner, 1996; Tsuruoka and Willis, 1996). These findings add to the accumulating evidence indicating that various pathophysiological conditions, including experimental neuropathy, may significantly change the brainstem-spinal control of nociceptive signals and that the direction of change may depend on various experimental factors. Dependence of hypersensitivity on maintained injury discharge versus afferent barrage at the time of injury Previous studies indicate that the afferent barrage induced by experimental nerve injury may trigger enhanced pain sensitivity, revealed by the preemptive effect of local lidocaine treatment in some models of neuropathy (e.g. Kauppila and Pertovaara, 1991; Luukko et al., 1994; for a review see Jensen and Nikolajsen, 2000, this volume). The injury discharge may trigger the enhanced responsiveness due to a number of neurochemical changes in the central nervous system and particularly in the spinal dorsal horn; these plastic changes presumably involve an important contribution of N-methyl-D-aspartate (NMDA) receptors (Coderre et al., 1993; see Moore et al., 2000, this volume; Sandktihler et al., 2000, this volume). This is supported in two different spinal nerve ligation-induced models of neuropathy by the finding that a single administration of an NMDA receptor antagonist systemically prior to, but not after, the nerve injury dose-dependently attenuated the

234

(A)

Mechanical

hyperalgesia

Tactile

allodynia

iceiflt---~+~. 0.05) among the three patient groups. Also, in the posteroinferior region, there was a large increase in the incidence of stimulation-evoked pain sites in the PSP patients compared with the other two groups (P < 0.001) (Fig. 3B). However, in contrast to the findings in the Vc, the incidence of temperature sensations evoked in the posteroinferior region of PSP patients was found to

Fig. 4. Scatter plots in the sagittal plane showing motor group of patients. Open circles represent sites posterior border of the Vc and y = 0 represents the mm. Reproduced with permission from Dostrovsky

Temperature

stimulation evoked sensations NSP, non-stroke pain group;

of pain or innocuous PSP, post-stroke pain

be significantly lower than in the NSP (P < 0.001) or movement disorder (P -C 0.001) patients. The locations of sites where pain, warm and cold sensations were evoked in each of the three groups were reconstructed in the sagittal and coronal planes. An example of the distribution of these sites in the sagittal plane in the motor group is shown in Fig. 4. The vertical and horizontal axes represent the posterior and inferior borders of the Vc, respectively. The distribution of sites for each modality was quite similar, except that cold sites were almost exclusively located posterior to the Vc. Most sites were concentrated in the region l-3 mm inferior and posterior to the inferior and posterior border of the Vc. In the mediolateral axis, the majority of the pain sites that were located within Vc extended about 2-mm medial or lateral to the representation of the face/hand border. In contrast, most of the warm sites were located in the medial-inferior quadrant (P < 0.0001). A

the locations of the stimulation sites evoking pain, warm or cold sensations in the within the Vc. The x-axis is parallel to the AC-PC line, x = 0 (mm) represents the ventral border of the Vc. A, anterior; I, inferior; P, posterior; S, superior. Scale is in et al. (2000).

251 PAIN

WARM

A’

-P -10

-8

-8

6 0

l

-8

V

-10

COLD

lo-

D

8640 9

-8 -10

V

8

10

252 similar distribution was observed for the cold sites, although it failed to reach statistical significance. The findings of this recent study confirm and extend the observations reported previously (Lenz et al., 1993b, 1998; Davis et al., 1996) showing that the major thalamic region from where pain and temperature sensations can be evoked is the posteroinferior region. However, in view of the absence of histological confirmation of stimulation sites and the problem of activation of axons of passage, it is difficult to conclude with certainty which nuclei are involved in mediating these sensations. It is likely that some of these sensations are probably due to activation of the VMpo nucleus. As summarized above, it has recently been shown in the monkey that the VMpo contains a large number of nociceptiveand cooling-specific neurons and is a major target of spinothalamic tract neurons originating in lamina I of the spinal and medullary dorsal horn (Craig et al., 1994). Our recent study, which showed that in this region one can find cooling-specific neurons and that at such sites microstimulation elicits cold sensations, provides strong support for an involvement of the VMpo in mediating the sensation of cold (Davis et al., 1999). However, many of our effective stimulation sites were lateral to the presumed location of the VMpo and probably activated neurons in the VP1 and/or axons of spinothalamic tract neurons terminating in the VP1 and/or Vc or possibly also axons of VMpo neurons ascending to cortex. The observations in this and other studies that stimulation within and ventroposterior to the Vc can evoke painful sensations strongly suggest that it is sufficient to activate the component of the pathway passing through the lateral thalamus in order to produce the sensation of pain. The pain evoked by such stimulation is frequently experienced by the patients as natural and very unpleasant. This suggests that the so-called ‘affective motivational aspects’ of pain, which are thought to be mediated by medial thalamic structures are, or can also be, elicited by activity in pathways ascending through the lateral thalamus. The very large incidence of stimulation-evoked pain both within Vc and in the posteroinferior region of PSP patients compared with the other two groups confirms and extends the findings of Davis et al. (1996) and Lenz et al. (1998). These find-

ings suggest that in PSP patients, there is a marked alteration in the processing of nociceptive signals. These changes may involve reduction in inhibitory processes in the thalamus, increases in efficacy of excitatory nociceptive inputs and/or sensitization of thalamic nociceptive neurons. The alterations in processing giving rise to these changes may also or alternatively occur at the cortical level. It is well known from animal studies that lesions in the somatosensory system can result in major changes in the properties of deafferented neurons and give rise to alterations in the somatotopic representation of the body. It is thus likely that damage to the ascending pain pathways that most likely occurs following stroke in PSP patients results in plasticity that gives rise to the chronic pain and to the increased incidence of stimulation evoked pain sites in the thalamus. These findings thus suggest that alterations in the thalamocortical processing and modulation of nociceptive signals may be involved in mediating central pain. Also of interest was the finding of a marked reduction in temperature sites. This might be due to damage to the temperature pathway, as it is well documented that PSP patients frequently have deficits in temperature perception. Alternatively, as suggested by Lenz et al. (1998) on the basis of their similar findings, perhaps the activity of innocuous thermoreceptive neurons, gives rise to painful sensations in these patients. Bursting activity in the thalamus of chronic pain patients Deafferentation and damage to the CNS have been reported to give rise to an abnormal bursting firing pattern in some neurons at various levels of the somatosensory system (Loeser and Ward, 1967; Loeser et al., 1968; Black, 1974; Albe-Fessard and Longden, 1983; Lombard and Larabi, 1983). It has generally been assumed that such activity may be the source of central pain and that regions containing neurons firing in a bursting pattern may be contributing to the chronic pain (e.g. see Boivie, 1994; Lenz and Dougherty, 1997). Over the past 10 years, we and several other groups have reported the existence of bursting cells in the medial and lateral thalamus of awake chronic pain patients undergoing functional stereotactic surgery, and several of these reports have

253 suggested that this activity may be related to the patient’s pain (Lenz et al., 1987,1989, 1994b; Hirayama et al., 1989; Rinaldi et al., 1991; Jeanmonod et al., 1996; Lenz and Dougherty, 1997). These proposals are based on the assumption that the bursting activity observed is not normally present in these regions. During sleep, thalamic neurons tend to fire in a characteristic bursting firing pattern which ceases during wakefulness to be replaced by a more regular firing pattern (see Steriade et al., 1990, 1997). This bursting activity occurs when the cell hyperpolarizes leading to deinactivation of low-threshold calcium channels which subsequently open when the cell depolarizes thus generating a low-threshold calcium spike (LTS) associated with a burst of sodium-generated action potentials. The bursting induced by the LTS has a very characteristic pattern that consists of progressively increasing interspike intervals in each burst. Furthermore, the first interspike interval in this burst has an inverse relationship with the number of spikes in the burst so that as the number of spikes in the burst increases the duration of the first interspike interval becomes shorter (McCarley et al., 1983; Domich et al., 1986). This type of bursting which has been extensively studied in experimental animals has also been shown to occur in the lateral thalamus of patients when they fall asleep (Tsoukatos et al., 1997; Zirh et al., 1998). It is generally assumed that LTS-induced bursting normally occurs only during sleep and thus that its presence in an awake patient signifies a pathological condition. However, we have observed LTS bursting activity in the medial and lateral thalamus of both pain and non-pain patients, suggesting that such activity is not unique to pain and may in fact occur normally in some thalamic neurons during wakefullness. The findings giving rise to these suggestions are described in further detail below. Our recently published analysis of bursting cells in the lateral thalamus was obtained during recordings in 116 functional stereotactic procedures in 91 awake patients (Radhakrishnan et al., 1999). Over 600 cells with bursting activity were recorded in these patients. Approximately one-third of these cells were recorded during 39 procedures in patients who were suffering from pain and the remaining cells were recorded from patients with movement disorders. Fig. 5 shows the number of bursting cells per

track (burst index) for these two groups and for the various subgroups. None of the differences were statistically significant. In order to determine whether the bursting activity was likely due to a LTS we analyzed in detail the firing pattern for a subset of 79 cells (47 from pain patients and 32 from non-pain patients) by plotting the relationship of successive interspike intervals within a burst and the effect of increasing burst size (see examples in Fig. 6). In addition, regression analysis of the first interspike interval as a function of burst size indicated that there was a statistically significant correlation of the first interspike interval with burst size (Fig. 6). This analysis revealed that the bursting activity of at least 50% of the cells identified as bursting cells in each of the two groups of patients was likely due to an underlying calcium spike. The bursting cells were most commonly found in the thalamic region located dorsal and/or anterior to the cells of the motor thalamus that responded to movements. Lower densities of bursting cells were found in the motor thalamus (Vop, Vim) and in the region posterior and/or ventral to the Vc. Surprisingly, the lowest concentration of bursting cells was in the Vc. Interestingly, this pattern was the same for both pain patients and motor patients, suggesting that the activity of these cells was not related to the chronic pain. In the medial thalamus, there have also been reports of the existence of many bursting cells in the awake chronic pain patient. We now have evidence that many bursting cells can also be found in the medial thalamus of epilepsy patients (unpublished observations). In both groups, the firing pattern is consistent with generation by a calcium spike. Also Jeanmonod et al. (1996) have reported the existence of bursting cells in the medial thalamus in patients who did not have chronic pain. Bursting activity in the awake state could be indicative of an abnormal activity resulting from pathophysiology, as has been suggested by Jeanmond (Jeanmonod et al., 1996) and recently by Llinas and colleagues (Llinas et al., 1999). Alternatively, bursting activity may occur normally in some cells and regions of the thalamus in various states other than sleeping (e.g. Sherman and Guillery, 1996). We propose that both situations may

254

AD

PP

Other Pain

PD

Other Tremor

Patient

MS

Pain Non-Pain Group Group

Groups

Fig. 5. Comparison of the burst indices from the pain and non-pain groups of patients. The pain group consisted of patients with anesthesia dolorosa (AD), phantom pain (PP), and other pain disorders (deafferentation pain, trigeminal neuralgia, post-stroke pain, arthritic pain and spinal cord injury). The non-pain group consisted of Parkinson’s disease (PD), other tremor conditions (essential tremor, cerebellar tremor and ballistic tremor) and multiple sclerosis (MS). The left panel of the figure shows the data (mean i S.E.M.) from individual patient groups and the right panel shows the collective data from the pain and non-pain groups, The values in parentheses denote the number of explorations for each group. Reproduced with permission from Radhakrishnan et al. (1999).

occur. It is likely that deafferentation results in decreased excitation and/or increased inhibitory inputs to thalamic cells thus leading to hyperpolarization and LTS bursting. Clearly, many of the pain patients in our group had sustained some central damage that could have led to such changes. Indeed, both we and Lenz and colleagues (Lenz et al., 1994b, 1998) have observed bursting activity in regions presumed to have been deafferented in pain patients. Deafferentation-induced bursting could also have occurred in the motor disorder group, in particular in the multiple sclerosis patients. However the occurrence of bursting cells in these various patient groups, if the result of pathology, should occur in different parts of the thalamus depending on the pathology. Our overall analysis failed to find differential localization of bursting cells in different patient groups. Thus, it seems likely that cells in some thalamic regions

typically fire in bursts in the awake state, even in the absence of pathophysiology, since we usually find bursting activity in similar locations in patients with different types of disorders. In order to claim that bursting activity is related to pain, one would, at the very least, have to show that the bursting activity was in a region involving the ascending pain pathway. At the present time, we do not know with certainty which thalamic regions are involved in mediating acute pain and chronic pain. Although the Vc is likely to play a role, the majority of the cells there are involved in the relay of non-painful tactile sensations. Furthermore, at least in our study and that of Ohye and Narabayashi (1972), only a few of the bursting cells were located in the Vc. In summary, our observations reveal that bursting activity is present in the thalamus of awake non-

255

Anesthesia

Dolorosa --A-ts

2 ISl(n=l7) 3 ISI(n=16) 4 ISI(n=ll)

8

z.

6

3 4 2

+

0

OJ 0

1

2

3

4

0

5

1

3

4

Ordinal number of ISI in the burst

Number of ISIS in the burst Essential Tremor Slope : -0.48 r : 1.00

5

2

E4

El

10 -

-

I ISI(n=13) ZISl(n=ll) 3 ISl(n=3) 4 ISl(n=l)

E"

g

3

tij .L LL

2

g

6

41

g=+Ic---

201

0 0

1

2

3

4

0

1

2

Ordinal number

Number of ISIS in the burst

3

4

of ISI in the burst

Fig. 6. Examples of firing pattern in a bursting cell in a chronic pain patient (anesthesia dolorosa) and a motor disorder patient (essential tremor). The left panel of the figure shows a plot of the mean of the first interspike interval (ISI) for bursts of different sizes plotted as a function of the total number of interspike intervals in the burst. The right panel illustrates the relation of successive ISIS within a burst for bursts of 2, 3, 4 and 5 spikes. In both cases, the abscissa represents the number of spikes in the burst. Note that the ISI increases for successive intervals within a burst and that the first ISI decreases as the number of spikes in the burst increases as expected for LTS mediated bursts. Modified with permission from Radhakrishnan et al. (1999).

pain patients and is similar with respect to incidence, location and the intraburst firing patterns to that observed in pain patients. Although bursting remains an attractive mechanism to explain chronic pain, the present data indicate that bursting activity is a more general phenomenon occurring in other patient groups. If nociceptive neurons in the pain pathway that are normally silent started to fire in a bursting fashion as a result of pathological changes, then this could explain the occurrence of chronic pain. How-

ever, it has yet to be demonstrated that in chronic pain patients, there exist bursting thalamic cells that are involved in the pain pathway. Abbreviations CL COLD HPC MDvc

central lamina lamina ventral

lateral nucleus I thermoreceptive-specific cell I polymodal nociceptive cell caudal part of the medial dorsal

256

NS NSP Pf PSP STT vc vcpc VMpo VP VP1 WDR

nucleus nociceptive-specific cell non-stroke pain patient parafascicular nucleus post-stroke pain patient spinothalamic tract ventral caudal nucleus ventral caudal parvocellularis nucleus posterior part of the ventral medial nucleus ventral posterior nucleus ventral posterior inferior nucleus wide dynamic range cell

References Al-Chaer, E.D., Feng, Y. and Willis, W.D. (1998) A role for the dorsal column in nociceptive visceral input into the thalamus of primates. J. Neurophysiol., 79: 3143-3150. Albe-Fessard, D. and Longden, A. (1983) Use of an animal model to evaluate the origin of and protection against deafferentation pain. In: J.J. Bonica, U. Lindblom and A. Iggo (Eds.), Advances in Pain Research and Therapy. Raven Press, New York, pp. 691-700. Apkarian, A.V. and Shi, T. (1994) Squirrel monkey lateral thalamus. I. Somatic nociresponsive neurons and their relation to spinothalamic terminals. J. Neurosci., 14: 6779-6795. Black, R.G. (1974) A laboratory model for trigeminal neuralgia. Adv. Neurol., 4: 651-659. Bogousslavsky, J., Regli, F. and Uske, A. (1988) Thalamic infarcts: clinical syndromes, etiology, and prognosis. Neurology, 38: 837-848. Boivie, J.O. (1994) Central pain. In: P.D. Wall and R. Melzack (Eds.), Texbook of Pain. Churchill Livingstone, Edinburgh, pp. 871-902. Bushnell, M.C. and Duncan, G.H. (1989) Sensory and affective aspects of pain perception: is medial thalamus restricted to emotional issues. Exp. Brain Res., 78: 415-418. Bushnell, M.C., Duncan, G.H. and Tremblay, N. (1993) Thalamic VPM nucleus in the behaving monkey. I. Multimodal and discriminative properties of thermosensitive neurons. J. Neurophysiol., 69: 7399152. Casey, K.L. (1966) Unit analysis of nociceptive mechanisms in the thalamus of the awake squirrel monkey. J. Neurophysiol., 29: 727-750. Chandler, M.J., Hobbs, S.F., Fu, Q.-G., Kenshalo Jr., D.R., Blair, R.W. and Foreman, R.D. (1992) Responses of neurons in ventroposterolateral nucleus of primate thalamus to urinary bladder distension. Brain Res., 571: 26-34. Craig, A.D. (1995) Supraspinal projections of lamina 1 neurons. In: J.-M. Besson, G. Guilbaud and H. Ollat (Eds.), Forebrain Processing of Pain. Libbey Eurotext, Paris, pp. 13-25. Craig, A.D. (1998) A new version of the thalamic disinhibition hypothesis of central pain. Pain Forum, 7: I-14.

Craig, A.D. (2000) The functional anatomy of lamina 1 and its role in post-stroke central pain. In: J. Sandktihler, B. Bromm and G.F. Gebhart (Eds.), Nervous System Plasticity and Chronic Pain, Progress in Brain Research, Vol. 129. Elsevier Science, Amsterdam, pp. 137-151. Craig, A.D. and Dostrovsky, J.O. (1997) Processing of nociceptive information at supraspinal levels. In: T.L. Yaksh (Eds.), Anesthesia: Biologic Foundations. LippincottRaven, Philadelphia, PA, pp. 625-642. Craig, A.D. and Dostrovsky, J.O. (1999) Medulla to Thalamus. In: P.D. Wall and R. Melzack (Eds.), Textbook of Pain. Churchill-Livingstone, Edinburgh, pp. 183-214. Craig, A.D. and Zhang, E.-T. (1996) Anterior cingulate projection from MDvc (a lamina I spinothalamic target in the medial thalamus of the monkey). Sot. Neurosci. A&r:, 22: 1 I I. Craig, A.D., Bushnell, M.C., Zhang, E.-T. and Blomqvist, A. (1994) A thalamic nucleus specific for pain and temperature sensation. Nature, 372: 770-173. Davis, K.D., Kiss, Z.H.T., Tasker, R.R. and Dostrovsky, J.O. (1996) Thalamic stimulation-evoked sensations in chronic pain patients and in nonpain (movement disorder) patients. J. Neurophysiol., 75: 1026-1037. Davis, K.D., Lozano, R.M., Manduch, M., Tasker, R.R., Kiss, Z.H.T. and Dostrovsky, J.O. (1999) Thalamic relay site for cold perception in humans. J. Neurophysiol., 81: 1970- 1973. Domich, L., Oakson, G. and Steriade, M. (1986) Thalamic burst patterns in the naturally sleeping cat: a comparison between cortically projecting and reticularis neurones. .I. Physiol., 379: 429-449. Dostrovsky, J.O., Wells, F.E.B. and Tasker, R.R. (1992) Pain sensations evoked by stimulation in human thalamus. In: R. Inoka, Y. Shigenaga and M. Tohyama (Eds.), Processing and Inhibition of Nociceptive Information, International Congress Series $89. Excerpta Medica, Amsterdam, pp. 115- 120. Dostrovsky, J.O., Manduch, M., Davis, K.D., Tasker, R.R. and Lozano, A.M. (2000) Thalamic stimulation-evoked pain and temperature sites in pain and non-pain patients. In: M. Devor, M. Rowbotham and Z. Wiesenfeld-Hallin (Eds.), Proceedings of the 9th World Congress on Pain. IASP Press, Seattle, pp. 419-425. Duncan, G.H., Bushnell, MC., Oliveras, J.-L., Bastrash, N. and Tremblay, N. (1993) Thalamic VPM nucleus in the behaving monkey. III. Effects of reversible inactivation by lidocaine on thermal and mechanical discrimination. J. Neurophysiol., 70: 2086-2096. Gybels, J.M. and Sweet, W.H. (1989) Neurosurgical treatment of persistent pain. Physiological and pathological mechanisms of human pain. Pain Headache, 11: 1-442. Halliday, A.M. and Logue, V. (1972) Painful sensations evoked by electrical stimulation in the thalamus. In G.G. Somjen (Ed.), Neurophysiology Studied in Man. Excerpta Medica, Amsterdam, pp. 221-230. Hassler, R. and Riechert, T. (1959) Klinische und anatomische Befunde bei stereotaktischen Schmerzoperationen im Thalamus. Arch. Psychiattz, 200: 93-122. Hirayama, T., Dostrovsky, J.O., Gore&i, .I., Tasker, R.R. and Lenz, EA. (1989) Recordings of abnormal activity in patients

257 with deafferentation and central pain. Stereotact. Funct. Neumsurg., 52: 120-126. Jeanmonod, D., Magnin, M. and Morel, A. (1996) Low-threshold calcium spike bursts in the human thalamus. Common physiopathology for sensory, motor and limbic positive symptoms. Brain, 119: 363-375. Kenshalo Jr., D.R., Giesler Jr., G.J., Leonard, R.B. and Willis, W.D. (1980) Responses of neurons in primate ventral posterior lateral nucleus to noxious stimuli. .I. Neurophysiol., 43: 15941614. Lenz, EA. and Dougherty, P.M. (1997) Pain processing in the human thalamus. In: M. Steriade, E.G. Jones and D.A. McCormick (Eds.), Thalamus, Experimental and Clinical Aspects, Vol II. Elsevier Science, Amsterdam, pp. 617-652. Lenz, EA., Tasker, R.R., Dostrovsky, J.O., Kwan, H.C., Gore&i, J., Hirayama, T. and Murphy, J.T. (1987) Abnormal singleunit activity recorded in the somatosensory tbalamus of a quadriplegic patient with central pain. Pain, 31: 225-236. Lenz, EA., Kwan, H.C., Dostrovsky, J.O. and Tasker, R.R. (1989) Characteristics of the bursting pattern of action potentials that occurs in the thalamus of patients with central pain. Brain Res., 496: 357-360. Lenz, EA., Gracely, R.H., Rowland, L.H. and Dougherty, P.M. (1994a) A population of cells in human thalamic principal sensory nucleus respond to painful mechanical stimuli. Neurosci. Lett., 180: 46-50. Lenz, EA., Seike, M., Lin, YC., Baker, F.H., Rowland, L.H., Gracely, R.H. and Richardson, R.T. (1993a) Neurons in the area of human thalamic nucleus ventralis caudalis respond to painful heat stimuli. Brain Res., 623: 235-240. Lenz, EA., Seike, M., Richardson, R.T., Lin, Y.C., Baker, F.H., Khoja, I., Jaeger, C.J. and Gracely, R.H. (1993b) Thermal and pain sensations evoked by microstimulation in the area of human ventrocaudal nucleus. .I. Neurophysiol., 70: 200-212. Lenz, EA., Kwan, H.C., Martin, R., Tasker, R., Richardson, R.T. and Dostrovsky, J.O. (1994b) Characteristics of somatotopic organization and spontaneous neuronal activity in the region of the thalamic principal sensory nucleus in patients with spinal cord transection. J. Neurophysiol., 72: 1570-1587. Lenz, EA., Gracely, R.H., Baker, EH., Richardson, R.T. and Dougherty, P.M. (1998) Reorganization of sensory modalities evoked by microstimulation in region of the thalamic principal sensory nucleus in patients with pain due to nervous system injury. J. Comp Neural., 399: 125-138. Lenz, EA., Lee, J.-I., Garonzik, I.-M., Rowland L.H., Dougherty, P.M. and Hua, S.E. (2000) Human thalamus reorganization related to nervous system injury and dystonia. In J. Sandkiihler, B. Bromm and G.F. Gebhart (Eds.), Nervous System Plasticity and Chronic Pain, Progress in Brain Research, Vol. 129. Elsevier Science, Amsterdam, pp. 259-273. Llinas, R.R., Ribary, U., Jeanmonod, D., Kronberg, E. and Mitra, PP. (1999) Neuropsychiatric syndrome characterized by mag-

netoencephalography. Proc. Natl. Acad. Sci. USA, 96: 1522215227. Loeser, J.D. and Ward Jr., A.A. (1967) Some effects of deafferentation on neurons of the cat spinal cord. Arch. Neurol., 17: 629-636. Loeser, J.D., Ward, A.A.J. and White Jr., L.E. (1968) Chronic deafferentation of human spinal cord neurons. J. Neurosurg., 29: 48-50. Lombard, MC. and Larabi, Y. (1983) Electrophysiological study of cervical dorsal horn cells in partially deafferented rats. In: J.J. Bonica, U. Lindblom and A. Iggo (Eds.), Advances in Pain Research and Therapy. Raven Press, New York, pp. 147-154. McCarley, R.W., Benoit, 0. and Barrionuevo, G. (1983) Lateral geniculate nucleus unitary discharge in sleep and waking: state- and rate-specific aspects. J. Neurophysiol., 50: 798-818. Ohye, C. and Narabayashi, H. (1972) Activity of thalamic neurons and their receptive fields in different functional states in man. In: G.G. Somjen (Ed.), Neurophysiology Studied in Man. Excerpta Medica, Amsterdam, pp. 79-84. Radhakrishnan, V., Tsoukatos, J., Davis, K.D., Tasker, R.R., Lozano, A.M. and Dostrovsky, J.O. (1999) A comparison of the burst activity of lateral thalamic neurons in chronic pain and non-pain patients. Pain, 80: 567-575. Rinaldi, P.C., Young, R.F., Albe-Fessard, D. and Chodakiewitz, J. (1991) Spontaneous neuronal hyperactivity in the medial and intralaminar thalamic nuclei of patients with deafferentation pain. J. Neurosurg., 74: 415-421. Sherman, S.M. and Guillery, R.W. (1996) Functional organization of thalamocortical relays. J. Neurophysiol., 76: 13671395. Steriade, M., Jones, E.G. and Llinas, R.R. (1990) Thalamic Oscillations and Signalling. Wiley, New York. Steriade, M., Jones, E.G. and McCormick, D.A. (1997) Thalamus. Elsevier Science, Oxford. Tasker, R.R. (1984) Stereotaxic surgery. In: P.D. Wall and R. Melzack (Eds.), Textbook of Pain. Churchill Livingstone, Edinburgh, pp. 639-655. Tsoukatos, .I., Kiss, Z.H.T., Davis, K.D., Tasker, R.R. and Dostrovsky, J.O. (1997) Patterns of neuronal firing in the human lateral thalamus during sleep and wakefulness. Exp. Brain Res., 113: 273-282. Willis, W.D. (1997) Nociceptive functions of thalamic neurons. In: M. Steriade, E.G. Jones and D.A. McCormick (Eds.), Thalamus, Experimental and Clinical Aspects, Vol. II. Elsevier Science, Amsterdam, pp. 373-424. Willis, W.D., Jr. (1985) The Pain System. The neural basis of nociceptive transmission in the mammalian nervous system. In: PL. Gildenberg (Ed.), Pain and Headache. Karger, Basle, pp. 1-346. Zirh, T.A., Lenz, EA., Reich, S.G. and Dougherty, P.M. (1998) Patterns of bursting occurring in thalamic cells during parkinsonian tremor. Neuroscience, 83: 107-121.

J. Sandkiihler, B. Bromm and GE Gebhart (Eds.) Progress in Brain Research, Vol. 129 0 2000 Elsevier Science B.V. All rights reserved

CHAPTER 18

Plasticity of pain-related neuronal activity in the human thalamus EA. Lenz ‘T*,J.-I. Lee 2, I.M. Garonzik I, L.H. Rowland ‘, P.M. Dougherty 1 and S.E. Hua ’ 2 Department

I Department of Neurosurgery,

of Neurosurgery, Samsung Medical

Johns Hopkins University, Baltimore, MD, USA Center; Sungkyunkwan University School of Medicine,

Introduction

Studies of plasticity of the somatosensory system in non-human primates have focussed on maps of cortical function determined by examining neuronal receptive fields (RFs) (Kaas, 1991). Studies in humans can explore both maps determined from RFs (RF maps) and maps of the locations and quality of sensations evoked by stimulation of the brain (projected field or PF maps) (Lenz et al., 1994a, 1998a). While RF maps are a reflection of the organization of inputs to the central nervous system PF maps give an indication of the image of the body contained in the thalamus and cortex. There are numerous studies of cortical plasticity secondary to injuries of the nervous system (Kaas, 1991) and motor behavior (Jenkins et al., 1990; Nudo et al., 1996; Recanzone et al., 1992). Although much of the reorganization of cortical maps may reflect changes in thalamic organization (Pons et al., 1991), there are relatively few studies of thalamic plasticity (Garraghty and Kaas, 1991; Rasmusson, 1996a,b; see Dostrovsky, 2000, this volume). We now review our studies of human thalamic plasticity as studied by single cell

*Corresponding author: EA. Lenz, Department of Neurosurgery, Meyer Building 7-113, Johns Hopkins Hosptial, 600 North Wolfe Street, Baltimore, MD 21287-7713, USA. Tel.: +l (410) 9552257; Fax: +l (410) 614-9877; E-mail: [email protected]

Seoul, Korea

recordings and microstimulation in patients undergoing operations for the treatment of chronic pain or movement disorders. Medically intractable chronic pain due to nervous system injury or medically intractable dystonia can be treated, respectively, by implantation of deep brain stimulating electrodes in the principal sensory nucleus of the thalamus (ventral caudal, Vc) (Hosobuchi, 1986; Young et al., 1985) or by thalamotomy (Cardoso et al., 1995; Tasker et al., 1988). During such operations, the location of the target is first defined by radiologic studies. Thereafter, microelectrode explorations may be used to confirm the target predicted by the radiologic studies. Physiologic exploration with the microelectrode involves both recording of neuronal activity and stimulation at microampere current levels as previously described (Lenz et al., 1994a,b). The borders of Vc were explored in patients with chronic pain to determine the location for implantation of the deep brain stimulating electrode for treatment of pain. Vc was also explored in dystonia or tremor patients to determine the anterior and inferior borders of Vc, which predict the borders of the thaiamic nucleus ventral intermediate (Vim) and the nucleus ventral oral posterior (Vop), the nuclei to be lesioned during thalamotomy for dystonia or tremor (Vitek and Lenz, 1998; Zirh et al., 1999). Therefore Vc was thoroughly examined in both patient groups. Analysis was restricted to the cutaneous core area and the posterior-inferior area of the region

260

of Vc. The core was defined as the area where the majority of cells responded to innocuous, cutaneous, mechanical stimuli and probably corresponded to Vc (Lenz et al., 1993a, 1994b, 1998b), the human analog of monkey ventral posterior (VP) (Hirai and Jones, 1989). The posterior inferior area is the cellular area below and behind the core and probably corresponds to the posterior subnucleus of Vc (nucleus ventral caudal portae, Vcpor), the inferior subnucleus of Vc (ventral caudal parvocellular nucleus, Vcpc), posterior nucleus, Craig’s nucleus ventral medial posterior (Vmpo) (Craig et al., 1994) and magnocellular medial geniculate (Lenz et al., 1993a; Mehler, 1966). These physiologic studies provide a unique opportunity to examine thalamic neuronal activity in patients with nervous system injury and abnormal motor behavior. We now report changes in the representation of cutaneous structures in the core of the Vc nucleus of the thalamus in such patients. These studies demonstrate that thalamic responses can be altered by nervous system injury, such as spinal cord transection or amputation, by the experience of painful stimuli or by the abnormal motor behaviors such as dystonia, which is often painful (Tolosa and Marti, 1997). Evidence that the region of Vc is involved in pain processing Several lines of evidence demonstrate that the region of Vc is important in human pain-signaling pathways. Studies of patients at autopsy following

lesions of the spinothalamic tract (STT) show the most dense STT termination in Vc (Bowsher, 1957; Mehler et al., 1960; Mehler, 1962, 1966). Additionally, terminations are observed posterior to Vc in the magnocellular medial geniculate (Mehler, 1962, 1969), limitans, and Vc portae nuclei (Mehler, 1966) and inferior to Vc in Vcpc (Mehler, 1966). STT terminations are found in monkey Vmpo which appears, by immunohistochemistry, to have a human analog (Craig et al., 1994). Cells in Vc have been identified with a differential response to painful thermal and mechanical stimuli (Lee et al., 1999; Lenz et al., 1994~) and with a response to innocuous cool and mechanical stimuli (Lenz and Dougherty, 1998). Fig. 1 shows an example of a cell in Vc responding to noxious thermal (Fig. 1D) and mechanical stimuli (Fig. lB,C). Cells in the posterior inferior region have been identified with a significant selective response to noxious heat stimuli (Lenz et al., 1993b) and to cold stimuli (Davis et al., 1999). These reports extend to humans the results of numerous monkey studies in which cells within VP (Kenshalo et al., 1980; Gautron and Guilbaud, 1982; Casey and Morrow, 1983; Chung et al., 1986; Bushnell and Duncan, 1987; Bushnell et al., 1993; Apkarian and Shi, 1994) and posterior and inferior to VP respond to noxious stimuli (Casey, 1966; Apkarian et al., 1991; Apkarian and Shi, 1994; Craig et al., 1994). Cells in the region of Vc that respond to noxious stimuli probably signal pain based on lesioning and stimulation studies. Blockade of the activity in this

Fig. 1. Activity of a cell (061093) in Vc responding to painful mechanical and thermal stimuli. (A) Location of the cell (arrow) relative to the positions of trajectories, nuclear boundaries, and other recorded cells. The ACPC line is indicated by the horizontal line and the trajectories are shown by the oblique lines (left: anterior; up: dorsal). Nuclear location was approximated from the position of the ACPC line. Lateral location of the cell (in millimeters) is indicated above each map. Trajectories have been shifted along the ACPC line until the most posterior cell with a cutaneous RF is aligned with the posterior border of Vc. Since cells responding to innocuous sensory stimuli may be located posterior to Vc (Apkarian and Shi, 1994), this map represents a first approximation of nuclear location and dimensions. The locations of cells are indicated by ticks to the right of each trajectory. Cells with cutaneous RFs are indicated by long ticks, those without definable RFs by short ticks. Filled circles attached to the long ticks indicate that somatic sensory testing was carried out. The scale is as indicated. The shape of action potentials recorded at the beginning of the recording on this cell during application of the brush (upper) and at the end of the recording, during a 12°C stimulus (lower). Data were collected from upgoing stroke of the action potential by using voltage threshold of 0.15 pV. The RF and PF for the natural, surface and deep, nonpainful, tingling sensation evoked by TMIS at the recording site (threshold -15 PA) are also shown, (B) Response to the brush, large clip, medium clip, and small clip. (C) The response of the neuron to progressive increase in pressure applied with the nonpenetrating towel clip, indicated by the number of steps. (D) Responses to heat stimuli at 42, 45, and 48°C. (E) Responses to cold stimuli at 12, 18, and 24°C. The upper trace in each panel is a footswitch signal indicating the onset and duration of the stimulus in panels B and C and the thermode signal in panels D and E. The scales for the axes for all histograms (binwidth 100 ms) are indicated in each panel. Adapted from Lee et al. (1999) with permission.

261 region by injection of local anesthetic into monkey VP, corresponding to human Vc (Hirai and Jones, 1989), significantly interferes with the monkey’s ability to discriminate temperature in both the innocuous and noxious range (Duncan et al., 1993). Stimulation within Vc and posterior-inferior to it can evoke the sensation of pain (Halliday and Logue,

1972; Dostrovsky et al., 1991; Lenz et al., 1993a) and thermal sensations (Lenz et al., 1993a; Davis et al., 1999). Thus, there is strong evidence that the region of Vc is involved in pain signaling pathways: (i) it receives input from pain signaling pathways, (ii) it contains cells that respond to noxious stimuli,

16 -

5 set

5 set C

3 steps J------u~u~u

6

stn

262 (iii) stimulation can evoke pain, and i.e. temporary lesioning of monkey VP disables the discrimination of pain and temperature. Therefore studies of this region are critical to understanding acute and chronic pain sensations, Plasticity of this area may have a significant impact on processing of chronic pain sensations. Reorganization of Vc after nervous system injury: RF and PF maps The region of the principal sensory nucleus of thalamus (Vc) was explored during stereotactic surgical procedures for treatment of patients with pain following spinal cord transection (spinal patients, n = 5) or of patients with movement disorders (n = 23) (Lenz et al., 1994a). RFs of thalamic single neurons and locations of sensations evoked by stimulation (PFs) were determined by standard methods. Many cells in the representation of the anesthetic part of the body did not have RFs (Lenz et al., 1994a). Therefore, ‘the region of Vc’ was defined as the cellular thalamic region where sensations were evoked at less than 25 LA. The ‘region of Vc’ in spinal patients was subdivided into different areas, according to RF and PF locations. Areas that were distant from the representation of the anesthetic part of the body were termed ‘spinal control’ areas while those that were adjacent to or included in the representation of the area of absolute sensory loss were termed ‘borderzone/anesthetic’ areas. The ‘region of Vc’ in movement disorder patients was termed the ‘control’ area. Fig. 2 shows RF and PF maps in a patient with a Ts spinal cord transection. In control and spinal control areas (Fig. 2A) the locations of

RFs and PFs were usually well matched. However, in borderzone/anesthetic areas of the thalamus (Fig. 2B), there was frequently a mismatch between the location of RFs and PFs (RF/PF mismatch). RFs are located on the chest and abdominal wall, above the level of the spinal transection, while PFs are located in the lower extremity, below the spinal transection. Since the anesthetic part of the body was normally represented, the perceptual image of that part of the body is unchanged. While the perceptual image of the anesthetic part of the body is relatively constant, the RF has changed to represent the borderzone of the anesthetic part of the body. Similar changes are seen in patients with amputations (Lenz et al., 1998a). Borderzone/anesthetic areas of thalamus often exhibited increased representations of the border of the anesthetic part of the body in comparison to the representation of the same parts of the body in control and spinal control areas. Two sites were said to have a consistent RF if the RF of both sites included the same part of the body. The length of a trajectory with consistent RFs is the distance along the trajectory where each RF continues to include the same part of the body. The part of the body chosen for any trajectory length was the part that maximized the length of the trajectory with a consistent RF. Human thalamic somatotopy in Vc is a function of laterality of the parasagittal plane. Rows of trajectories were aligned in parasagittal planes (Lenz et al., 1988a). The maximal distance along a trajectory over which the RF or the PF stays consistent is longer for body parts with larger representations (Lenz et al., 1994a). Lengths of trajectory with a consistent RF in a particular part of the body were significantly longer in borderzone anesthetic than in control or spinal control areas (Fig. 3).

Fig. 2. Map of receptive and projected fields for trajectories in the region of the Vc in a patient with spinal cord transection at Ts. (A) A trajectory in the 15 mm lateral parasagittal plane (labelled Lat 15 mm) through the ‘region of Vc’ which represents the arm. (B) A trajectory 2 mm lateral to the first (Lat 17 mm). In A and B the upper panel shows the position of the trajectory relative to nuclear boundaries as predicted radiologically. The anterior commissure-posterior commissure (AC-PC) line is indicated by the horizontal line in the panel while the trajectories are shown by the oblique lines. The locations of cells are indicated by ticks to the right of the trajectory. Cells with receptive fields (RF) are indicated by long ticks; those without are indicated by short ticks. Stimulation sites are shown to the left of the trajectory. Long ticks indicate a somatosensory response to stimulation while short ticks indicate no response to microstimulation. Scales are as indicated. The lower panel in A and B shows paired figurines for sites as numbered in the middle panel. The figurine to the right indicates the RF, NR indicates that the cell had no RP. The figurine to the left indicates the projected field (PF) for TMS at that site while the number below the figurine indicates the threshold in uA. At all sites along both trajectories where sensations were evoked that sensation was described as tingling. Adapted from Lenz et al. (1994a) with permission.

263

s4 PF NR

1

RF

PF

RF

RF

PF

I

5

NR I 6

PF

RF

264

CUTANEOUS

MONKEY

3 .

VP- CORONAL

DEEP rbc

1

Receptive Field (RF) Fig. 3. Lengths of trajectories with consistent receptive fields (RFs) in controls and in patients with spinal transection. Results for cutaneous RFs are shown in the large plot while deep RFs are plotted in the lower inset labelled deep. On both plots the dots indicate control areas, filled circles represent borderzone/anesthetic areas and open circles represent spinal control areas. Lengths include the part of a trajectory bounded by RFs located in the same body part. These lengths may include cells without RFs but not cells with RFs outside the body part indicated by the label along the horizontal axis. The upper inset is a cartoon (adapted from Jones and Friedman, 1982, with permission) showing representation of body parts in monkey VP, corresponding to human Vc (Hirai and Jones, 1989). Adapted from Lenz et al. (1994a) with permission.

Neurons with RFs adjacent to the area of sensory loss in amputation patients (n = 3) occupied a larger part of the thalamic homunculus (Lenz et al., 1998a) than found for the same part of the body in patients with movement disorders (Lenz et al., 1988b, 1994a). This result is consistent with somatotopic reorganization of afferent inputs from the limb. Similarly the large area over which PFs include the

stump (cf. Lenz et al., 1994a, 1998b) suggest that there has been reorganization of the perceptual image of the limb in the central nervous system (Jensen and Rasmussen, 1994). It has also been reported that phantom sensations can be evoked by stimulation of the region where stump RFs are located in the region of Vc (Davis et al., 1998). Thus, in the case of amputations and spinal cord injury the alteration in the

image of the body in the thalamus is less than that of the inputs from that part of the body. See Vierck and Light (2000, this volume) and Yezierski (2000, this volume) for animal models of spinal cord injuries. Reorganization of Vc after nervous system injury: modalities of sensations evoked by thalamic stimulation Our studies have shown that stimulation of the somatic sensory thalamus is more likely to evoke pain in patients with chronic pain (n = 12) after nervous system injury than in patients without somatic sensory abnormalities (patients with movement disorders, n = 10) (Lenz et al., 1998b). Patients were trained preoperatively to use a standard questionnaire to describe the location (projected field) and quality of sensations evoked by threshold microstimulation intraoperatively. The region of Vc was divided on the basis of projected fields into areas representing the part of the body where the patients experienced chronic pain (pain affected) or did not experience chronic pain (pain unaffected) and into a control area located in the thalamus of patients with movement disorders and no experience of chronic pain. The region of the Vc was also divided into a core region and a posterior inferior region. The core was defined as the region above a standard radiologic horizontal (anterior commissure-posterior commissure line, AC-PC line) where the majority of cells responded to innocuous somatosensory stimulation. The posterior inferior area was a cellular area posterior and inferior to the core. In both the core and posterior inferior regions, the proportion of sites where threshold microstimulation evoked pain was larger in pain affected and unaffected areas than in control areas (Fig. 4). The number of sites where thermal (warm or cold) sensations were evoked was correspondingly smaller, so that the total of pain plus thermal sites was not significantly different across all areas (Lenz et al., 1998b). Therefore, sites where stimulation evoked pain in patients with neuropathic pain may correspond to sites where thermal sensations were evoked by stimulation, in patients without somatic sensory abnormality. In the posterior inferior region the number of sites where cold was evoked by stimulation decreased significantly while the number of sites

where pain was evoked increased significantly. In the core region the number of sites where warm was evoked decreased while sites where pain was evoked increased. The present results suggest that pain is evoked in patients with neuropatbic pain by stimulation at sites where thermal sensations would normally be evoked. Therefore, the present data suggest that the SIT or elements to which the STT projects signal pain rather than thermal sensations in patients with neuropathic pain. This is consistent with the finding that stimulation of the STT evokes pain in patients with neuropathic pain (Tasker, 1982), but evokes nonpainful thermal sensations in patients who do not have neuropathic pain (Tasker, 1988). Anterolateral cordotomy relieves pain in a much greater proportion of patients with somatic pain than it does in patients with neuropathic pain (Tasker et al., 1980; Sweet et al., 1994). The failure of cordotomy to relieve neuropathic pain might be anticipated from the occurrence of central pain in patients with impaired function of the STT (Cassinari and Pagni, 1969; Beric et al., 1988; Boivie et al.? 1989; Andersen et al., 1995). These results suggest that the generator for pain in patients with central pain is the terminus of the STT. In patients with central pain, anatomic evidence of damage to STT is a common finding (Cassinari and Pagni, 1969) and loss of STT function, indicated by impaired thermal and pain sensibility, is a uniform finding (Beric et al., 1988; Boivie et al., 1989; Andersen et al., 1995). In patients with central pain, pain is more common than in controls while threshold microstimulation-evoked cold sensations are correspondingly less common. These findings suggest that there has been a reorganization so that cold modalities are relabeled to signal pain in the thalamus of patients with central pain. This reorganization might occur as a response to STT injury since dramatic changes in thalamic anatomy can result from interruption of sensory input (Rausell et al., 1992; Ralston et al., 1996). The relationship between thalamic stimulation-evoked cold and pain in patients with central pain may explain the perception of cold as pain (cold hyperalgesia) which can occur in these patients (Boivie et al., 1989)

266

A

Region of Vc

32

”

Control n=l39

Pain .affected n=lO8

Pain ;y$cted

Posterior Inferior Region

0 Control n=89

C

Pain unaffected n=64

Core Region

32

4

Pain affected n=80

r

Control n=50

Pain affected n=28

Pain unaffected n=73

Fig. 4. Percentages of pain, cold and warm sensations evoked by stimulation in the core (C) and posterior inferior areas separately (B) and combined (A). These percentages are shown for movement disorder patients (control) and for areas of thalamus representing the part of the body where the patient does (pain affected) or does not experience chronic pain (pain unaffected). Adapted from Lenz et al. (1998b) with permission.

267

Reorganization of RF and PF maps of Vc in patients with dystonia The reorganization of RF and PF maps in Vc occurs both in response to injuries of the nervous system and in response to motor behaviors (Nudo et al., 1996). We have studied the organization of Vc in patients with dystonia, an abnormal motor behavior characterized by sustained muscle contractions that lead to twisting movements and abnormal postures (Fahn, 1988). A wide range of observations suggest that sensory inputs play a significant role in dystonia (Hallett, 1995). For example, the map of the hand representation in the primary sensory cortex (area 3b) is altered in monkeys with dystonia-like movements resulting from overtraining in a gripping task (By1 et al., 1997). We have studied whether similar reorganization occurs in the somatic sensory thalamus of patients with dystonia (dystonia patients) (Lenz and Byl, 1999). Recordings of neuronal activity and microstimulation-evoked responses were studied in the cutaneous core of Vc in 11 dystonia patients who underwent stereotactic thalamotomy. Fifteen patients with essential tremor who underwent similar procedures were used as controls. The cutaneous core of Vc was defined as the part of the cellular thalamic region where the majority of cells had receptive fields (RFs) to innocuous cutaneous stimuli. A map for a patient with dystonia is shown in Fig. 5. The trajectory traversed the representation of cutaneous structures in the hand in the core of Vc (sites 14 to 22). In this patient, 55% (5/9) cells in the cutaneous core of Vc had multiple-part RFs (thenar or hypothenar and palm, fourth/fifth and thumb/second digit). This patient had multiple-part PFs at 88% (7/X) of sites in the cutaneous core of Vc. Multiple-part RFs and PFs were more common in dystonia patients (Fig. 5), than in patients with tremor. Multiple-part PFs were more common than multiple-part RFs. In the patient shown in Fig. 5, the RF representation of the fifth digit covered 1.6 mm (sites 18 to 22) while the PF distribution of the fifth finger (sites 14 to 20) covered 3.1 mm. These are large distributions considering that the RF representation of the fifth digit is normally the smallest of all of the digits (Lenz et al., 1994a; Zirh et al., 1996). Examples of mismatches are seen at sites 17 and 22.

The proportion of RFs including multiple parts of the body was greater in dystonia patients (29%) than in patients with essential tremor (11%). Similarly, the percentage of projected fields (PFs) including multiple body parts was higher in dystonia patients (71%) than in patients with essential tremor (41%). A match at a thalamic site was said to occur if the RF and PF at that site included a body part in common. Such matches were significantly less prevalent in dystonia patients (33%) than in patients with essential tremor (58%). The average length of the trajectory where the PF included a consistent, cutaneous RF was significantly longer in patients with dystonia than in control patients with essential tremor. Summary of reorganization of the human principal thalamic sensory thalamic nucleus In primates there are well documented alterations in thalamic anatomy and physiology after peripheral nerve injury. The distributions of thalamic cytochrome oxidase (CO) staining and calcium binding proteins are both altered (Rausell et al., 1992) in monkeys with a C&--T4 dorsal rhizotomy (Sweet, 1981; Levitt, 1985). In the affected arm area of rhizotomized animals there is a reduction in the density of large cells, and of parvalbumin and CO staining, all characteristic of the terminal zone for dorsal column inputs. There is corresponding increase in the calbindin staining in the arm area, characteristic of the terminal zone for STT inputs. The area of Vc occupied by cells with RFs representing the stump is dramatically increased in Vc of patients with amputations (Lenz et al., 1998a). After cervical dorsal rhizotomy, large numbers of cells without RFs are encountered (Lombard et al., 1979; Albe-Fessard and Lombard, 1983) in the forelimb region of the monkey VP Following adult digit amputation (Rasmusson, 1996a,b) increased representation of the stump is found with large RFs including adjacent digits (Rasmusson, 1996a,b). The thalamic representation of the border of the anesthetic part is increased in monkeys with nerve sections (Garraghty and Kaas, 1991). Thus results of studies in humans are consistent with those in animals. Reorganization of cortical activity has also been studied after interruption of peripheral nerves. These

268 B

C

PF

.4J iI&8\ ‘I c ‘Q 4 RF

PF

PF

RF

RF

17

10

I I I

18

gi?l?&

10

10

'I\\

4!u& 8

19 b 20

.12 '\\

II 13

I I

Lb

10

I 21

5

I 22

\

14 8

15

i

5

I

23

NR 8

I 16

I

studies have shown reorganization of cortical somatotopy over distances of up to 2 mm in macaques (Kaas et al., 1983; Merzenich et al., 1983a) (for

I NR

cortical reorganization in human amputees see Flor, 2000, this volume). Many cells did not have RFs, suggesting that incomplete reactivation occurs in

269

cortex after peripheral nerve section (Rasmusson, 1982; Kaas et al., 1983; Kelahan and Doestch, 1984; Wall and Cusick, 1984). Following a Cz-Tb dorsal rhizotomy in macaques (Pons et al., 1991), shifts of 1 to 2 cm were observed in the map of inputs to cortex. This is an order of magnitude larger than shifts observed in more limited nerve sections (Kaas et al., 1983; Kelahan and Doestch, 1984; Wall and Cusick, 1984). Similar changes are observed after central nervous system injury. In patients with spinal transection, the numbers of cellular RFs representing the border of the anesthetic part of the body are increased (Lenz et al., 1994a). Following transection of the dorsal columns at the T3-Ts level, activity in simian VP (Pollin and Albe-Fessard, 1979) shows a decrease in the percentage of cells with hindlimb RFs and an increase in the percentage of cells with forelimb RFs (see also Vierck and Light, 2000, this volume; Yezierski, 2000, this volume, for animal models of spinal cord injury). Many cells have large RFs and respond to high threshold inputs, consistent with inputs from STT. Therefore, loss of input due to peripheral or central nervous system injury leads to significant reorganization of the thalamic representation of inputs from different parts of the body. Human studies in patients with dystonia also show alterations in RF and PF maps. In monkeys, sensory cortex has been studied during dystonia-like movements induced by repetition of a motor task involving rapid opening and closing of a manipulandum (Hallett, 1995; By1 et al., 1996a, 1997). These monkeys developed hand cramps characterized by disordered motor coordination and posturing reminiscent of dystonia. In these animals, representations of the hand surface are remodeled

in the primary sensory cortex, area 3b (By1 et al., 1996a). The cutaneous RFs extend across multiple digits and the whole hand. This somatotopic reorganization is strikingly different from the normal representation of the hand of the adult monkey, which is defined by small, distinct, orderly, topographical RFs (Merzenich et al., 1987; Kaas, 1991). The present results demonstrate that similar changes in the thalamic cutaneous representation occur in dystonia patients. Alteration in RF maps has also been observed in studies of non-human primates trained to different motor behaviors. For example, behavioral training at a constant skin locus (e.g., tip of one digit, [Recanzone et al., 19921) and behavioral training in which stimuli move across the skin (Jenkins et al., 1990) produced changes in the representation of the body in somatic sensory cortex. Cortical plasticity induced by behavior can be characterized by an increased differentiation of the representation of the body part, including unusually small RFs, and by an increased area of somatic sensory representation of the body part that is involved in the behavior (Jenkins et al., 1990; Recanzone et al., 1992). Alternatively, de-differentiation of the representation of the body can occur and is characterized by unusually large RFs and a decreased area of somatic sensory representation (Merzenich et al., 1983b; Wang et al., 1995). Reorganization of the motor cortical representation can also result from learned repetitive motor tasks (Nudo et al., 1992, 1996). The learning hypothesis for focal hand dystonia suggests that repetitive sensory, nearly simultaneous inputs lead to a degradation of the somatic sensory representation of the hand. This mechanism may apply in patients who perform repetitive jobs (e.g.,

Fig. 5. Parasagittal map of cutaneous receptive fields (RFs) and projected fields (PFs) in the region of the Vc in a patient with dystonia. (A) Position of the trajectory relative to the anterior commissure-posterior commissure (AC-PC) line. The AC-PC line is indicated by the nearly horizontal line in the panel; the trajectory is shown by the oblique line. Anterior: left side of figure; dorsal: top side. The positions of the nuclei are inferred from the AC-PC line and therefore are only an approximate indication of nuclear location. (B) Location of the cells, stimulation sites, and trajectories (S2) relative to the AC-PC. The locations of the stimulation sites are indicated by ticks to the left of the trajectory and those of recorded cells by ticks to the right of the trajectory. Cells with receptive fields are indicated by long ticks and those without by short ticks. Scales are as indicated. Each site where a cell was recorded, or stimulation was carried out, or both is indicated by the same number in B and C. (C) Site numbers, with PF to the left of the vertical and RF to the right for a site. NR indicates no activity related to active movement or sensory stimulation. A phasic dystonia response is an increase in cellular activity occurring at the time of a transient increase in the patient’s dystonia, usually related to active movement. Adapted from Lenz and By1 (1999) with permission.

270 data entry clerks, musicians) under conditions of high cognitive drive (By1 et al., 1996a,b). A similar mechanism could apply to patients with generalized dystonia if the dystonic movements lead to nearly simultaneous stimulation of the afferents from different cutaneous structures or from different muscle groups. It is unclear, however, whether the observed changes are merely a consequence of dystonia. Recent studies of reorganized areas of Vim demonstrate that thalamic activity leads and might drive EMG activity in dystonia, that stimulation in these areas can increase dystonia, and that lesions of these areas can decrease dystonia (Lenz et al., 1999). These findings make it unlikely that cortical changes are responsible for the alteration in thalamic activity but suggest that reorganization of thalamic structures contributes to dystonia. Studies of the thalamus in patients with nervous system injury and dystonia demonstrate large changes in the perceptual representation of the body in the thalamus, as revealed by somatotopic maps of projected fields (Lenz et al., 1994a, 1998a). Patients with amputations show increases in the thalamic area from which stimulation evokes sensations in the part of a limb that has become the stump (Lenz et al., 1998a). In patients with spinal transections and dystonia, the incidence of mismatches between neuronal RFs and threshold microstimulation-evoked PFs at the same sites is much higher than in patients with movement disorders (Lenz et al., 1994a). In patients with spinal transection these mismatches occur because sensations in the anesthetic part are evoked by stimulation at sites where cellular RFs represent parts of the body proximal to the anesthetic part. In these patients, stimulation in the thalamic area where the representation of the anesthetic part of the body is usually located evokes sensations in the anesthetic part (Lenz et al., 1994a). These results suggest that the perceptual representation of the body in the thalamus reorganizes less than the representation of inputs to the thalamus. Although the central image of the body is relatively constant in the face of altered input, our studies show that changes in the modality organization of this image can change dramatically (Lenz et al., 1998a). This plasticity of modality organization may contribute to the development of chronic pain in patients with nervous system injury or with dystonia.

Acknowledgements This research was supported by grants to FAL from the Eli Lilly Corporation and the National Institutes of Health (NS38493, NS40059). References Albe-Fessard, D.G. and Lombard, M.C. (1983) Use of an animal model to evaluate the origin of and protection against deafferentation pain. Adv. Pain Res. TheK, 5: 691-700. Andersen, P., Vestergaard, K., Ingeman-Nielsen, M. and Jensen, T.S. (1995) Incidence of post-stroke central pain. Pain, 61: 187-193. Apkarian, A.V. and Shi, T. (1994) Squirrel monkey lateral thalamus. I. somatic nociresponsive neurons and their relation to spinothalamic terminals. J. Neurosci., 14: 6779-6795. Apkarian, A.V., Shi, T., Stevens, R.T., Kniffki, K.-D. and Hodge, C.J. (1991) Properties of nociceptive neurons in the lateral thalamus of the squirrel monkey. Sot. Neurosci. Abstr, 17: 838-838. Beric, A., Dimitrijevic, M.R. and Lindblom, U. (1988) Central dysesthesia syndrome in spinal cord injury patients. Pain, 34: 109-116. Boivie, J., Leijon, G. and Johansson, I. (1989) Central poststroke pain-a study of the mechanisms through analyses of the sensory abnormalities. Pain, 37: 173-185. Bowsher, D. (1957) Termination of the central pain pathway in man: the conscious appreciation of pain. Bruin, 80: 606-620. Bushnell, M.C. and Duncan, G.H. (1987) Mechanical response properties of ventroposterior medial thalamic neurons in the alert monkey. Exp. Bruin Res., 67: 603-614. Bushnell, M.C., Duncan, G.H. and Tremblay, N. (1993) Thalamic VPM nucleus in the behaving monkey. I. Multimodal and discriminative properties of thermosensitive neurons. J. Neurophysiol., 69: 739-752. Byl, N.N., Merzenich, M.M. and Jenkins, W.M. (1996a) A primate genesis model of focal dystonia and repetitive strain injury: I. learning-induced dedifferentiation of the representation of the hand in the primary somatosensory cortex in adult monkeys. Neurology, 47: 508520. Byl, N.N., Hamati, D. and Wilson, F. (1996b) The sensory consequences of repetitive strain injury in musicians: focal dystonia of the hand. J. Back Musculoskeletal. Rehab., 7: 2739. Byl, N.N., Merzenich, M.M., Cheung, S., Bedenbaugh, P., Nagarajian, S.S. and Jenkins, W.M. (1997) A primate genesis model of focal dystonia and repetitive strain injury: II the effect of movement strategy on the de-differentiation of the hand representation in the primary somatosensory cortex of owl monkeys. Sot. Neurosci. Abstc, 22: 1055. Cardoso, F., Jankovic, J., Grossman, R.G. and Hamilton, W. (1995) Outcome after stereotactic thalamotomy for dystonia and hemiballismus. Neurosurgev, 36: 501-508. Casey, K.L. (1966) Unit analysis of nociceptive mechanisms in

271 the thalamus of the awake squirrel monkey. .I. Neurophysiol., 29: 727-750. Casey, K.L. and Morrow, T.J. (1983) Ventral posterior thalamic neurons differentially responsive to noxious stimulation of the awake monkey. Science, 22 1: 615-617. Cassinari, V. and Pagni, CA. (1969) Central Pain. A Neurosurgical Survey. Harvard University Press, Cambridge, MA. Chung, J.M., Lee, K.H., Surmeier, D.J., Sorkin, L.S., Kim, I. and Willis, W.D. (1986) Response characteristics of neurons in the ventral posterior lateral nucleus of the monkey thalamus. J. Neurophysiol., 56: 370-390. Craig, A.D., Bushnell, MC., Zhang, E.T. and Blomqvist, A. (1994) A specific tbalamic nucleus for pain and temperature sensation in macaques and humans. Nature, 372: 770-773. Davis, K.D., Kiss, Z.H.T., Luo, L., Tasker, R.R., Lozano, A.M. and Dostrovsky, 3.0. (1998) Phantom sensations generated by thalamic microstimulation. Nature, 391: 385-387. Davis, K.D., Lozano, A.M., Manduch, M., Tasker, R.R., Kiss, Z.H.T. and Dostovsky, J.O. (1999) Thalamic relay site for cold perception in humans. J. Neurophysiol., 81: 1970-1973. Dostrovsky, J.O., Wells, F.E.B. and Tasker, R.R. (1991) Pain evoked by stimulation in human thalamus. In: Y. Sjigenaga (Ed.), International Symposium on Processing Nociceptive Information. Elsevier, Amsterdam. Dostrovsky, J.O. (2000) Role of thalamus in pain. In J. Sandkiihler, B. Bromm and G.F. Gebhart (Eds.), Nervous System Plasticity and Chronic Pain, Progress in Bruin Research, Vol. 129. Elsevier, Amsterdam, pp. 245-257. Duncan, G.H., Bushnell, MC., Oliveras, J.L., Bastrash, N. and Tremblay, N. (1993) Thalamic VPM nucleus in the behaving monkey. III. effects of reversible inactivation by lidocaine on thermal and mechanical discrimination. J. Neurophysiol., 70: 2086-2096. Fahn, S. (1988) Concept and classification of dystonia. In: S. Falm (Ed.), Advances in Neurology. Raven Press, New York, pp. l-8. Flor, H. (2000) The functional organization of the brain in chronic pain. In J. Sandkilhler, B. Bromm and G.F. Gebhart (Eds.), Nervous System Plasticity and Chronic Pain, Progress in Bruin Research, Vol. 129. Elsevier, Amsterdam, pp. 313322. Garraghty, P.E. and Kaas, J.H. (1991) Functional reorganization in adult monkey thalamus after a peripheral nerve injury. NemoReport, 2: 747-750. Gautron, M. and Guilbaud, G. (1982) Somatic responses of ventrobasal thalamic neurones in polyarthritic rats. Brain Res., 237: 459-471. Hallett, M. (1995) Is dystonia a sensory disorder?. Ann. Neural., 38: 139-140. Halliday, A.M. and Logue, V. (1972) Painful sensations evoked by electrical stimulation in the thalamus. In: G.G. Somjen (Ed.), Nemophysiology Studied in Man. Excerpta Medica, Amsterdam, pp. 221-230. Hirai, T. and Jones, E.G. (1989) A new parcellation of the human thalamus on the basis of histochemical staining. Brain Res. Rev., 14: l-34. Hosobuchi, Y. (1986) Subcortical electrical stimulation for con-

trol of intractable pain in humans. J. Neurosurg., 64: 543553. Jenkins, W.M., Merzenich, M.M., Ochs, M.T., Allard, T. and Guic-Robles, E. (1990) Functional reproganization of primary somatosensory cortex in adult owl monkeys after behaviorally controlled tactile stimulation. J. Neurophysiol., 63: 82- 104. Jensen, T.S. and Rasmussen, P. (1994) Phantom pain and related phenomena after amputation. In: PD. Wall and R. Melzack (Eds.), Textbook ofPain. Churchill Livingstone, New York, pp. 65 l-665. Jones, E.G. and Friedman, D.P. (1982) Projection pattern of functional components of thalamic ventrobasal complex on monkey somatosensory cortex. J. Neurophysiol., 48: 521-544. Kaas, J.H. (1991) Plasticity of sensory and motor maps in adult mammals. Annu. Rev. Neurosci., 14: 137-167. Kaas, J.H., Merzenich, M.M. and Killackey, H.P. (1983) The reorganization of somatosensory cortex following peripheral nerve damage in adult and developing animals. Annu. Rev. Neurosci., 6: 325-356. Kelahan, A.M. and Doestch, G.S. (1984) Time dependent changes in the functional organization of somatosensory cerebral cortex following digit amputation in adult raccoons. Somatosens. Res., 2: 49-81. Kenshalo, D.R., Giesler, G.J., Leonard, R.B. and Willis, W.D. (1980) Responses of neurons in primate ventral posterior lateral nucleus to noxious stimuli. J. Neurophysiol., 43: 15941614. Lee, J.-I., Antezanna, D., Dougherty, P.M. and Lenz, EA. (1999) Responses of neurons in the region of the thalamic somatosensory nucleus to mechanical and thermal stimuli graded into the painful range. J. Camp. Neural., 410: 541-555. Lenz, EA. and Byl, N.N. (1999) Reorganization of the cutaneous core of the human thalamic principal thalamic somatosensory nucleus in patients with dystonia. .I. Neurophysiol., 82: 32043212. Lenz, EA. and Dougherty, PM. (1998) Cells in the human principal thalamic sensory nucleus (Ventralis Caudalis - Vc) respond to innocuous mechanical and cool stimuli. J. Neutrphysiol., 79: 2227-2230. Lenz, EA., Tasker, R.R., Kwan, H.C., Schnider, S., Kwong, R., Murayama, Y., Dostrovsky, J.O. and Murphy, J.T. (1988a) Single unit analysis of the human ventral thalamic nuclear group: correlation of thalamic ‘tremor cells’ with the 3-6 Hz component of parkinsonian tremor. J. Neurosci., 8: 754-764, Lenz, EA., Dostrovsky, J.O., Tasker, R.R., Yamashiro, K., Kwan, H.C. and Murphy, J.T. (1988b) Single-unit analysis of the human ventral thalamic nuclear group: somatosensory responses. J. Neurophysiol., 59: 299-316. Lenz, EA., Seike, M., Lin, Y.C., Baker, F.H., Richardson, R.T. and Gracely, R.H. (1993a) Thermal and pain sensations evoked by microstimulation in the area of the human ventrocaudal nucleus (Vc). J. Neurophysiol., 70: 200-212. Lenz, EA., Seike, M., Lin, Y.C., Baker, F.H., Rowland, L.H., Gracely, R.H. and Richardson, R.T. (1993b) Neurons in the area of human thalamic nucleus ventralis caudalis respond to painful heat stimuli. Brain Res., 623: 235-240. Lenz, EA., Kwan, H.C., Martin, R., Tasker, R., Richardson, R.T.

212

and Dostrovsky, J.O. (1994a) Characteristics of somatotopic organization and spontaneous neuronal activity in the region of the thalamic principal sensory nucleus in patients with spinal cord transection. J. Neurophysiol., 12: 1570-1587. Lenz, EA., Gracely, R.H., Hope, E.J., Baker, F.H., Rowland, L.H., Dougherty, P.M. and Richardson, R.T. (1994b) The sensation of angina can be evoked by stimulation of the human thalamus. Pain, 59: 119-125. Lenz, EA., Gracely, R.H., Rowland, L.H. and Dougherty, P.M. (1994~) A population of cells in the human principal sensory nucleus respond to painful mechanical stimuli. Neurosci. Len., 180: 46-50. Lenz, EA., Zirh, A.T., Garonzik, I.M. and Dougherty, P.M. (1998a) Neuronal activity in the region of the principal sensory nucleus of human thalamus (ventralis caudalis) in patients with pain following amputations. Neuroscience, 86: 1065-108 1. Lenz, EA., Gracely, R.H., Baker, F.H., Richardson, R.T. and Dougherty, P.M. (1998b) Reorganization of sensory modalities evoked by microstimulation in the region of the thalamic principal sensory nucleus in patients with pain due to nervous system injury. J. Comp. Neurol., 399: 125-138. Lenz, EA., Seike, MS., Jaeger, C.J., Reich, S.G., Lin, Y.C., DeLong, M.R. and Vitek, J.L. (1999) Thalamic single neuron activity in patients with dystonia: dystonia-related activity and somatic sensory reorganization. J. Neurophysiol., 82: 23122392. Levitt, M. (1985) Dysesthesias and self-mutilation in humans and subhumans: a review of clinical and experimental studies. Bruin Res. Rev., 10: 247-290. Lombard, M.C., Nashold, B.S. and Pelissier, T. (1979) Thalamic recordings in rats with hyperalgesia. Adv. Pain Res. Ther, 3: 767-772. Mehler, W.R. (1962) The anatomy of the so-called ‘pain tract’ in man: an analysis of the course and distribution of the ascending fibers of the fasciculus anterolateralis. In: J.D. French and R.W. Porter (Eds.), Basic Research in Paraplegia. Thomas, Springfield, pp. 26-55. Mehler, W.R. (1966) The posterior thalamic region in man. Conjin. Neural., 27: 18-29. Mehler, W.R. (1969) Some neurological species differences: a posteriori. Ann. N.Y. Acad. Sci., 167: 424-468. Mehler, W.R., Feferman, M.E. and Nauta, W.H.J. (1960) Ascending axon degeneration following anterolateral cordotomy. An experimental study in the monkey. Bruin, 83: 718-750. Merzenich, M.M., Kaas, J.H., Wall, J.T., Nelson, R.J., Sur, M. and Fellernan, D. (1983a) Topographic reorganization of somatosensory cortical areas 3b and 1 in adult monkeys following restricted deafferentation. J. Neurosci., 10: 639-665. Merzenich, M.M., Kaas, J.H., Wall, J., Nelson, R.J., Sur, M. and Felleman, D. (1983b) Topographic reorganization of somatosensory cortical areas 3B and 1 in adult monkeys following restricted deafferentation. Neuroscience, 8: 33-55. Merzenich, M.M., Nelson, R.J., Kaas, J.H., Stryker, M.P., Jenkins, W.M., Zook, J.M., Cynader, MS. and Schoppmann, A. (1987) Variability in hand surface representations in areas 3b and 1 in adult owl and squirrel monkeys. J. Comp. Neuml., 258: 281-296.

Nudo, R.J., Jenkins, W.M., Merzenich, M.M., Prejean, T. and Grenda, R. (1992) Neurophysiological correlates of hand preference in primary motor cortex of adult squirrel monkeys. J: Neurosci., 12: 2918-2947. Nudo, R.J.. Milliken, G.W., Jenkins, W.M. and Merzenich, M.M. (1996) Use dependent alterations of movement representations in primary motor cortex of adult squirrel monkeys. J. Neurosci., 16: 785-807. Pollin, B. and Albe-Fessard, D.G. (1979) Organization of somatic thalamus in monkeys with and without section of dorsal spinal tracts. Bruin Res., 173: 431-449. Pons, T.P., Garraghty, P.E., Ommaya, A.K., Kaas, J.H., Taub, E. and Mishkin, M. (1991) Massive cortical reorganization after sensory deafferentation in adult macaques. Science, 252: 1857-1860. Ralston, H.J., Ohara, P.T., Meng, X.W., Wells, J. and Ralston, D.D. (1996) Transneuronal changes in the inhibitory circuitry of the macaque somatosensory thalamus following lesions of the dorsal column nuclei. J. Comp. Neurol., 371: 325-335. Rasmusson, D. (1996b) Changes in the organization of ventroposterior lateral thalamic nucleus after digit removal in the adult raccoon. J Comp. Neurol., 364: 92-103. Rasmusson, D.D. (1982) Reorganization of raccoon somatosensory cortex following removal of the fifth digit. J. Comp. Neural., 205: 3 13-326. Rasmusson, D.D. (1996a) Changes in the response properties of neurons in the ventroposterior lateral thalamic nucleus of the raccoon after peripheral deafferentation. J. Neurophysiol., 75: 2441-2450. Rausell, E., Cusick, C.G., Taub, E. and Jones, E.G. (1992) Chronic deafferentation in monkeys differentially affects nociceptive and non-nociceptive pathway distinguished by specific calcium-binding proteins and down-regulates gammaaminobutyric acid type A receptors at thalamic levels. Proc. Natl. Acad. Sci. USA, 89: 2571-2575. Recanzone, G.H., Merzenich, M.M., Jenkins, W.M., Grajski, K.A. and Dinse, H.R. (1992) Topographic reorganization of the hand representation in cortical area 3b of owl monkeys trained in a frequency-discrimination task. J. Neurophysiol., 67: 1031-1056. Sweet, W.H. (1981) Animal models of chronic pain: their possible validation from human experience with posterior rhizotomy and congenital analgesia. Pain, 10: 275-295. Sweet, W.H., Poletti, C.E. and Gybels, G.M. (1994) Operations in the brainstem and spinal canal with an appendix on the relationship of open and percutaneous cordotomy. In: P.D. Wall and R. Melzack (Eds.), Textbook of Pain. Churchill and Livingstone, New York, pp. 1113-l 136. Tasker, R.R. (1982) Identification of pain processing systems by electrical stimulation of the brain. Hum. Neurobiol., 1: 261212. Tasker, R.R. (1988) Percutaneous Cordotomy: The Lateral High Cervical Technique. In: H.H. Schmidek and W.H. Sweet (Eds.), Operative Neurosurgical Techniques Indications, Methods, and Results. W.B. Saunders, Philadelphia, pp. 11911205.

273

Tasker, R.R., Doorly, T. and Yamashiro, K. (1988) Thalamotomy in generalized dystonia. Adv. Neural., 50: 615-63 1. Tasker, R.R., Organ, L.W. and Hawrylyshyn, P. (1980) Deafferentation and causalgia. In: J.J. Bonica (Ed.), Pain. Raven Press, New York, pp. 305-329. Tolosa, ES. and Marti, M.J. (1997) Adult onset idiopathic torsion dystonias. In: R.L. Watts and W.C. Keller (Eds.), Movement Disorders. McGraw Hill, New York, pp. 427-441. Vierck, C.J. and Light, A.R. (2000) Allodynia and hyperalgesia within dermatomes caudal to a spinal cord injury in primates and rodents. In J. Sandktihler, B. Bromm and C.F. Gebhart (Eds.), Nervous System Plasticity and Chronic Pain, Progress in Bruin Research, Vol. 129. Elsevier, Amsterdam, pp. 41 l428. Vitek, J.L. and Lenz, EA. (1998) Contemporary pallidotomy for treatment of dystonia and other movement disorders. In: J. Jankovic, R.G. Grossman and G. Krauss (Eds.), Pullidotomy for treatment of Parkinson’s disease and other movement disorders. Lippincott-Raven, New York. Wall, J.T. and Cusick, C.G. (1984) Cutaneous responsiveness in primary somatosensory (S-l) hindpaw cortex before and after partial hindpaw deafferentation in adult rats. J. Neurosci., 4:

1499-1515. Wang, X., Merzenich, M.M., Sameshima, K. and Jenkins, W. (1995) Representation remodeling of hand hand surface map in adult cortex is input timing dependent. Nature, 378: 71-75. Yezierski, R.P. (2000) Pain following spinal cord injury: pathophysiology and central mechanisms. In: J. Sandkiihler, B. Bromm and G.F. Gebhart (Eds.), Nervous System Plasticity and Chronic Pain, Progress in Brain Research, Vol. 129. Elsevier, Amsterdam, pp. 429-449. Young, R.F., Kroening, R., Fulton, W.. Feldman, R.A. and Chambi, I. (1985) Electrical stimulation of the brain in treatment of chronic pain. J. Neurosurg., 62: 389-396. Zirh, A.T., Reich, S.G., Dougherty, P.M. and Lenz, EA. (1999) Stereotactic thalamotomy in the treatment of essential tremor of the upper extremity: re-assessment including a blinded measure of outcome. J. Neural. Neurosurg. Psychiatry, 66: 772-775. Zirh, A.T., Lenz, EA., Garonzik, I.M., Richardson, R.T., Ringkamp, M., Rowland, L.H. and Dougherty, P.M. (1996) Characteristics of neuronal activity in the region of the thalamic nucleus Vc in patients with amputations. Sot. Neurosci. Abstz, 22: 1054.

J. Sandkiihler, B. Bromm and G.F. Gebhart (Eds.) Progress in Brain Research, Vol. 129 0 2ooO Elsevier Science B.V. All rights reserved

CHAPTER

19

Concepts of pain mechanisms: the contribution of functional imaging of the human brain Kenneth L. Casey* University of Michigan, Neurology Service, VA. Medical Centel;

Ann Arbor;

MI 48105,

KiA

Introduction

Importance

The term ‘brain imaging’ has been used to refer to different methods, including the recording of patterns of changes of voltage, current, or magnetic fields recorded from the surface of the head. The field of functional magnetic resonance imaging (fMRI) has developed new methods for the rapid detection of activity changes that accompany the shift from deoxyhemoglobin to oxyhemoglobin during increases in regional cerebral blood flow (rCBF). In this essay, we shall emphasize the method of detecting rCBF changes by positron emission tomography (PET). Although the other imaging methods provide more detailed temporal resolution (evoked potentials, magnetoencephalography, see Bromm et al., 2000, this volume) and spatial localization (fMRI), PET currently has the advantage of combining statistical quantification with the simultaneous detection of patterns of regional activity throughout the forebrain and brainstem. In addition, PET provides the opportunity for determining quantitatively the location and degree of binding of membrane receptors by various ligands and putative neurotransmitters and modulators, opening up the emerging field of functional imaging neuropharmacology.

Neither normal nor pathological pain can be understood without knowledge about forebrain mechanisms. In humans, the forebrain anatomically dominates and physiologically controls nociceptive processing. Because of the large volume of the human forebrain in relation to that of the spinal cord, descending modulatory influences assume greater importance in humans than in other species, such as the laboratory rat, in which the forebrain is less anatomically dominant. The rodent forebrain comprises 44% and the spinal cord 35% of CNS volume. In contrast, the human forebrain occupies 85% of the CNS and the spinal cord 2% only (Swanson, 1995). After complete rhizotomy, at the spinal cord level, only 13% of the synapses in laminae 1 through 3 of the primate dorsal horn degenerate (Ralston and Ralston, 1979). Of the remaining 77%, many innervate the superficial dorsal horn via the corticospinal tract (Ralston and Ralston, 1985), which contains from 500 x IO3 to 1 x lo6 fibers (Blinkov and Glezer, 1968; Towe, 1995). The human spinothalamic tract, in contrast, contains only an estimated 2-5 x IO3 fibers (Blinkov and Glezer, 1968). Corticothalamic influences are also likely to dominate in the human. In the cat, for example, approximately 50% of the estimated 5-9000 synapses on thalamocortical somatosensory projection neurons are presumed to be of cortical origin while only 15% are formed by ascending afferent fibers (Liu et al., 1995). These observations highlight the fact that normal or patho-

* Corresponding author: K.L. Casey,University of Michigan, Chief, Neurology Service, V.A. Medical Center, Ann Arbor, MI 48105, USA. Tel.: +l-734-761-7562; Fax: +l-734-769-703s; E-mail: [email protected]

of forebrain

278

logical changes in forebrain activity may have profound effects on pain and nociceptive processing at all levels of the CNS. The physiological basis of PET and fMR1 Synaptic activity generates increases in CBF. This physiological fact is the basis for both positron emission tomography (PET) and functional magnetic resonance imaging (fMR1). Roy and Sherrington reported the first evidence that brain activity increases global CBF over 100 years ago (Roy and Sherrington, 1890). Subsequent radioactive tracer techniques revealed increases in rCBF during sensory stimulation or the performance of motor tasks (Roland, 1993). Recent studies using the technique of optical imaging demonstrate that cortical blood flow responses occur within 3 s of sensory stimulation and are initially restricted to the 3-500 pm dimensions of cortical columns before spreading to involve the surrounding 3-5 mm of cortical tissue (Malonek and Grinvald, 1996; Wang et al., 1996; MacVicar, 1997; Vanzetta and Grinvald, 1999). The biochemical coupling of rCBF and synaptic activity is unknown and is still an area of active investigation. Blocking the production of nitric oxide has no effect on synaptically induced rCBF responses in the rat somatosensory system (Adachi et al., 1994). There is evidence that adenosine may be a critical link in this process, but it is likely that the action of several mediators may be important (Dimagl et al., 1994). Vanzetta and Grinvald (1999) have recently demonstrated the tight coupling of rCBF with rapid increases in cortical oxidative metabolism following sensory stimulation (Vanzetta and Grinvald, 1999). Studies of regional cerebral glucose utilization (rCGU) show that rCBF and rCGU are normally tightly coupled and that the coupling occurs within the synaptic neuropil. The degree of coupling may vary among regions and under special experimental conditions (Mraovitch et al., 1992), but is reliably present in the normal mammal. Reductions in rCBF are observed in many PET studies (Seitz and Roland, 1992). The physiological significance of reduced rCBF is uncertain; it does not necessarily indicate the presence of inhibitory synaptic activity because both inhibitory and excitatory synaptic activity can contribute to increases

in synaptic metabolism. Indeed, the observations of Auker et al. (1983) suggest that this is likely to be the case (Auker et al., 1983). It is possible that some of the observed reductions in rCBF reflect autoregulatory mechanisms for global CBF and that these do not affect neuronal function; others may signal the removal of synaptic excitation (disfacilitation) by an inhibitory process located outside the area of rCBF decrease. In any event, it is not now possible to establish the valence of synaptic activity by rCBF estimation methods. Technical and analytical issues in PET Image voxel intensities are normalized to global cerebral activity with the use of a linear proportional model to remove baseline differences in global cerebral blood flow between scans and subjects (Fox and Raichle, 1984). In our facility, CBF images are aligned onto the coordinates of a standard stereotactic atlas (Talairach and Toumoux, 1988), using anatomical landmarks identified within the PET images of each individual so that the CBF differences are compared within the same brain regions (Minoshima et al., 1992, 1993, 1994). To determine whether a task or a stimulus has produced an increase in regional CBF, the rCBF computed during a control condition is subtracted from that computed during the test condition. A voxel-by-voxel statistical subtraction analysis (Z-score) with adjustment for multiple comparisons is performed by estimating the smoothness of subtraction images (Friston et al., 1991) following threedimensional Gaussian filtering to enhance signal-tonoise ratio and compensating for residual anatomical variance. Typically, only those voxels with normalized CBF values larger than 60% of the global value are analyzed because these represent the gray matter of the brain. Voxels showing a significantly increased CBF compared to the average noise variance computed across all voxels (pooled variance) are identified (Worsley et al., 1992). The critical level of significance is determined by adjusting p = 0.05 using this information (Adler and Hasofer, 1976). The results of interest using this method are revealed primarily through the data analysis. In addition, volumes of interest (VOI) may be established within brain structures selected because

279 of a priori hypotheses and results of previously published PET studies. The size and shape of each VOI may be standardized across studies or determined separately according to functional criteria. We presently use a method similar to that described by Burton (Burton et al., 1993) in which voxels showing significant peak increases in CBF between comparison conditions are identified within the brain structure of interest and progressively expanded in three dimensions to include contiguous voxels that meet the statistical criterion established by the voxel-byvoxel Z-score analysis. To determine the statistical significance of rCBF increases, a paired t statistic is computed for each VOI from the average percentage increase in CBF across all subjects. Levels of significance are established based on the Bonferroni correction for multiple comparisons among VOI. This form of data analysis is guided by hypotheses. It is also possible to perform correlations between the intensity of the rCBF responses throughout the brain and some behavioral parameter of interest. Numerous studies have used PET measurements of rCBF changes to correlate regional activation patterns with behavioral performance in humans. This correlative method of analysis may be employed on a voxel-by-voxel (data driven) or VOI (directed hypothesis driven) basis. Recent PET studies have begun to establish quantitative stimulus-response relationships between increases in rCBF and measurements of behavioral performance. For example, Dettmers et al. (1995) were able to demonstrate a logarithmic relationship between the force of finger flexion and the increases in rCBF in the primary motor cortex, posterior cingulate motor area, and the ventral posterior supplementary motor area (Dettmers et al., 1995). These results suggest that similar stimulus-response relationships can be determined for other behavioral performance measures, including measurements of various aspects of pain. Indeed, this method has been used recently in studies of pain to reveal those structures with rCBF responses that correlate parametrically with the perceived intensity or unpleasantness of noxious stimulation (Derbyshire et al., 1997; Rainville et al., 1997; Coghill et al., 1999).

What we have learned about pain from PET and fMR1 Multiple specijc brain structures are consistently active during pain perception We have known for many years that multiple brain structures and pathways participate in the processing of nociceptive information because of the correlation of clinical findings with pathological anatomy in the human. Many subsequent behavioral, physiological, and anatomical studies in experimental animals have confirmed that nociceptive activity is widely distributed throughout the central nervous system. Nonetheless, the specific pattern of cortical activity that was reported in the first PET study of human pain perception could not have been predicted with confidence (Talbot et al., 1991). The report by Talbot and colleagues provided the first direct evidence that the perception of pain in humans correlated specifically with synaptic activity in multiple cortical areas: SI, SII, and the anterior cingulate cortex. Previous information, based largely on lesions or electrical stimulation, was incomplete because of uncertainty about their effects on adjacent or remote brain areas through degeneration or stimulation of orthodromic or antidromic pathways. The PET study of Talbot and colleagues, however, provided the first temporally immediate and spatially specific correlation of brain activity with the perception of pain. Many subsequent PET and fMRI studies have confirmed and extended these findings, showing that activity within several cortical and subcortical structures is consistently present during the perception of pain (for review, see: Chen, 1993; Casey et al., 1994a; Davis et al., 1995, 1997; Casey and Minoshima, 1997; Apkarian et al., 1999; Casey, 1999; Derbyshire, 1999; Ingvar, 1999). Across all studies, pain-related activity is found most frequently within the medial midbrain, thalamus, lentiform nucleus, cerebellum, and the insular, prefrontal, parietal (including SI and SII), and anterior cingulate cortices (Derbyshire, 1999). These results highlight the participation of combined sensory, motor, association, and limbic systems in the parallel mediation of the multiple components of the pain experience and response. Although variability among brain imaging studies remains, there is sufficient predictability to

280

provide us with a new opportunity to investigate the specific functional role of each of these activated structures. The results of functional brain imaging have forced us to think in terms of the activation of a network of interactive subsystems. Normal group differences in pain perception are reflected by differences in brain activation Because not all pains are alike, one of the first questions is whether different types of noxious stimuli activate detectably different regions of the forebrain. To address this question, we investigated whether there are differences in the spatial pattern and intensity of synaptically induced rCBF during the perception of different forms and intensities of innocuous and noxious thermal stimuli (Casey et al., 1996). In a study of warm discrimination, we applied two intensities of innocuous heat (36°C and 43°C) repetitively to the left forearm of normal subjects (N = 9). All subjects discriminated the 43°C from the 36°C stimulus; neither stimulus was rated painful. We found significant increases in rCBF to the 43°C stimuli in the contralateral ventral posterior thalamus, lenticular nucleus, medial prefrontal cortex (Brodmann’s areas 10 and 32), and the cerebellar vermis. In a separate nearly identical study. nine other subjects discriminated between noxious and innocuous heat stimuli (40°C and 50°C). We found significant rCBF increases to 50°C stimuli in the contralateral thalamus, lenticular nucleus, anterior cingulate cortex, premotor cortex, insula, and SII cortex. Responses just below the threshold for statistical significance were seen in the contralateral sensorimotor cortex (Ml/X). The ipsilateral premotor cortex and thalamus, and the medial dorsal midbrain and cerebellar vermis also showed significant rCBF increases. We compared these results to those obtained during continuous immersion of the left hand in innocuously cold (20°C) and painfully cold water (1°C). We found highly significant increases in rCBF in the contralateral lenticular nucleus, sensorimotor cortex (Ml /SI), premotor cortex, anterior cingulate cortex, and anterior insula. Contralateral thalamic responses were just below the threshold for statistical significance. Ipsilateral responses were seen in the thalamus, lateral prefrontal cortex, anterior cingulate cortex, insula, and cerebellar vermis. Each of the five

regions responsive in both the heat pain and cold pain conditions (cerebellar vermis, ipsilateral thalamus, contralateral premotor, anterior cingulate, and the region of the anterior insular cortex and lenticular nucleus) showed a higher increase in rCBF in the cold pain than in the heat pain study (p c 0.022). Thus, two forms of noxious stimulation that are different in temporal pattern, afferent fiber activation, and perceived spatio-temporal and qualitative characteristics, produce similar, but not identical, patterns of brain increases in rCBF. These pain-related response patterns are each quite different from the brain responses observed during the discrimination between two intensities of innocuous heat stimuli. The overlap in the spatial distribution of rCBF increases during noxious cutaneous heat and noxious deep cold stimulation suggests that there is a reproducible pattern of rCBF responses that is common to the perception of pain produced by different stimuli. However, differences in the intensity and spatial pattern of these pain-related rCBF increases probably reflect physiological differences in neuronal nociceptive processing related to the perception of these two forms of pain. To investigate further physiological differences related to differences in normal pain perception, we applied both cutaneous laser and intramuscular electrical stimulation to normal males and examined their forebrain activation patterns (Svensson et al., 1997). Subjects perceived both types of stimuli as being near pain threshold and of nearly equal intensity. Therefore, differences in brain activation cannot be attributed to differences in perceived intensity. Because both forms of stimulation are near pain threshold, the activated structures are probably among the most sensitive to painful stimulation. We found significant increases in rCBF to both noxious cutaneous and intramuscular stimulation in the contralateral secondary somatosensory cortex (SII) and inferior parietal lobule (Brodmann area, B40). We found comparable levels of rCBF increase also in the contralateral anterior insular cortex, thalamus, and ipsilateral cerebellum. Direct statistical comparisons between cutaneous and intramuscular stimulation showed no reliable differences between these two forms of noxious stimulation, indicating a substantial overlap in brain activation pattern. The similar cerebral activation patterns suggest that the

281

Fig. 1. Color-coded statistical maps of the sites of increases in regional cerebral blood flow (rCBF) during discrimination between the intensities of innocuous (40°C) and noxious (50°C) heat contact stimuli in male and female subjects. Images are presented in alternating sets of male (M) and female (F) maps for ease of direct visual comparison. All stimuli were applied to the subjects’ left forearm. The right hemisphere of the brain magnetic resonance image (MRI) template is on the left of the figure. This template is the brain of a single normal subject transformed onto the stereotactic coordinates of the human brain atlas used in this study (Talairach and Tournoux, 1988). Below each set of columns of images is the stereotactic location with respect to the commissural line (+, superior; -, inferior). The numbers by the color bar show the Z-scores corresponding to the standard deviation of each region from the mean CBF increase. The colored regions in this figure include all structures showing CBF increases above the mean at a p = 0.05 level of significance, uncorrected for multiple comparisons. (Reprinted from Paulson et al. (1998) with permission.)

perceived differences between acute skin and muscle pain are mediated by differences in the intensity and temporo-spatial pattern of neuronal activity within similar sets of forebrain structures.

We obtained similar results when we examined the question of gender influences on pain perception and brain activation (Paulson et al., 1998). Both male and female subjects rated 40°C contact heat

282

stimuli as warm and 50°C stimuli as painful. However, females rated the 50°C stimuli as significantly more intense than did the males (p = 0.0052). Both genders showed a bilateral activation of premotor cortex in addition to the activation of a number of contralateral structures, including the posterior insula, anterior cingulate cortex and the cerebellar vermis during heat pain. There was a high degree of overlap of responding areas, again showing that the perception of pain is mediated by a common set of forebrain structures (Fig. 1). Females, however, showed significantly greater activation of the contralateral prefrontal cortex when compared to the males by direct image subtraction. In addition, a VOI comparison revealed greater activation of the contralateral insula and thalamus in females. These pain-related differences in brain activation may reflect differences in gender, perceived pain intensity, or both factors. The functional speciJicity of pain-activated brain regions can be identified in imaging experiments We have just begun to identify the sources of variation in the normal pain activation of the human forebrain. However, as suggested by the results reviewed here and by others (Derbyshire, 1999), there is a sufficiently consistent background of information to support studies directed at revealing the functional significance of different regions of activation. Correlation analysis is an effective method of approaching this problem. This approach requires identifying and controlling a variable that is an important component of the pain experience. One of the first studies to apply this method is that of Rainville and colleagues (Rainville et al., 1997). These investigators used

hypnotic suggestion to uncouple the perception of heat pain unpleasantness from heat pain intensity in normal subjects. They were able to show a positive correlation of pain unpleasantness with the intensity of rCBF response in a far-anterior (dorsal perigenual) region of the anterior cingulate cortex; this correlation was not found in a specific examination of the SI cortex. It is possible, however, that the activation of other, as yet unstudied, brain regions also correlates with unpleasantness. In related investigations, Derbyshire and, more recently, Coghill and colleagues, have used correlation analyses to investigate the inter-regional distribution of information about heat pain intensity (Derbyshire et al., 1997; Coghill et al., 1999). Both groups found that information about pain intensity was widely distributed among many, but not all, pain-activated regions, including the cerebellum in the Coghill study. This finding indicates that information about intensity is distributed among structures that are highly heterogeneous in function. Other methods have been used also to reveal functional specificity among pain-activated regions. For example, Ploghaus and colleagues used fMRI responses in experiments designed to separate the perception of pain from the anticipation of pain (Ploghaus et al., 1999). They showed that the activation of some regions, previously identified in PET studies of pain, is better correlated with the anticipation of pain than with pain perception. Thus, activity in the far frontal cingulate cortex, anterior insula, and anterior cerebellar vermis was related to the anticipation of pain while the perception of pain was related to activity in adjacent regions (mid-anterior cingulate, mid-insula, paravermian cerebellum). We are at an early stage in the investigation of functional specificity, but there is little doubt that there will be

Fig. 2. PET study of a patient with ongoing central post-stroke pain of the left hemibody following a lacunar infarction at the lateral edge of the right ventral posterior thalamus 3 years ago. (See text for details.) (A) Structural MRI scan (left) shows the right thalamic infarction (arrow). There were no other lesions in this patient’s brain. PET image (right), taken at rest (no stimulation) shows right thalamic rCBF reduced compared to the left. (B) Quantitative analysis of the average (std. dev.; three samples at each site) resting rCBF in five spherical (4 mm in diameter) VOI chosen based on the results of previous studies (Casey et al., 1994b, 1996; Paulson et al., 1998) and placed at peaks of activation in this patient. Black bars show the normalized resting rCBF in structures on the non-stroke (left) side of the patient’s brain, contralateral to his asymptomatic side. Gray striped bars show these values in the same structures on the side of the patient’s stroke, contralateral to his painful dysesthesia. The p values are derived from paired c-test comparisons of each side in each structure. Note that both cortical and subcortical structures on the side of the lacunar stroke show relatively reduced resting rCBF except for the XI cortex, which shows a trend in the opposite direction; this selective effect argues against a generalized hemispheric effect of the stroke.

283

RESTING: NO STIMULATION right

T

m left

T

s2 ctx

T

thalamus

284

significant and accelerating progress in this direction in the immediate future. Neuropathic pain is mediated by unique patterns of ,forebrain activation

Neuropathic pain is caused by damage to the peripheral or central nervous system (Merskey and Bogduk, 1994). Imaging studies of the pathophysiology of neuropathic pain are at an early stage. One major problem is the inevitable heterogeneity of the clinical presentation of neuropathic pain. To circumvent this problem, some investigators have injected capsaicin intracutaneously in normal subjects to produce tactile allodynia, a common symptom among patients with neuropathic pain. Separate PET and fMRI studies of normal subjects revealed unique activations of the prefrontal cortex when the experience of tactile allodynia was compared with normal tactile sensation or with the ongoing pain caused by capsaicin alone (Iadarola et al., 1998; Baron et al., 1999). Studies of patients with neuropathic pain, however, have produced results that are more complex. Cesaro and colleagues used single photon emission computerized tomography (SPECT) to investigate the cerebral responses to pain in four patients with central poststroke pain (CPSP) (Cesaro et al., 1991). They found thalamic hyperactivity exclusively in two patients with hyperpathia when the abnormal, but not normal, side was stimulated (see also: Dostrovsky, 2000, this volume; Lenz et al., 2000, this volume, for thalamic mechanisms of pain). Peyron et al. (1998) also reported increased contralateral thalamic responsiveness during cold allodynia in patients with lateral medullary infarction (Wallenberg syndrome). These investigators also noted allodynia-related increased responses in the somatosensory, inferior parietal, insular, and medial prefrontal cortices but deactivation in the anterior cingulate cortex (Brodmann areas 24 and 32). In contrast, three other imaging studies revealed reduced thalamic activity as patients experienced ongoing peripheral (Hsieh et al., 1995; Iadarola et al., 1995) or central (Pagni and Canavero, 1995) neuropathic pain without stimulation. In ongoing studies of patients with central neuropathic pain (Casey et al., 1999) we have also observed thalamic hypoactivity at rest. However, there appear to be bilaterally abnormal cortical and subcor-

tical responses. One patient, a 67-year-old hypertensive right-handed man, experienced the sudden onset of constant, persistent, painful dysesthesiae of the left hemibody and face 3 years ago; these symptoms have persisted. The MRI showed a 2 x 4 x 7.5 mm lacunar infarction in the lateral right ventral posterior lateral (VPL) thalamus (Fig. 2A). Sensory examination revealed only deep pressure allodynia on the left (L/R: 2.1/4.3 Kg; p < 0.001) and elevated but symmetrical cutaneous heat pain thresholds (L/R: 49.4/5O.l”C). At rest, estimated regional cerebral blood flow (rCBF) was markedly and significantly reduced, compared to the left, in the right putamen and the right parietal, and insular cortices. The right thalamus also showed a strong trend of hypoactivity at rest (Fig. 2B). The right SII cortex, however, was unique among the structures sampled because it showed a strong trend toward hyperactivity at rest, while the patient was experiencing spontaneous pain. The patient rated noxious heat stimulation (55°C) as equally painful on either side. During noxious heat stimulation of the patient’s right (normal) side, there was, relative to rest, a slight reduction in rCBF in the left (contralateral) thalamus, putamen, and in the insular and parietal cortices. The left SII cortex, in contrast, showed a strong trend of excessive rCBF increase (p = 0.062 compared to the right). When noxious heat was applied to the patient’s left (abnormal) side, the right SII, parietal, and insular cortices gave weak rCBF responses compared to the left. However, there was a strong and significant (p < 0.02) rCBF increase in the right (contralateral) putamen and thalamus compared to the left. These results suggest that the pathological hypoactivity in the resting hemithalamus masks an underlying hyper-responsiveness to noxious stimulation. This may be due to a loss of resting inhibitory activity within the thalamus, leading to bilateral and widespread cortical abnormalities. Sensory stimulation evokes considerable inhibitory activity in the mammalian thalamus (Salt, 1989; Roberts et al., 1992). Inhibitory interneurons and synaptic profiles comprise a significant population of primate thalamic neurons (Ohara et al., 1989; Williamson et al., 1994) and these are highly susceptible to pathological changes following spinal cord injury (Ralston et al., 1996). However, the analysis of this case, and of the other cases cited here, demonstrates the com-

285

plexity of physiological changes that may follow the anatomical reorganization that is triggered by injury to the central or peripheral nervous system (Pons et al., 1991; Kaas et al., 1997; Jones and Pons, 1998).

(DE-FG02-87-ER60561). The authors gratefully acknowledge the support and encouragement of David E. Kuhl, M.D. and the expert technical assistance of Jill Rothley, Todd Hauser, Paul Kison, Edward McKenna. and Andrew Weeden.

Summary Functional imaging of the conscious human brain has a solid physiological basis in synaptically induced rCBF responses. We still do not know how these responses are generated, but recent studies have shown that the rCBF response is parametrically positively correlated with functional measures of neuronal activity. Technical advances in both NRI and PET imaging have improved the spatial and temporal resolution of imaging methods. Further advances may be expected in the near future. Consequently, we now have an important tool to apply to the study of normal and, most importantly, pathological pain. There is a tendency to expect too much of this exciting technique, but the problems we wish to address are complex and will require considerable time, effort, and patience. We now know that the CNS adapts to both peripheral and central nervous system injury, sometimes in beneficial ways, but sometimes with reorganization that is maladaptive. An understanding of the pathophysiology of neuropathic pain is further complicated by the new knowledge, emphasized by functional brain imaging, that pain and pain modulation is mediated, not by a simple pathway with one or a few central targets, but by a network of multiple interacting modules of neuronal activity. Simplified phrenological thinking, with complete psychological functions separate and localized, is appealing, but wildly misleading. It is far more realistic and productive to apply qualitative and quantitative spatial and temporal analyses to the distributed activity of the conscious, communicating human brain. This will not be quick and easy, but there is every reason for optimism in our search for a thorough and useful understanding of both normal and pathological pain. Acknowledgements Supported by grants from the U.S. Department of Veteran’s Affairs (Merit Review Grant), the Department of Health and Human Services (NIH: PO1 HD33986) and the Department of Energy

References Adachi, K., Takahashi, S., Melzer, P, Campos, K.L., Nelson, T., Kennedy, C. and Sokoloff, L. (1994) Increases in local cerebral blood flow associated with somatosensory activation are not mediated by NO. Am. .I. Physiol. Heart Circ. Physiol., 261: H2155-H2162. Adler, R.J. and Hasofer, A.M. (1976) Level crossings for random fields. Ann. Probabil., 4: 1-12. Apkarian, A.V., Darbar, A., Krauss, B.R., Gelnar, P.A. and Szeverenyi, N.M. (1999) Differentiating cortical areas related to pain perception from stimulus identification: temporal analysis of fMR1 activity. .I. Neurophysiol., 81: 2956-2963. Auker, C.R., Meszler, R.M. and Carpenter, D.O. (1983) Apparent discrepancy between single-unit activity and [ 14C]deoxyghcase labeling in optic tectum of the rattlesnake. J. Neurophysiol., 49: 1504-1516. Baron, R., Baron, Y.. Disbrow, E. and Roberts, T.P.L. (1999) Brain processing of capsaicin-induced secondary hyperalgesia - a functional MRI study. Neurology, 53: 548-557. Blinkov, S.M. and Glezer, 1.1. (1968) The Human Brain in Figures and Tables. Plenum Press, New York. Bromm, B., Scharein, E. and Vahle-Hinz, C. (2000) Cortex areas involved in the processing of normal and altered pain, In: J. Sandktihler, B. Bromm and G.F. Gebhart (Eds.), Nervous syvsfern Plasticity and Chronic Pain. Progress in Brain Research, Vol. 129. Elsevier, Amsterdam, pp. 289-302. Burton, H., Videen, T.O. and Raichle, M.E. (1993) Tactiievibration-activated foci in insular and parietal-opercular cortex studied with positron emission tomography: mapping the second somatqsensory area in humans. Somatosens. Mot. Rex. 10: 297-308. Casey, K.L. (1999) Forebrain mechanisms of nociception and pain: analysis through imaging. Proc. N&l. Acad. Sci. USA, 96: 7668-7674. Casey, K.L. and Minoshima, S. (1997) Can pain be imaged? In: T.S. Jensen, J.A. Turner and Z. Wiesenfeld-Hallin (Eds.). Proceedings qf the 8th World Congress on P&I. IASP Press, Seattle, WA, pp. 855-866. Casey, K.L., Minoshima, S., Koeppe, R.A.. Morrow, T.J. and Frey, K.A. (1994a) Imaging the brain in pain: potentials, limitations. and implications. In: B. Bromm (Ed.), From Nociception to Pain. Raven Press, New York. Casey, K.L., Minoshima, S., Berger, K.L., Koeppe, R.A., Morrow, T.J. and Frey, K..4. (1994b) Positron emission tomographic analysis of cerebral structures activated specifically by repetitive noxious heat stimuli. J. Neurophysiol., 71: 802-807. Casey, K.L., Minoshima, S., Morrow, T.J. and Koeppe, R.A. (1996) Comparison of human cerebral activation patterns dur-

286 ing cutaneous warmth, heat pain, and deep cold pain. J. Neurophysiol., 76: 57 l-58 1. Casey, K.L., Cross, D.J., Morrow, T.J. and Minoshima, S. (1999) Thalamocortical disinhibition in a case of central pain, IASP Abstracts (9th World Congress). Cesaro, P., Mann, M.W., Moretti, J.L., Defer, G., Roualdes, B., Nguyen, J.P. and Degos, J.D. (1991) Central pain and thalamic hyperactivity: a single photon emission computerized tomographic study. Pain, 47: 329-336. Chen, A.C.N. (1993) Human brain measures of clinical pain: a review. II. Tomographic imagings. Pain, 54: 133- 144. Coghill, R.C., Sang, C.N., Maisog, J.M. and Iadarola, M.J. (1999) Pain intensity processing within the human brain: a bilateral distributed mechanism. J. Neurophysiol., 82: 19341943. Davis, K.D., Wood, M.L., Crawley, A.P. and Mikulis, D.J. (1995) fMRI of human somatosensory and cingulate cortex during painful electrical nerve stimulation. Neuroreport, 7: 321-325. Davis, K.D., Taylor, S.J., Crawley, A.P., Wood, M.L. and Mikulis, D.J. (1997) Functional MRI of pain- and attentionrelated activations in the human cingulate cortex. J. Ne~lrophysiol., 77: 3370-3380. Derbyshire, S.W., Jones, A.K., Gyulai, E, Clark, S., Townsend, D. and Firestone, L.L. (1997) Pain processing during three levels of noxious stimulation produces differential patterns of central activity. Pain, 73: 431-445. Derbyshire, S.W.G. (1999) Meta-analysis of thirty-four independent samples studied using PET reveals a significantly attenuated central response to noxious stimulation in clinical pain patients. Cum Rev. Pain, 3: 265-280. Dettmers, C., Fink, G.R., Lemon, R.N., Stephan, K.M., Passingham, R.E., Silbersweig, D., Holmes, A., Ridding, M.C., Brooks, D.J. and Frackowiak, S.J. (1995) Relation between cerebral activity and force in the motor areas of the human brain. J. Neuruphysiol., 74: 802-8 15. Dirnagl, U., Niwa, K., Lindauer, U. and Villringer, A. (1994) Coupling of cerebral blood Row to neuronal activation: role of adenosine and nitric oxide. Am. .I. Physiol. Heart Circ. Physiol., 267: H296-H301. Dostrovsky, J.O. (2000) Role of thalamus in pain. In: J. Sandktihler, B. Bromm and G.F. Gebhart (Eds.), Nervous System Plasticit)? and Chronic Pain. Progress in Brain Research, Vol. 129. Elsevier, Amsterdam, pp. 245-257. Fox, P.T. and Raichle, M.E. (1984) Stimulus rate dependence of regional cerebral blood flow in human striate cortex, demonstrated by positron emission tomography. J. Neurophysiol., 5 1: 1109-l 120. Friston, K.J., Frith, C.D., Liddle, PF. and Frackowiak, R.S. (1991) Comparing functional (PET) images: the assessment of significant change. J. Cereb. Blood Flow Metab., I I : 690-699. Hsieh, J.-C., Belfrage, M., Stone-Elander, S., Hansson, P. and Ingvar, M. (1995) Central representation of chronic ongoing neuropathic pain studied by positron emission tomography. Pain, 63: 225-336. Iadarola, M.J., Max, M.B., Berman, K.F., Byas-Smith, M.G., Coghill, R.C., Gracely, R.H. and Bennett, G.J. (1995) Unilateral decrease in thalamic activity observed with positron

emission tomography in patients with chronic neuropathic pain. Pain, 63: 55-64. Iadarola, M.J., Berman, K.F., Zefhro, T.A., Byas-Smith, M.G., Gracely, R.H., Max, M.B. and Bennett, G.J. (1998) Neural activation during acute capsaicin-evoked pain and allodynia assessed with PET. Brain, 121: 93 l-947. Ingvar, M. (1999) Pain and functional imaging. Philos. Truns. R. Sot. Land. B Biol. Sci., 354: 1347-1358. Jones, E.G. and Pons, T.P. (1998) Thalamic and brainstem contributions to large-scale plasticity of primate somatosensory cortex. Science, 282: 1121-I 125. Kaas, J.H., Florence, S.L. and Neeraj, J. (1997) Reorganization of sensory systems of primates after injury. Neuroscientist, 3: 123-130. Lenz, EA., Lee, J.-I., Garonzik, I.-M., Rowland L.H., Dougherty, P.M. and Hua, S.E. (2000) Human thalamus reorganization related to nervous system injury and dystonia. In: J. Sandktihler, B. Bromm and G.F. Gebhart (Eds.), Nervous System Plasricin and Chronic Pain. Progress in Brain Research, Vol. 129. Elsevier. Amsterdam, pp. 259-273. Liu. X.-B., Honda, C.N. and Jones, E.G. (1995) Distribution of four types of synapse on physiologically identified relay neurons in the ventral posterior thalamic nucleus of the cat. J. Comp. Neural., 352: 69-91. MacVicar, B.A. (1997) Mapping neuronal activity by imaging intrinsic optical signals. Neuroscientist, 3: 38 I-388. Malonek, D. and Grinvald, A. (1996) Interactions between electrical activity and cortical microcirculation revealed by imaging spectroscopy: implications for functional brain mapping. Science, 272: 55 1-554. Merskey, H. and Bogduk, N. (1994) Classijcafion cf Chronic Puin: Descriptions of Chronic Pain Syndromes and DeJinitions of Pain Terms. IASP Press, Seattle, WA. Minoshima, S., Berger, K.L., Lee, K.S. and Mintun, M.A. (1992) An automated method for rotational correction and centering of three-dimensional functional brain images. J. Nucl. Med.. 33: 1579-1585. Minoshima, S., Koeppe, R.A., Mintun, M.A., Berger, K.L., Taylor, S.F., Frey, K.A. and Kuhl. D.E. (1993) Automated detection of the intercommissural line for stereotactic localization of functional brain images. J. Nucl. Med.. 34: 322-329. Minoshima, S., Koeppe. R.A.. Frey, K.A. and Kuhl, D.E. (1994) Anatomic standardization: linear scaling and nonlinear warping of functional brain images. J. Nucl. Med., 35: 1528- 1536. Mraovitch. S., Calando, Y., Pinard, E., Pearce, W.J. and Seylaz, J. (I 992) Differential cerebrovascular and metabolic responses in specific neural systems elicited from the centromedianparafascicular complex. Neuroscience. 49: 45 l-466. Ohara. P.T.. Chazal, G. and Ralston, H.J. III, (1989) Ultrastructural analysis of GABA-immunoreactive elements in the monkey thalamic ventrobasal complex. J. Comp. Neurol.. 283: 541-558. Pagni, CA. and Canavero, S. (1995) Functional thalamic depression in a case of reversible central pain due to a spinal intramedullary cyst. J. Neurosurg.. X3: 163-165. Paulson, P.E., Minoshima, S., Morrow. T.J. and Casey. K.L. (1998) Gender differences in pain perception and patterns of

287 cerebral activation during noxious heat stimulation in humans. Pain, 76: 223-229. Peyron, R., Garcia-Larrea, L., Gregoire, M.C., Convers, P.. Lavenne, F., Veyre, L., Froment, J.C., Mauguitre, F., Michel, D. and Laurent, B. (1998) Allodynia after lateral-medullary (Wallenberg) infarct. A PET study. Brain, 21: 345-356. Ploghaus, A., Tracey, I., Gati, J.S., Clare, S., Menon, R.S., Matthews, P.M. and Rawlins, J.N.P. (1999) Dissociating pain from its anticipation in the human brain. Science, 284: 19791981. Pons, T.P., Garraghty, PE., Ommaya, A.K.. Kaas, J.H., Taub, E. and Mishkin, M. (1991) Massive cortical reorganization after sensory deafferentation in adult macaques. Science, 252: 1857-1860. Rainville, P., Duncan, G.H., Price, D.D., Carrier, M. and Bushnell, M.C. (I 997) Pain affect encoded in human anterior cingulate but not somatosensory cortex. Science, 277: 968-97 1. Ralston, D.D. and Ralston, H.J. III, (1985) The terminations of corticospinal tract axons in the macaque monkey. J. Camp. Neural., 242: 325-337. Ralston, H.J. III, and Ralston, D.D. (1979) The distribution of dorsal root axons in laminae I, 11 and III of the macaque spinal cord: a quantitative electron microscope study. J. Cofnp. Neural., 184: 643-683. Ralston III, H.J., Ohara, PT., Meng, X.W., Wells, J. and Ralston, D.D. (1996) Transneuronal changes of the inhibitory circuitry in the macaque somatosensory thalamus following lesions of the dorsal column nuclei. J. Camp. Neural., 371: 325-335. Roberts, W.A., Eaton, S.A. and Salt, T.E. (1992) Widely distributed GABA-mediated afferent inhibition processes within the ventrobasal thalamus of rat and their possible relevance to pathological pain states and somatotopic plasticity. Exp. Brain Rex, 89: 363-372. Roland, P.E. (1993) Brain Activation. Wiley-Liss, New York. Roy, C.S. and Sherrington, C.S. (1890) On the regulation of the

blood-supply of the brain. J. Physiol. (Land.). 11: 85-108. Salt, T.E. (1989) Gamma-aminobutyric acid and afferent inhibition in the cat and rat ventrobasal thalamus. Neuroscience, 28: 17-26. Seitz, R.J. and Roland, PE. (1992) Vibratory stimulation increases and decreases the regional cerebral blood flow and oxidative metabolism: a positron emission tomography (PET) study. Acta Neural. Stand., 86: 60-67. Svensson, P., Minoshima, S., Beydoun, A., Morrow, T.J. and Casey, K.L. (1997) Cerebral processing of acute skin and muscle pain in humans. J. Neurophysiol., 78: 4.50-460. Swanson, L.W. (1995) Mapping the human brain: past, present, and future. Trends Neurosci., 18: 47 l-474. Talairach, J. and Toumoux, A. (1988) A Coplanar Stereofaxic Atlas of the Hwnan Brain. Thieme Medical Publishers, Inc, New York. Talbot, J.D., Marrett, S., Evans, A.C., Meyer, E., Bushnell, M.C. and Duncan, G.H. (1991) Multiple representations of pain in human cerebral cortex. Science, 251: 1355- 1358. Towe, A.L. (1995) Pyramidal tract tiber spectrum in rats, with comments on cat and man. J Brain Rex, 36: 393-398. Vanzetta, I. and Grinvald, A. (1999) Increased cortical oxidative metabolism due to sensory stimulation: implications for functional brain imaging. Science, 286: 1555- 1558. Wang, G., Tanaka, K. and Tanifuji, M. (I 996) Optical imaging of functional organization in the monkey inferotemporal cortex. Science, 272: 1665-1668. Williamson, A.M., Ohara, P.T., Ralston, D.D., Milroy, A.M. and Ralston III, H.J. (1994) Analysis of gamma-aminobutyric acidergic synaptic contacts in the thalamic reticular nucleus of the monkey. J. Camp. Neural., 349: 182-192. Worsley, K.J., Evans, A.C., Marrett, S. and Neelin, P (1992) A three-dimensional statistical analysis for CBF activation studies in human brain. J. Cereb. Blood Flow Metah., 12: 900-918.

J. Sandktihler, B. Bromm and GE Gebhart (Eds.) Progress in Brain Research, Vol. 129 0 2000 Elsevier Science B.V. All rights reserved

CHAPTER 20

Cortex areas involved in the processing of normal and altered pain Burl&u-t Bromm *, Eckehard Schareinand Christiane Vahle-Hinz Institute

for

Physiology,

University

Hospital

Introduction Most of the functional brain imaging techniques available at the end of the ‘decade of the brain’ do not measure neuronal activity itself, but changes in metabolism or local blood circulation as a consequence of enhanced neuronal activity under given mental tasks. In our opinion, in particular the multilead electroencephalogram (EEG) and the magnetoencephalogram (MEG) monitor directly and in real time brain processing mechanisms of external or internal information. With the fulminant development of highly sophisticated techniques in measuring and evaluating potentials and magnetic fields, brain source analyses emerged to be promising tools in the non-invasive investigation of mental states and their neuronal correlates under well-defined experimental conditions in humans. In principle, electrical activity in assemblies of neurons, e.g. in functional cortical columns (Szentagothai, 1973, or even in large single pyramidal cells, can be modeled by an ‘equivalent current dipole’ (briefly, ‘dipole’ or ‘brain source’), which is defined by its site, strength and orientation. With regard to physics, the dipole is part of a closed circuit: currents strive through the surrounding tissues back to their origin; these extracellular currents are the so-called ‘volume currents’. Volume currents

* Corresponding author: B. Bromm, Institute for Physiology, University Hospital Eppendorf, D-20246 Hamburg, Germany. Tel.: +49-40-42803-6170; Fax: +49-40-428034920; E-mail: [email protected]

Eppendor$

D-20246

Hamburg,

Germany

play the decisive role in the generation of the EEG but do, of course, also modify the magnetic fields (for details see Williamson et al., 1991). This chapter focuses on source analyses from simultaneously measured EEG and MEG signals with respect to pain. The specific approach used here is the application of individual MRI scans in the course of the brain source computations, presetting the space for mathematical solutions by the subject’s head morphology (Fuchs et al., 1998). Phasic pain was induced by brief radiant heat pulses from an infrared thulium laser stimulator (for review see Treede et al., 1995) or by intracutaneously applied electrical stimuli (for review see Scharein and Bromm, 1998). The laser stimulus is particularly useful for diagnostic examinations of normal and disturbed pain sensitivity in patients: it can be applied to any body area under study and norm data and norm variances are available (see also Bromm and Lorenz, 1998). The intracutaneous stimulus, best applied through the horny skin of the finger tip, is usually used when quick changes in pain perception are to be demonstrated, e.g. those caused by analgesics. Both pain models activate predominantly nociceptive A delta fibers in normal subjects, eliciting a well localizable pricking and burning painful sensation and inducing, in parallel, reliable changes of neuronal activity in defined cortex areas. In patients, these experimental ‘test pain stimuli’ are processed in a different way, as will be described below.

290

Some technical prerequisites EEG and MEG are parallel phenomena of the same

neuronal origin: electrical currents generate circular magnetic fields. The electrical currents accompanying brain activity, as measured by the EEG, are extremely weak, and the resulting magnetic fields are with 100 ff (femtoTesla) 8 orders of magnitude smaller than the magnetic field of the earth. Nevertheless, these tiny fields can be measured by means of supraconduction at very low temperatures and the quantum tunnel effect (supraconducting quantum interference device, SQUID; for technical details see: Romani and Narici, 1986; Williamson et al., 1991). Therefore, the measuring sensors or gradiometers are placed into cryostats and cooled down to, e.g. -267°C (fluid helium). Meanwhile, worldwide about hundred MEG machines are working on problems of functional brain imaging. We use a 2 cryostat system (Fa Philips, Hamburg: 31/7 gradiometers, coil distances 25 mm, baseline 80 mm). Brain source analysis evaluates the magnetic fields and electrical potential maps measured time point by time point over the head. Algorithms are based on the mathematical postulate for a ‘minimum norm approximation’ (cf. Wang et al., 1992; Fuchs et al., 1994), which means that nature generates signals with the smallest possible expenditure, a questionable assumption if we regard the fundamental principle of multiple processing and redundancy in higher brain functions. The simplest model for an activated brain area is the ‘one moving dipole’ with coordinates for its site and direction - following Szentagothai’s cortical columns. Source identification starts with a dipole preset anywhere, from which the resulting fields or potentials are calculated. Normally, 50 numerical iterations are sufficient to achieve agreement between calculated fields or potentials and the measured ones within a preset ‘goodness of fit (gof)’ *. Other models are fixed or

’ Since different terms are used in literature to characterize the accuracy of fits, here the exact physical definitions: Standard deviation (dev) describes the difference between measured and calculated fields in the known way. Residual variance (var) describes the difference in energy between measured and calculated fields: var = dev2. If both terms are expressed in % of mean global field power, we obtain var (%) = dev* x 100 (a). Good-

rotating dipoles with a preset position over a certain time period in which only their strengths and/or orientations are allowed to vary. More complicated computations use multiple dipoles or current density and dipole scan procedures which will be described below. Many problems in source analysis exist; the major ones should be addressed briefly in order to enable the reader to critically judge relevant publications. Most of these problems basically meet those of PET, SPECT and other imaging techniques addressed in this book as well. The central problem is the correct data transfer from the scanner (cryostat) system in which the values are measured to the individual head system in which brain sources are to be localized. This includes the exact head position control during the measurements - except for multilead EEG were the measuring sites are wired with the head. We solved this problem for MEG analysis by fixing small coils at the EEG electrode positions which are marked in the MRI scans; between blocks of measurement tiny currents are fed into these coils which broadcast calibrating fields into the cryostat system (Fuchs et al., 1998). The second decisive problem concerns the correct attribution of calculated sources to specific structures of the individual brain. Most researchers calculate the sources in a spherical space consisting of different shells for skin, bone and brain. In a last step the coordinates of the estimated source are mirrored into the standardized stereotaxic space (Talairach and Toumoux, 1988); recent studies used individual MRI scans for final positioning. Our approach is to use the 80,000 voxels of the individual MRI to define the space of mathematical solutions for source localization in the individual cortex. The morphological data are also used to calculate the volume currents (see above) using the Boundary Element Method (BEM; e.g. Hamalainen and Sarvas, 1989; Fletcher et al., 1995) in a head model more realistic than simple spheres. Individual treatment of data generates individual results. Therefore, repeated experiments in the same subject or patient are required to validate the

ness of fit (gof) describes the percentage the model: gof = (100 - var) (%).

variance

explained

by

291

findings. For comparison of data across different individuals we use a relative PAN system with the head markers Pre Auricular points and Nasion: x-axis from left to right ear, y-axis from its center (origin 0) to the nasion, and z-axis from 0 perpendicularly upon the X, y plane upwards, penetrating the head - on the average - approximately 1.5 cm behind the vertex. Since the dimensions of heads are different in X, y and z dimensions, the units of each axis were standardized by setting each extension to 100% (PAN-100 system). Not only the coordinates of identified brain sources but also those for neuroanatomically defined brain areas (SI, SII, cingulum) are described in terms of PAN-100 coordinates; this way a quick orientation of dipole sites in the individual brain is achieved. The PAN system differs from the BESA (Brain Electrical Source Analysis) system used by Scherg et al. (e.g. Scherg, 1992); transfer of the coordinates between both systems is difficult, when realistic head models are regarded (for details see Towle et al., 1993). Bilateral SII activity reflecting the sensory-discriminative pain component With the development of magnetoencephalography the investigation of primary (SI) and secondary somatosensory cortex areas (SII) in humans came to the fore. These areas are rather superficial and their columns result in equivalent current dipoles with a marked tangential direction, the fields of which can well be measured outside the head. The first arrival of somatosensory information in the cortex is reflected in the earliest magnetic SI responses to conventional electrical nerve stimulation and these have mostly been reported to occur strictly contralaterally to the site of stimulus application. In a thorough analysis with the CURRY software, the sites of the SI components N20m and P30m after median nerve stimulation could be differentiated with a spatial accuracy of better than 3 mm (Buchner et al., 1996). Similarly, Walter et al. (1992) demonstrated a somatotopy in magnetic SI responses with high intra-individual consistency in space. In our laboratory, repeated experiments performed with the same subjects using stimulation of the median, radial, tibial or peroneal nerves resulted in intra-individual re-detection accuracy in the range of 3 mm. Thus,

source analysis of MEG responses in SI may be used as functional landmarks for presurgical exploration or for the investigation of neuronal reorganization and neuroplasticity (Flor et al., 1995; Birbaumer et al., 1997). With pain-inducing stimuli, early SI responses have not been found so far, neither in EEG nor in MEG investigations (for review and reasoning see Bromm and Lorenz, 1998). Considerable evidence has been collected for the involvement of the secondary somatosensory cortex in the processing of pain. SII areas are located along the superior bank of the Sylvian fissure, lateral and inferior to the face representation in SI, anterior and medial to the primary auditory areas (Fig. 1). These areas correspond cytoarchitectonically to parts of Brodmann’s areas 40 and 43 and have mostly been investigated in monkeys and cats, where SII seems to exhibit some somatotopic organization, though with extensive overlaps; the face is represented laterally and the legs most medially (Brodal, 1985; Burton, 1986). The SII cortex is commonly divided into an anterolateral part, located directly on the inferior border of the SI area, which has a somatotopic organization, and a posterolateral part with no obvious somatotopy (e.g. Whitsel et al., 1969; Manzoni et al., 1990; Creutzfeldt, 1993). Neurons in the posterior part are polysensory, respond to noxious stimuli, and possess relatively large receptive fields. The somatotopically organized SII receives nociceptive information from the spinothalamic tract (SIT) via relay stations in the thalamic VP1 and VMpo nuclei. Information from the dorsal column pathway reaches SII mainly via the VP nuclei. ST1 projects to SI, the parietotemporal cortex, the posterior insula and deep cortical structures like the amygdala, the septum and the hippocampus. Since these projections are reciprocal, a variety of feedback loops are possible. Before the era of MEG research, only very limited knowledge about human SII organization existed; most data were derived from neurosurgical observations (Woolsey et al., 1979; Van Buren, 1983). However, already in their first MEG publication, Hari et al. (1983) described activation of SII by painful skin and tooth pulp stimulation. These observations meanwhile have been extended in various MEG centers using painful chemical stimuli of the nasal mucosa (Huttunen et al.. 1996a,b; Hari et al., 1997),

Fig. 1. Body representation in primary (SI) and secondary representation in SI, extending along the upper operculum deep fissure is in the range of 30 to 40 mm. To the left a coronal Brodmann’s areas 40 and 43 (composed with data from: Kandel

(SII) somatosensory cortex. SII lies lateral and inferior to the face into the superior bank of the Sylvian fissure. The depth of the Sylvian projection of SII and insula (medial). SII corresponds to portions of et al., 1991; Burton et al., 1995; Brodal, 1998).

intracutaneous electrical stimulation of the finger tip (Joseph et al., 1991; Howland et al., 1995), or laser stimuli applied to the hand (Kakigi et al., 1995, 1996; Laudahn et al., 1995; Mauguiere et al., 1997) or temple (Laudahn et al., 1995; Bromm et al., 1996). In full agreement, all authors found a bilateral activation in corresponding SII areas though stimuli were applied strictly unilaterally. An example is shown in Fig. 2 where painful intracutaneous stimuli were applied to either the left or right finger tip. The results document two facts: firstly, under both the conditions of only left or only right middle finger stimulation a bilateral activation was found. Secondly, the SII dipoles for left or right middle finger stimulation are slightly different indicating a somatotopic organization. Obviously, different ‘paired SII areas’ are responsible for pain processing from left or right finger. These findings are statistically significant, as documented in repeated experiments and

sessions with the same subject. Further experiments with laser stimuli applied to the temple, hand or foot, resulted in similar findings documenting a clear somatotopy of ‘paired bilateral activation’ in SII. A crucial function of SII in the interhemispheric transfer of nociceptive information is documented in humans by functional imaging techniques, such as fMR1 (Oshiro et al., 1998), PET (Xu et al., 1997; Svensson et al., 1998), as well as by direct intra-cortical recordings (Frot et al., 1999a,b). Interestingly, the average over a sample of 18 subjects showed no difference between ipsi- and contralateral SII activation time. Statistical analysis revealed a marked inter-individual variability in latency of peak mean global field power; however, repeated measurements in the same individuals indicate a high consistency in the time sequence of activation after unilateral stimulation. This fits with data which make a sequential activation via transcallosal fibers unlikely. Instead,

293

Fig. 2. Bilateral SII activation by painful stimulation of left or right middle finger. Dipoles at peak latency 130 ms are calculated in the individual brain by CURRY software and boundary element method (BEM). With high intra-individual constancy two different pairs of dipoles could be differentiated if left or right (marked) middle finger was stimulated. Experimental details: blocks of 60 stimuli with randomized intensities between 1.O and 2.5 fold individual pain threshold strength were intracutaneously applied to left or right finger tip. MEG was measured by a Philips biomagnetometer with cryostats positioned over C3 and C4 (from Kohlhoff et al., 1997).

data from animal experiments and observations in humans support the idea that pain-related activity in ipsi- and contralateral SII originates already from bilateral spinothalamic projections and thalamic relay stations (e.g. Willis, 1985; Creutzfeldt, 1993). Pain-evoked SII activity depends on the arousal level, the vigilance of the subject under investigation. The most striking effect can be observed when sedative drugs are administered, such as tranquilizers. In other words, SII activity appears to be ‘tonically preprimed’ by projections from subcortical and cortical structures controlling the arousal state of the brain.

Decreased vigilance or modification of attention to the painful event attenuates SII activity in response to the standard pain test stimuli. This has been shown by investigations in cooperation with the Department of Anesthesiology in Hamburg, where the effects of pain processing were studied with bolus injections of clonidine (3 pg/kg) which exhibits a strong sedative component, making the subjects drowsy and sleepy. Fig. 3 gives an example which shows the mean dipole strength (averages over all points of measurements) as a function of post-stimulus time for different experimental blocks before and after clonidine injection. The thick black line indicates stimulus (intracutaneous electrical stimuli)-induced SII activity before treatment. In this subject we see a double peak: the generator of the first maximum was identified in the medial wall of the Sylvian fissure, the second one in the parietal operculum. Injection of clonidine induced an increasing sedation down to sleep stage I, an attenuation of SII activity, and in parallel - a decrease of pain ratings (Bischoff et al., 2000). The time course of the clonidine effect is slow, in agreement with the known pharmacokinetics (Rudd and Blaschke, 1985). Twenty minutes after injection, subjects were asleep and an attenuation of SII activity was found. After 2 h full recovery occurred on the subjective and objective level of measurement. Comparable attenuation of SII activity was found with the dissociative anesthetic ketamine (racemate; 0.5 mg/kg body weight), which induces unconsciousness for 5-10 min. During this period, SII activity was drastically decreased and, of course, no pain ratings were available. The role of SII as a higher integrative center linked to conscious sensory perception is under debate in recent literature. Mima et al. (1998) found with MEG measurements in humans that SII responses were enhanced by attention more prominently than the SI responses. Distraction, in contrast, resulted in an attenuation of pain-related SII activity and pain ratings (Petrovic et al., 2000). Inactivation of SII by lidocaine or barbiturates (Jackson and Cauller, 1998) in the awake animal (rat) left the short-latency component of electrocutaneously evoked potentials in SI intact but almost completely eliminated the long-latency components in SI. SII obviously controls long-latency components in SI; important parts of the ‘cognitive’ signal processing

294 80 r

I

I

100

150

I 200

time [ms] Fig. 3. Clonidine-induced sedation decreases SII activity in response to painful stimuli. Dipole strengths versus post-stimulus time (intracutaneous pain model, left middle finger) before (pre) and 15, 30, 45, 60, and 120 min after iv. clonidine bolus (0.2 mg). The slow decrease in action and complete recovery after 2 h (posd) of SII activity was in parallel with the varying level of vigilance.

and conscious sensory perception depend crucially on the activity in SII. When summarizing all our findings concerning SII responses to painful stimuli, we may state that SII activity reflects parts of the sensory-discriminative component of pain: bilateral activity compares the hurting side with the non-hurting side. Somatotopy informs us about the site of the hurting event, and stimulus response characteristics allow to rate the strength, even the quality of pain. Moreover, if SII activity is decreased or even completely abolished under deep anesthesia, the brain is not able to explore the magnitude, the quality of pain, and the site where it hurts. As a consequence we may speculate that the subject is not able to feel pain, even if the cingulate gyrus is still active (see below) and may elicit all those motor and autonomic responses, such as withdrawal reflexes or their inhibition, sweating, changes in heart rate, blood pressure, and respiration, which all have to be treated by the anesthesist with specific means. The cingulate gyrus and the affective and attentional components of pain Best known for more than 20 years are the so-called long-latency brain responses induced by painful

stimuli which have mostly been investigated in evoked EEG measurements and appear at latencies of 200 ms and more. The late ‘pain-related’ brain potentials consist in particular of a large vertex positivity (P250) with a bilateral symmetrical spatial distribution. The amplitude of this positivity depends strongly on the magnitude of stimulus perception and its painfulness, if sources of distortion are well under control. This has been shown in a principal component analysis with different stimulus qualities and intensities (Bromm and Scharein, 1982). In patients with a loss in thin fiber and/or anterolateral tract function, the vertex positivity in particular was found to be reduced; in other patients with hyperalgesia, e.g. due to fibromyalgia, this component was significantly enlarged (for review see: Gibson et al., 1994; Bromm and Lorenz, 1998). A block of myelinated fibers, experimentally or by disease, caused the P250 to disappear, but revealed an ultralate P1250 as correlate of pure C-nociceptor mediation (Bromm and Treede, 1987). Moreover, efficient analgesics attenuate significantly the long-latency pain-related brain potential, a fact which allows the comparative evaluation of analgesic potency of drugs (for review see Scharein and Bromm, 1998). Brain generators describing the long-latency vertex positivity in response to painful laser stimuli have

295

been identified in the posterior cingulate gyms (CG) by means of brain electrical source analysis (Bromm and Chen, 1995). However, MEG studies failed to show distinct CG responses to phasic pain. The reason is that the dipoles of these long-latency brain signals exhibit a predominating radial direction with an origin near the center of the head. Radial dipoles generate magnetic fields which do not penetrate the head, at least not to a measurable degree. Volume currents shield the magnetic fields and, moreover, parietal cryostat positions do not discriminate these deep brain structures within the limit of their spatial resolution. In order to investigate these important generators in the cingulum, we therefore have to use multilead EEG recordings, which can be analyzed by the same means of CURRY software using the constraint of the individual head morphology in source identification. CURRY applied to multilead EEG maps (up to 64 electrodes) at latencies above 200 ms proved that for a period of around 60 ms all late activity can be explained by one single dipole deep in the brain around the sulcus longitudinalis, in the cingulate gyms (Fig. 4). The direction of this generator is radial, the current dipole points in positive direction, following the convention in physics. Interestingly, the dipole identified for a period of at least 60 ms in the cingulum is not constant in space but changes its site in a systematic way: in a latency range between 220 and 280 ms maximum cingulate activity appeared to be shifted from caudal CG towards the front and to disappear in the prefrontal cortex (Bromm et al., 1996). These results were found very reliably within repeated sessions in the same subject and were principally similar in different subjects. The dipole in the cingulate gyms explains the pain-related large vertex positivity measured in the series of studies summarized above. In other words, all findings collected for the pain-related large vertex potentials can be attributed to the cingulate activity shown here. As a consequence, the strength of CG dipoles covaries with the unpleasantness of the stimulus applied according to the principal component analysis cited above. Centrally acting analgesics are effective at these structures of CG as has been observed in a series of studies (Bromm and Scharein, 1982). In a recent MEG investigation, we observed strong attenuation of pain-induced CC activity by

the p,-opioid-receptor agonist nortilidine, whereas the NMDA-receptor-antagonist ketamine, in spite of its strong anesthetic action, exhibited only little effect upon the cingulum. To sum up, CG activity in response to pain-inducing stimuli can be attributed to the aversive emotional component of pain which describes its hurting, aching, torturing and most unpleasant character and which is attenuated by morphine-like drugs. Our finding that painful stimuli start cingulum activation in posterior parts, first published in a brain electrical source analysis study (Bromm and Chen, 1995), is in contradiction to the many PET investigations which unanimously underline the decisive role of the anterior cingulate gyms in the processing of experimental pain (see also Casey, 2000, this volume). In all our experiments, activity in the posterior CG is prominent with laser stimuli applied to temple, hand and foot, or electrical stimuli applied to the middle finger tip; the sites are comparable in all subjects and for all pain models used (see Freybott et al., 2000). The discrepancy may be explained by the much longer time needed for data collection in PET studies, which may obliterate cingulate activity at those areas which are active for long periods. More likely is the involvement of different parts of the CG in the processing of persistent tonic or of phasic pain which was used in our studies. In view of pain processing, the CG may be divided into two parts: a posterior one comprising Brodmann’s areas 23, 29, 30 and 31 and an anterior part encompassing areas 24, 25 and 32 (Zilles, 1990; Vogt and Gabriel, 1993). In Fig. 5 we have extracted from the many connections those which may carry nociceptive signals to the CG. The anterior CG (CGa) receives its dominant thalamic inputs from autonomic regions (mediodorsal nucleus, midline nuclei), from motor regions (ventral anterior and ventral medial nuclei), from limbic regions (anterior medial and mediodorsal nuclei) as well as sensory and integrated sensory/motor regions (intralaminar and mediodorsal nuclei). In addition, prominent connections exist with the amygdala and hippocampus, as well as with the insula, which in turn interacts with SII. Pain-relevant inputs are transmitted via the spinothalamic tract (STT) projections and their trigeminal analogue to the intralaminar nuclei (IL) and the mediodorsal nucleus (MD). They are sub-

296

220 ms 94.2 %

297 parietotemporal

cortex A

P’B

\ motor

autonomic

t S‘TT

b

Sk t PsDC

8

Fig. 5. Pain-related connections of the anterior (CGu) and posterior (CGp) cingulate gyrus. Nociceptive input to CGa is transmitted by the spinothalamic tract (XT) mainly via the mediodorsal nucleus (MD) and intralaminar nuclei (IL) of the thalamus as well as via the amygdala and the insula. Its major outputs activate motor and autonomic regions in the brainstem and spinal cord. STf neurons also transmit nociceptive information to CGp, via the parabrachial nucleus (PB) of the brainstem, the amygdala and the thalamic laterodorsal nucleus (LD). A fast nociceptive and multimodal input originates from the postsynaptic dorsal column tract (PSDC) via the superior colliculus (SC). CGp interacts with wide regions of the parietal and temporal cortices including SK

jetted in CGa to findassociative processing and initiate the efferent control of motor and autonomous nocifensive behavior. No wonder that most brain imaging investigations with tonic experimental as well as chronic pain document the prominent role of CGa in nociception and pain. It should be kept in mind that many of the connections are of reciprocal character and that not only somatic but also other sensory and motor information, such as ocular, and complex information is integrated in the anterior CG (Musil and Olson, 1993). These connections provide the basis for a function in motivation, affective coloring of perception, initiation of behavior, and modulation of autonomic function. These probably

play a role in behavioral responses to stimuli, not only in motor reactions, like withdrawal reflexes, but in particular in the initiation of sweating, changes in blood pressure and respiration, i.e. nocifensive reactions seen by each anesthesiologist and interpreted as the onset of pain. The posterior CG (CGp) receives its dominant inputs from the hippocampus and amygdala via the anterior thalamic nuclei and the laterodorsal nucleus (LD), and is intimately linked with the sensory association areas of the parietotemporal cortex including SII. The pathways relevant to processing of pain and especially those providing fast, direct transmission of pain signals are largely unknown. The LD receives multimodal inputs from the superior colliculus (SC), which in turn is activated by projections from the dorsal column nuclei, presumably including postsynaptic dorsal column neurons (PSDC). The PSDC projections originate in laminae IV and V of the dorsal horn and carry high threshold nociceptive as well as wide dynamic range responses (WDR) mainly transmitted from A delta fibers. It is conceivable, therefore, that phasic pain stimuli, such as the brief radiant heat pulse and the intracutaneous electrical stimulus, which are known to activate nociceptive A delta fibers, elicit activity in the CGp. Moreover, collaterals of the STT project through the parabrachial nucleus (PB), a visceroceptive region of the brainstem, and the amygdala into the laterodorsal nucleus (LD) and from there into CGp (Brodal, 1998). The output to and input from wide regions of the parieta1 and temporal cortices indicate that CGp plays an interactive role in monitoring of sensory inputs and movements, thereby providing maintenance of spatial orientation, memory, attention and evaluation of the significance of stimuli for the organism. In this context, it may be mentioned that the activity in the posterior CG strongly depends on the arousal state of the brain (Frederikson et al., 1995). We still have to discuss the ‘movement’ of the ‘pain-related’ generator within the cingulum from

Fig. 4. Cingulate activity in response to pain-inducing stimuli. For radial dipoles multilead EEG recordings were evaluated by CURRY software (calculations on the individual cortex anatomy with BEM for volume currents). At latencies between 220 and 280 ms, all late vertex positivity could be explained by one moving dipole in the cingulate gyrus (CG) which appears in posterior CG at 220 ms, then moving towards anterior CG (280 ms) and fading out in frontal cortex around 320 ms after stimulus onset; left/right differences are not significant. Left temple stimulation by pain-inducing infrared laser stimuli (modified from Bromm et al., 1996).

298 posterior to frontal parts. We do not believe that the generator moves via the many intracingulate connections, since the latency of appearance at neighboring sites is too long. In contrast, we have to stress that the sites of the dipole describe the maximum activity which may also result from activity in distinct pathways. Information through PSDC is expected to be conducted faster than via the polysynaptic STT route. Furthermore, phasic pain exhibits several perceptive components: a fast sensory-discriminative (seen in CGp) and a slower motivational-affective one (seen in CGa). A third possibility is indicated by the circuit via LD, amygdala and MD which allows projection of information from the posterior to the anterior CG. Our results, therefore, are in agreement with the wiring of afferent and efferent projections to and from the cingulate cortex and may reflect the multiple involvement of the limbic system in emotional experiences of stimulus recognition or in the coordination of behavioral responses to the pain-inducing stimuli applied. Allodynia due to neuronal reorganization central pain

in

Lesions at all levels of the central nervous system can cause central pain. Persistent central pain can change normal pain pathways into abnormal projections and recurrent neuronal circuits (cf. following chapters) which modify perception also of phasic pain test stimuli. In context to the main theme of this book a study in a patient may be added which indicates the possibility of modulations in thalamocortical projections from mechanosensation into pain due to neuroplastic changes in the course of persisting central pain. This case clarifies the need of the careful determination of brain areas involved in normal pain perception in order to look for pathological variations. The patient suffering from Wallenberg’s disease was examined with MEG in a follow-up study, in the acute phase 3 weeks after symptom onset and 1 year later. During this year, central pain had progressively developed (Fig. 6; Lorenz et al., 1998). The acute symptoms were typical for an infarction of the left posterior inferior cerebellar artery which supplies neuronal structures of the lateral brainstem, including the spinothalamic tract, the descending

tract of the trigeminal nerve, the central descending sympathetic tract and pontocerebellar pathways. Consequently, the patient showed loss of pain and temperature senses contralesionally in the right trunk and limbs and ipsilesionally in the face; furthermore ipsilesional Homer’s triad (miosis, ptosis, enophthalmus) and cerebellar ataxia, whereas sensory information related to touch, vibration and joint position remained intact. After anticoagulative treatment the patient’s state markedly improved; however, painful dysesthesia developed during the year under observation in the affected right leg with allodynic pain in response to light touch. In the first MEG investigation, we found normal tibia1 SI responses for left and right tibia1 nerve stimulation, consisting of the two most prominent peaks P40 and N80. During the development of allodynia, the component at 80 ms became larger when the affected leg was stimulated. MEG source analysis of this component now identified a new generator which was not localized in SI but in the posterior CG. This generator was not observed when the unaffected left leg was stimulated. Activation of CG already at 80 ms in response to electrical tibia1 nerve stimulation clearly is too early to be mediated through peripheral A delta nociceptors (for latencies see Treede et al., 1995). The findings in this patient rather suggest mediation by fast conducting A beta fibers and an abnormal projection via the medial lemniscal tract. It is known that A beta fibers can signal allodynia and hyperalgesia after peripheral nerve injury, probably on the basis of central sensitization at the dorsal horn or higher level in convergent, wide dynamic range (WDR) neurons (see Moore et al., 2000, this volume; Sandktihler et al., 2000, this volume). The PSDC pathway, therefore, may mediate fast activation of thalamic and cortical nociceptive regions in this case. On the other hand, thalamic regions receiving STT projections are disconnected from their inputs which may lead to deafferentation hypersensitivity and furthermore to an imbalance between different nociceptive regions, e.g. the medial and lateral pain systems (see Fig. 5; Cesaro et al., 1991). Thalamic neurons deprived of their inputs have been shown to generate bursting activity which has been suggested to subserve central pain (Lenz et al., 1989, 2000, this volume; Lenz and Dougherty, 1995; Jean-

299

Fig. 6. Abnormal cingulate activity in a case of central pain with allodynia. Shown are the SI dipoles of the magnetically measured N80m in response to conventional electrical right tibia1 nerve stimulation (inner ankle) which normally appears, together with the early P40, in the primary somatosensory cortex SI, contralateral to the stimulated side near the vertex. In this patient suffering from central pain with severe allodynia in the right leg after a brainstem lesion (Wallenberg syndrome), a second source at around 80 ms appeared in the cingulum when light electrical stimuli were applied to the affected leg which generated burning pain. Stimulation of the healthy side only showed the normal N80m in contralateral SI (from Lorenz et al., 1998).

monod et al., 1996). Interestingly, Dostrovsky (2000, this volume) presents data demonstrating that the correlation of bursting activity with central pain syndromes may have been overestimated. The concepts of central pain, therefore, are still under debate with respect to the importance and anatomical basis of the various postulated interactions and phenomena. However, there is a broad agreement that damage of the spinal and supraspinal pain and temperature pathways causes a dysbalance of the system which leads to disinhibited spontaneous nociceptive activity and pain in response to innocuous stimuli in the presence of decreased physiological exteroceptive function (for a comprehensive review see Boivie, 1995). Conclusions Combined analyses of multichannel MEG and EEG responses to pain-inducing standard test stimuli are

suitable tools for the exploration of normal and disturbed pain pathways and in particular for the identification of cortical areas relevant for the perception of pain. The results of brain source analysis described here are based on realistic head models and individual cortical gyrus morphology; they excellently underline the hypothesis of multiple representation of pain in the human cortex. Analyses of activity in the individual cortex morphology are mandatory since we know from neuroradiological studies about the considerable variations in individual head and brain anatomy in comparison to the standard Talairach atlas. Even within individuals, large asymmetries have been observed between left and right hemispheres; e.g. side differences of the central sulcus can amount to centimeters (for review see White et al., 1997). With appropriate software, the various components of pain can clearly be attributed to neuronal activity in different well-defined cortical structures.

300 With these techniques we are able to correlate cortical structures with the processing of sensorydiscriminative components of pain and others which reflect the motivational-affective one. In analogy to somatosensation we can define a lateral pain system with the cortical representatives contralateral SI and bilateral SII with characteristics of somatotopy, stimulus specificity and stimulus response dependency. This system strongly depends on the momentary arousal level of the brain; in states of deep sleep or of unconsciousness, the bilateral SII activity in response to standard pain stimuli is significantly attenuated. The medial system associated with the motivational-affective component of pain is represented in the anterior CG with afferent and efferent control of emotional experiences. With the phasic stimuli and the functional imaging techniques with high time resolution used in our studies, the important role of the posterior CG emerged. In intimate connections with the parietotemporal cortex CGp focuses attention towards the stimulus event and evaluates its significance for the organism. In parallel to the subjective pain ratings these structures can selectively be attenuated by specific drugs which act differently upon the sensory-discriminative, the motivational-affective and the attentional-evaluative components of pain. The involvement of these cortical structures depends on the individual pain perception and exhibits modifications due to neuroplasticity under chronic pain. This knowledge may be used to tailor analgesic treatments with respect to site and mode of action. Careful inspection of neuroplastic changes in these cortical centers of pain processing demonstrates all the many variations in how to perceive, how to manage and how to suppress pain. References Birbaumer, N., Lutzenberger, W., Montoya, P., Larbig, W., Unertl, K., Topfner, S., &odd, W., Taub, E. and Flor, H. (1997) Effects of regional anesthesia on phantom limb pain are mirrored in changes in cortical reorganization, J. Neurosci., 17: 5503-550s. Bischoff, P., Scharein, E., Schmidt, G.N., Von Knobelsdorff, G., Bromm, B. and Schulte am Esch, J. (2000) Topography of clonidine-induced EEG changes evaluated by principal component analysis (PSA). Ancesthesiology, 92: 1545-1552 Boivie, J. (1995) Pain syndromes in patients with CNS lesions and a comparison with nociceptive pain. In: B. Bromm and J.

Desmedt (Eds.), Pain and the Brain. New York: Raven Press, pp. 367-376. Brodal, A. (1985) Neurological Anatomy in Relation to Clinical Medicine. Oxford University Press, New York. Brodal, P. (1998) The Central Nervous System. Oxford University Press, Oxford. Bromm, B. and Chen, A.C. (1995) Brain electrical source analysis of laser evoked potentials in response to painful trigeminaJ nerve stimulation. Electroencephalogr Clin. Neurophysiol., 95: 14-26. Bromm, B. and Lorenz, J. (1998) Neurophysiological evaluation of pain. Electroencephalogr Clin. Neurophysiol., 107: 227253. Bromm, B. and Scharein, E. (1982) Principal component analysis of pain-related cerebral potentials to mechanical and electrical stimulation in man. Electroencephalogr Clin. Neumphysiol., 53: 94-103. Bromm, B. and Treede, R.D. (1987) Pain related cerebral potentials: late and ultralate components. Znt. J. Neurosci., 33: 15-23. Bromm, B., Lorenz, J. and Scharein, E. (1996) Dipole source analysis of brain activity in the assessment of pain. In: .I. Kimura and H. Shibasaki (Eds.), Recent Advances in Clinical Neurophysiology. Amsterdam: Elsevier, pp. 328-335. Buchner, H., Waberski, T.D., Fuchs, M., Drenckhahn, R., Wagner, M. and Wischmann, H. (1996) Postcentral origin of P22: evidence from source reconstruction in a realistically shaped head model and from a patient with a postcentral lesion. Electroencephalogr Clin. Neurophysiol., 100: 332-342. Burton, H. (1986) Second somatosensory cortex and related areas. In: E.G. Jones and A. Peters (Eds.), Cerebral Cortex. Plenum Press, New York, pp. 31-98. Burton, H., Fabri, M. and Alloway, K. (1995) Cortical areas within the lateral sulcus connected to cutaneous representations in areas 3b and 1: a revised interpretation of the second somatosensory area in macaque monkeys. J. Comp. Neural., 355: 539-562. Casey, K.L. (2000) Concepts of pain mechanisms: the contribution of functional imaging of the human brain, In: J. Sandktihler, B. Bromm and G.F. Gebhart (Eds.), Nervous System Plasticity and Chronic Pain. Progress in Brain Research, Vol. 129. Elsevier, Amsterdam, pp. 277-287. Cesaro, P., Mann, M.W., Moretti, J.L., Defer, G., Roualdts, B., Nguyen, J.P. and Degos, J.D. (1991) Central pain and hyperactivity: a single photon emission computerized tomographic study. Pain, 41: 329-336. Creutzfeldt, O.D. (1993) Cortex Cereb. MPI, Biophysikalische Chemie, Gottingen. Dostrovsky, J.O. (2000) Role of thalamus in pain. In: J. Sandkiihler, B. Bromm and G.F. Gebhart (Eds.), Nervous System Plasticity and Chronic Pain. Progress in Brain Research, Vol. 129. Elsevier, Amsterdam, pp. 245-257. Fletcher, D.J., Amir, A., Jewett, D.L. and Fein, G. (1995) Improved method for computation of potentials in a realistic head shape model. IEEE Trans. Biomed. Eng., 42: IO94- 1104. Flor, H., Elbert, T., Knecht, S., Wienbruch, C., Pantev, C., Birbaumer, N., Larbig, W. and Taub, E. (1995) Phantom-limb

301 pain as a perceptual correlate of cortical reorganization following arm amputation. Nature, 375: 482-484. Frederikson, M., Wik, G., Fischer, H. and Anerson, J. (1995) Affective and attentive neural networks in humans: a PET study of Pavlovian conditioning. Neuroreport, 7: 97-101. Freybott, K., Elhich, J., Scharein, E. and Bromm, B. (2000) Anterior versus posterior cingulate gyms processing phasic pain. Eur: J. Physiol. (Pjtigers Archiv), 439: R311 . Frot, M., Rambaud, L. and Mauguiere, F. (1999a) Intracortical recordings of early pain-related COz-laser evoked potentials in the human second somatosensory (SII) area. Clin. Neurophysiol., 110: 259-267. Frot, M., Mauguiere, F., Hayashi, T., Konishi, S., Hasegawa, I. and Miyashita, Y. (1999b) Timing and spatial distribution of somatosensory responses recorded in the upper bank of the sylvian fissure (SII) in humans. Short communication: mapping of somatosensory cortices with functional magnetic resonance imaging in anaesthetized macaque monkeys. Eur J. Neurosci., 11: 445 l-4456. Fuchs, M., Wischmann, H. and Wagner, M. (1994) Generalized minimum norm least squares reconstruction algorithms. ISBET Newsl., 5: 8-11. Fuchs, M., Drenckhahn, R., Wischmann, H. and Wagner, M. (1998) An improved boundary element method for realistic volume conductor modeling. IEEE Trans. Biomed. Eng., 45: 980-997. Gibson. S.J., Littlejohn, G.O., Gorman, M.M., Helme, R.D. and Granges, G. (1994) Altered heat pain thresholds and cerebral event-related potentials following painful COz laser stimulation in subjects with fibromyalgia syndrome. Pain, 58: 185-193. Hamimllainen, M.S. and Sarvas, J. (1989) Realistic conductivity geometry model of the human head for interpretation of neuromagnetic data. IEEE Trans. Biomed. Eng., 36: 165 17 1. Hari, R., Hamalainen, M., Kaukoranta, E., Reinikainen, K. and Teszner, D. (1983) Neuromagnetic responses from the second somatosensory cortex in man. Acta Neural. Scund., 68: 207212. Hari, R., Portin, K., Kettenmann, B., Jousmaki, V and Kobal, G. (1997) Right-hemisphere preponderance of responses to painful CO:! stimulation of the human nasal mucosa. fain, 72: 145-151. Howland, E.W., Wakai, R.T., Mjaanes, B.A., Balog, J.P. and Cleeland, C.S. (1995) Whole head mapping of magnetic fields following painful electric finger shock. Brain Res. Cogn. Brain Res., 2: 165-172. Huttunen, J., Wikstrom, H., Korvenoja, A., Seppalainen, A.M., Aronen, H. and Ilmoniemi, R. (1996a) Significance of the second somatosensory cortex in sensorimotor integration: enhancement of sensory responses during finger movements. Neuroreport, 7: 1009-1012. Huttunen, J., Kobal, G., Kaukoranta, E. and Har, R. (1996b) Cortical responses to painful CO2 stimulation of nasal mucosa: a magnetoencephalographic study in man. Electroencephalogr Clin. Neurophysiol., 64: 347-349. Jackson, M.E. and Cauller, L.J. (1998) Neural activity in SII

modifies sensory evoked potentials in SI in awake rats. Neuroreport, 9: 3379-3382. Jeanmonod, D., Magnin, M. and Morel, A. (1996) Low-threshold calcium spike bursts in the human thalamus. Common physiopathology for sensory, motor and limbic positive symptoms. Bruin, 119: 363-375. Joseph, J., Howland, E.W., Wakai, R., Backonja, M., Baffa, 0.. Potenti, EM. and Cleeland, C.S. (1991) Late pain-related magnetic fields and electric potentials evoked by intracutaneous electric finger stimulation. Electroencephalogr Clin. Neurophysiol., 80: 46-52. Kakigi, R., Koyama, S., Hoshiyama, M., Kitamura, Y., Shimojo, M. and Watanabe, S. (1995) Pain-related magnetic fields following painful CO2 laser stimulation in man. Neurosci. L&t., 192: 45-48. Kakigi, R., Koyama, S., Hoshiyama, M., Kitamura, Y., Shimojo, M. and Watanabe, S. (1996) Pain-related brain responses following COz laser stimulation: magnetoencephalographic studies. Electroencephulogr Clin. Neurophysiol., 47: 11 l-120. Kandel, E.R., Schwarz, J.H. and Jessell, T.M. (1991) Principles of Neural Science. Elsevier, New York. Kohlhoff, H., Scharein, E. and Bromm, B. (1997) Involvement of somatosensory cortices in the processing of pain, combined MEG and EEG recording. J. Neural. Sci.. 150: 101. Laudahn, R., Kohlhoff, H., Bromm, B. (1995) Magnetoencephalography in the investigation of cortical pain processing. In: B. Bromm and J.E. Desmedt (Eds.), Pain and the Brain: From Nociception to Cognition. Raven Press. New York, pp. 267-281. Lenz, F. and Dougherty, PM. (1995) Pain processing in the ventrocaudal nucleus of the human thalamus. In: B. Bromm and J. Desmedt (Eds.). Pain and the Brain: From Nociception to Cognition. Raven Press, New York, pp. 175-185. Lenz, F., Kwan, H.C., Dostrovsky. J.O. and Tasker, R.R. (1989) Characteristics of the bursting pattern of action potentials that occurs in the thalamus of patients with central pain. Brain Res., 496: 357-360. Lenz, EA., Lee, J.-I., Garonzik, I.-M., Rowland L.H., Dougherty, P.M. and Hua, S.E. (2000) Human thalamic reorganization related to nervous system injury and dystonia. In: J. Sandktihler, B. Bromm and G.F. Gebhart (Eds.), Nervous System Plasticity and Chronic Pain. Progress in Brain Research, Vol. 129. Elsevier, Amsterdam, pp. 259-273. Lorenz, J., Kohlhoff, H., Hansen, H.C., Kunze, K. and Bromm, B. (1998) Ab-fiber mediated activation of cingulate cortex as correlate of central post-stroke pain. Neuroreport, 9: 659-663. Manzoni, T., Barbaresi, P and Bemardi, S. (1990) Matching of receptive fields in the association projections from SI to SII of cats. J. Comp. Neural., 300: 331-345. Mauguiere, F., Merlet, I., Forss, N., Vanni, S., Jousmaki, V., Adeleine, P. and Hari, R. (1997) Activation of a distributed somatosensory cortical network in the human brain. A dipole modelling study of magnetic fields evoked by median nerve stimulation. Part I: Location and activation timing of SEF sources. Electroencephalogr: Clin. Neurophysiol., 104: 281289. Mima, T., Nagamine, T., Nakamura, K. and Shibasaki, H. (1998)

302

Attention modulates both primary and second somatosensory cortical activities in humans: a magnetoencephalographic study. J. Neurophysiol., 80: 2215-2221. Moore, K.A., Baba, H. and Woolf, C.J. (2000) Synaptic transmission and plasticity in the superficial dorsal horn. In: J. Sandktihler, B. Bromm and G.F. Gebhart (Eds.), Nervous System Plasticily and Chronic Pain. Progress in Brain Research, Vol. 129. Elsevier, Amsterdam, pp. 63-80. Musil, S.Y. and Olson, CR. (1993) The role of cat cingulate cortex in sensorimotor integration. In: B.A. Vogt and M. Gabriel (Eds.), Neurobiology of Cingulate Cortex and Limbic Thalamus: A Comprehensive Handbook. Birkhauser, Boston, MA, pp. 345-365. Oshiro, Y., Fufita, N., Tanaka, H., Hirabuki, N., Nakamura, H. and Yoshiya, I. (1998) Functional mapping of pain-related activation with echo-planar MRI: significance of the SII-insular region. Neutoreport, 9: 2285-2289. Petrovic, P., Petersson, K.M., Ghatan, P.H., Stone-Elander, S. and Ingvar, M. (2000) Pain-related cerebral activation is altered by a distracting cognitive task. Pain, 85: 19-30. Romani, G.L. and Narici, L. (1986) Principles and clinical validity of the biomagnetic method. Med. Prog. Technol., 11: 123-159. Rudd, P. and Blaschke, T.F. (1985) Antihypertensive agents and the drug therapy of hypertension. In: A. Goodman Gilman, L.S. Goodman, T.W. Rall and F. Murad (Eds.), The Pharmacological Basis of Therapeutics. 7th edn. MacMillan, New York, pp. 784-805. Sandkiihler, J., Benrath, J., Brechtel, C., Ruscheweyh, R. and Heir&e, B. (2000) Synaptic mechanisms of hyperalgesia. In: J. Sandktihler, B. Bromm and G.F. Gebhart (Eds.), Nervous System Plasticity and Chronic Pain. Progress in Brain Research, Vol. 129. Elsevier, Amsterdam, pp. 81-100. Scharein, E. and Bromm, B. (1998) The intracutaneous pain model in the assessment of analgesic efficacy. Pain Rev., 5: 216-246. Scherg, M. (1992) Functional imaging and localization of electromagnetic brain activity. Brain Topogr, 5: 103-I 11. Svensson, P., Johannsen, P., Jensen, T.S., Arendt-Nielsen, L., Nielsen, J., Stodkilde-Jorgensen, H., Gee, A.D., Baarsgaard, H.S. and Gjedde, A. (1998) Cerebral blood-flow changes evoked by two levels of painful heat stimulation: a positron emission tomography study in humans. Em .I. Pain, 2: 95107. Szentagothai, J. (1975) The ‘module-concept’ in cerebral cortex architecture. Brain Res., 95: 475-496. Talairach, J. and Touroux, P. (1988) Co-planar Stereofaxic Atlas of the Human Brain. Thieme, New York.

Towle, V.L., Bolanos, J., Suarez, D., Tan, K., Grzeszczuk, R., Levin, D.N., Cakmur, R., Frank, A. and Spire, J.P. (1993) The spatial location of EEG electrodes: locating the best-fitting sphere relative to cortical anatomy. Electroencephalogr Clin. Neurophysiol., 86: l-6. Treede, R.D., Lorenz, J., Kunze, K. and Bromm, B. (I 995) Assessment of nociceptive pathways with laser-evoked potentials in normal subjects and patients. In: B. Bromm and J. Desmedt (Eds.), Pain and the Brain: From Nociception to Cognition, Vol. 22. Raven, New York, pp. 377-392. Van Buren, J.M. (1983) Sensory response from stimulation of the inferior Rolandic and Sylvian regions in man. J. Neurosurg., 59: 119-130. Vogt, B. and Gabriel, M. (1993) Neurobiology of the Cingulate Cortex and Limbic Thalamus. : Birkhluser, Boston, MA. Walter, H., Kristeva, R., Knon; U., Schlaug, G., Huang, Y., Steinmetz, H., Nebeling, B., Herzog, H. and Seitz, R.J. (1992) Individual somatotopy of primary sensorimotor cortex revealed by intermodal matching of MEG, PET, and MRI. Brain Topegs, 5: 183-187. Wang, J.Z., Williamson, S.J. and Kaufmann, L. (1992) Magnetic source images determined by a lead-field analysis: tee unique minimum-norm least-squares estimation. IEEE Trans. Biomed. Eng., 39: 665-675. White, L.E., Andrews, T.J., Hulette, C., Richards, A., Groelle, M., Paydarfar, J. and Purves, D. (1997) Structure of the human sensorimotor system. I: Morphology and cytoarchitecture of the central sulcus. Cereb. Cortex, 7: 18-30. Whitsel, B.L., Petrucelli, L.M. and Werner, G. (1969) Symmetry and connectivity in the map of the body surface in somatosensory area II of primates. J. Neurophysiol., 32: 170-183. Williamson, S.J., Lu, Z.L., Karron, D. and Kaufman, L. (1991) Advantages and limitations of magnetic source imaging. Brain Topogr, 4: 169-180. Willis, W. (1985) The Pain System. The Neuronal Basis of Nociceptive Transmission in the Mammalian Nervous System. Karger, Basel. Woolsey, C.N., Erickson, T.C. and Gilson, W.E. (1979) Localization and somatic and motor areas of human cerebral cortex as determined by direct recording of evoked potentials and electrical stimulation. J. Neurosurg., 5 1: 476-506. Xu, X., Fukuyama, H., Yazawa, S., Mima, T., Hanakawa, T., Magata, Y., Kanda, M., Fujiwara, N., Shindo, K., Nagamine, T. and Shibasaki, H. (1997) Functional localization of pain perception in the human brain studied by PET. Neuroreport, 8: 555-559. Zilles, K. (1990) Cortex. In: G. Paxinos (Ed.), The Human Nervous System. Academic Press, London, pp. 757-802.

J. Sandkithler, B. Bromm and G.F. Gebhart (Eds.) Pmgress in Brain Research, Vol. 129 0 2000 Elsevier Science B.V. All rights reserved

CHAPTER 21

Regional brain oxygenation during phasic and tonic painful stimulation C. Forster *, R. Ringler and H.O. Handwerker Institutfiir

Physiologie

und experimentelk

Pathophysiologie, 91054 Erlangen,

Introduction In 1990 Ogawa et al. (Ogawa et al., 1990) described the principle of the BOLD effect (blood oxygenation level dependent contrast). It is now widely used for detecting regional changes in cerebral blood flow in functional magnetic resonance imaging (fMRI). Increased neuronal activity leads to vasodilatation in the affected brain region which is followed by an increase of the blood oxygenation level at the venous side. Due to the diamagnetic properties of oxygenated hemoglobin the MRI signal from the activated cortical areas increases slightly, and the increase can be used as a marker of activity changes. In the last few years the use of fMRI has developed rapidly and numerous papers have been published describing brain regions which are involved in the processing of various motor and sensory tasks. This technique has also been used for studying which brain regions are possibly involved in pain processing (see also Casey, 2000, this volume). In connection with fMRI acute, experimental pain was most often induced by electrical stimulation of peripheral nerves (Davis et al., 199.5, 1997) or by thermal stimulation (Berman et al., 1998; Apkarian et al., 1999; Becerra et al., 1999). The reports *Corresponding author: C. Forster, Institut fir Physiologie und experimentelle Pathophysiologie, Universittit Erlangen/Ntirnberg, Universititsstrasse 17, 91054 Erlangen, Germany. Tel.: +49-9131-8522492; Fax: +49-91318522497; E-mail: [email protected]

Universittit Germany

ErlangenjNiinzberg,

Universitiitsstrasse

17,

on brain areas activated in response to pain stimuli varied between different groups, probably because the different stimulation techniques produced pain perceptions and reactions varying in several dimensions, e.g. quality, intensity, spatial and temporal properties, and the affective component of pain (see also Bromm et al., 2000, this volume; Casey, 2000, this volume; Flor, 2000, this volume). Therefore, for a better understanding of the different brain regions involved in sensory processing of painful and nonpainful stimuli, different stimuli, which have some, but not all dimensions in common, should be applied in the same experiment. This can be achieved by using various forms of mechanical stimulation. A pneumatic device has been developed in our laboratory for delivering impact stimuli to pre-selected skin spots (Kohlloffel et al., 1991). By tuning of the impact velocity the quality can be changed from a non-painful to a painful tactile sensation, by changing the impact frequency burning or stinging pain can be induced. Another form of mechanical stimulation is tonic pressure stimuli which evoke deep dull pain sensations in the skin which increase with stimulus repetition. In this case the pain persists even after the end of the pressure. These stimuli can be used to determine brain areas which are similarly or differentially activated and thus are probably involved in the sensory discriminative aspects of perception. Increasing the intensity of the pain level should activate in addition cortical areas which are related to the affective and unpleasant components of pain.

304

Under pathophysiological conditions stimuli which are non-painful when applied to normal skin can induce distinct pain. An example will be shown of a patient suffering from acute Complex Regional Pain Syndrome (CRPS) in one hand, who perceived tactile stimuli as painful, and in whom there was activation of a brain region not normally involved in processing of tactile perceptions. Subjects Seven right-handed male volunteers (age: 22 to 55 years) took part in this study, all of them free from any neurological or systemic disorders. In addition, seven patients (six female, one male; age: 27 to 68) suffering from CRPS (diagnosed according to the IASP criteria; Stanton et al., 1995) were studied. Only three of the females, however, showed clear signs of hyperalgesia in the affected left arms. The study was approved by the local ethics committee, and all subjects gave their informed consent. All subjects were informed that they could stop the stimulation and imaging procedure at any time by pressing the alarm balloon. During the imaging procedure their heads were fixed and their ears closed with plugs to minimize interference from the noises of the MRI scanner. Each subject underwent a brief training session several days before the MRI measurement to become familiar with the experimental protocol and the type of stimuli. The individual pain thresholds for the phasic and tonic stimuli were determined; in the patient group this was done separately both for the affected and for the non-affected limb. Data acquisition Imaging was performed with a 1.5 tesla Magnetom Vision MRI scanner (Siemens, Erlangen, Germany). For each subject a global shimming procedure was performed to adjust the field homogeneity. Then a three-dimensional MPRAGE (magnetization prepared rapid gradient echo) data set of the individual brain was acquired which was later used to identify the affected brain regions. Functional T2*-weighted images were obtained using a multi-slice echo planar imaging technique (EPI) with the following parameters: 16 axial slices

parallel to a line from the top of the orbital cavity to the auditory meatus; repetition time TR = 114 ms; echo time TE = 62 ms; flip angle = 90”; scan time = 2.5 s per 16 slices; slice thickness = 5 mm; field of view = 220 x 220 mm; data matrix = 128 x 128 pixels which gave volume elements (voxels) with a resolution of 1.72 x 1.72 x 4 mm3; the interslice gap was 1 mm. High-resolution Tl-weighted anatomic images (TR = 600 ms; TE = 15 ms; flip = 90”; resolution = 0.86 x 0.86 x 5 mm3) with the same orientation as the EPI slices were collected from each subject. They were used to align the functional slices into the three-dimensional data set and to identify the anatomical structures. The slice orientation was chosen so as to get a maximum representation of the sensory cortices and limbic systems (Andersson et al., 1997). Experimental

protocol

In each experiment four sequences of measurements were performed with the stimulus conditions ‘finger tapping’, ‘impact pain’, ‘tonic pain’ and ‘touch’. ‘Finger tapping’ is an active motor task widely used in fMR1 studies. In this task the subjects are asked to tap the four finger tips in rapid succession to the tip of the thumb, which serves as a control for the experimental conditions and helps to determine the position of the motor cortex. In the CRPS patients each of the four stimulus paradigms were applied first to the non-affected hand and then to the affected one. All healthy volunteers were stimulated to the right hand. The minimum time between succeeding sequences was about 5 min during which neither stimulation nor imaging was done. During each test sequence 78 blocks of functional MRI data were obtained consisting of 16 axial slices and lasting 195 s. During each sequence four stimulation periods were performed, each of them lasting 15 s (six blocks of slices). Stimulation started at acquisition time 30 s (after block 12) 70 s (block 28) 120 s (block 48), and 150 s (block 60) (Fig. 1). Pain rating. Subjects were instructed to assessthe intensity of the pain perception on a scale from 0 (no pain or only touch sensation) to 10 (most intense pain imaginable). Ratings were given verbally at the end of each stimulus sequence.

305

et al., 1994; Kleinschmidt et al., 1995). All calculations were done with a software package developed by us and running under Windows NT. The first two image blocks of each sequence were discarded to eliminate transients that appear at the beginning of the EPI measurement. The time series of each voxel was correlated with a reference wave form representing the on-off pattern of the stimulus (Fig. 1). To consider possible delays between the onset of a stimulus and the change in the BOLD signal (Davis et al., 1998) the stimulus pattern was shifted stepwise from one to six blocks to the right. The maximum correlation coefficient Y of each voxel was used for the ongoing analysis. Detection of clusters Fig. 1. Method of fMR1. The middle trace shows an example of the fMRI signal of a selected region during the course of an experiment. This signal is correlated with the rest-activation sequence according to the experimental setup (lower trace). To detect the delay of the BOLD effect to the stimulus response the rest-activation sequence is shifted stepwise up to six times. The resulting maximum correlation map for each axial slice is the basis of the cluster algorithm.

Mechanical stimulation A device driven by pressurized air was employed to deliver mechanical impact stimuli to the dorsal side of the index or middle fingers. The system contains no ferromagnetic parts in order to be compatible to the MRI environment. This stimulus apparatus has been described in detail elsewhere (Kohlloffel et al., 1991). It consisted of a guiding barrel in which a plastic cylinder with a diameter of 5 mm was accelerated towards the skin. The velocity of the object defines the power of the impact and hence the strength of the stimulus. During a stimulation period of 15 s the impacts were delivered at a rate of l/s. Another stimulation device was used for the tonic pressure application: a plastic cylinder (diameter: 3 mm) was pinched with a constant force (2.5 N to 4.5 N) to the back of the second phalanx of the index or middle fingers. Data analysis The analysis was performed using standard methods which are based on a correlation technique (Friston

Since Y has to be calculated for a large number of voxels (up to 512 x 512) the probability increases of marking voxels as activated when they do not represent an activated area (false positive). Usually these faulty markers appeared at single voxels with no or only a small number of markers in the surrounding region, while real activation seemed to form clusters of voxels with high correlation. Based on this characteristic a cluster detection algorithm was constructed to discriminate between real activation and statistical noise: each cluster consisted of a starting voxel, which had a correlation r > i&h, and a surrounding region with voxels which had correlations r > rgg (Fig. 2). rr,@, and rg5 were chosen such that 0.5% of all voxels had r > rhgh while r > rgs was true for 5%. In addition, the voxels of the surrounding region had to be direct neighbors of the starting voxel or another voxel of the cluster. The cluster size had to pass the minimum size of k = 12 marked voxels. k was determined empirically: the size and shape of the clusters hardly changed if k was set between 10 and 15. Identification

of activated brain regions

The affected brain regions were identified using the individual 3-D data set. The Tl-weighted single slices, which had the same orientation as the functional images, were registered into the individual 3-D data set (Hastreiter et al., 1996). Briefly: at least four pairs of corresponding landmarks, which could

“T A

0

no stimulus 0.1

0.2 0.3 0.4 0.5 correlation coefficient

1 0.6

Fig. 2. Cluster algorithm. Left side: normalized histogram of the maximum correlation map for all activated voxels. The determined thresholds rhish and r95, 0.5% and 5%, respectively, of the most significant VOX&, are shown for a sequence of painful stimuli (dotted line) compared to an experiment without stimulation (solid line). Right side: illustration of the two-threshold-based cluster algorithm. A point s with T(S) 2 rhtsh is used as a starting point for the cluster algorithm. If the number of voxels found with r(R) 1 rss in the surrounding region R equals or is above the predelined number of voxels k necessary for a cluster then this cluster counts as an activated region.

be found in the Tl-data set as well as in the 3-D data set, were manually selected. These pairs were used to initiate a process which performed transformations and rotations of the Tl slices until their differences to the corresponding part of the 3-D data set were minimized. The same transformations were then used to fit the clusters into the three-dimensional data set. Slices of any orientation through the individual brain could now be produced to be compared with a printed brain atlas. Some measurements were done without the MPRAGE data set. In these cases a computerized brain atlas was used (Thurfjell et al., 1995). Each individual brain was fitted into the standard brain of this atlas by means of translation, rotation, linear and non-linear scaling. This was interactively repeated until markers of the atlas brain best fitted the corresponding structures in the brain under investigation. The quality of the approximation was visually controlled using the contours of the brain surface, succinct sulci, and the ventricular system. Statistics For the analysis of pain-related areas six brain regions were chosen in which most of the active clusters were found. These regions had been described before in fMR1, PET, and MEG studies employ-

ing noxious and motor stimulation (Talbot et al., 1991; Coghill et al., 1994; Bromm and Chen, 1995; Svensson et al., 1997; Tiille et al., 1999; Treede et al., 1999). (1) Primary motor cortex Ml: Brodmann area (BA) 4. (2) Other motor areas: BA 6, 8, 9. (3) Primary somatosensory cortex Sl: BA 1, 2, 3. (4) Secondary somatosensory cortex S2: supramarginal gyrus BA 40, 43. (5) Anterior cingulate cortex (ACC) and medial frontal cortex BA 24, 32. (6) Areas of the frontal lobe: BA 46, 50. The total number of activated voxels that were assigned to clusters was computed and the percentage of them in each region of interest was determined. Cortical areas activated by painful stimuli Impacts and tonic pressure, the two forms of painful stimuli, activated cortical areas which had been found in previous studies to be involved in pain processing, namely the somatosensory areas SI and SII, the insular region and the ACC (Bromm and Chen, 1995; Bushnell et al., 1999; Coghill et al., 1999; Treede et al., 1999; see also Bromm et al., 2000, this volume; Casey, 2000, this volume). In addition, we found activation of motor areas, MI and supplementary motor cortex, and of frontal areas. In many cases the activations were bilateral. Different patterns of activation were observed in different

307 subjects. Fig. 3 shows an example of activations in one healthy subject to whom phasic and tonic pain stimuli were applied to the right hand.

In this specimen record, phasic pain induced contralateral activation of SI and SII cortex, bilateral activation of the insular region and of the ACC. The tonic pain stimulus also induced some activation in the contralateral SI cortex, activated SII and the insular region bilaterally, and also the ACC contralaterally. Temporal dynamics of the BOLD signal The correlation method was used in this study to detect brain areas which were activated during the stimulations. Since this was done with a simple onoff pattern (see Fig. 1) the correlation coefficients in this case do not reflect a common variance of the time course of the stimulus and the real BOLD signal. In particular a delay between this search pattern and the BOLD signal is to be expected. The correlation coefficients of the starting points of the clusters found by this method were 0.737 =t 0.087 (mean f Std. Dev.) during the tapping tasks, 0.538 f 0.087 during touch, 0.63 f 0.076 during phasic pain, and 0.494f0.103 during tonic pain. The mean phase shift to reach the maximum r was two blocks during all stimulus conditions which means that the BOLD signal had a mean shift of 5 s as compared with the rest-stimulation pattern to get a maximum correlation. This should not be interpreted as the delay between stimulus onset and activation of a certain brain area; instead this mainly was an effect of the cross-correlation: the BOLD signal started to increase immediately at the beginning of a stimulation but it lasted several image blocks to reach its maximum. The differences of the phase

Fig. 3. Activation during phasic and tonic pain stimuli. Activated functional areas (Z-3) during painful stimulation found by the evaluation algorithm described above. 1 = cingulate; 2 = somatosensory cortex (Sl); 3 = S2 and insula region. Upper row: axial slices through the Sl region (left) and the S2 and insula region (right). Lower row: coronal slice showing insula and S2 area (left) and sagittal view of the pre- and post-central region including S2 area (right). (A) Phasic pain with activated regions in the cingulate and the Sl contralateral to the stimulated finger as well as a bilateral activation of the S2 and insula region. (B) Tonic pain with some activation in the cingulate and in the ipsilateral S2 and insula region.

308

shifts of the voxels within a single cluster was fl block and was probably caused by the sampling process (jitter). Differential tonic pain

activation patterns during phasic and

For quantitative comparisons the cluster areas in each region were calculated as percentage of total activation in each subject. These data are summarized in Fig. 4 from which it is clear that tonic and phasic pain led to differential activation patterns. Phasic pain activated mainly the somatosensory projection areas SI and SII in the contralateral hemisphere; however, activation was also observed in the ipsilateral somatosensory projection areas. A significant proportion of the total activation was found in the frontal cortex. In contrast, tonic pain stimuli

A

contralateral

medial

ipsilateral

20 % 15% 10% 5 %

induced only minor activation in the somatosensory projection areas. Those stimuli which were generally regarded to be more unpleasant than the phasic stimuli induced stronger activation in the ACC and frontal cortex. However, the pain ratings were similar after both types of stimuli (median rating of tonic stimulus: 7; phasic stimulus: 8). Activation pain

of cortical areas during pathological

The same battery of stimuli was applied to subjects suffering from Complex Regional Pain Syndrome (CRPS) in one hand. Three of these subjects were in the acute phase of the disease and showed strong hyperalgesia of the affected hand. The painful tonic and phasic stimuli were adjusted in these subjects according to the individual pain thresholds. That means that stimuli applied to the affected and to the contralateral side were of different intensities. Under these conditions activation in SI, SII and insular region was not remarkably different from that observed in healthy subjects. However, a patient suffering from strong mechanical hyperalgesia who felt pain even from light touch stimuli (allodynia) showed pronounced activation of the ACC, an area which was never activated in healthy subjects during touch stimulation (Fig. 5). Conclusions

B

20%{-

Fig. 4. Distribution of clusters activated during painful stimulation. Percent of activation in different areas proportional to the total number of activated voxels. The main functional anatomical areas that were activated during phasic pain (A) and tonic pain (B) are shown. Ipsilateral, medial, and contralareral indicate the hemispheres in view of the stimulation side.

Our results show that painful stimuli activate various brain areas in healthy subjects, as described earlier (Coghill et al., 1994, 1999; Bushnell et al., 1999; Treede et al., 1999). Not only the known somatosensory areas were involved, but also areas related to motor activity and more integrative areas. It has been discussed that this multiple activation pattern is a strong indication of parallel processing of nociceptive input to the cerebral cortex (Porro et al., 1998; Gelnar et al., 1999). Besides the lateral somatosensory system which receives input from the contralateral body side, ipsilateral cortical areas are involved, in particular SII and the insular cortex, the frontal cortex and midline structures, e.g. the ACC. Phasic and tonic pain stimuli activate these regions differentially. The lateral projection system, and in particular the contralateral SI cortex, is more

Fig. 5. Activation due to allodynia in a patient with CRPS: Activated areas are found in the cingulate (I), the contralateral

a touch stimulus on the affected right hand leads to a painful somatosensory cortex S 1 (2) and S2/insula region (3).

strongly activated by phasic than by tonic pain stimuli though the latter are usually regarded as more painful by the subjects. This system seems to be mainly concerned with stimulus transients. In contrast, the tonic pain stimuli used in this study were activating the ACC and frontal cortical areas more efficiently than the somatosensory areas. This is in agreement with the notion that the ACC is important for the unpleasant affective component of pain (Davis et al., 1997; Tolle et al., 1999). This conclusion is also supported by our findings on a patient suffering from allodynia within the context of a CRPS. Upon touch stimulation this patient did not show a striking activation pattern in the lateral projection system, but a strong activation of ACC (Fig. 5). We conclude from these results that the processing of nociceptive stimuli is not dependent on a single brain area. It is apparently bilateral, in particular in the somatosensory areas SI and SII, in the insular region, in midline structures and in the frontal cortices. It is well conceivable that perception of the various dimensions of painful stimuli depends on activation of all these areas. This notion is supported by findings from patients in whom parts of the cerebral cortex have been removed. Discrete injuries of the contralateral SI cortex or removal of the cingulate cortex did not abolish the ability of the patients to perceive painful stimuli (Davis et al., 1994; Kuroda et al., 1995; Fuchs et al., 1996; Pastoriza et al., 1996; Peyron et al., 2000). The nociceptive input to these

sensation.

areas is probably transmitted via several pathways in the CNS. This is supported by electrophysiological findings (Willis-WD, 1985; Kenshalo-DR et al., 1988; Kalliomaki et al., 1993) and clinical observations (Berthier et al., 1988; Greenspan and Winfield, 1992; Knecht et al., 1996) which showed that parallel transmission seems to be a major principle in the processing of painful stimuli. In summary there is obviously no ‘master region’ for pain processing, a fact which ensures the ability of the organism to detect tissue injury even if parts of the CNS fail. Acknowledgements This work was supported by the Deutsche Forschungsgemeinschaft (DFG), SFB 353. References Andersson, J.L., Lilja, A., Hartvig, P., Langstrom, B., Gordh, T., Handwerker, H.O. and Torebjork, E. (1997) Somatotopic organization along the central sulcus, for pain localization in humans, as revealed by positron emission tomography. Exp. Bruin Rex, 117: 192-199. Apkarian, A.V., Darbar, A., Krauss, B.R., Gelnar, PA. and Szeverenyi, N.M. (1999) Differentiating cortical areas related to pain perception from stimulus identification: temporal analysis of fMRI activity. J. Neurophysiol., 8 1: 2956-2963. Becerra, L.R., Breiter, H.C., Stojanovic, M., Fishman, S., Edwards, A., Comite, A.R., Gonzalez, R.G. and Borsook, D. (1999) Human brain activation under controlled thermal stim-

310 ulation and habituation to noxious heat: an fMRI study. Magn. Reson. Med., 41: 1044-1057. Berman, H.H., Kim, K.H.S., Talati, A. and Hirsch, .I. (1998) Representation of nociceptive stimuli in primary sensory cortex. Neuroreport, 9: 4179-4187. Berthier, M., Starkstein, S. and Leiguarda, R. (1988) Asymbolia for pain: a sensory-limbic disconnection syndrome. Ann. Neurol., 24: 41-49. Bromm, B. and Chen, AC. (1995) Brain electrical source analysis of laser evoked potentials in response to painful trigemina.l nerve stimulation. Electroencephalogr Clin. Neurophysiol., 95: 14-26. Bromm, B., Scharein, E. and Vahle-Hinz, C. (2000) Cortex areas involved in the processing of normal and altered pain. In: .I. Sandktlhler, B. Bromm and GE Gebhart (Eds.), Nervous System Plasticity and Chronic Pain. Progress in Brain Research, Vol. 129. Elsevier, Amsterdam, pp. 289-302. Bushnell, M.C., Duncan, G.H., Hofbauer, R.K., Ha, B., Chen, J.I. and Carrier, B. (1999) Pain perception: is there a role for primary somatosensory cortex?. Proc. Natl. Acad. Sci. USA, 96: 7705-7709. Casey, K.L. (2000) Concepts of pain mechanisms: the contribution of functional imaging of the human brain. In: J. Sandkuhler, B. Bromm and GE Gebhart (Eds.), Nervous System Plasticity and Chronic Pain. Progress in Brain Research, Vol. 129. Elsevier, Amsterdam, pp. 277-287. Coghill, R.C., Talbot, J.D., Evans, AC., Meyer, E., Gjedde, A., Bushnell, M.C. and Duncan, G.H. (1994) Distributed processing of pain and vibration by the human brain. J. Neurosci., 14: 4095-4108. Coghill, R.C., Sang, C.N., Maisog, J.H. and Iadarola, M.J. (1999) Pain intensity processing within the human brain: a bilateral, distributed mechanism. J. Neurophysiol., 82: 1934-1943. Davis, K.D., Hutchison, W.D., Lozano, A.M. and Dostrovsky, J.O. (1994) Altered pain and temperature perception following cingulotomy and capsulotomy in a patient with schizoaffective disorder. Pain, 59: 189-199. Davis, K.D., Wood, M.L., Crawley, A.P. and Mikulis, D.J. (1995) fMRI of human somatosensory and cingulate cortex during painful electrical nerve stimulation. Neuroreport, 7: 321-325. Davis, K.D., Taylor, S.J., Crawley, A.P., Wood, M.L. and Mikulis, D.J. (1997) Functional MRI of pain- and attentionrelated activations in the human cingulate cortex. L Neurophysiol., 77: 3370-3380. Davis, K.D., Kwan, C.L., Crawley, A.P. and Mikulis, D.J. (1998) Event-related fMRI of pain: entering a new era in imaging pain. Neuroreport, 9: 3019-3023. Flor, H. (2000) The functional organization of the brain in chronic pain. In: J. Sandktihler, B. Bromm and G.F. Gebhart (Eds.), Nervous System Plasticity and Chronic Pain. Progress in Brain Research, Vol. 129. Elsevier, Amsterdam, pp. 313322. Friston, K.J., Jezzard, P. and Turner, R. (1994) Analysis of functional MRI time-series. Hum. Brain Mapp., 1: 153-l 7 1. Fuchs, P.N., Balinsky, M. and Melzack, R. (1996) Electrical stimulation of the cingulum bundle and surrounding cortical

tissue reduces formalin-test pain in the rat. Brain Rex, 7432: 116-123. Gelnar, P.A., Krauss, B.R., Sheehe, P.R., Szeverenyi, N.M. and Apkarian, A.V. (1999) A comparative fMRI study of cortical representations for thermal painful, vibrotactile, and motor performance tasks. Neuroinmge, 10: 460-482. Greenspan, J.D. and Winfield, J.A. (1992) Reversible pain and tactile deficits associated with a cerebral tumor compressing the posterior insula and parietal operculum. Pain, 50: 29-39. Hastreiter, P., Hopfer, W. and Ertl, Th. (1996) Semi-automatic registration of SD-multi-modality brain images based on an information theoretic approach. In: Workshop on Digital Image Processing in Medicine, 3, pp. 132-137. Kalliomaki, J., Weng, H.R., Nilsson, H.J., Yu, Y.B. and Schouenborg, J. (1993) Multiple spinal pathways mediate cutaneous nociceptive C fibre input to the primary somatosensory cortex (Sl) in the rat. Brain Res., 6222: 271-279. Kenshalo-DR, J., Chudler, E.H., Anton, F. and Dubner, R. (1988) SI nociceptive neurons participate in the encoding process by which monkeys perceive the intensity of noxious thermal stimulation. Brain Res., 4542: 378-382. Kleinschmidt, A., Requardt, M., Merboldt, K.D. and Frahm, J. (1995) On the use of temporal correlation coefficients for magnetic resonance mapping of functional brain activation: individualized thresholds and spatial response delineation, Int. J. Imug. Syst. Tech., 63: 238-244. Knecht, S., Kunesch, E. and Schnitzler, A. (1996) Parallel and serial processing of haptic information in man: effects of parieta1 lesions on sensorimotor hand function. Neuropsycholagia, 34: 669-687. Kohlloffel, L.U., Koltzenburg, M. and Handwerker, H.O. (1991) A novel technique for the evaluation of mechanical pain and hyperalgesia. Pain, 46: 81-87. Kuroda, R., Yorimae, A., Yamada, Y., Furuta, Y. and Kim, A. (1995) Frontal cingulotomy reconsidered from a WGA-HRP and c-Fos study in cat. Acta Neurochir Suppl. (Wien), 64: 69-73. Ogawa, S., Lee. T.M., Kay, A.R. and Tank, D.W. (1990) Brain magnetic resonance imaging with contrast dependent on blood oxygenation. Proc. Natl. Acad. Sci. USA, 87: 9868-9872. Pastoriza, L.N., Morrow, T.J. and Casey, K.L. (1996) Medial frontal cortex lesions selectively attenuate the hot plate response: possible nocifensive apraxia in the rat. Pain, 64: I I17. Peyron, R., Garcia-Larrea, L., Gregoire, M.C., Convers, P., Richard, A., Lavenne, E, Barral, F.G., Mauguiere, F., Michel, D. and Laurent, B. (2000) Parietal and cingulate processes in central pain. A combined positron emission tomography (PET) and functional magnetic resonance imaging (fMR1) study of an unusual case. Pam, 84: 77-87. Porro, CA., Cettolo, V., Francescato, M.P. and Baraldi, P. (1998) Temporal and intensity coding of pain in human cortex. .I. Neurophysiol., 80: 3312-3320. Stanton, H.M., Janig, W., Hassenbusch, S., Haddox, J.D., Boas, R. and Wilson, P. (1995) Reflex sympathetic dystrophy: changing concepts and taxonomy. Pam, 63: 127-133. Svensson, P, Minoshima, S., Beydoun, A., Morrow, T.J. and

311 Casey, K.L. (1997) Cerebral processing of acute skin and muscle pain in humans. J. NeurophysioZ., 78: 450-460. Talbot, J.D., Marrett, S., Evans, AC., Meyer, E., Bushnell, M.C. and Duncan, G.H. (1991) Multiple representations of pain in human cerebral cortex. Science, 251: 1355-1358. Thurfjell, L., Bohm, C. and Bengtsson, E. (1995) CBA - an atlas based software tool used to facilitate the interpretation of neuroimaging data. Comp. Meth. Prog. Biomed., 47: 51-71. ToBe, T.R., Kaufmann, T., Siessmeier, T., Lautenbacher, S.,

Berthele, A., Munz, E, Zieglgansberger, W., Willoch, F., Schwaiger, M., Conrad, B. and Bartenstein, P. (1999) Region-specific encoding of sensory and affective components of pain in the human brain: a positron emission tomography correlation analysis. Ann. Neural., 45: 40-47. Treede, R.D., Kenshalo, D.R., Gracely, R.H. and Jones, A.K.P. (1999) The cortical representation of pain. Pain, 79: 105- 111. Willis-WD, J. (1985) Pain pathways in the primate. Prog. Clin. Biol. Rex, 176: 117-133.

J. Sandkiihler, B. Bromm and G.F. Gebhart (Eds.) Progress in Brain Research, Vol. 129 0 2000 Elsevier Science B.V. All rights reserved

CHAPTER

22

The functional organization of the brain in chronic pain Herta Flor * Central

Institute

of Mental

Health

JS, Neuropsychology

Introduction The functional organization of the brain in response to acute painful stimulation has been extensively described (for a summary see Bromm and Desmedt, 1995). Much less research has focused on alterations in brain responses related to chronic states of pain. In the past five years we have used various imaging methods to examine how the brain changes in response to chronic pain and how alterations in the brain might contribute to the experience of chronic pain. In this chapter, we will first focus on cortical reorganization and chronic pain, will then discuss the role of pain memories in primary and associative cortex, will summarize the findings on cortical reorganization and phantom limb pain, and will end the chapter with a discussion of treatment implications of the findings reported here. Cortical reorganization

and chronic pain

Cortical reorganization describes functional and structural changes in the representational maps of the sensory and motor areas of adult mammals related to an altered efficacy of sensory inputs, to changes in the availability of effecters, or due to injury (Kaas, 1991). Changes in the somatotopic organization of primary somatosensory cortex have previously been shown subsequent to deafferentation or related to

* Corresponding author: H. Flor, Central Institute of Mental Health J.5,Neuropsychology and Clinical Psychology Unit, 68342 Mannheim, Germany. E-mail: [email protected]

and Clinical

Psychology

Unit, 68342

Mannheim,

Germany

behaviorally relevant input (Merzenich et al., 1984; Pons et al., 1991; Recanzone et al., 1992). Chronic pain may be viewed as a source of highly relevant nociceptive input that might then lead to an altered representation of the body part affected by pain and to subsequent increased pain perception (cf. Birbaumer et al., 1995; Ramachandran and Hirstein, 1998; Flor and Turk, 2000). Chronic pain patients often show hyperalgesia and allodynia; for example, perception and pain thresholds as well as pain tolerance levels were found to be significantly lower in patients with chronic back pain or fibromyalgia (Lorenz et al., 1996; Kleinbijhl et al., 1999; Flor et al., 2000). Although peripheral (see Cummins et al., 2000, this volume; Sutherland et al., 2000, this volume; Reeh and Pethb, 2000, this volume) as well as spinal (see Moore et al., 2000, this volume; Sandkiihler et al., 2000, this volume; Svendsen et al., 2000, this volume) and thalamic mechanisms (see Dostrovsky, 2000, this volume; Lenz et al., 2000, this volume) have been implicated in some of these changes in nociception (cf. Guilbaud et al., 1993; Woolf and Doubell, 1994), cortical changes might also play a role in these alterations in nociceptive sensitivity (cf. Lorenz et al., 1998; see Bromm et al., 2000, this volume; Casey, 2000, this volume). To further test this hypothesis, intracutaneous electric stimulation was applied in chronic back pain patients, subchronic back pain patients, and in healthy controls (Flor et al., 1997a). Stimulation sites were the back - the site of pain - and the index finger - a site unrelated to the chronic pain problem. Evoked magnetic fields in the range of 40500 ms following the stimulus were assessed and brain source analysis was used to detect the localiza-

chronic

back pain

control

(J back A finger group back B finger

Fig. 1. Localization of the digits and the back in back pain patients and healthy controls in primary somatosensory cortex. Stimulation was on the left side of the body, the representations are on the hemisphere contralateral to the stimulation side. Please note the shift of the back representation of the back pain patients into a more medial position (i.e. towards the leg representation). The shift amounted to about 2-3 cm. Based on Flor et al. (1997a).

tion of the stimuli. In the time window before 100 ms, a range that has been shown to reflect activation of primary somatosensory cortex (SI) (Elbert et al., 1995), we found that the chronic back pain patients showed an elevated response (global field power) that was specific for the site of pain, whereas the finger stimulation did not yield higher brain responses. In a later time window, 200-300 ms, a non-specific increase of the activity related to stimulation of both sites - the finger and the back - occurred. This late non-specific increase in activation in pain patients has also been shown for stimuli that are not specific to the somatosensory system (cf. Flor et al., 1997a; Karl et al., 1999) and might indicate general disinhibition of the cortex in states of chronic pain. Dipole modeling of the early component of the evoked magnetic field around 70 ms showed that

a shift of the back representation in these patients towards the leg area had occurred - which may be interpreted as an expansion when it is taken together with the high root mean square amplitudes observed in those patients (see Fig. 1). These data suggest that chronic pain leads to an expansion of the cortical representation zone related to nociceptive input, much like the expansions of cortical representations that have been documented to occur with other types of behaviorally relevant stimulation. Nociceptive input is of high relevance for the organism and it might be useful to enhance the representation of this type of stimulation to prepare the organism for the adequate response. The amount of expansion of the back region was positively correlated with chronicity (r > 0.80) suggesting that pain-related cortical reorganization develops

315 over time. This type of cortical alteration may correspond to what Katz and Melzack (1990) have called a somatosensory pain memory. We differentiate the terms ‘explicit memory’ which refers to memories that are verbally coded and accessible to consciousness, such as autobiographical or semantic memory, and ‘implicit memories’ such as skill learning or some types of classical and operant conditioning which are usually not verbally reproducible and not conscious. Somatosensory pain memories may reflect a sensitization process that is implicit - it may lead to behavioral and perceptual changes, such as hyperalgesia and allodynia, in the absence of conscious awareness. It may therefore be impossible for the patient to counteract these pain memories that may lead to pain perception in the absence of peripheral stimulation since. Such an expansion of a representational zone is related to higher acuity in the perception of tactile input. Further research on brain processes related to chronic pain has focussed on the impact of specific types of implicit memories such as operant and respondent conditioning on pain perception. Learning and chronic pain Fordyce (1976) was one of the first researchers to point out the important role of operant conditioning and reinforcement learning for the development of chronic pain. He postulated that positive reinforcement for observable pain behaviors such as moaning and limping, but also negative reinforcement of pain behaviors such as inactivity and the intake of medication as well as the lack of reinforcement for well-behaviors such as being active or working, all contribute to the transition from an acute to a chronic pain problem, which may over time primarily be maintained by reinforcement contingencies. We examined to what extent operant learning also influences brain responses related to the processing of nociceptive input. In the first study, we examined to what extent the presence of a spouse that habitually reinforces pain can also influence painrelated cortical responses. We asked chronic back pain patients and their spouses to come to our lab and divided the spouses into two groups based on the West Haven-Yale Multidimensional Pain Inventory (Kerns et al., 1985). One group habitually ignored

the pain or punished their partners for expressing pain; these spouses were termed the non-solicitous group. The other group habitually reinforced pain behaviors, here termed the solicitous group. In the experiment we simply asked the spouses to sit in the laboratory and watch the patient while he or she received electric stimulation at either the site of pain (the back) or a site unrelated to the pain (the finger). We recorded the pain-related cortical response from 92 electrodes spaced evenly over the cortical surface. Subsequently we computed the global field power for the time window between 250 and 500 ms following the stimulus and subtracted the values for the presence as compared to the absence of the spouse. The difference in global field power served as the dependent measure of the brain response to pain in the two conditions. Stimulating the finger tip we found no difference in global field power whether the spouse was present or absent, neither between the pain patient groups or the patient groups and the healthy controls (see Fig. 2). A completely different picture emerged, however, for the back stimulation. Here, the difference was largest for the solicitous spouses, followed by the non-solicitous spouses and the healthy controls. Brain activation in response to painful stimuli was thus highly dependent on the presence of a reinforcing spouse - the more pain-reinforcing the spouse was, the higher was the brain response to pain, suggesting that the spouse served as a discriminative cue for the enhanced processing of the nociceptive stimulus. The main difference between these conditions was observed in the left frontal cortex, a region where emotional influences are known to impact on the sensory-discriminative aspects of pain. Direct verbal reinforcement of pain has been identified as an additional important modulator of the pain response. We trained chronic pain patients and healthy controls to increase or decrease their pain ratings of electric stimuli by reinforcing them verbally and by smileys on a computer screen if they changed their ratings in the desired direction. Concurrently we assessed somatosensory evoked responses to these painful stimuli from 64 electrodes. We observed equal learning of the reinforced response in both the chronic back pain patients and the healthy controls but a lack of extinction when the desired direction of the pain ratings was no longer

316

Global Field Power

3

spouse

7

2.5 -1 Interaction:

F(1,27)

presence/absence

= 5.73, p = .024

p = ,036

0,5 T

+ - healthy

-

chronic

-__-._____-. Finger

~.

>

Back

Fig. 2. Depicted is the difference in global field power between presence and absence of the spouse finger and the back for chronic back pain patients and healthy controls. Note the substantial increase the presence of the solicitous spouse during back stimulation.

reinforced in the chronic pain patients. When we examined the brain responses to these stimuli we found that the late responses (>200 ms) were unaltered and showed mainly habituation. However, the early response (N150) was affected by the conditioning and remained high in the chronic pain group that had been reinforced for higher pain ratings, thus indicating a direct effect of verbal reinforcement on the early cortical processing of nociceptive information. This lack of extinction in the cortical domain suggests that learning processes related to verbal and behavioral conditioning may exert long-lasting influences on the cortical response to pain-related stimuli and form implicit pain memories. Further evidence for these pain-related memories came from studies that used pain-related words such as aching and burning and compared them to bodyrelated words such as sweating and breathing and neutral words such as walking or moving. When evoked responses to these words were examined in chronic back pain patients or subchronic pain patients, the patients showed enhanced early evoked responses (NlOO) to the pain-related words, indicative of a classical or Pavlovian conditioning process that has transferred special meaning to these words (see Flor et al., 1997b; Knost et al., 1997). A direct classi-

for the electric in EEG activity

stimulation at the of the patients in

cal conditioning experiment that paired pseudowords (i.e. words the subjects had never heard and could not attach any prior meaning to) with electric shock yielded exactly the same results - i.e. subjects acquired an elevated NlOO response to the words that had been paired with shock with a preponderance of the response over the left (language-related) hemisphere (see Montoya et al., 1996). Later components of the evoked response such as the P300 and the Late Positive Complex were elevated in the chronic pain patients, but non-specific with respect to word class, reflecting the generally elevated level of excitation related to chronic pain mentioned before. Chronic pain states seem to lead to the development of somatosensory pain memories that manifest themselves in alterations in the somatotopic map in somatosensory cortex as well as altered processing in associative areas and may contribute to hyperalgesic states in the absence of peripheral nociceptive stimulation. These pain memories can be influenced by psychological processes such as operant and classical conditioning or attention (see Buchner et al., 1999) that may establish additional and potentially more widespread implicit memories and may enhance existing memories. In addition to local representational changes, chronic states of pain are as-

317

sociated with increased cortical excitation that may significantly contribute to cortical reorganization. Cortical reorganization

and phantom limb pain

Cortical reorganization also seems to play an important role in another type of pain that has remained an enigma - that is, phantom limb pain. Katz and Melzack (1990) have suggested that somatosensory pain memories are important in the understanding of this type of pain based on the observation that phantom limb pain patients frequently experience the type of pain they had reported before amputation. We therefore wondered if somatosensory pain memories would be present in human amputees and if they would be related to phantom limb pain. When we examined the topographic map in SI in upper limb amputees we observed a shift of the mouth

representation into the area of the former hand representation only in patients with phantom limb pain (Flor et al., 1995; Grtisser et al., 2000; see Fig. 3). Interestingly, the standard deviations of this shift are very small, indicating that this reorganizational change occurs in all patients with pain but not in patients with non-painful phantoms. Non-painful phantoms seem to be related more to activation in posterior parietal and second somatosensory cortex (see Flor et al., 2000). We found the exact same changes in motor cortex using functional magnetic resonance imaging or transcranial magnetic stimulation or motor cortical potentials again only in patients with phantom limb pain (cf. Lotze et al., 1999; Karl et al., 2000).

How are these changes related to preamputation pain? When we assess pain not only immediately prior to the amputation (which is not highly related

Fig. 3. Localization of the mouth and the digits in primary somatosensory cortex of unilateral upper extren city amputees. Note that the representation of the mouth on the amputation side C= hemisphere contralateral to the amputated limb) has !shifted into the hand reg ion. The arrow denotes the magnitude of the shift which amounted to about 1-2 cm.

318

to cortical reorganization and phantom limb pain) but go back in time and ask the patients to recall the duration and intensity of preamputation pain, we find a high correlation (r > 0.50) with both phantom limb pain and cortical reorganization. These data could explain the negative findings of Nikolajsen et al. (1997) or related studies that tested the role of preemptive analgesia in the prevention of phantom limb pain (see also Jensen and Nikolajsen, 2000, this volume; Wilder-Smith, 2000, this volume). If the patient sample contained a large number of persons that had chronic pain prior to the amputation, then a somatosensory pain memory might have formed that might have led to intracortical changes that are not reversible by blocking afferent input from the periphery or general anesthesia. There is some evidence that intracortical processes might play an important role in the cortical reorganizational processes observed in phantom limb pain. For example, Florence et al. (1998) recently showed that axonal sprouting in primary somatosensory cortex but not in the thalamus occurs in monkeys with longstanding hand amputations. Several studies examined the effects of correlated inputs or deafferentation, showing that the main changes occurred in those cortical layers that have intracortical connections and not in the layers that have thalamocortical input (cf. Diamond et al., 1994; Wang et al., 1995). Birbaumer et al. (1997) showed that about half of the phantom limb pain patients had pain reduction and a reduction of cortical reorganization due to brachial plexus anesthesia, suggesting that in some patients rapid changes in the balance of inhibitory and facilitatory processes may lead to the observed cortical changes, whereas in other patients neither pain reduction or cortical reorganization were observed, suggesting that firm new cortical connections may have developed. It is likely that inputs from the face now code both in the face area and in the hand area that was formerly occupied by inputs from the now amputated arm (cf. Dykes, 1997). How do these changes relate to the somatosensory pain memories we have discussed at the beginning? Let us assume that somatosensory pain memories occupy specific pain-coding zones in SI cortex. The work of Kenshalo (Kenshalo and Douglass, 1995; cf. Treede et al., 1999) suggests that there are regions in SI that are specific for pain that do not overlap with

regions coding for non-painful stimuli - detailed data on this in humans or primates are missing - however, we know about the role of SI cortex from PET and fMRI studies (see Casey et al., 1994; Gelnar et al., 1999; Bromm et al., 2000, this volume; Casey, 2000, this volume). These zones may expand as a result of chronic pain. If later a deafferentation occurs, the vacated space is substantially enlarged and we also find enhanced cortical excitability. Thus there is an increased likelihood for neighboring input to enter the vacated space and to specifically occupy the zones devoted to painful stimuli as well as for unmasking of normally inhibited connections. We may assume that the assignment of cortical representation areas to peripheral locations is based on early learning (we find no such shifts in congenital amputees; cf. Flor et al., 1998) and that it remains fairly constant after deafferentation. This would mean that the zones that originally subserved the now amputated hand or arm and that were activated by painful stimulation would still code for pain and they would code for pain in the region that originally activated them (cf. Dykes, 1997). Thus, input from neighboring territory would lead to a high likelihood of phantom pain rather than non-painful phantom sensation to occur. These hypotheses were entered into a model on the development of phantom limb pain that suggests that prior pain memories are an important precondition for phantom limb pain to occur although they are not the only factors (see Fig. 4). The work of Jgnig, Devor, and Sherman (cf. Michaelis et al., 1996; Devor, 1997; Sherman, 1997) has demonstrated that peripheral factors such as ectopic discharge or psychophysiological activation are also important in maintaining phantom limb pain. For example, Calford and Tweedale (1991) have shown that the loss of C-fiber input leads to an expansion of receptive fields in SI due to a loss of inhibition that is mediated by C-fibers. Thus, the selective loss of C-fibers that has been observed in peripheral deafferentation might lead to disinhibition and unmasking and further cortical reorganization. The selective input of A/3- and potentially A&fibers from the amputation zone might lead to an alteration of the nature of this input, which might in some way activate nociceptive neurons and thus increase painful input from the phantom. We have found that

319

.

Selective loss of C fibers

.

Random input from stump neuroma

.

Abnormal

.

Svmnathetic

changes in the DRG and dorsal horn activation

Fig. 4. A model of the development of phantom limb pain.

stimulation on the stump, both with slight touch and with pin prick, will lead to an increase in phantom sensation, especially in painful sensations. The elicitation of painful but not of non-painful phantoms is related to reorganization in SI cortex, suggesting that the neural network underlying painful phantoms has a focus in SI. We also observed substantial pain-related reorganization when we selectively stimulated A& rather than Ap-fibers using infrared laser pulses (Treede et al., 1995). Non-painful phantoms can be elicited from the contralateral side or from the foot, suggesting that non-painful phantoms are more related to changes on the thalamic or SII level where the hand and the foot region are adjacent rather than only the SI region as well as the posterior parietal cortex that is also more active in non-telescoped phantoms (cf. Flor et al., 1998). Microneurographic recordings from nerves supplying the former hand region have shown that considerable spontaneous activity is present in these nerves that seems to be of a random nature (cf. Nystrijm and Hagbarth, 1981), suggesting that this spontaneous activity might be an additional source

of activation of cortical reorganization, since random input seems to increase shifts in the cortical map (cf. Spitzer, 1997). Furthermore, changes in skin conductance level, muscle tension levels, and skin temperature as well as blood flow in the amputation stump, which may be related to sympathetic activation and motor activations and also be influenced by psychological processes, have been known to influence phantom limb pain as well as phantom limb sensations (Katz, 1992; Sherman, 1997). These peripheral changes may either activate peripheral nociceptors or sensitize them or they may act via their activation of spinal and supraspinal mechanisms. Somatosensory pain memories represented by alterations in the topographic map of SI cortex may thus underlie the development of phantom limb pain. Longstanding states of chronic pain prior to the amputation may be instrumental in the formation of these pain memories by inducing representational and excitability changes. Deafferentation does not alter the original assignment of cortical representation zones to peripheral input zones and lead to double coding. Peripheral factors such as loss of C-fiber activity, spontaneous activity from neuroma, or psy-

320

chophysiological activations influence the cortical representational changes. Learning processes are instrumental in the development and maintenance of these cortical changes. Treatment implications There are important treatment implications based on these findings. The prevention of somatosensory pain memories is important and might be accomplished by (a) the prevention of chronic pain and thus cortical reorganization by pharmacological and psychological interventions in order to make sure that pain memories are not being established, and (b) cortical reorganization related to amputation could further be prevented by giving NMDA antagonists and/or GABA agonists prior to and after the surgery. We are currently conducting a study that tests this intervention in conjunction with peridural anesthesia which is important in inhibiting perisurgical pain memories. In chronic pain states these cortical pain memories must be extinguished by pharmacological and/or behavioral interventions. We have recently found that extensive input to the cortical amputation zone related to intensive use of a myoelectric prosthesis reduces both cortical reorganization and phantom limb pain (Lotze et al., 1999). Thus extensive training with a myoelectric or sensorimotor prosthesis would seem to be useful. Based on the work of Jenkins et al. (1990) and Recanzone et al. (1992) we have devised a sensory training for patients who cannot be fitted with a prosthesis. Electrodes are closely spaced over the amputation stump in a region where their activity excites the nerve that supplied the amputated arm portion. Patients then have to discriminate the frequency and the location of the stimulation in an extended training that encompasses two hours per day over a two-week period. We found substantial improvements in both 2-point discrimination and phantom limb pain in the trained patients. These improvements were accompanied by changes in cortical reorganization indicating a normalization of the shifted mouth representation. These findings are in line with other evidence suggesting that behavioral training can have massive effects on cortical representations. Direct modification of cortical activity by a biofeedback application might also be of value.

To summarize it may be stated that psychological factors have a direct influence on physiological processes in nociception and that behavioral interventions may be powerful methods of influencing maladaptations of the organism such as those seen in pain-related reorganizational changes. Acknowledgements The preparation of this manuscript was facilitated by grants Fl 156/16 and Bi 195/24 from the Deutsche Forschungsgemeinschaft. References Birbaumer, N., Flor, H., Lutzenberger, W. and Elbert, T. (I 995) The corticalization of chronic pain. In: B. Bromm and J.E. Desmedt (Eds.), Pain and the Brain: From Nociception to Cognition. Raven Press, New York, pp. 33 l-344. Birbaumer, N., Lutzenberger, W., Montoya, P., Larbig, W., Unertl, K., Topfner, S., Grodd, W., Taub, E. and Flor, H. (1997) Effects of regional anesthesia on phantom limb pain are mirrored in changes in cortical reorganization, J. Neurosci., 17: 5503-5508. Bromm, B. and Desmedt, J. (1995) Pain and the Brain: From Nociception to Cognition. Raven Press, New York. Bromm, B., Scharein, E. and Vahle-Hinz, C. (2000) Cortex areas involved in the process of normal and altered pain. In: J. Sandkiihler, B. Bromm and G.F. Gebhart (Eds.), Nentous System Plasticity and Chronic Pain. Progress in Brain Research, Vol. 129. Elsevier, Amsterdam, pp. 289-302. Buchner, H., Reinartz, U., Waberski, T.D., Gobbele, R., Noppeney, U. and Scherg, M. (1999) Sustained attention modulates the immediate effect of de-afferentation on the cortical representation of the digits: source localization of somatosensory evoked potentials in humans. Neurosci. Left., 260: 5760. Calford, M.B. and Tweedale, R. (1991) C-fibers provide a source of masking inhibition to primary somatosensory cortex. Proc. R. Sot. Lond. B, Biol. Sri., 243: 269-275. Casey, K.L. (2000) Concepts of pain mechanisms: the contribution of functional imaging of the human brain. In: J. Sandktihler, B. Bromm and G.F. Gebhart (Ed%), Nervous System Plasticity and Chronic Pain. Progress in Brain Research, Vol. 129. Elsevier, Amsterdam, pp. 277-287. Casey, K.L., Minoshima, S., Berger, K.L., Koeppe, R.A., Morrow, T.J. and Frey, K.A. (1994) Positron emission tomographic analysis of cerebral structures activated specifically by repetitive noxious heat stimuli. J. Neurophysiol., 71: 802-807. Cummins, T.R., Dib-Hajj, S.D., Black, J.A. and Waxman, S.G. (2000) Sodium channels and the molecular pathophysiology of pain. In: J. Sandktihler, B. Bromm and G.F. Gebhart (Eds.), Nervous System Plasticity and Chronic Pain. Progress in Brain Research, Vol. 129. Elsevier, Amsterdam, pp. 3-19.

H

Devor, M. (1997) Phantom pain as an expression of referred and neuropathic pain. In: R.A. Sherman (Ed.), Phantom Pain. Plenum Press, New York, pp. 33-58. Diamond, M.E., Huang, W. and Ebner, EF. (1994) Laminar comparison of somatosensory cortical plasticity. Science, 265: 1885-1888. Dostrovsky, J.O. (2000) Role of thalamus in pain. In: J. Sandktihler, B. Bromm and G.F. Gebhart (Eds.), Nervous System Plasticity and Chronic Pain. Progress in Brain Research, Vol. 129. Elsevier, Amsterdam, pp. 245-257. Dykes, R.W. (1997) Mechanisms controlling neuronal plasticity in somatosensory cortex. Can. L Physiol. Pharmacol., 75: 535-545. Elbert, T., Junghofer, M., Scholz, B. and Schneider, S. (1995) The separation of overlapping neuromagnetic sources in first and second somatosensory cortices. Brain Topogr, 7: 275282. Flor, H. and Turk, D.C. (2000) A Biobehaviorul Perspective of Chmnic Pain and its Management. APA Press, Washington, D.C., in press. Flor, H., Elbert, T., Knecht, S., Wienbruch, C., Pantev, C., Birbaumer, N., Larbig, W. and Taub, E. (1995) Phantom limb pain as a perceptual correlate of cortical reorganization. Nature, 357: 482-484. Flor, H., Braun, C., Elbert, T. and Birbaumer, N. (1997a) Extensive reorganization of primary somatosensory cortex in chronic back pain patients. Neurosci. L&t., 224: 5-8. Flor, H., Knost, B. and Birbaumer, N. (1997b) Processing of pain-related information in chronic pain patients: electrocortical and peripheral correlates. Pain, 73: 413-421. Flor, H., Elbert, T., Mtihlnickel, W., Pantev, C., Wienbruch, C. and Taub, E. (1998) Cortical reorganization and phantom phenomena in congenital and traumatic upper extremity amputees. Exp. Bruin Res., 119: 205-212. Flor, H., Mtihlnickel, W., Karl, A., Denke, C., Fritzsche, K., Grtisser, S. and Taub, E. (2000) A neural substrate of non-H painful phantom limb phenomena. Neuroreport, 11: 14071411. Florence, S.L., Taub, H.B. and Kaas, J.H. (1998) Large-scale sprouting of cortical connections after peripheral injury in adult macaque monkeys. Science, 282: 1117-l 121. Fordyce, W.E. (1976) A Behavioral Approach to Chronic Pain and Illness. Mosby, St. Louis. Gelnar, PA., Krauss, B.R., Sheehe, P.R., Szeverenyi, N.M. and Apkarian, A.V. (1999) A comparative NRI study of cortical representations for thermal painful, vibrotactile, and motor performance tasks. Neuroimage, 10: 460-482. Grtisser, S., Mtihlnickel, W., Denke, C., Karl, A., Villringer, K. and Flor, H. (2000) The relationship of non-painful phantom phenomena and cortical reorganization. Neuroscience, in press. Guilbaud, G., Berkley, K.J., Benoist, J.M. and Gautron, M. (1993) Responses of neurons in thalamic ventrobasal complex of rats to graded distension of uterus and vagina and to uterine suprafusion with bradykinin and prostaglandin F2 alpha. Bruin Res., 614: 285-290. Jenkins, W.M., Merzenich, M.M., Ochs, M.T., Allard, T. and Guic-Robles, E. (1990) Functional reorganization of primary

somatosensory cortex in adult owl monkeys after behaviorally controlled tactile stimulation. J. Neurophysiol., 63: 82-104. Jensen, T.S. and Nikolajsen, L. (2000) Pre-emptive analgesia in postamputation pain: an update. In: J. Sandktthler, B. Bromm and G.F. Gebhart (Eds.), Nervous System Plasticity and Chronic Pain. Progress in Brain Research, Vol. 129. Elsevier, Amsterdam, pp. 493-503. Kaas, J. (1991) Plasticity of sensory and motor maps in adult mammals. Annu. Rev. Neurosci., 14: 137-167. Karl, A., Birbaumer, N. and Flor, H. (1999) Enhanced P3-amplitudes in amputees with phantom pain in a visual oddball paradigm. Psychophysiology, 36: S65. Karl, A., Birbaumer, N., Lutzenberger, W. Cohen, L. and Flor H. (2000) Reorganization of motor and somatosensory cortex in upper extremity amputees with phantom limb pain. Submitted for publication. Katz, J. (1992) Psychophysiological contributions to phantom limbs. Can. .I. Psychiatry, 27: 283-298. Katz, J. and Melzack, R. (1990) Pain ‘memories’ in phantom limbs: review and clinical observations. Pain, 43: 319-336. Kenshalo, K. and Douglass, D.K. (1995) The role of the cerebral cortex in the experience of pain. In: B. Bromm and J.E. Desmedt (Eds.), Pain and the Brain: From Nociception fo Cognition. Raven Press, New York, pp. 21-34. Kerns, R.D., Turk, D.C. and Rudy, T.E. (1985) The West HavenYale Multidimensional Pain Inventory (WHYMPI). Pain, 23: 345-356. Kleinbohl, D., Holzl, R., Moltner, A., Rommel, C., Weber, C. and Osswald, P.M. (1999) Psychophysical measures of sensitization to tonic heat discriminate chronic pain patients. Pain, 81: 35-43. Knost, B., Flor, H., Braun, C. and Birbaumer, N. (1997) Cerebral processing of words and the development of chronic pain. Psychophysiology, 34: 474-48 1. Lenz, F.A., Lee, J.-I., Garonzik, I.-M., Rowland L.H., Dougherty, PM. and Hua, S.E. (2000) Human thalamic reorganization related to nervous injury and dystonia. In: J. Sandktlhler, B. Bromm and G.F. Gebhart (Eds.), Nervous System Plasticity and Chronic Pain. Progress in Brain Research, Vol. 129. Elsevier, Amsterdam, pp. 259-273. Lorenz, J., Grasedyck, K. and Bromm, B. (1996) Middle and long latency somatosensory evoked potentials after painful laser stimulation in patients with fibromyalgia syndrome. Electroencephalogr Clin. Neurophysiol., 100: 165-168. Lorenz, J., Kohlhoff, H., Hansen, H.C., Kunze, K. and Bromm, B. (1998) Ab-fiber mediated activation of cingulate cortex as correlate of central post-stroke pain. Neuroreport, 9: 659-663. Lotze, M., Grodd, W., Birbaumer, N., Erb, M., Huse, E. and Flor, H. (1999) Does use of a myoelectric prosthesis reduce cortical reorganization and phantom limb pain?. Nat. Neurosci., 2: 501-502. Merzenich, M.M., Nelson, R.J., Stryker, M.P., Cynader, M.S., Schoppmann, A. and Zook, J.M. (1984) Somatosensory cortical map changes following digit amputation in adult monkeys. .I. Comp. Neural., 224: 591-605. Michaelis, M., Blenk, K.H., J&rig, W. and Vogel, C. (1996) Development of spontaneous activity and mechanosensitivity

322 in axotomized afferent nerve fibers during the first hours after nerve transection in rats. J. Neurophysiol., 74: 1020-1027. Montoya, P., Larbig, W., Pulvermiiller, E, Flor, H. and Birbaumer, N. (1996) Cortical correlates of classical semantic conditioning of pain. Psychophysiology, 33: 644-649. Moore, K.A., Baba, H. and Woolf, C.J. (2000) Synaptic transmission and plasticity in the superficial dorsal horn. In: J. Sandkiihler, B. Bromm and G.F. Gebhart (Eds.), Nervous System Plasticity and Chronic Pain. Progress in Brain Research, Vol. 129. Elsevier, Amsterdam, pp. 63-80. Nikolajsen, L., Ilkjaer, S., Christensen, J.H., Kroner, K. and Jensen, T.S. (1997) Randomized trial of epidural bupivacaine and morphine in prevention of stump and phantom pain in lower-limb amputation. Lancet, 350: 1353-1357. NystrBm, B. and Hagbarth, K.E. (1981) Microelectrode recordings from transected nerves in amputees with phantom limb pain. Neurosci. Leti., 27: 211-216. Pons, T.I?, Garraghty, P.E., Ommaya, A.K., Kaas, J.H., Taub, E. and Mishkin, M. (1991) Massive cortical reorganization after sensory deafferentation in adult macaques. Science, 252: 1857-1860. Ramachandran, VS. and Hirstein, W. (1998) The perception of phantom limbs. The D.O. Hebb lecture. Brain, 121: 16031630. Recanzone, G.H., Merzenich, M.M., Jenkins, W.M., Grajski, K.A. and Dinse, H.R. (1992) Topographic reorganization of the hand representation in cortical area 3b owl monkeys trained in a frequency-discrimination task. J. Neurophysiol., 67: 1031-1056. Reeh, P. and PethG, G. (2000) Nociceptor exitation by thermal sensitization - a hypothesis. In: J. Sandklihler, B. Bromm and G.F. Gebhart (Eds.), Nervous System Plasticity and Chronic Pain. Progress in Brain Research, Vol. 129. Elsevier, Amsterdam, pp. 39-50. Sandkiihler, J., Benrath, J., Brechtel, C., Ruscheweyh, R. and Heinke, B. (2000) Synaptic mechanisms of hyperalgesia. In: J. Sandklihler, B. Bromm and G.F. Gebhart (Eds.), Nervous Sys-

tern Plasticity and Chronic Pain. Progress in Brain Research, Vol. 129. Elsevier, Amsterdam, pp. 81-100. Sherman, R.A. (1997) Phantom Limb Pain. Plenum Press, New York. Spitzer, M. (1997) Noise-driven neuroplasticity in self-organizing feature maps: a neurocomputational model of phantom limbs. MD Comput., 14: 192-199. Sutherland, S.P., Cook, S.P. and McCleskey, E.W. (2000) Chemical mediators of pain due to tissue damage and ischemia. In: J. Sandkiihler, B. Bromm and G.F. Gebhart (Eds.), Nervous System Plasticity and Chronic Pain. Progress in Brain Research, Vol. 129. Elsevier, Amsterdam, pp. 21-38. Svendsen, F., Hole, K. and Tjolsen, A. (2000) Long-term potentiation in single WDR neurons induced by noxious stimulation in intact and spinalized rats. In: J. Sandklihler, B. Bromm and G.F. Gebhart (Eds.), Nervous System Plasticity and Chronic Pain. Progress in Brain Research, Vol. 129. Elsevier, Amsterdam, pp. 153-161. Treede, R.D., Lorenz, J., Kunze, K. and Bromm, B. (1995) Assessment of nociceptive pathways with laser-evoked potentials in normal subjects and in patients. In: B. Bromm and J. Desmedt (Eds.), Pain and the Brain: from Nociception to Cognition, Vol. 22. Raven Press, New York, pp. 377-392. Treede, R.D., Kenshalo, D.R., Gracely, R.H. and Jones, A.K. (1999) The cortical representation of pain. Pain, 79: 105- 111. Wang, X., Merzenich, M.M., Sameshima, K. and Jenkins, W.M. (1995) Remodeling of hand representation in adult cortex determined by timing of tactile stimulation. Nature, 378: 7175. Wilder-Smith, O.H.G. (2000) Pre-emptive analgesia and surgical pain. In: J. Sandkiihler, B. Bromm and G.F. Gebhart (Eds.), Nervous System Plasticity and Chronic Pain. Progress in Brain Research, Vol. 129. Elsevier, Amsterdam, pp. 505-534. Woolf, C.J. and Doubell, T.P. (1994) The pathophysiology of chronic pain - increased sensitivity to low threshold A betafiber inputs. Curr: Opin. Neurobiol., 4: 525-534.

J. Sandktihler, B. Bromm and G.F. Gebhart (Eds.) Progress in Brain Research, Vol. 129 0 2000 Elsevier Science B.V. All rights reserved

CHAPTER 23

Neuroplasticity

and clinical pain

Giancarlo Carli * Institute di Fisiologia Umana, Universidad degli Studi di Siena, Via Aido Moro, I-53100 Siena, Italy

Long-term

changes in neuronal excitability

It is well known that, in both vertebrates and invertebrates, repeated exposure to weak somatic stimuli leads to a gradual decrease of the response, i.e. habituation, while exposure to potentially threatening or noxious stimuli produces a progressive potentiation of the response, i.e. sensitization. Any strong repeated stimulus also involves the activation of general arousal systems, which affect the intensity of the behavioral response and thus have an important adaptive value. Habituation and sensitization are considered elementary forms of non-associative learning, whose mechanisms have been studied in some animal models at different levels of analysis: behavioral, cell-physiological, ultrastructural, and molecular. For instance, in Aplysia, it has been shown at the synaptic level that long-term habituation is associated with a decrease in the varicosities containing synaptic active zones and a decrease in the number of transmitter quanta released; sensitization acts in the opposite direction, increasing the strength of each active synapse (Bailey and Kandel, 1993). In mammals, persistent alterations of synaptic efficacy may be produced postsynaptically by alterations in patterns of presynaptic stimulation. Longterm potentiation and long-term depression refer to an increase and a decrease in synaptic strength, respectively; changes in postsynaptic cytosolic calcium *Corresponding author: G. Carli, Instituto di Fisiologia Umana, Universidad degli Studi di Siena, Via Aldo Moro, I-53100 Siena, Italy. Tel. 0039-0577-234038; 0577-234037; E-mail: [email protected]

Fax: 0039-

ion concentration and protein kinase activities play a key role in their generation (Gerber et al., 2000, this volume; Sandktihler et al., 2000, this volume). Thus, repetitive stimuli may lead to persistent ultrastructural and functional changes; these influence the effectiveness of neural connections (Linden and Connor, 1995) with a high spatial specificity, allowing fine-tuning of the transmission strength of each individual synapse depending on its input (Dodt et al., 1999). These use-dependent modifications are denoted plastic changes. Evidence is emerging that neurons in certain areas of the adult primate brain, such as the prefrontal, inferior, and posterior temporal cortex, are capable of dividing, which has led to the hypothesis that new additional neurons may play a role in structural plasticity (Gould et al., 1999). Tissue lesion and inflammation Plasticity-inducing stimuli, such as strong nociceptive stimulation or inflammation, elicit a series of neurobiological mechanisms, which lead to an increase in the level of background activity, expansion of receptive fields, enhanced responses to sensory stimulation, and changes in discharge patterns and temporal correlations of discharges in dorsal horn neurons (Woolf, 1983, 1989; Woolf and Thompson, 1991; Sandkiihler and Eblen-Zajjur, 1994). The increase in excitability involves a cascade reaction triggered by the release of glutamate, substance P, and calcitonin gene-related peptide (GCRP) from primary afferent nociceptive fibers, which interact at synaptic levels through their different receptors and prolong the duration of their excitatory effects. In consequence, there is a mobilization of Ca*’ from

326

intracellular stores resulting in an elevation of the calcium ion concentration. This in turn increases the synthesis of nitric oxide (NO), the production of arachidonic acid and subsequently of prostaglandins, and, in the long term, the activity of nuclear transcription factors. The modifications of the expression of immediate-early genes may in turn change the phenotype of spinal nociceptive neurons and thereby the function of the nociceptive system (Zimmermann and Herdegen, 1994; Millan, 1999). Following a cutaneous injury, there are two zones of hyperalgesia. The inner zone of primary hyperalgesia is characterized by lowered pain threshold and enhanced pain responses to noxious thermal and mechanical stimulation and is caused by a combination of sensitization of peripheral nociceptors and central neurons (Woolf, 1983). The surrounding zone of secondary hyperalgesia is characterized by a lowered pain threshold and enhanced pain response to mechanical, but not thermal, stimulation; it seems to be caused by central neuron sensitization alone but requires ongoing primary afferent nociceptor input for its maintenance (Raja et al., 1988; Treede and Magerl, 2000, this volume). It has been suggested (Treede et al., 1992) that allodynia and hyperalgesia observed in some chronic pain syndromes may be based on mechanisms similar to that of the secondary hyperalgesia observed after experimental injury to human skin. Lesions of the central and peripheral nervous systems The adult mammalian brain is subject to constant modifications due to the activity of the peripheral sensory pathways: once a particular peripheral input becomes inactive, e.g. after nerve section, its former postsynaptic targets can be put to use by fibers of adjacent, normally innervated skin. The axonal sprouts have the mobility to shift to vacated sites and the necessary chromoaffinity to establish new synaptic connections, through reactivation of mechanisms that operate during development. Nerve damage, as well as stimulation, is associated with the induction and peripheral release of several trophic factors, in particular neuropeptides (Millan, 1999). Some of these may have a variety of actions on local tissues which collectively encourage

wound healing, while others, upon their peripheral and/or central release, potentiate the transmission of nociceptive information. These mechanisms are thought to be crucial for the development of neuropathic pain, a condition in which there is an increased sensitivity to painful stimuli, innocuous stimuli felt as painful, and spontaneous pain arises. Axotomy of a peripheral nerve produces numerous changes in sensory neurons, such as bilateral changes in levels of mRNA for cholecystokinin, galanin, neuropeptide Y, and vasoactive intestinal polypeptide (Koltzenburg et al., 1999; Kerr et al., 2000, this volume). The specificity of the contralatera1 effect is apparent in terms of a similar, although reduced, type of response as well as spatial distribution. The neuronal mechanisms involved in nerve injury (not yet determined) could also be related to the development of contralateral inflammation, which occurs in rheumatoid arthritis and other inflammatory joint diseases (Donaldson, 1999). An interesting consequence of deafferentation is the appearance of spontaneous epilepsy-like activity in the spinal cord of rats at the level of the entry zone of the interrupted afferent fibers and in the ventrobasal complex of the contralateral and ipsilateral thalamus (Lombard and Larabi, 1983). The hyperactivity develops slowly, cannot be driven or inhibited by peripheral or central stimuli, is associated with disinhibitory mechanisms involving a loss of GABAergic tone, and is correlated with abnormal animal behavior indicating a state of hyperalgesia and possibly pain. Sleep and plasticity During the spindles and slow waves of sleep, a massive amount of calcium ions enters thalamic and cortical neurons, which might be able to produce changes in gene expression. However, it is becoming increasingly clear that the expression of genes normally associated with plasticity is induced mainly during the awake state and not during sleep, since only little information can be acquired during sleep. For instance, many transcription factors and the expression of several immediate-early genes are high during waking and low during sleep (Tononi and Cirelli, 1999). On the other hand, sleep onset is associated with functional deafferentation of somatosen-

327

sory and visual messages and most likely also of nociceptive input (Carli and Aloisi, 1993). However, sleep can occur (although it is disturbed) in cats under persistent, long-lasting nociceptive stimulation (Carli et al., 1987) and during chronic pain (Carli and Aloisi, 1993). In many chronic pain syndromes, patients complain about non-restorative sleep. For instance, in the case of women with fibromyalgia, the poorer sleepers tend to report significantly more pain and there is an association between sleep disturbance and pain intensity (Affleck et al., 1996). In addition, a night of poorer sleep is followed by a more painful day and a more painful day is followed by a night of poorer sleep (Affleck et al., 1996). Many fibromyalgic patients display an alpha-delta sleep, i.e. an abnormal EEG rhythm correlated with non-restorative sleep (Harding, 1998). Pain can differentially affect sleep stages (Carli et al., 1987) so that in fibromyalgic patients, stage 1 may be increased while all the other sleep stages are decreased (Branco et al., 1994). Thus, the reciprocal relationships between pain and sleep mechanisms are probably important, but are far from being understood. Psychological aspects Clinical pain is a complex experience with several qualities, including sensory-discriminative, motivational-affective, and cognitive-evaluative aspects (Melzack and Casey, 1968). It has been suggested that pain associated with tissue or nerve lesions may cause emotional changes, and that psychological factors, even in the absence of physical injuries, may promote chronic pain (Merskey, 1980). It has also been proposed, but not yet proven, that the pain dimensions are subserved by neurophysiologically specialized systems operating in parallel with reciprocal influences (Melzack and Casey, 1968). For effective pain management, it is critical to identify which dimensions are most important to a patient and their relative contribution to the global pain experience (Bromm et al., 2000, this volume) The relationships between psychological aspects and chronic pain have been thoroughly investigated. While there is no doubt that psychological factors can modulate the pain experience, and in some instances pre-date its onset, the most crucial questions

are related to their role in the acquisition and maintenance of chronic pain. A classical conditioning model of pain has been described by Linton et al. (1985) who postulated that the frequent concurrence of the unconditioned stimulus of pain with innocuous stimuli leads to the development of painrelated responses to these formerly neutral stimuli. In agreement with this hypothesis, it has been shown that stimuli associated with the experience of pain may excite pain-related cortical neurons and create a painful experience; this experience depends on the cortical reorganization process even in the absence of the original nociceptive stimulation at the periphery (Flor et al., 1997; Flor, 2000, this volume). Although many other mechanisms, such as operant conditioning, observational learning, cognitions, and beliefs, may act as mediators between painevoking situations and emotional and behavioral reactions, the most hotly debated question concerns chronic pain in the absence of any physical or psychiatric diagnosis, the so-called ‘chronic indeterminate pain’ (Magni, 1987). Since the theoretical formulation of ‘psychogenic pain’ by Engel (1959), review articles analyzing patients with chronic pain without organic lesions have underlined the recurrent association with depression (Magni, 1987). In particular, it has been postulated (Blumer and Heilbronn, 1982) that an underlying depressive disturbance might be manifested in a somatic form through the symptom of pain: this is likely true for some, but not all, patients affected by chronic indeterminate pain (Merskey, 1980). The Diagnostic and Statistical Manual of Mental Disorders IV (DSM-IV) makes an attempt to overcome these difficulties by using the category of ‘pain disorder’ for patients in whom pain represents the predominant focus of attention. This disorder includes a subcategory associated with psychological factors which are judged to play a significant role in the onset, severity, or maintenance of pain. The main limitation of this classification is that some of the patients afflicted by indeterminate pain do not present psychological or psychiatric symptoms, display an acceptable quality of life, and do not seek any therapy (Aaron et al., 1996). In conclusion, the category of indeterminate pain includes some subtypes of patients for whom no other symptom besides pain can be presently iden-

328

tified and who may display plastic changes in the nociceptive system. Plastic changes in the absence of chronic pain It is common experience that exposure to intense nociceptive stimuli, either continuous or repetitive, represents a frequent antecedent to chronic pain syndromes. Psychophysical tests, recordings of eventrelated potentials, and functional neuroimaging represent useful tools to document plastic changes in the nociceptive system (Bromm, 1984; Handwerker and Kobal, 1993). There are various disturbances which should be briefly analyzed. First of all, hyperalgesia is present in many pain syndromes, such as fibromyalgia, temporomandibular joint disorder, or headache, while in others it may not occur or is present in only one modality (mechanical, thermal, etc.). On the other hand, alteration of the reactivity of the nociceptive system has been described in the absence of recurrent or chronic pain. Increased pain thresholds have been described in some psychiatric syndromes, such as major depression, anorexia nervosa and bulimia, as well as in patients with borderline personality disorders (Lautenbacher and Krieg, 1994). In particular, patients with major depression display a lowered capacity to discriminate painful stimuli alone, while their performance is in the control range for nonpainful stimuli. This suggests that the relative pain insensitivity is not a result of cognitive processes, such as attention, but reflects sensory as well as affective processes (Dworkin et al., 1995). Moreover, these plastic changes are reversible, since there is a normalization of pain perception after recovery from the psychiatric syndrome. It should be underlined that many chronic pain patients, such as fibromyalgics, have a history of major depression and display hyperalgesia to pressure and heat stimuli. A relative reduction of pain and stress is elicited by baroceptor activation in normotensive and hypertensive subjects (Dworkin et al., 1979); hypertension is associated with increased nociceptive thresholds (Zamir and Shuber, 1980). Parental history is also critical since the offspring of essential hypertensive subjects display decreased pain sensitivity (France and Steward, 1995). In addition, in hypotensive individuals, baroceptor activation does not change the

pain threshold (Angrilli et al., 1997). These individual differences in blood pressure levels and corresponding baroceptor reactivity may explain, in some subjects, the differences in pain reactivity and pain coping. Finally, it has long been known that, after peripheral or central nervous system lesions, not all patients display spontaneous pain, although they may suffer from allodynia or hyperpathia. Thus, the occurrence of plastic changes in the nociceptive system is not necessarily linked to chronic pain. Concluding remarks It will be interesting to learn more about the most important factors, in both neuralgic and inflammatory pain, relating plastic changes to chronic pain. Of particular interest are the critical mechanisms modulating the duration of plastic changes, the interactions between the plastic changes which tend to facilitate the onset and maintenance of chronic pain, and the plastic changes which tend to prevent it. Acknowledgements Supported by a Consiglio Nazionale delle Ricerche (CNR) grant, by Minister0 della Ricerca Scientifica e Tecnologica (MURST) 40% grants and by the Universita’ degli Studi di Siena (Piano di Ateneo per la Ricerca - P.A.R.). References Aaron, L.A., Bradley, L.A., Alarcbn, G.S., Alexander, R.W., Triana-Alexander, M., Martin, M.J. and Alberts, K.R. (1996) Psychiatric diagnoses in patients with fibromyalgia are related to health care-seeking behaviour rather than to illness. Arthritis Rheum., 39: 436-445. Affleck, G., Urrows, S., Tennen, H., Higgins, P. and Abeles, M. (1996) Sequential daily relations of sleep, pain intensity, and attention to pain among women with tibromyalgia. Pain, 68: 363-368. Angrilli, A., Mini, A., Mucha, R.F. and Rau, H. (1997) The influence of low blood pressure and baroceptor activity on pain responses. Physiol. Behav., 62: 391-397. Bailey, C.H. and Kandel, E.R. (1993) Structural changes accompanying memory storage. Annu. Rev. Neurosci., 55: 397-426. Blumer, D. and Heilbronn, M. (1982) Chronic pain as a variant of depressive disease. The pain-prone disorder. J. Nerv. Merit. Dis., 170: 381-406.

329 Branco, J., Atalaia, A. and Pavia, T. (1994) Sleep cycles and alpha-delta sleep in fibromyalgia syndrome. J. Rheumatol., 21: 1113-1117. Bromm, B. (1984) Pain-related components of the cerebral potential. Experimental and multivariate statistical approaches. In: B. Bromm (Ed.), Pain Measurement in Man. Neurophysiologicul Correlates of Pain. Elsevier, Amsterdam, pp. 257290. Bromm, B., Scharein, E. and Vahle-Hinz, C. (2000) Cortex areas involved in the process of normal and altered pain. In: J. Sandktihler, B. Bromm and G.F. Gebhart (Eds.), Nervous System Plasticity and Chronic Pain, Progress in Brain Research, Vol. 129. Elsevier, Amsterdam, pp. 289-302. Carli, G. and Aloisi, A.M. (1993) Integrated complex responses following tonic pain. In: L. Vecchiet (Ed.), New Trends in Referred Pain and Hyperalgesia. Elsevier, Amsterdam, pp. 223-238. Carli, G., Montesano, A., Rapezzi, S. and Paluffi, G. (1987) Differential effects of persistent nociceptive stimulation on sleep stages. Behav. Brain Rex, 26: 89-98. Dodt, H.O., Eder, M., Frick, A. and Zieglglnsberger, W. (1999) Precisely localized LTD in the neocortex revealed by infraredguided laser stimulation. Science, 286: 110-l 13. Donaldson, L. (1999) Unilateral arthritis: contralateral effects. TINS, 22: 495-496. Dworkin, B., Filewich, R., Miller, N., Craigmyle, N. and Pickering, T. (1979) Baroceptor activation reduces reactivity to noxious stimulation: implications in hypertension. Science, 205: 1299-1301. Dworkin, R.H., Clark, W.C. and Lip&, I.D. (1995) Pain responsivity in major depression and bipolar disorder. Psychiatry Res., 56: 173-181. Engel, G.L. (1959) ‘Psychogenic’ pain and the pain-prone patient. Am. 1. Med., 26: 899-918. Flor, H. (2000) The functional organization of the brain in chronic pain. In: J. Sandktihler, B. Bromm and G.F. Gebhart (Eds.), Nervous System Plastic+ and Chronic Pain. Progress in Brain Research, Vol. 129. Elsevier, Amsterdam, pp. 313322. Flor, H., Braun, C., Elbert, T. and Birbaumer, N. (1997) Extensive reorganization of primary somatosensory cortex in chronic back pain patients. Neurosci. L&t., 224: 5-8. France, CR. and Steward, K.M. (1995) Parental history of hypertension and enhanced cardiovascular reactivity are associated with decreased pain ratings. Psychophysiology, 32: 571-578. Gerber, G., Youn, D.-H., Hsu, C.H., Isaev, D. and Randic, M. (2000) Spinal dorsal horn synaptic plasticity: involvement of group I metobatropic glutamate receptors. In: J. Sandkiihler, B. Bromm and G.F. Gebhart (Eds.), Nervous System Plasticity and Chronic Pain. Progress in Brain Research, Vol. 129. Elsevier, Amsterdam, pp. 115-134. Gould, E., Reeves, A.J., Graziano, M.S.A. and Gross, C.G. (1999) Neurogenesis in the neocortex of adult primates. Science, 286: 548-552. Handwerker, H.O. and Kobal, G. (1993) Psychophysiology of experimentally induced pain. Physiol. Rev., 73: 639-671.

Harding, S. (1998) Sleep in fibromyalgia patients: subjective and objective findings. Am. J: Med. Sci., 315: 367-376. Kerr, B.J., Wynick, D., Thompson, S.W.N. and McMahon, S.B. (2000) The biological role of galanin in normal and neuropathic states. In: J. Sandktihler, B. Bromm and G.F. Gebhart (Eds.), Nervous System Pfusticity and Chronic Pain. Progress in Brain Research, Vol. 129. Elsevier, Amsterdam, pp. 219230. Koltzenburg, M., Wall, P.D. and McMahon, S.B. (1999) Does the right side know what the left side is doing?. TINS, 22: 122-127. Lautenbacher, S. and Krieg, J.-C. (1994) Pain perception in psychiatric disorders: a review of the literature. J. Psychiatry Res., 28: 1099122. Linden, D.J. and Connor, J.A. (1995) Long-term synaptic depression. Annu. Rev. Neurosci., 18: 3 19-357. Linton, S.J., Melin, L. and Gotestam, K.G. (1985) Behavioral analysis of chronic pain and its management. In: M. Hersen, A. Bellack and H. Heisler (Eds.), Progress in Behavioral Modijicutions, Vol. 18. Academic Press, New York, pp. l-42. Lombard, M.C. and Larabi, Y (1983) Electrophysiological study of cervical dorsal horn cells in partially deafferented rats. In: J.J. Bonica (Ed.), Adtbances in Pain Research and Therupy, Vol. 5. Raven Press, New York, pp. 147-154. Magni, G. (1987) On the relationship between chronic pain and depression when there is no organic lesion. Pain, 3 I: 1-2 I. Melzack, R. and Casey, K.L. (1968) Sensory, motivational and central control determinants of pain: a new conceptual model. In: D. Kenshalo (Ed.), The Skin Senses. Thomas, Springfield, IL, pp. 423-439. Merskey, H. (1980) The role of psychiatrist in the investigation and treatment of pain. Ass. Res. Nerv. Men?. Dis., 58: 249260. Millan, M.J. (1999) The induction of pain: an integrative review. Prog. Neurobiol., 57: 1-164. Raja, S.N., Meyer, R.A. and Campbell, J.N. (1988) Peripheral mechanisms of somatic pain. Anesthesiology, 68: 571-590. Sandkiihler, J. and Eblen-Zajjur, A. (1994) Identification and characterization of rhythmic nociceptive and non-nociceptive spinal dorsal horn neurons in the rat. Neuroscience, 61: 9911006. Sandkiihler, J., Benrath, J., Brechtel, C., Ruscheweyh, R. and Heinke, B. (2000) Synaptic mechanisms of hyperalgesia. In: J. Sandktihler, B. Bromm and G.F. Gebhart (Eds.), Nervous System Plasticity and Chronic Pain. Progress in Brain Research, Vol. 129. Elsevier, Amsterdam, pp. 81-100. Tononi, G. and Cirelli, C. (1999) The frontiers of sleep. TINS, 22: 417-421. Treede, R.-D. and Magerl, W. (2000) Multiple mechanisms of secondary hyperalgesia. In: J. Sandkiihler, B. Bromm and G.F. Gebhart (Eds.), Nervous System Plasticity and Chronic Pain. Progress in Brain Research, Vol. 129. Elsevier, Amsterdam, pp. 331-341. Treede, R.-D., Meyer, R.A., Raja, S.N. and Campbell, J.N. (1992) Peripheral and central mechanisms of cutaneous hyperalgesia. Prog. Neurobiol., 38: 397-421.

330 Woolf, C.J. (1983) Evidence for a central component of postinjury pain hypersensitivity. Nature, 308: 686-688. Woolf, C.J. (1989) Recent advances in pathophysiology of acute pain. BI: .I Anaesth., 63: 139-146. Woolf, C.J. and Thompson, S.W.N. (1991) The induction and the maintenance of central sensitization is dependent on N-methyl-D-aspartic acid receptor activation; implications for

the treatment of postinjury pain hypersensitivity states. Pain, 44: 293-300. Zamir, N. and Shuber, E. (1980) Altered pain perception in hypertensive humans. Brain Rex, 201: 411-414. Zimmermann, M. and Herdegen, T. (1994) Control of gene transcription by Jun and Fos proteins in the nervous system. Am. Pain Sot. .I., 3: 33-48.

3. Sandkiihler, B. Bromm and G.F. G&hart (Eds.) Progress in Brain Research, Vol. 129 0 2000 Elsevier Science B.V. All rights reserved

CHAPTER 24

Multiple mechanisms of secondary hyperalgesia Rolf-Detlef Treede* and Walter Magerl Institute

of Physiology

and Pathophysiology,

Johannes

Introduction

Tissue injury causes enhanced pain sensitivity both at the site of tissue damage (primary hyperalgesia) and in the surrounding skin (secondary hyperalgesia). Whereas sensitization of primary nociceptive afferents largely accounts for primary hyperalgesia, secondary hyperalgesia is attributed to altered processing in the central nervous system (Treede et al., 1992b; Millan, 1999). The sensory characteristics of secondary hyperalgesia are similar to those of hyperalgesia in neuropathic pain following lesions of the peripheral or central nervous system and to those of referred hyperalgesia following lesions of visceral organs (Treede et al., 1992b; Millan, 1999; Woolf and Mannion, 1999). Therefore, to understand the mechanisms of secondary hyperalgesia is highly relevant for clinical practice. Altered neuronal response properties consistent with the characteristics of secondary hyperalgesia have been observed in the spinal cord (Simone et al., 1991; Dougherty et al., 1998; Pertovaara, 1998), the thalamus (Guilbaud et al., 1987; Sherman et al., 1997) and the primary somatosensory cortex (D.R. Kenshalo Jr., pers. commun.), but not in primary nociceptive afferents (for review see Treede et al., 1992b). These observations suggest that the fundamental mechanisms of secondary hyperalgesia may consist of modulation of neural transmission in the

*Corresponding author: R.-D. Treede, Institute of Physiology and Pathophysiology, Johannes Gutenberg University, Saarstr. 21, D-55099 Mainz, Germany. Fax: +49-6131-392.5902; E-mail: [email protected]

Gutenberg

Universi@,

Saarstl:

21, D-55099

Maim,

Germany

spinal cord, which is then projected to higher neural centers. Spinal cord dorsal horn neurons express several mechanisms that can be triggered by a noxious conditioning event (injury discharge) and lead to enhanced synaptic efficacy (central sensitization). These mechanisms subsequently lead to enhanced output from the spinal cord even if the primary afferent input is not sensitized. The duration of this central sensitization may last from seconds (wind-up) to many hours (long-term potentiation; see Gerber et al., 2000, this volume; Sandktihler et al., 2000, this volume; Svendsen et al., 2000, this volume). The neuropharmacological basis of these mechanisms consists of the interaction of excitatory amino acids (e.g., glutamate acting at NMDA- and mGlureceptors) and neuropeptides (e.g., substance P acting at NKl- and NK2-receptors) with each other and with descending inhibitory and facilitatory pathways (Dickenson, 1995; Basbaum, 1999; see also Moore et al., 2000, this volume; Pertovaara, 2000, this volume). These interactions are conveniently depicted in schematic diagrams of the synapse between a C-fiber afferent and a projection neuron in the spinal cord (Urban et al., 1994; Dickenson, 1995). However, some basic observations in secondary hyperalgesia and the related central sensitization are not compatible with models that consist of only one synapse between a primary nociceptive afferent and a nociceptive projection neuron (e.g., spinothalamic). The aim of this paper is to define the minimal conditions of complexity that must be fulfilled by a model of plasticity of spinal nociceptive transmission, in order to explain clinical and psychophysical observations in humans.

332

Secondary hyperalgesia and central sensitization

Secondary hyperalgesia is characterized by a leftward shift of the stimulus-response function for noxious mechanical stimuli such as pinpricks (LaMotte et al., 1991; Magerl et al., 1998). In contrast, pain elicited by noxious heat stimuli remains constant or may even be reduced when applied adjacent to the site of injury (Raja et al., 1984; Dab1 et al., 1993; Ali et al., 1996). Recordings from primary nociceptive afferents have failed to show changes in stimulus-response functions to thermal stimulation when their receptive fields were adjacent to an injury site (Thalhammer and LaMotte, 1982; Campbell et al., 1988; Schmelz et al., 1996) although these afferents are strongly sensitized to heat stimuli when their receptive fields are injured directly (Meyer et al., 1994). Recordings from spinal cord neurons showed leftward shifts in stimulus-response functions only for noxious mechanical stimuli; responses to noxious heat stimuli applied to skin in the vicinity of an injury site were unchanged or reduced (Simone et al., 1991; Dougherty et al., 1998; Pertovaara, 1998). Fig. 1 illustrates that secondary hyperalgesia to probing the skin with a 225 mN v. Frey hair is reflected in the responses of spinal neurons (both wide dynamic range and nociceptive specific neurons), but not primary nociceptive afferents (both A- and C-fibers). These lines of evidence indicate that central sensitization in the spinal cord dorsal horn may fully account for the characteristics of secondary hyperalgesia. This type of central sensitization, however, needs to be specific for mechanosensitive pathways, since sensitization of polymodal pathways would imply an enhanced sensitivity to mechanical and heat stimuli. An increase in heat sensitivity, however, has neither been found psychophysically, nor in neurophysiological recordings. Complementary roles for A- and C-fiber nociceptors in secondary hyperalgesia

In order to define the afferent pathways involved in inducing central sensitization and in mediating the hyperalgesia to noxious mechanical stimuli, we performed several psychophysical experiments using selective nerve block techniques. Compression of a nerve against a bony prominence is an established

technique to block impulse conduction in myelinated but not unmyelinated nerve fibers (Sinclair and Hinshaw, 1950). However, sensitivity to nerve compression varies across subgroups of myelinated fibers: low-threshold mechanoreceptors and thermoreceptors (‘cold fibers’) are blocked rapidly (Fruhstorfer, 1984), whereas conduction in A-fiber nociceptors is blocked significantly later (Ziegler et al., 1999). In order to differentiate the functions of A- and C-fiber nociceptors, it was essential to demonstrate complete functional block of first pain sensation in addition to other A-fiber-mediated modalities such as touch and cold. Shifts in the distributions of reaction times to noxious mechanical stimuli into the C-fiber range (about 1 s) were used for this purpose. The results of these experiments are reported in the following two sections. Central sensitization is induced by C-jber nociceptors

The first aim was to delineate the relative contributions of A- and C-fiber nociceptors to the induction of central sensitization. For this purpose, we induced secondary hyperalgesia by intradermal injections of capsaicin. Capsaicin excites polymodal A- and C-fiber nociceptors, and less well defined classes of chemospecific nociceptors (Szolcsanyi et al., 1988; Baumann et al., 1991; Holzer, 1991). The discharges of these primary afferents are sufficient to induce central sensitization in the spinal cord leading to neurogenic hyperalgesia without causing actual tissue damage (LaMotte et al., 1991). Fig. 2A,B shows that a conditioning injection of 40 Kg of capsaicin into the back of the hand was equally painful with or without complete A-fiber blockade. This finding indicates that the excitation of A-fiber nociceptors makes little if any contribution to the sensation elicited by the injection. In Fig. 2C,D we plotted the average pain elicited by mechanical stimulation of the skin by punctate probes (200 pm diameter) before and 30 min after the conditioning capsaicin injection. At this time, the nerve compression had been released and all A-fiber mediated sensory functions had fully recovered (see Fig. 3A). Mechanically evoked pain was increased by a factor of two, irrespective of whether capsaicin had been injected with or without A-fiber blockade. These

333

6 Injury

1” Zone

C

Secondary

to heat

Secondary I-- 3eralgesia /

\

hyperalgesia

Primary afferents 50

Hyperalgesia

2” Zone

to punctate mechanical stimuli Psychophysics

Spinal neurons

6

-I

IO-

WDR

HT

:/

post

0

01 Pre

post

ve

post

Fig. 1. Secondary hyperalgesia is characterized by mechanical but not heat hyperalgesia. (A) An injury or an inflammatory process induces increased pain sensitivity both within the injury site (primary hyperalgesia: shaded area) and in unaffected skin surrounding this site (secondary hyperalgesia: open area). (B) Hyperalgesia to heat is present in the primary but absent in the secondary zone (mean f sem, n = 8 subjects, data from (Raja et al., 1984). (C) Mechanical hyperalgesia in the secondary zone is due to central sensitization: response characteristics of primary nociceptive afferents with receptive fields in the zone of secondary hyperalgesia do not change significantly after adjacent capsaicin (data from Baumann et al., 1991). Dorsal horn neurons with receptive fields in the zone of secondary hyperalgesia enhance their response to mechanical stimuli (data from Simone et al., 1991). Pain ratings evoked by mechanical stimuli are enhanced in the zone of secondary hyperalgesia (data from Simone et al., 1991). Test stimulus: 225 mN v. Frey hair; conditioning stimulus: adjacent i.d. injection of 100 kg capsaicin. AMH = A-fiber mechanoheat nociceptor; CMH = C-fiber mechanoheat nociceptor; WDR = wide dynamic range neuron; HT = high threshold neuron; n.s. = not significant (ANOVA); * = significantly different (ANOVA).

findings indicate that A-fiber nociceptors also make little or no contribution to central sensitization in secondary hyperalgesia (Ziegler et al., 1999).

Secondary hyperalgesia to punctate stimuli is mediated by A-Jiber nociceptors

The second aim was to delineate the afferent fiber population that mediates the hyperalgesia to noxious

Side of nerve block

Contralateral control h

80 67 9 60 9e

80si 5 60se -

.p 40 F?

.g 40E -

;

z 20n

20

0

1

2 3 Time (min)

4

5

s’ Capsaicin *** i

WI* l

6

I

1

2 3 Time (min)

4

5

l

Baseline

*

*

20 50 100 200 200 pm probes -force

500 (mN)

20 50 100 200 200 pm probes -force

500 (mN)

Fig. 2. Secondary hyperalgesia can be induced by selective C-fiber stimulation. To induce secondary hyperalgesia, i.d. injections of capsaicin (40 ug in 13 ~1) were given into the radial nerve territories of the left hand after complete block of A-fiber nociceptors (A) and of the normal right hand (B). The pain evoked by the injections was not significantly different between the two sides. 20 min after the capsaicin injection, the nerve block was released and sensory function quickly returned to normal in the left hand. The magnitude of secondary hyperalgesia was tested 30 min after the capsaicin injections using punctate mechanical stimuli (200 urn diameter, 35-407 mN). Significant hyperalgesia of equal magnitude was demonstrated independently of whether the capsaicin injection had selectively activated C-fibers (C, previously blocked side) or both A- and C-fibers (D, control side). Means i sem, n = 9 subjects. Modified from Ziegler et al., 1999.

mechanical stimuli. The mechanical test stimuli used in our studies (200 pm diameter cylindrical probes) were adequate to activate both A- and C-fiber nociceptors (Greenspan and McGillis, 1994; Garell et al., 1996). These stimuli were used to test for secondary hyperalgesia during a differential A-fiber blockade that was characterized by a shift in the average reaction times to pinpricks from 180 ms to 1150 ms (Fig. 3A). From distribution histograms of the reaction times we derived a boundary between responses mediated by A-fiber nociceptors (first pain)

and C-fiber nociceptors (second pain) at 500 ms (Ziegler et al., 1999; cf. Campbell and LaMotte, 1983). When reaction times below 500 ms were completely absent, indicating blockade of A-fiber nociceptors, more than 70% of the noxious mechanical test stimuli were still perceived by the subjects (albeit at a lower subjective magnitude: compare Fig. 3B with 3C), indicating that C-fiber nociceptors were still conducting. Fig. 3B shows that the pain elicited by the 200 km diameter probes during a complete A-fiber block was

335

A

2000-

First pain reaction ***

times

B

Side of nerve block

Contralateral

control

side

202 IO:

3 1000: E. i

P

tioo-

E 'S di

2oo100~

P I Pre

P I Block

I Post

50 100 200 500 20 200 pm probes - force (mN)

20 50 100 200 500 200 pm probes - force (mN)

Fig. 3. Only A-fiber input is centrally facilitated in secondary hyperalgesia. (A) Complete blockade of A-fiber nociceptor function was verified by determining the reaction times to pin pricks. After compression of the superficial radial nerve for 65 min, all reaction times were beyond 500 ms, indicating that only C-fibers were conducting. After release of nerve compression, reaction times rapidly returned to baseline levels. (B) During A-fiber nociceptor blockade, the pain evoked by punctate mechanical stimuli (200 urn diameter) was significantly reduced. There was no enhancement of the remaining C-fiber-mediated pain 10 min after the adjacent capsaicin injection. (C) In the control hand without nerve block, capsaicin injection caused a twofold increase in pain ratings to punctate mechanical stimuli within 10 min. Means f sem, n = 9 subjects. Modified from Ziegler et al., 1999.

of small magnitude, but significantly different from zero and dependent on stimulus intensity. At 10 min after the conditioning capsaicin injection, this C-fiber mediated mechanical pain was not increased, but rather decreased. In contrast, at the same time after capsaicin injection, hyperalgesia was fully developed in normal skin (Fig. 3C), indicating that the time lag of 10 min between injection and sensory testing was sufficient for secondary hyperalgesia to develop. These findings indicate that the central sensitization of secondary hyperalgesia to punctate stimuli is specific for a mechanosensitive A-fiber pathway, whereas mechanosensitive C-fiber pathways are not facilitated. In separate experiments we have found that the facilitated A-fiber pathway is insensitive to the desensitizing action of capsaicin, whereas the mechanosensitive C-fiber pathway is blocked by topical treatment with capsaicin (Magerl et al., 1999). As sensitivity to heat stimuli and to capsaicin are closely correlated in polymodal nociceptive neurons (Szolcsanyi et al., 1988; Kirschstein et al., 1999) the absence of heat hyperalgesia and the lack of facilitation of the C-fiber pathway indicate that polymodal nociceptive afferent pathways are exempt from central facilitation in secondary hyperalgesia. Central facilitation in secondary hyperalgesia thus appears to be specific for afferents with the fol-

lowing properties: myelinated fibers, mechanically sensitive, heat-insensitive, and capsaicin-insensitive. These properties define a specific group of nociceptive afferents to be most likely to subserve the facilitated pathway in secondary hyperalgesia to punctate stimuli: A-fiber high-threshold mechanoreceptors (HTM). Formally, A-fiber low-threshold mechanoreceptors that normally subserve touch sensation have a similar set of properties (see next section). Differences between hyperalgesia to light touch and to punctate mechanical stimuli Psychophysical studies suggest that at least two separate mechanosensitive pathways are centrally facilitated in secondary hyperalgesia. Fig. 4 illustrates the experimental evidence for such a distinction. In a typical subject, lightly touching the skin in the vicinity of a conditioning injection of capsaicinelicited moderate pain (Fig. 4A). Because these stimuli are never painful in normal skin, this phenomenon fulfills the IASP definition of allodynia: “Pain due to a stimulus which does not normally provoke pain” (Merskey and Bogduk, 1994). Moreover, these test stimuli activate only low-threshold mechanoreceptors and not nociceptors (Leem et al., 1993b), suggesting that hyperalgesia to light

336 A IOOj 5i g

zp E .E I?

B

30 10-T

l

3-

I

I-,

3 -

o=* CW QT BR

oi,,,,

, , ,,,,,, ,

,,

10 20 50 100 200 500 200 pm probes -force (mN)

Fig. 4. Hyperalgesia to light touch and to punctate stimuli induced by a capsaicin injection in a normal subject. (A) Hyperalgesia to light touch: light tactile stimuli were not painful when moved across normal skin (open symbols). Adjacent to an i.d. injection of capsaicin (40 kg, 13 ul), the same stimuli were mildly painful (filled symbols). This difference was significant (ANOVA, Ft.12 = 176.4, p c 0.001). CW = cotton wisp 3 mN; QT = Q-tip 100 mN; BR = brush 400 mN. (B) Hyperalgesia to punctate stimuli: the stimulus-response function for punctate mechanical stimuli (200 urn diameter) was significantly shifted to the left following the capsaicin injection (ANOVA, Fr,t~ = 17.7, p < 0.001). This leftward shift signifies both a decrease in pain threshold and an increase in pain elicited by suprathreshold stimuli. Subject 304, male, 30 years old.

touch may be Al+fiber-mediated pain (cf. Woolf and Doubell, 1994). A characteristic feature of these test stimuli is that they are moved across the skin; static application of the same stimuli often does not evoke pain. These observations have led some authors to suggest the term ‘dynamic hyperalgesia’ for this phenomenon (Koltzenburg et al., 1992; Ochoa and Yamitsky, 1993). In the same area of secondary hyperalgesia surrounding a capsaicin injection, the painfulness of punctate mechanical stimuli (200 pm diameter) was significantly increased by a factor of three (Fig. 4B). Because these stimuli are also painful in normal skin, this phenomenon fulfills the IASP definition of hyperalgesia: “An increased response to a stimulus which normally is painful” (Merskey and Bogduk, 1994). Moreover, their intensity is sufficient to activate nociceptive afferents (Garell et al., 1996), suggesting that hyperalgesia to punctate mechanical stimuli may simply be enhanced nociceptormediated pain. The leftward shift of the stimulusresponse function for noxious mechanical stimuli, however, also includes a marked lowering of me-

chanical pain threshold; in this case the stimulus intensity at which 50% of stimuli were perceived as painful decreased from 38.1 mN to 9.5 mN (cf. LaMotte et al., 1991; Magerl et al., 1998; Ziegler et al., 1999). According to IASP taxonomy this aspect of the shift in stimulus-response function is sometimes labelled ‘allodynia’, because by definition all stimulus intensities below the normal pain threshold are not normally painful. This is a trivial statement, but the threshold for activation of primary nociceptive afferents is also considerably below the normal pain threshold (Adriaensen et al., 1983; Leem et al., 1993a). Thus, the lowering of the pain threshold depicted in Fig. 4B likely reflects an enhanced central response to normal activity of nociceptive afferents (capsaicin-insensitive A-fiber nociceptors, see previous section). If the term ‘allodynia’ is intended to indicate a situation where the perceived modality for a given stimulus is changed (from touch to pain, cf. Merskey and Bogduk, 1994), this term should be reserved for the touch-evoked pain depicted in Fig. 4A. Table 1 summarizes the fundamental differences between touch-evoked pain and hyperalgesia to punctate stimuli with respect to stimulus characteristics, afferent fiber class, threshold for induction, dependence on maintaining input, duration of effect, and size of the area of altered skin sensitivity. There is some evidence that the central neuropharmacology of the two phenomena is partly different: NMDA-receptor antagonists tended to be more effective against hyperalgesia to punctate stimuli (Park et al., 1995) whereas an AMPA/kainate antagonist tended to be more effective against hyperalgesia to light touch (Sang et al., 1998). In most experimental and clinical studies, however, these two types of secondary hyperalgesia were not assessed differentially. It should be evident from this list that it is useful to reserve the term ‘allodynia’ for dramatic threshold reductions that imply a switch in the primary afferents involved. For clinical sensory testing, there is a simple operational distinction between allodynia and hyperalgesia to punctate stimuli: test stimuli in Fig. 4A were of different force, but all were moved across the skin (dynamic hyperalgesia, allodynia) and were only sufficient to activate low-threshold mechanoreceptors. In contrast, stimuli in Fig. 4B were applied to very small skin areas without move-

337

TABLE Differences

1 between

hyperalgesia

to light touch Hyperalgesia

Test stimuli: intensity application prototype A-fiber block: partial complete Induction threshold Incidence Duration Size of affected area Maintaining input IASP taxonomy

and to punctate

mechanical

to light touch

below nociceptor moving, dynamic soft brush, Q-tip

threshold

blocked blocked high low minutes-hours small obligatory allodynia

stimuli

Hyperalgesia

to punctate

stimuli

References

above nociceptor threshold punctate contact area, static pin prick, v. Frey hair

13 4, 10, 13 6, 8, 13

persists blocked low high hours-days large relatively independent hyperalgesia

2, 3, 13 1 I, 8, 6, I, 6, I, 5, 6 9

*

12

11 13 12

* 1: Cervero et al. (1993), 2: Jorum et al. (2000), 3: Kilo et al. (1994), 4: Koltzenburg et al. (1992), 5: Koltzenburg et al. (1994), LaMotte et al. (1991), 7: Liu et al, (1998). 8: Magerl et al. (1998), 9: Merskey and Bogduk (1994), 10: Ochoa and Yarnitsky (1993), Sethna et al. (1998), 12: Treede and Cole (1993) 13: Ziegler et al. (1999).

ment (static hyperalgesia, hyperalgesia to punctate stimuli), and their intensity was sufficient to activate nociceptive afferents. A heterosynaptic sensitization

facilitation

model of central

As discussed above, secondary hyperalgesia is likely to result from sensitization of nociceptive neurons in the spinal cord dorsal horn. Most diagrams of central sensitization are simplified single-neuron models in which a critical target neuron, namely the nociceptive projection neuron, is thought to be sensitized by the concerted action of neurokinins and excitatory ammo acids (e.g., McMahon et al., 1993; Urban et al., 1994). However, these models are difficult to reconcile with the psychophysical findings on secondary hyperalgesia presented in this chapter and with recent electrophysiological data. Our psychophysical findings in secondary hyperalgesia indicate that the facilitated pathways (capsaicin-insensitive A-fibers) are separate from the facilitating pathway (capsaicin-sensitive C-fibers). Central sensitization in secondary hyperalgesia is limited to mechanoreceptor input, since hyperalgesia to heat is absent (Raja et al., 1984; LaMotte et al., 199 1) or heat pain perception may even be reduced (Ali et al., 1996). Moreover, as demonstrated above, C-fiber-mediated mechanical pain is not facilitated

6: 11:

at all; only A-fiber mechanoreceptive input is facilitated. These psychophysical findings are consistent with electrophysiological data from nociceptive projection cells of the spinal dorsal horn in rat, cat and monkey. Transmission of noxious heat after adjacent capsaicin injection is either unchanged (Simone et al., 199 1; Pertovaara, 1998) or even transiently inhibited (Dougherty et al., 1998) at the same time when a clear cut sensitization to mechanical stimuli is present. Thus, secondary hyperalgesia, which is a special form of central sensitization of nociceptive pathways, likely involves a more complex circuitry than assumed by single-neuron models. To account for this complexity, we had previously proposed that the critical sensitized neurons in the spinal cord may be a subclass of mechanosensitive intemeurons that are presynaptic to spinothalamic projection neurons (Treede and Magerl, 1995). The projection neurons themselves are not sensitized in this model, which accounts for the unchanged sensitivity to C-fiber-mediated pain in the zone of secondary hyperalgesia. The intemeurons receive modulating input from C-fiber nociceptors located inside the zone of tissue injury. This input leads to heterosynaptic facilitation (Fig. 5) of input from A-fiber low-threshold mechanoreceptors (LTMs) that terminate in adjacent non-injured skin; as a consequence, light touch may elicit a sensation of burning pain (cf. Torebjork et al., 1992). We have now extended this model to in-

338

Secondary zone

Primary zone .

_

_

.

-

_

.

(

-

.

.

-

-

_

-

-

,

:. touch _ _. _ _ _’: :-pinprick - _ _. -. - .’0

1 transduction

:

: normal :. .._.._ transduction .._....

i i

: ..__...

I .___-__ normal 1 :’ .-_. transmission :* -_..-...._ /

w-e

normal oathwav facilitating pathway facilitated pathway

Fig. 5. Model for cutaneous secondary hyperalgesia. An adjacent injury or capsaicin injection excites chemosensitive (capsaicin-sensitive) C-nociceptors (including polymodal C-fiber nociceptors, mechanically insensitive afferents, chemoheat nociceptors, chemospecific nociceptors). These afferents induce sensitization of nociceptive neurons in the spinal cord dorsal horn. This sensitization is heterosynaptic and leads to facilitated synaptic transmission of two other inputs onto nociceptive projection neurons (both of which are mechanosensitive). The first facilitated pathway (responsible for secondary hyperalgesia to light touch) involves low-threshold mechanoreceptors (A-fiber LTM) that normally signal touch sensation. The second facilitated pathway (responsible for secondary hyperalgesia to punctate stimuli) involves A-fiber nociceptors including high-threshold mechanoreceptors (A-fiber HTM) and A-fiber nociceptors with high heat thresholds (type I AMHs). C-fiber inputs themselves are not facilitated. The absence of heat hyperalgesia in the secondary zone suggests that inputs from heat sensitive nociceptors in general are not facilitated, including polymodaf A-fibers that normally signal first pain to heat (type II-AMHs) and polymodal C-fiber nociceptors. Modified from Ziegler et al., 1999.

elude a second facilitated pathway that accounts for hyperalgesia to punctate stimuli. Although this second pathway consists of nociceptive afferents, again the facilitated pathway (A-fiber nociceptors) does not overlap with the facilitating pathway (capsaicin-sensitive polymodal C-fiber nociceptors). Clinical implications The observations in experimentally induced secondary hyperalgesia are reminiscent of the different types of mechanical hyperalgesia seen in patients

with neuropathic pain. Dynamic hyperalgesia (tested by a soft brush) was abolished by A-fiber blockade, whereas static hyperalgesia was not (Ochoa and Yarnitsky, 1993; Koltzenburg et al., 1994). These A-fiber blocks were probably partial (criteria: loss of touch and cold sensation, but loss of first pain not verified), and thus the static hyperalgesia in neuropathic pain patients is also consistent with mediation by A-fiber nociceptors. The two separate primary afferent channels that mediate experimentally induced secondary hyperalgesia are also likely to be involved in patients suffering from neuropathic pain.

339

Under normal conditions, including acute pain, the modality-specific facilitation of a mechanically sensitive sensory channel in secondary hyperalgesia provides a spatial warning system in uninjured tissue that results in a functionally adequate guarding behavior. This behavior prevents the access of potentially injurious mechanical stimuli to the adjacent injury site, and thus avoids interference with the progress of wound healing. The functional segregation of the spatial warning system (capsaicin-insensitive A-fibers) and the system signaling the injury (capsaicin-sensitive C-fibers) prevents self-facilitation and maintenance of central sensitization by the facilitated pathway itself. This system may, however, become maladaptive, when either the capsaicin-sensitive C-fiber pathway develops sustained ongoing activity (Devor et al., 1992; Gracely et al., 1992; Treede et al., 1992a) or when the capsaicin-insensitive A-fiber pathway develops the capacity to produce self-facilitation (Ma and Woolf, 1996; Neumann et al., 1996; Baba et al., 1999). Such mechanisms may contribute to chronification of neuropathic pain. Assuming that the mechanisms of hyperalgesia (stimulus-evoked pain) and those of ongoing pain are related, consideration of the pharmacological properties of the A-fiber nociceptor pathway has important therapeutic implications. For example, A-fiber nociceptors, in contrast to C-fiber nociceptors, lack presynaptic opioid receptors (Taddese et al., 1995). Likewise, neuropathic pain states are often not sufficiently ameliorated by opioid treatment (Am& and Meyerson, 1988; Jadad et al., 1992; Dellemijn, 1999). Our model also predicts that the utility of topical capsaicin as a treatment for neuropathic pain should be limited to cases where the facilitating and maintaining focus (the capsaicin-sensitive ‘irritable’ C-nociceptor; see Fields et al., 1998) is accessible for inactivating treatment. These examples underline how understanding the diverse roles of anatomically and functionally distinct nociceptive subsystems will enable a more rational and mechanism-based treatment of neuropathic pain syndromes (Woolf and Mannion, 1999). Acknowledgements The work presented in this paper was supported by NATO (CRG95032540495).

References Adriaensen, H., Gybels, J., Handwerker, H.O. and van Hees, J. (1983) Response properties of thin myelinated (A-6) fibers in human skin nerves. J. Neurophysiol., 49: 111-122. Ah, Z., Meyer, R.A. and Campbell, J.N. (1996) Secondary hyperalgesia to mechanical but not heat stimuli following a capsaicin injection in hairy skin. Pain, 68: 401-411. Am&, S. and Meyerson, B.A. (1988) Lack of analgesic effect of opioids on neuropathic and idiopathic forms of pain. Pain, 33: 1 l-23. Baba, H., Doubell, T.P and Woolf, C.J. (1999) Peripheral inflammation facilitates A8 fiber-mediated synaptic input to the substantia gelatinosa of the adult rat spinal cord. L Neurosci., 19: 859-867. Basbaum, AI. (1999) Distinct neurochemical features of acute and persistent pain. Proc. Natl. Acad. Sci. USA, 96: 77397743. Baumann, T.K., Simone, D.A., Sham, C.N. and LaMotte, R.H. (1991) Neurogenic hyperalgesia: the search for the primary cutaneous afferent fibers that contribute to capsaicin-induced pain and hyperalgesia. J. Neurophysiol., 66: 212-227. Campbell, J.N. and LaMotte, R.H. (1983) Latency to detection of first pain. Brain Rex, 266: 203-208. Campbell, J.N., Khan, A.A., Meyer, R.A. and Raja, S.N. (1988) Responses to heat of C-fiber nociceptors in monkey are altered by injury in the receptive field but not by adjacent injury. Pain, 32: 327-332. Cervero, F., Gilbert, R., Hammond, R.G.E. and Tanner, J. (1993) Development of secondary hyperalgesia following nonpainful thermal stimulation of the skin - a psychophysical study in man. Pain, 54: 181-189. Dahl, J.B., Brennum, J., Arendt-Nielsen, L., Jensen, T.S. and Kehlet, H. (1993) The effect of pre- versus postinjury infiltration with lidocaine on thermal and mechanical hyperalgesia after heat injury to the skin. Pain, 53: 43-51. Dellemijn, P (1999) Are opioids effective in relieving neuropathic pain? Pain, 80: 453-462. Devor, M., Wall, P.D. and Catalan, N. (1992) Systemic lidocaine silences ectopic neuroma and DRG discharge without blocking nerve conduction. Pain, 48: 261-268. Dickenson, A.H. (1995) Spinal cord pharmacology of pain. Bx J. Anaesth., 75: 193-200. Dougherty, P.M., Willis, W.D. and Lenz, EA. (1998) Transient inhibition of responses to thermal stimuli of spinal sensory tract neurons in monkeys during sensitization by intradermal capsaicin. Pain, 77: 1299136. Fields, H.L., Rowbotham, M. and Baron, R. (1998) Postherpetic neuralgia: irritable nociceptors and deafferentation. Neurobiol. Dis., 5: 209-227. Fruhstorfer, H. (1984) Thermal sensibility changes during ischemic nerve block. Pain, 20: 355-361. Garell, PC., Mcgillis, S.L.B. and Greenspan, J.D. (1996) Mechanical response properties of nociceptors innervating feline hairy skin. J. Neumphysiol., 75: 1177-l 189. Gerber, G., Youn, D.-H., Hsu, C.H., Isaev, D. and Randic, M. (2000) Spinal dorsal horn synaptic plasticity: involvement of

340

group I metabotropic glutamate receptors. In: J. Sandktihler, B. Bromm and G.F. Gebhart (Eds.), Nervous Sysfem Plasticity and Chronic Pain. Progress in Brain Research, Vol. 129. Elsevier, Amsterdam, pp. 115-134. Gracely, R.H., Lynch, S.A. and Bennett, G.J. (1992) Painful neuropathy: altered central processing, maintained dynamically by peripheral input. Pain, 51: 175-194. Greenspan, J.D. and McGillis, S.L.B. (1994) Thresholds for the perception of pressure, sharpness, and mechanically evoked cutaneous pain: effects of laterality and repeated testing. Somatosens. Mot. Res., 11: 311-317. Guilbaud, G., Neil, A., Benoist, J.M., Kayser, V. and Gautron, M. (1987) Thresholds and encoding of neuronal responses to mechanical stimuli in the ventro-basal thalamus during carrageenin-induced hyperalgesic inflammation in the rat. Exp. Brain Res., 68: 311-318. Holzer, P (1991) Capsaicin: cellular targets, mechanisms of action, and selectivity for thin sensory neurons. Pharmacol. Relj., 43: 143-201. Jadad, A.R., Carroll, D., Glynn, C.J., Moore, R.A. and Mcquay, H.J. (1992) Morphine responsiveness of chronic pain: double-blind randomised crossover study with patient-controlled analgesia. Lance& 339: 1367-1371. Jorum, E., Wamcke, T., Ziegler, E.A., Magerl, W., Fuchs, P.N., Meyer, R.A. and Treede, R.-D. (2000) Secondary hyperalgesia to punctate stimuli is mediated by A-fiber nociceptors. In: M. Devor, M. Rowbotham and Z. Wiesenfeld-Hallin (Ed%), Proceedings of the 9th World Congress on Pain. Progress in Pain Research and Management, Vol. 16. IASP Press, Seattle, WA, pp. 215-223. Kilo, S., Schmelz, M., Koltzenburg, M. and Handwerker, H.O. (1994) Different patterns of hyperalgesia induced by experimental inflammation in human skin. Brain, 117: 385-396. Kirschstein, T., Greffrath, W., Btisselberg, D. and Treede, R.-D. (1999) Inhibition of rapid heat responses in nociceptive primary sensory neurons of the rat by vanilloid receptor antagonists J. Neurophysiol., 82: 2853-2860. Koltzenburg, M., Lundberg, L.E.R. and Torebjork, H.E. (1992) Dynamic and static components of mechanical hyperalgesia in human hairy skin. Pain, 51: 207-219. Koltzenburg, M., Torebjork, H.E. and Wahren, L.K. (1994) Nociceptor modulated central sensitization causes mechanical hyperalgesia in acute chemogenic and chronic neuropathic pain. Brain, 117: 579-591. LaMotte, R.H., Shain, C.N., Simone, D.A. and Tsai, E.-F.P. (199 1) Neurogenic hyperalgesia: psychophysical studies of underlying mechanisms. J. Neurophysiol., 66: 190-211. Leem, J.W., Willis, W.D. and Chung, J.M. (1993a) Cutaneous sensory receptors in the rat foot. J. Neurophysiol., 69: 16841699. Leem, J.W., Willis, W.D., Weller, S.C. and Chung, J.M. (1993b) Differential activation and classification of cutaneous afferents in the rat. J. Neurophysiol., 70: 241 t-2424. Liu, M.W., Max, M.B., Robinovitz, E., Gracely, R.H. and Bennett, G.J. (1998) The human capsaicin model of allodynia and hyperalgesia: sources of variability and methods for reduction. J. Pain Symptom Manage., 16: 10-20.

Ma, Q.-P. and Woolf, C.J. (1996) Progressive tactile hypersensitivity: an inflammationinduced incremental increase in the excitability of the spinal cord. Pam, 67: 97-106. Magerl, W., Wilk, S.H. and Treede, R.-D. (1998) Secondary hyperalgesia and perceptual wind-up following intradermal injection of capsaicin in humans. Pain, 74: 257-268. Magerl, W., Fuchs, P.N., Meyer, R.A. and Treede, R.-D. (1999) Secondary hyperalgesia to punctate stimuli in humans is mediated by capsaicin-insensitive A-fiber nociceptors. Absrracts, 9th World Congress on Pain, Wien, p. 407. McMahon, S.B., Lewin, G.R. and Wall, P.D. (1993) Central hyperexcitability triggered by noxious inputs. Cur,: Opin. Neurobiol., 3: 602-610. Merskey, H. and Bogduk, N. (1994) Classijcation of Chronic Pain. IASP Press, Seattle, WA. Meyer, R.A., Campbell, J.N. and Raja, S.N. (1994) Peripheral neural mechanisms of nociception. In: P.D. Wall and R. Melzack (Eds.), Textbook of Pam. Churchill Livingstone, Edinburgh, pp. 13-44. Millan, M.J. (1999) The induction of pain: an integrative review. Prog. Neurobiol., 57: 1-164. Moore. K.A., Baba, H. and Woolf, C.J. (2000) Synaptic transmission and plasticity in the superficial dorsal horn. In: J. Sandktihler, B. Bromm and G.F. Gebhart (Eds.), Nervous Sy‘stern Plasticity and Chronic Pam. Progress in Brain Research, Vol. 129. Elsevier, Amsterdam, pp. 63-80. Neumann, S., Doubell, T.P., Leslie, T. and Woolf, C.J. (1996) Inflammatory pain hypersensitivity mediated by phenotypic switch in myelinated primary sensory neurons. Nature, 384: 360-364. Ochoa, J.L. and Yamitsky, D. (1993) Mechanical hyperalgesias in neuropathic pain patients: dynamic and static subtypes. Ann. Neural., 33: 465-472. Park, K.M., Max, M.B., Robinovitz, E., Gracely, R.H. and Bennett, G.J. (1995) Effects of intravenous ketamine, alfentanil, or placebo on pain, pinprick hyperalgesia, and allodynia produced by intradermal capsaicin in human subjects. Pain, 63: 163-172. Pertovaara, A. (1998) A neuronal correlate of secondary hyperalgesia in the rat spinal dorsal horn is submodality selective and facilitated by supraspinal influence. Exp. Neurol., 149: 193-202. Pertovaara, A. (2000) Plasticity in descending pain modulatory systems. In: J. Sandktihler, B. Bromm and G.F. Gebhart (Eds.), Nervous System Plasticity and Chronic Pam. Progress in Brain Research, Vol. 129. Elsevier, Amsterdam, pp. 231-242. Raja, S.N., Campbell, J.N. and Meyer, R.A. (1984) Evidence for different mechanisms of primary and secondary hyperalgesia following heat injury to the glabrous skin. Brain, 107: 1 1791188. Sandktihler, J., Benrath, J., Brechtel, C., Ruscheweyh, R. and Heinke, B. (2000) Synaptic mechanisms of hyperalgesia. In: J. Sandktihler, B. Bromm and G.F. Gebhart (Eds.), Nervous System Plasticity and Chronic Pain. Progress in Brain Research, Vol. 129. Elsevier, Amsterdam, pp. 81-100. Sang, C.N., Hostetter, M.P., Gracely, R.H., Chappell, AS., Schoepp, D.D., Lee, G., Whitcup, S., Caruso, R. and Max,

341 M.B. (1998) AMPA/kainate antagonist LY293558 reduces capsaicin-evoked hyperalgesia but not pain in normal skin in humans. Anesthesiology, 89: 1060-1067. Schmelz, M., Schmidt, R., Ringkamp, M., Forster, C., Handwerker, H.O. and Torebjork, H.E. (1996) Limitation of sensitization to injured parts of receptive fields in human skin C-nociceptors. Exp. Bruin Res., 109: 141-147. Sethna, N.F., Liu, M.W., Gracely, R., Bennett, G.J. and Max, M.B. (1998) Analgesic and cognitive effects of intravenous ketamine-alfentanil combinations versus either drug alone after intradermal capsaicin in normal subjects. An&h. A&g., 86: 1250-1256. Sherman, S.E., Luo, L. and Dostrovsky, J.O. (1997) Altered receptive fields and sensory modalities of rat VPL thalamic neurons during spinal strychnine-induced allodynia. J. Neurophysiol., 78: 2296-2308. Simone, D.A., Sorkin, L.S., Oh, U., Chung, J.M., Owens, C., LaMotte, R.H. and Willis, W.D. (1991) Neurogenic hypemgesia: central neural correlates in responses of spinothalamic tract neurons. J. Neurophysiol., 66: 228-246. Sinclair, D.C. and Hinshaw, J.R. (1950) A comparison of the sensory dissociation produced by procaine and by limb compression. Bruin, 73: 480-498. Svendsen, F., Hole, K. and Tjolsen, A. (2000) Long-term potentiation in single WDR neurons induced by noxious stimulation in intact and spinalized rats. In: J. Sandktihler, B. Bromm and G.F. Gebhart (Eds.), Nervous System Plasticity and Chronic Pain. Progress in Brain Research, Vol. 129. Elsevier, Amsterdam, pp. 153-161. Szolcsanyi, J., Anton, E, Reeh, P.W. and Handwerker, H.O. (1988) Selective excitation by capsaicin of mechano-heat sensitive nociceptors in rat skin. Brain Res., 446: 262-268. Taddese, A., Nah, S.-Y. and McCleskey, E.W. (1995) Selective opioid inhibition of small nociceptive neurons. Science, 270:

1366-1369. Thalhammer, J.G. and LaMotte, R.H. (1982) Spatial properties of nociceptor sensitization following heat injury of the skin. Brain Res., 231: 257-265. Torebjork, H.E., Lundberg, L.E.R. and LaMotte, R.H. (1992) Central changes in processing of mechanoreceptive input in capsaicin-induced secondary hyperalgesia. J. Physiol., 448: 765-780. Treede, R.-D. and Cole, J.D. (1993) Dissociated secondary hyperalgesia in a subject with a large-fibre sensory neuropathy. Pain, 53: 169-174. Treede, R.-D. and Magerl, W. (1995) Modem concepts of pain and hyperalgesia: beyond the polymodal C-nociceptor. News Physiol. Sci., 10: 216-228. Treede, R.-D., Davis, K.D., Campbell, J.N. and Raja, S.N. (1992a) The plasticity of cutaneous hyperalgesia during sympathetic ganglion blockade in patients with neuropathic pain. Brain, 115: 607-621. Treede, R.-D., Meyer, R.A., Raja, S.N. and Campbell, J.N. (1992b) Peripheral and central mechanisms of cutaneous hyperalgesia. Prog. Neurobiol., 38: 397-421. Urban, L., Thompson, S.W.N. and Dray, A. (1994) Modulation of spinal excitability: co-operation between neurokinin and excitatory amino acid neurotransmitters. TINS, 17: 432-438. Woolf, C.J. and Doubell, T.P. (1994) The pathophysiology of chronic pain - increased sensitivity to low threshold A-betafibre inputs. Cum Opin. Neurobiol., 4: 525-534. Woolf, C.J. and Mannion, R.J. (1999) Neuropathic pain: aetiology, symptoms, mechanisms, and management. Luncet, 353: 1959-1964. Ziegler, E.A., Magerl, W.. Meyer, R.A. and Treede, R.-D. (1999) Secondary hyperalgesia to punctate mechanical stimuli: Central sensitization to A-fibre nociceptor input. Brain, 122: 2245-2257.

J. Sandkiihler, B. Bromm and G.F. Gebhart (Eds.) Progress in Brain Research, Vol. 129 0 2000 Elsevier Science B.V. All rights reserved

CHAPTER 25

Referred pain as an indicator for neural plasticity Lars Arendt-Nielsen ‘,*, Renk J. Laursen ’ and Asbjgrn M. Drewes 2 I Laboratory

for

Experimental ‘Department

Pain Research, of Medical

Center for Sensory-Motor Interaction, Aalborg University, Fredrik DK-9220 Aalborg, Denmark Gastroenterology, Aalborg Hospital, DK-9100 Aalborg, Denmark

Introduction Pain and relief of pain are great challenges in medicine. Pain originating from deep somatic structures and viscera represents a major part of patients’ pain complaints. Although deep pain is an important factor in many disorders such as injuries, degenerative diseases, and cancer, the mechanisms underlying muscle and visceral pain are poorly understood. Deep pain is a major clinical problem, and a further insight into the peripheral and central mechanisms is necessary to improve therapy. The study of referred pain from muscles and viscera may help to uncover such mechanisms. The focus of this paper is to discuss the possible mechanisms underlying referred pain from muscles and viscera (see also Coutinho et al., 2000, this volume; Hoheisel and Mense, 2000, this volume). Referred pain has been known and described for more than a century and has been used extensively as a diagnostic tool in the clinic. The Taxonomy Committee of the International Association for the Study of Pain (IASP) has not made a definition of the term. In this paper, the definition ‘pain felt at a site remote from the site of origin/stimulation’ will be used.

* Corresponding author: L. Arendt-Nielsen, Aalborg University, Center for Sensory-Motor Interaction, Laboratory for Experimental Pain Research, Fredrik Bajers Vej 7D-3, DK-9220 Aalborg, Denmark. Fax: +45-98154008; E-mail: [email protected]

Bajers

Vej 70-3,

Several theories regarding appearance of referred pain have been suggested, and basically, they state that nociceptive dorsal horn neurons receive convergent inputs from various tissues, thus higher centers cannot identify the actual input source. New theories, in which plasticity of dorsal horn neurons plays a central role, have replaced the rigidities. In clinical practice, classical signs of viscerally referred pain are: (1) the radiating pain in arms, particularly the left, during angina pectoris; (2) McBurney’s sign indicative of appendicitis; (3) superficial abdominal pain from gastric ulcers; (4) the pain of cholecystitis, which may radiate to the interscapular area, right scapula, shoulder, or back, and (5) pain from renal and urethral stones referred to the lower back. The referral of pain is also well known in deep pain, e.g., arthritis of the shoulder and hip to the areas around the elbow and knee. During the past decades, a systematic attempt to chart referred musculoskeletal pain areas in humans has been made. Some of these findings have been reproduced in experimental muscle pain studies in humans (GravenNielsen et al., 1997b; Laursen et al., 1997a,b). In summary, referred muscle pain (and the possibly related hyperalgesia) is manifested in somatic structures (skin, muscles, joints, and tendons) whereas visceral pain manifests in somatic as well as visceral structures (Fig. 1). These manifestations are of significantly clinical importance for the diagnosis of pain pathologies.

344

Clinical

Clinical

Evoked

Localised

Muscle

they have the disadvantage of involving several/all muscle groups in the region investigated, and often pain from other somatic tissues cannot be excluded. Finally, endogenous methods are not adequate to generate referred pain. In this paper, we will concentrate on exogenous models.

Referred

Referred

Skin

Evoked

-. ‘Skin

Muscle

Fig. I. A schematic figure of manifestation referred pain from muscle and viscera.

Clinical versus experimental pain

Muscle pain by algogenic substances

Viecera of localized

and

studies on referred

Further basic research into all aspects of referred pain is needed. Clinical research and in particular research in pain are often confounded by many factors, which make it difficult to look at certain aspects of the disease. Experimental models, which can be standardized by using healthy subjects, are useful in basic research, because they allow a study without confounding factors (Arendt-Nielsen, 1997). Human experimental pain research involves two separate topics: standardized activation of the nociceptive system, and measurements of the evoked responses (for review see: Bromm and Desmedt, 1995; Arendt-Nielsen, 1997; Bromm and Lorenz, 1998; Bromm et al., 2000, this volume; Treede and Magerl, 2000, this volume). The ultimate goal of advanced human experimental pain research is to obtain a better understanding of mechanisms involved in pain transduction, transmission, and perception under normal and pathophysiological conditions. Hopefully, this can provide better characterization, prevention, and management of pain and give more insight into the mechanisms underlying referred pain. Muscle pain Various methods can be used to induce experimental, referred muscle pain. Usually, the methods are classified into two groups: (1) endogenous (without external stimuli); and (2) exogenous (external stimuli) methods. High response rate characterizes human endogenous methods (e.g., ischemia and exercise), which are suitable to study general pain states. However,

A number of exogenous methods have been used to induce experimental human muscle pain. The most used and accepted method is intramuscular infusion of hypertonic saline (5%). Kellgren and Lewis introduced the method in 1938 (Kellgren, 1938; Lewis, 1938), and since then intramuscular infusion of hypertonic saline has been used extensively (most recently by Graven-Nielsen et al., 1997b). A variety of parameters have been shown to correlate to the infusion of hypertonic saline (e.g., saline concentrations, infusion rate and pressure, and amount of saline infused). In spite of this, the mechanisms responsible for the excitation of pain shortly after the infusion are still unknown; however, a direct excitation of afferents due to osmotic difference has been proposed (Graven-Nielsen et al., 1997~). Referred pain is felt in structures at a distance from the infusion site and appears with a delay of 20 s compared with local pain (Graven-Nielsen et al., 1997a). This pain is characterized as diffuse and unpleasant. Muscle pain by electrical stimulation

Intramuscular electrical stimulation (IMES) of muscle tissue has been used in various experimental and clinical settings. IMES offers the advantage that it can elicit referred muscle pain in an on and off manner. It is an easy method to use, and a high incidence of local (94%) and referred (78%) pain is generated (Laursen et al., 1997a). In our studies on IMES, we used 10 Hz for at least 10 s to generate referred muscle pain. Referred pain after IMES appears with different delays in the various studies (Laursen et al., 1997a,b), ranging from simultaneously with local pain to a delay of 43 s in average. A difference in stimulus intensities could account for the variances of referred pain onset (Laursen et al., 1997a). A more consistent delay of referred pain onset is char-

34.5

acteristic for hypertonic saline experiments (GravenNielsen et al., 1997b). The reason for the difference in time delay between the two models could be due to different excitation mechanisms of the nociceptive afferents and/or due to central mechanisms (temporal summation or hyperexcitability). However, IMES has some shortcomings compared with hypertonic saline as it bypasses the sensory nerve endings making investigations of receptor transduction mechanisms impossible. In general, significantly higher stimulus intensity was required to elicit referred pain compared with local pain, and a significantly positive correlation has been found between the stimulus intensity and local and referred pain intensity ratings (Laursen et al., 1997a). This is in accordance with previous experimental and clinical studies (Inman and Saunders, 1944; Torebjijrk et al., 1984; Graven-Nielsen et al., 1997a,b) and studies using direct, intraneural, electrical stimulation of muscle nociceptive afferents (Torebjiirk et al., 1984; Marchettini et al., 1996). A significant correlation between the size of local and referred pain areas and local and referred sensation/pain intensity ratings was observed (Laursen et al., 1997a). Similar observations have been seen in studies where sequential infusions of hypertonic saline into a muscle resulted in increasing number of subjects experiencing referred pain and increasing areas of referred pain, and where intraneural, electrical stimulation of muscle afferents at constant frequency and intensity evoked an expansion of the projected pain area over time (Marchettini et al., 1996). An increased nociceptive input to the dorsal horn neurons, which generates an expansion of receptive fields (Hu et al., 1992), may account for the expansion of referred areas observed during increased intramuscular stimulation (Laursen et al., 1997a). Manifestation

of referred muscle pain

Sensory manifestations of muscle pain are seen as diffuse aching pain in the muscle, pain referred to distant somatic structures, and modifications in superficial and deep tissue sensibility in the painful areas. These manifestations are different from cutaneous pain, which is normally superficial and localized around the injury with a sharp and burning

quality. Referred pain and sensibility changes in the painful structures have been known for many years, but the mechanisms responsible for these phenomena are not fully understood. Referred muscle pain is probably a central mechanism because it is possible to induce referred pain to limbs with complete sensory loss due to an anesthetic block (Feinstein et al., 1954). However, the lack of peripheral input from the referred pain area decreases the referred pain intensity (Laursen et al., 1997b) suggesting that the peripheral input from the referred pain area is involved, but not a necessary condition for referred pain. Hypothetically, convergence of nociceptive afferents on dorsal horn neurons may mediate referred pain, but studies by Hoheisel and Mense (1990) show a rare convergence between muscle and other deep tissues (muscle, etc.). Central hyperexcitability may facilitate the generation of referred pain. Animal studies show a development of new receptive fields by noxious muscle stimuli (Hoheisel et al., 1993). Recordings from a dorsal horn neuron with a receptive field located in the biceps femoris muscle show new receptive fields in the tibialis anterior muscle and in the foot after i.m. injection of bradykinin into the tibialis anterior muscle (Hoheisel et al., 1993). In the context of referred pain, the unmasking of new receptive fields could mediate referred pain due to central hyperexcitability (Mense, 1994). This has been suggested to be the phenomenon of secondary hyperalgesia in deep tissue. Similar findings are seen in humans having received an intradermal injection of capsaicin, and a rapid development of the central hyperexcitability (seen as secondary cutaneous hyperalgesia) is found. The time needed for unmasking may account for the time delay between local pain and the development of referred pain, and for the increased number of subjects developing referred pain during repeated hypertonic saline infusions or tonic infusion. A number of studies have found that the area of the referred pain correlated with the intensity (Graven-Nielsen et al., 1997b; Laursen et al., 1997a) and duration (Marchettini et al., 1996) of the muscle pain, which parallels the observations for cutaneous secondary hyperalgesia. It is obvious to propose that muscle pain conditions (Mense, 1994) can evoke central hyperexcitability. The relation between temporal summation and central hyperexcitability may

346

be indicated by the progressive increase in referred pain area during tonic i.m. infusion of hypertonic saline. Similar findings on temporal summation and central hyperexcitability are also seen for stimulation of viscera (see next section). From studies on cutaneous hyperalgesia, central summation of nociceptive input from muscles and referred pain areas is expected to be exaggerated in musculoskeletal pain conditions if central hyperexcitability is involved. Infusions of hypertonic saline have shown larger referred pain areas in fibromyalgia patients than in control subjects, and also proximal referral of pain was found in the patients, but not in the control subjects (Sorensen et al., 1998). This may reflect central hyperexcitability in fibromyalgia patients as hypertonic saline is infused into muscles with no clinical muscle pain (Sorensen et al., 1998). Moreover, the gain of temporal summation was increased in fibromyalgia patients as the pain threshold for repeated i.m. electrical stimulation and not single stimulation was decreased in fibromyalgia patients compared with control subjects (Sorensen et al., 1998). In a recent study, a similar manifestation of enlarged referred pain areas to intramuscular injection of hypertonic saline was found in chronic pain patients after whiplash injuries (Johansen et al., 1999). Such enlarged referred pain areas are also seen after visceral stimulation in patients with chronic visceral pain (see next section).

These modality specific somatosensory changes found in the referred muscle area are similar to findings in secondary hyperalgesic areas of the skin. The mechanisms behind sensibility changes may be of peripheral or central origin. Infiltration of the muscle tissue with anesthetics 30 min after injection of hypertonic saline completely reverses the cutaneous and muscular hyperalgesia (Vecchiet et al., 1988). The effect of a peripheral block on the hyperalgesia (Vecchiet et al., 1988) suggests that the hyperalgesia is caused by maintained peripheral input, which is also a necessary condition for referred pain. Alternatively, the mechanisms responsible for deep and cutaneous hyperalgesia after muscle pain may be caused by central hyperexcitability. Central hyperexcitability of dorsal horn neurons caused by the muscle pain may explain the expansion of pain with referral to other areas and probably also hyperalgesia in these areas. However, facilitated neurons cannot account for the decreased sensation to certain sensory stimuli in the referred area. Descending inhibitory control of the dorsal horn neurons may explain the decreased response to additional noxious stimuli in the referred pain area. Recently, we found that saline-induced muscle pain gave reason to deep tissue hypoalgesia in extra segmental areas (including the area of referred pain) distant from the pain focus (Graven-Nielsen et al., 1998). Alternatively, segmental inhibition may contribute to the decreased sensibility.

Hyperalgesia related to referred muscle pain

Visceral pain

The somatosensory sensibility in the referred pain area may give additional information about the mechanisms involved in generation of referred pain. In general, it is accepted that muscle pain can result in hyperalgesia in the referred somatic structures (Fig. lj. The somatosensory sensibility is affected by saline-induced muscle pain in cutaneous and deep structures in the area of local and referred pain. During saline-induced pain, the deep tissue sensibility may increase, decrease, or remain unaffected in the referred muscle pain area. Increased visual analog scale (VAS) response to electrical cutaneous stimulation and decreased sensibility to radiant heat stimulation has been reported.

Visceral pain is the most common form of pain experienced in medical practice. Correspondingly, characterization of pain in gastroenterology has been reported to be a very important aid in the diagnosis and assessment of organ dysfunction, Accordingly, abdominal pain and discomfort are among the most common symptoms responsible for patients seeking the care of a gastroenterologist. Visceral pain differs from cutaneous and muscle pain in many aspects: (1) the pain probably does not arise in all viscera, and localized stimuli sufficient to produce tissue damage (such as plucking endoscopic biopsy specimens and applying heat probes) are not always painful; (2) the pain appears to be non-specific regarding the nature of the stimulus; (3)

347

visceral pain is diffuse with poor deep localization and correspondingly, the verbal descriptors include every subgroup of the McGill Pain Questionnaire; (4) referred pain to somatic structures is often found in various parts of the body away from the diseased organ, and hyperalgesia in the referred pain area has been described; and finally, (5) visceral pain is generally accompanied by larger, emotional, autonomic, and motor responses than cutaneous pain. Despite its greater clinical importance, visceral pain has not been as extensively investigated and is less understood than cutaneous pain (see also Coutinho et al., 2000, this volume). One major problem is the limited access to the organs, which makes visual inspection of the eventual lesion and motor responses difficult. Several experimental procedures have been used to elicit referred visceral pain, although most methods have been designed for animals. Different criteria have been used for the ideal visceral pain model. Many researchers have focused on noxious stimuli assumed to be natural such as distension of the gut. Moreover, the stimulus must be minimally invasive, reliable in test-retest experiments, and quantifiable. Normally, the duration of an experimental stimulus is limited to a few seconds or minutes. In contrast, the pain in the majority of diseases lasts much longer, giving sufficient time for peripheral and central mechanisms to be involved. It should, therefore, be discussed whether any experimental study can imitate all aspects of chronic pain. In most gastrointestinal diseases, several mechanisms such as inflammation, distension, ischemia, and chemical transmitters interplay in the disease process, and the pain is probably a result of many components working together. Therefore, as the different experimental stimuli described below may evoke various pain responses, probably only a combination of these can mimic the clinical aspects of referred pain. Visceral sensations by thermal stimuli

Only few studies have used thermal stimuli to elicit visceral pain. Old studies have stimulated the esophagus with water in four subjects. Most researchers reported the cold water as a cold sensation, whereas warm water was experienced as cold by three of the four subjects. Water-filled bags have been introduced

into patients with a gastrostomy. Only temperatures outside 18-40°C produced referred sensations (as cold or warm). Visceral pain by chemical stimuli

Various substances such as hypertonic saline, acidic, basic, and salt solutions have been used in the past. Intraperitoneal administration of bradykinin in humans caused abdominal referred pain in approximately 80% of the subjects. Other human studies have applied algogenic substances to the mucosal surfaces in patients with a pre-existing gastro- or colostomy and found that these chemicals produced pain when the mucosa was inflamed. No pain was, however, seen following the same procedures, when the mucosa appeared to be normal (Ness and Gebhart, 1990). The utility of chemical models is limited, but improved models may be important in future research. Visceral pain by mechanical stimuli

Most previous models for experimental pain stimulation of the colon were based on balloon distension using either fixed-volume or fixed-pressure modalities. One of the advantages of these models is that the stimulus is assumed to be natural, as the pain in many diseases is probably released by distension of the gut. Although it should be emphasized that distension as a pain stimulus of the gut is of great value, several limitations exist: (1) the stimulated gut area is rather large and poorly defined, and more localized diseases such as peptic ulcers are badly mimicked; and (2) the distension may often only elicit discomfort as pain is not always reported at pressures acceptable in the human situation (Ness et al., 1990). The barostat, which allows simultaneous acquisition of volume and pressure in a balloon placed inside the gut, has been widely used in experiments of reflexes and muscular properties of the gut. The barostat has been used in several studies of pathophysiology, especially in functional bowel disorders, where specific visceral hyperalgesia (e.g., hyperalgesia to visceral, but not somatic stimuli) was shown in a large proportion of patients with the irritable bowel syndrome (IBS) (Mertz et al., 1995). The

348

inflation rate of the barostat balloon can be selected to give tonic or phasic distension, and especially regimes, such as repeated phasic distension, have been shown suitable to elicit rectal hyperalgesia in patients with IBS and thus discriminate the sensation from that in healthy control subjects. Besides rectal hyperalgesia, increased perception to distension has been shown in most other parts of the gut in patients with IBS, demonstrating the general visceral hypersensitivity in these patients. The importance of central mechanisms in functional gut disorders was recently shown by Munakata et al. (1997) where all patients developed rectal hyperalgesia following repetitive sigmoid distension despite great heterogeneity in baseline rectal sensitivity. Such induction of hyperalgesia in one area of the gut following conditioning of the more proximal parts most probably results from central sensitization. In patients with organic diseases, however, the distension studies were less consistent. Hyperalgesia to rectal distension was, for example, seen in patients with active ulcerative colitis and solitary rectal ulcer. In patients with Crohn’s disease and inflammation in the ileum, hypoalgesia to stimuli remote from the inflammation (rectum) was seen together with atypical referral pain areas (Bernstein et al., 1996). In patients with esophagitis due to gastroesophageal reflux disease, however, increased sensibility only to chemical (acid perfusion), but not mechanical (distension) stimuli of the esophagus, was seen (Fass et al., 1998). In the same study, the treatment resulted in macroscopic normalization of tissue injury in most patients, but the enhanced chemosensitivity persisted. Whether long-standing inflammation can cause persistent changes in the visceral nervous system is, however, not settled. Visceral pain by electrical stimuli

Many animal experiments used electrical stimuli directly on visceral nerves. The stimulation was clearly noxious in character, and behavioral as well as autonomic responses were elicited (Ness and Gebhart, 1990). Furthermore, stimulation of single visceral nerves can lead to excitation of neurons in widespread segments of the spinal cord, and such studies were important for the classification of viscereceptive neurons by their cutaneous receptive fields,

as well as viscera-visceral and viscera-somatic reflexes. In humans, electrical stimulation of the abdominal nerves and viscera has been used in the past (Ness and Gebhart, 1990), although the methods described in more recent papers have further refined the techniques (Frobert et al., 1995; Drewes et al., 1997, 1999a) (Fig. 2). In our experience a single constant current pulse is not adequate to evoke referred visceral pain. Only repetitive (5 pulses, 2 Hz) or continuous (2-4 s, 4 Hz) pulses should be used (Drewes et al., 1997). It is open for discussion whether the discomfort and pain elicited during electrical stimulation and other methods are comparable. Manifestation of referred visceral pain

Information on localization of visceral pain is an important guideline to the clinical diagnosis. Referred pain tends to be localized in areas corresponding to the dermatome and myotome that are supplied by the affected viscus (referred visceral pain), or remote from the inciting stimulus in an area supplied by the same neural segment as the injured organ (referred parietal pain). Viscerotomes have been defined by patterns of evoked or pathological, referred pain and hyperalgesia described in the clinical literature. The clinical picture is, however, not simple and often depends on whether visceral or somatoparietal impulses predominate. In patients with peptic ulcer, for example, the size and localization of the ulcer frequently differ from the clinical presumptions. Experimental models are, therefore, needed as they may help to clarify such aspects. Clinical pain following diseases in the stomach and duodenum is typically described focally in the mid-epigastrium. Correspondingly, experimental studies using distension of the stomach and duodenum were reported to give local discomfort in the abdomen with referral to the epigastrium or upper quadrants (Mertz et al., 1998). It is well known, however, that the pain is often atypical, especially with respect to the somatic referral (Haubrich, 1995). In our study (Drewes et al., 1997) with twelve healthy subjects, the following four local areas of the stomach mucosa were stimulated with electrical stimuli:

349

Fig. 2. Schematic illustration pair of biopsy forceps, which delivered.

of the experimental set-up can under visual inspection

for evoking be placed

the prepyloric region, the greater and lesser curvature, and the duodenal bulb. Pain was felt deep in the abdomen and was referred to very different areas of the abdominal wall, the back, and the retrosternal region. Such different manifestations of referred stomach pain correspond very well with the clinical manifestations of local diseases. In the colon, the sites of referred pain elicited by distension have previously been mapped in healthy subjects, although more recent studies have focused on the rectum and sigmoid colon, being more easily reached in distension models. Pain referred to the mid-abdomen and lower quadrants was mostly reported in the previous studies. In the clinical literature pain in the middle and lower parts of the abdominal wall was reported following diseases of the colon, although individual differences have frequently been described. In our study (Arendt-Nielsen et al., 1997), repeated burst electrical stimuli were applied to the mucosa of the colon in seven subjects with a stoma. All reported local pain deep in the abdomen and five had referred pain. A shift in the electrode position, however, resulted in marked differences for the direction of the referred pain, a

referred visceral pain from on the mucosa, and single,

the colon. The electrode repeated, or continuous

is a modified stimuli can be

finding that emphasizes the importance of the electrode position. Continuous pain stimuli resulted in an increasing area of referred pain, localized away from the stoma1 site, and in one patient the referred pain shifted from the abdominal area to the back. In the subsequent study (Drewes et al., 1999a) with eleven healthy subjects, we stimulated the intact colon with repeated burst stimuli during colonoscopy at the cecum, the hepatic and splenic flexures, and the rectosigmoid junction. Pain was reported in all subjects following successive increase in stimulus intensity and was described as going deep into the abdomen with very individual referral to somatic structures. Stimuli applied at the left side of the colon were referred to the periumbilical area and lower abdominal wall in agreement with distension studies and clinical observations. Stimuli at the right colon, however. resulted in pain referred to the upper- and mid-abdomen and left side of the abdominal wall. This is, however, in accordance with the embryonic development: segments of the embryonic hindgut that becomes the large intestine were originally at the lower end, together with the rest of the neural tract. Only later in fetal development, the

350 colon does rotate up and around so that its middle portion occupies the upper abdomen. Therefore, nerve fibers supplying the right colon mucosa may partly converge with somatic afferents from the left side of the abdominal wall. Hyperalgesia related to referred visceral pain

In general, referred visceral pain can result in hyperalgesia in the somatic areas, but also in other visceral structures (Fig. 1). In the clinical literature on referred visceral pain, hyperalgesia and allodynia in the referred pain areas have been described (Ness and Gebhart, 1990), and this was confirmed in animal studies (Ness and Gebhart, 1990; see Coutinho et al., 2000, this volume). Visceral hyperalgesia probably includes both peripheral and central components. Afferent fibers innervating the gut can be sensitized by endogenous chemicals, released or synthesized after tissue injury. This results in the development of hyperalgesia with decreased excitation thresholds of both low-threshold and high-threshold visceral afferents. In addition, recruitment of silent nociceptors may take place. In clinical studies, evidence for deep (mostly) and superficial hyperalgesia in somatic tissues was seen in patients with colics due to ureter stones. Moreover, trophic changes in the parietal tissues were seen (Giamberardino, 1999). These changes were still detectable in the patients long (months or years) after the original episode and confirmed in subsequent studies where the stones were removed by extracorporeal shock-wave lithotripsy (Giamberardino et al., 1994). This central hyperexcitability might share some features with cutaneous secondary hyperalgesia after cutaneous C-fiber stimulation. The degree and spread of hyperalgesia depend on the duration and hence central summation of visceral stimulation. In healthy subjects, Ness and Gebhart (1990) performed a repeated 30-s colonic distension ten times, separated by 4 min. They observed that both the pain intensity and the referred pain area increased during the sequential stimulation. The study has recently been replicated for repetitive bladder fillings (Ness et al., 1998) where as well the areas of referred pain as the evoked pain progressively increased for consecutive stimuli. Moreover, repeated balloon distension of the

stomach (Mertz et al., 1998) were shown to increase the somatic, referred pain area in healthy subjects, thus giving evidence for central neuronal excitability changes, probably mainly at the spinal cord level. These findings are consistent with animal experiments where repetitive distension of the gallbladder induced hyperpolarization of dorsal horn neurons. The increased pain and referred pain areas to repetitive stimulation could indicate that the repeated stimuli may evoke mechanisms related to central hyperexcitability. From human experimental studies, the NMDA-antagonist ketamine is known very efficiently to inhibit temporal summation, but it does not affect pain evoked by a single stimulus. This assumption is supported by recent animal studies in which NMDA-antagonists have proven effective in inhibiting visceral nociceptive mechanisms. We found (Arendt-Nielsen et al., 1997) that continuous, electrical stimulation from 30 to 120 s of the gut caused a progressive increase in the referred pain area as the stimulus time was increased again indicating a very dynamic character of the referred area. Often the subjects can hardly perceive a single stimulus whereas repeated stimuli are much more potent in generating pain from viscera. Previously, temporal summation was demonstrated to exacerbate pain using electrical stimuli in different areas of the esophagus (Frobert et al., 1995) and the gut (Arendt-Nielsen et al., 1997; Drewes et al., 1997). Afferent fibers repeatedly stimulated give a progressive increase in second-order neuronal responsiveness. In turn, this activates secondary messengers and presynaptic transmitter release, leading to positive feedback loops and to increased connectivity of the dorsal horn (central hyperexcitability) as seen for intensive excitation of muscle and cutaneous nociceptive afferents (Hoheisel et al., 1993). Central hyperexcitability may explain many aspects of chronic pain in inflammatory as well as functional visceral disorders. However, evidence for many of these mechanisms is not available in human visceral pain, but it is likely to exist. Recently, Munakata et al. (1997) demonstrated that repetitive sigmoid balloon distension in patients with IBS results in hyperalgesia and increased viscera-somatic referral compared with control subjects. Another older study reported increased and atypical areas of referred pain to visceral stimulation in patients

351

with irritable bowel syndrome compared with control subjects. Therefore, a central hyperexcitability may affect the processing of sensory input from the rectum in these patients. Mertz et al. (1998) showed increases in the referred areas to gastric distension in patients suffering from functional dyspepsia. In half of these patients, the referral was reported in aberrant locations. The abnormalities reported above are thus suggestive of altered central processing of the afferent visceral information in patients with functional gut disorders. This is similar to findings of increased areas of referred pain and increased gain of temporal summation of muscle pain in chronic fibromyalgia (Sorensen et al., 1998) and whiplash (Johansen et al., 1999) patients. Only one paper studied the referred pain pattern during experimental pain stimuli in organic diseases. Mertz et al. (1998) compared the response to tonic stimulus of the fundus in 10 patients with different organic diseases in the stomach to 1.5 healthy controls. Although they only found aberrant referral in 20% of the patients, the results indicate common mechanisms in organic and functional gut disorders, and more studies addressing pain mechanisms in different organic diseases are highly recommended. In systematic, clinical studies of patients with calculosis of the upper urinary tract, hyperalgesia in cutaneous and muscular tissue was seen in the referred pain area, with normalization after stone elimination (Giamberardino et al., 1994). As sensitivity changes in the referred pain area may increase our information of neuronal mechanisms for referred pain, human experimental studies may also be valuable. In the skin and other deep tissues such as the muscles, experimental pain was followed by sensory changes in the area of referred pain (Graven-Nielsen et al., 1997b), but to the authors’ knowledge, the study conducted by our group (Drewes et al., 1999b) was the first to investigate the sensitivity in the referred pain area following experimental visceral pain in man. In this study, the sensitivity in the referred pain area to electrical and mechanical stimuli as well as heat was tested during painful, continuous, electrical stimulation in the prepyloric area of the stomach. Hyperalgesia to heat was demonstrated, whereas the sensitivity to the other stimuli was unchanged. This modality specific somatosensory hyperalgesia is also seen in referred muscle pain areas and emphasizes

the importance of employing a multimodal sensory testing approach when these aspects are studied. As stated in the section discussing referred muscle pain, alterations in descending inhibitory control systems are the most probable explanation for the finding. Although most studies have addressed viscerosomatic hyperalgesia, animal studies have given evidence for viscera-visceral convergence between different organs (Giamberardino, 1999). Correspondingly, the Italian group showed that women with urinary calculosis had enhanced pain from the urinary tract in the menstrual cycle where pelvic congestion was suggested. Moreover, this pain was exaggerated in subjects with dysmenorrhea (Giamberardino, 1999). Although only anecdotally described in clinical studies, patients with pain due to organic diseases may report increased sensitivity to different clinical procedures, and the subjects await further studies. Modulation

of referred pain

During the last century, several theories on the origin of referred pain have been suggested (see next section). In order to illuminate possible mechanisms of referred pain, a number of case reports and experiments have been published on the effect of anesthetizing the referred pain area. In most cases, referred visceral pain has been investigated, and contradictory results have been demonstrated. Old studies have found that patients suffering from various diseases (e.g., angina pectoris, pleuritis, stomach ulcer, chronic cholecystitis, salpingitis, and kidney stones) experienced pain at structures (most often the skin) located at a distance from the affected organ(s), which could be partially, and in some cases, completely abolished by infiltrating the area with a local anesthetic. Similar findings were demonstrated in other studies where experimental pain (pressure stimulation of the diaphragm; distension of duodenum by a balloon; injection of hypertonic saline) produced pain referred to the skin at distant sites, which was abolished with the injection of local anesthetics. In most cases the referred pain returned once the anesthetic effect had diminished. None of these studies have been placebo controlled. On the contrary, a number of clinical cases and studies have not been able to demonstrate any effect of anesthetizing the referred pain area.

352 Several explanations have been offered to the divergent results obtained when an area of referred pain is anesthetized. (1) The variation in the number of structures (skin, subcutis, fascia, muscle, tendons, ligaments, and bone) anesthetized. This is probably a major source of error because referred pain areas, and especially visceral referred pain, tend to be located in the deep tissues where complete anesthesia of a referred pain area is difficult. (2) The duration and level of pain at the primary site. (3) The site of the primary pain (skin, viscera, and deep structures). (4) Whether sensory changes (hypersensitivity) take place at the referred pain site. Referral of visceral pain has divided referred pain into: (a) referred pain without hyperalgesia where no effect of anesthetizing the referred pain area is observed; and (b) referred pain with hyperalgesia (true parietal pain) where anesthetizing the referred pain area results in a cease or a considerably diminishing effect (Giamberardino and Vecchiet, 1996). In order to determine if it is possible to anesthetize referred muscle pain, we performed two experiments (Laursen et al., 1997b). Due to the obvious problems of blinding a study involving compression-ischemia nerve blocks and local anesthetics, it was decided initially to use a non-invasive technique to anesthetize the skin area of referred pain (Laursen et al., 1997b). The main purpose of the experiment was to quantify a potential reduction of referred pain induced by IMES. In the next experiment, the effect of a differential and complete nerve block of afferents from the referred pain area was tested. In a placebo-controlled experiment (Laursen et al., 1997b), an eutectic mixture of lidocaine and prilocaine was applied to the skin lying over the referred pain area for 90 min to quantify the skin component of referred muscle pain. A cream was chosen instead of injection as this method minimized the spread of the drug to subcutaneous tissues. A 22.7% reduction of the referred pain intensity was observed in the local anesthetic group compared with the placebo group. A similar result has been reported in one case where another technique of skin anesthetization was used. Thus these observations, where the superficial part of the referred pain area has been anesthetized, seem to suggest that referred pain is, to some degree, dependent on spontaneous input from cutaneous receptors.

To completely block all afferents from the referred pain area we combined two techniques: (1) differential nerve blocking with an inflated tourniquet between the site of stimulation and the corresponding distal referred area (Laursen et al., 1999b); and (2) intravenous regional analgesia (IVRA). The referred pain intensity was investigated during 60 min of progressive nerve blocking (Laursen et al., 1999a). Referred pain was elicited by IMES for 10 s with 5-min intervals throughout the experiments. Even though referred pain intensity was reduced by 40.2% while myelinated nerve fiber function was abolished, an additional reduction was not observed when the limb containing the referred pain area was completely anesthetized (myelinated and unmyelinated nerve fiber block) (Laursen et al., 1999b). This substantial observation suggests that referred pain has a peripheral component associated with intact myelinated nerve fiber function. Although human skin nociceptors are known not to have any resting activity, a reduction of activity from other skin receptors (e.g., thermal receptors and possibly low-threshold mechano-receptors) could explain the reduced referred pain in the skin anesthetic study (Laursen et al., 1997b). A clear indication that spinal and/or supraspinal mechanisms were involved in the appearance of referred pain was observed in the complete nerve block study (Laursen et al., 1999b), where an additional reduction of referred pain was observed; nevertheless, referred pain was still perceived. Several cases, in which the afferents supplying a limb have been completely blocked, have reported similar findings. In an old study, two patients with angina1 pain, which in part was referred to a left amputated arm, experienced moderate/complete relief of their referred pain, when the left brachial plexus was anesthetized. Therefore, there seems to be evidence that referred pain somewhat depends on an intact peripheral nervous system with some spontaneous input. Neurophysiological

mechanisms for referred pain

The mechanisms responsible for referred pain referral to adjacent anatomical segments are not known in detail. Several theories have been suggested and will briefly be summarized. These theories do not apply

353

Models for Referred Pain

Fig. 3. The different, possible mechanisms indicate connectivity changes in the dorsal Selzer and Spencer (1969).

of referred pain. Dorsal horn neurons are shown as open circles, and the shaded circles horn. The letters A-E refer to the explanation in the text. Part of the figure is modified from

equally well to referred muscle and viscera pain, and the basic mechanisms governing the two phenomena are to some extent different. The convergent-projection

theory

Ruth (1979) proposed that afferent fibers of different origin converge onto common spinal neurons (Fig. 3A). The core of this suggestion is that the nociceptive activity from the spinal cord is misapprehended as originating from other structures. This could explain the segmental nature of referred muscle pain and the increased referred pain intensity recorded when local muscle pain was intensified (Arendt-Nielsen et al., 1997; Drewes et al., 1997; Graven-Nielsen et al., 1997b; Laursen et al., 1997a). However, it does not explain the delay in the development of referred pain following local pain (Laursen et al., 1997a). In addition, referred pain has not been demonstrated to be a bi-directional phenomenon (e.g., muscle pain in the anterior tibia1 muscle produces pain in the ventral part of the ankle, but the opposite condition has not been demonstrated). Finally, the threshold for eliciting local and referred muscle pain is different (Torebjork et al., 1984; Laursen et al., 1997a).

The convergence-facilitation

theory

MacKenzie (1893) believed that viscera were totally insensitive and that non-nociceptive afferent input to the spinal cord created an irritable focus in the spinal cord (Fig. 3B). This focus would make input of other somatic natures appear in an abnormal fashion and in some cases even be perceived as referred pain. The theory was not recognized, mainly because it did not accept true visceral pain. In recent years, however, MacKenzie’s simple idea of an irritable focus has reclaimed awareness under another term, i.e., central sensitization. The somatosensory sensibility changes reported in referred pain areas could in part be explained by similar mechanisms in the dorsal horn neurons, and the delay in appearance of referred pain demonstrated in various studies could also be explained since the creation of central sensitization may take time. The axon-reflex theory

Bifurcation of afferents from two different tissues has been suggested as an explanation for referred pain (Fig. 3C) (Sinclair et al., 1948). Although, bifurcation of nociceptive afferents from different

354

tissues (muscle and skin and intervertebral discs and skin) exists, it is generally agreed that these types of neurons are rare. Moreover, a time delay in the appearance of referred pain, different thresholds for eliciting local and referred muscle pain, and somatosensory sensibility changes in the referred pain area cannot be explained by this theory. The thalamic-convergence theory

Theobald (1949) suggested that referred pain appeared as a summation of input from the injured area and the referred pain area within neurons in the brain, and not in the spinal cord (Fig. 3E). A recent study of referred pain in monkeys applying computer simulations has demonstrated several pathways, which converge on different cortical and sub-cortical neurons (Apkarian et al., 1995). Numerous experimental and clinical studies (see section on modulation of referred pain) have documented an effect of anesthetizing the area of referred pain, and therefore, referred pain may likely not be explained solely by a central mechanism. The above-mentioned theories lack some of the referred pain characteristics previously described in this chapter. Recently, Mense (1994) suggested an interesting theory, especially from a referred muscle pain point of view, which is known as the centralhyperexcitability theory (Fig. 3D). Recordings from a dorsal horn neuron in animals have revealed that noxious stimuli to a receptive field in a muscle generated within minutes receptive fields at a distance from the original receptive field (Hoheisel et al., 1993). The appearance of two new receptive fields could indicate that latent convergent afferents on the dorsal horn neuron are opened by noxious stimuli appearing from muscle tissue (Mense, 1994), and this facilitation of latent convergence connections could appear as referred pain. Recent observations from the same group have demonstrated that substance-P released from the terminal ends of primary afferents plays a role in the connectivity in the dorsal horn. Furthermore, an expansion of the receptive fields proximal to the existing receptive field was found in a study where experimental myositis is induced, and afterwards application of antagonists to three different neurokinin receptors is effective in preventing the induced hy-

perexcitability (Hoheisel et al., 1997). The idea of this theory falls in line with several of the characteristics of referred muscle pain (dependency on stimulus and a delay in appearance of referred pain compared with local pain). However, the proximal appearance of receptive fields, thought of as referred pain, is in contrast to the reports from a majority of the experimentally referred pain studies including healthy subjects. Clinical studies looking at the spread of experimentally induced referred pain in patients suffering from whiplash syndrome and fibromyalgia have demonstrated proximal as well as distal referral of pain (Johansen et al., 1999; GravenNielsen et al., 2000). We have never seen proximal spread of referred muscle pain in healthy volunteers. A possible explanation to the divergence in these observations could be that an already ongoing pain is necessary to induce a state of hyperexcitability in the spinal cord resulting in proximal and distal referral compared with the one-sided distal referral in healthy subjects. The hyperexcitability theory (Mense, 1994) is based on animal studies where receptive fields appeared within minutes. This does not fit exactly with the development of referred pain in humans, which occurs within seconds. We believe, however, that the idea of latent connections between dorsal horn neurons is convincing. An intact connection between the dorsal horn neurons and the potential receptive field/referred pain area seems to be mandatory if they are to appear with full magnitude. In order to explain the referred pain, which could not be anesthetized, supraspinal mechanisms that could mimic the mechanisms seen in the dorsal horn region cannot be excluded. If the processing of local and the referred pain is not done in the same supraspinal pathways and centers, neuro-imaging techniques (positron emission tomography and functional magnetic resonant imaging) will possibly be able to visualize the nociceptive processing of referred pain in humans. Conclusions Referred pain from muscles and visceral shares some common features: (1) the size of referred pain is related to the intensity and duration of ongoing/evoked pain;

355

(2) temporal summation is a potent mechanism for the generation of referred muscle and visceral pain; (3) central hyperexcitability is important for the extent of referred pain; (4) patients with chronic musculoskeletal and visceral pains have enlarged referred pain areas to experimental stimuli; proximal spread of referred muscle pain is only seen in patients with chronic musculoskeletal pain; (5) modality-specific somatosensory changes occur in referred areas emphasizing the importance of using a multimodal sensory test regime. Human experimental pain research has provided new possibilities to study referred pain quantitatively in volunteers and patients. Clinical studies and pharmacological modulation of experimentally induced referred pain may contribute with additional information concerning the underlying mechanisms. Better characterization and understanding of referred pain mechanisms and related hyperalgesia may help to optimize and rationalize pain management. Acknowledgements The Danish National Research Foundation is acknowledged for supporting the time spent to write this paper. References Apkarian, A.V., Brtiggemann, J., Shi, T. and Airapetiam, L.R. (1995) A thalamic model for true and referred visceral pain. In: G.F. Gebhart (Ed.), Visceral pain. Progress in Pain Research and Management, Vol. 5. IASP Press, Seattle, WA, pp. 217-259. Arendt-Nielsen, L. (1997) Induction and assessment of experimental pain from human skin, muscle and viscera. In: T.S. Jensen, J.A. Turner and Z. Wiesenfeld-Hallin (Eds.), Proceedings of the 8th World Congress on Pain. IASP Press, Seattle, WA, pp. 393-425. Arendt-Nielsen, L., Drewes, A.M., Hansen, J.B. and Tage Jensen, U. (1997) Gut pain reactions in man: an experimental investigation using short and long duration transmucosal electrical stimulation. Pain, 69: 255-262. Bernstein, C.N., Niazi, N., Robert, M., Mertz, H., Kodner, A., Munakata, J., Naliboff, B. and Mayer, E.A. (1996) Rectal afferent function in patients with inflammatory and functional intestinal disorders. Pain, 66: 151-161, Bromm, B. and Desmedt, J.E. (1995) Pain and the Brain -

from Nociception to Cognition. Advances in Pain Research and Therapy, Vol. 22. Raven Press, New York. Bromm, B. and Lorenz, J. (1998) Neurophysiological evaluation of pain. Electroencephalogl: Clin. Neurophysiol., 107: 227253. Bromm, B., Scharein, E. and Vahle-Hinz, C. (2000) Cortex areas involved in the processing of normal and altered pain. In: J. Sandktihler, B. Bromm and G.F. Gebhart (Eds.), Nervous System Plasticity and Chronic Pain. Progress in Brain Research, Vol. 129. Elsevier, Amsterdam, pp. 289-302. Coutinho, S.V., Su, X., Sengupta, J.N. and Gebhart, G.F. (2000) Role of sensitized pelvic nerve afferents from the inflamed rat colon in the maintenance of visceral hyperalgesia. In: J. Sandkiihler, B. Bromm and G.F. Gebhart (Eds.), Nervous System Plasticity and Chronic Pain. Progress in Brain Research, Vol. 129. Elsevier, Amsterdam, pp. 375-387. Drewes, A.M., Arendt-Nielsen, L., Jensen, J.H., Hansen, J.B., Krarup, H.B. and Tage-Jensen, U. (1997) Experimental pain in the stomach: a model based on electrical stimulation guided by gastroscopy. Gut, 41: 753-757. Drewes, A.M., Petersen, P., Nielsen, J. and Arendt-Nielsen, L. (1999a) An experimental pain model based on electrical stimulation of the colon mucosa. &and. L Gastroenterol., 34: 765-771. Drewes, A.M., Krarup, H.B., Hansen, J.B., Tage-Jensen, U. and Arendt-Nielsen, L. (1999b) Pain evoked by electrical stimulation of the prepyloric region of the stomach: cutaneous sensibility changes in the referred pain area. Pain Res. Munage., 4: 131-137. Fass, R., Naliboff, B., Higa, L., Johnson, C., Kodler, A., Munakata, J., Ngo, J. and Mayer, E.A. (1998) Differential effect of long-term esophageal acid exposure on mechanosensitivity and chemosensitivity in humans, Gastroenterology, 115: 1363-1373. Feinstein, B., Langton, J.N.K., Jameson, R.M. and Schiller, F. (1954) Experiments on pain referred from deep somatic tissues. J. Bone Joint Surg., 36: 98 l-997. Frobert, O., Arendt-Nielsen, L., Bak, P, Funch-Jensen, P. and Bagger, J.P. (1995) Oesophageal sensation assessed by electrical stimuli and brain evoked potentials - a new model for visceral nociception. Gut, 37: 603-609. Giamberardino, M.A. (1999) Recent and forgotten aspects of visceral pain. Eul: .I. Pain, 3: 77-92. Giamberardino, M.A. and Vecchiet, L. (1996) Pathophysiology of visceral pain. Curx Rev. Pain, 1: 23-33. Giamberardino, M.A., De Bigontina, P., Martegiani, C. and Vecchiet, L. (1994) Effects of extracorporeal shock-wave lithotripsy on referred hyperalgesia from renal/ureteral calculosis. Pain, 56: 77-83. Graven-Nielsen, T., Arendt-Nielsen, L., Svensson, P. and Jensen, T.S. (1997a) Quantification of local and referred muscle pain in humans after sequential i.m. injections of hypertonic saline. Pain, 69: 111-117. Graven-Nielsen, T., Arendt-Nielsen, L., Svensson, P. and Jensen, T.S. (1997b) Stimulus-response functions in areas with experimentally induced referred muscle pain - a psychophysical study. Bruin Res., 744: 121-128.

356 Graven-Nielsen, T., McArdle, A., Phoenix, J., Arendt-Nielsen, L., Jensen, T.S., Jackson, M.J. and Edwards, R.H. (1997~) In vivo model of muscle pain: quantification of intramuscular chemical, electrical, and pressure changes associated with saline-induced muscle pain in humans. Pain, 69: 137-143. Graven-Nielsen, T., Babenko, V., Svensson, P. and ArendtNielsen, L. (1998) Experimentally induced muscle pain induces hypoalgesia in heterotopic deep tissues, but not in homotopic deep tissues. Bruin Rex, 787: 203-210. Graven-Nielsen, T., Aspegren-Kendall, S., Hem&son, KG., Bengtsson, M., Sorensen, J., Johnson, A., Gerdle, B. and Arendt-Nielsen, L. (2000) Ketamine attenuates experimental referred muscle pain and temporal summation in fibromyalgia patients. 9th World Congress on Pain, Vienna, Austria (in press). Haubrich, W.S. (1995) Abdominal pain. In: W.S. Haubrich, F. Schaffner and J.E. Berk (Eds.), Gastroenterology, Vol. 1. W.B. Saunders, Philadelphia, PA, pp. 1 l-29. Hoheisel, U. and Mense, S. (1990) Response behaviour of cat dorsal horn neurones receiving input from skeletal muscle and other deep somatic tissues. J. Physiol. (Land.), 426: 265-280. Hoheisel, U. and Mense, S. (2000) The role of spinal nitric oxide in the control of spontaneous pain following nociceptive input. In: J. Sandktihler, B. Bromm and G.F. Gebhart (Eds.), Nervous System Plasticity and Chronic Pain. Progress in Brain Research, Vol. 129. Elsevier, Amsterdam, pp. 163-172. Hoheisel, U., Mense, S., Simon& D.G. and Yu, X.M. (1993) Appearance of new receptive fields in rat dorsal horn neurons following noxious stimulation of skeletal muscle: a model for referral of muscle pain?. Neurosci. Len., 153: 9-12. Hoheisel, U., Sander, B. and Mense, S. (1997) Myositis-induced functional reorganisation of the rat dorsal horn: effects of spinal superfusion with antagonists to neurokinin and glutamate receptors. Pain, 69: 219-230. Hu, J.W., Sessle, B.J., Raboisson, l?, Dallel, R. and Woda, A. (1992) Stimulation of craniofacial muscle afferents induces prolonged facilitatory effects in trigeminal nociceptive brainstem neurones. Pain, 48: 53-60. Inman, V.T. and Saunders, J.B.C.M. (1944) Referred pain from skeletal structures. J. Nerv. Merit. Dis., 99: 660-667. Johansen, M.K., Graven-Nielsen, T., Olesen, A.S. and ArendtNielsen, L. (1999) Generalised muscular hyperalgesia in chronic whiplash syndrome. Pain, 83: 229-234. Kellgren, J.H. (1938) Observation on referred pain arising from muscle. Clin. Sci., 3: 175-190. Laursen, R.J., Graven-Nielsen, T., Jensen, T.S. and ArendtNielsen, L. (1997a) Quantification of local and referred pain in humans induced by intramuscular electrical stimulation. Eul: J. Pain, 1: 105-I 13. Laursen, R.J., Graven-Nielsen, T., Jensen, T.S. and ArendtNielsen, L. (1997b) Referred pain is dependent on sensory input from the periphery - a psychophysical study. Eur: J. Pain, 1: 261-269. Laursen, R.J., Graven-Nielsen, T., Jensen, T.S. and ArendtNielsen, L. (1999a) The effect of compression and regional anaesthetic block on referred pain intensity in humans. Pain, 80: 257-263.

Laursen, R.J., Graven-Nielsen, T., Jensen, T.S. and ArendtNielsen, L. (1999b) The effect of differential and complete nerve block on experimental muscle pain in humans. Muscle Nerve, 22: 1564-1570. Lewis, T. (1938) Suggestions relating to the study of somatic pain. BMJ, 1: 321-325. MacKenzie, J. (1893) Some points bearing on the association of sensory disorders and visceral disease. Brain, 16: 321-353. Marchettini, P., Simone, D.A., Caputi, G. and Ochoa, J.L. (1996) Pain from excitation of identified muscle nociceptors in humans Brain Rex, 740(1-2): 109-116. Mense, S. (1994) Referral of muscle pain. APS J., 3: l-9. Mertz, H., Naliboff, B., Munakata, J., Niazi, N. and Mayer, E.A. (1995) Altered rectal perception is a biological marker of patients with irritable bowel syndrome. Gastroenterology, 109: 40-52. Mertz, H., Fullerton, S., Naliboff, B. and Mayer, E.A. (1998) Symptom and visceral perception in severe functional and organic dyspepsia. Gut, 42: 814-822. Munakata, J., Naliboff, B., Harraf, F., Kodler, A., Lembo, T., Chang, L., Silverman, D.H.S. and Mayer, E.A. (1997) Repetitive sigmoid stimulation induces rectal hyperalgesia in patients with irritable bowel syndrome. Gastroenterology, 112: 55-63. Ness, T.J. and Gebhart, G.F. (1990) Visceral pain: a review of experimental studies. Pain, 41: 167-234. Ness, T.J., Metcalf, A.M. and Gebhart, G.F. (1990) A psychophysiological study in humans using phasic colonic distension as a noxious visceral stimulus. Pain, 43: 377-386. Ness, T.J., Richter, H.E., Varner, R.E. and Fillingim, R.B. (1998) A psychophysical study of discomfort produced by repeated filling of the urinary bladder. Pain, 76: 61-69. Ruth, T.C. (1979) Pathophysiology of pain. In: T.C. Ruth and H.D. Patton (Eds.), The Brain and Neural Function, Vol. 20. W.B. Saunders, Philadelphia, PA, pp. 272-324. Selzer, M. and Spencer, W.A. (1969) Convergence of visceral and cutaneous afferent pathways in the lumbar spinal cord. Brain Res., 14: 331-348. Sinclair, D.C., Weddell, G. and Feindel, W.H. (1948) Referred pain and associated phenomena. Brain, 7 I : 184-2 Il. Sorensen, J., Graven-Nielsen, T., Henriksson, K.G., Bengtsson, M. and Arendt-Nielsen, L. (1998) Hyperexcitability in hbromyalgia. J. Rheumatol., 25: 152-155. Theobald, G.W. (1949) The role of the cerebral cortex in the apperception of pain. Lancet, 257: 41-47. Torebjork, H.E., Ochoa, J.L. and Schady, W. (1984) Referred pain from intraneural stimulation of muscle fascicles in the median nerve. Pain, 18: 145-156. Treede, R.-D. and Magerl, W. (2000) Multiple mechanisms of secondary hyperalgesia. In: J. Sandktlhler, B. Bromm and G.F. Gebhart (Eds.), Nervous System Plasticity and Chronic Pain. Progress in Brain Research, Vol. 129. Elsevier, Amsterdam, pp. 331-341. Vecchiet, L., Galletti, R., Giamberardino, M.A., Dragani, L. and Marini, F. (1988) Modifications of cutaneous, subcutaneous, and muscular sensory and pain thresholds after the induction of an experimental focus in the skeletal muscle. Clin. J. Puin, 4: 55-59.

J. Sandktihler, B. Bromm and GE Gebhart (Eds.) Progress in Brain Research, Vol. 129 0 2000 Elsevier Science B.V. All rights reserved

CHAPTER 26

Neurochemical

plasticity in persistent inflammatory

pain

Prisca Honor& Patrick M. Menning, Scott D. Rogers, Michael L. Nichols and Patrick W. Mantyh * Neurosystems

Centel; Departments of Preventive Sciences, Psychiatry Minneapolis, MN 55455, USA and VA Medical Centec

Introduction Chronic inflammation is responsible for a variety of persistent pain states including arthritis, back pain and temporomandibular joint disorder. Although significant progress has been made in understanding the peripheral inflammatory response, the neurochemical changes within the spinal cord that are involved in the generation and maintenance of chronic inflammatory pain are poorly understood. In response to persistent inflammatory pain, normally innocuous sensory stimuli are perceived as painful (allodynia) and mildly noxious sensory stimuli are perceived as highly painful (hyperalgesia). Both hyperalgesia and allodynia are thought to arise from a sensitization of peripheral nociceptors (peripheral sensitization) and spinal dorsal horn neurons (central sensitization) (Treede et al., 1992). In an effort to understand the mechanisms involved in the peripheral and central sensitization associated with inflammatory pain, the expression and internalization of the substance P receptor (SPR) in the spinal dorsal horn in four well-characterized and widely used experimental models of inflammatory pain in the rat were explored. These models of inflammation were produced by unilateral subcutaneous (s.c.) injection of formalin, carrageenan or * Corresponding author: P.W.Mantyh, NeurosystemsCenter, 18-208 Moos Tower, 515 Delaware Street, Minneapolis, MN 55455, USA. Tel: +l-612-626-0180; Fax: +l612-626-2565; E-mail: [email protected]

and Neuroscience, Universiq of Minnesota, Minneapolis, MN 55417, USA

complete Freund’s adjuvant (CFA) into the hindpaw. In addition, the neurochemical changes in an animal model of polyarthritis induced by CFA injection into the base of the tail were examined. Each of these inflammatory models is characterized by a different onset and time course of nociceptive inputs and responses. Using this approach, the previous electrophysiological and behavioral data were compared to the alterations in the amount and/or site of release of SP from primary afferent neurons, the number and location of SPR expressing spinal neurons that are activated by this released SP and the populations of neurons showing up-regulation of the SPR. Together, these data suggest that each type of inflammatory pain (acute, short-term and long-term) is characterized by a unique neurochemical signature within the spinal cord. Acute inflammatory

pain

Formalin induces a stereotypical biphasic response, consisting of an early short-lasting painful response followed by a prolonged period of tonic/persistent pain (Dubuisson and Dennis, 1977; Dickenson and Sullivan, 1987a,b). Although it is generally agreed that the first phase results from a direct action of formalin on nociceptive primary afferent fibers, the factors contributing to the second phase have not been fully defined. In our study, a significant SPR internalization in lamina I of the spinal dorsal horn was observed at the end of the first phase (8 mm), a time where there is almost no detectable periph-

358

era1 edema (see Fig. 1B). At 60 min following the initial injection of formalin, SPR internalization was still observed in a majority of lamina I SPR-IR neurons, at which time significant peripheral edema has developed (see Fig. 1B). It is of interest to compare this pattern of internalization with that produced by a single injection of capsaicin (Mantyh et al., 1995). Capsaicin induces a rapid, marked SPR internalization that is confined to lamina I neurons and this SPR internalization is largely resolved 60 min after capsaicin injection. In contrast, 60 min after formalin injection there remains a significant SPR internalization in lamina I neurons suggesting that there is an ongoing release of SP from primary afferent terminals. These results suggest that during both the first and second phases of the formalin response there is ongoing primary afferent input from C fibers and release of SP in the spinal cord. Similarly, others have also demonstrated that SPR antagonists could block both the first and second phases of the formalin response in spinal neurons (Chapman and Dickenson, 1993) and that peripheral injection of local anesthetics after the end of the first phase reduces electrophysiological, behavioral and anatomical correlates of the second phase (Dickenson and Sullivan, 1987a; Taylor et al., 1995; Puig and So&in, 1996; Abbadie et al., 1997a). This provides further evidence that a major component of the second phase of the formalin response is due to ongoing activity of primary afferent neurons. Together these results suggest that the neurochemical signature of acute inflammation (and persistent pain) is very similar, although with a longer time course, to what is produced by a brief chemical noxious stimulus, i.e. a stimulus that produces acute pain under non-pathological conditions. Short-term inflammatory

pain

Within 3 h following carrageenan injection, peripheral edema, allodynia and hyperalgesia had fully developed. In this condition, SPR intemalization was observed early, i.e. 10 min after carrageenan injection, in dendrites and in a few lamina I SPR-IR neurons. This indicates that SP release occurs in the early stage of carrageenan inflammation. However, there is a major difference between carrageenan-induced inflammation (short-term) and

25 -

0‘ Fig. 1. Quantification of SPR internalization and levels of SPR immunoreactivity in acute, short- and long-term inflammatory pain states. Results are expressed as percents of respective control values for SPR immunofluorescence levels in spinal lamina I (A) and percents of internalized SPR-IR lamina I neurons (B) in the spinal cord of animals 8 min and 1 h after formalin injection, 3 h after carrageenan injection, and 3 days after CFA injection and in CFA-induced polyarthritic rats (mean + s.e.m.). Note that in acute inflammatory pain there is on-going release of SP as measured by SPR internalization in lamina I neurons at both 8 and 60 min following formalin injection. In long-term inflammatory pain models (CFA and polyarthritis), while there is no tonic release of SP, there is a significant up-regulation of the SPR in lamina I neurons. In contrast, such changes were not observed in short-term inflammation (carrageenan). Student’s to respective controls. z-test, *** p < 0.001 compared

359 formalin-induced inflammation (acute). Specifically, formalin injection induced SPR internalization in lamina I neurons even at 1 h after injection, whereas there was no evidence of ongoing SPR intemalization at 3 h after carrageenan injection (see Fig. 1B). This lack of ongoing SPR internalization suggests that there is either no significant release of SP (using SP-induced SPR internalization as an assay of SP release) from primary afferents, even though a significant edema has developed at this time point, or that the system had desensitized, i.e. the receptor no longer responds. This absence of ongoing SPR internalization agrees with the observation that there is no spontaneous activity of dorsal horn neurons 3 h after carrageenan injection (Stanfa et al., 1992; see however Kocher et al., 1987) although a large number of spinal neurons express the Fos protein in both laminae I-II and III-IV 3 h after carrageenan injection (Honor6 et al., 1995). If spinal Fos induction is a reflection of neuronal activation (Hunt et al., 1987; Munglani and Hunt, 1995; Doyle and Hunt, 1999), it seems reasonable to conclude that there must be maintained nociceptive input from the periphery to the spinal cord during the development of carrageenan inflammation. Whether this is due to SP, released immediately after carrageenan injection (the Fos protein half-life is approximately 2 h), or to other neurotransmitters, notably excitatory amino acids is unclear. What is apparent at 3 h after carrageenan injection is that normally non-noxious mechanical stimulation of the carrageenan-inflamed hindpaw, which does not induce SPR internalization in normal animals, now induces massive SPR internalization in lamina I neurons (see Fig. 2A). Furthermore, noxious mechanical stimulation, which only induces SPR internalization in lamina I in normal animals, now induces SPR internalization in neurons located in laminae I-II and III-IV (see Fig. 2B,C). A key question raised by these observations is whether the activation of these laminae III-IV SPR-IR neurons is due to increased release and/or diffusion of SP from terminals that reside in laminae I-II or from de novo synthesis and release of SP from primary afferent neurons that terminate in laminae III-IV. Following peripheral inflammation, an increase in SP synthesis by small DRG neurons that normally synthesize SP (Donnerer et al.,

1993; Galeazza et al., 1995; Abbadie et al., 1996) as well as SP synthesis by large DRG neurons has been reported (Neumann et al., 1996). Release of SP by Al3 fibers could explain the SPR internalization observed in laminae III-IV after non-noxious stimulation in long-term inflammatory pain (Abbadie et al., 1997b). However, 3 h after carrageenan injection, there would appear to be insufficient time for de novo synthesis and transport of SP from primary afferent cell bodies to the terminals in the spinal cord. These data suggest that the SPR intemalization that is observed in lamina I following normally innocuous stimulation and the increased SPR internalization that is observed in laminae I-II and III-IV neurons following noxious stimulation is due primarily to peripheral sensitization that manifests itself by a greater release and diffusion of SP from primary afferent neurons that already expressed SP. This increase in SP release from primary afferent fibers could thus lead to SP diffusing significantly greater distances, resulting in a switch from primarily synaptic to volume neurotransmission (Agnati et al., 1995; Zoli et al., 1998). Because SP would diffuse and interact with SPR at both synaptic and extra-synaptic sites, this increased release of SP from primary afferent fibers could also explain why there is significantly greater SPR internalization in lamina I SPR neurons after noxious stimulation under inflammatory conditions. Based on these observations, we suggest that this neurochemical signature of short-term inflammation is characterized by a lack of spontaneous SP release from primary afferents as reflected by the lack of on-going SPR intemalization, a lack of SPR up-regulation and a switch from synaptic to volume transmission so that there is an increase in both the number and the location of SPR-IR spinal neurons that are activated in response to innocuous or noxious stimuli. Long-term

inflammatory

pain

CFA-induced unilateral inflammation and adjuvant-induced polyarthritis are two of the most commonly used models of long-term inflammatory pain. They elicit peak symptoms at 3 and 21 days, respectively. Similarities in the neurochemical signature of short- and long-term inflammatory pain include the lack of ongoing SPR internalization in basal un-

A

3 100 - Lamina I ’s . Non noxious

:

z 80i ; -0 6o zE E a E

Carrageenan

3hr.

40

C

floe - Lamina Ill ‘3j Noxious s 50) 80 _

: *

i ] n 60 zw m 5 40‘ E

Fig. 2. Quantification of SPR internalization in cell bodies of lamina I and lamina III SPR-IR neurons following non-noxious and noxious stimulation in carrageenanand CFA-induced inflammatory pain states. Results are expressed as percent of internalized SPR-IR neurons in lamina I after non-noxious mechanical stimulation (A) and in laminae I and III after noxious mechanical stimulation (B and C), in saline-, carrageenanand CFA-treated rats. Note that there is greater SPR internalization in lamina I in carrageenan-injected rats after non-noxious stimulation compared to CFA-injected rats (p < 0.01) and that this difference is also observed in lamina III neurons after noxious stimulation (p < 0.0001). In contrast, noxious mechanical stimulation induces a maximal SPR internalization in lamina I neurons in both carrageenanand CFA-injected rats. No SPR internalization is observed in laminae III-IV after non-noxious stimulation in any of the inflammatory pain states examined. One-way ANOVA and Fisher PLSD, ** p < 0.01, *** p < 0.001 compared to the saline-injected group.

stimulated condition (see Fig. 1B) and an increase in the numbers and location of the spinal neurons that showed SPR internalization in response to either normally non-noxious or noxious stimuli (see Fig. 2). The major difference in the spinal cords of animals with short- vs. long-term inflammation is that in long-, but not short-term inflammation, there is a significant up-regulation of the SPR on neurons in lamina I of the spinal cord (see Fig. 1A). The increase in SPR mRNA observed in the spinal cord several days after peripheral inflammation has been

reported to be blocked by morphine or SPR antagonists (Noguchi et al., 1988; McCarson and Krause, 1994, 1995, 1996) suggesting that SP release and/or SPR activation is necessary for SPR up-regulation. However, in both chronic inflammatory pain, which is associated with an increase in SP in primary afferents (Lembeck et al., 1981; Donaldson et al., 1992) and after nerve injury, which is associated with a decrease of SP in primary afferents (Noguchi et al., 1989; Garrison et al., 1993), there is an increase in SPR immunoreactivity in lamina I of the spinal dorsal horn (Abbadie et al., 1996). Additionally, while

361 CAMP has been reported to be involved in the regulation of SPR expression, SPR activation leads to the production of inositol phosphates, suggesting that SP is not the major regulator of SPR expression. These findings suggest that while SP could contribute to SPR up-regulation, other neurotransmitters, acting directly on SPR-IR neurons or indirectly via the release of yet unknown factors, must be involved. If there is a significant up-regulation of the SPR in lamina I neurons in long-term inflammation, does it alter the response properties of these neurons? Several electrophysiological studies have shown that the response of spinal cord neurons to peripheral stimuli increases in an inflammatory pain state (Haley et al., 1990; Simone et al., 1991; Dougherty et al., 1992; Stanfa et al., 1992; Urban et al., 1993; Neugebauer et al., 1994; see also Moore et al., 2000, this volume; Sandktihler et al., 2000, this volume). This increased responsiveness is hypothesized to be largely mediated by a facilitated transmission through the NMDA receptor. SPR activation leads to the generation of diacyl glycerol and inositol triphosphate, inducing an increase in intracellular calcium and a synergistic facilitation of the activity of the protein kinase C. In turn, protein kinase C induces phosphorylation of the NMDA receptors, counteracting the magnesium block and allowing NMDA receptors to operate at a more negative potential (for review see: Urban et al., TABLE

1994; Yaksh et al., 1995; Urban and Gebhart, 1998; Millan, 1999). These data suggest that SPR activation enhances NMDA receptor-mediated events and that the co-joint activation of SPR and NMDA receptors leads to increased neuronal excitability (see also Gerber et al., 2000, this volume; Sandktihler et al., 2000, this volume). The SPR up-regulation observed in long-term inflammatory pain states may therefore contribute to the central sensitization observed in long-term inflammatory pain. Conclusions Previous experimental and clinical studies have suggested that there are distinctive differences between acute and chronic pain including the shift from the sensitization of primary afferent neurons to a sensitization of spinal cord neurons. What is unique about the present approach is the ability to visualize and quantify neurochemical changes at the single and intracellular level as a pain ‘moves’ from the acute to the long-term state (see Table 1). These results suggest that SPR internalization might serve as a marker of the contribution of ongoing primary afferent input to acute/persistent pain states. Using a similar approach to understand the changes that other neurotransmitter/receptor systems undergo as a pain moves from acute to chronic should provide

1

Neurochemical

signature

of inflammatory

pain Experimental

models

of inflammatory

Acute formalin, SPR internalization in lamina I Synaptic transmission Volume transmission SPR internalization in lamina I after non-noxious stimulation SPR internalization in laminae I-III after non-noxious stimulation SPR up-regulation in lamina I SPR up-regulation on the contralateral

side

pain Short term

8 min

formalin,

60 min

carrageenan.

3 h

Long

term

CFA,

3 days

polyarthritis,

Yes Yes No ND

Yes Yes No ND

No Yes Yes Yes

No Yes Yes Yes

No ND ND ND

ND

ND

Yes

Yes

ND

No No

No No

No No

Yes No

Yes Yes

In acute inflammatory pain, there is on-going release of substance Although on-going release of SP is absent in short-term inflammatory somatosensory stimulation with a resulting SPR internalization being inflammatory pain, the same pattern of SP release and SPR activation of a significant up-regulation of the SPR in lamina I neurons.

21 days

P (SP) which induces SPR internalization in lamina I neurons. pain, SP is released in response to both noxious and non-noxious observed in neurons located in laminae I and III-IV. In long-term is observed as with short-term inflammation, but with the addition

362

significant insight into the mechanisms involved in the generation and maintenance of chronic pain and may lead to novel therapies to control different pain states. Acknowledgements This study was supported by the National Institute of Health, NINDS 23970, NIT-IDA 11986, NIH Training Grant DE07288, DE08973, DE14627, a Department of Veterans Affairs Merit Review, the Spinal Cord Society and the Association Fraqaise pour la Recherche Therapeutique. References Abbadie, C., Brown, J.L., Mantyh, P.W. and Basbaum, A.1. (1996) Spinal cord substance P receptor immunoreactivity increases in both inflammatory and nerve injury models of persistent pain. Neuroscience, 70: 201-209. Abbadie, C., Taylor, B.K., Peterson, M.A. and Basbaum, Al. (1997a) Differential contribution of the two phases of the formalin test to the pattern of c-fos expression in the rat spinal cord: studies with remifentanil and lidocaine. Pain, 69: lOl110. Abbadie, C., Trafton, J., Liu, H., Mantyh, P.W. and Basbaum, A.I. (1997b) Inflammation increases the distribution of dorsal horn neurons that internalize the neurokinin-1 receptor in response to noxious and non-noxious stimulation. J. Neurosci., 17: 8049-8060. Agnati, L.F., Zoli, M., Stromberg, I. and Fuxe, K. (1995) Intercellular communication in the brain: wiring versus volume transmission. Neuroscience, 69: 71 l-726. Chapman, V. and Dickenson, A.H. (1993) The effect of intrathecal administration of RP67580, a potent neurokinin 1 antagonist on nociceptive transmission in the rat spinal cord. Neurosci. Lett., 157: 149-152. Dickenson, A.H. and Sullivan, A.F. (1987a) Peripheral origins and central modulation of subcutaneous formalin-induced activity of rat dorsal horn neurons. Neurosci. Left., 83: 207211. Dickenson, A.H. and Sullivan, A.F. (1987b) Subcutaneous formalin-induced activity of dorsal horn neurones in the rat: differential response to an intrathecal opiate administered preor post-formalin. Pain, 30: 349-360. Donaldson, L.F., Harmar, A.J., McQueen, D.S. and Se&l, J.R. (1992) Increased expression of preprotachykinin, calcitonin gene-related peptide, but not vasoactive intestinal peptide messenger RNA in dorsal root ganglia during the development of adjuvant monoarthritis in the rat. Brain Rex Mol. Bruin Res., 16: 143-149. Donnerer, J., Schuligoi, R., Stein, C. and Amann, R. (1993) Upregulation, release and axonal transport of substance P and calcitonin gene-related peptide in adjuvant inflammation and

regulatory function of nerve growth factor. Regul. Pept., 46: 150-154. Dougherty, P.M., Sluka, K.A., So&in, L.S., Westlund, K.N. and Willis, W.D. (1992) Enhanced responses of spinothalamic tract neurons to excitatory amino acids parallel the generation of acute arthritis in the monkey. Brain Res., 17: l-13. Doyle, CA. and Hunt, S.P (1999) Substance P receptor (neurokinin-1).expressing neurons in lamina I of the spinal cord encode for the intensity of noxious stimulation: a c-Fos study in rat. Neuroscience, 89: 17-28. Dubuisson, D. and Dennis, S.G. (1977) The formalin test: a quantitative study of the analgesic effects of morphine, meperidine, and brain stem stimulation in rats and cats. Pain, 4: 161174. Galeazza, M.T., Gamy, M.G., Yost, H.J., Strait, K.A., Hargreaves, K.M. and Seybold, V.S. (1995) Plasticity in the synthesis and storage of substance P and calcitonin gene-related peptide in primary afferent neurons during peripheral inflammation, Neuroscience, 66: 443-458. Garrison, C.J., Dougherty, P.M. and Carlton, S.M. (1993) Quantitative analysis of substance P and calcitonin gene-related peptide immunohistochemical staining in the dorsal horn of neuropathic MK-801~treated rats. Brain Rex, 607: 205-214. Gerber, G., Youn, D.-H., Hsu, C.H., Isaev, D. and Randic, M. (2000) Spinal dorsal horn synaptic plasticity: Involvement of group I metabotropic glutamate receptors. In: J. Sandkiihler, B. Bromm and G.F. Gebhart (Eds.), Nervous System Plasticity and Chronic Pain. Progress in Brain Research, Vol. 129. Elsevier, Amsterdam, pp. 115-133. Haley, J.E., Sullivan, A.F. and Dickenson, A.H. (1990) Evidence for spinal N-methyl-D-aspartate receptor involvement in prolonged chemical nociception in the rat. Bruin Res., 518: 21% 226. Honor& P., Buritova, J. and Besson, J.M. (1995) Carrageeninevoked c-Fos expression in rat lumbar spinal cord: the effects of indomethacin. Em J. Pharmacol., 272: 249-259. Hunt, S.P., Pini, A. and Evan, G. (1987) Induction of c-fos-like protein in spinal cord neurons following sensory stimulation. Nature, 328: 632-634. Kocher, L., Anton, F., Reeh, P.W. and Handwerker, H.O. (1987) The effect of carrageenin-induced inflammation on the sensitivity of unmyelinated skin nociceptors in the rat. Pain, 29: 363-373. Lembeck, F., Donnerer, J. and Colpaert, EC. (1981) Increase of substance P in primary afferent nerves during chronic pain. Neuropeptides, 1: 175-180. Mantyh. P.W., Allen. C.J., Ghilardi, J.R.. Rogers, S.D., Mantyh, C.R., Liu, H., Basbaum, A.I., Vigna, S.R. and Maggio, J.E. (1995) Rapid endocytosis of a G protein-coupled receptor: substance P evoked internalization of its receptor in the rat striatum in vivo. PNAS, 92: 2622-2626. McCarson, K.E. and Krause, J.E. (1994) NK-1 and NK-3 type tachykinin receptor mRNA expression in the rat spinal cord dorsal horn is increased during adjuvant or formalin-induced nociception. J. Neurosci., 14: 7 12-720. McCarson, K.E. and Krause, J.E. (1995) The formalin-induced expression of tachykinin peptide and neurokinin receptor mes-

363 senger RNAs in rat sensory ganglia and spinal cord is modulated by opiate preadministration. Neuroscience, 64: 129-139. McCarson, K.E. and Krause, J.E. (1996) The neurokinin-1 receptor antagonist LY306,740 blocks nociception-induced increases in dorsal horn neurokinin-1 receptor gene expression. Mol. Pharmacol., 50: 1189-l 199. Millan, M.J. (1999) The induction of pain: an integrative review. Prog. Neurobiol., 51: 1-164. Moore, K.A., Baba, H. and Woolf, C.J. (2000) Synaptic transmission and plasticity in the superficial dorsal horn. In: J. Sandkiihler, B. Bromm and GE Gebhart (Eds.), Nervous System Plasticity and Chronic Pain. Progress in Brain Research, Vol. 129. Elsevier, Amsterdam, pp. 63-80. Munglani, R. and Hunt, S.P (1995) Molecular biology of pain. Br J. Anaesth., 75: 186-192. Neugebauer, V., Lucke, T., Grubb, B. and Schaible, H.G. (1994) The involvement of N-methyl-D-aspartate (NMDA) and nonNMDA receptors in the responsiveness of rat spinal neurons with input from the chronically inflamed ankle. Neurosci. Lat., 170: 237-240. Neumann, S., Doubell, T.l?, Leslie, T. and Woolf, C.J. (1996) Inflammatory pain hypersensitivity mediated by phenotypic switch in myelinated primary sensory neurons. Nature, 384: 360-364. Noguchi, K., Morita, Y., Kiyama, H., Ono, K. and Tohyama, M. (1988) A noxious stimulus induces the preprotachykinin-A gene expression in the rat dorsal root ganglion: a quantitative study using in situ hybridization histochemistry. Brain Rex, 464: 31-35. Noguchi, K., Senba, E., Morita, Y., Sato, M. and Tohyama, M. (1989) Prepro-VIP and preprotachykinin mRNAs in the rat dorsal root ganglion cells following peripheral axotomy. Brain Res. Mol. Brain Res., 6: 327-330. Puig, S. and So&in, L.S. (1996) Formalin-evoked activity in identified primary afferent fibers: systemic lidocaine suppresses phase-2 activity. Pain, 64: 345-355. Sandktthler, J., Benrath, J., Brechtel, C., Ruscheweyh, R. and Heinke, B. (2000) Synaptic mechanisms of hyperalgesia. In:

J. Sandktihler, B. Bromm and G.F. Gebhart (Eds.), Nervous System Plasticity and Chronic Pain. Progress in Brain Research, Vol. 129. Elsevier, Amsterdam, pp. 81-100. Simone, D.A., So&in, L.S., Oh, U., Chung, J.M., Owens, C., LaMotte, R.H. and Willis, W.D. (1991) Neurogenic hyperalgesia: central neural correlates in responses of spinothalamic tract neurons. J. Neurophysiol., 66: 228-246. Stanfa, L.C., Sullivan, A.F. and Dickenson, A.H. (1992) Alterations in neuronal excitability and the potency of spinal mu, delta and kappa opioids after carrageenan-induced inflammation. Pain, 50: 345-354. Taylor, B.K., Peterson, M.A. and Basbaum, A.1. (1995) Persistent cardiovascular and behavioral nociceptive responses to subcutaneous formalin require peripheral nerve input. J. Neurosci., 15: 7575-7584. Treede, R.D., Meyer, R.A., Raja, S.N. and Campbell, J.N. (1992) Peripheral and central mechanisms of cutaneous hyperalgesia. Prog. Neurobiol., 38: 391-421. Urban, L., Dray, A., Nagy, I. and Maggi, CA. (1993) The effects of NK-1 and NK-2 receptor antagonists on the capsaicin evoked synaptic response in the rat spinal cord in vitro. Regul. Pept., 46: 413-414. Urban, L., Thompson, S.W.N. and Dray, A. (1994) Modulation of spinal excitability: co-operation between neurokinin and excitatory amino acid neurotransmitters. Trends Neurosci., 17: 432-438. Urban, M.O. and Gebhart, G.F. (1998) The glutamate synapse: a target in the pharmacological management of hyperalgesic pain states. In: Glutamate Synapse as a Therapeutical Target: Molecular Organization and Pathology gf the Glutamate Synapse, Vol. 116, pp. 407-420. Yaksh, T.L., Chaplan, S.R. and Malmberg, A.B. (1995) Future directions in the pharmacological management of hyperalgesic and allodynic pain states: the NMDA receptor. NIDA Res. Monogr, 147: 84-103. Zoli, M., Torri, C., Ferrari, R., Jansson, A., Zini, I., Fuxe, K. and Agnati, L.F. (1998) The emergence of the volume transmission concept. Brain Res. Rev., 26: 136-147.

J. Sandkiihler, B. Bromm and G.F. Gebhart (Ed%) Progress in Brain Research, Vol. 129 0 2000 Elsevier Science B.V. All rights reserved

CHAPTER

27

Modelling the prolonged effects of neonatal pain D. Alvares, C. Torsney,B. Beland, M. Reynolds and M. Fitzgerald * Department

of Anatomy

and Developmental

Biology,

Introduction Alterations in normal activity patterns during development can permanently alter the future pattern of connections within the CNS. While this is well established in the auditory or visual system, it is not clear whether aberrant sensory activity, such as arise from pain and injury, during a critical period of early life has long-term consequences upon somatosensory function. Critically ill paediatric patients are frequently exposed to pain as a result of their disease processes or intensive care therapies (Chambliss and Anand, 1997). Preterm infants in intensive care show prolonged hyperalgesia both within an area of local tissue damage (Fitzgerald et al., 1988a, 1989) and secondary hyperalgesia in the contralateral limb following local ischaemic injury (Andrews and Fitzgerald, 1994). These findings are supported by other reports of ongoing infant pain behaviour in intensive care (Johnston and Stevens, 1996; Franck and Miaskowski, 1999; Porter et al., 1999b). However, there are also reports of prolonged sensory disturbances and altered pain perception lasting well beyond the infant period in children that have undergone early pain and trauma (Porter et al., 1999a). The relatively mild surgery of neonatal circumcision results in increased pain behaviour in infants three months later * Corresponding author: M. Fitzgerald, Department of Anatomy and Developmental Biology, University College London, London WClE 6BT, UK. Tel.: +44-207391-1303; Fax: +44-207-383-1929; E-mail: [email protected]

University

College

London,

London

WClE

6B1: UK

(Taddio et al., 1995, 1997), while early intensive care leads to complex changes in pain perception and somatisation (Grunau et al., 1994a,b). More recently, birth trauma has been linked to increased acute stress responses to painful stimulation in infancy (Taylor et al., 2000). The key to understanding such changes lies in the plasticity of the developing nervous system. Repetitive painful experiences, prolonged tissue or nerve damage in newborn rats can lead to long-lasting neurobehavioural sequelae not observed when the same stimuli are applied to adults (Reynolds and Fitzgerald, 1995; Anand et al., 1999; De Lima et al., 1999). Anatomical, neurochemical and electrophysiological experiments in rats have identified significant neuroplasticity in peripheral and central sensory pathways following nerve injury in infancy that last into adulthood and are likely to underlie these alterations in behaviour. Developing sensory neurons are more vulnerable to axotomy than their adult counterparts. Section/ligation of the sciatic nerve results in the death of 75% of axotomised dorsal root ganglion neurons in newborn rats compared to 30% in adult rats (Himes and Tessler, 1989; Lewis et al., 1999). As a result, the central dorsal root terminals of nearby intact nerves sprout in the spinal cord to occupy areas that are normally the exclusive territory of the sectioned nerve (Fitzgerald, 1985a,b; Fitzgerald et al., 1990; Shortland and Fitzgerald, 1994). These new sprouts form inappropriate functional connections with dorsal horn cells in areas far outside their normal termination area (Shortland and Fitzgerald, 1991), such that the nervous system becomes permanently distorted with disproportionately large areas

366

devoted to inputs surrounding the denervated skin (see also Moore et al., 2000, this volume). These plastic changes result from direct peripheral axotomy or nerve damage in infancy, which does not necessarily occur in clinical situations. In this article we consider the long-term effects of more restricted neonatal tissue damage and investigate whether skin wounding and inflammation alone are sufficient, at a critical stage of development, to alter the future development of sensory connections.

Early skin wounding Our laboratory has established a model of early injury which is a defined skin wound in the hindpaw of the newborn rat pup (Reynolds and Fitzgerald, 199.5; De Lima et al., 1999). While this wound heals rapidly, local sensory nerve terminals show a profound sprouting response, which long outlasts the injury (at least 12 weeks in the rat) (see Fig. 1). The effect is most dramatic when wounds are performed at birth and decreases progressively with

Wild type P6 mouse wounded at birth

P6 contralateral hindpaw skin section - normal innervation of the skin Fig. 1. Skin sections through the dorsal hindpaw from 7-day-old mice pups immunostained with PGP 9.5, a pan-neuronal marker, showing the sensory innervation. Upper panel. Skin that had been wounded at birth (see text for methods). The wound is now completely healed but hyperinnervation of the region is clear. Lower panel. Control, unwounded skin.

367 age at wounding. Behavioural studies show that the hyperinnervation is accompanied by long-lasting hypersensitivity and lowered mechanical threshold in the injured region (Reynolds and Fitzgerald, 1995; De Lima et al., 1999). Such a response in human infants could lead to prolonged sensory disturbance to both tactile and noxious stimulation. One possibility is that damage of skin sensory terminals in the wound triggers a local collateral sprouting response from nearby damaged nerve branches (Diamond et al., 1992). However, many of the sensory fibres in the wound area appear to be drawn up from deep tissues and non-cutaneous nerve bundles (see Fig. 1) which implies sprouting from distant intact axons. Damage of nerve terminals may not therefore be required for this response and skin inflammation alone may be sufficient. Previous work has already shown that there is no sympathetic involvement; the hyperinnervated fibres are both large diameter (RT97-positive) and small to medium diameter (CGRP-positive), the response being only partially reduced by pretreatment with the C-fibre neurotoxin capsaicin (Reynolds and Fitzgerald, 1995). To fully understand and therefore prevent this sprouting and the accompanying hypersensitivity, we need to isolate the factors that trigger and maintain it. To search for such factors (Marsh et al., 1999) we need to know the time of onset and pattern of sprouting following skin injury.

7 days

Time course of hyperinnervation Methods Newborn Sprague-Dawley rat pups were halothane anaesthetised and bilateral dorsal foot wounds of full thickness were made to hindpaws by pinching skin with forceps and cutting a 2 mm by 2 mm flap of skin. Pups were allowed to recover and returned to their mother. Control pups were anaesthetised only. At various intervals following surgery pups were sacrificed and tissue prepared for sectioning and immunostaining with PGP 9.5 (1 : 1000, Ultraclone).

Fig. 2. Sections of rat dorsal hindpaw skin immunostained with PGP 9.5 at 3, 5 and 7 days following skin wounding at birth (for methods, see text).

Results The results are illustrated in Fig. 2. The inflammatory response is evident at 24 h after wounding, with an en-

larged wounded area distorting the uniform layer of the epidermis, such that in some animals it is hardly apparent. The layered structure of the dermis is also

368

distorted such that only a homogeneous mass of inflamed tissue is visible. Three days after wounding skin structure is still disrupted and a scab has formed. As the wound contracts the inflammatory and necrotic area is pushed laterally. The wounded area itself has few hair follicles. At this stage only a few nerve fibres are seen, mainly around remnants of hair follicles and deep within the dermis. Greater numbers of fibres appear at 5 days both in and around the wound. The epidermis seems almost fully restored and smooth in appearance. At P7 hyperinnervation is clear in and around the wounded area. Thick, coarse fibres can be seen deep within the dermis, leading to finer-branched fibres superficially and penetrating the epidermis. The keratinised scab has already been shed or is near shedding leaving an apparently normal epidermal layer. Some incoming fibres to the wounded area are from collaterals from adjacent hair follicle fibres, whereas others appear to arise from more distant nerve bundles that follow the course of blood vessels and granular cells, probably mast cells.

The first signs of re-innervation of a neonatal skin can be observed 3 days after wounding. At 5 days there are signs of aberrant and excessive growth and by 7 days the area in and around the wound is clearly hyperinnervated.

The role of a diffusible growth factor is supported by the fact that the effect can be modelled in an explant co-culture system where neonatal dorsal root ganglion sensory neurons show increased neurite outgrowth in the presence of wounded skin compared to control unwounded skin (Reynolds et al., 1997). Sensory innervation of different sensory modalities (nociception, mechanoreception, etc.) is regulated by different members of the neurotrophin gene family, reflecting the expression by subpopulation of sensory neurons of different trk receptors. NGF overexpression produces an increase in size and numbers of nerves in the dermal layers (Davis et al., 1997) and its regulation of neurite outgrowth and target innervation is independent of its regulation of cell survival (Lewis et al., 1999). NGF is upregulated in inflamed and neonatally wounded skin (Constantinou et al., 1994) and collateral sprouting following partial skin denervation is NGF-dependent (Diamond et al., 1992). NGF, therefore, appears to be a prime candidate for triggering the hyperinnervation described here. It was surprising, therefore, that anti-NGF has no effect upon the neurite outgrowth from DRG co-cultured with wounded skin (Reynolds et al., 1997). Since this model may not completely mimic the situation ‘in vivo’, here we tested the effect of systemic injections of anti-NGF administered to newborn rats upon the hyperinnervation that follows neonatal wounding.

A possible role for NGF in hyperinnewation

Methods

A critical step in understanding this hyperinnervation and hypersensitivity is to identify the factors released by wounded skin that trigger sensory terminal growth (Andrews and Fitzgerald, 1994). Sensory neurons continue to express growth-associated protein (GAP43) mRNA for some weeks postnatally indicating their potential for growth and reorganisation of their terminals over this period (Chong et al., 1992). Nevertheless, the excessive growth observed here must be triggered by the numerous growth factors, neurotrophic factors and cytokines that are released on neonatal skin wounding and the accompanying inflammatory reaction. In addition, the response is unaffected by local anaesthetic sensory blockade at the time of injury, indicating that neural activity is not critical (De Lima et al., 1999).

Injections of 5 p.l/g anti-NGF (n = 10) or control injections of sheep serum (n = 10) were given i.p. on the day of birth, PO, and again at P2. In some cases (n = 3) they were given daily until P7. The anti-NGF serum was raised in a sheep against HPLC-purified NGF from mouse salivary glands and neutralised the growth produced by NGF (10 rig/ml) at a dilution 1 : 4000, but not BDNF or NT3, in a chick DRG assay. The potency of the antibody-treated rat pup serum was tested at P2 and P7. Serum diluted at 1 : 40 or 1 : 80 effectively blocked neurite outgrowth from cultured explants of PO dorsal root ganglion cells grown in 10 rig/ml NGF (for methods, see Reynolds et al., 1997) (Fig. 3B). Under halothane anaesthesia, a full-thickness skin wound was made on the dorsal surface of the foot at birth on Pl (n = 5

Conclusion

369

2-

0 7

WOUNDED SKIN

II

NORMAL SKIN

B.

No serum

wound

no wound anti-NGF treatment

wound anti-NGF treatment

Fig. 3. (A) Density of skin innervation in the 7-day-old rat dorsal hindpaw in controls and in pups that had been wounded Hyperinnervation occurred equally in wounded pups that had been treated with anti-NGF treatment or with serum control Serum from anti-NGF-treated pups was effective in blocking PO rat DRG neurite outgrowth (for methods, see text).

antibody treatment, IZ = 5 serum controls). At P7 animals were killed with an anaesthetic overdose and the hindpaws removed for sectioning and immunostaining with PGP 9.5 (for details, see Reynolds and Fitzgerald, 1995). Unwounded control animals were anaesthetised only (n = 5 antibody treatment, n = 5

at birth. only. (B)

serum controls). The density of skin innervation was measured by capturing and digitising the microscope image using a Leica imaging software and calculating the percentage of a known area of epidermis and dermis skin occupied by axons.

370

Results Fig. 3A shows the effect of ‘in vivo’ anti-NGF treatment on wound hyperinnervation. The density of nerve fibres in the wound area at P7 is the same in serum-treated and anti-NGF-treated skin. Both are significantly greater than unwounded controls. Conclusion Systemic anti-NGF treatment has no effect on wound hyperinnervation. These ‘in vivo’ data support earlier ‘in vitro’ data (Reynolds et al., 1997) and suggest that NGF is not a critical factor in the sprouting response. Early inflammation Skin wounding will involve a degree of sensory terminal damage and therefore could be viewed as partially neuropathic. It is still not clear to what extent pain and tissue inflammation alone are sufficient, at a critical stage of development, to alter the future development of affected sensory neurons and their central connections. Central changes following peripheral inflammation at birth have been reported but have involved very severe tissue damage caused by large volumes of inflammatory agents which may have neuropathic and systemic consequences (Ling and Ruda, 1999; Cleland et al., 1999). Following a limited inflammatory lesion, nociceptive thresholds fall (hyperalgesia) and previously non-noxious stimuli may become painful (allodynia) in both adults (see Willis, 1992) and in human infants and rat pups (Andrews and Fitzgerald, 1994; Jiang and Gebhart, 1998; Teng and Abbott, 1998; Marsh et al., 1999). Such hyperalgesic and allodynic pain is the most common type of pain seen in clinical practice. There are good reasons to suggest that the infant sensory nervous system could be particularly susceptible such peripheral stimulation. Newborn rats generally display an exaggerated response to mechanical

and thermal noxious stimulation compared to adults (Falcon et al., 1996; Fitzgerald et al., 1988b) and the response to experimental inflammatory agents such as formalin has a tenfold higher sensitivity in neonatal rats compared to weanlings (Teng and Abbott, 1998). This is paralleled in the hyperexcitable responses of sensory dorsal horn neurons at this age (Fitzgerald, 1997; Jennings and Fitzgerald, 1998). Furthermore, some of the key systems known to be involved in the plasticity of adult cord following tissue injury, such as the NMDA receptor, are overexpressed and more effective in the immature spinal cord (see Fitzgerald and Alvares, 1999). In addition, inflammation triggers changes in adult sensory neurone transmitter phenotype (Neumann et al., 1996), probably as a result of upregulation of neurotrophins (Woolf et al., 1994) and the postnatal period is a critical time of neurotrophin dependency in the development of sensory neuron phenotype (McMahon et al., 1996). The effects of early injammation on the development of pain behavioural responses Here we have tested whether an inflammatory lesion applied at birth produces long lasting effects upon mechanical and thermal sensory thresholds and pain behaviour to later inflammatory challenge. Methods Newborn rat litters were halotbane anaesthetised and injected with 1 pi/g (10 pl) of either 2% carageenan (n = 7 litters) or saline vehicle (n = 7 litters) into the plantar surface of one hindpaw. Other controls were anaesthetised only (n = 3 litters). On recovery from anaesthetic the pups were returned to their mothers. In one series of experiments (n = 4 litters), mechanical reflex withdrawal thresholds were tested with von Frey hairs on the dorsal surface of the paw at 2 h and 6 h, and 2,4,7, 10,14, 17,21 and 28 days after carageenan injection. Paw diameter was also

Fig. 4. The effect of dorsal hindpaw injection of 2% carageenan at birth upon paw diameter and van Frey hair thresholds for the following 28 days. The top two traces show the differences in paw diameter between the left and right feet in (upper tracej animals injected with carageenan in the left paw at birth and (lower trace) control, untreated animals. The bottom two traces show the von Frey thresholds in the same two groups of animals.

371

0

PI litter carageenan

inflamed left hindpaw non-inflamed right hindpaw

Time

Pi litter control

Time

I

non-inflamed right hindpaw

6hr

P2

P4

PlO

P7

PI4

P17

P21

P20

Time

II

right hindpaw

6hr

P2

P4

P7

Time

PlO

Pi4

P17

P21

P20

372 measured at these times. In a second series (n = 3 litters), pups were left until postnatal day (P) 21 and P28, when mechanical thresholds were tested and P24 and P31 when withdrawal latencies to noxious thermal stimulation were tested using the Hargreaves test. At 6-8 weeks, pups were either reinjected with 2% carageenan or complete Freunds adjuvant (CFA) (0.5 pi/g, 100 ~1) and the inflammatory response, mechanical and heat thresholds were tested again. Results Fig. 4 shows that a single injection of 10 ~1 2% carageenan causes a profound and long-lasting inflammation in the neonate. The hindpaw is measurably swollen and enlarged for 2 weeks. Despite this substantial stimulus, no difference was found in all three groups (neonatal carageenan, saline, anaesthetic only) in mechanical or heat thresholds (either between left and right paws or between inflamed and control groups) at any stage tested. In addition, the reapplication of 2% carageenan or CFA in these rats, when they had reached maturity, caused normal inflammatory, hyperalgesic and allodynic responses that did not differ from controls. Conclusion A substantial inflammatory response lasting for 2 weeks follows from a single injection of 2% carageenan into one hindpaw within 24 h of birth. This causes no lasting effect on behavioural sensory thresholds or inflammatory pain responses in adulthood. It is important to emphasise, however, that electrophysiological and neuroanatomical analysis may reveal changes in sensory connections that are not evident in reflex behavioural tests and that more research is required in this area. References Anand, K.J., Coskun, V., Thrivikraman, K.V., Nemeroff, C.B. and Plotsky, PM. (1999) Long-term behavioral effects of repetitive pain in neonatal rat pups. Physiol. Behav., 66(4): 621-631. Andrews, K. and Fitzgerald, M. (1994) The cutaneous with drawal reflex in human neonates: sensitization, receptive fields and the effects of contralateral stimulation. Pain, 56: 95-101.

Chambliss, C.R. and Anand, K.J. (1997) Pain management in the pediatric intensive care unit. Curr: @in. Pediarr:, 9: 246-253. Chong, M.S., Fitzgerald, M., Winter, J., Hu-Tsai, M., Emson, P.C., Weise, U. and Woolf, C.J. (1992) GAP-43 mRNA in rat spinal cord and dorsal root ganglia neurons: developmental changes and re-expression following peripheral nerve injury. Em J. Neurosci., 4: 883-895. Cleland, CL., Ritter, SM., Hawkins, A.R., Broghammer, A.M. and Gebhart, G.F. (1999) Neonatal inflammation causes permanent dose-dependent changes in thermal pain sensitivity in rats. Proc. World Congress on Pain, IASP Press, Seattle, p. 413. Constantinou, J., Reynolds, M.L., Woolf, C.J., Safieh-Garabedian, B. and Fitzgerald, M. (1994) Nerve growth factor levels in developing rat skin; upregulation following skin wounding. Neuroreport, 5: 2281-2284. Davis, B.M., Fundin, B.T., Albers, K.M., Goodness, T.P., Cronk, K.M. and Rice, EL. (1997) Overexpression of nerve growth factor in skin causes preferential increases among innervation to specific sensory targets. J. Camp. Neural., 387: 489-506. De Lima, J., Alvares, D., Hatch, D. and Fitzgerald, M. (1999) Sensory hyperinnervation following skin wounding: the effect of bupivacaine sciatic nerve blockade. BK J. Anaesth., 83: 662-664. Diamond, J., Holmes, M. and Coughlin, M.J. (1992) Endogenous NGF and nerve impulses regulate the collateral sprouting of sensory axons in the skin of the adult rat. Neuroscience, 12: 1454-1466. Falcon, M., Guendellman, D., Stolberg, A., Frenk, H. and Urea, G. (1996) Development of thermal nociception in rats, Pain, 67: 203-208. Fitzgerald, M. (1985a) The sprouting of saphenous nerve terminals in the spinal cord following early postnatal sciatic nerve section in the rat. J. Camp. Neurol., 240: 407-413. Fitzgerald, M. (1985b) The postnatal development of cutaneous afferent fibre input and receptive field organization in the rat dorsal horn. J. Physiol., 364: t-18. Fitzgerald, M. (1997) Neonatal pharmacology of pain. In: J.-M. Besson and A.H. Dickenson (Eds.), The Pharmacology of Pain. Handbook of Experimental Pharmacology, Vol. 130. Springer, Berlin, pp. 447-465. Fitzgerald, M. and Alvares, D. (1999) Building blocks of pain: the regulation of key molecules in spinal sensory neurones during development and following peripheral axotomy. Pain, 82: I-15. Fitzgerald, M., Millard, C. and Macintosh, N. (1988) Hyperalgesia in premature infants. Lancer, 1: 292. Fitzgerald, M., Shaw, A. and McIntosh, N. (1988b) The postnatal development of the cutaneous flexor reflex: a comparative study in premature infants and newborn rat pups. Dev. Med. Child Neurol., 30: 520-526. Fitzgerald, M., Millard, C. and McIntosh, N. (1989) Cutaneous hypersensitivity following peripheral tissue damage in newborn infants and its reversal with topical anaesthesia. Pain, 39: 31-36. Fitzgerald, M.. Woolf, C.J. and Shortland, P. (1990) Collateral sprouting of the central terminals of cutaneous primary affer-

373 ent neurons in the rat spinal cord: pattern, morphology and influence of targets. J. Camp. Neural., 300: 370-385. Franck, L.S. and Miaskowski, C. (1999) Measures of neonatal responses to painful stimuli: a research review. J. Pain .Syrnp. Manage., 14: 343-378. Grunau, R.V., Whitfield, M.F. and Petrie, J.H. (1994a) Pain sensitivity and temperament in extremely low-birth-weight premature toddlers and preterm and full-term controls. Pain. 58: 341-346. Grunau, R.V., Whitfield, M.F., Pettie, J.H. and Fryer, E.L. (1994b) Early pain experience, child and family factors, as precursors of somatization: a prospective study of extremely premature and fullterm children. Pain, 56: 353-359. Himes, B.T. and Tessler, A. (1989) Death of some DRG neurons and plasticity of others following sciatic nerve section in adult and neonatal rats. .I. Comp. Neural., 284: 215-230. Jennings, E. and Fitzgerald, M. (1998) Postnatal changes in responses of rat dorsal horn cells to afferent stimulation: a fibre induced sensitisation. J. Physiol., 509: 859-867. Jiang, MC. and Gebhart, G.F. (1998) Development of mustard oil-induced hyperalgesia in rats. Pain, 77: 305-313. Johnston, C.C. and Stevens, B.J. (1996) Experience in a neonatal intensive care unit affects pain response. Pediatrics, 98: 925930. Lewis, SE., Mannion, R.J., White, EA., Coggeshall, M., Martin, J.L., Dillmann, W.H. and Woolf, C.J. (1999) A role for HSP27 in sensory neuron survival. J. Neurosci., 19: 8945-8953. Ling, Q.-D. and Ruda, M.A. (1999) Neonatal persistent pain alters spinal neural circuitry and the response to peripheral inflammation and hyperalgesia. Proc. World Congress on Pain. IASP Press, Seattle, p. 412. Marsh, D., Dickenson, A.H., Hatch, D. and Fitzgerald, M. (1999) Epidural opioid analgesia in infant rats II: responses to carageenan and capsaicin. Pain, 82: 33-38. McMahon, S.B., Mendell, L.M., Phillips, H.S. and Wall, PD. (Eds.) (1996) Neurotrophins and Sensory Neurons: Role in Development, Maintenance and Injury. Philos. Trans. R. Sot. Lond. B. Biol. Sci., 351: 361-467. Moore, K.A., Baba, H. and Woolf, C.J. (2000) Synaptic transmission and plasticity in the superficial dorsal horn. In: J. Sandktihler, B. Bromm and G.F. Gebhart (Eds.), Nervous System Plasticity and Chronic Pain. Progress in Brain Research, Vol. 129. Elsevier, Amsterdam, pp. 63-80.

Neumann, S., Doubell, TX, Leslie, T. and Woolf, C.J. (1996) Inflammatory pain hypersensitivity mediated by phenotypic switch in myelinated primary sensory neurons. Nature, 384: 360-364. Porter, EL., Grunau, R.E. and Anand, K.J. (1999a) Long-term effects of pain in infants. J. Dev. Behav. Pediatr, 20: 253-261. Porter, EL., Wolf, C.M. and Miller, J.P. (1999b) Procedural pain in newborn infants: the influence of intensity and development. Pediatrics, 104: 13-20. Reynolds, M. and Fitzgerald, M. (1995) Long term sensory hyperinnervation following neonatal skin wounds. L Comp. Neurol., 358: 487-498. Reynolds, M.L., Alvares, D., Middleton, J. and Fitzgerald, M. (1997) Neonatally wounded skin induces NGF-independent sensory neurite outgrowth in vitro. Deb: Bruin Res., 102: 275283. Shortland, P. and Fitzgerald, M. (1991) Functional connections formed by saphenous nerve terminal sprouts in the dorsal horn following neonatal nerve section. Eur: J. Neurosci., 3: 383396. Shortland, P. and Fitzgerald, M. (1994) Neonatal sciatic nerve section results in a rearrangement of the central terminals of saphenous and axotomized sciatic nerve afferents in the dorsal horn of the spinal cord of the adult rat. Eu,: J. Neurosci., 6: 75-83. Taddio, A., Goldbach, M., Ipp, M., Stevens, B. and Koren, G. (1995) Effect of neonatal circumcision on pain responses during vaccination in boys. Lancer, 345: 291-292. Taddio, A., Katz, J., Ilersich, A.L. and Koren, G. (1997) Effect of neonatal circumcision on pain response during subsequent routine vaccination, Lancer, 349: 599-603. Taylor, A., Fisk, N.M. and Glover, V. (2000) Mode of delivery and subsequent stress response. Luncet, 355: 120. Teng, C.J. and Abbott, EV. (1998) The formalin test: a doseresponse analysis at three developmental stages. Pain, 76: 337-347. Willis, W.D. (Ed.), (1992) Hyperulgesiu and Allodyniu. Raven Press, New York. Woolf, C.J., Safieh-Garabedian, B., Ma, Q.-P., Crilly, P and Winter, J. (1994) Nerve growth factor contributes to the generation of inflammatory sensory hypersensitivity. Neuroscience, 62: 327-331.

J. Sandkiihler, B. Bromm and GE G&hart (Ed%) Progress in Brain Research, Vol. 129 0 2CKJOElsevier Science B.V. All rights reserved

CHAPTER 28

Role of sensitized pelvic nerve afferents from the inflamed rat colon in the maintenance of visceral hyperalgesia S.V. Coutinho ‘, X. Su, J.N. Sengupta2 and G.F. Gebhart* Department

of Pharmacology,

College

of Medicine, The University Iowa City, IA 52242, USA

Introduction Peripheral tissue injury results in hyperalgesia that often long outlasts the initiating injury. The mechanisms underlying cutaneous hyperalgesia have been extensively examined in humans and non-human animals following experimentally induced injury or inflammation (see Treede et al., 1992; Carstens, 1995 for reviews). Many investigators have documented that tissue damage or inflammation sensitizes nociceptors (i.e., increased spontaneous activity, reduced response thresholds and increased response magnitudes; e.g., Perl, 1985; Reeh et al., 1986; Baumann et al., 1991; Handwerker et al., 1991; LaMotte et al., 1992). Sensitization of nociceptors likely accounts for some characteristics of cutaneous hyperalgesia at early stages when tissue damage and inflammation are prevalent. However, the appearance of hyperalgesia at a site distant from the injury and the *Corresponding author: G.F. Gebhart, Department of Pharmacology, College of Medicine, The University of Iowa, Bowen Science Building, Iowa City, IA 52242, USA. Fax: +I-319-335-8930; E-mail: [email protected] ’ Current address: UCLA/CURE Neuroenteric Disease Program, WLA VA Medical Ctr., Bldg. 115, Room 223, 11301 Wilshire Blvd., Los Angeles, CA 90073, USA. 2 Current address: Division of Gastroenterology and Hepatology, Medical College of Wisconsin, Milwaukee, WI 53226, USA.

of Iowa,

Bowen

Science

Building,

persistence of hyperalgesia, even after the affected tissue has been locally anesthetized, suggests that changes in the CNS must also contribute to the expression of hyperalgesia (Hardy et al., 1950; Woolf, 1983, 1984; Wall and Woolf, 1984; Coderre and Melzack, 1985, 1987; LaMotte et al., 1991; Torebjork et al., 1992). Whether hyperalgesia is mediated by peripheral afferent mechanisms or central neuron hyperexcitabihty has been debated since the divergent observations of Lewis (1942) and Hardy et al. (1950). This discrepancy was apparently reconciled by a series of experiments carried out by Lah4otte and colleagues (Baumann et al., 1991; LaMotte et al., 1991; Simone et al., 1991). Their observations strongly suggest that the hyperalgesia produced by intrademral injection of capsaicin in humans cannot be attributed solely to peripheral sensitization of nociceptors, and favor central sensitization as an important mechanism underlying hyperalgesia. Although, until recently, most studies on hyperalgesia focused on cutaneous systems, it is becoming increasingly apparent, based on animal studies and clinical observations, that visceral organs may also develop an analogous hypersensitivity to pain. In humans, heightened sensitivity to pain from the gastrointestinal tract may arise as a consequence of infection or inflammation, or may present itself as a symptom of functional bowel disorders in the absence of a demonstrable pathology (see Mayer and Gebhart, 1994; Arendt-Nielsen et al., 2000, this volume for review). Additionally, in non-human animals, application of irritant chemicals to the viscera,

376

including the gut, has been shown to facilitate nociceptive reflexes (Ness et al., 1991; Gebhart, 1993; Langlois et al., 1994; Rice and McMahon, 1994; Burton and Gebhart, 1995; Rice, 1995; Coutinho et al., 1996; Ide et al., 1997). Numerous electrophysiological studies have documented that, akin to their cutaneous counterparts, visceral afferent fibers in the gastrointestinal and lower urinary tracts also possess the ability to sensitize as a consequence of peripheral inflammation (e.g., Habler et al., 1990, 1993; Sengupta et al., 1996; Su et al., 1997a,b; see also Sengupta and Gebhart, 1994 for review). That these enhanced afferent discharges likely evoke profound central changes is suggested by several observations. Prolonged noxious visceral stimulation or peripheral visceral inflammation results in increased excitability of visceroceptive neurons in the dorsal horn and an expansion of their peripheral receptive fields (McMahon, 1988; Ness and Gebhart, 1991; Cervero et al., 1992; Garrison et al., 1992; Gebhart, 1993; Willis, 1993). Additionally, noxious visceral stimulation or injury is known to cause increased expression of c-fos and nitric oxide synthase (NOS) in the spinal cord (e.g., Traub et al., 1992, 1993; Coutinho and Gebhart, 1996; Martinez et al., 1998). The altered patterns of somatovisceral referral in patients diagnosed with irritable bowel syndrome also favor a role for central sensitization in mediating the visceral hyperalgesia underlying functional bowel disorders as well (see Mayer and Gebhart, 1994). The relative contributions of peripheral afferent and central neuron mechanisms to visceral hyperalgesia have, however, not been examined. We previously documented that intracolonic instillation of zymosan reliably produces a visceral hyperalgesia in the rat (Coutinho et al., 1996). Furthermore, maintenance of hyperalgesia in this model is dependent on activity at both spinal NMDA and nonNMDA receptors, and increased production of nitric oxide (NO) due to induction of the neuronal isoform of NOS in the spinal cord (Coutinho and Gebhart, 1996). Additionally, zymosan-produced visceral hyperalgesia appears to be under the influence of a descending pain facilitatory system from the brainstem (Coutinho et al., 1998). The goal of the present experiments was to assess the relative contribution of peripheral afferent mechanisms to zymosan-pro-

duced visceral hyperalgesia. Portions of these data have previously been reported in the form of an abstract (Coutinho et al., 1997). Methods Electrophysiological

experiments

Experiments were performed on male SpragueDawley rats (Harlan, Indianapolis, IN) weighing 400-525 g. Rats were anesthetized with pentobarbital sodium (45 mg/kg, i.p., Nembutal@); catheters were placed in the femoral artery and vein. Anesthesia was maintained with supplemental i.v. (femoral vein) doses of 5-10 mg kg-’ h-’ . Blood pressure was continuously monitored through a catheter passed into the descending aorta via the left common carotid artery. Rats were paralyzed with pancuronium bromide (0.3 mg/kg, i.v.) and artificially ventilated with room air (3-4 ml stroke volume, 55-60 strokes/min). Paralysis was maintained during the course of the experiment with supplemental doses of pancuronium bromide (0.2-0.3 mg kg-’ hh’). A heating pad with circulating hot water and an overhead feedback controlled heat lamp (thermoprobe in the thoracic esophagus) were used to maintain core body temperature at 36°C. At the end of the experiment, the rat was killed with an iv. overdose of pentobarbital. A 3-4 cm long incision was made along the left flank to expose the lower abdominal viscera. The urinary bladder was emptied and catheterized (PE100) through the fundus. The urethra was ligated at its point of entry to the penis and the bladder was constantly emptied through the catheter. The left testis, vas deferens and seminal vesicle were tied off and removed. The major pelvic ganglion and pelvic nerve were accessed by reflecting the prostate lobe laterally. Two lengths of Teflon-coated stainless-steel wire with stripped ends were wrapped around the pelvic nerve following its isolation from the surrounding fatty tissue, and sealed with non-reactive gel (Wacker Silicone, Adrian, MI). The hypogastric, pudendal and femoral nerves were isolated and transected. The sciatic nerve was transected at the ischial notch. The rat was suspended by thoracic vertebral and ischial clamps after a laminectomy (T13Sl) had been performed to expose the lumbosacral spinal cord. The dura was carefully removed, the dor-

377

sal skin was reflected laterally, and the spinal cord was overlaid with mineral oil (37°C) to form a pool. The Sl dorsal root was transected close to its point of entry into the spinal cord. The distal cut end of the dorsal root was teased into fine filaments on a black microbase plate immersed in the pool of mineral oil. Single unit electrical activity was recorded monopolarly, amplified using a low-noise AC differential amplifier, displayed on a storage oscilloscope, and continuously monitored by analog delay. The descending colon and rectum were distended by pressure-controlled air inflation of a 7-8 cm long flexible latex balloon tied around a flexible tube. The balloon was lubricated, inserted intra-anally and positioned such that the end of the balloon was 1 cm proximal to the anus. Phasic colorectal distension (CRD, 80 mmHg, 20 s) was given by opening a solenoid gate to a constant-pressure air reservoir. Intracolonic pressure was continuously monitored using a pressure control device (Bioengineering, University of Iowa, Iowa City, IA). Peristimulus-time histograms, intracolonic pressure and blood pressure were displayed online continuously. The frequency of action potential occurrence was quantified (1 s bin width) online using the Spike2/CED 1401 data acquisition program. Experimental protocol Pelvic nerve afferent fibers in the S 1 dorsal root were identified by electrical stimulation of the pelvic nerve (single square wave pulse, 0.5 ms, 3-8 mA). Fibers that responded to CRD were further characterized by generating a stimulus-response function (SRF) to graded, increasing intensities of distending pressures (5, 10, 20, 30, 40, 60, 80, 100 mmHg; 30 s duration; 4 min interstimulus interval). Resting activity of each fiber was determined over 60 s before distension, and the response to distension was calculated as the increase in discharge during distension over resting activity, if any. Following generation of the control SRF, the balloon was removed and zymosan (2 ml in 30% ethanol, Sigma Chemical Co., St. Louis, MO) or the vehicle (30% ethanol) was instilled into the colon. 30 min after intracolonic treatment, the balloon was re-inserted and SRFs were generated again (30 min, 1, 2 and 3 h post-treatment).

Electromyographic

experiments

Rats were anesthetized with pentobarbital sodium and Teflon-coated stainless-steel wire electrodes were sewn into the external oblique musculature, just superior to the inguinal ligament for electromyographic (EMG) recording. The electrode leads were tunneled subcutaneously and exteriorized to the base of the neck, where they were anchored to the neck musculature with the aid of a suture. The response that was quantified was the visceromotor response, a contraction of the abdominal and hind limb musculature during CRD in awake rats (Ness and Gebhart, 1988a). CRD was produced as described above. EMG activity in the external oblique musculature (visceromotor response) was quantified by recording the number of spikes crossing a pre-set voltage threshold as described earlier (Coutinho et al., 1996). Each distension trial lasted 60 s and EMG activity was quantified for 20 s before distension (baseline), during distension, and for 20 s after distension. The increase in total number of spikes during distension over baseline was recorded as the response. Experimental protocol On the day of testing, two stable control responses to CRD (80 mmHg, 20 s, 4 min interstimulus interval) were obtained prior to any treatment. The animals were then briefly anesthetized with halothane and their colons were washed with ethanol (30%, 1 ml, approximately 30 s) to break the mucous barrier, followed by a saline rinse (1 ml). Either zymosan (1 ml, 25 mg/ml) or an equal volume of vehicle (saline) was then instilled into the colon through a gavage needle inserted intra-anally to a depth of about 7-8 cm. Following intracolonic treatment, two responses to CRD were obtained again at each of several times (1, 2, 3, 4, 5, 6 and 24 h) to determine the time-course of hyperalgesia. Data analysis and statistics Unless indicated otherwise, data are represented as mean f SEM. Data for visceromotor responses (VMR) are normalized as percentage of control (% control), computed as percent of the mean of the two responses to CRD obtained in the same

378

animal prior to intracolonic treatment. Changes in the VMR following intracolonic treatment were statistically analyzed by repeated measures ANOVA. Except for resting activity, data from afferent fiber recordings are normalized as % control of the response to 100 mmHg prior to intracolonic treatment. Data from afferent fiber recordings were also analyzed by repeated measures ANOVA followed by Student’s modified t test with the Bonferroni correction for multiple comparisons if indicated. P < 0.05 was considered statistically significant in all cases. Results

,25Hz]

low threshold

/

high threshold 5 Hz]

lo:],: 0

500

1000 time (set)

1500

2000

Fig. 1. Typical responses of a low-threshold fiber and a highthreshold fiber to graded intensities of CRD (5-100 mmHg, 30 s). Data are presented as peristimulus-time histograms (1 s bin width); phasic distension pressures are shown below.

Fiber sample

A total of 54 mechanosensitive pelvic nerve afferent fibers innervating the colon were identified. Of 51 fibers tested, 25 (49%) were unmyelinated C-fibers (mean CV = 1.87 f 0.07 m/s) and 26 (51%) were thinly myelinated Ah-fibers (mean CV = 9.59 l 1.24 m/s). The conduction velocities of three fibers was not determined. All fibers were mechanosensitive and exhibited incrementing responses to increasing intensities of graded CRD. The relative proportions of mechanosensitive A6 and C colonic afferent fibers are similar to those reported earlier (Sengupta and Gebhart, 1994; Su et al., 1997a). Responses to CRD

Of the 54 fibers that responded to CRD, 45 were characterized further for responses to increasing intensities of CRD. Six of these 45 fibers were tested 3-4 days after intracolonic instillation of zymosan. Of the remaining 39 fibers, two exhibited no resting activity in the presence of the balloon in the descending colon; 37 fibers had an ongoing discharge (mean = 0.5 f 0.1 imps/s). There was no statistically significant difference between the resting activities of C-fibers (mean = 0.4 f 0.13 imps/s, n = 19) and A&-fibers (mean = 0.6 f 0.13 imps/s, n = 18) All fibers studied exhibited monotonically increasing, time-locked responses to graded intensities of phasic CRD. On the basis of their response thresholds, afferent fibers were classified as low-threshold (LT) (cl0 mmHg, n = 31) or high-threshold (HT) (~30 mmHg, n = 4). Representative examples of re-

sponses of LT and HT fibers to increasing intensities of phasic distension are illustrated in Fig. 1. The LT fiber responded to the lowest intensity of CRD tested (5 mmHg), whereas the HT fiber did not respond until the intensity of the distending stimulus was increased to 30 mmHg. SRFs for individual fibers are illustrated in Fig. 2; the insets illustrate the mean SRFs for each group. Effect

of

intracolonic zymosan

In the present study, colons were qualitatively examined at the end of each experiment. Following treatment with intracolonic zymosan, but not vehicle, colons appeared hyperemic and engorged. The effect of acute intracolonic instillation of zymosan was tested on 18 fibers (16 LT, 2 HT). All fibers studied were sensitized 30 min after intracolonic zymosan (i.e., exhibited an increase in response magnitude to graded CRD). The time-course of sensitization was studied in 14 of these 18 fibers. 13/14 fibers were maximally sensitized 30 min after intracolonic treatment with zymosan; one fiber was maximally sensitized at 1 h. There was no progressive enhancement of responses to CRD after 1 h. With time, response magnitudes were diminished and reverted to pre-zymosan control levels between 1 and 3 h. An example of a LT fiber that was sensitized by intracolonic zymosan is illustrated in Fig. 3. Both resting activity as well as responses to distension were enhanced 30 min after intracolonic instillation of zymosan, but reverted to control levels by 3 h. The effect of intracolonic zymosan was also tested

379

If

mean imps/s 30 7

KES” imps/s 30 1

20

20

10 lmpsls

n

-,

801

-

j”’

5

10

20

10 n=31 100

30

40

pressure

60

80

imps/s

0 L-?=df

8o 1

5

100

5

(mmHg)

10

n=4 100

20

30

pressure

40

60

80

100

(mmHg)

Fig. 2. Individual stimulus-response functions (SRFs) of low-threshold afferent fibers (left panel) and high-threshold fibers (right panel) to graded intensities of CRD. The insets illustrate the mean SRFs.

pre-zymosan

control 20 Hd

30 min after zymosan

2 hrs after zymosan

+

-‘. ; m-

3 hrs after zymosan

mmflg 100 0

I, 0

r! 500

n

n 1000

n

n 1500

n, 2000

time (set)

Fig. 3. Representative example of the responses of a low-threshold pelvic nerve afferent fiber to graded CRD (S-100 mmHg, 30 s) before and at various times after intracolonic instillation of zymosan. Data are presented as peristimulus-time histograms (1 s bin width); distending pressures are shown below. Note that both resting activity and responses to distension were enhanced 30 min after zymosan treatment and reverted to control levels by 3 h.

on 2 HT fibers. Response thresholds were clearly reduced at 30 min in both cases. The time-course of sensitization was followed in 1 HT fiber (Fig. 4), and was found to be similar to that of LT fibers. Data for all 14 fibers are summarized in Fig. 5. The resting

activity of these fibers also increased significantly 30 min after treatment with zymosan (Fig. 5 inset), but recovered to pre-zymosan activity during the l3 h recovery period (typically by 1 h). Response magnitudes were significantly increased 30 min after

380 pre-zymosan

control

5 Hz]

30 min after zymosan

1 hr after zymosan

Jd

L

i-l

n

I

.I b

,,

1

I,

,,

2 hrs after zymosan

mmHg 100

50 0 3,

0

500

n 1000

n

,

1500

n

n,

2000

time (set)

Fig. 4. Example of the responses of a high-threshold various times after intracolonic instillation of zymosan. pressures are shown below. Note the shift in response reverted to control levels by 2 h

pelvic nerve afferent fiber to graded CRD (5-100 mmHg, 30 s) before and at Data are presented as peristimulus-time histograms (1 s bin width); distending threshold at 30 min and 1 h after zymosan treatment. Responses of the fiber

intracolonic zymosan, although possible increases in response magnitudes to lower intensities of distension (5, 10 mmHg) were likely masked by the increase in resting activity. Responses to distension reverted to near control levels by l-3 h. In one experiment, the effect of intracolonic zymosan was studied on a mechanically insensitive (‘silent’) afferent fiber. This fiber had no ongoing activity and did not respond to 100 mmHg CRD prior to intracolonic zymosan (Fig. 6). 30 min after intracolonic instillation of zymosan, the fiber developed resting activity and began to encode intensities of CRD (40-100 mmHg). At 3 h after intracolonic zymosan, this fiber was still spontaneously active, but no longer clearly responsive to CRD. The mechanosensitive properties of six fibers were tested 30 and 60 min following intracolonic instillation of the vehicle (30% ethanol). Resting activities of these fibers (mean = 1.08 f 2.4 imp/s before) were unchanged (mean = 1.29 & 2.4 imp/s, 30 min after) by intracolonic instillation of vehicle. Response magnitudes were also unchanged (Fig. 7). Because visceromotor responses to colonic dis-

tension are increased significantly for more than 3 h after intracolonic zymosan treatment (see below), we also examined six fibers 3-4 days after zymosan treatment. As presented in Fig. 8, the stimulusresponse function of these fibers virtually overlaps the function of the pretreatment fibers (n = 35) illustrated in Fig. 2, which received no treatment. Effect of intracolonic zymosan on visceromotor responses to CRD in awake animals

In separate experiments, the effect of intracolonic instillation of zymosan on the visceromotor response (VMR) to a noxious intensity of CRD (80 mmHg) was tested in four animals. Similar to the effect on the response characteristics of afferent fibers innervating the colon, intracolonic zymosan resulted in significantly enhanced VMRs to CRD (i.e., produced visceral hyperalgesia; Fig. 9). However, in contrast to the sensitization of afferent fibers, the VMR to CRD exhibited progressive sensitization with time. Zymosan-produced visceral hyperalgesia was apparent at 1 h, increased to a robust level at 3 h, a time

381

30 min after zymosan

5

10

20

30

40

60

80

100

CRD (mmHg)

Fig. 5. Summary of the effects of intracolonic zymosan on the response characteristics of mechanosensitive pelvic nerve afferent fibers innervating the colon. Responses to distension are presented as % control; 100% = response to 100 mmHg CRD before intracolonic instillation of zymosan. Fibers (n = 14) exhibited a significant increase in response magnitude 30 min after treatment with zymosan (P < 0.05 vs. pre-zymosan control, ANOVA). The resting activity of these fibers was also significantly enhanced 30 min after treatment with zymosan (*, inset).

when afferent fibers were no longer sensitized, and continued to persist even at 24 h. Responses to 80 mmHg were unchanged in control animals (n = 4) that received intracolonic saline. We (D. Beck and G.F. Gebhart, unpublished data) have subsequently followed the time course of visceral hyperalgesia after intracolonic zymosan and found the VMR to be greater than pre-zymosan, control responses to distensions for up to four days.

Discussion We have previously demonstrated that intracolonic instillation of zymosan produces colonic inflammation and facilitates the VMR to CRD in awake rats, (i.e., results in a visceral hyperalgesia; Coutinho et al., 1996). The present study was designed to examine the relationship between alterations in the mechanosensitive properties of pelvic nerve afferent fibers innervating the colon and the facilitated VMR to distension following instillation of zymosan

into the colon. A significant finding of the present study was the absence of correspondence between the duration of afferent fiber sensitization and visceral hyperalgesia. In the rat, afferent information from the colon is conveyed to the thoracolumbar (T13-Ll) and lumbosacral (L6-S2) segments of the spinal cord via the lumbar splanchnic and pelvic nerves, respectively. It has been documented in the cat that afferent fibers in both pathways exhibit similar responses to mechanical stimuli (Blumberg et al., 1983; Janig and Koltzenburg, 1986, 1990; Htibler et al., 1990). In addition, responses of visceroceptive neurons in the thoracolumbar and lumbosacral spinal segments to noxious CRD in the rat have been found to be qualitatively indistinguishable (Ness and Gebhart, 1987, 1988~). However, it appears that pelvic nerve afferent fibers innervating the colon of the rat, and the lumbosacral spinal cord, make significantly greater contributions to the processing of colonic nociceptive information than do splanchnic nerve afferent fibers and the thoracolumbar spinal cord (Ness and Gebhart, 1988b; Traub et al., 1993; Kolhekar and Gebhart, 1994; Coutinho and Gebhart, 1996). In support, we have observed that the VMR to CRD in rats is abolished following transection of the pelvic nerve (S. Coutinho and G.F. Gebhart, unpublished observation). Hence, in evaluating the contribution of peripheral afferent mechanisms to zymosan-produced visceral hyperalgesia, we studied the mechanosensitive properties of pelvic nerve afferent fibers innervating the colon. Several studies have investigated the effect of experimental inflammation of the colon or bladder on mechanosensitive pelvic nerve afferent fibers (Habler et al., 1990, 1993; Sengupta et al., 1996; Su et al., 1997a,b). A common finding is that both HT and LT visceral afferent fibers can sensitize as a consequence of inflammation. Consistent with this observation, we found that all LT and HT colonic mechanosensitive fibers tested were sensitized 30 min following intracolonic instillation of zymosan (i.e., exhibited enhanced responses to CRD). In addition, these fibers also exhibited an increase in resting activity. In contrast to treatment with intracolonic zymosan, responses to distension as well as resting activity were unaltered following treatment with intracolonic vehicle (30% ethanol). The time-course of

382 pre-zymosan

control

IOHz]

I

30 min after zymosan

1 hr after zymosan

3 hrs after zymosan

mmHg 100 0 I,

n 0

n

n

500

1000

n

,

l-l

n,

1500

2000

time (set)

Fig. 6. Responses of a mechanically insensitive afferent fiber to graded CRD (5-100 mmHg, 30 s) before, and at various times after intracolonic instillation of zymosan. Data are presented as peristimulus-time histograms (1 s bin width); distending pressures are shown below. Note that the fiber exhibited neither resting activity nor responses to distension prior to intracolonic instillation of zymosan. At 30 and 60 min after treatment with zymosan, there was a marked increase in spontaneous activity, and the fiber began to encode distending pressures

zymosan-produced sensitization was studied in fourteen fibers; there was no progressive sensitization with time and within l-3 h response characteristics had reverted to control, uninflamed levels. The mechanically insensitive, ‘silent’ afferent fiber that was fortuitously identified in one experiment also exhibited a similar time-course of sensitization, although it

still exhibited some residual spontaneous activity at 3 h. These observations are similar to those of Htibler et al. (1993) who studied the effect of mustard oil or turpentine oil on pelvic nerve afferent fibers innervating the urinary bladder of the cat. Prior to inflammation, none of the afferent fibers displayed resting mean

s 2% 0 ..

.z 20 al g 15. 2 al IO2 :: 5? o-

imp/s 30 - 0 ‘zs 25 - .--;‘:ir; l 3-4 days post-zymosan control

control 30 min 60 min

1

I

f+j 20 s 15. al g IOgc

50

c 0

5 5

IO

20

pressure

30

40

60

80

100

(mmHg)

Fig. 7. Summary of the effects of intracolonic vehicle (30% ethanol) on the response characteristics of mechanosensitive pelvic nerve fibers (n = 4) innervating the colon. Responses to distension are presented as mean impulses/s & SEM before (control) and at 30 and 60 min after ethanol treatment.

IO

20

30

pressure

40

60

80

100

(mmHg)

Fig. 8. Summary of the effects of intracolonic the response characteristics of mechanosensitive fibers (n = 6) 3-4 days after zymosan treatment. to distension are presented as mean impulse/s comparison, the mean stimulus-response function fibers presented in Fig. 2 is included (control).

zymosan on pelvic nerve Responses ZIZ SEM. For from the 35

383 % control 300 ,

m 2 8 ;

250 200

‘0 2 150 :: 100 %

,o.

‘5 0

intracolonic saline intracolonic zvmosan

I

C

I

I

I

2

3

I

I

I

4

5

6

24

time (hours)

Fig. 9. Time-course of behavioral hyperalgesia. Responses to a noxious intensity of CRD (80 mmHg, 20 s) were recorded prior to, and following intracolonic instillation of zymosan or saline at the times indicated. Data are presented as mean f SEM, calculated as % control of the response to 80 mmHg CRD before intracolonic treatment. Zymosan treatment resulted in a significant facilitation of the visceromotor response to 80 mmHg (P < 0.05 vs. saline control, ANOVA).

activity. Both treatments activated pelvic nerve afferent fibers at short latency. In addition, mustard oil produced an enhanced mechanosensitivity in some fibers during the first 30 min. At later time-points, however, all fibers studied in mustard oil-treated animals exhibited desensitization to distension of the bladder. In contrast, instillation of turpentine oil resulted in increased responsiveness to distension in all fibers studied. One hour following intracolonic instillation of zymosan, but not saline, the VMR to noxious CRD was facilitated compared to the pre-treatment control response. In contrast to the short time-course of afferent fiber sensitization (30 min to 3 h), however, there was a progressive increase in the magnitude of the hyperalgesia with time. Additionally, although afferent fibers from the zymosan-inflamed colon were no longer sensitized at 3 h, zymosanproduced visceral hyperalgesia was significant and robust at 3 h and continued to persist at near maximum even at 24 h. These observations suggest that the visceral hyperalgesia in this model is initiated by sensitization of peripheral afferent fibers, but is independent of a persistent afferent drive from sensitized mechanosensitive afferent fibers. Although the mechanically insensitive fiber studied in the present experiments also desensitized with time, it still exhibited some ongoing activity at 3 h. It is possible that low-level residual spontaneous discharges in sensi-

tized ‘silent’ afferent fibers may contribute to the maintenance of visceral hyperalgesia in this model. However, although currently available data suggest that approximately 35-40% of afferent fibers in the pelvic nerve are mechanically insensitive (‘silent’), less than 10% of them appear to be sensitized by inflammation (Habler et al., 1990; Sengupta and Gebhart, 1994) These results favoring central mechanisms in the maintenance of visceral hyperalgesia are in agreement with several studies demonstrating the involvement of central changes in the maintenance of cutaneous hyperalgesia (Woolf, 1983, 1991; Coderre and Melzack, 1985; Dickenson and Sullivan, 1990; Haley et al., 1990, 1992; Woolf and Thompson, 1991; see also Coderre et al., 1993 for review). The observation that hyperalgesia continues to persist even after afferent input from the affected area has been silenced by local anesthesia, lends further credence to the role of central mechanisms in the maintenance of hyperalgesia (Hardy et al., 1950; Woolf, 1983, 1984; Wall and Woolf, 1984; Coderre and Melzack, 1985, 1987; LaMotte et al., 1991; Torebjork et al., 1992). Central changes produced by a peripheral injury manifest as increased excitability of spinal neurons characterized by increased spontaneous activity, reduced thresholds and increased responsiveness, and expansion in size of peripheral receptive fields (Price et al., 1978; Kenshalo et al., 1982; Menetrey and Besson, 1982; McMahon and Wall, 1984; Schaible et al., 1987; Cervero et al., 1988; Hoheisel and Mense, 1989; Hylden et al., 1989; Woolf and King, 1990; Simone et al., 1991; Dougherty et al., 1992; see also Svendsen et al., 2000, this volume; Gerber et al., 2000, this volume; Sandkiihler et al., 2000, this volume). These changes are believed to be partly dependent on activity at spinal excitatory amino acid receptors and increased production of NO (Woolf, 1983, 1991; Coderre and Melzack, 1985; Dickenson and Sullivan, 1990; Haley et al., 1990, 1992; Woolf and Thompson, 1991; see also Svendsen et al., 2000, this volume; Sandktihler et al., 2000, this volume; Hoheisel and Mense, 2000, this volume). In support, we have previously documented that zymosan-produced visceral hyperalgesia is significantly attenuated by intrathecally administered NMDA and non-NMDA receptor antagonists as well as by inhibitors of NOS (Coutinho et al., 1996; Coutinho and Gebhart, 1996).

384

Additionally, colonic inflammation also induces the increased expression of nNOS in the lumbosacral spinal cord and activates descending pain facilitatory systems from the brainstem (Coutinho and Gebhart, 1996). Moreover, chemical irritation of the colon has also been shown to result in the expansion of convergent cutaneous receptive fields of spinal visceroceptive neurons (Ness and Gebhart, 1991; Gebhart, 1993). In summary, the present results suggest that zymosan-produced visceral hyperalgesia continues to persist after mechanosensitive colonic afferent fibers are no longer sensitized, supporting the concept that central mechanisms are responsible for maintenance of hyperalgesia. The present results, however, are not definitive on this point. Although we encountered no mechanosensitive fibers that exhibited sensitization lasting more than 3 h, we did not (could not) continuously record from fibers for 6, 12 or 24 h to determine whether a progressive inflammation produced additional waves of sensitization. Maintenance of hyperalgesia may not require persistent sensitization, but only enhanced or episodic increases in spontaneous input. Spontaneous activity was increased in the present experiments, significantly at times early after intracolonic instillation of zymosan, and modestly (not statistically significant) at 3 h. Input from silent afferents could also be important to maintenance of hyperalgesia. The one silent afferent encountered was active 3 h after zymosan treatment; how long such activity might be sustained is not known. Finally, non-mechanosensitive mucosal, muscle and/or serosal receptors not studied here could have been affected by zymosan and have contributed to both the development and maintenance of hyperalgesia. Accordingly, several factors not studied or encountered in the present experiments could bridge the apparent discontinuity between the duration of peripheral sensitization and duration of visceral hyperalgesia reported here. Arguments opposing a role for sustained peripheral input as necessary to the maintenance of hyperalgesia are given above. There are clear and long-lasting central nervous system consequences and contributions to the maintenance of hyperalgesia (e.g., see Urban and Gebhart, 1999) and resolution of the relative roles of peripheral and central mechanisms awaits experimental study.

Acknowledgements The authors gratefully acknowledge the assistance of Michael Burcham in preparation of graphics and Susan Birely in preparation of the manuscript. Supported by NS 19912. References Akoev, G.N., Filippova, L.V. and Sherman, N.O. (1996) Mast cell mediators excite the afferents of cat small intestine. Neuroscience, 11: 1163-l 166. Anderson, G.D., Hauser, S.D., McGarity, K.L., Bremer, M.E., Isakson, P.C. and Gregory, S.A. (1996) Selective inhibition of cyclooxygenase (COX)-2 reverses inflammation and expression of COX-2 and interleukin 6 in rat adjuvant arthritis. J. Clin. Invest., 97: 2672-2679. Arendt-Nielsen, L., Laursen, R.J. and Drewes, A.M. (2000) Referred pain as an indicator for neural plasticity. In: .I. Sandktihler, B. Bromm and G.F. Gebhart (Eds.), Nervous System Plastic& and Chronic P&z. Progress in Brain Research, Vol. 129. Elsevier, Amsterdam, pp. 343-356. Baumann, T.K., Simone, D.A., Shain, C.N. and LaMotte, R.H. (1991) Neurogenic hyperalgesia: the search for the primary cutaneous afferent fibers that contribute to capsaicin-induced pain and hyperalgesia. J. Neurophysiol., 66: 212-221. Blumberg, H., Haupt, P., Janig, W. and Kohler, W. (1983) Encoding of visceral noxious stimuli in the discharge patterns of visceral afferent fibers from the colon, PJIuegers Arch., 398: 33-40. Burton, M.B. and Gebhart, G.F. (1995) Effects of intracolonic acetic acid on responses to colorectal distension in the tat. Brain Res., 672: 11-82. Carstens, E. (1995) Neural mechanisms of hyperalgesia: peripheral or central sensitization?. NIPS, 10: 260-265. Castagliuolo, I.. Lamont, J.T., Qiu, B., Fleming, S.M., Bhaskar, K.R., Nikulasson, S.T., Kometsky, C. and Pothoulakis, C. (1996) Acute stress causes mucin release from rat colon: role of corticotropin releasing factor and mast cells. Am. J. Physiol., 271: G884-892. Cervero, F., Handwerker, H.O. and Laird, J.M.A. (1988) Prolonged noxious stimulation of the rat’s tail: responses and encoding properties of dorsal horn neurons. .I. Physiol. (Land.), 404: 419-436. Cervero, F., Laird, J.M.A. and Pozo, M.A. (1992) Selective changes of receptive field properties of spinal nociceptive neurons induced by noxious visceral stimulation in the cat. Pain, 51: 335-342. Coderre, T.J. and Melzack, R. (1985) Cutaneous hyperalgesia: contribution of the peripheral and central nervous systems to the increase in pain sensitivity after injury. Bruin Res., 404: 95-106. Coderre, T.J. and Melzack, R. (1987) Increased pain sensitivity following heat injury involves a central mechanism. Behav. Bruin Rex, 15: 259-262.

385 Coderre, T.J., Katz, J., Vaccarino, A.L. and Melzack, R. (1993) Contribution of central neuroplasticity to pathological pain: review of clinical and experimental evidence. Pain, 52: 259285. Coutinho, S.V. and Gebhart, G.F. (1996) Involvement of nitric oxide in zymosan-produced visceral hyperalgesia in the rat. Sot. Neurosci. Abstc, 22: 1812. Coutinho, S.V., Meller, S.T. and Gebhart, G.F. (1996) Intracolonic zymosan produces visceral hyperalgesia in the rat that is mediated by spinal NMDA and non-NMDA receptors. Brain Rex, 73: 7-15. Coutinho, S.V., Sengupta, J.N. and Gebhart, G.F. (1997) Sensitized afferents from the inflamed colon are not involved in the maintenance of visceral hyperalgesia. Sot. Neurosci. Abstx, 23: 1002. Coutinho, S., Urban, M.O. and Gebhart, G.F. (1998) Role of glutamate receptors and nitric oxide in the rostra1 ventromedial medulla in visceral hyperalgesia. Pain, 78: 59-60. Dickenson, A.H. and Sullivan, A.F. (1990) Differential effects of excitatory amino acid antagonists on dorsal horn nociceptive neurones in the rat. Brain Rex, 506: 31-39. Dougherty, P.M., Sluka, K.A., Sorkin, K.N., Westlund, K.N. and Willis, W.D. (1992) Neural changes in acute arthritis in monkeys, 1. Parallel enhancement of responses of spinothalamic tract neurons to mechanical stimulation and excitatory amino acids. Brain Rex Rev., 17: I-13. Garrison, D.W., Chandler, M.J. and Foreman, R.D. (1992) Viscerosomatic convergence onto feline spinal neurons from esophagus, heart and somatic fields: effect of inflammation. Pain, 49: 373-382. Gebhart, G.F. (1993) Visceral pain mechanisms. In: C.R. Chapman and K.M. Foley (Eds.), Current and Emerging Issues in Cuncer Pain: Reseurch and Practice. Raven Press. New York, pp. 99-l 1 I. Gerber, G., Youn, D.-H., Hsu, C.H., Isaev, D. and Randic, M. (2000) Spinal dorsal horn synaptic plasticity: involvement of group I metabotropic glutamate receptors. In: J. Sandkiihler, B. Bromm and G.F. Gebhart (Eds.), Nervous System Plasticity and Chronic Pain. Progress in Brain Research. Vol. 129. Elsevier, Amsterdam, pp. 115-133. Habler, H.-J., Jlnig, W. and Koltzenburg, M. (1990) Activation of unmyehnated afferent fibers by mechanical stimuli and inflammation of the urinary bladder in the cat. J, Physiol. (Land.), 425: 545-562. Habler, H.-J., Jlnig, W. and Koltzenburg, M. (1993) Receptive properties of myelinated primary afferents innervating the inflamed urinary bladder of the cat. J. Neurophysiol., 69: 395405. Haley, J.E., Sullivan, A.F. and Dickenson, A.H. (1990) Evidence for spinal N-methyl-D-aspartate receptor involvement in prolonged chemical nociception in the rat. Brain Rex, 5 18: 218226. Haley, J.E., Dickenson, A.H. and Schachter, M. (1992) Electrophysiological evidence for a role of nitric oxide in prolonged chemical nociception in rat. Neuropharmacology, 3 1: 25 I258. Handwerker. H.O., Forster, C. and Kirchhoff, C. (1991) Dis-

charge properties of human C-fibers induced by itching and burning stimuli. J. Neurophysiol., 66: 307-3 15. Hardy, J.D., Wolff, H.G. and Goodell, H. (1950) Experimental evidence on the nature of cutaneous hyperalgesia. J. C/in. Invest., 29: 115-140. Hoheisel, U. and Mense, S. (1989) Long-term changes in discharge behavior of cat dorsal horn neurones following noxious stimulation of deep tissues. Pain, 36: 231-247. Hoheisel, U. and Mense, S. (2000) The role of spinal nitric oxide in the control of spontaneous pain following nociceptive input. In: J. Sandkiihler, B. Bromm and G.F. Gebhart (Eds.), Nervous System Plasticity and Chronic Pain. Progress in Brain Research, Vol. 129. Elsevier, Amsterdam, pp. 163- 172. Hylden, J.L.K.. Nahin, R.L., Traub, R.J. and Dubner, R. (1989) Expansion of receptive fields of lamina I projection neurones in rats with unilateral adjuvant-induced inflammation: the contribution of dorsal horn mechanisms. Pain, 37: 229-243. Ide, Y., Maehara, Y., Tsukahara, S., Kitahata, L.M. and Collins, J.G. (1997) The effects of an intrathecal NMDA receptor antagonist (AP5) on the behavioral changes induced by colorectal inflammation with turpentine in rats. Life Sci., 60: 1359-1363. J&rig, W. and Koltzenburg, M. (1986) The neural basis of consciously perceived sensations from the gut. In: F. Cervero and J.F.B. Morrison (Eds.), Visceral Sensution. Progress in Brain Research, Vol. 67. Elsevier, Amsterdam, pp. 255-277. Janig, W. and Koltzenburg, M. (1990) On the function of spinal primary afferents supplying the colon and urinary bladder. L Auton. New Syst., 30: 589-596. Kenshalo Jr., D.R., Leonard, R.B., Chung, J.M. and Willis, W.D. (1982) Facilitation of the responses of primate spinothalamic cells to cold and mechanical stimuli by noxious heating of the skin. Pain, 12: 141-152. Kolhekar, R. and Gebhart, G.F. (I 994) NMDA and quisqualate modulation of visceral nociception in the rat. Bruin Res., 651: 215-226. Kumazawa, T. and Mizumura, K. (1980) Chemical responses of polymodal receptors of the scrotal contents in dogs. J. Physiol. (Land.), 299: 2 19-23 I. LaMotte, R.H., Shain, C.N., Simone, D.A. and Tsai, E.-F.P. (1991) Neurogenic hyperalgesia: psychophysical studies of underlying mechanisms. J. Neurophysiol., 66: 190-21 I, LaMotte, R.H., Lundberg, L.E.R. and Torebjork, H.E. (1992) Pain, hyperalgesia and activity in nociceptive C units in humans after intradermal injection of capsaicin. J. Physiol. (Land.), 448: 749-764. Langlois, A., Diop, P., Riviere, P.J.M., Pascaud, X. and Junien, J.-L. (1994) Effect of fedotozine on the cardiovascular pain reflex induced by distension of the irritated colon in the anesthetized rat. EM J. Pharmacol., 27 I : 245-25 1, Lewis, T. (1942) Pain. hdacmillan, London, Longhurst, J.C. and Dittmann, L.E. (1987) Hypoxia, bradykinin and prostaglandins stimulate ischemically sensitive visceral afferents. Am. J. Physiol., 253: H556-H567. Martinez, V., Wang, L.., Mayer, E.A. and Tache, Y. (1998) Proximal colon distension increases Fos expression in the lumbosacral spinal cord and activates sacral parasympathetic

386 NADPH-d-positive neurons in rats. J. Camp. Neural., 390: 31 l-321. Mayer, E.A. and Gebhart, G.F. (1994) Basic and clinical aspects of viscera1 hyperalgesia. Gasttuenterology, 107: 27 l-293. McMahon, S.B. (1988) Neuronal and behavioral consequences of chemical inflammation of rat urinary bladder. Agents Actions, 25: 231-233. McMahon, S.B. and Wall, PD. (I 984) Receptive fields of rat lamina I projection cells move to incorporate a nearby region of injury. Pain, 19: 235-247. Menetrey, D. and Besson, J.-M. (1982) Electrophysiological characteristics of dorsal horn cells in rats with cutaneous inflammation resulting from chronic arthritis. Pain, 13: 343364. Nadaud, S. and Soubrier, E (1996) Molecular biology and molecular genetics of nitric oxide synthases. Clin. Exp. Hypertension, 18: 113-143. Ness, T.J. and Gebhart, G.F. (1987) Characterization of neuronal responses to noxious visceral and somatic stimuli in the medial lumbosacral spinal cord of the rat. J. Neurophysiol., 57: 18671892. Ness, T.J. and Gebhart, G.F. (1988a) Colorectal distension as a noxious visceral stimulus: physiologic and pharmacologic characterization of pseudaffective reflexes in the rat. Brain Res., 450: 153-169. Ness, T.J. and Gebhart, G.F. (1988b) Colorectal distension as a noxious visceral stimulus: physiologic and pharmacologic characterization of pseudaffective reflexes in the rat. Bruin Rex, 45: 153-169. Ness, T.J. and Gebhart, G.F. (1988~) Characterization of neurons responsive to noxious colorectal distension in the T13-L2 spinal cord of the rat. J. Neurophysiol., 6: 1419-1438. Ness, T.J. and Gebhart, G.F. (1991) Central mechanisms of visceral pain. Can. J. Physiol. Pharmacol., 69: 627-634. Ness, T.J., Randich, A. and Gebhart, G.F. (1991) Further behavioral evidence that colorectal distension is a ‘noxious’ viscera1 stimulus in rats. Neurosci. Lett., 13 I : 1 13-l 16. Paya, M., Garcia, PP., Coloma, J. and Alcaraz, M.J. (1997) Nitric oxide synthase and cyclooxygenase pathways in the inflammatory response induced by zymosan in the rat air pouch. BI: J. Pharmacol., 120: 1445-1452. Per], E.R. (1985) Unravelling the story of pain. In: H.L. Fields, R. Dubner and F. Cervero (Eds.), Advances in Pain Research and Therapy, Vol. 9. Raven Press, New York, pp. 1-29. Price, D.D., Hayes, R.L., Ruda, M. and Dubner, R. (1978) Spatial and temporal transformations of input to spinothalamic tract neurons and their relation to somatic sensations. J. Neurophysiol., 4 I : 933-946. Reeh, PW., Kocher, L. and Jung, S. (1986) Does neurogenic inflammation alter the sensitivity of unmyelinated nociceptors in the rat?. Brain Rex, 384: 42-50. Rice, A.S.C. (1995) Topical spinal administration of a nitric oxide synthase inhibitor prevents the hyper-reflexia associated with a rat mode1 of persistent visceral pain. Neurosci. Len., 187: 111-114. Rice, A.S.C. and McMahon, S.B. (1994) Pre-emptive intrathecal administration of an NMDA receptor antagonist (AP5) pre-

vents hyper-reflexia in a model of persistent visceral pain. Pain, 57: 335-340. Sandktlhler, J., Benrath, J., Brechtel, C., Ruscheweyh, R. and Heinke, B. (2000) Synaptic mechanisms of hyperalgesia. In: J. Sandkiihler, B. Bromm and G.F. Gebhart (Eds.), Nervous System Plasticity and Chronic Puin. Progress in Brain Research, Vol. 129. Elsevier, Amsterdam, pp. 8 l- 100. Schaible, H.-G., Schmidt, R.F. and Willis, W.D. (1987) Enhancement of the responses of ascending tract cells in the cat spinal cord by acute inflammation of the knee joint. Exp. Bruin Rex, 66: 489-499. Schepelmann, K., Messlinger, K., Schaible, H.G. and Schmidt. R.F. (1992) Inflammatory mediators and nociception in the joint: excitation and sensitization of slowly conducting afferent tibers of cat’s knee by prostaglandin 12. Neuroscience, 50: 237-247. Sengupta, J.N. and Gebhart, G.F. (1994) Gastrointestinal afferent fibers and sensation. In: L.R. Johnson (Ed.), Physiology of the Gastrointestinal Tract, 3rd edn. Raven Press, New York, pp. 483-519. Sengupta, J.N., Su, X. and Gebhart, G.F. (1996) Kappa, but not mu or delta, opioids attenuate responses to distension of afferent fibers innervating the rat colon. Gnstroenterology, 11 I: 968-980. Simone, D.A., Sorkin, L.S., Oh, U., Chung, J.M., Owens, C., LaMotte, R.H. and Willis, W.D. (1991) Neurogenic hyperalgesia: central neural correlates in responses of spinothalamic tract neurons. J. Neurophysiol.. 66: 228-246. Su, X., Sengupta, J.N. and Gebhart, G.F. (1997a) Effects of kappa opioid receptor-selective agonists on responses of pelvic nerve afferents to noxious colorectal distension. J. Neurophysiol., 78: 1003-1012. Su, X., Sengupta, J.N. and Gebhart, G.F. (1997b) Effects of opioids on mechanosensitive pelvic nerve afferent fibers innervating the urinary bladder of the rat. J. Neurophy.sinl., 77: 1566-1580. Svendsen, F., Hole, K. and Tjolsen, A. (2000) Long-term potentiation in single WDR neurons induced by noxious stimulation in intact and spinalized rats. In: J. Sandkiihler, B. Bromm and G.F. Gebhart (Eds.), Nervous System Plasticity and Chronir Pam. Progress in Brain Research, Vol. 129. Elsevier. Amsterdam, pp. 153-161. Taiwo, Y.O. and Levine, J.D. (1989) Prostaglandin effects after elimination of indirect hyperalgesic mechanisms in the skin of the rat. Bruin Res., 492: 397-399. Tomlinson, A.. Appleton, I.. Moore, A.R., Gilroy, D.W., Willis, D., Mitchell, J.A. and Willoughby, D.A. (1994) Cyclooxygenase and nitric oxide synthase isoforms in rat carrageenin-induced pleurisy. BI: J. Pharmucol., 113: 693-698. Torebjork, H.E.. Lundberg, L. and LaMotte, R.H. (1992) Central changes in processing of mechanoreceptive input in capsaicin-induced secondary hyperdlgesia in humans. J. Physiol., 448: 765-780. Traub, R.J., Pechman, P., Iadarola, M.J. and Gebhart, G.F. ( 1992) Fos-like proteins in the lumbosacral spinal cord following noxious and non-noxious CRD in the rat. Pam, 49: 393-403. Traub, R.J., Herdegen. T. and Gebhart, G.F. (1993) Differential

387

expression of c-fos and c-jun in 2 regions of the rat spinal cord following noxious colorectal distension. Neurosci. Left., 160: 121-125. Treede, R.-D., Meyer, R.A., Raja, S.N. and Campbell, J.N. (1992) Peripheral and central mechanisms of cutaneous hyperalgesia. Prog. Neurobiol., 38: 397-42 1. Urban, M.O. and Gebhart, GE (1999) Supraspinal contributions to hyperalgesia. Proc. Natl. Acad. Sci., 96: 7681-7692. Vane, J.R., Mitchell, J.A., Appleton, I., Tomlinson, A., Bishop, B.D., Croxtall, J. and Willoughby, D.A. (1994) Inducible isoforms of cyclooxygenase and nitric oxide synthase in inflammation Proc. Nat/. Acud. Sci., 91: 2046-2050. Wall, P.D. and Woolf, C.J. (1984) Muscle but not cutaneous C-afferent input produces prolonged increases in the excitability of the flexion reflex in the rat. J. Physiol. (Land.), 356: 443-458. Wang, J.F., Khasar, S.G., Ahlgren, S.C. and Levine, J.D. (1996) Sensitization of C-fibers by prostaglandin E2 in the rat is inhibited by guanosine 5’.0-(2thiodiphosphate), 2’,5’-dideoxyadenosine and Walsh inhibitor peptide. Neumscience, 7 1: 259-263. Willis, W.D., Jr. (1993) Central sensitization and plasticity fol-

lowing intense noxious stimulation. In: E.A. Mayer and H.E. Raybould (Eds.), Basic and Clinical Aspects of Chronic Abdominnl Pain. Elsevier, Amsterdam, pp. 202-217. Woolf, C.J. (1983) Evidence for a central component of postinjury pain hypersensitivity. Nature, 306: 686-688. Woolf, C.J. (1984) Long term alterations in the excitability of the flexion reflex produced by peripheral tissue injury in the chronic decerebrate rat. Pain, 18: 325-343. Woolf, C.J. (1991) Central mechanisms of acute pain. In: M.R. Bond, J.E. Charlton and C.J. Woolf (Eds.), P&n Research and Clinical Management. Pmt. Vlth World Congress on Pain. Vol. 4. Elsevier, Amsterdam, pp. 25-34. Woolf. C.J. and King, A.E. (1990) Dynamic alterations in the cutaneous mechanoreceptive fields of dorsal horn neurons in the rat spinal cord. J. Neurosci., IO: 27 17-2726. Woolf, C.J. and Thompson, S.W.N. (1991) The induction and maintenance of central sensitization is dependent on N-methyl-D-aspartic acid receptor activation: implications for the treatment of postoperative pain. Pain, 44: 293-301. Yun, H.-Y., Dawson, V.L. and Dawson, T.M. (1996) Neurobiology of nitric oxide. Crit. Rev. Neurobiol.. 10: 291-316.

I. Sandktihlt?r, B. Bromm and G.F. Gebhart (Eds.) Progress in Brain Research, Vol. 129 0 2000 Elsevier Science B.V. All rights reserved

CHAPTER

29

Cellular and neurochemical remodeling of the spinal cord in bone cancer pain Prisca Honor6 ‘, Matthew J. Schwei ‘, Scott D. Rogers ‘, JaneenL. Salak-Johnson‘, Matthew P. Finke ‘, Margaret L. Ramnaraine2, Denis R. Clohisy 2 and Patrick W. Mantyh ‘3” ’ Neurosystems

Center;

2 Department

Departments 18-208 of Orthopaedic

of Preventive Sciences, Psychiatry, Neumscience and Cancer Center; Universig of Minnesota, Moos Tower; 515 Delaware Street, Minneapolis, MN 55455, USA and VA Medical Center; Minneapolis, MN 55417, USA Surgery and Cancer Center; University of Minnesota, Minneapolis, MN 55455, USA

Introduction Over 1 million patients suffer from cancer-related pain each year. Pain is the first symptom of cancer in 20-50% of all cancer patients and 75-90% of advanced or terminal cancer patients must cope with chronic pain syndromes related to failed treatment and/or tumor progression (Portenoy and Lesage, 1999). The two most difficult cancer pains to treat are those related to invasion of peripheral nerves and destruction of bone. They account for approximately 75% of all chronic cancer pain (Banning et al., 1991; Coleman, 1998; Foley, 1999). Episodes of intense pain, or breakthrough pain occur in both of these cancer-induced pain states and represent a serious and debilitating clinical problem (Mercadante and Arcuri, 1998; Portenoy et al., 1999). The greatest obstacle to developing new treatments for persistent cancer pain and/or optimally coordinating existing treatments is a paucity of knowledge of the basic neurobiology of cancer pain. There * Corresponding author: P.W. Mantyh, Neurosystems Center, 18-208 Moos Tower, 515 Delaware Street, Minneapolis, MN 554.55, USA. Tel.: +I-612-626-0180; Fax: + l-6 12-626-2565; E-mail: [email protected]

is no well-accepted animal model of cancer pain and the majority of what we know about the neurochemistry of cancer pain has been obtained from clinical studies on how to best manage pain in cancer patients. Studies on the sensory and sympathetic innervation of human tumors suggest that there is relatively little direct neural innervation of tumors (O’Connell et al., 1998). However, malignant cells are known to secrete prostaglandins, cytokines, epidermal growth factor, transforming growth factor and platelet-derived growth factor, many of which have been shown to excite primary afferent nociceptors (Hingtgen and Vasko, 1994; Suzuki and Yamada, 1994; Vasko et al., 1994; Watkins et al., 1994; Hingtgen et al., 1995; Safieh-Garabedian et al., 1995; Vasko, 1995; Woolf et al., 1997; see also Reeh and Pethii, 2000, this volume; Sutherland et al., 2000, this volume). Additionally, macrophages, which can represent more than 20-30% of the cells in the tumor mass (McBride, 1986), produce factors such as tumor necrosis factor and interleukin-1 that have been reported to excite primary afferent neurons (Watkins et al., 1994; Safieh-Garabedian et al., 1995; Sorkin et al., 1997; Woolf et al., 1997). To determine the neurochemical mechanisms that give rise to cancer pain, we have developed a model

390

of bone cancer pain that shares many similarities with human cancer bone pain (Mercadante, 1997). Following development of the model, we have characterized the extent of cancer-induced bone destruction, the sensory innervation of the bone, the animal behavior indicative of pain, and the neurochemical changes that occur in the spinal cord and primary afferent neurons that may be involved in the generation and maintenance of cancer pain.

A model for bone cancer pain? In assessing any experimental animal model, it is important to determine how well the model approximates the human disease. The most common symptom of bone metastases in humans is bone pain. Bone destruction which causes this pain can lead to pathological fractures and/or hypercalcemia (Lipton, 1997; Fulfaro et al., 1998). Over weeks or months, as the tumor grows and stimulates bone destruction, the pain intensifies and can incapacitate the affected individual. The severity and the frequency of breakthrough pain, intermittent episode of extreme pain which occurs spontaneously or more commonly by weight bearing or movement of the affected bone, is correlated with the extent of bone destruction (Mercadante and Arcuri, 1998; Portenoy et al., 1999). The murine model appears to share many features of human bone cancer pain. The osteolytic sarcoma cell line used in this study aggressively destroys bone (see Fig. 1) and provides localized pathologic findings similar to those found in humans with osteolytic bone cancer (Clohisy et al., 1995, 1996a). Mice with bone cancer exhibit painful behavior in the form of guarding of the affected limb. This guarding behavior is correlated with the extent of bone destruction.

Severe acute pain is also observed in mice after significant bone destruction has occurred, as normally non-noxious palpation of the affected bone results in behaviors indicative of severe pain and this severe pain is again correlated with the extent of bone destruction.

What factors contribute to bone cancer pain? Primary afferent sensory neurons innervate mineralized bone and the periosteum, a fibrous tissue covering the outside surface of the bone. What is clear from the present and previous studies is that while mineralized bone and the marrow are innervated by sensory neurons, this innervation is sparse compared to the sensory innervation of the periosteum (Bjurholm et al., 1988a,b; Hill and Elde, 199 1; Hukkanen et al., 1992; Tabarowski et al., 1996). The sensory innervation of the tumor itself has been explored and the general conclusion is that while the occasional sensory fiber can be observed innervating the tumor, this innervation is sparse and when present, is usually associated with the blood vessels that vascularize the tumor (O’Connell et al., 1998). In light of the sensory innervation of the bone and periosteum and our understanding of tumor and bone biology, what drives bone cancer pain and why does this type of pain increase so dramatically over time? Most osteolytic tumors release a variety of factors that induce excessive osteoclast activity (Taube et al., 1994; Clohisy et al., 1995, 1996a,b; Clohisy and Ramnaraine, 1998). With continued tumor-driven osteoclast bone resorption, the osteoclasts and tumor cells may increasingly come in contact with sensory nerve fibers in the bone and the richly innervated periosteum. Many primary afferent neurons that in-

Fig. 1. Quantification of bone destruction following injection of osteolytic sarcoma cells into the femoral intramedullary space. Hematoxylin/eosin staining of normal (A) and 21.day sarcoma-bearing (B) femora showing the replacement within the intramedullary space of the darkly stained marrow cells with the more lightly stained sarcoma cells and areas where significant bone destruction has occurred (arrow in B). Higher magnification of the distal portion of the normal (Al) and sarcoma-bearing femora (Br ) illustrating that in normal bone, there is a clear separation of normal bone and marrow cells (A)) whereas in animals with osteolytic sarcoma cells in the intramedullary space (Br ), the sarcoma cells have largely replaced the marrow cells and induced destruction through the bone (arrowhead) and beyond (arrow). Radiographs of the femur (C) showing the progressive loss of bone due to tumor growth. Bone destruction was quantified on a 0 to 3 scale based on the loss of bone. Images 0 to 3 are examples of each state of destruction: 0 = normal bone; 1 = minor loss of bone in medullary canal (arrow); 2 = substantial loss of bone in medullary canal with some destruction of the femoral head (arrow); 3 = substantial loss of bone in medullary canal with major structural destruction of the distal end of the femur (arrow), scale bar = 2 mm in A, B, and C and 200 (*m in Al and Br

391

nervate the periosteum have been shown to express acid sensing ion channels (Olson et al., 1998) and it is probable that the decrease in pH produced by both nearby osteoclasts (pH 4.5-4.8) and tumor cells

(pH 6.0 and 7.0) directly sensitize and/or excite Periosteal primary afferent fibers (see Reeh and Pethe, 2000, this volume). In addition, as already stated, tumor cells release an array of growth factors, CY-

392 tokines, and chemokines, many of which have been reported to also directly excite an/or sensitize primary afferent fibers (see Sutherland et al., 2000, this volume). In the present study, we used two markers, substance P receptor (SPR) internalization and c-Fos expression in lamina I neurons, to indirectly demonstrate that primary afferent neurons are sensitized following extensive tumor-induced bone destruction. In previous studies, it has been shown that, whereas in the normal animal only noxious stimulation results in the release of SP and the subsequent internalization of the SPR in lamina I neurons, in animals with either persistent inflammatory or neuropathic pain, normally non-noxious or noxious somatosensory stimulation now induces SPR internalization in lamina I neurons (Mantyh et al., 1995; Abbadie et al., 1997; Honore et al., 1999). Similarly, whereas in the normal animal, noxious stimulation is required to induce c-Fos expression in lamina I neurons (Hunt et al., 1987; Abbadie and Besson, 1993; Honori: et al., 1995; Doyle and Hunt, 1999), following extensive bone destruction, non-noxious palpation of the tumorous bone induces c-Fos expression (see Fig. 2). These data suggest that there is sensitization of primary afferent neurons in animals with bone cancer and this sensitization is correlated with the extent of bone destruction and growth of the tumor. Bone cancer also induces a profound reorganization of the spinal cord that may be reflective of a central sensitization that is frequently observed in persistent pain states. In the segments of the spinal cord that receive primary afferent input from the bone with cancer, and only in these segments, we observed massive astrocyte hypertrophy without neuronal loss (see Figs. 2 and 3) and expression

of the pro-hyperalgesic peptide dynorphin and c-Fos protein in a population of neurons located in the deep spinal laminae (see Fig. 2). What is remarkable about these alterations is the magnitude and localized nature of these changes. Hypertrophy of spinal astrocytes has previously been reported following sciatic nerve injury (Garrison et al., 1991, 1994; Colburn et al., 1997). However, examination of the sciatic nerve that innervates the femur with cancer revealed no sign of direct physical injury and astrocyte hypertrophy was not observed in sham injected animals or when the tumor was growing exclusively outside the bone. Whether the massive astrocyte hypertrophy is involved in the generation and maintenance of bone cancer pain is unclear but it has previously been shown that astrocytes express glutamate/aspartate transporters and thus are intimately involved in regulating the extracellular levels of excitatory amino acids (Hansson and Riinnback, 1991; Shao and McCarthy, 1994; Shibata et al., 1997). Additionally, astrocytes that have undergone hypertrophy have been shown to release a variety of cytokines and growth factors that can dramatically alter the surrounding neurochemical environment (Maimone et al., 1993; Murphy et al., 1993; Pechan et al., 1993; Derocq et al., 1996; Grimaldi et al., 1997; Shafer and Murphy, 1997; Bruno et al., 1998). Indeed, the population of neurons that show up-regulation of the pro-hyperalgesic neuropeptide dynorphin and increased c-Fos expression are in the deep laminae (see Fig. 2) and in close proximity to the astrocytes showing marked hypertrophy (see Figs. 2 and 3). While changes in c-Fos and dynorphin have been associated with changes observed in inflammatory and neuropathic pain states (Ruda et al., 1988; Weihe et al., 1988;

Fig. 2. Nemochemical changes in the dorsal horn of the spinal cord 21 days following unilateral injection of an osteolytic sarcoma in the intramedullary space of the femur. Confocal images of coronal sections of the L4 spinal cord illustrate the distribution of: (A) the astrocyte marker glial fibrillary acidic protein (GFAP); (B) dynorphin (DYN) with arrows indicating the cell bodies expressing this pro-hyperalgesic peptide; (C) c-Fos protein in the basal unstimulated state; (D) c-Fos protein at 1 h following normally non-noxious palpation of the knee; (E) substance P (SP); (F) protein kinase Cy isoform (PKCy). Note that the major changes occur in the spinal cord ipsilateral to the cancer-bearing femur and include an increase in GFAP (A), DYN (B), basal c-Fos expression (C) and increased expression of c-Fos in neurons located in laminae I and II following normally non-noxious palpation (D). In contrast, levels of SP (E), a peptide contained in primary afferent neurons that is frequently up-regulated in persistent pain states and PKCy (F), a kinase that is expressed in a subset of spinal neurons in lamina II and that is frequently up-regulated in neuropathic pain states, remained unchanged. These images are from 60-urn-thick tissue sections projected from 10 optical sections acquired at 5-(*m intervals with a 20x lens, scale bar = 200 urn.

393

395 Noguchi et al., 1991; Abbadie and Besson, 1993; Wagner et al., 1993; Honor6 et al., 1995; Catheline et al., 1999), these data suggest that bone cancer induces a profound neurochemical reorganization of the spinal cord that is directly correlated with the extent of cancer-induced bone destruction. Neurochemical signature of inflammatory, neuropathic, and bone cancer pain In recent years, significant progress has been made in demonstrating that in different persistent pain states there are strikingly different neurochemical changes that occur in primary afferent neurons and the spinal cord. These neurochemical differences mirror the fact that many analgesics are most efficacious in blocking a specific type of persistent pain but inefficacious with others (Fields, 1988). In comparing bone cancer pain to inflammatory or neuropathic pain, both the neurochemical changes that take place and the effective analgesics in treating humans suggest that the mechanisms involved in the generation and maintenance of bone cancer pain are unique. For example, whereas SP levels in primary afferent neurons rise in inflammatory pain (Lembeck et al., 1981; Donaldson et al., 1992) and decrease in neuropathic pain (Noguchi et al., 1989; Garrison et al., 1993), they are not altered in bone cancer pain (see Fig. 3). Even more striking are the changes observed in astrocyte hypertrophy in the spinal cord. Astrocyte hypertrophy in the spinal cord is uncommon in most models of inflammatory pain and is only observed in neuropathic pain states when there has been significant injury to the peripheral nerve (Garrison et al., 1991, 1994). In contrast, although there is no evidence of direct injury to the peripheral

nerve, massive astrocyte hypertrophy is observed in the bone cancer model (see Fig. 3). Conclusions The unique neurochemical reorganization of the spinal cord in bone cancer is mirrored by the clinical experience that analgesics that are efficacious in the relief of inflammatory or neuropathic pain are frequently ineffective at relieving advanced bone cancer pain. Understanding the distinct neurochemical events that are involved in the generation and maintenance of different persistent pain states should provide a mechanistic approach for understanding and developing novel therapies for unique persistent pain states such as cancer pain. Acknowledgements This work was supported by a Merit Review from the Veterans Administration, NIH grants NS23970, AGI 1852, NIHDA 11986 and the Roby C. Thompson, Jr. Endowment in Musculoskeletal Oncology. References Abbadie, C. and Besson, J.-M. (1993) c-fos expression in rat lumbar spinal cord following peripheral stimulation in adjuvan&induced arthritis and in normal rats. Bruin Res., 607: 195-204. Abbadie, C.. Trafton, J., Liu, H., Mantyh, PW. and Basbaum, A.I. (1997) Inflammation increases the distribution of dorsal horn neurons that internalize the neurokinin-1 receptor in response to noxious and non-noxious stimulation. J. Neurosci., 17: 8049-8060. Banning, A., Sjogren, P. and Henriksen, H. (I 99 1) Pain causes in 200 patients referred to a multidisciplinary cancer pain clinic. Pain, 45: 45-48.

Fig. 3. Confocal images showing the increase in the astrocyte marker glial fibrillary acidic protein (GFAP) in coronal sections of the L4 spinal cord 21 days following injection of osteolytic sarcoma cells into the intramedullary space of the femur. In (A, B and C) the GFAP is bright orange and in (D and E) GFAP is green and the NeuN staining (which labels neurons) is in red. A low-power image (A) shows that the up-regulation of GFAP is almost exclusively ipsilateral to the femur with cancer with a small increase in the contralateral spinal cord in lamina X. Higher magnification of GFAP contralateral (B, D) and ipsilateral (C, E) to the femur with cancer shows that on the ipsilateral side, there is marked hypertrophy of astrocytes characterized by an increase in both the size of the astrocyte cell bodies and the extent of the arborization of their distal processes. Additionally, this increase in GFAP (green) is observed without a detectable loss of neurons, as NeuN (red) labeling remained unchanged (D and E). These images, from 60.pm-thick tissue are projected from 6 optical sections acquired at 4-pm intervals with a 20x lens, scale bar = 200 pm (A), are projected from 12 optical sections acquired at 0.8pm intervals with a 100x lens, scale bar = 20 pm, (B and C) and are projected from 10 optical sections acquired at O&pm intervals with a 60x lens, scale bar = 30 urn (D and E).

396

Bjurholm, A., Kreicbergs, A., Brodin, E. and Schultzberg, M. (1988a) Substance P- and CGRP-immunoreactive nerves in bone. Peptides, 9: 165- 17 1. Bjurholm, A., Kreicbergs. A., Terenius, L., Goldstein. M. and Schultzberg. M. (1988b) Neuropeptide Y-, tyrosine hydroxylase- and vasoactive intestinal polypeptide-immunoreactive nerves in bone and surrounding tissues. J. Auton. Nen! Sysr., 25: 119-125. Bruno, V., Battaglia, G.. Casabona. G., Copani, A., Caciagli, F. and Nicoletti, F. (1998) Neuroprotection by glial metabotropic glutamate receptors is mediated by transforming growth factor-beta L Neurosci., 18: 9594-9600. Catheline, G., Le Guen, S., Honor& P. and Besson, J.-M. (1999) Are there long-term changes in the basal or evoked Fos expression in the dorsal horn of the spinal cord of the mononeuropathic rat?. Pain, 80: 347-357. Clohisy, D.R. and Ramnaraine, M.L. (1998) Osteoclasts are required for bone tumors to grow and destroy bone. J. Orthopued. Re,v,, 16: 660-666. Clohisy, D.R., Ogilvie, C.M. and Ramnaraine, M.L. (1995) Tumor osteolysis in osteopetrotic mice. J. Orthopaed. Rex, 13: 892-897. Clohisy, D.R., Ogilvie, C.M., Carpenter. R.J. and Ramnaraine, M.L. (1996a) Localized, tumor-associated osteolysis involves the recruitment and activation of osteoclasts. J. Orthopued. Rex, 14: 2-6. Clohisy, D.R., Palkert, D., Ramnaraine, M.L., Pekurovsky, 1. and Oursler. M.J. (1996b) Human breast cancer induces osteoclast activation and increases the number of osteoclasts at sites of tumor osteolysis. J. Orthopaed. Res., 14: 396-402. Colburn. R.W., DeLeo, J.A., Rickman, A.J., Yeager, M.P., Kwon, P. and Hickey, W.F. (1997) Dissociation of microglial activation and neuropathic pain behavior following peripheral nerve injury in the rat. J. Neuroimmunol., 79: 163-175. Coleman, R.E. (1998) How can we improve the treatment of bone metastases further?. Curr: Opin. Oncol.. 10: S7-13. Derocq. J.M., Segui, M., Blazy, C., Emonds-Alt, X., Le Fur, G., Brelire, J.C. and Casellas, P. (1996) Effect of substance P on cytokine production by human astrocytic cells and blood mononuclear cells: characterization of novel tachykinin receptor antagonists. FEBS Lett., 399: 321-325. Donaldson. L.F., Harmar, A.J., McQueen. D.S. and Seckl, J.R. ( 1992) Increased expression of preprotachykinin, calcitonin gene-related peptide, but not vasoactive intestinal peptide messenger RNA in dorsal root ganglia during the development of adjuvant monoarthritis in the rat. Mol. Brain Res., 16: 143149. Doyle, C.A. and Hunt, S.P. (1999) Substance P receptor (neurokinin-I)-expressing neurons in lamina I of the spinal cord encode for the intensity of noxious stimulation: a c-Fos study in rat. Neuroscience. 89: 17-28. Fields, H.L. (1988) Can opiates relieve neuropathic pain?. Puin, 35: 365-367. Foley, K.M. (1999) Advances in cancer pain. Arch. Neural., 56: 413-417. Fulfaro. F.. Casuccio, A., Ticozzi, C. and Ripamonti, C. (1998) The role of bisphosphonates in the treatment of painful

metastatic bone disease: a review of phase III trials. Pain, 78: 157-169. Garrison, C.J., Dougherty, P.M., Kajander. K.C. and Carhon, S.M. (1991) Staining of glial fibrillary acidic protein (GFAP) in lumbar spinal cord increases following a sciatic nerve constriction injury. Bruin Rex, 565: 1-7. Garrison, C.J., Dougherty, P.M. and Carhon, S.M. (1993) Quantitative analysis of substance P and calcitonin gene-related peptide immunohistochemical staining in the dorsal horn of neuropathic MK-801.treated rats. Brain Rex, 607: 205-214. Garrison, C.J., Dougherty, P.M. and Carlton, S.M. (1994) GFAP expression in lumbar spinal cord of naive and neuropathic rdtS treated with MK-801. Exp. Neural., 129: 237-243. Grimaldi, M., Florio, T. and Schettini, G. (1997) Somatostatin inhibits interleukin-6 release from rat cortical type I astrocytes via the inhibition of adenylyl cyclase. Biochem. Biophy. Rex Commun., 235: 242-248. Hansson, E. and Ronnbkk, L. (1991) Receptor regulation of the glutamate. GABA and taurine high-affinity uptake into astrocytes in primary culture. Bruin Rex, 548: 2 15-22 I. Hill, E.L. and Elde, R. (1991) Distribution of CGRP-, VIP-, DbH-, SP-, and NPY-immunoreactive nerves in the periosteum of the rat. Cell Tissue Rex. 264: 469-480. Hingtgen. C.M. and Vasko, M.R. (1994) Prostacyclin enhances the evoked-release of substance P and calcitonin gene-related peptide from rat sensory neurons. Brain Res., 655: 5 I-60. Hingtgen, CM., Waite, K.J. and Vasko, M.R. (1995) Prostaglandins facilitate peptide release from rat sensory neurons by activating the adenosine 3’,5’-cyclic monophosphate transduction cascade. J. Neurosci.. 15: 541 l-5419. Honor& P., Buritova, J. and Besson, J.-M. (1995) Carrageeninevoked c-Fos expression in rat lumbar spinal cord: the effects of indomethacin. Eu,: J. Phurmucol., 272: 249-259. Honor& P., Menning, P.M., Rogers, SD., Nichols, M.L.. Basbaum, AI., Besson, J.-M. and Mantyh, P.W. (1999) Spinal substance P receptor expression and internalization in acute. short-term, and long-term inflammatory pain states. 1. Neurosci., 19: 7670-7678. Hukkanen, M.. Konttinen, Y.T., Rees, R.G., Santavirta, S., Terenghi, G. and Polak, J.M. (1992) Distribution of nerve endings and sensory neuropeptides in rat synovium, meniscus and bone. Int. .I. Tissue React., 14: I-10. Hunt, S.P.. Pini, A. and Evan, G. (1987) Induction of c-fos-like protein in spinal cord neurons following sensory stimulation. Nature, 32X: 632-634. Lembeck, F., Donnerer, J. and Colpaert, EC. (1981) Increase of substance P in primary afferent nerves during chronic pain. Newopeptides, I: 175-l 80. Lipton, A. (1997) Bisphosphonates and breast carcinoma. Guncer, 80: 16681673. Maimone, D., Cioni, C., Rosa, S., Macchia, G., Aloisi. F. and Annunziata, P. (1993) Norepinephrine and vasoactive intestinal peptide induce IL-6 secretion by astrocytes: synergism with IL- I beta and TNF alpha. J. Newoimmunol., 47: 73-8 1. Mantyh, P.W., DeMaster, E., Malhotra, A., Ghilardi, J.R.. Rogers, S.D., Mantyh. CR., Liu. H., Basbaum, A.I., Vigna, S.R., Maggio, J.E. and Simone, D.A. (1995) Receptor endocytosis

397

and dendrite reshaping in spinal neurons after somatosensory stimulation. Science, 268: 1629-1632. McBride, W.H. (1986) Phenotype and functions of intratumoral macrophages. Biochim. Biophys. Acta, 865: 27-41. Mercadante, S. (1997) Malignant bone pain: pathophysiology and treatment. Pain, 69: I- 18. Mercadante, S. and Arcuri, E. (1998) Breakthrough pain in cancer patients: pathophysiology and treatment. Cancer Treat. Rev., 24: 425-432. Murphy, T.H., Blatter, L.A., Wier, W.G. and Baraban, J.M. (1993) Rapid communication between neurons and astrocytes in primary cortical cultures J. Neuvosci., 13: 267222679. Noguchi, K., Senba, E., Morita, Y., Sato, M. and Tohyama, M. (1989) Prepro-VIP and preprotachykinin mRNAs in the rat dorsal root ganglion cells following peripheral axotomy. Mol. Brain Res., 6: 327-330. Noguchi, K., Kowalski, K., Traub, R., Solodkin, A., Iadarola, M.J. and Ruda, M.A. (1991) Dynorphin expression and Foslike immunoreactivity following inflammation induced hyperalgesia are colocalized in spinal cord neurons. Mol. Brain Rex, IO: 227-233. O’Connell, J.X., Nanthakumar, S.S., Nielsen, G.P. and Rosenberg, A.E. (I 998) Osteoid osteoma: the uniquely innervated bone tumor. Mod Pathol., 11: 175-180. Olson, T.H., Rield, M.S., Vulchanova, X.R. and Elde, R. (I 998) An acid sensing ion channel (ASIC) localizes to small primary afferent neurons in rats. Neuroreport, 9: 1109-l 113. Pechan, P.A., Chowdhury, K., Gerdes, W. and Seifert, W. (1993) Glutamate induces the growth factors NGF, BFGF, the receptor FGF-RI and c-fos mRNA expression in rat astrocyte culture. Neurosci. Lett., 153: 11 l-l 14. Portenoy, R.K. and Lesage, P. (I 999) Management of cancer pain. Lancer, 353: 1695- 1700. Portenoy, R.K., Payne, D. and Jacobsen, P (1999) Breakthrough pain: characteristics and impact in patients with cancer pain. Pain, 81: 129-134. Reeh, I? and Petho, G. (2000) Nociceptor excitation by thermal sensitization - a hypothesis. In: J. Sandktihler, B. Bromm and G.F. Gebhart (Eds.), Nervous System Plastic& and Chronic Pain. Progress in Brain Research, Vol. 129. Elsevier. Amsterdam, pp. 39-50. Ruda, M.A., ladarola, M.J., Cohen, L.V. and Young, W.S. (1988) In situ hybridization histochemistry and immunocytochemistry reveal an increase in spinal dynorphin biosynthesis in a rat model of peripheral inflammation and hyperalgesia. PNAS. 85: 622-626. Safieh-Garabedian, B., Poole, S., Allchorne, A., Winter, J. and Woolf, C.J. (1995) Contribution of interleukin-1 beta to the inflammation-induced increase in nerve growth factor levels and inflammatory hyperalgesia. BI: J. Pharmacol., 115: 12651275.

Shafer, R.A. and Murphy, S. (1997) Activated astrocytes induce nitric oxide synthase-2 in cerebral endothelium via tumor necrosis factor alpha. Glia, 21: 370-379. Shao, Y. and McCarthy, K.D. (1994) Plasticity of astrocytes. Glia, 11: 147-155. Shibata, T., Yamada, K., Watanabe, M., Ikenaka, K., Wada, K., Tanaka, K. and moue, Y. (1997) Glutamate transporter GLAST is expressed in the radial glia-astrocyte lineage of developing mouse spinal cord. J. Neurosci., 17: 9212-9219. Sorkin, L.S., Xiao, W.H., Wagner. R. and Myers, R.R. (1997) Tumor necrosis factor-alpha induces ectopic activity in nociceptive primary afferent fibers. Neuroscience, 8 1: 255-262. Sutherland, S.P., Cook, S.P. and McCleskey, E.W. (2000) Chemical mediators of pain due to tissue damage and ischemia. In: J. Sandktihler. B. Bromm and G.F. Gebhart (Eds.). Nervous Systern Plastici and Chronic Pain, Progress in Brain Research, Vol. 129. Elsevier, Amsterdam, pp. 21-38. Suzuki, K. and Yamada, S. (1994) Ascites sarcoma 180. a tumor associated with hypercalcemia, secretes potent boneresorbing factors including transforming growth factor alpha, interleukin-I alpha and interleukin-6. Bone Mineral., 27: 219233. Tabarowski, Z., Gibson-Berry, K. and Felten, S.Y. (1996) Noradrenergic and peptidergic innervation of the mouse femur bone marrow. Acta Histochem., 98: 453-457. Taube, T., Elomaa. I., Blomqvist, C., Beneton, M.N. and Kanis, J.A. (1994) Histomorphometric evidence for osteoclast-mediated bone resorption in metastatic breast cancer. Bone, 15: 161-166. Vasko, M.R. (1995) Prostaglandin-induced neuropeptide release from spinal cord. Prog. Brain Rex. 104: 367-380. Vasko, M.R., Campbell, W.B. and Waite, K.J. (1994) Prostaglandin Ez enhances bradykinin-stimulated release of neuropeptides from rat sensory neurons in culture. J. Neurosci.. 14: 4987-4997. Wagner, R., DeLeo, J.A., Coombs. D.W., Willenbring, S. and Fromm, C. (1993) Spinal dynorphin immunoreactivity increases bilaterally in a neuropathic pain model. Brain Res., 629: 323-326. Watkins, L.R., Wiertelak, E.P., Goehler, L.E., Smith, K.P., Martin, D. and Maier, S.F. (1994) Characterization of cytokine-induced hyperalgesia. Brain Rex, 654: 15-26. Weihe, E.. Millan, M.J., Leibold, A., Nohr, D. and Herz, A. (I 988) Colocalization of pro-enkephalin and pro-dynorphinderived opioid peptides in laminae IV/V spinal neurons revealed in arthritic rats. Neurosci. Lett., 85: 187-192. Woolf, C.J., Allchome. .4., Safieh-Garabedian, B. and Poole, S. (1997) Cytokines. nerve growth factor and inflammatory hyperalgesia: the contribution of tumour necrosis factor alpha. BI: J. Pharmacol.. 121: 417-424.

J. Sandkitbler, B. Bromm and GE Gebhart (Eds.) Prop& in Brain Research, Vol. 129 0 2COO Elsevier Science B.V. All rights reserved

CHAPTER 30

Altered spinal processing in animal models of radicular and neuropathic pain E. Carstens* Section

of Neurobiology,

Physiology

and Behavios

Universiq

Introduction Peripheral nerve damage often leads to the development of neuropathic pain, which is characterized by spontaneous pain sensation, hyperalgesia and allodynia in the affected limb. Even more common is low-back pain, sometimes accompanied by sciatica, which may occur in as much as 80% of the population at some time during life and which has a large economic impact in terms of work disability (Kelsey and White, 1980). Until recently, little was known concerning the neural mechanisms underlying neuropathic or low-back pain. In the past 10 years, several animal models of chronic pain following peripheral nerve or dorsal root injury have been introduced. Manifestations of neuropathic and radicular pain in these models generally include thermal hyperalgesia and/or mechanical allodynia in the affected limb, as well as indicators of spontaneous pain such as altered gait, limb-guarding, autotomy, and others, which are reminiscent of the symptoms seen in humans. These symptoms generally appear shortly after the experimental nerve or dorsal root injury and persist for several weeks or months. Presumably, pathophysiological alterations in the peripheral and/or spinal processing of sensory information underlie the expression of these behavioral changes. However, there *Corresponding author: E. Carstens, Section of Neurobiology, Physiology and Behavior, University of California, 1 Shields Avenue, Davis, CA 95616, USA. Tel.: +l-916-752-6640; Fax: +l-916-752-5582; E-mail: [email protected]

of Cd(foforn&/ Shields.4venuev Davis,CA 95616,

(ISA

appear to be sometimes glaring inconsistencies in the development of pathophysiological changes in spinal nociceptive processing as compared to the behavioral symptoms. The main aims of this paper are to review certain models of neuropathic and radicular pain, and to critically examine some of the mismatches between behavioral symptoms of pain and neurophysiological changes in the spinal cord that occur following the peripheral nerve or dorsal root injury.

Animal models of radicular

pain

Compression and/or inflammation of dorsal roots and dorsal root ganglia (DRGs) can contribute to radicular pain (Garfin et al., 1995). Several studies using animal models of radicular pain have appeared recently. In the first such studies, lumbar dorsal roots (Kawakami et al., 1994a,b) or DRGs (Chatani et al., 1995) in rats were loosely ligated on one side using either double silk (4-O) or double chromic gut sutures. Dorsal root constriction (DRC) with silk did not produce thermal hyperalgesia but did result in a moderate (approximately 20%) increase in mechanical sensitivity on the operated compared to non-operated side; this resolved after 12 wk (Kawakami et al., 1994a,b). There was also a moderate motor paresis immediately following surgery that resolved within 10 days. DRC using chromic gut sutures, or placement of four pieces of chromic gut alongside the dorsal roots which were held with two chromic gut ligatures, produced motor paresis and thermal hyperalgesia, but no mechanical allodynia. The authors suggested that the thermal hyperalgesia following chromic gut may have been

400

due to an inflammatory response triggered by some factor in the chromic gut (e.g., chromium ions or pyrogallol). The inflammation might be associated with heightened levels of phospholipase A2 (Franson et al., 1992), and, indeed, it was recently reported that phospholipase A2 levels were significantly increased in the L4-L.5 dorsal root/DRGs from rats that had received DRC with chromic gut sutures (Lee et al.. 1998). The DRC-induced increase in phospholipase A2 levels was significantly attenuated by steroid treatment (Lee et al., 1998), as was thermal hyperalgesia (Hayashi et al., 1998). Another potential contributing factor during inflammation might be local acidification; this is supported by the observation that continuous experimental acidification of the sciatic nerve produced thermal hyperalgesia in rats (Maves et al.. 1995). In a recent study of rats receiving DRC with 1 or 2 loose silk (7-O) ligatures around L4-L6 dorsal roots (Tab0 et al., 1999), we confirm the pronounced mechanical allodynia following DRC which, however, exhibited a slower recovery period of approximately 22 wk (Fig. 1A). Furthermore, we also observed a pronounced ‘mirror image’ mechanical allodynia on the side contralateral to DRC (Fig. 1B). Such changes were not observed in sham-operated or suture control animals that received lengths of suture placed alongside the dorsal roots, arguing that the results were not due to inflammation consequent to a foreign body reaction. After an initial brief period of

thermal hypoalgesia (Fig. IC), there was no further change in thermal withdrawal thresholds on either side (Fig. lC,D), also in agreement with the results of Kawakami et al. (1994a,b) using silk sutures. Tight ligation of L4-L6 dorsal roots with a clip produced severe thermal and mechanical hypoalgesia (Kawakami et al., 1994a,b), and it is conceivable that the brief thermal hypoalgesia observed in our study resulted from a less-severe, reversible compression of afferent fibers. We also observed a significant reduction in weight bearing by the hind limbs ipsilateral to DRC that resolved within 5 wk, at a time when mechanical allodynia peaked. This was accompanied only during the first wk by a compensatory increase in weight bearing by the opposite limb (Fig. 3E). Reduced weight bearing suggests that it may have been painful for the rat to place weight on the affected limb. Thus, the time courses of mechanical hypersensitivity differ for the weight-bearing vs. von Frey tests, indicating that DRC results in a prolonged mechanical allodynia that requires adjustment in hind limb weight bearing only during the initial weeks following injury. The relative time courses of the various behavioral measures are depicted schematically in Fig. 2. In sum, our results and those of Kawakami et al. (1994a,b) indicate that DRC using silk ligatures produces a pronounced and prolonged bilateral mechanical allodynia, accompanied initially by altered gait and reluctance to use the affected limb that resolved more quickly.

Fig. I. Behavioral effects of DRC. (A) Graph plots mean von Frey mechanical withdrawal thresholds on the operated side vs. time followng surgery for each group. Rats stood on a wire mesh floor and graded von Frey filaments were pressed against the ventral paw. The threshold bending force eliciting paw withdrawal was determined. Mean thresholds at 100%: 87.5 & 2 I g ( I -ligation), 85.1 f 19.3 g (2-ligation) and 81.6 & 34.1 g (sham). Mean thresholds were significantly reduced relative to pre-surgery baseline (p < 0.05. paired f-test) at 3 days to 22 wk following surgery in I- and 2-ligation groups but not shams. Error bars: SEM; ‘: significantly different from pre-surgery baseline (p < 0.05, paired t-test). (B) Graph as in (A) showing delayed reduction in von Frey withdrawal threshold on non-operated side. Mean thresholds at 100%: 99. I + 24.9 g (l-ligation). 89.6 i 22.3 (2.ligation) and 86.3 i 29.4 g (sham). Maximum reduction in threshold was to 26%. (C) Graph plots mean thermal paw withdrawal latencies (Hargreaves et al., 1988) in each group vs. time. Rats stood on a glass plate warmed to 30°C ZL 1°C. An infrared beam was focused on the ventral paw from below, and the latency until the paw reflexively moved was measured. At 3 days and I wk following surgery there was a significant (/I < 0.05, paired r-test) increase in latency (hypoalgesia) on the operated side in I- and 2.ligation groups but not shams. followed by recovery to pre-surgical baseline. (D) Graph as in (A) showing lack of significant change in latencies on the non-operated side. Error bars for sham group are omitted for clarity. (E) Graph plots mean weight bearing by each hind limb vs. time following surgery for ligated rats. Rats stood in an enclosure with each hind paw centered on a separate force plate to measure the weight borne by that limb. Data for l-ligation (N = 3) and 2-ligation groups (N = 6) were similar and have been pooled. Weight bearing (expressed as % of total body weight) on the operated side was significantly reduced relative to pre-surgery baseline (p < 0.05, paired r-test) for 5 wk following surgery. At 3 days following surgery there was a significant (p < 0.05, paired r-test) compensatory increase in weight bearing on the opposite side. (F) Graph as in (A) showing no change in weight bearing in shams. From Tabo et al. (1999) with permission of Elsevier Publishing Co.

Operated

Non-Operated

Side

Side

-oSingle Ligation -oDouble Ligation ..r.-.Sham

60-

01

1

Pre

’

3D

1

’

1W 2W

1

Time

c

1

3W

5W

’

7W

’

’

1

1

9W 12W 16W22W

0

’

I Pre

3D

1W

2W

Post-Ligation

3W

Time

5W

7W

9W i2W

16W22W

Post-Ligation

_ ** Operated -*--.A..-

Side Double Sham

Non-Operated -o-*--.A.-.

Ligation

1

1

Pre

3D

1W

2W

3W

Time

5W

*

4

9W 12W 16W 22W

I

I

I

I

I

I

I

Pre

3D

IW

2W

3W

5W

7W

Time

Ligation

(N=9)

Single Ligation Double Ligation Sham

T

Post-Ligation

Single+Double -+ -o-

7W

T

Side

F

Sham

50

Operated Side Non-Operated Side

I

I

I

I,

9W 12W 16W22W

Post-Ligation

(N=iO)

--A-A-

Operated Side Non-Operated Side

40

01

Pre

3D

IW

2W

3W

5W

7W

9W 12W 16W 22W

Time Post-Surgery

Pre

3D

IW

2W

3W

5W

7W

9W 12W 16W22W

Time Post-Surgery

402

Sensitive

+ _ Weight-bearing

bv affected

hindlimb

More Sensitive Less Sensitive

Sensitive

‘\ \’ :\~J -Thermal

Nociceotion:

+

DRC

5 Wk Post

22 Wk Post

TIME(Weeks) Fig. 2. Schematic

depicting

time course

of behavioral

and electrophysiological

Compression and/or inflammation of DRGs appears to produce more pronounced behavioral effects compared to DRC. Loose ligation of the L5-L6 DRGs with chromic gut sutures on one side in rats resulted in a marked thermal hyperalgesia and mechanical allodynia (Chatani et al., 1995). Compression of DRG and dorsal roots with an L-shaped steel rod implanted in the intervertebral foramen of L4L5 ganglia on one side in rats resulted in mechanical allodynia from post-surgical days l-28, as well as a mild thermal hyperalgesia (Song et al., 1997; Hu and Xing, 1998). Injection of carrageenan into the intervertebral foramen also resulted in thermal hyperalgesia (Hu and Xing, 1998). Functional changes in primary afferents and dorsal horn neurons after dorsal root/DRG injury There are relatively few electrophysiological studies of the effects of acute compression of dorsal root or DRG, and until our recent report (Tab0 et al., 1999) there were no prior studies of altered dorsal horn processing following chronic DRC. Acute high-threshold mechanical stimulation of the dor-

changes

following

DRC.

sal root only transiently excites dorsal root fibers, whereas acute compression of DRG activates dorsal root fibers at a lower mechanical threshold throughout the duration of compression (Howe et al., 1977; Sugawara et al., 1996). Acute mechanical compression (40 g for 3 min) of the DRG excited widedynamic-range (WDR)-type spinal cord dorsal horn neurons throughout the duration of the compression stimulus, while similar compression of the dorsal roots only transiently excited the same dorsal horn neurons (Hanai et al., 1996). In a recent study, recordings were made from dorsal root fibers in animals exposed to chronic DRG compression for 5-20 days (Hu and Xing, 1998). A significantly higher fraction of dorsal root fibers exhibited abnormal spontaneous firing and enhanced sensitivity to mechanical or chemical stimulation of the injured DRG in these animals compared to controls. These data collectively indicate that the DRG is more sensitive than dorsal roots to a mechanical compression stimulus. Activation of afferent fibers by mechanical stimulation of the DRG might conceivably contribute to pain under conditions such as disc hemiation or spinal stenosis (Garfin et al., 1995). There is even less information regarding longer-

403

A 40 “C

150.

8 2? 8 100. 52

j

$50. -

j

2 Ligations Operated Side

n I

46 “C

52 “C

i

OlLl 0

15

20

25

30

Time (set) Non-Operated

150.

-T-

Side

E zi

loo-

s $ -

50.

0-l0

: : 9. :

:

.

15

20

a-1 30

0

25

Time(sec)

:f

;

:

15

20

-A-

Q c z 22 2 .E

30

L_

0

:

j

15

20

25

30

:I;

Heat 52 “C 5 set

C

Heat Response, Operated Side Heat Response, Non-Operated Side Spont. Act., Operated Side Spont. Act., Non-Operated Side

fl 49

/

800 -

0

25

Heat 46 “C 5 set

-o-o-A-

1000 -

j

:

Heat 40 “C 5 set

B

:

800 -

400 -

38

40

42

44 Temperature

46

48

50

52

Operated Side

Non-Operated Side

(‘Cc)

Fig. 3. Example of WDR neurons from rat which had received DRC (2-ligations) unilaterally 22 wk earlier. (A) Peristimulus-time histograms (PSTHs; bin width 1 s) of responses to graded noxious heat stimuli for a neuron ipsilateral to (upper row of PSTHs) and contralateral to (lower row of PSTHs) DRC. Bars with dashed vertical lines indicate duration of heat stimuli. Examples of each neuron’s action potential waveform are shown in boxes. (B) Graph plotting responses of neurons shown in (A) vs. temperature. The neuron ipsilateral to ligation had a higher spontaneous firing rate and less steep stimulus-response function. (C) Receptive fields of units shown in (A). Shading depicts areas within which application of von Frey filament at the indicated bending force elicited a response. From Tabo et al. (1999) with permission of Elsevier Publishing Co.

404

term changes following chronic dorsal root or DRG injury. Two wk following DRC, there was a 4-fold increase in c-fos immunoreactive neurons in lumbar spinal cord ipsilateral, and smaller (48%) increase contralateral, to the DRC (Kawakami et al., 1994a), possibly indicative of a sensitization of dorsal horn neurons. The number of c-fos-positive neurons had returned to control levels by 12 wk following DRC, but there were slight reductions in numbers of neurons staining for substance P and CGRP ipsilaterally 8 and 12 wk after DRC (Kawakami et al., 1994a). This latter observation might be explained by a reduction in the number of fibers in the injured spinal nerve and dorsal root (Kawakami et al., 1994b; see below). Since no other studies have examined possible changes in dorsal horn function in animals that had been subjected to a chronic dorsal root injury, we pursued this question by recording from WDR-type dorsal horn neurons in rats that had received DRC either 5 or 22 wk earlier (Tab0 et al., 1999). Recordings were made from WDR-type dorsal horn neurons ipsilateral and contralateral to the DRC, as well as from sham-operated rats (5 or 22 wk following surgery). Measured parameters included spontaneous activity, size of the mechanosensitive receptive field, and responses to graded noxious thermal stimuli (3852°C). In the 22-wk post-DRC rats, neurons ipsilateral to the DRC showed significantly higher spontaneous activity compared to those contralateral to the DRC or those recorded in sham animals. Most notably, receptive field areas of neurons ipsilateral to the DRC were significantly larger compared to those of neurons contralateral to DRC or from shams. An example is shown in Fig. 3. Fig. 3A shows peristimulus-time histograms (PSTHs) of neurons recorded ipsi- (upper row) and contralateral to the DRC (lower row), respectively. Note the higher ongoing activity prior to heat in the upper left PSTH. The graph of Fig. 3B shows that the neuron ipsilateral to DRC showed higher spontaneous activity, although both it and the neuron contralateral to DRC encoded increasing stimulus temperatures. Overall, there were no significant differences in stimulusresponse functions for graded noxious heat in neurons ipsi- vs. contralateral to DRC. Importantly, the receptive field area, shown in Fig. 3C, was much larger for the neuron ipsilateral to DRC compared

to the neuron on the contralateral side. By contrast, in the 5-wk post-DRC rats the ipsilateral receptive field areas were not different from those contralateral to the DRC or in shams (Fig. 4) even though the DRC-injured rats exhibited maximum mechanical hypersensitivity at this time (Fig. 3A). Therefore, the time course of changes in mechanical receptive field areas of dorsal horn neurons is too slow to account for the rapidly occurring changes in mechanical sensitivity following DRC (Fig. 2), nor are there any neuronal changes on the opposite side that could account for the ‘mirror image’ increase in mechanical sensitivity seen contralateral to the DRC (Fig. 4). Models of neuropathic pain after peripheral nerve injury and functional changes in dorsal horn neurons The most commonly used animal models of neuropathic pain involve rats that received either loose chronic constriction of the whole sciatic nerve, typically with four chromic gut ligatures (‘Bennett’ model; Bennett and Xie, 1988), tight ligation of onethird to one-half of the whole sciatic nerve (‘Seltzer’ model; Seltzer et al., 1990), or tight ligation of L4 and L5 spinal nerves peripheral to the DRG (‘Chung’ model; Kim and Chung, 1992). The three models share a common symptomatology of spontaneous pain, hyperalgesia and mechanical allodynia, although there are differences among the models in the degree of expression of various symptoms (e.g., Kim et al., 1997b). There are now several published studies of spinal dorsal horn neurons from rats that had previously received one of these peripheral nerve injuries: Bennett model (Palecek et al., 1992a,b; Sotgiu et al., 1992, 1994, 1995; Laird and Bennett, 1993; Sotgiu, 1993; Sotgiu and Biella, 1995, 1997; Biella et al., 1997), Seltzer model (Behbehani and DollbergStolik, 1994; Takaishi et al., 1996), Chung model (Leem et al., 1995; Pertovaara et al., 1997; Chapman et al., 1998; Suzuki et al., 1999), as well as one study of effects of tight ligation of the L7 spinal nerve in monkeys (Palecek et al., 1992a). Spinal WDR and nociceptive-specific-type dorsal horn neurons exhibited increased spontaneous firing in the Bennett model (Palecek et al., 1992b; Sotgiu et al., 1992, 1994; Laird and Bennett, 1993; Sotgiu, 1993;

405

B

1 Ligation 22 W,Operated Side (N=14) 1 Ligation 22 W,Non-Operated Side (N=14) Sham 22 W, Operated Side (N=32)

m 0

500-

2 Ligations 22 W, Operated Side (N=28) 2 Ligations 22 W, Non-Operated Side (N=29) *

400-

0.7

4

12

76

Force (g) m 0

-500

m

ET

0-

I

0.7

4

12

76

Force (g)

1 Ligation 5 W, Operated Side (N=16) 1 Ligation 5 W, Non-Operated Side (N=18) Sham 5 W, Operated Side (N=15) Sham 5 W, Non-Operated Side (N=l6)

E 400 .E.

g

200

ii

100

2 0

0.7

4

12

76

Force (g) Fig. 4. Enlarged receptive fields of dorsal horn neurons 22 wk after DRC. (A) Bar graph plots mean receptive field areas of dorsal horn neurons in the indicated groups (22 wk) as a function of the bending force of each von Frey filament used (indicated by shading). At each bending force, mean receptive field areas were significantly larger for units ipsilateral to DRC (l-ligation) compared with units contralateral to the DRC or with units from sham-operated animals (p < 0.05, t-test). (B) Graph as in (A) for units from 2-ligation group (22 wk). At each bending force, areas on the operated side were significantly larger (1-7 < 0.05, r-test) compared to those on non-operated side. (C) Graph as in (A) for 5-wk DRC (l-ligation) and sham-operated groups; for each bending force there were no significant differences in mean areas among groups. From Tabo et al. (1999) with permission of Elsevier Publishing Co.

Sotgiu and Biella, 1997), Seltzer model (Behbehani and Dollberg-Stolik, 1994) and Chung model (Leem et al., 1995; Pertovaara et al., 1997; Chapman et al., 1998) as well as after tight ligation of the L7 spinal nerve in monkeys (Palecek et al., 1992a). They also showed expanded mechanical receptive field area (Seltzer model: Behbehani and Dollberg-Stolik, 1994; Takaishi et al., 1996), increased percentage without a mechanical receptive field (Bennett model: Laird and Bennett, 1993; Seltzer model: Takaishi et al., 1996), increased response to mechanical stimuli (Bennett model: Palecek et al., 1992b; Laird and Bennett, 1993; Chung model: Leem et al., 1995; Per-

tovaara et al., 1997; L7 ligation in monkey: Palecek et al., 1992a), or decreased response to mechanical stimuli (Chung model: Chapman et al., 1998), increased incidence of responsiveness to extraterritorial nerve stimulation (Bennett model: Sotgiu and Biella, 1995), increased afterdischarge (Bennett model: Palecek et al., 1992b; Laird and Bennett, 1993; Sotgiu and Biella, 1997), abnormal response to cooling (Chung model: Chapman et al., 1998; L7 ligation: Palecek et al., 1992a) and reduced inhibition by electrical stimulation of the midbrain periaqueductal gray (Chung model: Pertovaara et al., 1997). WDR neurons in nerve-injured animals did

406

not show increased responsiveness to noxious thermal stimuli in any of the three models (Palecek et al., 1992b; Laird and Bennett, 1993; Takaishi et al., 1996; Pertovaara et al., 1997; Chapman et al., 1998), with the exception of one study reporting enhanced responses of spinothalamic neurons to noxious heat in primates with tight ligation of the L7 spinal nerve (Palecek et al., 1992a). It is of interest to compare the changes in behavior with changes in properties of spinal dorsal horn neurons at different times following peripheral nerve injury. In an earlier study from our laboratory using the Seltzer model (Takaishi et al., 1996), we demonstrated a significant bilateral thermal hyperalgesia which increased in severity from the 1st to 16th wk after injury. This was accompanied by a bilateral mechanical allodynia that was more profound ipsilaterally after 1 and 5 wk, but which increased in severity and was nearly equal on both sides after 16 wk. Electrophysiological experiments conducted on the behaviorally tested rats revealed no changes in properties of WDR-type dorsal horn neurons ipsior contralateral to the injury after 5 wk. There was a significant increase in the mean area of neuronal mechanical receptive fields in rats tested 16 wk after nerve injury, compared to neurons recorded in shamoperated rats. Interestingly, the increase in receptive field area was marked for neurons ipsilateral to the nerve injury, but there was also a lesser but still significant expansion of receptive fields contralateral to the injury. This was consistent with the bilateral mechanical allodynia observed at 16 wk. There were no significant changes in the responses of the WDR neurons to graded noxious thermal stimuli at 5 or 16 wk following injury, although the trend was toward lesser responsiveness. We did not find evidence for increased spontaneous firing of neurons ipsilateral to the nerve injury, but did obtain evidence for unresponsive neurons. It is therefore notable that the marked behavioral allodynia and thermal hyperalgesia observed after 1 and 5 wk were not reflected by any corresponding change in spinal neurons. At the later time point, there was a more reasonable correspondence between the bilateral receptive field expansion of dorsal horn neurons and the bilateral mechanical allodynia. However, there was still a mismatch between the thermal hyperalgesia still present and the marked absence of any change in thermal

responsiveness of dorsal horn neurons. The results of this study are similar those cited earlier with DRC (Tab0 et al., 1999). A recent study by Chapman et al. (1998) examined changes in WDR-type neurons in rats at two time points (7-10 and 14-17 days) following tight ligation of L5-L6 spinal nerves (Chung model). The rats exhibited ipsilateral mechanical and cold allodynia that peaked at days 9-12 following injury. Neurons ipsilateral to the nerve injury exhibited a significantly higher incidence of spontaneous activity compared to sham-operated controls, and the level of spontaneous firing increased with time following injury. There was also an increase in the incidence of neurons responding to mechanical and cold stimuli after lo-14 days. Curiously, however, the incidence and magnitude of mechanically and cold-evoked responses was significantly lower compared to shams at the later time point. Again, there is a notable mismatch between the behavioral allodynia and the reduced mechanosensitivity of dorsal horn neurons in this nerve injury model. A third study employing the Bennett model examined properties of rat spinothalamic neurons at 7, 14 and 28 days following nerve injury (Palecek et al., 1992b). The rats exhibited thermal hyperalgesia and mechanical allodynia behaviorally at each of these time points. In electrophysiological recordings from identified spinothalamic tract neurons, spontaneous firing was significantly elevated in the nerve-injured rats compared to controls, although the magnitude of spontaneous firing progressively decreased with time following injury. There was a tendency for neurons from nerve-injured rats to be more responsive to low-threshold mechanical stimuli, and to exhibit abnormally long afterdischarges. However, neuronal responses to graded noxious thermal stimuli were not significantly different from controls. Thus, while enhanced mechanosensitivity of spinothalamic neurons might account for mechanical allodynia, there was no change in thermal sensitivity of these neurons that could readily account for the thermal hyperalgesia. Dorsal horn mechanisms underlying and neuropathic pain

radicular

How can we reconcile the different time courses for the development of manifestations of pain in

407

the behavioral models, and for the changes in dorsal horn function, that occur following peripheral nerve or dorsal root/DRG injury? It seems reasonable that a variety of factors, each having a different time course, may contribute to dynamic neuroplastic changes that underlie the altered behavior. In the following, we assume that similar mechanisms apply to both peripheral nerve and dorsal root/DRG injury, although this assumption may be questionable. The afferent barrage provided by nociceptive primary afferent fibers at the onset and following injury of dorsal roots (Howe et al., 1977; Hanai et al., 1996; Sugawara et al., 1996; Hu and Xing, 1998) or peripheral nerve (e.g., Kajander and Bennett, 1992; Kajander et al., 1992) might conceivably trigger central sensitization via NMDA receptors residing on postsynaptic dorsal horn neurons, leading to enhanced cellular excitability (Dickenson et al., 1997; Gerber et al., 2000, this volume; Moore et al., 2000, this volume; Sandktihler et al., 2000, this volume). NMDA antagonists can reduce signs of neuropathic pain (e.g., Chaplan et al., 1997; Kim et al., 1997a; Burton et al., 1999). Central sensitization can occur quickly (Gerber et al., 2000, this volume; Moore et al., 2000, this volume; Sandktihler et al., 2000, this volume), within minutes to hours, consistent with the rapid development of mechanical allodynia. Signs of central sensitization, such as enhanced mechanical responsiveness, were observed 7-10 days following sciatic or L5-L7 spinal nerve injury (Palecek et al., 1992b; Chapman et al., 1998). Such changes might subside, or be superseded by a progressive loss of afferent input, based on the progressive reduction in dorsal horn neuronal mechanosensitivity 14-17 days after nerve injury (Chapman et al., 1998), and our observation that dorsal horn neurons exhibited no major functional changes 5 wk following partial sciatic nerve ligation (Takaishi et al., 1996) or DRC (Tabo et al., 1999; see Fig. 2). Excessive activation of dorsal horn neurons might also lead to cell damage or death via excitotoxicity, as possibly reflected by the appearance of ‘dark’ neurons in superficial dorsal horn following nerve injury (Sugimoto et al., 1990). If such degenerating neurons normally function as inhibitory intemeurons, their loss might eventually lead to hyperexcitability of dorsal horn neurons via disinhibition. Following the establishment of central sensitiza-

tion, or disinhibition, dorsal horn neurons would give abnormally large responses to non-noxious stimuli (resulting in allodynia) as well as to noxious stimuli (resulting in hyperalgesia). In addition, previously ineffective (‘subliminal’) synaptic inputs onto the hyperexcitable dorsal horn neuron might now be suprathreshold, contributing to expansion of receptive fields (Woolf and King, 1989). A role for the neuropeptide, substance P, in the mediation of neuropathic pain receives support from a recent study (Nichols et al., 1999). Signs of neuropathic pain in the Chung model were significantly reduced in rats in which spinal neurons expressing the substance P (NK-1; neurokinin-1) receptor were selectively ablated after having internalized substance P conjugated to the neurotoxin, saporin. In the Bennett model, spinal cord substance-P levels progressively decrease between 3 and 14 days after nerve injury, followed by a progressive, incomplete recovery at 28 and 70 days (Cameron et al., 1997). Spinal substance-P levels also decreased slightly following DRC with chromic gut (Kawakami et al., 1994a). Decreased substance P might be associated with degeneration of primary afferents (see below). On the other hand, spinal NK-1 receptor levels increased following constriction of the sciatic nerve, peaking at day 4 and progressively declining toward baseline after 28 days (Goff et al., 1998). In the face of reduced substance-P levels, an upregulation of NK-1 receptors would appear necessary to allow dorsal horn neurons to give enhanced responses to peripheral nociceptive stimuli in mediating hyperalgesia. What factors might contribute to pain that persists for weeks, months, or even indefinitely, following nerve injury? Such injuries would be expected to produce some degeneration of afferent fibers, resulting in a partial denervation. Indeed, DRC with either silk or chromic gut both reduced the number of myelinated fibers in spinal nerve and dorsal root (Kawakami et al., 1994b). Degeneration of peripheral fibers also results following loose sciatic nerve ligation (Munger and Bennett, 1990; Basbaum et al., 1991; Carlton et al., 1991). Partial deafferentation might lead to a variety of central alterations such as reduced synaptic input to dorsal horn neurons as observed in some studies or, alternatively, denervation supersensitivity (e.g., upregulation of NK-1 recep-

408

tors). Partial denervation might also trigger sprouting of large myelinated fibers into lamina II (Woolf et al., 1992), as observed following tight ligation of L5-L6 spinal nerves (Lekan et al., 1996). If the sprouts form functional synapses with nociceptive dorsal horn neurons, then activation of mechanoreceptors giving rise to the sprouting myelinated fibers might excite lamina-II neurons, thereby giving rise to an aberrant signal that is interpreted as pain. Such synaptic remodeling might be expected to occur over a longer time course compared to more rapidly occurring processes such as sensitization. Such longterm changes are consistent with the late (>5 wk) appearance of expanded dorsal horn neuronal receptive fields that we reported following partial sciatic nerve ligation (Takaishi et al., 1996) or DRC (Tab0 et al., 1999). Conclusions The observed changes in spinal nociceptive transmission do not correlate well with behavioral symptoms of pain that develop following injury to dorsal roots, DRGs or peripheral nerves. Part of the problem may be due to qualitative behavioral differences among the models, compounded by inconsistencies regarding the parameters of dorsal horn neuronal function that have been investigated. Nonetheless, the emerging pattern is that thermal hyperalgesia, common to most chronic pain models, is not associated with increased thermal sensitivity of dorsal horn neurons, at least in rats. This is not likely to be due to a sampling bias, since considerable evidence indicates that the well-studied WDR and spinothalamic tract neurons reasonably reflect pain behavior (Willis, 1985). The association between enhanced responses of dorsal horn neurons to mechanical stimuli, and behavioral assessment of mechanical allodynia, is somewhat better although major inconsistencies exist in some studies, including our own (Fig. 2). A variety of mechanisms might contribute to altered spinal nociceptive processing following nerve injury. For example, central sensitization or upregulation of NK-1 receptors might be involved in fairly early changes underlying the rapid onset of hyperalgesia and allodynia. Other mechanisms such as transsynaptic degeneration, or synaptic remodeling within the dorsal horn, may underlie long-lasting,

pathophysiological changes associated with persistent pain. A more complete understanding of the various neuroplastic changes underlying radicular and neuropathic pain will aid in further developing rational strategies to prevent or reverse these processes, thereby providing more choices in dealing with the difficult problem of chronic pain. Acknowledgements Supported by a grant from the California TobaccoRelated Disease Research Program # 6RT-023 1. References Basbaum, A.I., Gautron, M., Jazat, F., Mayes, M. and Guilbaud, G. (1991) The spectrum of fiber loss in a model of neuropathic pain in the rat: an electron microscopic study. Pain, 47: 3.59367. Behbehani, M.M. and Dollberg-Stolik, 0. (1994) Partial sciatic nerve ligation results in an enlargement of the receptive field and enhancement of the response of dorsal horn neurons to noxious stimulation by an adenosine agonist. Pain, 58: 421428 Bennett, G.J. and Xie, Y.K. (1988) A peripheral mononeuropathy in rat that produces disorders of pain sensation like those seen in man. Pain, 33: 87-107. Biella, G., Riva, L. and Sotgiu, M.L. (1997) Interaction between neurons in different laminae of the dorsal horn of the spinal cord. A correlation study in normal and neuropathic rats. Eus J. Neurosci., 9: 1017-1025. Burton, A.W., Lee, D.H., Saab, C. and Chung, J.M. (1999) Preemptive intrathecal ketamine injection produces a long-lasting decrease in neuropathic pain behaviors in a rat model. Reg. Anesth. Pain Med., 24: 208-213. Cameron, A.A., Cliffer, K.D., Dougherty, P.M., Garrison, C.J., Willis, W.D. and Carlton, S.M. (1997) Time course of degenerative and regenerative changes in the dorsal horn in a rat model of peripheral neuropathy. .I. Camp. Neural., 379: 428-442. Carlton, S.M., Dougherty, P.M., Pover, C.M. and Coggeshall, R.E. (1991) Neuroma formation and numbers of axons in a rat model of experimental peripheral neuropathy. Neurosci. Lett., 131: 88-92. Chaplan, S.R., Malmberg, A.B. and Yaksh, T.L. (1997) Efficacy of spinal NMDA receptor antagonism in formalin hyperalgesia and nerve injury evoked allodynia in the rat. J. Phnrmucol. Exp. The%, 280: 829-838. Chapman, V., Suzuki, R. and Dickenson, A.H. (1998) Electrophysiological characterization of spinal neuronal response properties in anaesthetized rats after ligation of spinal nerves L5-L6. J. Physiol., 507: 881-894. Chatani, K., Kawkami, M., Weinstein, J.N., Meller, S.T. and Gebhart, G.F. (1995) Characterization of thermal hyperalge-

sia, c-fos expression, and alterations in neuropeptides after mechanical irritation of the dorsal root ganglion. Spine, 20: 277-290. Dickenson, A.H., Chapman, V. and Green, G.M. (1997) The pharmacology of excitatory and inhibitory amino acid-mediated events in the transmission and modulation of pain in the spinal cord. Gen. Phamcol., 28: 633-638. Franson, R.C., Saal, J.S. and Saal, J.A. (1992) Human disc phospholipase A2 is inflammatory. Spine, 175: 129-132. Garfin, S.R., Rydevik, B., Lind, B. and Massie, J. (1995) Spinal nerve root compression. Spine, 20: 1810-1820. Gerber, G., Youn, D.-H., Hsu, C.H., Isaev, D. and Randic, M. (2000) Spinal dorsal horn synaptic plasticity: Involvement of group I metabotropic glutamate receptors. In: J. Sandktihler, B. Bromm and G.F. Gebhart (Eds.), Nervous System Plasticity and Chronic Pain. Progress in Brain Research, Vol. 129. Elsevier, Amsterdam, pp. 115-134. Goff, J.R., Burkey. A.R., Goff, D.J. and Jasmin, L. (1998) Reorganization of the spinal dorsal horn in models of chronic pain: correlation with behaviour. Neuroscience, 82: 559-574. Hanai, F., Matsui, N. and Hongo, N. (1996) Changes in responses of wide dynamic range neurons in the spinal dorsal horn after dorsal root or dorsal root ganglion compression. Spine, 21: 1408-1415. Hargreaves, K., Dubner, R., Brown, F., Flores, C. and Joris, J. (1988) A new and sensitive method for measuring thermal nociception in cutaneous hyperalgesia. Pain, 32: 77-88. Hayashi, N., Weinstein, J.N., Meller, ST., Lee, H.-M., Pratt, K.F. and Gebhart, G.F. (1998) The role of steroids and their effects on phospholipase A2. Spine, 23: 119 1- 1196. Howe, J.E., Loeser, J.D. and Calvin, J.H. (1977) Mechanosensitivity of dorsal root ganglia and chronically injured axons: a physiological basis for the radically pain of nerve root compression. Pain, 3: 25-41. Hu, S.-J. and Xing, J.-L. (1998) An experimental model for chronic compression of dorsal root ganglion produced by intervertebral foramen stenosis in the rat. Pain, 77: 15-23. Kajander, KC. and Bennett, G.J. (1992) Onset of a painful peripheral neuropathy in rat, a partial and differential deaf ferentation and spontaneous discharge in A8 and A6 primary afferent neurons. J. Neurophysiol., 68: 734-744. Kajander, K.C., Wakisaka, S. and Bennett, G.J. (1992) Spontaneous discharge originates in the dorsal root ganglion at the onset of a painful peripheral neuropathy in the rat. Neurosci. Left., 138~ 225-228. Kawakami, ‘M., Weinstein, J.N., Spratt, K.F., Chatani, K.-I., Traub, R.J., Meller, S.T. and Gebhart, G.F. (1994a) Experimental lumbar radiculopathy. Immunohistochemical and quantitative demonstrations of pain induced by lumbar nerve root irritation of the rat. Spine, 19: 1780-1794. Kawakami, M., Weinstein, J.N., Chatani, K.-I., Spratt, K.F., Meller, S.T. and Gebhart, G.F. (1994b) Experimental lumbar radiculopathy. Behavioral and histologic changes in a model of radicular pain after spinal nerve root irritation with chromic gut ligatures in the rat. Spine, 19: 1795-l 802. Kelsey, J.A. and White, A.A. (1980) Epidemiology and impact of low-back pain. Spine, 5: 133-142.

Kim, K.J. and Chung, J.-M. (1992) An experimental model for peripheral neuropathy produced by segmental spinal nerve ligation in the rat. Pain, 50: 3.55-363. Kim, K.J., Na, H.S., Yoon, J.W., Han, H.C., Ko, K.H. and Hong, S.K. (1997a) NMDA receptors are important for both mechanical and thermal allodynia from peripheral nerve injury in rats. Neuroreport, 8: 2149-2153. Kim, K.J., Yoon, Y.W. and Chung, J.M. (1997b) Comparison of three rodent neuropathic pain models. Exp. Bruin Rex, 113: 200-206. Laud, J.M.A. and Bennett, G.J. (1993) An electrophysiological study of dorsal horn neurons in the spinal cord of rats with an experimental peripheral neuropathy. J. Neurophysiol., 69: 2072-2085. Lee, H.-M., Weinstein, J.N., Meller, S.T., Hayashi, N., Pratt, K.F. and Gebhart, G.F. (1998) The effect of epidural injection of betamethasone or bupivacaine in a rat model of lumbar radiculopathy. Spine, 23: 877-885. Leem, J.W., Park, E.S. and Paik, KS. (1995) Electrophysiological evidence for the antinociceptive effect of transcutaneous electrical stimulation on mechanically evoked responsiveness of dorsal horn neurons in neuropathic rats. Neurosci. L&t., 192: 197-200. Lekan, H.A., Carlton, S.M. and Coggeshall, R.E. (1996) Sprouting of A-beta fibers into lamina II of the rat dorsal horn in peripheral neuropathy. Neurosci. Mt., 208: 147-150. Maves, T.J., Gebhart, G.F. and Meller, S.T. (1995) Continuous infusion of acidified saline around the rat sciatic nerve produces thermal hyperalgesia. Neurosci. Lett., 194: 45-48. Moore, K.A., Baba, H. and Woolf, C.J. (2000) Synaptic transmission and plasticity in the superficial dorsal horn. In: J. Sandktihler, B. Bromm and G.F. Gebhart (Eds.), Nervous Systern Plasticity and Chronic Pain. Progress in Brain Research, Vol. 129. Elsevier, Amsterdam, pp. 63-80. Munger, B.L. and Bennett, G.J. (1990) The peripheral axonal pathology in the constrictive model of peripheral neuropathy. Anat. Rec., 226: 70. Nichols, M.L., Allen, B.J., Rogers, S.D., Ghilardi, J.R., Honore, P., Luger, N.M., Finke, M.P., Li, J., Lappi, D.A., Simone, D.A. and Mantyh, P.W. (1999) Transmission of chronic nociception by spinal neurons expressing the substance P receptor. Science, 286: 1558-1561. Palecek, J.V., Dougherty, P.M., Kim, K.Y., Paleckova, V., Lekan, H., Chung, J.M., Carlton, S.M. and Willis, W.D. (1992a) Responses of spinothalamic tract neurons to mechanical and thermal stimuli in an experimental model of peripheral neuropathy in primates. L Neurophysiol., 68: 1951-1966. Palecek, J.V., Paleckova, V., Dougherty, PM., Carlton, SM. and Willis, W.D. (1992b) Responses of spinothalamic tract cells to mechanical and thermal stimulation of skin in rats with experimental peripheral neuropathy. J. Neurophysiol., 67: 1562-1573. Pertovaara, A., Kontinen, V.K. and Kalso, E.A. (1997) Chronic spinal nerve ligation induces changes in response characteristics of nociceptive spinal dorsal horn neurons and their descending regulation originating in the periaqueductal gray in the rat. Exp. Neural., 147: 428-436.

410

Sandkiihler, J., Benrath, J., Brechtel, C., Ruscheweyh, R. and Heinke, B. (2000) Synaptic mechanisms of hyperalgesia. In: J. Sandktihler, B. Bromm and G.F. Gebhart (Eds.), Nervous System Plasticity and Chronic Pain. Progress in Brain Research, Vol. 129. Elsevier, Amsterdam, pp. 81-100. Seltzer, Z., Dubner, R. and Shir, Y. (1990) A novel behavioral model of neuropathic pain disorders produced in rats by partial sciatic nerve injury. Pain, 43: 205-218. Song, X.-J., Hu, S.-J., Zhang, J.-M., Greenquist, K.W. and LaMotte, R.H. (1997) Cutaneous hyperalgesia during chronic compression of dorsal root ganglion. Sot. Neurosci. Abstr:, 23: 1259. Sotgiu, M.L. (1993) Descending influence on dorsal horn neuronal hyperactivity in a rat model of neuropathic pain. Neuroreport, 4: 2 1-24. Sotgiu, M.L. and Biella, G. (1995) Spinal expansion of saphenous afferents after sciatic nerve constriction in rats. Neuroreport, 6: 2305-2308. Sotgiu, M.L. and Biella, G. (1997) Role of input from saphenous afferents in altered spinal processing of noxious signal that follows sciatic nerve constriction in rats. Neurosci. Mt., 223: 101-104. Sotgiu, M.L., Lacerenza, M. and Marchettini, P. (1992) Effect of systemic lidocaine on dorsal horn neuron hyperactivity following chronic peripheral nerve injury in rats. Somatosens. Mot. Res., 9: 227-233. Sotgiu, M.L., Biella, G. and Riva, L. (1994) A study of ongoing activity in dorsal horn units following sciatic nerve constriction Newweport, 5: 2609-2612. Sotgiu, M.L., Biella, G. and Riva, L. (1995) Poststimulus after-

discharges of spinal WDR and NS units in rats with chronic nerve constriction. Neuroreport, 6: 1021-1024. Sugawara, O., Atsuta, Y., Iwahara, T., Muramoto, T., Watakabe, M. and Takemitsu, Y. (1996) The effects of mechanical compression and hypoxia on nerve root and dorsal root ganglia. An analysis of ectopic firing using an in vitro mode. Spine, 21: 2089-2094. Sugimoto, T., Bennett, G.J. and Kajander, K.C. (1990) Transsynaptic degeneration in the superficial dorsal horn after sciatic nerve injury effects of a chronic constriction injury, transection, and strychnine. Pain, 42: 205-213. Suzuki, R., Chapman, V. and Dickenson, A.H. (1999) The effectiveness of spinal and systemic morphine on rat dorsal horn neuronal responses in the spinal nerve ligation model of neuropathic pain. Pain, 80: 215-228. Tabo, E., Jinks, S.L., Eisele Jr., J.H. and Carstens, E. (1999) Behavioral manifestations of neuropathic pain and mechanical allodynia, and changes in spinal dorsal horn neurons, following L4-L6 dorsal root constriction. Pain, 80: 503-520. Takaishi, K., Eisele Jr., J.H. and Carstens, E. (1996) Behavioral and electrophysiological assessment of hyperalgesia and changes in dorsal horn responses following partial sciatic nerve ligation in rats. Pain, 66: 297-306. Willis, W.D. (1985) The Pain System. Karger, Basel. Woolf, C.J. and King, A.E. (1989) Subthreshold components of the cutaneous mechanoreceptive fields of dorsal horn neurons in the rat lumbar spinal cord. J. Neurophysiol., 62: 907-9 16. Woolf, C.J., Shortland, P. and Coggeshall, R.E. (1992) Peripheral nerve injury triggers central sprouting of myelinated afferents. Nature, 355: 75-77.

J. Sandkiitrler, B. Bromm and GE Gebhat (Eds.) Progress in Brain Research, Vol. 129 0 2000 Elsevier Science B.V. All rights reserved

CHAF’TER 3 1

Allodynia and hyperalgesia within dermatomes caudal to a spinal cord injury in primates and rodents Charles J. Vierck, Jr. I,* and Alan R. Light 2

’ Department

’ Department of Neuroscience, University of Florida Brain Institute, of Cellular and Molecular Physiology, University of North Carolina,

Categories of central pain following spinal cord injury Despite the common occurrence of pain of central origin that is referred to dermatomes supplied by segments below the level of a spinal cord injury (Botterell et al., 1953; Nepomuceno et al., 1979; Yezierski, 1996), this disorder has not been analyzed extensively in laboratory animals. In contrast, there are well characterized laboratory animal models that describe changes in behavior directed toward dermatomes supplied by segments adjacent to and directly affected by a spinal lesion (Hao et al., 1991; Yezierski et al., 1998). These border zone phenomena are probably related to (if not dependent upon) excitatory effects of a lesion on neurons in neighboring segments (Hao et al., 1992; Yezierski and Park, 1993). However, mechanisms underlying pain that is referred to remote caudal dermatomes are yet to be established. The latter type of sensory disorder is referred to as deafferentation zone pain, consistent with referral of abnormal sensations to portions of the body which have lost all or a portion of their afferent drive of (projections to) supraspinal somatotopic representations. This does not imply that deafferentation is a sufficient condition for development of central pain. It describes the distribution *Corresponding author: C.J. Vierck, Jr., Department of Neuroscience, University of Florida, College of Medicine, Gainesville, FL. 32610-0244, USA. Fax: +l (352) 3928513; E-mail: [email protected]

Gainesville. FL 32610-0244, School of Medicine, Chapel

USA Hill, NC 27599,

USA

of referred sensations to dermatomes with reduced input to supraspinal levels as a result of a spinal lesion. Models of deafferentation

zone pain

A difficulty for an animal model of deafferentation zone pain is that such pain in human beings is referred to body regions that have been shown to be relatively insensitive to peripheral stimulation. In particular, sensitivity to innocuous thermal and normally painful stimulation is reliably diminished within the region of pain referral for patients with

post-stroke central pain (Boivie et al., 1989; Bowsher, 1996). These and other findings suggest that interruption of the spinothalamic tract at any level of the neuraxis can produce central pain (Cassinari and Pagni, 1969). In related experiments with laboratory animals, investigators have noted the occurrence and timing of overgrooming or autotomy in a region af-

fected by transection of a peripheral nerve (Wall et al., 1979) or interruption

of the spinothalamic tract

(Levitt and Levitt, 1981). An assumption behind this approach is that an animal excessively grooms or inflicts damage on tissue in an attempt to elimi-

nate pain referred to that region. When the region is analgesic or anesthetic, autotomy would not be discouraged, because pain would not be produced by the animal’s actions.

Problems with autotomy as a measure of deafferentation zone pain are that self-destructive behavior may or may not be driven by an abnormal sen-

412 sation, and the quality, intensity and timing of such a sensation that might be experienced by the animal are unknowable (Rodin and Kruger, 1984). One possibility is that autotomy results from an absence of sensation. However, this is an unlikely explanation in all cases, because autotomy is susceptible to modulation (Coderre et al., 1986; Kauppila, 1998). Alternatively, an animal could be responding to a nonpainful dysesthetic sensation (e.g., tingling, itching). Substantial deafferentation regularly produces some form of dysesthesia, either painful or not. For example, limb amputation produces phantom sensations in virtually all cases, but referred pain of sufficient intensity to seek treatment occurs at some point in time for an estimated 10% of amputees (Jensen and Rasmussen, 1984). Similarly, spinal cord injury regularly produces phantom sensations; some form of pain occurs later for most patients, and an estimated 25% develop severe deafferentation zone pain (Botterell et al., 1953; Nepomuceno et al., 1979). Clearly it is important to distinguish between painful and non-painful sensations experienced in a region of deafferentation, but this distinction cannot be made on the basis of autotomy alone. Furthermore, autotomy is not reliably produced by anterolateral spinal lesions (Vierck and Luck, 1979; Saade et al., 1990; Vierck et al., 1990a, 1995; Vierck and Light, 1999). Autotomy is more prevalent among animals housed singly in a confined space (Rodin and Kruger, 1984) and social isolation predisposes animals to repetitive autotomous behaviors in the absence of neural injury. If we cannot rely upon observation of spontaneous behaviors to indicate the presence of deafferentation zone pain, and if nociceptive sensitivity is diminished for stimulation in the region of presumed pain, then an appropriate laboratory animal model of deafferentation zone pain is problematic. However, extensive observations of humans who have received surgical lesions of the anterolateral spinal column for relief of chronic pain in a limited distribution suggest otherwise. This procedure is seldom utilized at present, because surgical interruption of the spinothalamic tract at any level of the neuraxis often results in ‘recovery’ of contralateral pain sensitivity over time. The functional restoration usually is associated with dysesthetic pain within the region of early postoperative relief of clinical pain and reduced sensitivity to nociceptive stimulation (Horax, 1929;

Hyndman and Van Epps, 1939; Cassinari and Pagni, 1969; White and Sweet, 1969). Thus, clinical experience with surgical interruption of the spinothalamic pathway strongly reinforces findings from studies of post-stroke pain. Central pain occurs in the distribution of sensory deficits that implicate damage to the spinothalamic tract as a causal factor. After recovery from interruption of the spinothalamic tract, pain elicited within contralateral regions affected by the lesion is qualitatively distinct from similarly evoked pain in intact individuals (Wycis and Spiegel, 1962; White, 1968; Eide et al., 1996). Also, regions with input contralateral and caudal to a cordotomy can be abnormally responsive to certain forms of stimulation (e.g., repetitive stimulation; Walker, 1943; King, 1957; Nathan and Smith, 1979; Eide et al., 1996). These findings are important for modeling of deafferentation zone pain. In an extensive study of sensory capacities following anterolateral cordotomy for clinical pain, Lahuerta et al. (1994) reported that “Most patients (with central pain) noted disagreeable abnormal sensations on cutaneous stimulation within the analgesic area.” Thus, tests of cutaneous sensitivity could be appropriate for detection of aversive dysesthesias referred to regions affected by interruption of the spinothalamic tract. Effects of anterolateral cordotomy on operant escape responses of monkeys Several studies with monkeys have evaluated changes in sensitivity to nociceptive stimulation of the legs over time after interruption of the spinothalamic tract at a thoracic level (Vierck and Luck, 1979; Vierck et al., 1990a). The monkeys were trained preoperatively to terminate (escape) electrocutaneous stimulation of either lateral calf by pulling on a manipulandum with either hand. Following interruption of pathways in one lateral spinal column, substantial contralateral decrements in sensitivity (reduced escape frequencies and elevated latencies) were obtained for a wide range of stimulus intensities. Postoperative sensitivity to contralateral stimulation was clearly less than for preoperative testing, in contrast to postoperative sensitivity to ipsilateral stimulation, which did not differ from preoperative performance or was enhanced. The contralateral deficit was observed for escape responses to previously nociceptive

413

stimulus levels but not for detection of innocuous stimulation, as determined by a testing paradigm involving food reinforcement for responses to low levels of stimulation (Greenspan et al., 1986). These results established an important first component of an animal model of spinothalamic tractotomy. Consistent with evidence from human beings, there should be a diminution in nociceptive sensitivity and a sparing of non-nociceptive sensitivity contralateral to a lesion involving one anterolateral spinal column. Our working assumption for the primate model was that nociceptive sensitivity would return over time following anterolateral cordotomy for at least some of the animals (Vierck, 1991). In an initial study with new-world Cebus albifrons monkeys (Vierck and Luck, 1979), recovery to preoperative levels of responsivity occurred for each animal. A subsequent study with old-world Macaca nemestrina monkeys (Vierck et al., 1990a) produced a result consistent with reports involving long-term clinical observations of human patients with chronic pain (i.e., delayed recovery in approximately 50% of the cases; White and Sweet, 1969). More generally, regardless of whether recovery occurred and was sustained, sensitivity contralateral to the lesion was exceedingly variable over long postoperative testing periods. To illustrate this lability, different patterns of change in contralateral sensitivity over months of testing after cordotomy are depicted in Fig. 1 for four monkeys; two that were classified as recovered and two that were categorized as unrecovered, based upon their level of sensitivity to contralateral stimulation. Fig. 1 shows the time course of changes in operant response speed for electrocutaneous stimulation of the left (contralateral) and right (ipsilateral) legs following anterolateral cordotomy on the right side. Escape speed for each trial consisted of the maximum trial duration minus the operant response latency, so that high speeds represent short trial durations and a high level of aversion. For these plots, response speeds were averaged across four stimulus intensities presented in each testing session (0.1, 0.4, 1.1 and 2.5 mA/mm2). These graphs reveal overall levels of aversion for stimulus intensities that ranged for normal animals (preoperatively) from a level near threshold for escape (eliciting escape on 50% of the trials) to an intensity that produced minimal response latencies on each trial (near 200

ms). The horizontal lines represent average response speed over the last 4 weeks of preoperative testing. Each data point for postoperative testing represents a cascaded average over 10 testing sessions (week 2) or 20 sessions (all subsequent points). That is, the first postoperative value is averaged over weeks 1-2; subsequent points are averaged over weeks l-4,3-6, 5-8, etc. This reveals gradual changes in contralatera1 sensitivity over biweekly periods but obscures day-to-day variability. Despite the eventual categorization as recovered or unrecovered, contralateral sensitivity (large symbols in Fig. 1) returned to or near preoperative levels for each animal at some point in the postoperative testing period. Note that ipsilateral sensitivity (small symbols in Fig. 1) was relatively stable, providing a control for the possibility that the contralateral changes resulted from spontaneous shifts in behavioral bias toward responding. Ipsilateral sensitivity was generally increased over preoperative levels (reaching statistical significance for animals eventually classified as recovered). Methodological nociception

considerations for evaluating

A considerable advantage of the animal model in understanding mechanisms of recovery from cordotomy is that progression of a disease process is not a factor, as it is in clinical cases (e.g., cancer). In addition, it is feasible to test the animals daily, to plot changes in sensitivity with good resolution over time. Other features of the animal model are pertinent to evaluation of the results. Electrocutaneous stimulation is advantageous for training and testing animals on an escape paradigm, because the evoked sensation disappears abruptly when the appropriate response occurs. For forms of stimulation that offset gradually (e.g., ramped heat) and/or produce prolonged after-sensations, it is difficult for an animal to associate occurrences of an operant response with termination of a nociceptive sensation. Electrocutaneous intensities ranging from levels that do not elicit escape to levels that reliably produce escape at minimal latencies have been evaluated by testing human observers (e.g., Vierck et al., 1983, 1995; Vierck and Cooper, 1984; Cooper et al., 1986). The threshold for escape by monkeys is comparable to pain threshold for humans, and there are many

414 0.9

A.

% 0.5 0.4 1 ; 0.3 J ,j ;I: 00 : ,I Pre

2

/ 1 1 -r-1 , -~,~_I 6 10 14 18 22

1 _ ,T , I-7 30 34 38 42

71 26

bst0P.

1 r r--, 46 50

1 T 1 r-i 7 T- r~ 54 58 62 66 70

wssks

30 B

25

f 2.0 cn 1.5 % p r3 0.5 00

I - lpsi ( 7 _ .-TIPreopm r 1 - ,r- r7 Pre

2

6

10

14

18

22

26

30

+ Contra. e-717- r---l~-, 1~1 ~-1-1 n-1 734

Postop.

36

42

46

50

54

58

62

66

70

74

weeks

1.8

C.

16

+

Postop.

Contra.

weeks

20.

i br): 1.’ % 0.6 I lt3 0.4 I

7-

Pre 2

6

--1-

TT

r

-1-7

-r-r--,-

-r-r

~-

7-

r-r-7-7

-7

1

10 14 18 22 26 30 34 38 42 46 50 54 56 62 66 70 74 76 82 Postop.weeks

415

discriminable levels of nociceptive intensity (just noticeable differences or JNDs) across the stimulus range. Human beings describe sensations subthreshold for pain as tingling (sometimes itching). Painful sensations elicited by higher intensities have a realistic thermal component for the parameters utilized (50 ms of 60 Hz stimulation, separated by 200 ms interstimulus intervals). It is important to recognize that escape responses for normal animals on this paradigm occur exclusively in relationship to first pain. The animals respond to the high intensities before second pain would occur. The effects of cordotomy could differ for methods of stimulation that preferentially activate unmyelinated nociceptors (e.g., Cooper et al., 1986; Yeomans et al., 1996; Lu et al., 1997). Sensations elicited by the nociceptive stimulus intensities range from very weak to strong pain for human observers (Vierck et al., 1983), indicating that the shifts in operant responsivity shown in Fig. 1 represent substantial changes in nociceptive sensitivity. Finally, the effects of anterolateral cordotomy on escape behavior do not parallel effects on flexion reflexes that are often utilized as behavioral indices of nociceptive sensitivity. Flexion reflexes can be depressed bilaterally in cases of a strictly contralateral effect of cordotomy on operant escape responses (Vierck et al., 1990a; Vierck and Light, 1999), and alterations of reflex sensitivity do not parallel changes in pain sensitivity after cordotomy in humans (Garcia-Larrea et al., 1993). Effects of spinal lesion size and distribution contralateral nociceptive sensitivity

on

Initial goals of the animal model were to relate alterations of cellular interactions to the time course of re-

covery from cordotomy (e.g., Bullitt et al., 1988) and to identify histologically the lesion configuration(s) that permit or promote the return of contralateral nociceptive sensitivity. For example, recovery of nociceptive sensitivity could depend upon incomplete interruption of the spinothalamic tract. Incomplete injuries to peripheral nerves can result in exaggerated responses to cutaneous stimulation (Seltzer et al., 1990; Bennett, 1993), possibly implicating partial deafferentation in the generation of pathological pain processing (Kingery and Vallin, 1989; Palecek et al., 1992). If these principles apply to partial deafferentation of rostra1 projection targets of the spinothalamic tract in the brain stem, thalamus (see also Dostrovsky, 2000, this volume), hypothalamus, cerebral cortex (see also Bromm et al., 2000, this volume; Casey, 2000, this volume) and elsewhere, an exaggerated and qualitatively altered response to spared input could result. That is, stimulus-response functions for contralateral stimulation would be expected to reveal aversive responses at some point in time to normally innocuous stimulation (allodynia) and/or exaggerated responses to normally nociceptive stimulation (hyperalgesia). Delayed appearance of these effects could depend upon neural reorganization in response to partial deafferentation. The lesions shown in Fig. 1 indicate that the incidence of contralateral recovery of pain sensitivity cannot be explained as the result of greater sparing of spinothalamic tract axons. The largest lesion (panel D) is associated with gradual return of nociceptive sensitivity to levels of responsivity greater than during preoperative testing. The smaller lesions for animals shown in Fig. 1 (panels A and B) were associated with substantial hypoalgesia after months of postoperative testing. In general, prolonged return of contralateral sensitivity in this study of Mucaca

Fig. 1. Average escape speed across stimulation intensities is shown for 4 monkeys after surgical lesions of the right anterolateral column at an upper thoracic level. Stabilized preoperative speeds are shown for the left and right sides @re), and the average preoperative speeds for both sides are shown as thick horizontal lines. Escape speeds are calculated from escape latencies (see Vierck et al., 1990a) so that enhanced postoperative sensitivity is represented by values above the preoperative mean, and decreased sensitivity is shown by values below preoperative. In this study, cascaded postoperative means were plotted for stimulation of each hindlimb across 4 weeks of daily sessions (5 per week), to reveal gradual changes in sensitivity. Two animals (A and B) were classified as unrecovered, based on substantial decreases in response speed for contralateral stimulation after more than a year of testing. Two other animals (C and D) were classified as recovered, based upon a progressive increase in contralateral sensitivity that was maintained over time. Responses to ipsilateral stimulation were generally increased by anterolateral cordotomy, especially for large lesions associated with recovery of contralateral sensitivity (e.g., panel D).

416

A.

LEFT SIDE FROM RIGHT THALAMUS* 60

I3 Cervical E3 Lumbar

I, II

C.

III, IV

B.

Silver Sliver

V

Ventral

El Cervical

El Cervical El Lumbar

I, II

D.

LEFT SIDE FROM LEFT THALAMUS

60;

RIGHT SIDE FROM LEFT THALAMUS 60

V

Ventral

RIGHT SIDE FROM RIGHT THALAMUS

60

TMB

ill, IV

TMB TMB

13 Cervical •I Lumbar

Silver Silver

50

$ d a t *

40

30 I 20.

0 I, II

III, IV

V

Ventral

I, II

Ill, IV

V

Ventral

Fig. 2. Cell counts on the left (panels A and C) and right sides (panels B and D) of spinal segments at the cervical and lumbar enlargements of a monkey after injection of the left thalamus with HRP and the right thalamus with apo-HRP conjugated to gold beads. The number of cells containing HRP reaction product (TMB) or silver grains (apo-HRP, demonstrated with silver intensification) were averaged over four cervical or lumbar segments for four regions of the gray matter (laminae I and II, laminae III and IV, lamina V and all more ventral laminae). Both injections labeled a small number of ipsilateral cells (panels C and D). The anterolateral spinal lesion cat T6 on the right side) did not appreciably interrupt ipsilateral transport of label. The right anterolateral lesion substantially reduced the number of labeled cells in laminae I-V of the left (contralateral) lumbar segments (panel A; injection in the right thalamus). Controls for this lesion effect were provided by cervical transport on either side from the opposite thalamus (panels A and B) and lumbar transport of label from the left thalamus to the right side (panel B).

nemestrina monkeys (Vierck et al., 1990a) was associated with lesions that were among the largest. In order to more formally evaluate the sparing of spinothalamic tract projection cells in monkeys with an anterolateral spinal lesion, four animals received thalamic injections of retrograde tracers in the thalamus. Fig. 2 shows data from one animal that satisfied criteria of complete bilateral filling of nucleus ventralis posterolateralis (VPL) and adjacent thalamic terminations of the spinothalamic tract. The injections occurred 9 months after a lesion of the

right anterolateral column at T6. In brief, recordings from the thalamus under surgical anesthesia were utilized to locate VPL on each side. Horseradish peroxidase (HRP) was injected into the left thalamus at the site of maximal evoked responses from the right foot, and gold particles conjugated to apo-HRP were injected at multiple sites in the right thalamus, in and around the focus of evoked activity from stimulation of the left foot. After 3 days survival, the animal was killed, the spinal cord was processed, and cervical and lumbar neurons were evaluated for HRP reac-

417

tion product and gold particles intensified with silver deposition. This method permitted identification of neurons on both sides of the spinal cord that projected to the left thalamus (HRP reaction product) and the right thalamus (silver grains). Four blocks of tissue from segments in the cervical and lumbar enlargements were sectioned, and the labeled neurons in each block were counted and assigned to laminar regions. The principal effect of a lesion on the right side was to virtually eliminate the projection of lamina I, II and V neurons from the lumbar enlargement on the left to thalamus on the opposite side. An average of 1.3 lumbar neurons in laminae I and II on the left projected past the anterolateral lesion to the right thalamus, in contrast to 40 or more neurons that transported label contralaterally from thalamus to cervical segments and to lumbar segments on the right. Thus, the lesion did not spare projections from the superficial dorsal horn that shift dorsally in the lateral spinal columns at high spinal levels (Ralston and Ralston, 1992). Similarly, an average of 1.5 lamina-V neurons on the left were labeled in lumbar segments after thalamic injection on the right. In contrast, 25 to 32 neurons were labeled in lamina V on the right (cervical and lumbar sections) and on the left side of cervical segments, after contralateral thalamic injections. Although there was an almost complete lack of labeled dorsal neurons in lumbar segments on the left after injection of the right thalamus, an average of 11.5 ventral neurons was labeled. Thus, there was slight sparing of spinothalamic projections to the right thalamus from ventral neurons in lumbar segments on the left, despite the presence of an anterolateral lesion on the right. In addition, small numbers of spinal projections to the ipsilatera1 thalamus from lumbar segments were identified. These originated predominately in the ventral gray matter. As expected, the anterolateral lesion spared some spinothalamic projections to the thalamus. In addition to a slight sparing of direct spinothalamic projections by anterolateral cordotomy on the right side, input from the contralateral (left) side projects indirectly to the right thalamus via dorsal and ventral spinal pathways on the left (e.g., via spinal lemniscal and spinoreticular pathways). These pathways receive nociceptive and non-nociceptive input (Vierck et al., 1986; Willis and Coggeshall,

1991). The partial deafferentation of the somatosensory thalamus by anterolateral cordotomy (Berkley, 1980) poses a question that is pertinent to the issue of recovery from contralateral hypoalgesia following anterolateral cordotomy. Which pathways are responsible for rostra1 conduction of nociception after contralateral interruption of the spinothalamic tract? That is, which spinal pathways must be interrupted to eliminate sensitivity to nociceptive stimulation (i.e., produce analgesia)? Effects of sequential lesions to the spinothalamic tract and then other spinal pathways have been evaluated in monkeys (Vierck and Luck, 1979). Following recovery of sensitivity to nociceptive stimulation contralateral to an anterolateral lesion, interruption of dorsal spinal pathways did not eliminate cutaneous pain sensitivity that had returned after anterolateral cordotomy. Also, a large lesion sparing only one anterolateral column has been reported to preserve nociceptive sensitivity bilaterally in a human patient (Noordenbos and Wall, 1976; Danziger et al., 1996). In contrast, ventral hemisection resulted in an enduring loss of sensitivity to nociceptive stimulation in monkeys. Similarly, bilateral ventral lesions in human beings have been reported to produce analgesia (Triggs and Beric, 1992). These findings are of particular interest in comparison to bilateral anterolateral cordotomy, which does not eliminate pain sensitivity in monkeys or humans (White and Sweet, 1969; Vierck and Luck, 1979; Lahuerta et al., 1994). It appears that axons in the ventral half of the spinal cord, not restricted to the spinothalamic tract, are critical for elicitation of pain by nociceptive stimulation of dermatomes well below the level of the lesion(s). Extensive lesions of the ventral spinal cord are not an option for control of clinical pain, because of involvement of motor and autonomic pathways. Furthermore, large bilateral lesions of the ventral spinal cord that eliminate cutaneous nociception are associated with severe central pain and allodynia (Triggs and Beric, 1992). The allodynia appears to result from interruption of ventral pathways and sparing of dorsal spinal pathways. In addition, development of central pain following spinal cord contusion injury has been suggested to depend upon incomplete injuries that spare some rostra1 transmission (Beric, 1993). Thus, the concept that deafferentation zone

418

pain can develop following complete spinal transection may be incorrect and related to difficulties in discriminating complete from incomplete injuries (Davidoff et al., 1987; Beric et al., 1988). In contrast to large ventral spinal lesions that can be associated with development of central pain and allodynia, superficial lesions of the anterolateral spinal cord that interrupt spinothalamic axons but do not extend medially to involve the gray matter appear to produce an enduring contralateral hypoalgesia for monkeys (Vierck and Luck, 1979; Vierck et al., 1990a). Similarly, Nathan and Smith (1979) have presented a series of patients who obtained long-term relief of chronic pain contralateral to anterolateral lesions that were shown histologically to be uniformly small and superficial. These results suggest that anterolateral cordotomy can be effective for long-term attenuation of pain that is otherwise intractable, if the lesion is appropriately placed and restricted to transection of axons located superficially in the spinal white matter. Based upon the evidence presented above, the following conclusions are offered regarding effects of anterolateral cordotomy in primates. (1) As expected, unilateral interruption of the spinothalamic tract produces a strictly contralateral hypoalgesia. (2) Repeated measurements over time after unilateral cordotomy show that there can be substantial swings in nociceptive sensitivity for stimulation contralateral to the lesion. (3) Recovery of contralateral sensitivity is especially evident following anterolatera1 lesions that extend medially into the gray matter, but recovery is not strictly dependent upon the configuration of these lesions. (4) The variability in effects of cordotomy between and within animals over time suggests that some adventitious consequence of medially extensive lesions influences sensitivity to contralateral stimulation. Effects of hemotoxicity at the site of anterolateral cordotomy on sensitivity to ipsilateral stimulation The amount of bleeding into the surgical cavity could be one determinant of the functional effects of cordotomy. Application of blood, hemoglobin or a ferrous compound to CNS gray matter results in a focus of epileptic-like discharge among cortical neurons (Willmore et al., 1986) and produces an

ischemic infarct when injected into the spinal cord (Sadrzadeh et al., 1987). These effects are present in models of border zone pain following ischemia (Xu et al., 1992) or excitotoxic injury (Yezierski et al., 1998) within spinal gray matter. In these models, attention has been directed to overgrooming and hypersensitivity within dermatomes supplied by segments near the level of the spinal lesion. However, it is quite possible that involvement of pathways in the core spinal region (e.g., the propriospinal system of diffuse projections) could influence sensitivity to stimuli applied to dermatomes remote from the site of ischemic injury (Sandktihler, 1996; Liu et al., 1998; Wall et al., 1999). Therefore, the influence of bleeding into anterolateral lesion cavities was tested in a rodent model of nociceptive sensitivity (Vierck and Light, 1999). After training Long-Evans rats to escape electrocutaneous stimulation of either hindlimb by pressing a lever with the right forelimb, the rats received thoracic spinal lesions of one of two types. Some anterolateral lesion cavities were filled with gelfoam pledgets soaked in the animal’s blood, and there was no intentional introduction of blood into the lesion cavities of other animals. Intentional introduction of blood into the lesion cavities consistently resulted in an ipsilateral hypersensitivity that was apparent as soon as the animals could be tested after surgery (approximately 1 week) and generally lasted for months thereafter. Significant ipsilateral hypersensitivity had been observed after recovery of contralateral sensitivity in monkeys, but these effects were observed throughout the postoperative testing interval for all rats after introduction of blood into the lesion cavity. Ipsilateral hypersensitivity was not seen consistently in monkeys that did not show long-term recovery of contralateral sensitivity or in rats without intentional introduction of blood into the lesion cavity. The occurrence of ipsilateral hypersensitivity after anterolateral cordotomy, without the presence of a source of pathological pain in the periphery, puts clinical observations of mirror image pain (Nathan, 1956; Bowsher, 1988; Nagaro et al., 1993) in a new perspective. Following cordotomy for lateralized chronic pain, patients sometimes develop a persistent dysesthesia ipsilateral to the lesion. In these cases, it is possible that relief of contralateral pain has unmasked ipsilateral pain originating from the

419 contralateral peripheral source. For example, pathological involvement of a peripheral nerve produces bilateral effects on spinal circuits that could result in bilateral sensory abnormalities (Koltzenburg et al., 1999). However, there was no peripheral pathology in the monkeys or rats that became hypersensitive to stimulation ipsilateral to anterolateral spinal lesions. Another potential explanation for ipsilateral hypersensitivity is that the anterolateral lesions interrupted descending spinal pathways which exert inhibitory effects on caudal spinal circuits (Jones and Light, 1992; Sandkiihler et al., 1993; Smith et al., 1995), resulting in disinhibition of nociceptive transmission. However, comparing animals with and without intentional hemotoxicity, there were no discernable differences in configuration of white matter damage that could account for the presence of ipsilateral allodynia and hyperalgesia. These effects apparently depended upon introduction of blood into the lesion cavity, without additional destruction of spinal white matter that could be observed with standard histological techniques. An implication of the ipsilateral hypersensitivity is that hemotoxic effects on the spinal gray matter change the excitability of distant neurons in nociceptive pathways. It is apparent that responsivity within segments near the distribution of a spinal lesion can be increased by excitotoxicity (Yezierski et al., 1998). However, the relevance of gray matter damage to responsivity within segments well outside the distribution of a lesion has not been emphasized previously. At this point, it is not clear whether remote effects of gray matter involvement should be attributed to interruption of propriospinal conduction or to excitatory propagation rostrally and/or caudally from the lesion site (e.g., via Lissauer’s tract). Sensitivity to contralateral anterolateral cordotomy

stimulation

following

Despite the consistent enhancement of ipsilateral sensitivity by cordotomy plus hemotoxicity, recovery from a contralateral attenuation of nociceptive sensitivity was not obviously enhanced by intentionally introducing blood into the lesion cavity. In retrospect, however, restricting attention to sustained recovery of nociceptive sensitivity ignores a common effect of anterolateral cordotomy on contralat-

era1 sensitivity. Even monkeys that were classified as unrecovered on the basis of sustained late postoperative performance had one or more periods in which contralateral sensitivity was near preoperative levels (e.g., Fig. lA,B). Classification of animals as recovered would have been almost unanimous for medially extensive lesions, if the criterion had been contralateral sensitivity that was close to the preoperative level at any time in the postoperative period of testing. Furthermore, averaging performance over weeks to reveal long cycles of increased and decreased sensitivity in the monkey study obscured a considerable day-to-day variability that is characteristic for stimulation contralateral to an anterolateral lesion. This is exemplified in Fig. 3, which presents daily means of responsivity for rats after anterolatera1 cordotomy with hemotoxicity. Testing for extended periods before and after surgery provides the opportunity to regard each animal as a case study and permits secure descriptions of within-animal variability, in addition to the more common comparisons between animals. Panels A and B in Fig. 3 typify the operant escape performance of rats with substantial and enduring contralateral hyposensitivity (on the average) and ipsilateral hypersensitivity after cordotomy plus hemotoxicity. As in Fig. 1, the response measure is operant escape speed, averaged across stimulus intensities. There were five intensities, ranging from below normal escape threshold to a value that reliably produced escape by unlesioned animals at minimal latencies. In contrast to Fig. 1, performance during sequential daily testing sessions is plotted, rather than averaging across days. Contralateral sensitivity is represented by bars, and ipsilateral sensitivity is shown as triangles. The average preoperative response speed is depicted by a horizontal line. Notice the extreme variability in contralateral responsivity for both animals, ranging from apparent analgesia on some days to levels of responsivity on other days that were equal to or greater than the average preoperative performance for that animal. That is, during some sessions the animals did not terminate stimulation during 5 s trial periods at any stimulus intensity (or they responded only occasionally to the highest intensity), but they responded frequently to all intensities during other sessions. Of particular interest is a late period (around the 80th day of post-

420

5

10

15

20

25

30

Postopemtive

35

40

45

50

55

60

Days

65

0

70

75

Contra.

80

+

65

90

95

100

Ipsi.

2.5 P '" m $ g e ," 0 W

2.0 1.5 1.0 0.5 0.0 5

10

15

20

25

30

35

Postoperative

3

40

45

50

55

60

Days

65

70

0

75

60

Contra.

65

90

95

100

--t Ipsi.

2.0

'Ir 00 1.5 B (I) 0

1.0

:: yI::

0.5 0.0 20

30

40

Postopemtlve

50

60

70

Days

60

90

100

0

Contra.

110

120

-t

130

Ipsi.

D.

Postopemtive

Days

0Contra.

-lpsi.

140

150

421

operative testing) in which these animals (and two others) were above normal levels of responsivity. Because the electrocutaneous stimulus (current) was monitored on each trial, we can be sure that these variations in responsivity were not a consequence of drifts in stimulation intensity. Surgery for these animals was within several days, so that their postoperative testing schedules were similar. The period of heightened contralateral responsivity overlapped for the four rats (but was not identical in onset or duration) and was likely triggered by some environmental event. After the fact, we determined that the timer controlling light in the room housing the animals had malfunctioned during this period of time, producing 24 h of light for several days. Therefore, it is possible that stress precipitated a period of heightened sensitivity contralateral to anterolateral cordotomy (Davis and Martin, 1947; Bowsher, 1996). Ipsilateral sensitivity, however, was not similarly affected. Panel C in Fig. 3 depicts a pattern of delayed and intermittent contralateral hyposensitivity that was observed for some rats following cordotomy with hemotoxicity. Early in the postoperative testing period, contralateral sensitivity appeared to be normal for approximately half the rats, and then contralateral hyposensitivity appeared later, for periods of varying length. The lesions for these animals tended to include the entire lateral column on one side and extend medially to involve gray matter (as in Fig. 3C). Other animals with similar lesions were more consistently hyposensitive in the early postoperative period. Thus, delayed appearance of contralateral hyposensitivity was not attributable to the distribution of white matter damage. Some other factor - for example, ischemic involvement of the gray matter and associated hyperactivity of spinal neurons (Yezierski and Park, 1993) - could have been maximal in the early postoperative period. Furthermore, excitotoxicity could wax and wane over time, accounting for

periods of relative hyper- and hyposensitivity. This is a difficult hypothesis to test, because histological techniques give. only one point in time per animal. However, significant attenuation of functional variability by a spinally acting treatment over an extended period would provide very useful information concerning mechanisms of hypersensitivity. Panel D in Fig. 3 presents a lesion that extended through the dorsal gray matter to bilaterally involve lateral column white matter. This is the only animal with bilaterally decreased nociceptive sensitivity in the early postoperative interval (even though the spinothalamic tract was largely spared on the left side). Another unusual finding with this animal was that nociceptive sensitivity recovered within a month (bilaterally), indicating that extensive damage to the spinal gray matter contributes to recovery from cordotomy. The bilateral involvement of the core spinal region is reminiscent of compression injuries, where bilateral cellular loss within the gray matter is certain, and central pain and hypersensitivity are common (Pagni, 1984; Milhorat et al., 1996). Thus, it is proposed that involvement of gray matter and propriospinal systems contributes to the development of dysesthesia and pain, particularly when extensive damage to the gray matter is associated with partial interruption of the spinothalamic pathway. This suggests that prevention or attenuation of secondary cellular excitotoxicity consequent to spinal cord injury could prevent development of central pain, either by minimizing damage to the core spinal region or by attenuating abnormal activity among neurons near the injury. Episodic allodynia and hyperalgesia for segments caudal to anterolateral cordotomy For both monkeys and rats, anterolateral cordotomy results in periods of apparently normal sensitivity,

Fig. 3. Escape speed is shown for 4 rats after surgical lesions of the right anterolateral column at an upper thoracic level with hemotoxicity. Average escape speed for stimulation of the left (contralateral) and right (ipsilateral) hindpaw is shown for single postoperative testing sessions that included 8 trials at each of 5 electrocutaneous stimulus intensities. The average preoperative escape speed for stimulation of both hindpaws is shown as a horizontal line. Animals A and B revealed substantial decreases in escape speed for contralateral stimulation over 100 days of postoperative testing, but there were days in which contralateral sensitivity was near to or greater than preoperative levels. Average postoperative escape speed for animals C and D was below preoperative levels for stimulation of the contralateral hindpaw, but contralateral sensitivity was frequently at or above preoperative levels.

422 4.0 3.5

l -PREOP. MIN. A.. PREOP. MAX. -e-IPSI. MIN. -e-IPSI. MAX.

3.5

- A -. PREOP. MAX. -E--CONTRA.

3.0

+

3 3.0 f

2.5

i

2.0

l -. PREOP. MIN.

,

0.5

MIN.

CONTRA.

MAX.

0.5 , A

0.0 0.05

0.10

0.20

0.40

Stimulus

Intensity (mA)

0.80

0.0 1 0.05

0.10

0.20

0.40

0.80

Stimulus Intensity (mA)

Fig. 4. Stimulus-response (S-R) functions are averaged over 5 preoperative and 5 postoperative sessions in which operant escape speeds were lowest (min.) or highest (max.) for stimulation of each hindpaw of 14 animals that received unilateral spinal lesions with hemotoxicity. Panel A shows that postoperative allodynia and hyperalgesia for ipsilateral stimulation were evident during maximally sensitive sessions, but the slope of the ipsilateral S-R functions was not altered postoperatively. In contrast, the S-R functions for contralateral stimulation were considerably flattened (panel B). Allodynia was apparent for the sessions in which contralateral sensitivity was high.

periods of contralateral hyposensitivity and periods of ipsilateral hypersensitivity. The plots that were utilized to reveal the time course of these changes in overall sensitivity averaged performance across all stimulus intensities delivered in a testing session (Fig. 3) or across weeks (Fig. 1). Another way to compare preoperative and postoperative sensitivity is to plot stimulus-response (S-R) functions. This provides opportunities to determine whether postoperative changes in sensitivity are restricted to certain stimulus intensities (e.g., indicating that allodynia and/or hyperalgesia has occurred). Of particular interest is the relative form of S-R functions during periods of maximal and minimal sensitivity to stimulation contralateral and ipsilateral to a lesion. S-R functions for relatively sensitive and insensitive periods were obtained by sorting the average response speeds (across stimulus intensities) for each animal, testing period (preoperative and postoperative) and leg stimulated. Stimulus-responses functions for each animal were then averaged across the five least responsive and five most responsive sessions for each leg and testing period. Minimal and maximal S-R functions for all animals with unilateral cordotomy plus hemotoxicity (n = 14) are shown in Fig. 4. Panel A of Fig. 4 shows S-R functions for stim-

ulation of the right hindpaw (ipsilateral to cordotomy). During sessions in which responsivity was lowest for stimulation of the ipsilateral leg, there was no difference between preoperative and postoperative performance, and escape speed increased nearly linearly for the stimulus intensities presented. In contrast, there was a considerable difference in S-R functions for ipsilateral stimulation during the most responsive preoperative and postoperative sessions. The ipsilateral hypersensitivity that was revealed by averaging response speed across stimulus intensities (Fig. 3) was dependent upon a subset of postoperative sessions. The lowest stimulus intensity (0.05 mA) was subthreshold for escape responding during all preoperative sessions but did elicit escape during postoperative testing sessions in which the rats were maximally responsive (revealing allodynia). In addition, postoperative escape speed was greater than preoperative speed for the higher stimulus intensities (revealing hyperalgesia) during sessions in which responsivity was maximal. Panel B of Fig. 4 shows minimal and maximal S-R functions for stimulation of the left hindpaw (contralateral to cordotomy). In contrast to ipsilateral S-R relationships, where the slopes of preoperative and postoperative functions were similar, the

423 2.5 ,

/ A.

0.Q

LI PRE

1 0.8 ;

B.

q PRE

POST

2.0 !

Contra.

POST

Ipsi.

Contra.

Ipsi.

Fig. 5. Panel A shows escape speeds averaged across all preoperative and postoperative testing periods for each of 14 animals with spinal lesions confined to one lateral column with hemotoxicity. Nociceptive sensitivity was significantly reduced contralaterally and significantly increased ipsilaterally. Between animal standard deviations are shown for each condition. Panel B shows within-animal variability across the preoperative and postoperative testing periods for stimulation of the left (contralateral) and right (ipsilateral) hindpaws. Coefficients of variability (CVs) were calculated for each animal and then averaged (between-animal standard deviations are shown for the CVs). Escape speeds were considerably more variable for postoperative stimulation of the contralateral hindpaw than for any other condition.

contralateral slopes were smaller (flatter) postoperatively than preoperatively. During postoperative sessions in which sensitivity was minimal, the animals were nearly analgesic for contralateral stimulation, responding only occasionally at the highest stimulus intensity. During postoperative sessions in which sensitivity was maximal, the animals were allodynic, as evidenced by responses to the lowest stimulus intensity. These animals rarely responded to the lowest intensity during preoperative testing, even during sessions in which they were most responsive. Thus, for certain periods following anterolateral cordotomy, escape responses to low levels of contralateral stimulation were enhanced, providing evidence for abnormal nociceptive processing and showing that cordotomy does not globally depresses contralateral pain sensitivity. It is not necessary to show prolonged recovery of contralateral sensitivity to model a manifestation of central pain following spinal cord injury. It is apparent from plots showing the time course of changes in operant pain sensitivity after cordo-

tomy, such as those shown in Figs. 1 and 3, that contralateral responsivity is labile. Dramatic changes can occur over time, even from day to day. There are examples of adjacent sessions in which an animal is nearly analgesic or is allodynic for contralateral stimulation, and there are gradual fluctuations in sensitivity that occur over weeks. The simplest way to demonstrate changes in sensitivity, regardless of the periodicity, is shown in Fig. 5B, where coefficients of variability (CVs) for escape speed were calculated for each of 14 animals, before and after cordotomy with hemotoxicity. For reference, response speeds averaged across all preoperative and postoperative sessions for the same group of animals are shown in Fig. 5A. The averaged, within-animal CVs for postoperative stimulation contralateral to anterolatera1 cordotomy were substantially larger than preoperative values (nearly tripled), and this difference was comparable for the comparison of contralateral with ipsilateral stimulation after cordotomy. Labile nociceptive sensitivity was characteristic of all animals for contralateral stimulation, providing a model

424

of aberrant nociceptive processing for stimulation within the deafferentation zone. Using variability as a principal measure of abnormal processing and plotting of stimulus-response functions for operant escape responses provides an opportunity to evaluate a variety of therapeutic procedures. For this purpose, it is important to utilize methods permitting repetitive evaluation (e.g., behavioral testing or chronic physiological recording) to model deafferentation zone pain sensitivity. Extreme variability in the magnitude of pain is well-recognized and is characteristic following injuries to peripheral nerves or to the CNS that result in a pathological pain condition (Bowsher, 1996; Eide et al., 1996). For example, a recent case report documents considerable moment-to-moment variations in pain levels in a patient with spinal cord injury (Ness et al., 1998). Similarly, spinal lesions involving the dorsal columns produce exaggerated variability in motor control and in performance on certain tests of non-nociceptive somatosensory capabilities (Vierck, 1982; Vierck et al., 1990b). That is, interruption of the spinothalamic pathway results in substantial variability in pain sensitivity, and interruption of the dorsal column pathway produces abnormal variability on sensory and motor tests that depend upon transmission via that pathway. These observations indicate that deafferentation represents an important mechanistic component of fluctuations in pain sensitivity following peripheral or central reduction of nociceptive input. After reduction of input from the spinothalamic pathway, rostra1 networks of partially deafferented neurons become unstable, as demonstrated by neurophysiological recordings in the somatosensory thalamus of humans with central pain (Lenz et al., 1994; see also Lenz et al., 2000, this volume) and animals with spinal lesions (Weng et al., 2000). It is important to understand mechanisms contributing to this instability, to identify which patterns of activity are correlated with episodes of pain and hypersensitivity, and to develop means of limiting or eliminating abnormal patterns that coexist with pain. It is not sufficient to ask whether or not abnormal activity is present among deafferented neurons in an individual with central pain; a more pertinent question is whether a certain type of abnormal activity is associated with periods in which pain (and/or hypersensitivity) is greatest.

The contrast in variability of postoperative sensitivity to contralateral and ipsilateral stimulation (Fig. 5) is instructive in terms of spinal or supraspinal sources of episodic modulation. Because the propriospinal system is a diffusely connected system (Nathan and Smith, 1959), it could be affected bilaterally by hemotoxic influences on a portion of the core spinal region. Hemotoxicity produced an ipsilateral allodynia and hyperalgesia that was episodic, but the variability of operant responsivity to ipsilateral stimulation was not substantially increased. Also, the effect of hemotoxicity on ipsilateral nociceptive sensitivity continued for months, suggesting that perioperative damage to the gray matter produced an enduring ipsilateral effect. Similarly, maximal contralateral sensitivity occurred episodically for stimulation that elicited activity in partially deafferented rostra1 structures, and the variability in contralateral sensitivity was extreme. Thus, allodynia for contralateral stimulation may depend upon or develop as a result of trauma to the core spinal region, but fluctuations in contralateral sensitivity likely depend as well upon supraspinal mechanisms. For example, partially deafferented supraspinal neurons could be especially sensitive to fluctuations in levels of abnormal activity at the spinal lesion site, if conducted rostrally. Alternatively, levels of arousal or stress may abnormally modulate partially deafferented structures containing an altered intrinsic circuitry that normally regulates excitatory and inhibitory influences. Surgical interruption of the spinothalamic tract with hemotoxicity in primates results in a deficiency in GABA-ir profiles in the somatosensory thalamus (Ralston et al., 2000), including thalamic reticular axon terminals and local circuit presynaptic dendritic profiles. In addition, interruption of spinothalamocortical projections would reduce inhibitory influences within the primary somatosensory cortex that normally result from nociceptive stimulation (Tommerdahl et al., 1998). These disinhibitory conditions could contribute to the allodynia observed for stimulation contralateral to cordotomy. Conclusions

Persistent pain of central nervous system origin is a common consequence of spinal cord injury and is

425 highly refractory to treatment. In order to understand mechanisms and develop treatments for this condition, an effective laboratory animal model is needed. Clinical observations indicate that interruption of the spinothalamic tract is a prerequisite for development of deafferentation zone pain referred to segments well below the site of spinal cord injury. Therefore, a series of studies has evaluated nociceptive responses of monkeys and rats before and after an anterolateral spinal lesion. Attention to the variability of responses to a wide range of stimulus intensities across testing sessions has shown that contralateral sensitivity oscillates from nearly analgesic to allodynic. All animals cycled between these states of hyper- or hypo-sensitivity. At different periods (either early or late after anterolateral cordotomy), one of these states could predominate for weeks or months, giving the impression that a given animal had recovered or had not recovered from cordotomy. These variable patterns are likely the reason that estimates of the incidence of chronic pain from injury to peripheral nerves or the spinal cord injury range from 10 to nearly lOO%, depending upon applied criteria of severity or longevity. The animal data indicate that, after interruption of the spinothalamic tract, allodynia occurs episodically for each case, but persistent allodynia occurs in a subset of cases. Thus, a manifestation of central pain following spinal cord injury can be studied in an animal model by evaluating operant escape responses that reflect cerebral processing of nociception, and by studying factors that increase or decrease occurrences of allodynia. Interruption of the spinothalamic tract partially deafferents rostra1 neurons in central pain pathways, eventuating in abnormal sensitivity which can be enhanced potentially by a variety of influences, including input from non-nociceptive afferents. It is presumed that certain patterns of activity among neurons in pain pathways are interpreted as pain. An important issue concerning development of these abnormal patterns of activity is whether interruption of the spinothalamic tract is sufficient. Information from our animal model and observations of human beings indicate that superficial lesions of lateral spinal cord white matter that do not extend into the gray matter produce an enduring contralateral hypoalgesia, with minimal allodynia and/or central pain. Extension of a lesion into the core spinal region

results in a higher incidence of recovery from cordotomy (i.e., increased sensitivity over time). Hemotoxic influences at the lesion site produce ipsilateral allodynia and hyperalgesia, suggesting that involvement of the propriospinal system enhances aversion for input to spinal segments remote from the spinal injury. In addition, maximizing hemotoxic influences increases the likelihood that allodynic episodes will be observed for contralateral stimulation. Considerably more investigation of remote effects of injuries involving the core spinal region is needed. For example, it is important to know whether damage to the propriospinal system has effects on nociceptive sensitivity as a result of (1) a loss or disruption of input to caudal spinal neurons and/or supraspinal structures (a physiological deafferentation), (2) reduced rostra1 and/or caudal inhibition, or (3) excitatory influences that emanate from the site of damage. Acknowledgements The research of the authors that is summarized in this paper was supported by NIH grants NS-14899 and NS-07261 and BSCIRTF funds from the state of Florida. References Bennett, G.J. (1993) An animal model of neuropathic pain: a review. Muscle Nerve, 16: 1040-1048. Beric, A. (1993) Central pain: ‘new’ syndromes and their evaluation. Muscle Nerve, 16: 1017-1024. Beric, A., Dimitrijevic, M.R. and Lindblom, U. (1988) Central dysaesthesia syndrome in spinal cord injury patients. Pain, 34: 109-l 16. Berkley, K.J. (1980) Spatial relationships between the terminations of somatic sensory and motor pathways in the rostra1 brainstem of cats and monkeys, I. Ascending somatic sensory inputs to lateral diencephalon. J. Comp. Neurol., 193: 283317. Boivie, J., Leijon, G. and Johansson, I. (1989) Central poststroke pain - a study of the mechanisms through analyses of the sensory abnormalities. Pain, 37: 173-185. Botterell, E.H., Callaghan, J.C. and Jousse, A.T. (1953) Pain in paraplegia: clinical management and surgical treatment. Proc. R. Sot. Med., 281: 281-288. Bowsher, D. (1988) Contralateral mirror-image pain following anterolateral chordotomy. Pain, 18: 63-65. Bowsher, D. (1996) Central pain: clinical and physiological characteristics. J. Neural. Neurosurg. Psychiatry, 61: 62-69. Bromm, B., Scharein, E. and Vahle-Hinz, C. (2000) Cortex areas involved in the processing of normal and altered pain. In: J.

426

Sandkiihler, B. Bromm and GE Gebhart (Eds.), Nervous System Plasticity and Chronic Pain. Progress in Brain Research, Vol. 129. Elsevier, Amsterdam, pp. 289-302. Bullitt, E., Stofer, W.D., Vierck Jr., C.J. and Perl, E.R. (1988) Reorganization of primary afferent nerve terminals within the dorsal horn of the primate spinal cord caudal to the level of anterolateral chordotomy. J. Camp. Neurol., 270: 549-558. Casey, K.L. (2000) Concepts of pain mechanisms: the contribution of functional imaging of the human brain. In: J. Sandkiihler, B. Bromm and G.F. Gebhart (Eds.), Nervous System Plasticity and Chronic Pain. Progress in Brain Research, Vol. 129. Elsevier, Amsterdam, pp. 277-287. Cassinari, V. and Pagni, C.A. (1969) Central Pain: A Neurosurgical Survey. Harvard University Press, Cambridge. Coderre, T.J., Grimes, R.W. and Melzack, R. (1986) Deafferentation and chronic pain in animals: an evaluation of evidence suggesting autotomy is related to pain. Pain, 26: 61-84. Cooper, B.Y., Vierck Jr., C.J. and Yeoman& D.C. (1986) Selective reduction of second pain sensations by systemic morphine in humans. Pain, 24: 93-l 16. Danziger, N., Willer, J.C., Pidoux, B., Dormont, D., Samson, Y., Fournier, E. and Wall, P.D. (1996) A clinical and neurophysiological study of a patient with an extensive transection of the spinal cord sparing only a part of one anterolateral quadrant. Brain, 119: 1835-1848. Davidoff, G., Roth, E., Guarracini, M., Sliwa, J. and Yarkony, G. (1987) Function-limiting dysesthetic pain syndrome among traumatic spinal cord injury patients: a cross-sectional study. Pain, 29: 39-48. Davis, L. and Martin, J. (1947) Studies upon spinal cord injuries. Nature and treatment of pain. J. Neurosurg., 4: 483-49 1. Dostrovsky, J.O. (2000) Role of thalamus in pain. In: J. Sandkiihler, B. Bromm and G.F. Gebhart (Eds.), Nervous System Plasticiq and Chronic Pain. Progress in Brain Research, Vol. 129. Elsevier, Amsterdam, pp. 245-257. Eide, PK., Jorum, E. and Stenehjem, A.E. (1996) Somatosensory findings in patients with spinal cord injury and central dysaesthesia pain. J. Neural. Neurosurg. Psychiatry, 60: 41 l415. Garcia-Larrea, L., Charles, N., Sindou, M. and Mauguibre, F. (1993) Flexion reflexes following anterolateral corodomy in man: dissociation between pain sensation and nociceptive reflex RIB. Pain, 55: 139-149. Greenspan, J.D., Vierck Jr., C.J. and Ritz, L.A. (1986) Sensitivity to painful and non-painful electrocutaneous stimuli in monkeys: effects of anterolateral chordotomy. J. Neurosci., 6: 380-390. Hao, J.-X., Xu, X.J., Aldskogius, H., Seiger, A. and Wiesenfeld-Hallin, Z. (1991) The excitatory amino acid receptor antagonist MK-801 prevents the hypersensitivity induced by spinal cord ischemia in the rat. Exp. Neural., 113: 182-191. Hao, J.-X., Xu, X.-J., Yu, Y.-X., Seiger, A. and WiesenfeldHallin, Z. (1992) Transient spinal cord ischemia induces temporary hypersensitivity of dorsal horn wide dynamic range neurons to myelinated, but not unmyelinated, fiber input. J. Neurophysiol., 68: 384-39 I.

Horax, G. (1929) Experiences with chordotomy. Arch. Surg., 18: 1140-1164. Hyndman, O.R. and Van Epps, C. (1939) Possibility of differential section of the spinothalamic tract. Arch. Surg., 38: 10361053. Jensen, T. and Rasmussen, P. (1984) Amputation. In: P.D. Wall and R. Melzack (Eds.), Textbook of Pain. Churchill Livingstone, New York, pp. 402-412. Jones, S.L. and Light, A.R. (1992) Serotoninergic medullary raphespinal projection to the lumbar spinal cord in the rat: a retrograde immunohistochemical study. J. Comp. Neural., 322: 599-610. Kauppila, T. (1998) Correlation between autotomy-behavior and current theories of neuropathic pain. Neurosci. Biobehav. Rev., 23: 111-129. King, R.B. (1957) Postchordotomy studies of pain threshold. Neurology, 7: 610-614. Kingery, W.S. and Vallin, J.A. (1989) The development of chronic mechanical hyperalgesia, autotomy and collateral sprouting following sciatic nerve section in rat. Pain, 38: 321-332. Koltzenburg, M., Wall, P.D. and McMahon, S.B. (1999) Does the right side know what the left is doing?. Trends Neurosci., 22: 122-127. Lahuerta, J., Bowsher, D., Buxton, P.H. and Lipton, S. (1994) Percutaneous cervical cordotomy: a review of 181 operations in 146 patients, including a study on the location of ‘pain fibers’ in the second cervical spinal cord segment of 29 cases. J. Neurosurg., 80: 975-985. Lenz, EA., Kwan, H.C., Martin, R., Tasker, R., Richardson, R.T. and Dostrovsky, J.O. (1994) Characteristics of somatotopic organization and spontaneous neuronal activity in the region of the thalamic principal sensory nucleus in patients with spinal cord transection. J. Neurophysiol., 72: 1570-1587. Lenz, EA., Lee, J.-I., Garonzik, I.-M., Rowland L.H., Dougherty, P.M. and Hua, S.E. (2000) Human thalamus reorganization related to nervous system injury and dystonia. In: J. Sandktiler, B. Bromm and G.F. Gebhart (Eds.), Nervous System Plastirity and Chronic Pain. Progress in Brain Research, Vol. 129. Elsevier, Amsterdam, pp. 259-273. Levitt, M. and Levitt, J. (1981) The deafferentation syndrome in monkeys: dysesthesias of spinal origin. Pain, 10: 129-147. Liu, X.-G., Morton, C.R., Azkue, J.J., Zimmermann, R. and Sandkiihler, J. (1998) Long-term depression of C-fibre-evoked spinal field potentials by stimulation of primary afferent A&-fibres in the adult rat. Eur: J. Neurosci., 10: 3069-3075. Lu, Y., Pirec, V. and Yeoman& D.C. (1997) Differential antinociceptive effects of spinal opioids on foot withdrawal responses evoked by C fibre or A6 nociceptor activation. BK J. Pharmacol., 121: 1210-1216. Milhorat, T.H., Kotzen, R.M., Mu, H.T.M. and Milhorat, R.H. (1996) Dysesthetic pain in patients with syringomyelia. Neurosurgery, 38: 940-947. Nagaro, T., Amakawa, K., Arai, T. and Ochi, G. (1993) Ipsilateral referral of pain following cordotomy. Pain, 55: 275276.

427

Nathan, P.W. (1956) Reference o n at the spinal level. J. Net&. Neurosurg. Psychiatry, LOO. asciculi proprii of the Nathan, P.W. and Smith, M.C. spinal cord in man. Brain, 82: 610-668. Nathan, P.W. and Smith, M.C. (1979) Clinico-anatomical correlation in anterolateral cordotomy. In: J.J. Bonica, J.C. Liebeskind and D.G. Albe-Fessar (Eds.), Advances in Pain Research and Therapy. Raven Press, New York, pp. 921-926. Nepomuceno, C., Fine, P.R., Richards, J.S., Gowens, H., Stover, S.L., Rantanuobol, U. and Houston, R. (1979) Pain in patients with spinal cord injury. Arch. Phys. Med. Rehabil., 60: 605609. Ness, T.J., San Pedro, E.C., Richards, J.S., Kezar, L., Liu, H.-G. and Mountz, J.M. (1998) A case of spinal cord injury-related pain with baseline rCBF brain SPECT imaging and beneficial response to gabapentin. Pain, 78: 139-144. Noordenbos, W. and Wall, P.D. (1976) Diverse sensory functions with an almost totally divided spinal cord. A case of spinal cord transection with preservation of part of one anterolateral quadrant. Pain, 2: 185-196. Pagni, C.A. (1984) Central pain due to spinal cord and brain stem damage. In: P.D. Wall and R. Melzack (Eds.), Textbook of Pain. Churchill Livingstone, New York, pp. 481-495. Palecek, J., Dougherty, PM., Kim, S.H., Paleckova, V., Lekan, H., Chung, J.M., Carlton, S.M. and Willis, W.D. (1992) Responses of spinothalamic tract neurons to mechanical and thermal stimuli in an experimental model of peripheral neuropathy in primates. J. Neurophysiol., 68: 1951-1966. Ralston, D.D., Dougherty, PM., Lenz, EA., Weng, H.-R., Vierck, C.J. and Ralston, H.J. (2000) Plasticity of the inhibitory circuitry and neuronal responses in the primate somatosensory thalamus (VB) following lesions of the dorsal column and spinothalamic pathways. In: M. Devor (Ed.), Proceedings of IX Congress of Pain, 7 (in press). Ralston III, H.J. and Ralston, D.D. (1992) The primate dorsal spinothalamic tract: evidence for a specific termination in the posterior nuclei (Po/SG) of the thalamus. Pain, 48: 107-l 18. Rodin, B.E. and Kruger, L. (1984) Deafferentation in animals as a model for the study of pain: an alternative hypothesis. Brain Res. Rev., 7: 213-228. Saade, N.E., Atweh, SE, Jabbur, S.J. and Wall, P.D. (1990) Effects of lesions in the anterolateral columns and dorsolateral funiculi on self-mutilation behavior in rats. Pain, 42: 3 13-321. Sadrzadeh, SM., Anderson, D.K., Panter, S.S., Hallaway, PE. and Eaton, S.A. (1987) Hemoglobin potentiates central nervous system damage. J. Clin. Invest., 79: 662-664. Sandkiihler, J. (1996) Neurobiology of spinal nociception: new concepts. In: G. Carli and M. Zimmerman (Eds.), Towards the Neurobiology of Chronic Pain. Progress in Brain Research, Vol. 110. Elsevier, Amsterdam, pp. 208-224. Sandkiihler, J., Stelzer, B. and Fu, Q.-G. (1993) Characteristics of propriospinal modulation of nociceptive lumbar spinal dorsal horn neurons in the cat. Neuroscience, 54: 957-967. Seltzer, Z., Dubner, R. and Shir, Y. (1990) A novel behavioral model of neuropathic pain disorders produced in rats by partial sciatic nerve injury. Pain, 43: 205-218. Smith, M.S., Schambra, U.B., Wilson, K.H., Page, SO., Hulette,

C., Light, A.R. and Schwimr, D.A. (1995) Alpha 2-adrenergic receptors in human spinal cord: specific localized expression of mRNA encoding alpha 2-adrenergic receptor subtypes at four distinct levels. Brain Res. Mol. Brain Res., 34: 109-I 17. Tommerdahl, M., Delemos, K.A., Favorov, O.V., Metz, C.B., Vierck Jr., C.J. and Whitsel, B.L. (1998) Response of anterior parietal cortex to different modes of same-site skin stimulation. .I. Neurophysiol., 80: 3272-3283. Triggs, W.J. and Beric. A. (1992) Sensory abnormalities and dysaesthesias in the anterior spinal artery syndrome. Brain, 115: 189-198. Vierck Jr., C.J., (1982) Plasticity of somatic sensations and motor capabilities following lesions of the dorsal spinal columns in monkeys. In: A.R. Morrison and PL. Strick (Eds.), Changing Concepts of the Nervous System. Academic Press, New York, pp. 151-169. Vierck Jr., C.J., (1991) Can mechanisms of central pain syndromes be investigated in animal models? In: K.L. Casey (Ed.), Pain and Central Nervous System Disease: The Central Pain Syndromes. Raven Press, New York, pp. 129- 14 1. Vierck Jr., C.J. and Cooper, B.Y. (1984) Guidelines for assessing pain reactions and pain modulation in laboratory animal subjects. In: L. Kruger and J. Liebeskind (Eds.), Neural Mechanisms of Pain. Raven Press, New York, pp. 305-322. Vierck Jr., C.J. and Light, A.R. (1999) Effects of combined hemotoxic and anterolateral spinal lesions on nociceptive sensitivity. Pain, 83: 447-457. Vierck Jr., C.J. and Luck, M.M. (1979) Loss and recovery of reactivity to noxious stimuli in monkeys with primary spinothalamic chordotomies, followed by secondary and tertiary lesions of other cord sectors. Brain, 102: 233-248. Vierck Jr., C.J., Cooper, B.Y. and Cohen, R.H. (1983) Human and non-human primate reactions to painful electrocutaneous stimuli and to morphine. In: R.L. Kitchell and H.H. Erickson (Eds.), Animal Pain Perception and Alleviation. American Physiological Society, Washington, DC, pp. 117-132. Vierck Jr., C.J., Greenspan, J.D., Ritz, L.A. and Yeomans, D.C. (1986) The spinal pathways contributing to the ascending conduction and the descending modulation of pain sensations and reactions. In: T.L. Yaksh (Ed.), Spinal Systems of Afferent Processing. Plenum Press, New York, pp. 275-329. Vierck Jr., C.J., Greenspan, J.D. and Ritz, L.A. (1990a) Long term changes in purposive and reflexive responses to nociceptive stimulation in monkeys following anterolateral chordotomy. J. Neurosci., 10: 2077-2095. Vierck Jr., C.J., Whitsel, B.L., Kulics, A.T. and Cooper, B.Y. (1990b) Alterations of a cortical network of neurons following interruption of the dorsal spinal columns. In: F. Seil (Ed.), Advances in Neural Regeneration Research: Proceedings of the Third International Neural Regeneration Research Symposium. Neurology and Neurobiology, Vol. 60. Wiley-Liss, New York, pp. 335-368. Vierck Jr., C.J., Lee, CL., Willcockson, H.H., Kitzmiller, A., DiRuggiero, D., Bullitt, E. and Light, A.R. (1995) Effects of anterolateral spinal lesions on escape responses of rats to hindpaw stimulation. Somatosens. Mot. Res., 12: 163-174. Wall, P.D., Devor, M., Inbal, R., Scadding, J.W., Schonfeld, D.,

428

Seltzer, Z. and Tomkiewicz, M.M. (1979) Autotomy following peripheral nerve lesions: experimental anesthesia dolorosa. Pain, 6: 103-l 13. Wall, PD., Lidierth, M. and Hillman, I? (1999) Brief and prolonged effects of Lissauer tract stimulation on dorsal horn cells. Pain, 83: 579-589. Walker, A.E. (1943) Central representation of pain. Pruc. Assoc. Res. New. Ment. Dis., 23: 63-85. Weng, H.-R., Lee, J.I., Lenz, EA., Schwartz, A., Vierck, C., Rowland, L. and Dougherty, P.M. (2000) Functional plasticity in primate somatosensory thalamus following chronic lesion of the ventral lateral spinal cord. J. Neurosci. (in press). White, J.C. (1968) Operations for the relief of pain in the torso and extremities: evaluation of their effectiveness over long periods. In: A. Soulairac, .I. Cahn and J. Charpentier (Eds.). Pain. Academic Press, New York, pp. 503-5 19. White, J.C. and Sweet, W.H. (1969) Pain and the Neurosurgeon: A Forty-Year Experience. Charles C. Thomas, Springfield. Willis, W.D. and Coggeshall, R.E. (1991) Sensory mechanisms of the spinal cord. Plenum Press, New York. Willmore, L.J., Triggs, W.J. and Gray, J.D. (1986) The role of

iron-induced hippocampal peroxidation in acute epileptogenesis. Bruin Rex, 382: 422-426. Wycis, H.T. and Spiegel, E.A. (1962) Long range results in the treatment of intractable pain by stereotaxic midbrain surgery. J. Neurosurg., 19: 101-107. Xu, X.-J., Hao, J.-X., Aldskogius, H., Seiger, A. and WiesenfeldHallin, Z. (1992) Chronic pain-related syndrome in rats after ischemic spinal cord lesion: a possible animal model for pain in patients with spinal cord injury. Pain, 48: 279-290. Yeomans, D.C., Cooper, B.Y. and Vierck Jr., C.J. (1996) Effects of systemic morphine on responses of primates to first or second pain sensations. Pain, 66: 253-264. Yezierski, R.P. (1996) Pain following spinal cord injury: the clinical problem and experimental studies. Pain, 68: 185- 194. Yezierski, R.P. and Park, S.-H. (1993) The mechanosensitivity of spinal sensory neurons following intraspinal injections of quisqualic acid in the rat. Neumsci. L.&t.. 157: 115-l 19. Yezierski, R.P., Liku, S., Ruenes, G.L., Kajander, K.J. and Brewer, K.L. (1998) Excitotoxic spinal cord injury: behavioral and morphological characteristics of a central pain model. Pain, 75: 141-155.

J. Sandkiihler, B. Bromm and GE Gebhat-t (Eds.) Progress in Brain Research, Vol. 129 0 2000 Elsevier Science B.V. All rights reserved

CHAPTER 32

Pain following spinal cord injury: pathophysiology and central mechanisms Robert P. Yezierski * University

of Miami,

Department

of Neurological Surgery and The Miami Miami, FL 33136, USA

The condition of pain following spinal cord injury Painful sensations are a frequent and troublesome sequela of paraplegia and quadriplegia following partial or complete lesions of the spinal cord (see review Yezierski, 1996). The condition of pain following spinal cord injury (SCI) was first described over 100 years ago, and the origin of various SC1 pain syndromes is based on the nature of the lesion, neurological structures damaged, and secondary pathophysiological changes of surviving tissue (Beric, 1990; Bonica, 1991; Davidoff and Roth, 1991; Tasker et al., 1991). In spite of recent progress directed towards understanding the mechanism(s) of spinal injury pain, it still remains a major challenge for health professionals (Bonica, 1991; Yezierski, 1996; Eide, 1998). To understand the impact of spinal injury pain on the health care community one needs only examine a sampling of studies which report the incidence of painful sensations at a rate of 60-80% for all SC1 patients with nearly 40% reporting severe pain to the extent they would trade any chance of functional recovery for relief of pain (Nepomuceuno et al., 1979; Beric, 1990; Britell and Mariano, 1991; Tasker

* Corresponding author: R.F?Yezierski, University of Miami, Department of Neurological Surgery and The Miami Project, 1600 N.W., 10th Avenue, R-48, Miami, FL 33136, USA. Fax: + I-305-243-4427; E-mail: [email protected]

Project,

1600 N. W, 10th Avenue,

R-48,

et al., 1991; Levi et .al., 1995; Widerstrom-Noga et al., 1999). The prevalence of pain coupled with the number of new injuries each year underscores the challenge to develop new treatments for patients requiring pain management. Through the use of different experimental models, valuable insights related to the mechanism(s) responsible for the onset of pain following injury have been obtained (WiesenfeldHallin et al., 1994; Yezierski, 1996; Christensen and Hulsebosch, 1997). Continued use of these models will hopefully lead to the identification of appropriate therapeutic targets and the development of novel treatment strategies. In this chapter efforts will be made to review the results of experimental and clinical studies that provide insights into possible mechanism(s) underlying selected pain states following spinal injury. A special emphasis will be placed on studies related to central dysesthetic pain, perhaps the most disabling of all sensory complications associated with SC1 (Nepomuceuno et al., 1979; Tunks, 1986; Davidoff et al, 1987; Beric et al., 1988; Davidoff and Roth, 1991; Tasker et al., 1991). Pain of musculoskeletal, radicular, visceral, and psychogenic origins all play a significant role in the clinical sequela of spinal injury and are discussed elsewhere (Tunks, 1986; Britell and Mariano, 1991; Summers et al., 1991). Similarly the epidemiological and clinical characteristics of different SC1 pain syndromes have been previously reviewed (Tunks, 1986; Bonica, 1991; Davidoff and Roth, 1991; Nashold, 1991; Yezierski, 1996; Siddall et al., 1997; Rintala et al., 1998).

430

1’

Spinal Injury

] ----I *

Neurochemical

Anatomical

AAs (glutamate, GABA) ionic (Na+,Ca”, CI-) peptides (dynorphin, Sub P) 2nd messengers (cGMP, NO, c-fos, NFkB); cytokines (TNF, IL-ID) enzymes (calpain, PLAz, PKC)

I

necrosis, apoptosis, gliosis, demyelination, cytoskeletal damage, deafferentation, sprouting

Physiological 0

excitability,

RF, background

activity,

gain,

after

discharge

+

Clinical/Behavioral allodynia,

hyperalgesia,

pain

Fig. 1. Summary of components in the spinal injury cascade responsible for the onset of pain following injury. Evidence supporting the basic concept of this cascade follows from results of clinical studies as well as those obtained from the ischemic, lesion and excitotoxic models of spinal cord injury (see text). The four major components of the cascade (neurochemical, excitotoxicity, anatomical and inflammation) are represented as being interactive and collectively lead to changes in the physiological state of spinal and supraspinal neurons. The end point of the cascade is the onset of clinical symptoms, e.g., allodynia, hyperalgesia, and pain. Abbreviations: EAAs = excitatory amino acids; Sub P = substance P; cGMP = cyclic guanidine monophosphate; NO = nitric oxide; NFkB = nuclear factor kappa B; PKC = protein kinase C; TNF = tumor necrosis factor; IL-l,3 = interleukin-lg; PL.42 = phospholipase A2; iNOS = inducible nitric oxide synthase; COX-2 = cyclooxygenase-2; RF = receptive field.

One of the first things to consider when discussing potential mechanisms of pain following spinal injury is the initial sequence of events triggered by ischemic or traumatic insult to the cord. Obviously there is significant structural damage to the cord parenchyma leading to a reorganization of spinal and supraspinal circuits responsible for the integration and processing of sensory information. Ischemic or traumatic insult to the cord also brings about changes in the expression of intrinsic chemical systems responsible for maintaining the homeostatic balance between inhibitory and excitatory circuits. Equally important is the cascade of cellular events affecting signaling, transduction and survival pathways of spinal neurons. Collectively, these injury-induced effects have a profound impact on the excitability and background discharges of spinal sensory neurons which ultimately affect both evoked and resting sensibilities. It is noteworthy that many of the patho-

physiological changes described following spinal injury parallel descriptions of events thought to be responsible for the development of pain following peripheral nerve and/or tissue injury (Dubner, 1991; see also Moore et al., 2000, this volume). It was a result of this observation that a common central injury cascade was proposed for the initiation of pain-related behaviors following central or peripheral injury (Yezierski, 1996). This proposal is by no means novel, as Livingston (1943) was one of the first to advance the concept that different pain syndromes may share a common patbophysiology. The different components of this central cascade are shown in Fig. 1 and include anatomical, neurochemical, excitotoxic, and inflammatory events that collectively interact to influence the functional state of spinal sensory neurons leading to the onset of different clinical pain states (allodynia, hyperalgesia, pain).

431

Pathophysiology

of spinal cord injury

The most obvious pathological characteristics associated with traumatic or ischemic injury to the spinal cord include but are by no means limited to the dramatic loss of neurons, damage to surrounding white matter, astrocytic scarring, syrinx formation, and breakdown of the spinal blood brain barrier (Kakulas et al., 1990; Bunge et al., 1993a; see also Vierck and Light, 2000, this volume). Also contributing to the progression of tissue damage are secondary injury cascades that include excitotoxic and inflammatory processes (Young, 1987; Tator and Fehlings, 1991; Hsu et al., 1994; Regan and Choi, 1994; Bethea et al., 1998). Finally, with the aid of specific histological and immunohistological stains one can follow the temporal profile of glia activation (astrocytes, microglial) and the infiltration of macrophages and other inflammatory cell types. Up-regulation of messenger RNA for c-fos, TNF-alpha and dynorphin have also been described following SC1 (Yakovlev and Faden, 1994). From this discussion it should be clear that the pathological sequela of SC1 is by no means simple nor is it restricted to the site of insult as pathological consequences of SC1 have been observed throughout the full extent of the neuraxis (Brewer et al., 1997; Jain et al., 1998; Ness et al., 1998; Morrow et al., 1999). One of the initial consequences associated with stroke, hypoxia-ischemia and traumatic brain injury is the well documented excitotoxic effects of excitatory amino acids (EAAs) (Regan and Choi, 1994). Similarly, evidence supports the involvement of glutamate in the secondary pathology, including neuronal degeneration, cavitation and edema, of ischemic and traumatic spinal injury (Faden and Simon, 1988; Nag and Riopelle, 1990; Hao et al., 1991a; Yezierski et al., 1993; Wrathall et al., 1994). For this reason glutamate has been viewed as one of several putative chemical mediators contributing to the ‘central cascade’ of secondary pathological changes following spinal injury (Tator and Fehlings, 1991). Three lines of evidence support the role of EAAs in the destructive cascade initiated by SCI: (a) following traumatic or ischemic SC1 there are significant increases in tissue content of glutamate and aspartate (Panter et al., 1990; Simpson et al, 1990; Marsala et al., 1994); (b) EAA receptor agonists exacerbate the neurodegenerative effects of spinal injury (Faden and Simon, 1988; Nag

and Riopelle, 1990); and (c) MK-801 and NBQX, selective NMDA and non-NMDA antagonists, respectively, provide significant neuroprotection following SC1 (Gomez-Pinilla et al., 1989; Hao et al., 1991b; Wrathall et al., 1994; Liu et al., 1997). Although EAA contents rise to toxic levels for only a brief period following injury, this dramatic change in tissue content of endogenous signaling molecules is thought to trigger an injury cascade that includes the production of inflammatory cytokines, prostanoids, as well as the up and down regulation of cellular messengers and transcription factors that can severely compromise the anatomical and functional integrity of spinal neurons. The excitotoxic model of spinal cord injury In recent years a number of experimental models have been used in the study of SC1 (Bunge et al., 1993b; Lighthall and Anderson, 1994; Christensen et al., 1996), each with distinctive characteristics related to specific aspects of the human condition. An important feature of these models is that each is based on a critical component of the primary injury (e.g., trauma or ischemia). Two approaches used to study spinal injury pain include the photochemical (Wiesenfeld-Hallin et al., 1994) and hemisection (Christensen et al., 1996) models. The weight drop or contusion model is the oldest and most widely used model of SCI, but has only recently been used in studies related to the altered sensation following injury (Siddall et al., 1995, 1999). A final approach involves the use of selected spinal lesions to study the central mechanisms of injury-induced pain (Levitt, 1989; Vierck, 1991; Vierck and Light, 2000, this volume). All of these experimental strategies share the distinction of producing pathological and/or behavioral changes associated with human SCI, and each provides unique opportunities to study different aspects of the spinal and supraspinal mechanisms responsible for central pain of spinal origin. Although the contusion and ischemic models of SC1 share many pathological characteristics with the human condition, the extent of tissue damage produced in these models makes it difficult to evaluate specific neural substrates responsible for the onset of injury-induced abnormal sensations. For this reason an alternative approach was developed (Yezierski et al., 1993). Out of consideration for the well

432

documented elevation of EAAs following SCI, the excitotoxic model was developed to simulate this chemical change and evaluate the involvement of non-NMDA receptors in the pathological sequela of SC1 (Yezierski et al., 1993; Liu et al., 1997). Extending the results of previous investigators (Pisharodi and Nauta, 1985; Nag and Riopelle, 1990; Urea and Urea, 1990) Yezierski and colleagues used the technique of intraspinal microinjection to evaluate the anatomical and functional consequences resulting from injection of the AMPA/metabotropic receptor agonist quisqualic acid (QUIS) (Yezierski et al., 1993; Yezierski and Park, 1993). Additionally, the precision of the intraspinal injection technique made it possible to evaluate the contribution of specific neuronal populations to different behavioral outcome measures. The results also showed that using different injection strategies (i.e., volume, depth), it was possible to produce graded patterns of neuronal loss in specific regions of the spinal gray matter (Yezierski and Park, 1993; Yezierski et al., 1998a). While the varied morphological changes following QUIS injections supported a role of non-NMDA mediated mechanisms in ‘secondary injury’, of special significance was the relationship between excitotoxic cell loss and the emergence of spontaneous and evoked pain behaviors commonly associated with models of chronic neuropathic pain (Levitt, 1985; Vierck, 1991; Zeltser and Seltzer, 1994; Siddall et al., 1995; Christensen et al., 1996). Although the pathological findings following

QUIS injections were believed to be initiated by excitotoxic events, the contributions of other components of the ‘central injury cascade’ cannot be ignored. For example, QUIS injections resulted in the activation of the nuclear factor-kappa B (NF-kB) family of transcription factors (Bethea et al., 1998), a critical step in the inducible regulation over 150 genes involved in inflammatory, proliferative and cell death responses of cells (Baeuerle and Baltimore, 1991; Kaltschmidt et al., 1993; Pahl, 1999). Examples of genes regulated by activation of NF-kB include those responsible for encoding tumor necrosis factor (TNF), enzymes for cyclooxygenase (COX-2), nitric oxide synthase (iNOS), and prostaglandin synthase-2, interleukins (IL) 6 and lb, dynorphin, and intercellular and vascular cell adhesion molecules (Baeuerle and Baltimore, 199 1; O’Neil and Kaltschmidt, 1997). Thus, the excitotoxic model originally developed to evaluate the pathological role of non-NMDA mediated events in spinal injury, led to the discovery of an interrelationship between excitotoxic and inflammatory components of injury, and a possible link with the regulation of genes that may play an important role in spinal sensory processing. Recently, a more detailed study related to the inflammatory component of the QUIS injury model has shown that following QUIS injections there is an upregulation of mRNA for cytokines (IL-lp, TNF-a), COX-2, iNOS as well as the death-inducing ligands TRAIL and CD-95 (Plunkett et al., 2000).

Fig. 2. The effects of quisqualic acid (QUIS) injections or sham surgery on responses to mechanical and thermal stimuli delivered to the hind paws. Animals were pre-tested over a period of 9-10 days prior to receiving intraspinal injections of 125 mM QUIS (A,C) or sham surgery (B,D). Animals received unilateral injections (1.2 ul) directed at the dorsal horn and intermediate gray (right side) in spinal segments ranging from T12 to L2. Post-injection testing commenced 8-10 days after surgery and continued for a period of 44 days. (A,B) Results of testing with mechanical stimuli delivered to the hind paws ipsilateral and contralateral to the side of QUIS injections (A) or following sham surgery (B). Each point on the graph represents the mean threshold for all animals on each day of testing. Error bars represent standard errors around the mean. Stimulus intensity in grams is represented on the y-axis and days pre- and post-injection/surgery on the x-axis. Statistical comparisons were made between the mean pre-injection baseline value and data obtained on each day of post-injection testing: * = P < 0.05. No significant differences were observed between pre- and post-surgery response thresholds in animals with sham surgery (B). (C) Results of the paw flick test assessing the sensitivity to a radiant heat stimulus delivered to the hind paws (same group of ten animals tested in A). During post-injection testing there was a significant decrease in withdrawal latencies to thermal stimulation (bilaterally). Each point on the graph represents the mean of three trials for all animals on each day of testing. Error bars represent standard errors around the mean. Time in seconds is represented on the y-axis and days on the x-axis. Statistical comparisons were made between the mean pre-injection baseline value and data obtained on each day of post-injection testing: * = P < 0.05. (D) Responses to the paw flick test following sham surgery for the same group of seven animals tested in (B). No significant differences were observed between pre- and post-surgery response latencies. (Reprinted with permission from Yezierski et al., 1998a.)

433

no post-injection paresis and/or paralysis of hind limbs). Initial responses evaluated over a two-week period during pre-injection testing were elicited by stimulus intensities of lo-35 g (mean baseline value 21.0 f 9.8 g). Following QUIS injections in spinal segments T12-L2 stimulus intensities required to elicit hind limb responses were significantly lower (1.5-8.0 g) than pre-injection values (Fig. 2A). The time course for the onset of mechanical allodynia

Behavioral consequences of excitotoxic spinal injury Mechanical hypersensitivity

The evaluation of animals to varying intensities of mechanical stimuli was carried out in animals meeting inclusion criteria for behavioral testing (i.e., no signs of early excessive grooming behavior and

MECHANICALTEST 40. 36.

3 +

Len RQhl

32. 28;

Days

THERMALTEST

I

434 was IO-12 days. A control group of seven animals undergoing sham surgeries had pre-surgery response thresholds similar to the pre-injection values of QUIS animals (17.9 f 4.4 g) and showed no significant change in stimulus intensity required to elicit responses throughout the post-surgery evaluation period (Fig. 2B). Animals evaluated for responses to mechanical stimuli were followed for a period of 34 days post-injection. The fact that significant effects (relative to baseline) were observed throughout this time period (with no signs of recovery) underscores the chronic nature of the behavioral effect. Responses to thermal stimuli Responses to the Hargreaves et al. (1988) thermal detection task were evaluated in the same animals undergoing mechanical testing. Differences in responses to thermal stimulation were found starting approximately IO-12 days post-QUIS injections and lasted throughout the evaluation period of 34 days. Pre-injection withdrawal latencies averaged 13.2 f 0.8 s while post-injection values were in the range of 8-l 1 s. As with mechanical stimulation, thermal stimuli were delivered to the glabrous skin of the hind paws in animals receiving QUIS injections in spinal segments Tl2-L2. Thus, responses reflecting a hypersensitivity to thermal stimulation were observed in dermatomes remote from those represented by segments receiving QUIS injections. Similar to the results with mechanical testing no preferential effects were observed between left and right hind paws (Fig. 2C). A control group of seven animals undergoing sham surgeries had pre-surgery response latencies similar to pre-injection values of QUIS animals (13.2 f 0.2 s) and showed no significant changes in response latencies throughout the post-surgery evaluation period (Fig. 2D). Efforts to correlate the pattern of neuronal loss with thermal and mechanical hypersensitivity proved unsuccessful; no such correlation was found with the onset of thermal hyperalgesia or mechanical allodynia. In fact, the results support the conclusion that animals with neuronal loss anywhere in the superficial or deep dorsal horn could be expected to exhibit changes, albeit not of the same magnitude, in thermal and mechanical sensitivity. Furthermore, it is of special significance that the behavioral responses to

mechanical and thermal stimuli were evaluated in the hind paws (following injections in spinal segments Tl2-L2). These results suggest that the effects of injury are distributed several segments (and bilaterally) from the site of injection. The fact that all animals exhibited bilateral changes in sensitivity (some without obvious signs of contralateral neuronal loss) indicates that the effects of unilateral injury are distributed along both sides of the cord. These results are not surprising given the complexities of propriospinal connections and the likely involvement of propriospinal circuits in the distribution of descending influences (see also Pertovaara, 2000, this volume) from bulbospinal pathways. Excessive grooming behavior Beginning on the second day post-QUIS injection animals were inspected daily for signs of excessive grooming (e.g., removal of hair, superficial skin damage). Based on results in over 100 animals this behavior typically targets dermatomes associated with spinal segments at or caudal to the site of QUIS injections (Yezierski et al., 1993, 1998a). Excessive grooming behavior is a progressive condition and therefore a classification scheme for different phases of this behavior was developed: (a) Class I, hair removal over contiguous portions of a dermatome; (b) Class II, extensive hair removal combined with signs of damage to the superficial layers of skin; (c) Class III, hair removal and damage to dermal layers of skin; and (d) Class IV, subcutaneous tissue damage (experiment terminated). Excessive grooming behavior was correlated with a lesion sparing the superficial laminae of the dorsal horn and is viewed as a variant of the well described ‘deafferentation autotomy’ (Yezierski et al., 1998a). Support for the conclusion that activity in the superficial dorsal horn may contribute to this behavior was found in animals with injections selectively eliminating different regions of the gray matter. Examples of neuronal loss consistent with the conclusion that the superficial dorsal horn ipsilateral to the site of grooming is important in the onset of excessive grooming behavior are shown in Fig. 3. Fig. 3A shows a cord where the superficial region and neck of the dorsal horn on the side of injection were eliminated and this animal did not

435

Fig. 3. Patterns of neuronal loss following intraspinal injection of 125 mM quisqualic acid (QUIS). All injections were made on the right side of the spinal cord. (A) Neuronal loss throughout the dorsal horn following injection of 0.6 pl of QUIS at a depth of 300 urn in spinal segment L2 (survival period 30 days). (B) G rooming-type damage represented by neuronal loss in the neck of the dorsal horn (arrows) following injection of 0.6 ~1 of QUIS at a depth of 900 urn in spinal segment T13 (survival period 32 days). (C) Neuronal loss throughout the dorsal horn (ipsilateral to injection) and in the neck of the dorsal horn (arrows) contralateral to injection site (L4) where 1.2 pl of QUIS was injected at depths of 600 urn and 1200 urn (survival period 18 days). The pattern of neuronal loss contralateral to injection represents grooming-type damage. (D) Bilateral grooming-type damage represented by neuronal loss in the dorsal horn partially sparing the superficial laminae (arrowheads) following injection of 1.2 ~1 of QUIS at depths of 600 urn and 1200 pm in spinal segment T13 (survival period 34 days). (E) Bilateral grooming-type damage represented by neuronal loss throughout the neck of the dorsal horn following injection of 1.2 ul of QUIS at depths of 600 urn and 1200 urn in spinal segment Ll (survival period 27 days). Note partial and complete sparing of the superficial laminae (arrowheads) ipsilateral and contralateral, respectively, to injection site. (F) Bilateral grooming-type damage represented by neuronal loss below the superficial laminae ipsilateral and contralateral following injection of 1.2 ul of QUIS at depths of 600 pm and 1200 urn in spinal segment L3 (survival period 28 days). Note sparing of superficial laminae (arrowheads) contralateral and ipsilateral to the site of injection. Scale bar in (E) equals 190 urn in (A-C), (E-F) and 320 urn in (D). (Reprinted with permission from Yezierski et al., 1998a.)

exhibit excessive grooming behavior (30 days survival). By contrast, the pattern of neuronal loss in Fig. 3B included the neck of the dorsal horn with sparing of the superficial region. This animal exhibited excessive grooming behavior ipsilateral to the side of neuronal loss. This pattern of neuronal loss is referred to as ‘grooming-type damage’. Results com-

parable to those in Fig. 3B were found in nearly 90% of animals with excessive grooming behavior. Also supporting our conclusion were animals with extensive neuronal loss throughout the superficial and deep laminae of the dorsal horn ipsilateral to injection sites and additionally had neuronal loss on the contralateral side of the cord restricted to the neck

436

of the dorsal horn (Fig. 3C). These animals exhibited excessive grooming behavior targeting skin regions contralateral to the injection site. Finally, animals exhibiting bilateral grooming behavior typically had bilateral grooming-type damage that included damage to the neck of the dorsal horn with partial or complete sparing of neurons in the superficial laminae (Fig. 3D-F). While the onset of self-directed behaviors have been described following lesions of the spinal cord (Frommer et al., 1977; Levitt, 1985), the results following QUIS injections were the first to have a dermatomal relationship that correlated with an injury site restricted to the gray matter of the spinal cord. Although the clinical relevance of autotomy has been controversial (Rodin and Kruger, 1984; Levitt, 1985; Coderre et al., 1986), Mailis (1996) described compulsive, self-injurious behavior (SIB) in humans with neuropathic pain. In this study it was concluded that “. . compulsive targeted self-injurious behavior in humans with neuropathic pain and painful dysethesiae is consistent with the concept that animal autotomy may result from chronic neuropathic pain after experimental peripheral or CNS lesions”. As discussed by Vierck (1991; see also Vierck and Light, 2000, this volume), however, the presence of self-directed behaviors (e.g., autotomy or excessive grooming behavior) does not necessarily support the conclusion that pain is the eliciting stimulus. These behaviors could be generated by paresthetic sensations or dysesthesias. Regardless of the eliciting stimulus when these behaviors are present, it is not unreasonable to conclude that they reflect the presence of an abnormal sensation. Without exception excessive grooming behavior targets peripheral dermatomes associated with spinal segments at or adjacent to the site of injury (Fig. 4). This distribution, in general, coincides with the dermatomal map described in the rat (Takahashi and Nakajima, 1996). Of special interest in QUIS-injected animals is the parallel between the delayed onset of excessive grooming and a similar temporal profile of central pain in patients with SCI. QUIS-induced excessive grooming is also similar to the behavioral agitation described by Yaksh (1989) following intrathecal administration of strychnine and bicuculline, where animals bite and scratch themselves after injections of these inhibitory amino

Fig. 4. Topographic distribution of areas targeted for excessive grooming behavior as a function of spinal segments injected with quisqualic acid (QUIS). Excessive grooming behavior is directed towards skin areas in dermatomes represented by spinal segments at or caudal to those receiving QUIS injections. The areas outlined on the line drawing summarize the location of all areas affected by excessive grooming behavior (Class IIV) following injections in segments shown in the inset. As injection sites move from rostra1 (TlO-Tll) to caudal (L3L4), the location of sites target for excessive grooming behavior move down the body from thoracic to hind limb dermatomes. (Reprinted with permission from Yezierski et al., 1998a.)

acid antagonists. In many respects the biting and scratching is also similar to that observed following intrathecal injections of substance P, somatostatin, or alumina gel (see references in Levitt, 1985). The similarity between QUIS-induced excessive grooming behavior and pain behaviors following intrathecal injection of these substances suggests that the central mechanism responsible for these behaviors involves disruption of local inhibitory pathways and/or the emergence of abnormal ‘focal generators’ within the injured cord (see below). In conclusion there are three important similarities between excessive grooming behavior and the well documented clinical condition of junctional pain in patients with spinal injury: (a) delayed onset; (b) spontaneous nature; and (c) dermatomal distribution relative to site of injury. The delayed onset of excessive grooming behavior suggests that the neural mechanism is not simply an inhibitory release phenomenon, but instead requires significant changes (over time) in the functional state of spinal (and possibly supraspinal) sensory neurons. It is hypothesized that this behavior may be due to a loss of spinal nociceptive neurons, thus creating an imbalance between normal gating and biasing mechanisms within spinal and supraspinal somatosensory pathways (Melzack

437 and Loeser, 1978). Combined with a loss of segmental and/or supraspinal inhibitory influences, spinal neurons become hyperactive and these focal pattern generators are responsible for producing paraesthetic and/or dysesthetic sensations referred to the affected dermatome. Functional correlate of behavioral changes following excitotoxic spinal cord injury Evaluation of a possible neural correlate of QUISinduced behavioral changes was undertaken by examining the functional properties of dorsal horn neurons, including cells belonging to the spinomesencephalic tract (SMT). Supportive of a spinal mechanism for the evoked and spontaneous behavioral changes following QUIS injections was the finding that spinal neurons adjacent to the injury site undergo significant functional changes, including a shift to the left in the stimulus-response function, an increase in the level of background activity, and an increase in the duration of afterdischarge responses following removal of a stimulus (Yezierski and Park, 1993). The fact that these changes were observed in cells belonging to the SMT supported the notion of a possible surpraspinal component to the observed pain behaviors, the fact that cells had peripheral receptive fields overlapping with areas targeted for excessive skin grooming supports the contention that this behavior is not related to an insensate area/region of skin. Changes in the response characteristics of spinal neurons similar to those observed following QUIS injuries have been reported for cells following ischemic (Hao et al., 1992) and hemisection (Hulsebosch et al., 1997) injury of the spinal cord. In the excitotoxic model these changes appeared within 47 days of injury and were especially prevalent in animals with excessive grooming behavior. Afterdischarges lasting 5-15 min and ‘wind-up’ of background discharges with repeated stimulation were novel characteristics of neurons in QUIS-injected animals. The increased excitability, bursting discharges, and long afterdischarge responses of neurons in QUIS-injected animals are reminiscent of the abnormal functional characteristics of neurons recorded in patients with chronic pain following SC1 (Loeser et al., 1968; Edgar et al., 1994). Elimination

of this activity by computer-assisted DREZ (dorsal root entry zone) results in a significant reduction of spontaneous pain. Further supporting the hypothesis of a spinal pain generator are results showing that local anesthetic applied to the proximal stump of a spinal transection results in transient relief of pain (Pollock et al., 1951). Spatial profile of spinal cord damage required for the onset of pain behaviors following excitotoxic spinal injury In initial studies evaluating the behavioral consequences of QUIS-induced spinal injury little attention was given to the longitudinal extent of neuronal loss required to produce these behaviors. While it is acknowledged that the cellular and molecular events accompanying injury are undoubtedly important in inducing pain behaviors, we asked the question if the longitudinal extent over which neurons are affected by the injury process is also important. To this end we carried out an analysis of the number of sections with grooming-type damage (as determined from counts of cresyl violet stained sections) in animals with and without excessive grooming behavior. The results suggested that there exists a critical extent of damage along the rostrocaudal axis of the cord which is required to precipitate the onset of this spontaneous pain behavior. The evaluation of Long Evans (LE) (n = 19) and Sprague-Dawley (SD) (n = 69) rats consistently showed that there was a significant difference in the extent of groomingtype damage between grooming and non-grooming animals. In animals with injury-induced grooming behavior the longitudinal extent of grooming-type damage was 5175 urn (SD) and 5400 urn (LE). By contrast, damage in non-grooming animals was 3375 urn (SD) and 3750 urn (LE). Based on these data a hypothetical model of tissue damage versus pain behavior is proposed (Fig. 5). In this model there is a gradual progression of cord damage (including the influence of cellular and molecular changes) away from the injury epicenter. As the injury evolves the extent of tissue damage approaches threshold, triggering the onset of pain behaviors. This model implies that it is not only the injury cascade per se that is responsible for the onset of central pain following spinal injury. but the longitudinal extent over

THRESHOLD FOR PAIN BEHAVIOR

Fig. 5. Hypothetical progression of the central injury cascade from the epicenter (EPZ) of an ischemic, traumatic or excitotoxic insult to the spinal cord. In this model the extent of injury and/or the area of cord influenced by different components of the injury cascade expands to include 2” and 3” areas of injury. - . If left untreated the amount of cord damage will continue to expand until it exceeds the threshold required for the onset of pain behavior.

which these events influence neurons is also important. Although a fundamentally simple concept, these results suggest that limiting the amount of neuronal damage either by neuroprotective or rescuing strategies could have a beneficial effect as an intervention in the treatment of spinal cord injury pain. To test the above hypothesis rats were injected with QUIS and simultaneously administered intraperitoneal injections of the NMDA antagonist and NOS inhibitor agmatine. Agmatine is a cationic amine formed by the decarboxylation of arginine. It interacts with various neurotransmitter receptors including nicotine, NMDA, alpha2-adrenergic, and imidazoline. The fact that agmatine has biological activity as an antagonist of the NMDA receptor and inhibitor of NOS led to the proposal that it meets many criteria for a neurotransmitter-neuromodulator (Reis and Regunathan, 1998). In previous studies agmatine was shown to be neuroprotective following ischemic injury in the brain and following traumatic and excitotoxic injury in the spinal cord (Gilad et al., 1996; Yezierski et al., 1998b; Yu et al., 2000). Additionally, agmatine has beneficial effects on pain behaviors in the dynorphin model of chronic allodynia, Chung model of spinal nerve ligation, carrageenan-evoked muscle hyperalgesia and the CFA model of hyperal-

gesia (Fairbanks et al., 2000). We have shown that agmatine administered at the time of QUIS injury delays or prevents the onset of excessive grooming behavior (Fig. 6).The results of a 16day treatment with agmatine (100 mg/kg, i.p.) showed that over a survival period of 40 days the final area of skin involvement targeted for excessive grooming was significantly reduced (Fig. 6A), the onset time of spontaneous grooming was significantly increased (Fig. 6B), the severity of grooming behavior was significantly lower (Fig. 6C), and the longitudinal extent of grooming-type damage was significantly less than in animals treated with saline (Fig. 6D). Neuroprotective effects similar to those with agmatine have also been obtained with the potent antiinflammatory IL-10 (Brewer et al., 1999). In this study a one time systemic injection of 5 pg of IL-10 significantly reduced neuronal loss in the excitotoxic model of injury. These results are consistent with a study showing the neuroprotective and behavioral effects of IL-10 following traumatic spinal cord injury (Bethea et al., 1999a). To determine if there was a behavioral correlate to the neuroprotective effects of IL-10 in the QUIS model we tested IL-10 against QUIS-induced excessive grooming behavior. The results of this evaluation showed that IL-10 significantly increased the onset time and reduced

439

FINAL

AREA

OF SKIN

DAMAGE

ONSET

OF GROOMING

C

GROOMING

GROOMING DAMAGE IN THE DORSAL HORN

SEVERITY

A: Q-Groom+Ag C: QUIS+Ag

Ll

(n=8) (n=7)

B: Q-Groom+Saline D: QUIS+Saline

(n=8) (n=8)

Fig. 6. Effects of agmatine treatment on excessive grooming behavior following excitotoxic spinal cord injury. (A) Final area of skin damage caused by excessive grooming behavior. (B) Onset time for the start of excessive grooming behavior following the intraspinal injection of quisqualic acid (QUIS). (C) Final classification of excessive grooming behavior, i.e. severity of grooming, following intraspinal injection of QUIS. (D) Longitudinal extent of grooming-type damage in the spinal gray matter following the intraspinal injection of QUIS. The type of treatment and number of animals in each group is indicated in the inset at the bottom of the figure. (A) QUIS-induced grooming plus agmatine treatment (100 mg/kg, i.p., 14 days). (B) QUIS-induced grooming plus saline treatment. (C) QUIS injection plus agmatine treatment (100 mg/kg, i.p, 14 days). (D) QUIS injection plus saline treatment (see text for details). ** = P < 0.01 (A) versus (B); ### = P < 0.01 (C) versus (D).

the area of excessive grooming behavior following QUIS injections (Bethea et al., 1999b). The effect of IL-10 on the QUIS-induced pain behavior is consistent with the effects of IL-10 on dynorphin-induced allodynia (Laughlin et al., 1999). The fact that agmatine and IL-10 can be used effectively as preventive treatments for injury-induced pain behaviors underscores the potential of these therapeutic interventions as preemptive strategies of pain management (see also Jensen and Nikolajsen, 2000, this volume) for

patients predisposed to progressive tissue damage in the spinal cord (e.g., syringomyelia, spinal cord injury). The above studies with agmatine and IL-10 lend support to the ‘neuroprotective hypothesis’ of SC1 pain, but of greater importance and potential clinical relevance, agmatine is also an effective treatment for the progression of injury-induced excessive grooming behavior (Fig. 6A,C,D). Similar results have also been obtained with IL-10 (data not shown). These

440

results further underscore the importance of inflammatory mediators, NMDA receptors and nitric oxide in influencing the functional state of spinal neurons. Agmatine (100 mg/kg, i.p.) administered at the time of grooming onset significantly reduced the final area of skin damage, the severity of grooming, and the extent of grooming-type damage in the spinal cord. These results are similar to those obtained with transplantation of adrenal medullary tissue which was also found to reverse the progression of excessive grooming behavior (Brewer and Yezierski, 1998). Adrenal chromaffin cells are known to produce a wide array of potentially analgesic agents including the NMDA antagonist histogranin and neurotrophic and growth factors including FGF-2 and TGF-l3 (Sagen, 1996; Brewer and Yezierski, 1998). These results are encouraging as they suggest that it is possible to ‘turn down’ an ongoing pathological process over a critical length of the cord and retard or reverse an ongoing pain behavior. The exact nature of this ‘process’ and the precise therapeutic target(s) remain the focus of future investigation.

In addition to the above results obtained with the excitotoxic model of spinal injury a number of other mechanisms have been proposed to explain the onset of pain following injury. A brief review of these mechanisms is presented below. Over the past 40 years a number of mechanisms have been proposed to explain the condition of central pain following SCI: (a) loss of balance between different sensory channels (Beric et al., 1988); (b) loss of spinal inhibitory mechanisms (Melzack and Loeser, 1978; Wiesenfeld-Hallin et al., 1994); (c) the presence of pattern generators within the injured cord (Pollock et al., 195 1; Melzack and Loeser, 1978; Yezierski and Park, 1993) or supraspinal relay nuclei (Lenz et al., 1991); and (d) synaptic plasticity (see also Gerber et al., 2000, this volume; Moore et al., 2000, this volume; Sandktihler et al., 2000, this volume).

classic pain transmission system in the anterolateral quadrant of the cord can play a significant role in the onset of dysesthesia pain. This paradox has been discussed previously (Kendall, 1949; Boivie, 1992; Pagni, 1998). In the studies by Beric and colleagues, preservation of the modalities of touch and vibration (i.e., dorsal column function) in the absence of pain and temperature sensibilities (i.e., spinothalamic function), combined with evoked potential studies consistent with these findings, were common in SC1 patients with dysesthetic pain syndrome (Beric, 1988, 1992). As a result it was proposed that an imbalance in sensory information conveyed by the dorsal column medial lemniscal (DCML) and anterolateral systems (ALS) has an important role in post-traumatic central pain. The paradox of this hypothesis lies in the fact that disruption of pathways in the anterolateral quadrant of the cord has historically been used to eliminate chronic pain (Pagni, 1998). Dysesthetic sensations, however, may occur following spinothalamic tractotomies directed at the spinal or mesencephalic trajectories of these pathways (Pagni, 1998). In fact, severe spinal lesions with total destruction of ascending sensory systems are usually not followed by pain syndromes, but mild, moderate or severe disruption of the ALS with partial or complete sparing of the DCML pathway is most frequently associated with central pain. Additional support for the ‘imbalance hypothesis’ comes from reports of stroke patients with pain or dysesthesia. These patients invariably have an absence of anterolateral sensibilities coupled with preserved dorsal column function (Boivie and Leijon, 1991). Pagni (1998) concluded that lesions of the spinothalamic system intended to relieve pain regardless of level may sometimes produce new pains or aggravate existing pains. In view of these clinical reports it was proposed that dysesthesias following SC1 result from the central misinterpretation of residual dorsal column input which functions in the absence of suppression via an integrated spinothalamic tract system (Beric et al., 1988; Tasker et al., 1991).

Imbalance of sensory pathways

Loss of inhibitory tone

Perhaps the most intriguing of the above hypotheses evolved from the observation that damage to the

A critical factor in the onset of pain-like behaviors following SC1 is believed to be a loss of inhibitory

Mechanism(s) of pain following spinal injury

441

tone within the injured spinal cord (WiesenfeldHallin et al., 1994). The loss of spinal inhibitory control would allow for the recruitment of surrounding neurons and the intensification and spread of abnormal sensations, including pain. Support for this conclusion comes from four lines of evidence: (a) following ischemic injury there is an increased excitability of WDR neurons that is reversed by the GABAs receptor agonist baclofen (Hao et al., 1992b); (b) hypersensitivity to peripheral stimuli can be produced in normal animals by the intrathecal administration of GABA receptor antagonists (Yaksh, 1989; Hao et al., 1994); (c) there are decreased numbers of GABA-positive neurons following ischemic spinal cord injury (Zhang et al., 1999); and (d) administration of drugs that prolong the action of inhibitory neurotransmitters (i.e., cyclic antidepressants) are effective in the short-term treatment of spinal injury pain (Tunks, 1986; Leijon and Boivie, 1991; Boivie, 1994). Although existing evidence supports the decreased inhibitory influence of GABAergic neurotransmission in altering the functional properties of neurons in the injured cord, not to be overlooked in this process is a decreased influence of supraspinal and propriospinal inhibitory pathways that may be disrupted following injury. An important response (in brain and spinal cord) following ischemia and trauma is an increase in tissue content of glutamate. Given the well documented involvement of NMDA receptors in altering the excitability of spinal neurons (Haley and Wilcox, 1992; Woolf, 1992) it is reasonable to propose, along with a failed GABAergic inhibitory system, that increased NMDA receptor activation (secondary to injury-induced release of glutamate) could play a role in the cascade of physiological and behavioral changes following spinal injury. Blockade of acute allodynia by the competitive NMDA receptor antagonist MK-801 provides support for this hypothesis (Hao et al., 1991a). Relevant to this discussion is the clinical report of Eide et al. (1995) showing that central dysesthesia pain after traumatic spinal cord injury is dependent on NMDA receptor activation. Pattern generators of pain Another component of the pathophysiological sequela of SC1 which is thought to contribute to the

onset of altered sensations is the emergence of a ‘pattern generating mechanism’ (spinal and supraspinal). Evidence supporting this hypothesis led Melzack and Loeser (1978) to conclude that not all postinjury pains are due to noxious input; some may be due to deafferentation and/or loss of inhibitory control and subsequent changes in the firing patterns, including burst activity and long afterdischarges of neuronal pools lying adjacent to a site of injury. Loss of inhibitory control (spinal and supraspinal) is an important component of this model, allowing for the recruitment of surrounding neurons and the intensification and spread of pain. Several observations are consistent with the ‘pattern generating’ hypothesis: (a) the existence of abnormal focal hyperactivity in the spinal cord and thalamus of spinal injured patients (Loeser et al., 1968; Lenz, 1991; Edgar et al., 1994); (b) the effectiveness of spinal anesthesia in alleviating pain when delivered proximal to the site of injury (Pollock et al., 1951; Botterell et al., 1953); and (c) abnormal responses and prolonged afterdischarges of spinal sensory neurons following experimental SC1 (Hao et al., 1992a; Yezierski and Park, 1993; Christensen and Hulsebosch, 1997). Consistent with the excitability hypothesis and the involvement of neurons in the superficial dorsal horn in spontaneous and evoked pain behaviors, are clinical reports of abnormal focal hyperactivity within the superficial laminae of the injured cord (Edgar et al., 1994). In this report it was concluded that the origin of post-traumatic central pain is in the dorsal root entry zone (DREZ) at depths of l-2 mm, representing Rexed’s laminae I-II. Microcoagulation of these hyperactive areas results in a significant diminution of pain. A variant of the DREZ technique utilizing intramedullary recordings of C-fiber evoked responses to guide DREZ lesioning was recently described (Falci et al., 1999). This technique allows for the somatotopic mapping of the DREZ with regard to the generation of central deafferentation pain. In 25 patients where this technique was used, 84% received complete pain relief. Local anesthetics (which shut down local ectopic activity) delivered to the spinal cords of SC1 patients with dysesthetic pain syndrome also result in the temporary relief of pain (Pollock et al., 1951; Botterell et al., 1953). Two recent reports are consistent with the involvement of the superficial laminae in the persistence of neuro-

442

pathic pain (Malmberg et al., 1997; Mantyh et al., 1997). Thus, the superficial dorsal horn may be important not only in the onset but also the progression of neuropathic pain following SCI. Although changes in the excitability and response properties of spinal sensory neurons are thought to be important in the central response to XI, another aspect of the central injury cascade that should not be ignored is the involvement of supraspinal structures (e.g., diencephalon). The contribution of abnormal spinal input together with the effects of deafferentation (secondary to loss of spinal projection neurons) could result in the development of abnormal generators at supraspinal sites (Lenz, 1991; Vierck, 1991; see also Lenz et al., 2000, this volume). Hyperactive neuronal activity in thalamus and cortex of rats exhibiting extensive self-directed behavior following dorsal rhizotomy (Lombard et al., 1979), and the onset of autotomy following injections of epileptogenic agents into the mesencephalic central gray matter in cats (Black, 1974), provide additional support for the view that changes in the excitability of supraspinal neurons play a role in the onset of pain behaviors following peripheral and/or central injury. Recently, the report of elevated blood flow in thalamic nuclei (possibly reflecting changes in the functional state of neurons) lend further support to a thalamic involvement in the supraspinal response to SC1 and to the mechanism of injury induced pain (Morrow et al., 1999). Synaptic plasticity A final consideration for the mechanism of chronic pain following spinal injury is synaptic plasticity. A significant response to injury of central or peripheral origin is the cellular response believed to contribute to the sensitization of spinal neurons. Synaptic plasticity in the CNS is clearly a part of the central sensitization underlying acute pain, and together with the long-term changes in spinal connectivity represent a potential mechanism for persistent pain (see also Moore et al., 2000, this volume). Unfortunately, we are only at the beginning of understanding the biochemical, molecular and functional events underlying long-term plasticity and its role in the clinical condition of chronic pain. Included in the events thought to be involved in producing long-term changes are:

(a) phosphorylation of regulatory proteins (as well as dephosphorylation of others) (see Malmberg, 2000, this volume); (b) positive and negative regulation of gene transcription (see Berthele et al., 2000, this volume); (c) induced synthesis of proteins (as well as reduced synthesis of others); (d) strengthening of some and weakening of other connections (synaptic longterm potentiation, sprouting and pruning) (see Gerber et al., 2000, this volume; Sandkiihler et al., 2000, this volume); and (e) the death or rescuing of neurons. These neuronal changes share initiating events such as glutamate-induced elevations in calcium, maintaining events (nitric oxide synthesis, activation of protein kinases), perpetuating forces (activation of CREB, NP-kB), and terminal results (altered synaptic efficacy). Two critical initiating events in the process of injury-induced synaptic plasticity are an increase in intracellular calcium followed by the activation of calcium-dependent enzymes, including members of the calpain family of neutral proteinases believed to participate in many intracellular processes such as turnover of cytoskeletal proteins, and the regulation of kinase activities and transcription factors. Calpain activation has been linked to the breakdown of microtubule-associated protein (MAP) along with the activation of enzymes triggering programmed cell death (Suzuki et al., 1987). Cytoskeletal proteins, including MAP, are involved in maintaining structural integrity, which is essential for normal cellular function and survival of adult neurons. Degradation of these skeletal elements may contribute to neuronal dysfunction after CNS injury (Siman et al., 1985; Blomgren et al., 1995; Banik et al., 1997). Activation of calpains after CNS injury is therefore thought to be an important factor in determining the extent of cellular dysfunction and likelihood of cell death (Springer et al., 1997; Happel et al., 1981). Consistent with this hypothesis, administration of calpain inhibitors provide significant neuroprotection associated with glutamate excitotoxicity (Caner et al., 1994). These studies underscore the importance of glutamate in excitotoxic neuronal damage and the potential contribution to the establishment of neuronal vulnerability and dysfunction. Evidence supporting the hypothesis that cytoskeletal degradation is extremely sensitive to glutamate- and calcium-mediated excitotoxic events have shown MAP,

443

spectrin, and neurofilament proteins undergoing proteolysis after CNS injury involving excitotoxic insult (Banik et al., 1997). Of relevance to the discussion of spinal injury, SC1 results in the rapid degradation of MAP-2 and spectrin (Springer et al., 1997) a process known to be initiated by excessive glutamate receptor activation (Siman et al., 1985). The proposed relationship between degradation of membrane cytoskeletal proteins and central sensitization stems from a body of evidence suggesting that high-frequency afferent stimulation produces postsynaptic changes mediated by glutamate or aspartate (or both). For example, following high-frequency stimulation glutamate binding increases and this correlates with the induction of LTP in hippocampal slice preparations (Lynch and Baudry, 1984). Electron microscopic studies from rats with LTP reveal a rounding of dendritic spines and an increase in the number of synapses (Lee et al., 1980). These changes are not seen in preparations lacking LTP. The relationship between calcium and increased glutamate binding sites follows from the observation that calcium induced a substantial increase in the maximal number, but not the affinity, of glutamate binding in sites in the hippocampus, striatum and cortex (Lynch and Baudry, 1984). Like LTP, ‘kindling’ also produces an increase in glutamate binding sites (Savage et al., 1982). From these results it was concluded that the postsynaptic face of neuronal connections is quite plastic and can be substantially changed by physiological activity. The above studies led to the hypothesis that calcium activates a membrane-associated calpain which breaks up a portion of the cytoskeletal network, producing structural and chemical changes in the region of the postsynaptic membrane (Baudry et al., 1981; Lynch and Baudry, 1984). These events result in previously occluded glutamate receptors being exposed, thereby increasing the size of the postsynaptic response to released transmitter. Continued high bursts of activity or injury increase the availability of glutamate leading to a larger calcium influx and more widespread activation of calcium-dependent proteinases. These events have relevance to the altered functional properties of neurons following SC1 since intracellular calcium and calpain activity increase significantly after SC1 (Li et al., 199.5; Banik et al., 1997).

Other biochemical events potentially contributing to the initiation of injury-induced plasticity include: (a) the calcium activation of phospholipase A2; (b) hydrolysis of phospholipid precursors to arachidonate; and (c) the release of the cis-unsaturated fatty acid oleate which directly stimulates PKC (Linden et al., 1987).PKC activation has recently been shown to evoke mechanical allodynia and thermal hyperalgesia (Palecek et al., 1999) providing further evidence along with that of Malmberg et al. (1997; see Malmberg, 2000, this volume) that PKC is involved in the modulation of nociceptive information in the spinal cord and the onset of neuropathic pain. The production of arachidonic acid potentially exacerbates the injury process by increasing extracellular levels of aspartate and glutamate by inhibiting sodium-dependent uptake (Breukel et al., 1997) and by stimulating exocytosis of glutamate in synergy with PKC activation. Activation of the arachidonic acid cascade also leads to synthesis of eicosanoids which regulate neuronal ion channels and the formation of superoxide free radicals (Piomelli, 1994). In parallel with the above events are changes in the local expression of cytokines, chemokines and adhesion molecules (Hsu et al., 1994). The role of the inflammatory cytokine IL- 1b along with protein synthesis, and activation of NF-kB in synaptic plasticity and chronic pain behaviors in the dynorphin model of neuropathic pain was recently described by Wilcox and colleagues (Laughlin et al., 1999). The increased synthesis and release of inflammatory mediators also triggers the rapid induction of COX-2 and oxygenation of arachidonic acid to prostaglandin endoperoxide H2, which is converted to biologically active end-products by individual synthases or reductases. Simultaneously, since iNOS is coexpressed with COX-2 (Salvemini et al., 1994; Nogawa et al., 1998) and because COX-2 is a heme-containing enzyme, its enzymatic activity is potentiated by nitric oxide (NO), a gas with high affinity for heme iron (Ignarro, 1991). Therefore COX-2 induction could be another mechanism by which NO exerts its pathogenic effects following CNS injury; NO produced by iNOS has been shown to activate COX-2 and increase the output of proinflammatory prostaglandins (Salvemini et al., 1994; see, however, Hoheisel and Mense, 2000, this volume). A final participant in the cellular cascade of synaptic

444

plasticity is the high lipid content of CNS membranes, the integrity of which is a prerequisite for normal brain function. The effects of traumatic or ischemic injury leads to oxygen deprivation and compromised lipid metabolism including a decomposition of membrane-bound phospholipids and the release of free fatty acids (Rehncrona et al., 1982). Changes in free fatty acids lead to mitochondrial dysfunction (Lazarewicz et al., 1988), edema (Chan and Fishman, 1978) and production of leukotrienes along the lipoxygenase pathway, some of which are potentially toxic. Conclusions From the above discussion it is clear that there are many possible scenarios by which the biochemical cascade initiated by injury can influence not only the progression of neuronal damage, but also changes in the physiological properties responsible for the response of neurons to peripheral stimuli. The present discussion addressed three important components initiated by injury, each of which potentially contributes to the anatomical and functional plasticity of spinal neurons. Firstly, the influx and/or mobilization of calcium is responsible for the activation of calpain and phospholipase A2. The downstream effects of activating these calcium-dependent enzymes on cytoskeletal proteins, glutamate binding sites, release of glutamate, ion channels, production of free radicals and PKC activation have been associated with events ranging from the strengthening of synaptic efficacy to compromised neuronal function and cell death. Secondly, expression of cytokines, chemokines and adhesion molecules leads to the induction of COX-2 and iNOS and subsequent production of prostanoids and NO, two cellular messengers thought to play an important role in CNS inflammation and sensory processing (Haley and Wilcox, 1992; Meller and Gebhart, 1994). Thirdly, injury-induced release of free fatty acids increases phospholipase A2 activity, compromised mitochondrial function and potential activation of cell death programs. Finally, it is unlikely that any one mechanism discussed in this chapter is solely responsible for the onset of central pain following SCI. Depending on the nature of injury and the progression of

pathological and biochemical changes along the rostrocaudal axis of the cord, it is probable that each of the proposed mechanisms contributes to the onset of this condition. Continued research directed towards specific components of the spinal injury cascade should provide a better understanding of spinal and supraspinal mechanism(s) responsible for this condition and the future development of novel therapeutic strategies. Acknowledgements The author would like to thank the dedicated collaborative assistance of Drs. Shanliang Liu, Kori Brewer, Chen-Guang Yu, Jeffery Plunkett and Laurel Gorman. Expert technical assistance was provided by Gladys Ruenes and Dimarys Sanchez. The work of the author was supported by the Hollfellder Foundation, Paralyzed Veterans of America, Department of Defense, U.S. Army and by funds from the Miami Project and State of Florida Spinal Cord Injury Trust. References Baeuerle, PA. and Baltimore, D. (1991) The physiology of NFkB transcription factor. In: P Cohen and J.G. Foulkes (Eds.), Molecular Aspects of Cellular Regulation-Hormonal Contml of Gene Transcription. Elsevier, Amsterdam, pp. 409432. Banik, N.L., Matzelle, D.C., Wilford, G.G., Osborne, A. and Hogan, L. (1997) Increased calpain content and progressive degradation of nemofilament protein in spinal cord injury. Brain Rex, 752: 301-306. Baudry, M.. Smith, E. and Lynch, G. (1981) Influences of temperature, detergents, and enzymes on glutamate receptor binding and its regulation by calcium in rat hippocampal membranes. Mol. Phamcol., 20: 280-286. Beric, A. (1990) Altered sensation and pain in spinal cord injury. In: M.R. Dimitrijevic, P.D. Wall and U. Lindblom (Eds.), Recent Achievements in Restorative Neurology. Karger, Base], pp. 27-36. Beric, A. (1992) Pain in spinal cord injury. In: L.S. Illis (Ed.), Spinal Cord Dysfunction. Oxford University Press, New York, pp. 156-165. Beric, A., Dimitrijevic, M. and Lindblom, U. (1988) Central dysesthesia syndrome in SC1 patients. Pain, 34: 109-l 16. Berthele, A., Schadrack, J., Castro-Lopes, J.M., Conrad, B., Zieglgansberger, W. and Tolle, T.R. (2000) Neuroplasticity in the spinal cord of monoarthritic rats: from metabolic changes to the detection of interleukin-6 using mRNA differential display. In: J. Sandkiihler, B. Bromm and G.F. Gebhart (Eds.),

445 Nervous System Plasticity and Chronic Pain. Progress in Brain Research, Vol. 129. Elsevier, Amsterdam, pp. 191-203. Bethea, J.R., Castro, M., Lee, T.T., Dietrich, W.D. and Yezierski, R.P. (1998) Traumatic spinal cord injury induces nuclear factor kappa B activation. J. Neurosci., 18: 3251-3260. Bethea, J.R., Nagashima, H., Acousta, MC., Briceno, C., Gomez, E, Marcillo, A., Loor, K., Green, J. and Dietrich, W.D. (1999a) Systemically administered interleukin-10 (IL-lo) reduces tumor necrosis factor-alpha production and significantly improves functional recovery following traumatic spinal cord injury in rats. J. Neurotrauma, 16: 851-863. Bethea, J.R., Yu, C.-G., Plunkett, J.A. and Yezierski, R.P. (1999b) Effects of interleukin-10 (IL-lo) on pain behaviors following excitotoxic spinal cord injury. Sot. Neurosci. Abstr, 25: 1443. Black, R.G. (1974) A laboratory model of trigeminal neuralgia. In: International Symposium on Pain. Advances in Neurology. Raven Press, New York. Blomgren, K., Kasashima, S., Saido, T.C., Karlsson, J.-O., Elmered, A. and Hagberg, H. (1995) Fodrin degradation and subcellular distribution of calpains after neonatal rat cerebral hypoxic-ischemia. Brain Res., 684: 143-149. Boivie, J. (1992) Hyperalgesia and allodynia in patients with CNS lesions. In: W.D. Willis (Ed.), Hyperalgesia and Allodynia. Raven Press, New York, pp. 363-373. Boivie, J. (1994) Central pain. In: PD. Wall and R. Melzack (Eds.), Textbook of Pain (4rd ed.). Churchill Livingston, New York, pp. 871-902. Boivie, J. and Leijon, G. (1991) Clinical findings in patients with central post stroke pain. In: K.L. Casey (Ed.) Pain and Central Nervous System Disease: The Central Pain Syndromes. Raven Press, New York, pp. 65-75. Bonica, J.J. (1991) Semantic, epidemiologic and educational issues of central pain. In: K.L. Casey (Ed.), Pain and Central Nervous System Disease: The Central Pain Syndromes. Raven Press, New York, pp. 13-29. Botterell, E.H., Callaghan, H.C. and Jousse, A.T. (1953) Pain in paraplegia: clinical management and surgical management. Proc. R. Sot. Med., 47: 281-288. Breukel, A.I.M., Besselsem, E., Lopes da Silva, F.H. and Ghijsen, W.E.J.M. (1997) Arachidonic acid inhibits uptake of amino acids and potentiates PKC effects on glutamate, but not GABA, exocytosis in isolated hippocampal nerve terminals. Brain Res., 773: 90-97. Brewer, K.L. and Yezierski, R.P. (1998) Effects of adrenal medullary transplants on pain-related behaviors following excitotoxic spinal cord injury. Brain Res., 798: 83-92. Brewer, K., Yezierski, RI? and Bethea, J.R. (1997) Excitotoxic spinal cord injury induces diencephalic changes in gene expression. Sot. Neurosci. Abstr, 23: 438. Brewer, K.L., Bethea, J.R. and Yezierski, R.P (1999) Neuroprotective effects of interleukin-10 following excitotoxic spinal cord injury. Exp. Neurol., 159: 484-493. Britell, C.W. and Mariano, A.J. (1991) Chronic pain in spinal cord injury. Phys. Med. Rehab., 5: 71-82. Bunge, M.B., Holets, V.R., Bates, M.L., Clarke, T.S. and Watson, B.D. (1993a) Characterization of photochemically in-

duced spinal cord injury in the rat by light and electron microscopy. Exp. Neural., 127: 76-93. Bunge, R.P., Puckett, W.R., Becerra, J.L., Marcillo, A. and Quencer, R.M. (1993b) Observations on the pathology of human spinal cord injury. A review and classification of 22 new cases with details from a case of chronic cord compression with extensive focal demyelination. Adv. Neurol., 59: 75-89. Caner, H., Collins, J.L., Harris, S.M., Kassell, N.K. and Lee, KS. (1994) Attenuation of AMPA-induced neurotoxicity by a calpain inhibitor. Brain Res., 607: 354-356. Chan, PH. and Fishman, R.A. (1978) Brain edema: induction in cortical slices by polyunsaturated fatty acids. Science, 201: 358-360. Christensen, M.D. and Hulsebosch, C.E. (1997) Chronic pain after spinal cord injury. J. Neurotrauma, 14: 517-543. Christensen, M.D., Everhart, A.W., Pickeman, J. and Hulsebosch, C.E. (1996) Mechanical and thermal allodynia in chronic central pain following spinal cord injury. Pain, 68: 97-107. Coderre, T.J., Grimes, R.W. and Melzack, R. (1986) Deafferentation and chronic pain in animals: an evaluation of evidence suggesting autotomy is related to pain. Pain, 26: 61-84. Davidoff, G. and Roth, E. (1991) Clinical characteristics of central (dysesthetic) pain in spinal cord injury patients. In: K.L. Casey (Ed.), Pain and Central Nervous System Disease: The Central Pain Syndromes. Raven Press, New York, pp. 77-83. Davidoff, G., Roth, E., Guarracini, M., Sliwa, J. and Yarkony, G. (1987) Function limiting dysesthetic pain syndrome among traumatic SC1 patients: a cross sectional study. Pain, 29: 3948. Dubner, R. (1991) Neuronal plasticity and pain following peripheral tissue inflammation or nerve injury. In: M.R. Bond, J.E. Charlton and C.J. Woolf (Eds.), Proceedings of the VIth World Congress on Pain. Elsevier, Amsterdam, pp. 263-276. Edgar, R.E., Best, L.G., Quail, P.A. and Obert, A.D. (1994) Computer-assisted DREZ microcoagulation: posttraumatic spinal deafferentation pain. J. Spinal Disord., 6: 48-56. Eide, P.K. (1998) Pathophysiological mechanisms of central neuropathic pain after spinal cord injury. Spinal Cord, 36: 601612. Eide, P.K., Stubhaug, A. and Stenehjem, A.E. (1995) Central dysesthesia pain after traumatic spinal cord injury is dependent on N-methyl-D-aspartate receptor activation. Neurosurgery, 37: 1080-1087. Faden, AI. and Simon, R.P. (1988) A potential role for excitotoxins in the pathophysiology of spinal cord injury. Ann. Neural., 24: 623-626. Falci, S., Best, L., Lammertse, D. and Starnes, C. (1999) Surgical treatment of spinal cord injury pain using a new technique of intramedullary electrical analysis. J. Spinal Cord Med., 22: 39. Fairbanks, CA., Schreiber, K.M., Brewer, K.L., Yu, C.G., Stone, L.S., Kitto, K.F., Nguen, H.O., Grochocholski, B.M., Shoeman, D., Kehl, L.J., Regunathan, S., Reis, D., Yezierski, R.P. and Wilcox, G.L. (2000) Agmatine reverses pain induced by inflammation, neuropathy and spinal cord injury. Proc. Natl. Acad. North America, in press. Frommer, G.P., Trefz, B.R. and Casey, K.L. (1977) Somatosen-

446 sory function and cortical unit activity in cats with only dorsal column fibers. Exp. Bruin Rex, 27: 113-129. Gerber, G., Youn, D.-H., Hsu, C.H., Isaev, D. and Randic, M. (2000) Spinal dorsal horn synaptic plasticity: involvement of group I metabotropic glutamate receptors. In: J. Sandkiihler, B. Bromm and G.F. Gebhart (Eds.), Nervous System Plasticity and Chronic Pain. Progress in Brain Research, Vol. 129. Elsevier, Amsterdam, pp. 1 15- 134. Gilad, G.M., Salame, K., Rabey, J.M. and Gilad, V.H. (1996) Agmatine treatment is neuroprotective in rodent brain injury models. Life Sci., 58: 41-46. Gomez-Pinilla, F., Tram, H., Cotman, C.W. and Nietosampedro, M. (1989) Neuroprotective effect of MK-and U-50488H after contusive spinal cord injury. Exp. Neural., 104: 118-124. Haley, J.E. and Wilcox, G.L. (1992) Involvement for excitatory amino acids and peptides in the spinal mechanisms underlying hyperalgesia. In: W.D. Willis (Ed.), Hyperulgesiu and Allodyniu. Raven Press, New York, pp. 281-293. Hao, J.-X., Xu, X.-J., Aldskogius, H., Seiger, A. and Wiesenfeld-Hallin, Z. (1991) The excitatory amino acid receptor antagonist MK-801 prevents the hypersensitivity induced by spinal cord ischemia in the rat. Exp. Neurol., 114: 182-191. Hao, J.-X., Xu, X.-J., Aldskogius, H., Seiger, A. and Wiesenfeld-Hallin, Z. (1992a) Transient spinal cord ischemia induces temporary hypersensitivity of dorsal horn wide dynamic range neurons to myelinated, but not unmyelinated, fiber input. J. Neurophysiol., 68: 384-391. Hao, J.-X., Xu, X.-J., Yu X, Y.-, Seiger, A. and WiesenfeldHallin, Z. (1992b) Baclofen reverses the hypersensitivity of dorsal horn wide dynamic range neurons to mechanical stimulation after transient spinal cord ischemia: implications for a tonic GABAergic inhibitory control of myelinated fiber input. Neurophysiology, 68: 392-396. Hao, J.-X., Xu, X.-J. and Wiesenfeld-Hallin, Z. (1994) Intrathecal-aminobutyric acids (GABAn) receptor antagonist CGP 45448 induces hypersensitivity to mechanical stimuli in the rat. Neurosci. L&t., 182: 299-302. Happel, R.D., Smith, K.P., Powers, J.M., Banik, N.L., Hogan, E.L. and Balentine, J.D. (1981) Calcium accumulation in experimental spinal cord trauma. Bruin Rex, 2 11: 476-479. Hargreaves, K., Dubner, R., Brown, E, Flores, C. and Joris, J. (1988) A new and sensitive method for measuring thermal nociception in cutaneous hyperalgesia. Pain, 32: 77-88. Hoheisel, U. and Mense, S. (2000) The role of spinal nitric oxide in the control of spontaneous pain following nociceptive input, In: J. Sandkiihler, B. Bromm and G.F. Gebhart (Eds.), Nervous System Plusticity and Chronic Pain. Progress in Brain Research, Vol. 129. Elsevier, Amsterdam, pp. 163-172. Hsu, C.Y., Lin, T.-N., Xu, J., Chao, J. and Hogan, E.L. (I 994) Kinins and related inflammatory mediators in central nervous system injury. In: S.K. Salzman and A.L. Faden (Eds.), The Neurobiology of Central Nervous System Trauma. Oxford Press, New York, pp. 145-I 54. Hulsebosch, C.E., Christensen, M.D., Peng, Y.B. and Willis, W.D. (1997) Central sensitization of wide dynamic range neurons in chronic central after spinal cord hemisection. APS Abstr., A-l 12.

Ignarro, L.J. (1991) Heme-dependent activation of guanylate cyclase by nitric oxide: a novel signal transduction mechanism. Blood Vessels, 28: 67-73. Jain, N., Florence, S. and Kaas, J.H. (1998) Reorganization of somatosensory cortex after nerve and spinal cord injury. News Physiol. Sci., 14: 144-149. Jensen, ST. and Nikolajsen, L. (2000) Pre-emptive analgesia in postamputation pain: an update. In: J. Sandkiihler, B. Bromm and G.F. Gebhart (Eds.), Nervous System Plasticity and Chronic Pain. Progress in Brain Research, Vol. 129. Elsevier, Amsterdam, pp. 493-503. Kakulas, B.A., Smith, E., Gaekwad, U.F., Kaelan, C. and Jacobsen, P. (1990) The neuropathology of pain and abnormal sensations in human spinal cord injury derived from the clinicopathological data base at the Royal Perth Hospital. In: M.R. Dimitrijevic, PD. Wall and U. Lindblom (Eds.), Recent Achievements in Restorative Neurology. Karger, Basel, pp. 3741. Kaltschmidt, B., Baeuerle, PA. and Kaltschmidt, C. (1993) Potential involvement of the transcription factor NF-kB in neurological disorders. Mol. Aspects Med., 14: 171-190. Kendall, D. (1949) Some observations on central pain. Bruin, 62: 253-273. Laughlin, T., Bethea, J., Yezierski, R. and Wilcox, G. (2000) Involvement of the pro-inflammatory cytokine IL- 18 in dynorphin induced allodynia. Pain, 84: 159-167. Lazarewicz, J.W., Wroblewski, J.T., Palmer, M.E. and Costa, E. (1988) Activation of N-methyl-D-aspartate-sensitive glutamate receptors stimulates arachidonic acid release in primary cultures of cerebellar granule cells. Neuropharmucology, 27: 769. Lee, S.K., Schottler, F., Oliver, M. and Lynch, G. (1980) Brief bursts of high-frequency stimulation produce two types of structural change in rat hippocampus. J. Neurophysiol., 44: 247-258. Leijon, G. and Boivie, J. (1991) Pharmacological treatment of central pain. In: K.L. Casey (Ed.), Pain and the Centrul Nervous System Disease: The Central Pain Syndromes. Raven Press, New York, pp. 257-266. Lenz, E ( 199 1) The thalamus and central pain syndromes: human and animal studies. In: K.L. Casey (Ed.), Pain und Central Nervous System Disease: The Central Pain Syndromes. Raven Press, New York, pp. 171-182. Lenz, EA., Lee, J.-I., Garonzik, I.-M., Rowland L.H., Dougherty, P.M. and Hua, S.E. (2000) Human thalamus reorganization related to nervous system injury and dystonia. In: J. Sandktihler, B. Bromm and G.F. Gebhart (Eds.), Nervous System Plasticity and Chronic Pain. Progress in Brain Research, Vol. 129. Elsevier, Amsterdam, pp. 259-273. Levi, R., Hultling, C., Nash, M.S. and Seiger, A. (1995) The Stockholm spinal cord injury study, I, Medical problems in a regional SC1 population, Paraplegia, 33: 308-315. Levitt, M. (1985) Dysesthesias and self-mutilation in humans and subhumans: a review of clinical and experimental studies. Brain Res. Rev., IO: 247-290. Levitt, M. (I 989) Postcordotomy spontaneous dysesthesias in

441

macaques: recurrence after spinal cord transection. Bruin Rex, 481: 41-56. Li, Z.H., Hogan, E.L. and Banik, N.L. (1995) Role of calpain in spinal cord injury: increased calpain immunoreactivity in rat spinal cord after impact trauma. Neurochem. Rex, 21: 441448. Lighthall, J.W. and Anderson, T.E. (1994) In vivo models of experimental brain and spinal cord trauma. In: S.K. Salzman and A.I. Faden (Eds.), The Neurobiology of Central Nervous System Truuma. Oxford Press, New York, pp. 3-l 1. Linden, D.J., Sheu, F.-S., Murakami, K. and Roouttenberg, A. (1987) Enhancement of long-term potentiation by cis-unsaturated fatty acid: relation to protein kinase c and phospholipase A2. J. Neutosci., 7: 3183-3792. Liu, S., Ruenes, G. and Yezierski, R.P. (1997) NMDA and non-NMDA receptor antagonists protect against excitotoxic injury in the rat spinal cord. Bruin Rex, 756: 160-167. Livingston, W.K. (1943) Pain Mechanisms. Macmillan, New York. Loeser, J.D., Ward, A.A. and White, L.E. (1968) Chronic deafferentation of human spinal cord neurons. J. Neurosurg., 29: 48-50. Lombard, M.C., Nashold, B.S. and Pelissier, T. (1979) Thalamic recordings in rats with hyperalgesia. In: J.J. Bonica, J.C. Liebeskind and D.G. Albe-Fessard (Eds), Advances in Pain Research and Therapy. Raven Press, New York, pp. 767-772. Lynch, G. and Baudry, M. (1984) The biochemistry of memory: a new and specific hypothesis. Science, 224: 1057-1063. Mailis, A. (1996) Compulsive targeted self-injurious behavior in humans with neuropathic pain: a counterpart of animal autotomy? Four case reports and literature review. Pain, 64: 569-578. Malmberg, A. (2000) Protein kinase subtypes involved in injury-induced nociception. In: J. Sandkiihler, B. Bromm and G.F. Gebhart (Eds.), Nervous System Plasticity and Chronic Pain. Progress in Brain Research, Vol. 129. Elsevier, Amsterdam, pp. 51-59. Malmberg, A.B., Chen, C., Tonegawa, S. and Basbaum, A.I. (1997) Preserved acute pain and reduced neuropathic pain in mice lacking PKCgamma. Science, 278: 275-279. Mantyh, PW., Rogers, S.D., Honore, I?, Allen, B.J., Ghilardi, J.R., Li, J., Daughters, R.S., Lappi, D.A., Wiley, R.G. and Simone, D.A. (1997) Inhibition of hyperalgesia by ablation of lamina I spinal neurons expressing the substance P receptor. Science, 278: 279-283. Marsala, M., So&in, L.S. and Yaksh, T.L. (1994) Transient spinal ischemia in rat: characterization of spinal cord blood flow, extracellular amino acid release and concurrent histopathological damage. J. Cereb. Blood Flow Metab., 14: 6 14-624. Meller, ST. and Gebhart, G.F. (1994) Nitric oxide (NO) and nociceptive processing in the spinal cord. Pain, 52: 127-146. Melzack, R. and Loeser, J.D. (1978) Phantom body pain in paraplegics: evidence for a central ‘pattern generating mechanism’ for pain. Pain, 4: 195-210. Moore, K.A., Baba, H. and Woolf, C.J. (2000) Synaptic transmission and plasticity in the superficial dorsal horn. In: J.

Sandktihler, B. Bromm and G.F. Gebhart (Eds.), Nervous System Plasticity and Chronic Pain. Progress in Brain Research, Vol. 129. Elsevier, Amsterdam, pp. 63-80. Morrow, T.J., Paulson, P.E., Brewer, K.L., Yezierski, R.P. and Casey, K.L. (2000) Chronic, selective forebrain responses to excitotoxic dorsal horn injury. Exp. Neurol., 161: 220-226. Nag, S. and Riopelle, R.J. (1990) Spinal neuronal pathology associated with continuous intrathecal infusion of N-methyl-D-aspartate in the rat. Acta Neuropathol., 81: 713. Nashold, B.S., Jr. (1991) Paraplegia and pain. In: B.S. Nashold, Jr. and J. Ovelmen-Levitt (Eds.), Advances in Pain Research and Therapy, Vol. 19, Deafferentation Pain Syndromes: Pathophysiology and Treatment. Raven Press, New York, pp. 301319. Nepomuceuno, C., Fine, P.R., Richards, J.S., Gowens, H., Stover, S.L., Rantanuabol, U. and Houston, R. (1979) Pain in patients with spinal cord injury. Arch. Phys. Med. Rehab., 60: 605609. Ness, T.J., Pedro, E.C.S., Richards, J.S., Kezar, L., Liu, H.-G. and Mountz, J.M. (1998) A case of spinal cord injury-related pain with baseline rCBF brain SPECT imaging and beneficial response to gabapentin. Pain, 78: 139-143. Nogawa, S., Forster, C.. Zhang, E. Nagayama, M., Ross, M.E. and Ladecola, C. (1998) Interaction between inducible nitric oxide synthase and cyclooxygenase-2 after cerebral ischemia. Proc. Natl. Acad. Sci. USA, 95: 10966-10971. O’Neil, L.A.J. and Kaltschmidt, C. (1997) NF-kB: a crucial transcription factor for glial and neuronal cell function, TINS, 20: 252-258. Pagni, C.A. (1998) Central Pain: A Neurosurgical Challenge. Torino: Edizioni Minerva Medica. Pahl, H.L. (1999) Activators and target genes of Rel/NF-KB transcription factors. Oncogene, 18: 6853-6866. Palecek, J., Paleckova, V. and Willis, W.D. (1999) The effect of phorbol esters on spinal cord amino acid concentrations and responsiveness of rats to mechanical and thermal stimuli. Pain, 80: 597-605. Panter, S.S., Yam, S.W. and Faden, AI. (1990) Alterations in extracellular amino acids after traumatic spinal cord injury. Ann. Neural., 27: 96-99. Pertovaara, A. (2000) Plasticity in descending pain modulatory systems. In: J. Sandkiihler, B. Bromm and G.F. Gebhart (Eds.), Nervous System Plasticity and Chronic Pain. Progress in Brain Research, Vol. 129. Elsevier, Amsterdam, pp. 23 l-242. Piomelli, D. (1994) Arachidonic acid in cell signaling. Curr Opin. Cell Biol., 5: 274-280. Pisharodi, M. and Nauta, W.J.W. (1985) An animal model for neuron-specific spinal cord lesions by the microinjection of N-methylaspartate, kainic acid, and quisqualic acid. Appl. Neurophysiol., 48: 226-233. Plunkett, J.A., Yu, C.G., Easton, J.M., Bethea, J.R. and Yezierski, R.P. (2000) Effects of interleukin-10 (IL-IO) on pain behavior and gene expression following excitotoxic spinal cord injury in the rat. Exp. Neurol., submitted. Pollock, L.J., Brown, M., Boshes, B., Finkelman, I., Chor, H., Arieff, A.J. and Fir&l, J.R. (1951) Pain below the level of

448

injury of the spinal cord. AMA Arch. Neural. Psychiat., 65: 319-322. Regan, R. and Choi, D.W. (1994) Excitoxicity and central nervous system trauma. In: S.K. Salzman and A.L. Faden (Eds.), The Neurobiology of Central Nervous System Trauma. Oxford Press, New York, pp. 173-181. Rehncrona, S., Westerberg, E., Akesson, B. and Siedjo, B.K. (1982) Brain cortical fatty acids and phospholipids during and following complete and severe incomplete ischemia. J. Neurosci., 38: 84-93. Reis, D.J. and Regunathan, S. (1998) Agmatine: a novel neurotransmitter?. Adv. Pharmacol., 42: 645-649. Rintala, D.H., Loubser, P.G., Castro, J., Hart, K.A. and Fuhrer, M.J. (1998) A comprehensive assessment of chronic pain in a community-based sample of men with spinal cord injury. J. Phys. Med. Rehab., 19: 604-614. Rodin, B.E. and Kruger, L. (1984) Deafferentation in animals as a model for the study of pain: alternative hypothesis. Brain Res. Rev., 7: 313-328. Sagen, J. (1996) Chromaffin cell transplants in the CNS: basic and clinical update. In: R.P. Lanza and W.L. Chick (Eds.), Yearbook of Cell and Tissue Transplantation. Kluwer, Dordrecht, pp. 71-89. Salvemini, D., Misko, T.P, Masferrer, J.L., Seibert, K., Currie, M.G. and Needleman, P. (1994) Nitric oxide activates cyclooxygenase enzymes. Proc. Natl. Acad. Sci. USA, 90: 72407244. Sandktlhler, J., Benrath, J., Brechtel, C., Ruscheweyh, R. and Heinke, B. (2000) Synaptic mechanisms of hyperalgesia. In: J. Sandktihler, B. Bromm and G.F. Gebhart (Eds.), Nervous System Plasticity and Chronic Pain, Progress in Brain Research, Vol. 129. Elsevier, Amsterdam, pp. 81-100. Savage, D.D., Werling, L.L., Nadler, J.V. and McNamara, J.O. (1982) Selective increase in L-[4H]glutamate binding to a quisqualate-sensitive site on hippocampal synaptic membranes after angular bundle kindling. Eur: J. Phannacol., 85: 255256. Siddall, P., Xu, CL. and Cousins, M. (1995) Allodynia following traumatic spinal cord injury in the rat. Neuroreport, 6: 12411244. Siddall, P.J., Taylor, D.A. and Cousins, M.J. (1997) Classification of pain following spinal cord injury. Spinal Cord, 45: 69-75. Siddall, P.J., Xu, C.L., Floyd, N. and Keay, K.A. (1999) C-jos expression in the spinal cord in rats exhibiting allodynia following contusive spinal cord injury. Brain Res., 85 1: 281286. Siman, R., Noszek, J. and Kegerise, C. (1985) Calpain I activation is specifically related to excitatory amino acid induction of hippocampal damage. J. Neumsci., 9: 1579-1590. Simpson, R.K., Robertson, C.S. and Goodman, J.C. (1990) Spinal cord ischemia-induced elevation of amino acids: extracellular measurements with microdialysis. Neurochem. Res., 15: 635-639. Springer, J.E., Azbill, R.D., Kennedy, S.E., George, J. and Geddes, J.W. (1997) Rapid calpain I activation and cytoskeletal protein degradation following traumatic spinal cord injury: at-

tenuation with riluzole pretreatment. J. Neurochem., 69: 15921600. Summers, J.D., Rapoff, M.A., Varghese, G., Porter, K. and Palmer, R.E. (1991) Psychosocial factors in chronic spinal cord injury pain. Pain, 47: 183-189. Suzuki, K., Imajoh, S., Emori, Y., Kawasaki, H., Minami, Y. and Ohno, S. (1987) Calcium-activated neutral protease and its endogenous inhibitor: activation at the cell membrane and biological function. FEBS Lett., 220: 27 1-277. Takahashi, Y. and Nakajima, Y. (1996) Dermatomes in the rat limbs as determined by antidromic stimulation of sensory C-fibers in spinal nerves. Pain, 67: 197-202. Tasker, R.R., de Carvalho, G. and Dostrovsky, J.O. (1991) The history of central pain syndromes, with observations conceming pathophysiology and treatment. In: K.L. Casey (Ed.), Pain and Central Nervous System Disease: The Central Pain Syndromes. Raven Press, New York, pp. 31-58. Tator, C.H. and Fehlings, M.G. (1991) Review of the secondary injury theory of acute spinal cord trauma with emphasis on vascular mechanisms. J. Neurosurg., 75: 15-26. Tunks, E. (1986) Pain in spinal cord injured patients. In: R.F. Bloch and M. Basbaum (Eds.), Management of Spinal Cord Injuries. William and Wilkins, Baltimore, pp. 180-2 11. Urea, G. and Urea, R. (1990) Neurotoxic effects of excitatory amino acids in the mouse spinal cord: quisqualate and kainate but not N-methyl-D-aspartate induce permanent neural damage. Brain Res., 529: 7-15. Vierck, C.J., Jr. (1991) Can mechanisms of central pain syndromes be investigated in animal models? In: K.L. Casey (Ed.), Pain and Central Nervous System Disease: The Central Pain Syndromes. Raven Press, New York, pp. 129-141. Vierck, C.J. and Light, A.R. (2000) Allodynia and hyperalgesia within dermatomes caudal to a spinal cord injury in primates and rodents. In: J. Sandktihler, B. Bromm and G.F. Gebhart (Eds.), Nervous System Plasticity and Chronic Pain. Progress in Brain Research, Vol. 129. Elsevier, Amsterdam, pp. 411428. Widerstrom-Noga, E.G., Cuervo, E., Broton, J.G., Duncan, R.C. and Yezierski, R.P. (1999) Self-reported consequences of spinal cord injury: results of a postal survey. Arch. Phys. Med. Rehab., 80: 580-586. Wiesenfeld-Hallin, Z., Hao, J.-X., Aldskogius, H., Seiger, A. and Xu, X.-J. (1994) Allodynia-like symptoms in rats after spinal cord ischemia: an animal model of central pain. In: J. Boivie. P. Hansson and U. Lindblom (Eds.), Touch, Temperature and Pain in Health and Disease: Mechanisms and Assessments. Progress in Pain Research and Management, Vol. 4, IASP Press, Seattle, WA, pp. 455-472. Woolf, C.J. (1992) Excitability changes in central neurons following peripheral damage. In: W.D. Willis (Ed.), Hyperalgesia and Allodynia. Raven Press, New York, pp. 221-243. Wrathall, J.R., Choiniere, D. and Teng, Y.D. (1994) Dose dependent reduction of tissue loss and functional impairment after spinal cord trauma with the AMPA/kainate antagonist NBQX. J. Neurosci., 14: 6598-6607. Yakovlev, A.G. and Faden, AI. (1994) Sequential expression of c:fos protooncogene, TNF-alpha, and dynorphin genes

449 in spinal cord following experimental traumatic injury. Mol. Chem. Neuropathol., 24: 179-190. Yaksh, T.L. (1989) Behavioral and autonomic correlates of the tactile evoked allodynia produced by spinal glycine inhibition: effects of modulatory receptor systems and excitatory amino acid antagonists. Pain, 47: 111-123. Yezierski, R.P (1996) Pain following spinal cord injury: the clinical problem and experimental studies. Pain, 68: 185-194. Yezierski, R.P. and Park, S.H. (1993) The mechanosensitivity of spinal sensory neurons following intraspinal injections of quisqualic acid in the rat. Neurosci. L&t., 157: 115-l 19. Yezierski, R.P., Santana, M., Park, D.H. and Madsen, P.W. (1993) Neuronal degeneration and spinal cavitation following intraspinal injections of quisqualic acid in the rat. J. Neurotrauma, 10: 445-456. Yezierski, R.P., Liu, S., Ruenes, G.L., Kajander, K.J. and Brewer, K.L. (1998a) Excitotoxic spinal cord injury: behavioral and morphological characteristics of a central pain model. Pain, 75: 141-155. Yezierski, R.P., Brewer, K.L., Fairbanks, CA. and Wilcox, G.L. (1998b) Neuroprotective effects of intraspinal agmatine fol-

lowing excitotoxic spinal cord injury. Neurosci. Abstr, 24: 577. Young, W. (1987) The post-injury responses in trauma and ischemia: secondary injury or protective mechanism?. CNS Trauma, 4: 27-52. Yu, C.G., Marcillo, A., Fairbanks, C.A., Wilcox, G.L. and Yezierski, R.P. (2000) Agmatine improves locomotor function and reduces tissue damage following traumatic spinal cord injury. NeuroReport, in press. Zeltser, R. and Seltzer, Z. (1994) A practical guide for the use of animal models in the study of neuropathic pain. In: J. Boivie, P. Hansson and U. Lindblom (Eds.), Touch, Temperature and Pain in Health and Disease: Mechanisms and Assessments. Progress in Pain Research and Management. Vol. 4. IASP Press, Seattle, WA, pp. 295-338. Zhang, A.-L., Hao, J.-X., Seiger, A., Xu, X.-J., WiesenfeldHallin, Z., Grant, G. and Aldskogius, H. (1999) Decreased GABA immunoreactivity in spinal cord dorsal horn neurons after transient spinal cord ischemia in the rat. Bruin Rex, 656: 187-190.

J. SandHthler, B. Bromm and GE Gebhart (Eds.) Progress in Brain Research, Vol. 129 0 2000 Elsevier Science B.V. All tights resewed

CHAPTER 34

A role for the endogenous cannabinoid system in the peripheral control of pain initiation Antonio Calignano ‘, Giovanna La Rana ‘, Patrick Loubet-Lescouli6 2 and Daniele Piomelli 2** 2 Department

’ Department of Pharmacology,

of Pharmacology, University University of California,

Introduction Although most pain signals arise from the stimulation of nociceptors located in peripheral tissues, the earliest stages of pain modulation are generally thought to occur within the central nervous system (CNS). This view has been challenged, however, by multiple evidence indicating that functionally active antinociceptive mechanisms are also present outside the CNS. For example, it is now recognized that endogenous opioid peptides released upon activation of immune cells during inflammation inhibit pain transmission through the interaction with opioid receptors on peripheral sensory nerve endings (see, for example: Stein et al., 1990; Inoue et al., 1998; Kolesnikov and Pastemak, 1999; Machelska et al., 1999; see also Ingram, 2000, this volume). This concept has both theoretical and clinical relevance, as it implies that peripherally applied opioid drugs may provide significant pain control without producing the undesirable effects that inevitably follow the recruitment of brain opioid receptors (Stein, 1991). A role in the control of pain initiation has also been proposed for cannabinoid receptors,

* Corresponding author: D. Piomelli, 360 Med Surge II, University of California, Irvine, CA 92697-4625, USA. Tel.: +l-949-824-6180; Fax: +l-949-824-6305; E-mail: [email protected]

of Naples, Naples 80131, Italy 360 Med Surge II, Irvine, CA 92697-4625,

USA

the molecular target of the Cannabis constituent A’-tetrahydrocannabinol (A9-THC), and for their attending system of endogenous ligands. The aim of this review is to outline the experimental evidence supporting such a role. To place this evidence into perspective, we will first describe the pathways of formation and inactivation of the endogenous cannabinoids (endocannabinoids) and the pharmacological properties of the receptors they activate. We will then focus on the implication of the endocannabinoid system in the central processing of pain signaling. Finally, we will turn our attention to the antinociceptive functions served by these compounds in peripheral organs and tissues and delineate several unresolved issues that still await experimentation. Endocannabinoid

formation

and inactivation

The endocannabinoids, which include anandamide (arachidonylethanolamide) and 2arachidonylglycerol (2-AG), are a class of lipid compounds found in the brain and other tissues (Devane et al., 1992; Di Marzo et al., 1994; Mechoulam et al., 1995; Sugiura et al., 1995; Stella et al., 1997) (Fig. 1). In contrast to classical neurotransmitters and neuropeptides, anandamide and 2-AG are produced upon demand by receptor-stimulated cleavage of lipid precursors, and are released from neurons and other cells immediately after their production. In the case of anandamide, the membrane precursor is represented by

/\/OH

472

NH Anandamide

n

2-Arachidonylglycerol

N-arachidonyl

PE

MMMpjlNpo”

Phospholipase D

Palmitylethanolamide

Fig. 1. Chemical structures of anandamide, 2-arachidonylglycerol and palmitylethanolamide.

an N-acylated species of phosphatidylethanolamine (PE), N-arachidonyl PE (Fig. 2). Activation of cell surface receptors triggers a chain of biochemical events that culminates in the enzymatic cleavage of IV-arachidonyl PE by an unknown phospholipase D and in the extracellular release of anandamide (Di Marzo et al., 1994; Cadas et al., 1996; Sugiura et al., 1996; Giuffrida et al., 1999). An illustration of this receptor-dependent process has been provided by in vivo microdialysis studies, which suggest that anandamide may be released in the brain striatum upon activation of Dz-type dopamine receptors and that such release may be involved in counterbalancing the stimulatory effects of dopamine on motor activity (Giuffrida et al., 1999; Beltramo et al., 2000). The most likely route of 2-AG biosynthesis involves the same enzymatic cascade responsible for the generation of second messengers, inositol trisphosphate and 1,2 diacylglycerol (DAG) (Fig. 3). Phospholipase C acting on phosphatidylinositol bisphosphate produces DAG, which is converted to 2-AG by a DAG-lipase activity (Stella et al., 1997). Additional pathways of 2-AG formation may implicate the hydrolysis of lysophospholipids or triacylglycerols (for a review, see Piomelli et al., 1998) (Fig. 3). Regardless of the mechanism involved, 2-AG biosynthesis may be triggered by neural activity or by occupation of membrane receptors. For example, in the rat hippocampus, 2-AG levels are strongly increased by stimulation of the Schaffer

ii

Anandamide

Ii

Amidohydrolase

Arachidonic Acid

Ethanolamine

Fig. 2. Hypothetical model illustrating two primary pathways of anandamide formation and inactivation. Hydrolysis of N-arachidonyl phosphatidylethanolamine (TV-arachidonyl PE) by an unknown phospholipase D produces anandamide, which is then released into the external milieu. After release, anandamide is taken up by cells via a selective carrier system (not shown in figure), and broken down by anandamide amidohydrolase to yield arachidonic acid and ethanolamine.

collaterals, an excitatory fiber tract that utilizes glutamate as a neurotransmitter and represents an essential component in the processing of information by the hippocampus (Stella et al., 1997). After release, anandamide and 2-AG are thought to accumulate back into cells via a high-affinity transport system which, unlike amine and amino acid

473

1Phospholipase

Cl

&FPhospholipid

r-2

Lysophospholipid

1,2 diacylglycerol

2-arachidonylglycerol Fig. 3. Hypothetical model illustrating the pathways of 2-AG formation. (Left) Phospholipase C (PLC) acts on phospholipids generating 1,2 diacylglycerol. This diglyceride is cleaved in turn by an unknown 1,2 diacylglycerol-lipase, producing 2arachidonylglycerol. (Right) Alternatively, a phospholipase At activity may form a lysophospholipid, which is further hydrolyzed by a lysophospholipid PLC to yield 2-arachidonylglycerol,

transporters, is both energy- and Naf-independent (Di Marzo et al., 1994; Beltramo et al., 1997; Hillard et al., 1997). Although this carrier system remains uncharacterized at the molecular level, its selectivity for anandamide and 2-AG over many other fatty acid ethanolamides and esters is reasonably well documented (Piomelli et al., 1999). Once inside cells,

both anandamide and 2-AG are substrates for enzymes which catalyze their hydrolytic breakdown. The enzyme responsible for anandamide hydrolysis, referred to as anandamide amidohydrolase or fatty acid amide hydrolase (Desarnaud et al., 1995; Hillard et al., 1995; Ueda et al., 1995), has been purified and cloned (Cravatt et al., 1996; Giang and Cravatt,

474

1997). This enzyme is a membrane-bound amidase with relatively low substrate selectivity, which corresponds with the N-acylethanolamine amidohydrolase activity identified, before the discovery of anandamide, by Schmid and coworkers (Schmid et al., 1985). 2-AG may also serve as a substrate for anandamide amidohydrolase in vitro (Di Marzo et al., 1998; Goparaju et al., 1998, 1999). There is evidence, however, that a different enzyme, possibly a monoacylglycerol lipase, mediates 2-AG hydrolysis in intact cells (Beltramo and Piomelli, 2000). In conclusion, the available data suggest that anandamide and 2-AG are released upon demand from neurons and other cells through distinct receptor-dependent mechanisms. This property, along with the short life spans of anandamide and 2-AG, suggest that these endocannabinoid lipids may act near their site of synthesis to modulate the effects of primary neurotransmitters and hormones.

many peripheral cells, including dorsal root ganglion neurons (see below) and inflammatory cells (Munro et al., 1993; Galiegue et al., 1995). By contrast, CB2 receptors are apparently absent from the normal brain, but they are abundantly expressed in immune cells such as B-lymphocytes, natural killer cells and monocytes (Galiegue et al., 1995). When evaluating the therapeutic potential for CB 1 and CB2 receptor agonists, an important factor to consider is that these drugs cause a rapid receptor desensitization (Bouaboula et al., 1999; Jin et al., 1999). As shown for other G protein-coupled receptors, this process may be mediated by the G protein-coupled receptor kinase/8-arrestin pathway, and may be responsible for the loss of pharmacological activity that is associated with the prolonged administration of cannabinoid agonists in vitro and in vivo (Sim et al., 1996). Central modulation

Cannabinoid

of pain signaling

receptors

Two cannabinoid receptor subtypes, termed CB 1 and CB2, have been cloned, and shown to belong to the superfamily of G protein-coupled membrane receptors (Devane et al., 1988; Matsuda et al., 1990; Munro et al., 1993). Although CBl and CB2 receptors share only about 44% sequence identity, they have sufficient similarity to recognize with comparable affinities a variety of structurally distinct agonists, such as the compounds CP55940 (a bicyclic cannabinoid) and WIN55212-2 (an aminoalkylindole) (see, for a review: Howlett, 1995; Matsuda, 1997). It has been possible, however, to develop excellent subtype-selective antagonists/inverse agonists - such as the compounds SR141716A for the CBl receptor (Rinaldi-Carmona et al., 1994) and SR144528 for the CB2 receptor (Rinaldi-Carmona et al., 1998) - while subtype-preferring agonists are just now beginning to emerge (Hanus et al., 1999). CBl and CB2 receptors differ quite dramatically not only in their molecular architecture, but also in their tissue and cell distribution. The CBl receptor is predominantly found in brain neurons, where it is highly represented in areas involved in the processing of movement, cognition and pain (Herkenham et al., 1990, 1991a; Matsuda et al., 1993). In addition to the CNS, CBl receptors are also present in

The reduction in pain behaviors observed after systemic administration of cannabinoid drugs is partly mediated by the activation of CB 1 receptors located in the brain and spinal cord. Brain sites that participate in cannabinoid analgesia have been identified by local microinjections of CBl receptor agonists, and include the amygdala, thalamus, superior colliculus, periaqueductal gray, and rostra1 ventromedial medulla (Lichtman and Martin, 1996; Martin et al., 1996; Tsou et al., 1996; Meng et al., 1998). In addition to the brain, CBl receptors are also found in the dorsal horn and lamina X of the spinal cord (Herkenham et al., 1991b; Tsou et al., 1998), where they may be located on nerve terminals of afferent sensory neurons, on intrinsic spinal neurons and/or on terminals of supraspinal neurons (Hohmann et al., 1999a). The presence of cannabinoid receptors in pain-processing areas of the CNS and the ability of cannabinoid antagonists to produce hyperalgesia (see, for example, Richardson et al., 1998) have led to the suggestion that endocannabinoids may participate in the intrinsic regulation of pain signaling. Further support for this hypothesis has been provided by two findings. First, mutant CBl-/- mice are hypoalgesic in the hot plate and formalin tests, a phenotype that may be interpreted as an overcompensation triggered by the inactivation of an intrinsic

475 pain-controlling system (Zimmer et al., 1999). Second, electrical stimulation in the periaqueductal gray causes a 40% increase in local anandamide outflow, as indicated by in vivo microdialysis studies (Walker et al., 1999). If confirmed, these findings would indicate that spontaneous or stimulated release of anandamide contributes to neuronal excitability in pain-controlling circuits within the CNS (Meng et al., 1998), and that drugs that prevent anandamide inactivation may be useful as analgesics. However, the results obtained thus far show that at least one such compound, the anandamide transport inhibitor AM404, does not change nociceptive threshold in the hot plate test at doses that elicit significant inhibitory effects on motor activity (Beltramo et al., 2000). Therefore, additional data are required to clarify the analgesic potential of this new class of pharmacological agents. Peripheral modulation

of pain signaling

In addition to the CNS, CBl receptors are also expressed in peripheral sensory neurons of the dorsal root ganglia (DRG). A small percentage of these CBl-expressing cells also contain the pronociceptive neuropeptides, substance P and a-calcitonin gene-related peptide (CGRP) (Hohmann and Herkenham, 1999b). Although quantitatively limited, the presence of CBl receptors on CGRP-containing neurons is likely to be functionally significant, since CB 1 agonists strongly decrease capsaicinevoked CGRP release from dorsal horn tissue in vitro (Richardson et al., 1998). Immunohistochemical experiments indicate that CBl receptors are present on central terminals of primary sensory afferents as well as on their peripheral counterparts. These experiments show in fact that CBl receptors synthesized in the cell bodies of DRG neurons are transported by axonal flow to sensory terminals in peripheral tissues (Hohmann and Herkenham, 1999a). In further support of this notion, administration of CBl receptor agonists in the skin alleviates the behavioral responses evoked by the chemical irritant, formalin (Calignano et al., 1998). Injections of dilute formalin into the rodent hind paw elicits a pain behavior consisting of two phases of licking and flexing of the injected limb. The first phase, which involves the acute activation of sensory fibers,

starts immediately after formalin injection and lasts for approximately 10 min. This phase is followed by a quiescent period lasting for lo-15 min, and then by a second, more sustained phase of pain behavior accompanied by central sensitization and inflammation (Dubuisson and Dennis, 1977). In mice, anandamide blocks the early phase of pain behavior when it is injected into the paw together with formalin, whereas both phases are prevented by synthetic cannabinoid agonists, such as WIN55212-2 and HU210 (Fig. 4a and data not shown). These analgesic effects are prevented by the CB I receptor antagonist SR141716A, but not by either the CB2 receptor antagonist SR144528 or the opioid receptor antagonist naloxone (Fig. 4a and data not shown). The rapid biological inactivation undergone by anandamide in tissues (see above) provides a likely explanation for why there is no effect of this compound during the late phase of the formalin response. This interpretation is supported by the fact that methanandamide, an inactivation-resistant anandamide analog, inhibits pain behavior during the entire duration of the test (Fig. 4a). Additional support for the peripheral antinociceptive actions of anandamide is provided by the ability of this compound to alleviate visceral pain in the turpentine model of urinary bladder inflammation. Systemically administered anandamide inhibits the viscera-visceral hyper-reflexia elicited by the injection of turpentine into the bladder. Unfortunately, the effects of cannabinoid receptor antagonists were not tested in these studies (Jaggar et al., 1998b). Several results indicate that the analgesic actions of anandamide on the formalin response are mediated by peripheral CBl receptors. First, local (intraplantar) anandamide administration is not accompanied by the classic signs of central cannabimimetic activity, such as catalepsy or immobility. Second, anandamide is approximately loo-fold more potent when applied locally than systemically (Fig. 4b). Third, when radioactively labeled anandamide is injected into the paw, most of the radioactivity (94%) is confined to the injected limb. Thus, the ability of anandamide to inhibit formalin-evoked nociception may be accounted for by an interaction of this compound with local CBl receptors, possibly present on the peripheral endings of sensory neurons (Hohmann and Herkenham, 1999b; Hohmann et al., 1999b). This conclusion agrees well with the

476

b

Formalin Drug

-

-

I 0.001 AEA

A&V SRI

AEAJ SR2

WIN

Fig. 4. Inhibitior nc 3f formalin-evoked nociception 50 kg intraplantar, i.pl.), WIN-55212-2 (WIN, 500 the CBl antagonist SR141716A (SRI, 0.1 mg per (b) Dose-dependent effects of anandamide following P < 0.01 (n = 12-18 for each condition). ODen Calignano et al. (1998), with permission.

WIN/ SRI

MAEA

I 0.01

I 0.1

Dose

, 1

I 10

I 100

(mglkg)

in mice by anandamide and other CBl agonists. (a) Effects of anandamide (AEA, Kg Lpi.) and methanandamide (MAEA, 50 kg i.pl.), in the absence or presence of kg intravenous, i.v.) or the CB2 antagonist SR144528 (SR2, 0.1 mg per kg i.v.). i.pl. (squares), i.v. (triangles) or intraperitoneal (i.p., circles) administrations. *, bars, earlv uhase, closed bars, late phase of the formalin response. Modified from I

I

finding that locally applied cannabinoid receptor agonists reduce capsaicin-induced thermal nociception in rhesus monkeys (Ko and Woods, 1999) and inhibit CGRP release from superfused rat skin in vitro (Richardson et al., 1998). An intriguing question raised by these studies is whether anandamide (and/or other endocannabinoids), generated spontaneously or as a result of tissue injury, participates in the intrinsic control of pain initiation. The presence of relatively high levels of anandamide in unstimulated skin suggests that this may be possible. Gas chromatography/mass spectrometry (GC/MS) analyses indicate indeed that rat skin contains about 50 pmol of anandamide per g of tissue, a concentration 5 to 15 times greater than that measured in plasma (Calignano et al., 1998; Giuffrida and Piomelli, 1998; Giuffrida et al., 2000). Another line of evidence supporting a role for the endocannabinoids in the processing of peripheral pain information is derived from experiments using the CB 1 receptor antagonist/inverse agonist SR141716A. Administration of SR141716A potentiates formalin-evoked pain in mice (Fig. 5). This hyperalgesic effect is particularly pronounced after local injection of the drug, and is observed during both the early and the late phases of the test (Fig. 5) (Calignano et al., 1998). However, SR141716A does not produce hyperalgesia in rats (A.S. Rice, pers.

commun.), indicating that the extent to which the endocannabinoid system participates in peripheral analgesia may depend on experimental conditions and/or species differences. Palmitylethanolamide

as a peripheral analgesic

As we have seen above, anandamide may be generated from the enzymatic cleavage of the phospholipid precursor N-arachidonyl PE (Fig. 2). N-arachidonyl PE belongs to a family of N-acylated PE species varying in the fatty acyl group linked to the primary amino group of PE. These N-acyl PE species give rise to fatty acid ethanolamides, such as palmitylethanolamide (PEA) and oleylethanolamide, the physiological roles of which are still unclear (see, for a review, Piomelli et al., 1998). PEA (Fig. 1) was originally isolated in 1963 by S. Udenfriend and coworkers (Bachur et al., 1965) and its anti-inflammatory properties are fairly well-characterized. However, its mechanism of action is still unclear (Aloe et al., 1993; Facci et al., 1995; Mazzari et al., 1996). In mice, PEA is as potent as anandamide at inhibiting formalin-evoked pain behavior (Fig. 6a) (Calignano et al., 1998). This effect cannot be accounted for by attenuation of inflammatory hyperalgesia, as it is not associated with a reduction of formalin-evoked edema. In six mice, the average

477

1501 b

10

15

20

25

30

10

15

20

25

30

Injection

(min)

5*

?oo2‘5 8 f

50-

n” 5

Time after formalin

Fig. 5. Hyperalgesic effects of CBl and CB2 antagonists on formalin-evoked nociception. (a) Effects of systemic administration of the CBl antagonist SR141716A (filled bars) and the CB2 antagonist SR144528 (hatched bars). The responses to formalin alone are illustrated by the empty bars. The antagonists were administered by i.v. injection at the dose of 0.1 mg per kg. (b) Hyperalgesic effects of local administration of SR141716A (filled bars). 10 pg of SR141716A were administered by i.pl. injection. *, P < 0.01 (n = 6). Modified from Calignano et al. (1998), with permission.

paw volumes are 0.18 & 0.003 ml under control conditions, 0.37 f 0.006 ml 30 min after injection of formalin, and 0.35 f 0.006 ml after injection of formalin plus PEA (50 ug). PEA-induced analgesia is reversed by pretreatment with the CB2 receptor antagonist SR144528 (Fig. 6a), whereas the CBl receptor antagonist SR141716A and the opioid receptor antagonist naloxone are ineffective (Fig. 6a and data not shown). Like anandamide, PEA is more potent when injeeted locally (by intraplantar injection) than when administered systemically (Fig. 6b). Together, these findings suggest that the antinocicep-

tive activity of PEA may be mediated by CBZlike cannabinoid receptors in skin. Importantly, PEA has similar antinociceptive actions when administered in rats, where systemic applications of this compound alleviate both formalin-evoked pain behaviors and turpentine-evoked bladder hyper-reflexia (Jaggar et al., 1998a). Despite the fact that the actions of PEA are prevented by the CB2-selective receptor antagonist SR144528, it is important to emphasize that PEA does not interact with the cloned CB2 receptor expressed in immune cells (Showalter et al., 1996). Thus, the structural relationship between the CB2like receptor activated by PEA and other cannabinoid receptor subtypes, if any, remains at present unresolved. The cellular localization of the putative PEA receptor is also unknown. There is evidence that CB2 receptors may be present on peripheral neurons (Pertwee, 1999). However, PEA does not prevent nociceptive responses that require sensory neuron activation. For example, PEA does not inhibit capsaicin-induced pain and does not affect the behavioral response to acute thermal stimuli in the hot plate test (A. Calignano, G. La Rana and D. Piomelli, unpubl. results). This absence of effect suggests that the putative PEA receptor may not be localized on neurons, but rather on non-neuronal cells that participate in triggering the pain response elicited by chemical tissue damage. Is PEA produced in the skin and does it act locally to regulate pain initiation? As is the case with anandamide, GC/MS analyses reveal the presence of relatively high concentrations of PEA in unstimulated skin (692 & 119 pmol per g of tissue, II = 8). In addition, systemic administration of the CB2 receptor antagonist SR144528 causes a selective enhancement of the early phase, but not the late phase, of the formalin response (Fig. 7). The selectivity of this effect is unlikely to result from a rapid elimination of SR144528 following the early phase, because the drug reverses PEA-evoked analgesia during both early and late-phase pain behavior (Fig. 6). Thus, a plausible interpretation of these results is that locally generated PEA may participate in reducing the early stages of nociception following chemical skin damage.

478

200-

5 '0 '5m 2 .c 2

b

_

150-

IOO-

50-

0-

ti

Formalin Drug

-

+

+

+

-

PEA

PEAI SRI

Formalin AEAIPEA Antagonist

_

+ _

+ + -

+ + SRI

PEN SR2

SEA

OEA

Dose

(mglkg)

+ + SR2

Fig. 6. Inhibition of formalin-evoked nociception in mice by PEA. (a) Effects of palmitylethanolamide (PEA, 50 Kg i.pl.), stearylethanolamide (SEA, 50 kg i.pl.) and oleylethanolamide (O&l, 50 pg i.pl.), in the absence or presence of the CBl antagonist SR141716A (SRI, 0.1 mg per kg i.v.) or the CB2 antagonist SR144.528 (SR2, 0.1 mg per kg i.v.). (b) Dose-dependent effects of PEA following i.pl. (squares), i.v. (triangles) or i.p. (circles) administrations. *, P < 0.01 (n = 12-18). Open bars, early phase and closed bars, late phase of the formalin response. Modified from Calignano et al. (1998). with permission.

Synergistic inhibition

of pain

When anandamide and PEA are administered together in equal doses, they inhibit the early phase of formalin-evoked pain behavior with a potency that is about loo-fold greater than either of the compounds separately (Fig. 7a). A comparable enhancement is observed during the late phase, when anandamide produces no effect when given alone (Fig. 7b). Pretreatment with either a CBl or CB2 receptor antagonist completely prevents this response, suggesting that the cannabinoid receptors act in a truly synergistic manner (Fig. 7~). This synergism between anandamide and PEA is

reminiscent of those seen between other analgesic compounds (see, for example, Yaksh and Reddy, 1981), and appears to be restricted to peripheral tissues. Thus, direct injection of PEA into the third cerebral ventricle has no effect on the nociceptive threshold in the hot plate test, and does not enhance the inhibitory activity of anandamide when administered by the same route (Table 1). These findings suggest that peripheral CBl receptors activated by anandamide and CB2-like receptors activated by PEA may cooperate in the control of pain initiation. However, the mechanism underlying this cooperation is at present unknown.

479 TABLE

zooa

1

Effects

of anandamide

and PEA in the hot-plate

test

Time (min.)

Vehicle

AEA hot-plate latencies (s)

PEA

AEA/PEA

5 10 15 20 30 40 60

24f2 24f2 22f3 25 f 2 25zt3 251t3 261t2

24&t 23 zt3 29 f 3 35f2* 40+3* 30f3 28zt3

22*3 23 dz 3 2112 25zt4 26f5 22dc4 24f2

25 f2 26zt4 3Of3 37*3* 39f2* 35 f5 26+3

l

P

ZlSOb '5m 2 IOO.c h

0

A

l

O-h, 0

0.01

1

0.1

10

b

2001

100

1

Latencies to jump were measured after administration of vehicle (10% DMSO damide (AEA, 10 kg), PEA (10 Kg) or (10 ug each). Results are expressed as six experiments. Asterisks indicate P < subsequent Dunnett’s test.

intracerebroventricular in saline, 5 ~1). anananandamide plus PEA the mean i s.e.m. of 0.01 by ANOVA and

Conclusions and perspectives

0.01

I

I

I

I

0.1

1

IO

100

+ + SRl

+ + SR2

Dose (mg/paw)

C

2501 2ocz

.g 1502 2 .~‘O~ 2 50 o-Formalin AENPEA Antagonist

-

+ -

+ + -

Fig. 7. Synergistic effects of anandamide and PEA on formalin-evoked pain behavior: (a) early phase; (b) late phase. Equal amounts of anandamide and PEA, indicated in the abscissa, were injected into the paw (c). The CBl antagonist SR141716A (SRI) and the CB2 antagonist SRI44528 (SR2) prevent the effects of anandamide plus PEA (0.1 p.g each). Antagonists were given by i.v. injection (0.1 mg per kg); *, P < 0.01 (n = 18). Open bars, early phase, closed bars, late phase of the formalin response. Modified from Calignano et al. (1998), with permission.

Although cannabinoid analgesia has been generally associated with the activation of cannabinoid receptors in the CNS, the evidence reviewed here indicates that cannabinoid receptor agonists can exert potent antinociceptive effects by interacting with receptors located in peripheral tissues. Two distinct receptor subtypes have been implicated in this response: CBl receptors that are activated by anandamide and blocked by the CBl receptor antagonist SR141716A; and CB2-like receptors that are activated by PEA and blocked by the CB2 receptor antagonist SR144528. The presence of CBl receptors on terminals of sensory neurons tallies well with these pharmacological findings, suggesting that CBl agonists may produce peripheral antinociception by preventing the transmission of pain signals to the CNS. On the other hand, the molecular structure and cellular localization of the CB2-like receptor activated by PEA are still unknown. Yet, the potent antinociceptive effects of PEA suggest that this putative receptor may represent an important target for analgesic drugs and underscore the need for its molecular identification. Additional experimentation is also needed to characterize in greater detail the pharmacological activity of PEA in vitro and in vivo, including the potential effects of this compound on a wider variety of experimental models of acute and persistent pain.

480

Abbreviations 2-AG AEA CBl CB2 CGRP CNS DAG DRG OEA PE PEA SEA

2-arachidonyl glycerol anandamide cannabinoid receptor 1 cannabinoid receptor 2 calcitonin gene-related peptide central nervous system 1,2-diacylglycerol dorsal root ganglia oleylethanolamide phosphatidyl ethanolamine palmitylethanolamide sterarylethanolamide

Acknowledgements We thank H. Kim, F. Nava and A. Giuffrida for critical reading of the manuscript, and Roche Bioscience for an unrestricted grant (to D.P.). References Aloe, L., Leon, A. and Levi-Montalcini, R. (1993) A proposed autacoid mechanism controlling mastocyte behaviour. Agents Actions, 39: C145-147. Bachur, N.R., Masek, K., Melmon, K.L. and Udenfriend, S. (1965) Fatty acid amides of etbanolamine in mammalian tissues. J. Biol. Chem., 240: 1019-1024. Beltramo, M. and Piomelli, D. (2000) Carrier-mediated transport and enzymatic hydrolysis of the endogenous cannabinoid, 2-arachidonylglycerol. Neuroreport, 11: 1231-1235. Beltramo, M., Stella, N., Calignano, A., Lin, S.Y., Makriyannis, A. and Piomelli, D. (1997) Functional role of high-affinity anandamide transport, as revealed by selective inhibition. Science, 277: 1094-1097. Beltramo, M., Rodriguez de Fonseca, F., Navarro, M., Calignano, A., Gorriti, M.A., Grammatikopoulos, G., Sadile, A.G., Giuffrida, A. and Piomelli, D. (2000) Reversal of dopamine Dz-receptor responses by an anandamide transport inhibitor. J. Neurosci., 20: 3401-3407. Bouaboula, M., Dussossoy, D. and Casellas, P. (1999) Regulation of peripheral cannabinoid receptor CB2 phosphorylation by the inverse agonist SR 144528. Implications for receptor biological responses. J. Biol. Chem., 274: 20397-20405. Cadas, H., Gaillet, S., Beltramo, M., Venance, L. and Piomelli, D. (1996) Biosynthesis of an endogenous cannabinoid precursor in neurons and its control by calcium and CAMP. J. Neurosci., 16: 3934-3942. Calignano, A., La Rana, G., Giuffrida, A. and Piomelli, D. (1998) Control of pain initiation by endogenous cannabinoids. Nature, 394: 277-281.

Cravatt, B.F., Giang, D.K., Mayfield, S.P., Boger, D.L., Lemer, R.A. and Gilula, N.B. (1996) Molecular characterization of an enzyme that degrades neuromodulatory fatty-acid amides. Nature, 384: 83-87. Desamaud, F., Cadas, H. and Piomelli, D. (1995) Anandamide amidohydrolase activity in rat brain microsomes. Identification and partial characterization. J Biol. Chem., 270: 6030-6035. Devane, W.A., Dysarz, F.A.d., Johnson, M.R., Melvin, L.S. and Howlett, A.C. (1988) Determination and characterization of a cannabinoid receptor in rat brain. Mol. Pharmacol., 34: 605613. Devane, W.A., Hanus, L., Breuer, A., Pertwee, R.G., Stevenson, L.A., Griffin, G., Gibson, D., Mandelbaum, A., Etinger, A. and Mechoulam, R. (1992) Isolation and structure of a brain constituent that binds to the cannabinoid receptor. Science, 258: 1946-1949. Di Marzo, V., Fontana, A., Cadas, H., Schinelli, S., Cimino, G., Schwartz, J.C. and Piomelli, D. (1994) Formation and inactivation of endogenous cannabinoid anandamide in central neurons. Nature, 372: 686-691. Di Marzo, V., Bisogno, T., Sugiura, T., Melck, D. and De Petrocellis, L. (1998) The novel endogenous cannabinoid 2-arachidonoylglycerol is inactivated by neuronaland basophil-like cells: connections with anandamide. Biochem. J., 331: 15-19. Dubuisson, D. and Dennis, S.G. (1977) The formalin test: a quantitative study of the analgesic effects of morphine, meperidine, and brain stem stimulation in rats and cats. Pain, 4: 16 l174. Facci, L., Dal Toso, R., Romanello, S., Buriani, A., Skaper, S.D. and Leon, A. (1995) Mast cells express a peripheral cannabinoid receptor with differential sensitivity to anandamide and palmitoylethanolamide. Proc. Natl. Acad. Sci. USA, 92: 33763380. Galiegue, S., Mary, S., Marchand, J., Dussossoy, D., Carriere, D., &rayon, P., Bouaboula, M., Shire, D., Le Fur, G. and Casellas, P. (1995) Expression of central and peripheral cannabinoid receptors in human immune tissues and leukocyte subpopulations. Eur: J. Biochem., 232: 54-61. Giang, D.K. and Cravatt, B.F. (1997) Molecular characterization of human and mouse fatty acid amide hydrolases. Proc. Narl. Acad. Sci. USA, 94: 2238-2242. Giuffrida, A. and Piomelli, D. (1998) Isotope dilution GC/MS determination of anandamide and other fatty acylethanolamides in rat blood plasma. FEBS Mt., 422: 373376. Giuffrida, A., Parsons, L.H., Kerr, T.M., Rodriguez de Fonseca, E, Navarro, M. and Piomelli, D. (1999) Dopamine activation of endogenous cannabinoid signaling in dorsal striatum. Nat. Neurosci., 2: 358-363. Giuffrida, A., Rodriguez de Fonseca, F. and Piomelli, D. (2000) Quantification of bioactive acylethanolamides in rat plasma by electrospray mass spectrometry. Anal. Biochem., 280: 87-93. Goparaju, S.K., Ueda, N., Yamaguchi, H. and Yamamoto, S. (1998) Anandamide amidohydrolase reacting with 2-arachidonoylglycerol, another cannabinoid receptor ligand. FEBS Len., 422: 69-73. Goparaju, SK., Ueda, N., Taniguchi, K. and Yamamoto, S.

481

(1999) Enzymes of porcine brain hydrolyzing 2arachidonoyll glycerol, an endogenous ligand of cannabinoid receptors. Biochem. Phannacol., 57: 417-423. Harms, L., Breuer, A., Tchilibon, S., Shiloah, S., Goldenberg, D., Horowitz, M., Pertwee, R.G., Ross, R.A., Mechoulam, R. and Fride, E. (1999) HU-308: a specific agonist for CB(2), a peripheral cannabinoid receptor. Proc. Nail. Acad. Sci. USA, 96: 14228-14233. Herkenham, M., Lynn, A.B., Little, M.D., Johnson, M.R., Melvin, LX, De Costa, B.R. and Rice, KC. (1990) Cannabinoid receptor localization in brain. Proc. Natl. Acad. Sci. USA, 87: 1932-1936. Herkenham, M., Groen, B.G., Lynn, A.B., De Costa, B.R. and Richfield, E.K. (1991a) Neuronal localization of cannabinoid receptors and second messengers in mutant mouse cerebellum. Brain Res., 552: 301-310. Herkenham, M., Lynn, A.B., Johnson, M.R., Melvin, L.S., De Costa, B.R. and Rice, K.C. (1991b) Characterization and localization of cannabinoid receptors in rat brain: a quantitative in vitro autoradiographic study. J. Neurosci., 11: 563-583. Hillard, C.J., Wilkison, D.M., Edgemond, W.S. and Campbell, W.B. (1995) Characterization of the kinetics and distribution of N- arachidonylethanolamine (anandamide) hydrolysis by rat brain. Biochim. Biophys. Acta, 1257: 249-256. Hillard, C.J., Edgemond, W.S., Jarrahian, A. and Campbell, W.B. (1997) Accumulation of N-arachidonoylethanolamine (anandamide) into cerebellar granule cells occurs via facilitated diffusion. J. Neurochem., 69: 631-638. Hohmann, A.G. and Herkenbam, M. (1999a) Cannabinoid receptors undergo axonal flow in sensory nerves. Neuroscience, 92: 1171-1175. Hohmann, A.G. and Herkenham, M. (1999b) Localization of central cannabinoid CB 1 receptor messenger RNA in neuronal subpopulations of rat dorsal root ganglia: a double-label in situ hybridization study. Neuroscience, 90: 923-93 1. Hohmann, A.G., Briley, E.M. and Herkenham, M. (1999a) Preand postsynaptic distribution of cannabinoid and mu opioid receptors in rat spinal cord. Bruin Res., 822: 17-25. Hohmann, A.G., Tsou, K. and Walker, J.M. (1999b) Cannabinoid suppression of noxious heat-evoked activity in wide dynamic range neurons in the lumbar dorsal horn of the rat. J. Neurophysiol., 81: 575-583. Howlett, A.C. (1995) Pharmacology of cannabinoid receptors. Annu. Rev. Phamacol. Toxicol., 35: 607-634. Ingram, S.L. (2000) Cellular and molecular mechanisms of opioid action. In J. Sandkiihler, B. Bromm and G.F. Gebhart (Eds.), Nervous System Plasticity and Chronic Pain. Progress in Brain Research, Vol. 129. Elsevier, Amsterdam, pp. 483492. Inoue, M., Kobayashi, M., Kozaki, S., Zimmer, A. and Ueda, H. (1998) Nociceptin/orphanin FQ-induced nociceptive responses through substance P release from peripheral nerve endings in mice. Proc. Natl. Acad. Sci. USA, 95: 10949-10953. Jaggar, S.I., Hasnie, ES., Sellaturay, S. and Rice, AS. (1998a) The anti-hyperalgesic actions of the cannabinoid anandamide and the putative CB2 receptor agonist palmitoylethanolamide in visceral and somatic inflammatory pain. Pain, 76: 189-199.

Jaggar, S.I., Sellaturay, S. and Rice, A.S. (1998b) The endogenous cannabinoid anandarnide, but not the CB2 ligand pahntoylethanolamide, prevents the viscera-visceral hyper-reflexia associated with inflammation of the rat urinary bladder. Neurosci. L&t., 253: 123-126. Jin, W., Brown, S., Roche, J.P., Hsieh, C., Celver, J.P., Kovoor, A., Chavkin, C. and Ma&e, K. (1999) Distinct domains of the CBI cannabinoid receptor mediate desensitization and internalization. J. Neurosci., 19: 3773-3780. Ko, M.C. and Woods, J.H. (1999) Local administration of delta9tetrahydrocannabinol attenuates capsaicin-induced thermal nociception in rhesus monkeys: a peripheral cannabinoid action. Psychophamacology (Berl.), 143: 322-326. Kolesnikov, Y.A. and Pastemak, G.W. (1999) Peripheral blockade of topical morphine tolerance by ketamine. Eur: J. Pharmacol., 374: Rl-2. Lichtman, A.H. and Martin, B.R. (1996) Delta 9-tetrahydrocannabinol impairs spatial memory through a cannabinoid receptor mechanism. Psychophamacology (Berl.), 126: 125131. Machelska, H., Pfluger, M., Weber, W., Piranvisseh-Volk, M., Daubert, J.D., Dehaven, R. and Stein, C. (1999) Peripheral effects of the kappa-opioid agonist EMD 61753 on pain and inflammation in rats and humans. 1 Pharmacol. Exp. Ther, 290: 354-361. Martin, W.J., Hohmann, A.G. and Walker, J.M. (1996) Suppression of noxious stimulus-evoked activity in the ventral posterolateral nucleus of the thalamus by a cannabinoid agonist: correlation between electrophysiological and antinociceptive effects. J. Neurosci., 16: 6601-6611. Matsuda, L.A. (1997) Molecular aspects of cannabinoid receptors. Crit. Rev. Neurobiol., 11: 143-166. Matsuda, L.A., Lolait, S.J., Brownstein, M.J., Young, A.C. and Bonner, T.I. (1990) Structure of a cannabinoid receptor and functional expression of the cloned cDNA. Nature, 346: 561564. Matsuda, L.A., Bonner, T.I. and Lolait, S.J. (1993) Localization of cannabinoid receptor mRNA in rat brain. J. Camp. Neural., 327: 535-550. Mazzari, S., Canella, R., Petrelli, L., Marcolongo, G. and Leon, A. (1996) N-(2.hydroxyethyl)hexadecanamide is orally active in reducing edema formation and inflammatory hyperalgesia by down-modulating mast cell activation. ELK J. Pharmacol., 300: 227-236. Mechoulam, R., Ben-Shabat, S., Hanus, L., Ligumsky, M., Kaminski, N.E., Schatz, A.R., Gopher, A., Almog, S., Martin, B.R. and Compton, D.R. et al. (1995) Identification of an endogenous 2-monoglyceride, present in canine gut, that binds to cannabinoid receptors. Biochem. Pharmacol., 50: 83-90. Meng, I.D., Manning, B.H., Martin, W.J. and Fields, H.L. (1998) An analgesia circuit activated by cannabinoids. Nature, 395: 381-383. Munro, S., Thomas, K.L. and Abu-Shaar, M. (1993) Molecular characterization of a peripheral receptor for cannabinoids. Nature, 365: 61-65. Pertwee, R.G. (1999) Evidence for the presence of CB 1 cannabinoid receptors on peripheral neurones and for the existence of

482 neuronal non-CBl cannabinoid receptors. Life Sci., 65: 597605. Piomelli, D., Beltramo, M., Giuffrida, A. and Stella, N. (1998) Endogenous cannabinoid signaling. Neurobiol. Dis., 5: 462473. Piomelli, D., Beltramo, M., Glasnapp, S., Lin, S.Y., Goutopoulos, A., Xie, X.Q. and Makriyannis, A. (1999) Structural determinants for recognition and translocation by the anandamide transporter. Pmt. Nail. Acad. Sci. USA, 96: 5802-5807. Richardson, J.D., Kilo, S. and Hargreaves, K.M. (1998) Cannabinoids reduce hyperalgesia and inflammation via interaction with peripheral CB 1 receptors. Pain, 75: 11 l-l 19. Rinaldi-Carmona, M., Barth, F., Heaulme, M., Shire, D., Calandra, B., Congy, C., Martinez, S., Maruani, J., Neliat, G., Caput, D., Ferrara, P., Soubrie, I?, Breliere, J.C. and Le Fur, G. (1994) SR141716A, a potent and selective antagonist of the brain cannabinoid receptor. FEBS L&t., 350: 240-244. Rinaldi-Carmona, M., Barth, F., Millan, J., Derocq, J.M., Casellas, P., Congy, C., Oustric, D., Sarran, M., Bouaboula, M., Calandra, B., Portier, M., Shire, D., Breliere, J.C. and Le Fur, G.L. (1998) SR 144528, the first potent and selective antagonist of the CB2 cannabinoid receptor. J. Pharmacol. Exp. Thes,

284: 644-650.

Schmid, P.C., Zuzarte-Augustin, M.L. and Schmid, H.H. (1985) Properties of rat liver N-acyletbanolamine amidohydrolase. J. Biol. Chem., 260: 14145-14149. Showalter, V.M., Compton, D.R., Martin, B.R. and Abood, M.E. (1996) Evaluation of binding in a transfected cell line expressing a peripheral cannabinoid receptor (CB2): identification of cannabinoid receptor subtype selective ligands. .I Pharmacol. Exp. TheK, 278: 989-999.

Sim, L.J., Hampson, R.E., Deadwyler, S.A. and Childers, S.R. (1996) Effects of chronic treatment with delta9-tetrahydrocannabinol on cannabinoid-stimulated [35S]GTPgammaS autoradiography in rat brain. J. Neurosci., 16: 8057-8066. Stein, C. (1991) Peripheral analgesic actions of opioids. J. Pain Symptom Manage., 6: 119-124. Stein, C., Hassan, A.H., Przewlocki, R., Gramsch, C., Peter, K. and Herz, A. (1990) Gpioids from immunocytes interact with receptors on sensory nerves to inhibit nociception in

inflammation. Proc. Natl. Acad. Sci. USA, 87: 5935-5939. Stella, N., Schweitzer, P and Piomelli, D. (1997) A second endogenous cannabinoid that modulates long-term potentiation. Nature,

388: 773-778.

Sugiura, T., Kondo, S., Sukagawa, A., Nakane, S., Shinoda, A., Itoh, K., Yamashita, A. and Waku, K. (1995) 2-Arachidonoylglycerol: a possible endogenous cannabinoid receptor ligand in brain. Biochem. Biophys. Res. Commun., 215: 89-97. Sugiura, T., Kondo, S., Sukagawa, A., Tonegawa, T., Nakane, S., Yamashita, A. and Waku, K. (1996) Enzymatic synthesis of anandamide, an endogenous carmabinoid receptor ligand, through N-acylphosphatidylethanolamine pathway in testis: involvement of Ca(*+)-dependent transacylase and phosphodiesterase activities. Biochem. Biophys. Rex Commun., 218: 113-117. Tsou, K., Lowitz, K.A., Hohmann, A.G., Martin, W.J., Hathaway, C.B., Bereiter, D.A. and Walker, J.M. (1996) Suppression of noxious stimulus-evoked expression of FOS protein-like immunoreactivity in rat spinal cord by a selective cannabinoid agonist. Neuroscience, 70: 791-798. Tsou, K., Brown, S., Sanudo-Pena, M.C., Ma&e, K. and Walker, J.M. (1998) Immunohistochemical distribution of cannabinoid CBl receptors in the rat central nervous system. Neuroscience, 83: 393-41 I. Ueda, N., Kurahashi, Y., Yamamoto, S. and Tokunaga, T. (1995) Partial purification and characterization of the porcine brain enzyme hydrolyzing and synthesizing anandamide. J. Biol. Chem.,

270: 23823-23827.

Walker, J.M., Huang, SM., Strangman, N.M., Tsou, K. and Sanudo-Pena, M.C. (1999) Pain modulation by release of the endogenous cannabinoid anandamide. Proc. Natl. Acad. Sci. USA, 96: 12198-12203. Yaksh, T.L. and Reddy, S.V. (1981) Studies in the primate on the analgetic effects associated with intrathecal actions of opiates, alpha-adrenergic agonists and baclofen. Anesthesiology, 54: 45 1-467. Zimmer, A., Zimmer, A.M., Hohmann, A.G., Herkenham, M. and Bonner, T.I. (1999) Increased mortality, hypoactivity, and hypoalgesia in cannabinoid CBl receptor knockout mice. Proc.

Natl.

Acad.

Sci. USA, 96: 5780-5785.

.I. Sand&ihler, B. Bromm and GE Gebhat (Eds.) Progress in Brain Research, Vol. 129 0 2000 Elsevier Science B.V. All rights resewed

CHAPTER 35

Cellular and molecular mechanisms of opioid action SusanL. Ingram * Vellum Institute,

Oregon

Introduction: controversial in chronic pain

Health

Sciences

University

3181 SW Sam Jackson

aspects of opioid use

The effective treatment of chronic pain is an intriguing problem for many physicians. Although opioid administration is an effective treatment and is the treatment of choice for acute pain, opioids are ineffective in many instances of chronic pain (reviewed in: Dellemijn, 1999; Savage, 1999). The reasons for this discrepancy are not understood but may be a product of our poor understanding of the cellular and molecular mechanisms of both chronic pain and opioid analgesia. Endogenous opioids act at every level of the pain system from the primary afferent terminals to synapses in brain processing centers, and we are just beginning to understand the multiple mechanisms by which opioids modulate cellular activity. This review will not address all of the known actions of opioids (for a recent review see Akil et al., 1998), but will highlight some new data on the molecular mechanisms of opioid action in the brainstem periaqueductal gray area (PAG). The PAG The PAG is an organized processing center for the integration of noxious, threatening and stressful stimuli and the production of defense responses (reviewed in

* Corresponding author: S.L. Ingram, Vollum Institute, Oregon Health SciencesUniversity, 3 181 SW Sam Jackson Park Road, Portland OR 97201, USA. Tel.: +1-503494-672 1; Fax: + I-503-494-8230; E-mail: [email protected]

Park Road,

Portland,

OR 97201,

USA

Bandler and Shipley, 1994). It is abundant in opioid peptides and receptors (Atweh and Kuhar, 1977) and is an important site in the descending control of nociception (Basbaum and Fields, 1984; Fields et al., 1988). The PAG is the most effective site of stimulation-produced analgesia (Mayer and Liebeskind, 1974), suggesting a role for endogenous pain-relieving substances in this area (see also Pertovaara, 2000, this volume). The PAG is subdivided into functional columns that run in a rostrocaudal orientation (Bandler and Shipley, 1994). These columns comprise two distinct systems that modulate the flow of nociceptive information. The dorsal system consists of the dorsal and lateral columns of the PAG and the ventral system encompasses the ventrolateral columns. Various techniques have suggested that the ventral system is sensitive to opioids. Stimulation-produced analgesia of the ventral PAG but not the dorsal PAG is inhibited by naloxone (Gebhart and Toleikis, 1978; Cannon et al., 1982). In addition, microinjections of opioid agonists only produce antinociception when injected into the ventral PAG (Yaksh et al., 1976; Lewis and Gebhart, 1977). Opioid-mediated analgesia in the PAG is dependent on a descending relay system through the rostroventral medial medulla (RVM) and subsequent projections onto nociceptive neurons in the dorsal horn of the spinal cord (Morgan et al., 1992). The disinhibition antinociception

model of descending

The most widely recognized model of descending antinociception is the disinhibition model of Fields

484

PAG

RVM

Spinal Cord PAG

Morphine

lntrathecal

Morphine

Systemic

Morphine

Fig. 1. Opioids act at many levels of the descending pain pathway. Schematic illustration of some of the possible connections linking opioid-sensitive circuitry in the PAG and spinal cord with ON-cells in RVM. Open circles indicate excitatory and closed circles inhibitory connections. Dashed line indicates that a particular pathway may be multisynaptic. PAG morphine: morphine microinjected into the PAG appears to activate an inhibitory input to the ON-cell either directly or indirectly. Intrathecal morphine: morphine given intrathecally appears to have a disfacilitating, but not inhibitory action on ON-cells. Systemic morphine: when given systemically, morphine would act at both the PAG and spinal cord, in addition to having a direct effect upon ON-cells in the RVM. (Adapted from Neuroscience 47(4): 863-87 1, Morgan et al. (1992) with permission from Elsevier Science.)

and colleagues (Fig. 1) (Basbaum and Fields, 1978; Basbaum and Fields, 1984). This model is based on observations that opioid suppression of tonic inhibitory inputs from GABA interneurons is a major mechanism of opioid-induced excitation in many areas of the central nervous system (Nicoll et al., 1980; Fields et al., 1983; Cohen et al., 1992). Excitation of neurons in the nucleus raphe magnus (NRM) is responsible for the ultimate inhibition of nociceptive processing in the spinal cord (Behbehani and Fields, 1979; see also Pertovaara, 2000, this volume). Evidence for disinhibition hypothesis in PAG

There are substantial indirect data that support the disinhibition hypothesis in both the PAG and the NRM. Opioid receptors and peptides are abundant in the PAG and the NRM (Atweh and Kuhar, 1977). Enkephalin terminals form synaptic contacts with GABA neurons in the PAG (Wang et al., 1994), and there are abundant connections between GABA cells and PAG neurons that project to the NRM (Reichling and Basbaum, 1990). In addition, microinjections of

GABA* agonists into the PAG inhibit and GABA antagonists potentiate morphine-induced analgesia (Zambotti et al., 1982; Moreau and Fields, 1986; Depaulis et al., 1987). More direct evidence for the disinhibition hypothesis in the PAG is provided by a series of studies by Osborne et al. (1996). The majority of PAG-RVM output neurons identified with retrograde labeling were unaffected by opioid application, suggesting that opioids do not act on PAG output neurons. These data, in conjunction with studies that show that l.r-opioids directly hyperpolarize a subpopulation of PAG neurons (Behbehani et al., 1990; Chieng and Christie, 1994a) and inhibit GABA-mediated synaptic potentials (Chieng and Christie, 1994b), indicate that p-opioids primarily activate the descending antinociceptive pathway through disinhibition of PAG output neurons. However, since opioids also hyperpolarize a small percentage of PAG-NRM output neurons (Osborne et al., 1996) and inhibit glutamate synaptic potentials (Chieng and Christie, 1994b), disinhibition may not be the only mechanism by which opioids modulate this system.

485

Classic mechanisms of opioid action Opioids inhibit nociceptive transmission at many different levels of the ascending and descending pain pathways. Activation of opioid receptors results in three classical actions: inhibition of Ca2+ channels, inhibition of adenylyl cyclase and membrane hyperpolarization via activation of K+ channels (for review, see: Duggan and North, 1984; Grudt and Williams, 1995). Opioid agonists were first shown to inhibit voltage-gated Ca2+ currents in primary afferent neurons (Gross and Macdonald, 1987; Schroeder et al., 1991) and may be an important mechanism in opioid inhibition of peripheral pain (see also Sutherland et al., 2000, this volume). p-Opioids also inhibit voltagedependent Ca 2+ channels (see also Ito et al., 2000, this volume) in a subpopulation of PAG neurons (Kim et al., 1997). This contrasts with GABAn and nociceptin-mediated inhibition of Ca2+ channels in all PAG neurons (Connor and Christie, 1998). Therefore, it appears that p-opioid inhibition of Ca2+ channels may play a selective role in certain PAG pathways. Opioid inhibition of adenylyl cyclase has been studied in detail in neurons and cell lines. It is commonly used as a biochemical tool to study opioid receptor activation and the development of tolerance and dependence associated with chronic administration of opioids (Collier and Francis, 1975; Sharma et al., 1975, 1977; Law et al., 1982). Adenylyl cyclase in PAG neurons is inhibited by both p- and &opioid agonists (Noble and Cox, 1996), but the functional significance of this inhibition is not known. However, there is mounting evidence that adenylyl cyclase and protein kinase A are involved in the regulation of neurotransmitter release (see section on synaptic plasticity below). Opioid-induced hyperpolarization occurs via Gi-mediated activation of inwardly rectifying K+ channels (Kta) in many areas of the central nervous system, such as the locus coeruleus (Williams et al., 1982), substantia nigra (Lacey et al., 1988), ventral tegmental area (Johnson and North, 1992), and PAG (Chieng and Christie, 1994a). More recently, it has been determined that opioids also activate other voltage-sensitive potassium channels (Wimpey and Chavkin, 1991).

Opioid inhibition neurotransmitter

of presynaptic release

Although many studies have shown that opioids inhibit presynaptic GABA release, only a couple of studies have attempted to address the underlying mechanism (Capogna et al., 1993; Vaughan et al., 1997). Blockers of Kra and Ca 2+ channels do not block opioid inhibition of neurotransmitter release. On the basis of these results, it was postulated that opioids inhibit presynaptic release by a direct action on the exocytotic release process (Capogna et al., 1993). However, a study in the PAG determined that opioids activate a voltage-sensitive potassium channel (K”) in addition to Kia that is preferentially sensitive to blockade by 4-AP and dendrotoxin (Vaughan et al., 1997). Interestingly, these blockers did not affect opioid inhibition of glutamate release or baclofen-mediated inhibition of GABA release, suggesting that this K, channel may be selectively involved in opioid inhibition of GABA release. It is uncertain how a potassium channel regulates synaptic release, but it is probably through hyperpolarization of the terminal membrane. Many second-messenger systems couple G-protein activation to potassium channels. Protein kinase A and protein kinase C stimulate the release of GABA and glutamate in the PAG (Vaughan et al., 1997; see also Malmberg, 2000, this volPhos holipids ‘i

0

Arachidonic

Acid

\

Prostaglandins

Fig. 2. Schematic of arachidonic acid metabolism. Crosses indicate inhibitors that block opioid inhibition of GABA release in the PAG. (Reprinted by permission from Nature, 390: 611-614 Vaughan et al. (1997) Macmillan Magazines Ltd.)

486

ume). However, inhibitors of PKA and PKC do not block opioid inhibition of GABA release, suggesting that these pathways do not play a role in the opioid transduction mechanism. More recently it has been demonstrated that opioid receptors couple to additional signaling cascades such as the arachidonic acid (AA) cascade (Fukuda et al., 1996) and the extracellular signal regulated kinase (ERK) cascade (Fukuda et al., 1996; Li and Chang, 1996). PLA2 inhibitors, which inhibit the production of AA, block CL-opioid inhibition of GABA release in the PAG (Fig. 2) (Vaughan et al., 1997). In addition, AA and the 1Zlipoxygenase metabolites of AA (12-HETE and hepoxilin A3), but not metabolites of the 5lipoxygenase, mimic

a.

NSAIDs In~amtnatoty

the action of opioids. These results suggest that 12-lipoxygenase metabolites activate K, channels involved in opioid inhibition of GABA release from PAG terminals. Opioid/NSAID

synergy

Peripheral mechanisms

Opioids and non-steroidal anti-inflammatory drugs (NSAIDs) are used to treat pain resulting from inflammation and hyperalgesia in peripheral tissues. NSAIDs and opioids produce a greater analgesia when prescribed together. The synergistic effect in peripheral tissues is thought to result from actions

b.

OPIOIDS

PGE2

OPIOIDS

ATP Primary ufferent

CAMP 0

Nerve terminal+

n

n”

/

“ivINHIBITION EXCITATION Fig. 3. (a) A peripheral mechanism of opioid, prostanoid and NSAID interaction in primary afferent nerves via antagonistic actions on adenylyl cyclase (AC). Prostanoids (PGE2 and PG12) formed via cyclooxygenase (COX) act on prostanoid receptors (EP) to stimulate AC. The CAMP formed shifts the activation of the hyperpolarization-activated cation current, Ih. increasing excitability and action potential frequency of the nerve. Under these, but not basal conditions, opioids acting on u-receptors functionally antagonize this pathway by reducing AC activity. NSAIDs interact with this pathway by inhibiting the formation of prostanoids. (b) A central mechanism of interaction between opioids and NSAIDs in nerve terminals in the PAG that is independent of formation of prostanoids. Opioids stimulating u-receptors inhibit neurotransmitter release by stimulating phospholipase AZ (PI&). This leads to the formation of 1Zlipoxygenase (12-LOX) metabolites of arachidonic acid (AA) that enhance the activity of voltage-sensitive potassium channels (KY) that inhibit neurotransmitter release. NSAIDs and specific 5lipoxygenase (5.LOX) inhibitors block alternative pathways of AA metabolism, causing a shunt to enhance formation of 12-LOX metabolites, further enhancing the efficacy of opioids. This mechanism can account for opioid and NSAID synergism and the naloxone-sensitive analgesic actions of NSAIDs in the central nervous system. (Reprinted from ZnjIamm. Res., 48: l-4 Christie et al. (1999) with permission from Birkhauser Verlag AG.)

487 on a common signal transduction pathway between NSAIDs and opioids (Dahl and Kehlet, 1991; Malmberg and Yaksh, 1993; Maves et al., 1994), although other mechanisms have been proposed (McCormack, 1994). Prostanoids exacerbate inflammation in part by stimulating adenylyl cyclase activity (Levine and Taiwo, 1989). Therefore, NSAIDs, which inhibit the synthesis of prostanoids, and opioids, which inhibit adenylyl cyclase, functionally produce the same cellular response. The mechanisms by which changes in CAMP concentration modulate the activity of nociceptive neurons are not entirely understood. One such mechanism may be that ion channels are intracellular sites of convergence between prostanoid and opioid receptor transduction systems in primary afferents (Fig. 3a; Ingram and Williams, 1994). The CAMP-dependent hyperpolarization-activated cation channel (Ii,) is important in many cell types as a regulator of neuronal excitability (Brown and DiFrancesco, 1980; Bobker and Williams, 1989). Under normal conditions, opioids have no effect on Ii, activation. However, PGE2 stimulation of Ii, is inhibited by p-opioids, suggesting that opioids only modulate the channel under conditions of inflammation and augmented primary afferent activity (Ingram and Williams, 1994).

Central mechanisms Opioid/NSAID synergy via modulation of adenylyl cyclase does not appear to be pertinent in the central nervous system where NSAIDs produce analgesia in the absence of inflammation and prostanoid release (Bjorkman et al., 1992; Malmberg and Yaksh, 1993; Tortorici et al., 1996). Microinjections of NSAIDs into the PAG produce analgesia on their own (Tortorici and Vanegas, 1995) and potentiate the effects of morphine (Maves et al., 1994). Although there are a modest number of PGE2 binding sites in the PAG, the opioid receptor antagonist, naloxone, attenuates analgesia produced by NSAIDs microinjected into the PAG and RVM (Bjorkman et al., 1992; Tortorici et al., 1996), indicating that NSAIDs may facilitate an opioid transduction pathway. This additional mechanism for a synergistic interaction between opioids and NSAIDs has recently been described. As AA is metabolized by three enzymatic pathways (5lipoxygenases, 12lipoxygenases and cyclooxygenases), it was postulated that blockade of 5-lipoxygenase and cyclooxygenase might shunt AA down the 1Zlipoxygenase pathway to increase the potency of opioid inhibition (Fig. 3b; Vaughan et al., 1997). Indeed, coapplication of morphine (a weak partial agonist in the PAG) and the cyclooxyge-

b 60$T E $ 40r 5 I g 20fi k E I

0

I

10 Time (min)

I

I

20

30

Or 1

I 10

I

I 1000

1 10000

Cb”R, (nM) Fig. 4. Potentiation of opioid inhibition of GABAergic synaptic transmission by cyclooxygenase and 5-lipoxygenase blockade. (a) Time course of miniature inhibitory postsynaptic current (mZPSC) rate during superfusion of morphine (10 PM), plus indomethacin (10 FM), then indomethacin plus naloxone (1 FM). (b) Concentration-response relationship for percentage mIPSC rate inhibition by DAMGO (a selective p-receptor agonist) in the absence (open circles) and presence (closed circles) of indomethacin (10 ELM); and by morphine in the absence (open squares) and presence of indomethacin (closed squares, 10 FM), aspirin (closed diamonds, 30 kM), caffeic acid (closed triangles, 30 FM). Each point shows the mean (k s.e.m.) of responses of 4-6 different neurons. (Reprinted by permission from Nature, 390: 611-614 Vaughan et al. (1997) Macmillan Magazines Ltd.)

488

nase inhibitors, indomethacin or aspirin, produce a larger inhibition together than when applied separately (Fig. 4a). Furthermore, indomethacin produces a three-fold increase in the potency of the p-receptor agonist DAMGO (Fig. 4b). Selective cyclooxygenase inhibitors determined that the cyclooxygenase 1 subtype (not cyclooxygenase 2) mediate the synergistic actions of opioids and NSAIDs in the midbrain PAG (Vaughan, 1998). Similar effects were also seen with inhibitors of the 5lipoxygenase pathways. Thus, there appear to be at least two mechanisms by which opioids and NSAIDs provide synergistic analgesia. The first is via adenylyl cyclase inhibition, a common signal transduction pathway for opioid and prostanoid receptors. The second is via shunting AA metabolism through the 12-lipoxygenase pathway, potentiating opioid activation of a voltage-dependent potassium channel. It is not known whether these mechanisms occur in other tissues. NSAIDs also increase the analgesic potency of opioids in the dorsal horn of the spinal cord (Malmberg and Yaksh, 1993), suggesting that a mechanism similar to that shown in Fig. 3a may be evident in the spinal cord. However, naloxone-sensitive NSAID analgesia has also been observed in the spinal cord, suggesting that the mechanism shown in Fig. 3b may be involved. Either way, evidence is mounting that NSAIDs, specifically cyclooxygenase- 1 inhibitors, CHRONIC OPIOIDS

L

;,

used in conjunction with opioids could be used to treat different types of pain more effectively. It will be interesting to determine if 5lipoxygenase inhibitors are also effective analgesic agents. Synaptic plasticity in the opioid system: tolerance and dependence Chronic morphine treatment produces long-term changes (i.e. tolerance and dependence) in the opioid system, indicating that opioid-sensitive terminals in the brain are very plastic. Opioid tolerance is characterized as a diminished responsiveness to the inhibitory actions of opioids and is thought to involve the functional uncoupling between the opioid receptors and their effecters. Opioid dependence is characterized by withdrawal behaviors or rebound responses after administration of an opiate antagonist and cannot be explained by opioid receptor/effecter uncoupling. Withdrawal rebound results in increased neuronal excitability in several brain areas (Fry et al., 1980; Johnson and Duggan, 1981) and an increase in adenylyl cyclase activity (Avidor-Reiss et al., 1997). Functional and biochemical studies have suggested a role for the PAG in the expression of many withdrawal signs (reviewed in Christie et al., 1997).

ACUTE OPlOlDS

PLA2 y@(&+ I

9

I

0

a-API UDTX

I

\

I

PKA

I \ STIMULATES

RELEASE

I

J INHIBITS

RELEASE

I

Fig. 5. A schematic of synaptic plasticity in a GABA nerve terminal in the PAG. Acute administration of opioids act on u-receptors to activate phospholipase A2 (PUS) that in turn activates a 4-AP and dendrotoxin-sensitive potassium channel to inhibit the release of GABA from the terminal. Chronic administration of opioids results in increased efficacy of opioid inhibition of adenylyl cyclase such that removal of opioid inhibition of AC (withdrawal) causes a rebound excitation of AC activity and stimulation of protein kinase A (PKA) that stimulates the release of GABA. (Adapted from and reprinted by permission from J. Neumsci., 18: 10269-10276 (1998))

489

Synaptic plasticity in PAG

Opioid efficacy for inhibition of GABA release (both spontaneous and evoked GABA release) in the PAG is increased after chronic administration of opioids (Ingram et al., 1998). Removal of the opioids or administration of naloxone results in a rebound increase in GABA release that is not sensitive to 4-AP and dendrotoxin blockade of K, channels. These results suggest that acute and chronic effects of opioids are mediated by different transduction systems. Inhibitors of PKA block and CAMP analogs occlude the stimulatory effects of naloxone on GABA release, suggesting that the observed opioid supersensitivity is primarily caused by enhanced cAMP/PKA regulation of GABA release in the PAG of dependent animals (Fig. 5). Chronic opioid-induced adaptations in adenylyl cyclase signaling systems have been known for a long time (Sharma et al., 1975; Nestler and Tallman, 1987; Avidor-Reiss et al., 1997), but the physiological significance of this signaling system has remained elusive (Christie et al., 1997). These recent results might explain how hypertrophy of the adenylyl cyclase system mediates withdrawal behaviors elicited from the PAG. After chronic opioid treatment, opioid supersensitivity would maintain disinhibition of descending PAG output neurons, alleviating pain. However, removal of the opioids or application of an opioid receptor antagonist would increase GABA release and inhibit the descending antinociceptive pathways, producing withdrawal behaviors. Thus, synaptic plasticity in the GABA system of the PAG may play a major role in the modulation of nociceptive input with prolonged opioid use. These mechanisms may also be responsible for withdrawal rebound and sensitization in other opioid-sensitive nerve terminals in areas such as the ventral tegmental area (Bonci and Williams, 1996, 1997; Manzoni and Williams, 1999) and the nucleus accumbens (Chieng and Williams, 1998). Concluding remarks The development of secondary hyperalgesia and sensitization after peripheral injury and inflammation is thought to occur through changes in supraspinal systems (reviewed in Urban and Gebhart, 1999). One

recent study has determined that chronic inflammation activates neurons in the NRM in such a way that potentiates the effects of opioids on the descending antinociceptive pathway (Hurley and Hammond, 2000). Although previous studies determined that peripheral inflammation can increase opioid levels (Williams et al., 1995) and K-opioid receptor binding (Millan et al., 1987) in the PAG, the functional significance of these changes to the mechanisms of opioid analgesia is not understood. One intriguing possibility is that peripheral inflammation or other types of chronic pain alter the signal transduction pathways that regulate neurotransmitter release, contributing to secondary hyperalgesia and sensitization. These changes may explain why some chronic pain syndromes are impervious to inhibition by typical opioid therapies. References Akil, H., Owens, C., Gutstein, H., Taylor, L., Curran, E. and Watson, S. (1998) Endogenous opioids: overview and current issues. Drug Alcohol Depend., 51: 127-140. Atweh, S.F. and Kuhar, M.J. (1977) Autoradiographic localization of opiate receptors in rat brain, II. The brain stem. Bruin Rex, 129: 1-12. Avidor-Reiss, T., Nevo, I., Saya, D., Bayewitch, M. and Vogel, Z. (1997) Opiate-induced adenylyl cyclase superactivation is isozyme-specific. J. Biol. Chem., 272: 5040-5047. Bandler, R. and Shipley, M.T. (1994) Columnar organization in the midbrain periaqueductal gray: modules for emotional expression?. Trends Neurosci., 17: 379-389. Basbaum, AI. and Fields, H.L. (1978) Endogenous pain control mechanisms: review and hypothesis. Ann. Neurol., 4: 451462. Basbaum, AI. and Fields, H.L. (1984) Endogenous pain control systems: brainstem spinal pathways and endorphin circuitry. Annu. Rev. Neurosci., 7: 309-338. Behbehani, M.M. and Fields, H.L. (1979) Evidence that an excitatory connection between the periaqueductal gray and nucleus raphe magnus mediates stimulation produced analgesia. Brain Res., 170: 85-93. Behbehani, M.M., Jiang, M. and Chandler, S.D. (1990) The effect of [MET]enkephalin on the periaqueductal gray neurons of the rat: an in vitro study. Neuroscience, 38: 373-380. Bjorkman, R.L., Hedner, T., Hallman, K.M., Henning, M. and Hedner, J. (1992) Localization of the central antinociceptive effects of diclofenac in the rat. Brain Res., 590: 66-73. Bobker, D.H. and Williams, J.T. (1989) Serotonin augments the cationic current Is in central neurons. Neuron, 2: 1535-1540. Bonci, A. and Williams, J.T. (1996) A common mechanism mediates long-term changes in synaptic transmission after chronic cocaine and morphine. Neuron, 16: 631-639.

490

Bonci, A. and Williams, J.T. (1997) Increased probability of GABA release links adenylyl cyclase and chronic morphine treatment. .I. Neurosci., 17: 7966803. Brown, H. and DiFrancesco, D. (1980) Voltage-clamp investigations of membrane currents underlying pace-make activity in rabbit sino-atrial node. J. Physiol. (Lo&.), 308: 331-35 1. Cannon, J.T., Prieto, G.J., Lee, A. and Liebeskind, J.C. (1982) Evidence for opioid and non-opioid forms of stimulation-produced analgesia in the rat. Brain Rex, 243: 315-321. Capogna, M., Gahwiler, B.H. and Thompson, SM. (1993) Mechanism of u-opioid receptor-mediated presynaptic inhibition in the rat hippocampus in vitro. J. Physio2. (Land.), 470: 539558. Chieng, B. and Christie, M.J. (1994a) Hyperpolarization by opioids acting on mu-receptors of a sub-population of rat periaqueductal gray neurones in vitro. BI: J. Phnrmacol., 113: 121-128. Chieng, B. and Christie, M.J. (1994b) Inhibition by opioids acting on u-receptors of GABAergic and glutamatergic postsynaptic potentials in single rat periaqueductal gray neurones in vitro. BI: J. Pharmacol., 113: 303-309. Chieng, B. and Williams, J.T. (1998) Increased opioid inhibition of GABA release in nucleus accumbens during morphine withdrawal, J. Neurosci., 18: 7033-7039. Christie, M.J., Williams, J.T., Osborne, P.B. and Bellchambers, C.E. (1997) Where is the locus in opioid withdrawal?. Trends Pharmacol. Sk., 18: 134-140. Christie, M.J., Vaughan, C.W. and Ingram, S.L. (1999) Opioids, NSAIDs and 5-lipoxygenase inhibitors act synergistically in brain via arachidonic acid metabolism. 1n&mm. Res., 48: l-4. Cohen, G.A., Doze, V.A. and Madison, D.V. (1992) Opioid inhibition of GABA release from presynaptic terminals of rat hippocampal interneurons. Neuron, 9: 325-335. Collier, H.O. and Francis, D.L. (1975) Morphine abstinence is associated with increased brain cyclic AMP. Nature, 255: 159162. Connor, M. and Christie, M.J. (1998) Modulation of Ca2+ channel currents of acutely dissociated rat periaqueductal grey neurons. J. Physiol. (Land.), 509: 47-58. Dahl, J.B. and Kehlet, H. (1991) Non-steroidal anti-inflammatory drugs: rationale for use in severe postoperative pain. Br J. Anaesth., 66: 703-7 12. Dellemijn, P (1999) Are opioids effective in relieving neuropathic pain?. Pain, 80: 453-462. Depaulis, A., Morgan, M.M. and Liebeskind, J.C. (1987) GABAergic modulation of the analgesic effects of morphine microinjected in the ventral periaqueductal gray matter of the rat. Bruin Rex, 436: 223-228. Duggan, A.W. and North, R.A. (1984) Electrophysiology of opioids. Phamtacol. Rev., 35: 219-281. Fields, H.L., Vanegas, H., Hentall, I.D. and Zorman, G. (1983) Evidence that disinhibition of brain stem neurones contributes to morphine analgesia. Nature, 306: 684-686. Fields, H.L., Barbaro, N.M. and Heinricher, M.M. (1988) Brain stem neuronal circuitry underlying the antinociceptive action of opiates. In: H.L. Fields and J.-M. Besson (Eds.), Pain

Modulation. Progress in Brain Research, Vol. 77. Elsevier, Amsterdam, pp. 245-257. Fry, J.P., Herz, A. and Zieglgansberger, W. (1980) A demonstration of naloxone-precipitated opiate withdrawal on single neurones in the morphine-tolerant dependent rat brain. Br J. Phannacol., 68: 585-592. Fukuda, K., Kato, S., Morikawa, H., Shoda, T. and Mori, K. (1996) Functional coupling of the 6-, n-. and tc-opioid receptors to mitogen-activated protein kinase and arachidonate release in Chinese hamster ovary cells. J: Neurochem., 67: 1309-1316. Gebhart, G.F. and Toleikis, J.R. (1978) An evaluation of stimulation-produced analgesia in the cat. Exp. Neural., 62: 570579. Gross, R.A. and Macdonald, R.L. (1987) Dynorphin A selectively reduces a large transient (N-type) calcium current of mouse dorsal root ganglion neurons in cell culture. Proc. Nutl. Acad. Sci. USA, 84: 5469-5473. Grudt, T.J. and Williams, J.T. (1995) Opioid receptors and the regulation of ion conductances. Rev. Neurosci., 6: 279-286. Hurley, R.W. and Hammond, D.L. (2000) The analgesic effects of supraspinal n and delta opioid receptor agonists are potentiated during persistent inflammation. J. Neurosci., 20: 12491259. Ingram, S.L. and Williams, J.T. (1994) Opioid inhibition of Ih via adenylyl cyclase. Neuron, 13: 179- 186. Ingram, S.L., Vaughan, C.W., Bagley, E.E., Connor, M. and Christie, M.J. (1998) Enhanced opioid efficacy in opioid dependence is caused by an altered signal transduction pathway. .I. Neurosci., 18: 10269-10276. Ito, S., Okuda-Ashitaka, E., Imanishi, T. and Minami, T. (2000) Central roles of nociceptin/orphanin FQ and nocistatin: allodynia as a model of neural plasticity. In: J. Sandktihler, B. Bromm and G.F. Gebhart (Eds.), Nervous System Plasticity and Chronic Pain. Progress in Brain Research, Vol. 129. Elsevier, Amsterdam, pp. 205-218. Johnson, S.M. and Duggan, A.W. (1981) Evidence that the opiate receptors of the substantia gelatinosa contribute to the depression, by intravenous morphine, of the spinal transmission of impulses in unmyelinated primary afferents. Brain Res., 207: 223-228. Johnson, S.W. and North, R.A. (1992) Opioids excite dopamine neurons by hyperpolarization of local interneurons. J. Neurosci., 12: 483-488. Kim, C.-J., Rhee, J.-S. and Akaike, N. (1997) Modulation of high-voltage activated Ca 2+ channels in the rat periaqueductal gray neurons by u-type opioid agonist. J. Neurophysiol., 77: 1418-1424. Lacey, M.G., Mercuri, N.B. and North, R.A. (1988) On the potassium conductance increase activated by GABAu and dopamine D2 receptors in rat substantia nigra neurones. J. Physiol. (L.ond.), 401: 437-453. Law, PY., Horn, D.S. and Loh, H.H. (1982) Loss of opiate receptor activity in neuroblastoma x glioma NG108-15 hybrid cells after chronic opiate treatment. Mol. Pharmacol., 22: 1-4. Levine, J.D. and Taiwo, Y.O. (1989) Involvement of the mu-opi-

491

ate receptor in peripheral analgesia. Neuroscience, 32: 571575. Lewis, V.A. and Gebhart, G.F. (1977) Evaluation of the periaqueductal central gray (PAG) as a morphine-specific locus of action and examination of morphine-induced and stimulationproduced analgesia at coincident PAG loci. Bruin Rex, 124: 283-303. Li, L.-Y. and Chang, K.-J. (1996) The stimulatory effect of opioids on mitogen-activated protein kinase in Chinese hamster ovary cells transfected to express u-opioid receptors. Mol. Pharmacol., 50: 599-602. Malmberg, A. (2000) Protein kinase subtypes involved in injury-induced nociception. In: J. Sandktihler, B. Bromm and G.F. Gebhart (Eds.), Nervous System Plasticity and Chronic Pain. Progress in Brain Research, Vol. 129. Elsevier, Amsterdam, pp. 51-59. Malmberg, A.B. and Yaksh, T.L. (1993) Pharmacology of the spinal action of ketorolac, morphine, ST-91, U50488H, and L-PIA on the formalin test and an isobolographic analysis of the NSAID interaction. Anesthesiology, 79: 270-28 1, Manzoni, O.J. and Williams, J.T. (1999) Presynaptic regulation of glutamate release in the ventral tegmental area during morphine withdrawal. J. Neurosci., 19: 6629-6636. Maves, T.J., Pechman, P.S., Meller, S.T. and Gebhart, G.F. (1994) Ketorolac potentiates morphine antinociception during visceral nociception in the rat. Anesthesiology, 80: 1094-l 101. Mayer, D.J. and Liebeskind, J.C. (1974) Pain reduction by focal electrical stimulation of the brain: an anatomical and behavioral analysis. Brain Rex, 68: 73-93. McCormack, K. (1994) Non-steroidal anti-inflammatory drugs and spinal nociceptive processing. Pain, 59: 9-43. Millan, M.J., Morris, B.J., Colpaert, EC. and Herz, A. (1987) A model of chronic pain in the rat: high-resolution neuroanatomical approach identifies alterations in multiple opioid systems in the periaqueductal grey. Bruin Res., 416: 349-353. Moreau, J.L. and Fields, H.L. (1986) Evidence for GABA involvement in midbrain control of medullary neurons that modulate nociceptive transmission. Bruin Res., 397: 37-46. Morgan, M.M., Heinricher, M.M. and Fields, H.L. (1992) Circuitry linking opioid-sensitive nociceptive modulatory systems in periaqueductal gray and spinal cord with rostra1 ventromedial medulla. Neuroscience, 47: 863-871. Nestler, E.J. and Tallman, J.F. (1987) Chronic morphine treatment increases cyclic AMP-dependent protein kinase activity in the rat locus coeruleus. Mol. Pharmacol., 33: 127-132. Nicoll, R.A., Alger, B.E. and Jahr, C.E. (1980) Enkephalin blocks inhibitory pathways in the vertebrate CNS. Nature, 287: 22-25. Noble, F. and Cox, B.M. (1996) Differential desensitization of uand 6-opioid receptors in selected neural pathways following chronic morphine treatment. BI: J. Pharmacol., 117: 161-169. Osborne, P.B., Vaughan, C.W., Wilson, H.I. and Christie, M.J. (1996) Opioid inhibition of rat periaqueductal grey neurones with identified projections to rostra1 ventromedial medulla in vitro. J. Physiol. (Land.), 490: 383-389. Pertovaara, A. (2000) Plasticity in descending pain modulatory systems. In: J. Sandktihler, B. Bromm and G.F. Gebhart (Eds.),

Nervous System Plasticity and Chronic Pain. Progress in Brain Research, Vol. 129. Elsevier, Amsterdam, pp. 231-242. Reichling, D.B. and Basbaum, A.I. (1990) Contribution of brainstem GABAergic circuitry to descending antinociceptive controls, II. Electron microscopic immunocytochemical evidence of GABAergic control over the projection from periaqueductal gray to the nucleus raphe magnus in the rat. .I. Camp. Neurol., 302: 378-393. Savage, S.R. (1999) Opioid use in the management of chronic pain. Med. Clin. North Am., 83: 761-786. Schroeder, J.E., Fischbach, P.S., Zheng, D. and McCleskey, E.W. (1991) Activation of mu opioid receptors inhibits transient high- and low-threshold Ca2+ currents, but spares a sustained current. Neuron, 6: 13-20. Sharma, S.K., Klee, W.A. and Nirenberg, M. (1975) Dual regulation of adenylate cyclase accounts for narcotic dependence and tolerance. Proc. Natl. Acad. Sci. USA, 72: 3092-3096. Sharma, S.K., Klee, W.A. and Nirenberg, M. (1977) Opiatedependent modulation of adenylate cyclase. Proc. Nutl. Acad. Sci. USA, 74: 3365-3369. Sutherland, S.P., Cook, S.P. and McCleskey, E.W. (2000) Chemical mediators of pain due to tissue damage and ischemia. In: J. Sandktihler, B. Bromm and G.F. Gebhart (Eds.), Nervous System Plasticity and Chronic Pain. Progress in Brain Research, Vol. 129. Elsevier, Amsterdam, pp. 21-38. Tortorici, V. and Vanegas, H. (1995) Anti-nociception induced by systemic or PAG-microinjected lysine-acetylsalicylate in rats. Effects on tail-flick related activity of medullary off- and on-cells. Eur .I. Neurosci., 7: 1857-1865. Tortorici, V., Vasquez, E. and Vanegas, H. (1996) Naloxone partial reversal of the antinociception produced by dipyrone microinjected into the periaqueductal gray of rats. Possible involvement of medullary off- and on-cells. Brain Res., 725: 106-l 10. Urban, M.O. and Gebhart, G.F. (1999) Central mechanisms in pain. Med. Clin. North Am., 83: 585-596. Vaughan, C.W. (1998) Enhancement of opioid inhibition of GABAergic synaptic transmission by cycle-oxygenase inhibitors in rat periaqueductal grey neurones. Br J. Pharmacol., 123: 1479-1481. Vaughan, C.W., Ingram, S.L., Connor, M.A. and Christie, M.J. (1997) How opioids inhibit GABA-mediated neurotransmission. Nature, 390: 611-614. Wang, Q.-P., Guan, J.-L. and Nakai, Y. (1994) Immunoelectron microscopy of enkephalinergic innervation of GABAergic neurons in the periaqueductal gray. Bruin Res., 665: 39-46. Williams, F.G., Mullet, M.A. and Beitz, A.J. (1995) Basal release of Met-enkephalin and neurotensin in the ventrolateral periaqueductal gray matter of the rat: a microdialysis study of antinociceptive circuits. Bruin Res., 690: 207-216. Williams, J.T., Egan, TM. and North, R.A. (1982) Enkephalin opens potassium channels on mammalian central neurones. Nature, 299: 74-77. Wimpey, T.L. and Chavkin, C. (1991) Opioids activate both an inward rectifier and a novel voltage-gated potassium conductance in the hippocampal formation. Neuron, 6: 28 l-289. Yaksh, T.L., Yeung, J.C. and Rudy, T.A. (1976) Systematic

492 examination in the rat of brain sites sensitive to the direct application of morphine: observation of differential effects within the periaqueductal gray. Bruin Rex, 114: 83-103. Zambotti, F., Zonta, N., Parenti, M., Tommasi, R., Vicentini,

L., Conci, E and Mantegazza, P. (1982) Periaqueductal gray matter involvement in the muscimol-induced decrease of morphine antinociception. Naunyn-Schmiedebergs Arch. Pharma col., 318: 368-369.

J. Sand!&iler, B. Bromm and GE Gebhart (Eds.) Progress in Brain Research, Vol. 129 0 2000 Elsevier Science B.V. All rights reserved

CHAPTER 36

Pre-emptive analgesia in postamputation pain: an update Troels StaehelinJensen1,3,*and Lone Nikolajsen 2,3 ’ Department

of Neurology,

2 Department

of Anesthesiology, and 3 Danish Pain Research DK-8000 Aarhus C. Denmark

Introduction A key issue in current pain research and treatment is the transition of acute pain into chronic pain states and the identification of risk factors for such transition. Based on a series of experimental studies on neuronal hyperexcitability and the induction of sensitisation, the question was raised in 1988 (Wall, 1988) whether neuronal hyperexcitability and pain could be prevented by an analgesic treatment before injury as opposed to after the injury. This issue, termed pre-emptive analgesia has been explored extensively both experimentally and clinically within the last decade (for review see: McQuay, 1992, 1994; Woolf and Chong, 1993; Dahl, 1994; Kalso, 1997). The nociceptive system is not a fixed static system, but a dynamic neuronal network, which continuously alters its response characteristics depending on the prior exposure to noxious activity (for review see: Coderre and Katz, 1997; Woolf and Salter, 2000; Berthele et al., 2000, this volume; Gerber et al., 2000, this volume; Moore et al., 2000, this volume; Sandktihler et al., 2000, this volume). In classical studies, Woolf and others observed that the functional plastic changes with sensitisation of second-order neurons could be prevented by a neural blockade or by an opioid administration before, but not after injury (for review see Woolf, 1992). The

* Corresponding author: T.S. Jensen, Department of Neurology, Aarhus University Hospital, DK-8000 Aarhus C, Denmark. Tel.: +45-8949-3283; Fax: +45-8949-3300; E-mail: tsj @panter.soci.aau.clk

Cente,:

Aarhus

University

Hospital,

clinical counterpart of such pre-emptive analgesia was initiated by a study carried out by Bach and colleagues in 1988 (Bach et al., 1988) in a group of amputees, where epidural treatment before amputation apparently reduced the incidence of phantom pain. This important landmark has since been the focus for a series of further experimental and clinical studies on the effect of blockade of noxious input to the nervous system and its possible influence on neuronal hyperexcitability, sensitisation and chronic pain. Some of these studies but not all support the idea of a pre-emptive effect in amputees to avoid the subsequent development of phantom pain. Although amputation only represents a small minority among patients undergoing surgery it is extremely important to clarify if phantom pain indeed can be prevented because such an observation will have a significant bearing not only for amputees but for future strategies in anaesthesiology and pain treatment. Several studies have dealt with other types of postoperative pain (for review see: Dahl, 1994; McQuay, 1994). In the present chapter we will examine the experimental and clinical evidence for a pre-emptive analgesic effect on nerve injury pain with emphasis on phantom and stump pains as consequences of limb amputation. Clinical aspects of postamputation

pain

A detailed description of postamputation phenomena can be found elsewhere. For the present purposes a few clinical highlights, however, are necessary. It is useful to distinguish between the following three elements of the phantom complex (Jensen and Niko-

494

lajsen, 1999): phantom limbpain: pain referred to the missing limb; phantom limb sensation: non-painful sensations referred to the missing limb; stump pain: pain localised to the stump. Although these phenomena can be separated and patients usually are able to distinguish between each they often coexist in the same patient.

of pain is considerably reduced with time (Jensen et al., 1983; Houghton et al., 1994; Nikolajsen et al., 1997a). Character and location

Phantom pain develops in about 70% of patients following amputation irrespective of the cause of amputation (Jensen and Nikolajsen, 1999). Gradual remission of pain over time is the rule, but for a significant group of amputees (approximately 10%) the pain persists in a severe form requiring treatment. Treatment of phantom pain, however, is difficult and there is at present no single therapy offering universal benefit. Most recent studies agree that the incidence of phantom pain is in the range of 60-80% (Houghton et al., 1994; Nikolajsen et al., 1997a; Wartan et al., 1997). The incidence does not seem to be influenced by age, gender, side or level of amputation (Jensen et al., 1983) or cause of amputation (Sherman and Sherman, 1983; Houghton et al., 1994), but phantom pain is less frequent in young children, and in congenital amputees (Flor et al., 1998; Wilkins et al., 1998).

Phantom pain is highly variable from patient to patient with some patients having only a few short-lasting spells of pain sensation and others suffering from daily excruciating pains lasting for months or years. However, only few patients are in constant pain. The description of pains varies considerably from shooting, stabbing, pricking, boring, squeezing pains to throbbing and burning types of pains (Jensen et al., 1983; Sherman and Sherman, 1983; Nikolajsen and Jensen, 2000). Phantom pains are most distinct in distal parts of the body (hand, fingers, foot, toes, etc.). A characteristic feature is a gradual shrinkage of the phantom as time passes by after amputation, a phenomenon known as telescoping. As for phantom pain character, the location also varies from patient to patient. It is important to take this variability both in character, intensity and location into account in clinical studies and assessdifferent aspects (intensity, number of attacks, duration of attacks, etc.) and ensure constancy in cross-over trials and similarities between groups in parallel study trials.

Time course

Mechanisms

The onset of phantom pain is usually in the first week following amputation. In a prospective study, 48% developed their pain within the first 24 h, and 83% within 4 days. In less than 10% is phantom pain delayed for more than 1 week (Jensen et al., 1983). Another prospective study showed similar results (Nikolajsen et al., 1997a). However, case reports suggest that the onset of phantom pain may be delayed until several years after amputation. For example, Rajbhandari and colleagues (Rajbhandari et al., 1999) described the development of diabetic neuropathic pain in a leg amputated 44 years earlier. Although phantom pain may diminish with time and in some cases even disappear, prospective studies show that 2 years after amputation the incidence of postamputation pains are almost the same as at onset. However, both duration, frequency, and intensity

The mechanisms responsible for stump and phantom pain in amputation are still under discussion and apparently both peripheral and central mechanisms are involved and contribute to the phantom pain perception. An understanding of underlying mechanism for phantom pain and other types of nerve injury pain may lead to new types of rational founded treatments.

Frequency

Peripheral generator

Several lines of evidence indicate that the periphery is a generator and a site for maintaining stump and phantom pain. (1) Neuromas are universal phenomena after a nerve cut. An extensive experimental literature shows that these neuromas express spontaneous and

abnormal evoked activity with an increased sensitivity to a variety of stimuli including mechanical stimuli (Devor et al., 1993; Devor and Seltzer, 1999). The ectopic and increased activity from the periphery is assumed to be the result of a novel expression and distribution of ion channels (Novakovic et al., 1998). Also in the DRG there are major changes in the expression of sodium channels with a switch of one channel type into another (Waxman, 1999; Cummins et al., 2000, this volume). (2) It is a common clinical experience that percussion of the stump or stump neuromas induces stump and phantom pain. In a classical microneurographic study in two amputees, Nystrom and Hagbarth (198 1) showed that tapping of neuromas was associated with an increased activity in afferent C-fibres and an increased pain sensation. Consistent with this observation it has been shown that there is an inverse, albeit small correlation between phantom pain intensity and pressure pain threshold of the stump in amputees early after the amputation (Nikolajsen et al., 2000b). It should be noted that such correlation can only be found in the early period after amputation but not later in the course, indicating that other than peripheral factors are involved in generating phantom pain. (3) Perineuromal injection of gallamine, a drug that increases neuronal activity by facilitating sodium conductance, produces pain, while a sodium channel blocker, lidocaine, blocks phantom pain at the site of neuroma (Chabal et al., 1989). An extensive literature has shown that changes both in the periphery of the nerve and in dorsal root ganglion cells may be involved in generating phantom pain (for review see Devor and Seltzer, 1999). These findings suggest that long-lasting pain may induce secondary changes in the nervous system that may persist after a nerve injury and predict post-injury pain. This observation is in accordance with previous studies showing that long-lasting and intense preamputation pain may increase the risk for phantom pain following amputation (Nikolajsen et al., 1997a). Central sensitisation It is well appreciated that peripheral sensitisation occurs as a result of spontaneous activity from sprout-

ing, regenerating nerve endings. This peripheral sensitisation in turn gives rise to secondary changes in otherwise silenced small dorsal root ganglia (DRG) cells. The increased barrage from neuromas (organised sprouts) and from DRG cells following nerve injury eventually sweeps centrally and induces long-term changes in central projecting neurons. It is conceivable that in certain patients, where an early afferent barrage and hence sensitisation has been particularly intense, central mechanisms may subsequently maintain sensitisation without any additional nociceptor input. The central sensitisation in the spinal dorsal horn induced by an increased barrage form C-fibres may originate from two sources: (1) a sustained C-fibre input recruit N-methyl-D-aspartate (NMDA) receptors located on second-order spinal neurons (Woolf and Thompson, 1991; see also Moore et al., 2000, this volume; Sandktihler et al., 2000, this volume; Svendsen et al., 2000, this volume); (2) in other cases a central reorganisation occurs because of sprouting of Ap-fibres from laminae III and IV into lamina II, forming a functional contact with second-order nociceptive neurons (Woolf et al., 1992). In both cases the result is an evocation of pain by Afi-fibre input, e.g. touch. However, while the former type of touch-evoked pain may respond to agents that counteract the NMDA receptor system, the latter is unlikely to be affected by such agents and may perhaps be more resistant to modulation at all. It is at present unclear whether central sensitisation can be present without a peripheral generator. The above observation that sprouting from lamina III to lamina II does occur suggests that this may indeed be the case. The central aspects of the hyperexcitability include chemical changes of cells and cell membranes in the dorsal horn of the spinal cord. The glutamate release cascade with release of glutamate, activation of NMDA receptors, increase of intracellular Ca2+ (see Gerber et al., 2000, this volume) and activation of nitrogen oxide synthase (see, however, Hoheisel and Mense, 2000, this volume) is particular important in this hyperexcitability. The clinical translation of such hyperexcitability phenomena may include expansion of painful areas and sensory abnormalities, a lowering of threshold to evoke noxious activ-

496 ity, a build of pain following repetitive stimulation, e.g. wind-up-like phenomena, after-sensations and a switch in modality to evoke pain (Jensen and Gottrup, 2000). While intravenous administered NMDA antagonists can block phantom pain and signs of sensitization oral administered NMDA antagonists are ineffective (Nikolajsen et al., 1996, 2000a). Cortical reorganisation The peripheral plasticity in the nervous system which sweeps into the dorsal horn of the spinal cord may subsequently project to more central relays in the thalamus and cortex and induce further changes at these sites. Recent magnetoencephalographic studies in amputees have shown a strong correlation between phantom limb pain and the extent of cortical reorganisation (Flor et al., 1995, 1998; Bromm et al., 2000, this volume; Casey, 2000, this volume; Flor, 2000, this volume). Currently, a study is carried out using the f-MRI technique to examine whether similar expansion of cortical representation occurs in those patients with pain as opposed to those without pain (M. Lindvig et al., 2000, pers. commun.). It has therefore been suggested that phantom limb pain could be the result of plastic changes in the somatosensory cortex (Larbig et al., 1996). Birbaumer et al. (1997) studied the effect of regional anaesthesia on cortical reorganisation in upper-limb amputees and found that a brachial plexus blockade abolished pain and reorganisation in three out of six amputees. Another aspect of central elements in phantom pain involves the concept of a neuromatrix in the brain proposed by Melzack (1995) in which a neuronal network in the brain produces a characteristic nerve-impulse pattern, which underlies the various body perceptions including phantom pains. So taken together, a cascade of events at several synaptic levels in the nervous system take place following a peripheral nerve injury and they may all contribute more or less to the development of pain. Biologically, a sustained noxious input to secondorder neurons may be expressed in several ways: cells expand their response repertoire to include noxious responding activity following non-noxious Ab-input, they show an increase in receptive fields, a reduction of firing threshold, after-discharges and the recruitment of silenced cells, etc. The clinical

counterpart of central sensitisation in terms of phantom pain is of course not clear, but it may include phenomena such as: spontaneous pain, increased pain following stimulation, tender neuromas, allodynia or hyperalgesia of the stump and invasion of such abnormal sensation to a distant dermatome. The latter phenomenon is occasionally seen in patients following repetitive surgery on stumps (Jensen and Nikolaj sen, 1999). Pre-emptive analgesia In trials on pre-emptive analgesia the effect of treatments (usually drugs) administered before injury are compared with the same type of treatment administered after injury. Studies of peri-operative analgesia usually compare pre-injury treatment with no treatment. An important point concerns the duration and the efficacy of the analgesic intervention. A shortlasting and incomplete analgesia is probably less effective than a long-lasting and profound nerve blockade in preventing sensitisation of noxious responding neurons upstream in the nervous system. There is a long list of studies that have examined the effect of a pre-emptive pain treatment in pain (see: Dahl and Kehlet, 1993; Dahl, 1994; McQuay, 1994). Experimental evidence for pre-emptive effect in neuropathic pain There is a long list supporting a pre-emptive analgesic effect in experimental pain including neuropathic pain (for review see Coderre and Katz, 1997; see also Wilder-Smith, 2000, this volume). The background for a pre-emptive effect is obtained from a series of classical studies in the mideighties by Woolf and others, demonstrating plastic changes in the nervous system following long-lasting and intense noxious activity (for review, see Woolf, 1992). See also: More et al. (2000) this volume. Peripheral injury or noxious stimulation either by capsaicin, thermal injury or repetitive C-fibre stimulation have all been shown to induce a sensitisation of spinal dorsal horn neurons. This sensitisation is manifested by a wide spectrum of peripheral and central phenomena. Dickenson and Sullivan (1987) observed that intrathecal administration of morphine in phase I of the formalin test inhibited dorsal horn

497

neuronal activity more than if administered after phase I. Similar effects were observed with the NMDA antagonist MK801 administered intrathetally (Yaksh, 1993), but subsequent studies have not unequivocally supported these findings. Brennan et al. (1997) in an incision model of the plantar hindpaw failed to find an effect of intrathecal opioids or local anaesthetics. In certain animal models of neuropathic pain it has been shown that noxious stimulation of the paw or the sciatic nerve in the rat before neurectomy significantly shortens the onset of autotomy and enhances its severity. An early literature in rodents showed that autotomy or hyperalgesia can be prevented if the initial nerve injury is preceded by an intense antinociceptive treatment. Several drugs have been examined, including local anaesthetics, clZadrenergic agonists, opioids and NMDA antagonists (Puke and Wiesenfeld-Hallin, 1993; Kalso, 1997). These studies have shown that local anaesthetics applied locally on the nerve before transection or constriction reduces hyperalgesia or other behavioural signs of pain. Similarly, morphine and NMDA antagonists given intrathecally or systemically have also been shown to prevent pain (for review see Kalso, 1997). Clinical evidence for prevention of postamputation pain At least three lines of evidence point to a possible role of pre-emptive pain treatment in amputees (Jensen and Nikolajsen, 1999). (1) Phantom pain is in many cases a replicate of pain experienced before amputation and patients with severe and long-lasting preamputation pain in the involved limb are more likely to develop postamputation pain. (2) Phantom pain represent the most radical example of neuropathic pain, a condition where prolonged noxious input impinging on central structures are known to induce long-term secondary changes. These plastic changes in the nervous system are experimentally amenable to modulation by pre-emptive measures (see above). (3) An early study in patients undergoing amputation has suggested that an effective pain treatment before amputation reduces the subsequent incidence

of phantom pain (Bach et al., 1988). This observation has been confirmed in some subsequent trials. The idea of a pre-emptive analgesic effect in postamputation pain was initiated by clinical observations that phantom pain in some cases is similar to previous experienced pain (Nathan, 1962; Jensen and Nikolajsen, 1999). This observation has led to the idea that preamputation pains may create an imprint in memorising structures of the central nervous system and that such imprint could be responsible for persistent pains after amputation. Some studies have shown that prior neuropathic pains have an influence on subsequent pains. Dworkin and colleagues in postherpetic neuralgia have noticed that severe postherpetic neuralgia often were preceded by severe zoster pains (Dworkin et al., 1992). Similarly, in amputees it has been shown that patients with severe phantom pain often have suffered from long-lasting and more severe preamputation pain than patients with less intense phantom pain (Jensen et al., 1985; Nikolajsen et al., 1997a). Preamputation pain seems to increase the risk of phantom pain after amputation, but the relation is not simple (Nikolajsen et al., 1997a). We prospectively followed 58 amputees (Jensen et al., 1983, 1985). Phantom pain was significantly more frequent after 8 days and 6 months, but not after 2 years in patients who had pain in the limb before the amputation compared to those who were free of pain. In a retrospective study by Houghton and colleagues (Houghton et al., 1994) similar results were found: in vascular amputees preamputation pain was related to phantom pain after 8 days, 6 months and 2 years, but in traumatic amputees preamputation pain was only related to phantom pain immediately after the amputation. In a prospective study of 60 amputees Nikolajsen et al. (1997a) found that intense preamputation pain was related to phantom pain after 1 week and 3 months but no relation was found later in the course. Several amputees presented statements as “I can still feel my ingrown toe-nail” and “I feel the hole, where the ulcer was” (Katz and Melzack, 1990; Hill et al., 1996; Nikolajsen et al., 1997a). Only few studies have examined the incidence with which preamputation pain persists as phantom pain. Katz and Melzack (1990) interviewed 68 amputees up to several years after the amputation. Fifty-seven per-

498

cent of those who reported having had preamputation pain before the amputation claimed that their phantom pain resembled the pain they had at the time of amputation. The incidence was much lower in two prospective studies (Jensen et al., 1983; Nikolajsen et al., 1997a). In the latter study patients were asked to describe their pain both before and after amputation using: (1) their own words, (2) specific pain descriptors, and (3) McGill Pain Questionnaire. Location of pain was also recorded. Although 42% of the patients claimed that their phantom pain was similar to the pain they had at the time of the amputation, actual similarity on the basis of (1) pain descriptors, (2) the patients’ own selection of pain words, and (3) on the basis of the McGill Pain Questionnaire were only found to be similar in a small minority of patients when comparing preamputation and postamputation recordings of pain. So although preamputation pain and a conditioning sensitisation of the nervous system play a role in phantom pain they are not the only mechanisms involved. Pre-emptive analgesic studies in postamputation pain While the issue of pre-emptive analgesia has been dealt with extensively by others (McQuay, 1992, 1994; Dahl, 1994; Kalso, 1997; see Wilder-Smith, 2000, this volume), its possible role in nerve injury has been less discussed. Peripheral nerve blockade

The effect of postoperative perineural analgesia to prevent phantom pain has been studied in a few trials (Table 1). Fischer and Meller (1991) introduced a catheter into the transected nerve sheath at the time of amputation and infused bupivacaine 0.25% at a rate of 10 ml/h for 72 h after amputation. None of the 11 patients studied developed phantom pain during the first year after amputation. In a retrospective study, Elizaga et al. (1994) found no difference in the incidence of phantom pain between two groups of patients who had received either bupivacaine 0.5% 2-6 ml/h for 72 h after amputation (9 patients) or opioid analgesics alone (12 patients). Pinzur and colleagues (Pinzur et al., 1996) conducted a randomised,

blinded and placebo-controlled trial in which 21 patients received either perineural infusion with bupivaCaine 0.5% or saline for three consecutive days after amputation. The incidence of phantom pain was similar in both groups 3 and 6 months after amputation. Epidural blockade

Bach et al. (1988) carried out the first study in which 25 patients scheduled for amputation of the lower limb were randomised by means of their year of birth to receive either epidural morphine, epidural bupivacaine 0.25% or both in combination for 3 days before the amputation (11 patients) or conventional analgesics such as opioids, paracetamol, dextropropoxyphene and acetylsalicylic acid (14 patients). All patients received epidural or spinal analgesia for the amputation and postoperatively their pain was treated with meperidine, paracetamol or ASA. Patients were interviewed about phantom pain after 1 week, 6 and 12 months. Pain was categorised as either present or not present and apparently interviewers were not blinded to the treatment. Six patients died during the follow-up period. The incidence of phantom pain was found reduced 6 months after amputation but not after 1 week or after 12 months in the epidural blocked group as compared to the non-blocked control group. At the 6-months follow-up, all 10 patients in the blockade group were free of pain while 5 out of 13 patients in the control group had phantom pain. In another study, Jahangiri et al. (1994) prospectively followed 24 patients undergoing limb amputation. In a non-randomised design patients received either an epidural infusion of bupivacaine, diamorphine and clonidine from 24 to 58 h before surgery and for at least 3 days after surgery (13 patients) or on demand opioid as analgesia (11 patients). Amputation was carried out under general anaesthesia. The presence of phantom pain was graded on a scale of l-10 and pain was considered significant when the score was >3. During follow-up two patients died. After 1 week, 6 and 12 months, the incidence of phantom pain was found significantly lower in the blocked group versus the unblocked control group. Schug and colleagues in a letter to the editor (Schug et al., 1994) presented data from a non-randornised trial. Methods of blinding and pain assess-

499 TABLE Effect

1 of perineural

or epidural

pre-emptive

analgesia

in amputees

Randomisation

Blinding

Bach et al., 1988

+

-

25

Jahangiri

-

-

-

-

Authors

Schug

et al., 1994 et al., 1994

Katsuly-Liapis

Nikolajsen

et al., 1996

et al., 1997a,b

Fischer

and Meller,

Elizaga

et al., 1994

Pinzur

1991

et al.. 1996

a B = blockade

group;

No. of patients

on phantom

pain

Treatment a

preoperative

intraoperative

B: I1 c: 14

+ -

+ +

24

B: 13 c: 11

+ -

+ -

+ -

+ -c

23

Bt: 8 B2: 7 C: 8

+

+ + -

+ + -

+ +

-

postoperative

Effect of treatment

+ +

Bl vsC

+

-

45

BI: 15 B2: 12 C: 18

+ + -

+ + Bt vsB2andC

+

+

60

B: 29 c: 31

+ +

-

-

-

11

B: 11 c: 0

+

t

-

-

21

B: 9 c: 12

+ -

-

21

B: 11 c: 10

+ -

-

+

C = control

+

-

-

group.

ment were not described. Twenty-three patients were divided into three groups. One group received an epidural infusion of bupivacaine 0.125% and fentanyl 0.0002% for 24 h before amputation and continued for at least 48 h after surgery (8 patients). Another group (7 patients) were operated in epidural anaesthesia and had postoperative epidural infusion of fentanyl and bupivacaine as described above. Finally, a group (8 patients) were operated under general anaesthesia and had systemic analgesia. After 1 year the incidence of phantom pain was significantly lower among the patients receiving pre-, intraand postoperative epidural analgesia, relative to patients who received general anaesthesia and systemic analgesia. Katsuly-Liapis and colleagues (Katsuly-Liapis et al., 1996) reported in abstract form a study where 45 patients were randomised into three groups to receive: (1) epidural analgesia with bupivacaine 0.25% and morphine for 3 days before amputation and continued for 3 days after the operation (15 patients), (2) epidural analgesia postoperatively (12 patients), or

(3) systemic analgesia with opioids and NSAID (18 patients). After 6 months the incidence of phantom pain was significantly lower in the group of patients who had epidural analgesia before, during and after amputation compared to the other two groups. No details about randomisation, blinding or pain assessment were presented. These data have to our knowledge so far not been published in an extended paper. In a blind and placebo-controlled trial (Nikolajsen et al., 1997b), 60 patients were randomly assigned to receive epidural bupivacaine (0.25% 4-7 ml/h) and morphine (0.16-0.28 mg/h) for 18 h before the amputation (29 patients) or epidural saline and systemic opioids (31 patients). Both groups underwent general anaesthesia for the amputation and all received epidural bupivacaine and morphine for postoperative pain management. Phantom pain both in terms of frequency and intensity was assessed after 1 week, 3 months, 6 months, and 12 months. Pain intensity was measured by means of a visual analogue scale (VAS) O-100. Fifty-six patients were available for

500 follow-up after 1 week; this number was reduced to 28 after 1 year, mainly because of deaths among the amputees (generally old people with generalised vascular diseases). The blindness in this study was ensured by two independent investigators, who were responsible for either the randomisation and preoperative pain treatment or for the postoperative pain treatment and follow-up. Patient blindness was secured by asking the patients after 6 months which type of treatment they had and they were not able at this time to distinguish between control and blockade treatment. After 1 week 52% in the blockade group and 56% in the control group had phantom pain. There was no difference in pain frequency and pain intensity at any time points between the two groups. So according to this study it was not possible to document a pre-emptive effect during an 18-h intense pain treatment before the operation. In the same group of patients the intense pre-injury pain treatment also failed to have any effect on subsequent stump sensitivity to mechanical and thermal stimuli (Nikolajsen et al., 1998). Discrepancy between experimental pre-emptive analgesia studies

and clinical

It is important to emphasise that despite failure to demonstrate a pre-emptive analgesic effect in many clinical studies, functional plastic changes of the nociceptive system by injury is an indisputable phenomenon. While there is firm experimental evidence for this notion, the clinical evidence for pre-emptive analgesia is much less clear. It is therefore of interest to speculate on differences between experimental and clinical studies that might explain this observation. TABLE 2 Factors influencing effectiveness of pre-emptive analgesia Genetic factors Pre-injury events Intensity of noxious injury Duration of noxious stimulus before during and after Type of injury (degree of inflammation and nerve injury) Duration of analgesia Efficacy of neural block Type of pharmacological blockade

Table 2 presents a list of possibilities why preemptive treatments may not work or only show a marginal effect in clinical studies. Clinical pains are much more complex than those in the laboratory. They may involve a mixture of inflammatory and neuropathic pain components and it can be difficult to dissect the contribution of each part. Patients are heterogeneous in terms of injury and pain experience before, during and after such injury. Another point concerns the duration of the analgesia. In animal experiment where the post-injury discharges are short lasting, a short pre-emptive pain treatment may be sufficient to prevent such discharges. It is also possible that an extended period might have been able to reveal a pre-emptive effect. The duration of post-injury neuronal discharges are not known. If post-injury neuronal activity from the periphery persists beyond the duration of analgesia this may mask a possible pre-emptive effect. Finally, an insufficient afferent blockade may lead to central sensitisation and prevent seeing a possible effect of the pre-emptive treatment (Dahl, 1994). It is possible that an extension of the analgesia to the peri-operative period may document an effect in terms of reduced pain and prevention of long-term sequelae. Conclusions The importance of sensitisation for chronic pain syndromes and the possible role of pre-injury pain for late post-injury pain has raised the issue whether phantom pain can be prevented. Early studies have shown that such prevention is possible by a preemptive analgesic treatment. However, subsequent randomised and controlled trials have failed to find evidence for such a pre-emptive effect. In fact it is not surprising that a short-lasting pre-injury treatment should prevent various aspects central sensitisation induced by an intense and long-lasting noxious input before, during and after the amputation. A future challenge lies in determining whether a long-lasting and intense pre-, peri- and postoperative analgesic treatment can in fact reduce the subsequent development of nerve injury pain including phantom pain. Moreover, it will be important to identify those patients that are likely to benefit from such a procedure.

501

References Bach, S., Noreng, M.F. and Tjellden, N.U. (1988) Phantom limb pain in amputees during the first 12 months following limb amputation, after preoperative lumbar epidural blockade. Pain, 33: 297-301. Berthele, A., Schadrack, J., Castro-Lopes, J.M., Conrad, B., Zieglgansberger, W. and Tolle, T.R. (2000) Neuroplasticity in the spinal cord of monoarthritic rats: from metabolic changes to the detection of interleukin-6 using mRNA differential display. In: J. Sandktlhler, B. Bromm and GE Gebhart (Eds.), Nervous System Plasticity and Chronic Pain. Progress in Brain Research, Vol. 129. Elsevier, Amsterdam, pp. 191-203. Birbaumer, N., Lutzenberger, W. and Montoya, P (1997) Effects of regional anaesthesia on phantom limb pain are mirrored in changes in cortical reorganization. J. Neurosci., 17: 55035508. Brennan, T.J., Umali, E.F. and Zahn, P.K. (1997) Comparison of pre- versus post-incision administration of intrathecal bupivacaine and intrathecal morphine in rat model postoperative pain. Anesthesiology, 87: 1517-1528. Bromm, B., Scharein, E. and Vahle-Hinz, C. (2000) Cortex areas involved in the processing of normal and altered pain. In: J. Sandktihler, B. Bromm and G.F. Gebhart (Eds.), Nervous System Plasticity and Chronic Pain. Progress in Brain Research, Vol. 129. Elsevier, Amsterdam, pp. 289-302. Casey, K.L. (2000) Concepts of pain mechanisms: the contribution of functional imaging of the human brain. In: J. Sandktihler, B. Bromm and G.F. Gebhart (Eds.), Nervous System Plasticity and Chronic Pain. Progress in Brain Research, Vol. 129. Elsevier, Amsterdam, pp. 277-287. Chabal, C., Jacobson, L. and Russell, L.C. (1989) Pain responses to perineuromal injection of normal saline, gallamine and lidocaine in humans. Pain, 36: 321-325. Coderre, T.J. and Katz, J. (1997) Peripheral and central hyperexcitability: differential signs and symptoms in persistent pain. Behav. Brain Sci., 20: 404-419. Cummins, T.R., Dib-Hajj, SD., Black, J.A. and Waxman, S.G. (2000) Sodium channels and the molecular pathophysiology of pain. In: J. Sandktihler, B. Bromm and G.F. Gebhart (Eds.), Nervous System Plasticity and Chronic Pain. Progress in Brain Research, Vol. 129. Elsevier, Amsterdam, pp. 3-19. Dahl, J.B. (1994) Neuronal plasticity and pre-emptive analgesia: implications for the management of post-operative pain. Dan. Med. Bull., 41: 434-442. Dahl, J.B. and Kehlet, H. (1993) The value of pre-emptive analgesia in the treatment of post-operative pain. B,: J. Anaesth., 70: 434-439. Devor, M. and Seltzer, Z. (1999) Pathophysiology of damaged nerves in relation to chronic pain. In: P.D. Wall and R. Melzack (Eds.), Textbook of Pain. 4th ed. Churchill Livingstone, Edinburgh, pp. 129-164. Devor, M., Govrin-Lippman, R. and Angelides, K. (1993) Na+ channels immunolocalization in peripheral mammalian axons and changes following nerve injury and neuroma formation. J. Neurosci., 135: 1976-1992. Dickenson, A.H. and Sullivan, A.F. (1987) Subcutaneous for-

malin-induced activity of dorsal horn neurones in the rat: different responses to an intrathecal opiate administered preor post-formalin. Pain, 30: 339-348. Dworkin, R.H., Hartstein, G. and Rosner, H.L. (1992) A high risk method for studying psychosocial antecedents of chronic pain: the prospective investigation of herpes zoster. J. Abnorm. Psychol., 101: 200-205. Elizaga, A.M., Smith, D.G., Sharar, S.R., Edwards, T. and Hansen, S.T. (1994) Continuous regional analgesia by intraneural block: effect on postoperative opioid requirements and phantom limb pain following amputation. J. Rehab. Res. Dev., 31: 179-187. Fischer, A. and Meller, Y. (1991) Continuous postoperative regional analgesia by nerve sheath block for amputation surgery - a pilot study. Anesth. Analg., 72: 300-303. Flor, H. (2000) The functional organization of the brain in chronic pain. In: J. Sandkiihler, B. Bromm and G.F. Gebhart (Eds.), Nervous System Plasticity and Chronic Pain. Progress in Brain Research, Vol. 129. Elsevier, Amsterdam, pp. 313322. Flor, H., Elbert, T. and Knecht, S. (1995) Phantom limb pain as a perceptual correlate of massive cortical reorganization in upper limb amputees. Nature, 375: 482-484. Flor, H., Elbert, T. and Mtihlnickel, W. (1998) Cortical reorganization and phantom phenomena in congenital and traumatic upper-extremity amputees. Exp. Brain Res., 119: 205-212. Gerber, G., Youn, D.-H., Hsu, C.H., Isaev, D. and Randic, M. (2000) Spinal dorsal horn synaptic plasticity: involvement of group I metabotropic glutamate receptors. In: J. Sandktihler, B. Bromm and G.F. Gebhart (Eds.), Nervous System Plasticity and Chronic Pain. Progress in Brain Research, Vol. 129. Elsevier, Amsterdam, pp. 115-134. Hill, A., Niven, C.A. and Knussen, C. (1996) Pain memories in phantom limbs: a case study. Pain, 66: 381-384. Hoheisel, U. and Mense, S. (2000) The role of spinal nitric oxide in the control of spontaneous pain following nociceptive input. In: J. Sandkiihler, B. Bromm and G.F. Gebhart (Eds.), Nervous System Plasticity and Chronic Pain. Progress in Brain Research, Vol. 129. Elsevier, Amsterdam, pp. 163-172. Houghton, A.D., Nicholls, G., Houghton, A.L., Saadah, E. and McCall, I. (1994) Phantom pain: natural history and association with rehabilitation. Ann. R. Coil. Surg. Eng., 76: 2225. Jahangiri, M., Jayatunga, A.P., Bradley and Dark, C.H. (1994) Prevention of phantom pain after major lower limb amputation by epidural infusion of diamorphine, clonidine and bupivaCaine. Ann. R. Coil. Surg. Engl., 76: 324-326. Jensen, T.S. and Gottrup, H. (2000) Assessment of neuropathic pain. In: C. Warheld. A. Rice, D. Justins and C. Eccleston (Eds.), Textbook of Clinical Pain Management. Arnold, London (in press). Jensen, T.S. and Nikolajsen, L. (1999) Phantom pain and other phenomena after amputation. In: P.D. Wall and R. Melzack (Eds.), Textbook of Pain. 4th ed. Churchill Livingstone, Edinburgh, pp. 799-8 14. Jensen, T.S., Krebs, B., Nielsen, J. and Rasmussen, P. (1983) Phantom limb, phantom pain and stump pain in amputees

502

during the first 6 months following limb amputation. Pain, 17: 243-256. Jensen, T.S., Krebs, B., Nielsen, J. and Rasmussen, P. (1985) Immediate and long-term, phantom limb pain in amputees: incidence, clinical characteristics and relationship to pre-amputation pain. Pain, 21: 267-278. Kalso, E. (1997) Prevention of chronicity. In: T.S. Jensen, J.A. Turner and Z. Wiesenfeld-Hallin (Eds.), Proceedings of the 8th World Congress on Pain. Progress in Pain Research and Management, Vol. 8. IASP Press, Seattle, WA, pp. 215-230. Katsuly-Liapis, I., Georgakis, P. and Tierry, C. (1996) Pre-emptive extradural analgesia reduces the incidence of phantom pain in lower limb amputees. BI: J. Anaesth., 76: 125. Katz, J. and Melzack, R. (1990) Pain memories in phantom limbs: review and clinical observations. Puin, 43: 319-336. Larbig, W., Montaya, P. and Flor, H. (1996) Evidence for a change in neural processing in phantom limb pain patients. Pain, 67: 275-283. McQuay, H.J. (1992) Pre-emptive analgesia. Br J. Anaesth., 69: l-3. McQuay, H.J. (1994) Do preemptive treatments provide better pain control. In: G.F. Gebhart, D.L. Hammond and T.S. Jensen (Eds.), Proceedings of the 7th World Congress on Pain. Progress in Pain Research and Management, Vol. 2. IASP Press, Seattle, WA, pp. 709-723. Melzack, R. (1995) Phantom-limb pain and the brain. In: B. Brommm and I.E. Desmedt (Eds.), Advances in Pain Research and Therapy, Vol. 22. Raven Press, New York, pp. 73-82. Moore, K.A., Baba, H. and Woolf, C.J. (2000) Synaptic transmission and plasticity in the superficial dorsal horn. In: J. Sandkiihler, B. Bromm and G.F. Gebhart (Eds.), Nervous Systern Plasticity and Chronic Pain. Progress in Brain Research, Vol. 129. Elsevier, Amsterdam, pp. 63-80. Nathan, P.W. (1962) Pain traces left in the central nervous system. In: C.A. Keele and R. Smith (Eds.), The Assessment of Pain in Man and Animals. E&S Livingstone, Edinburgh, pp. 129-134. Nikolajsen, L. and Jensen, T.S. (2000) Phantom limb pain. Curr: Rev. Pain, 4: 166-170. Nikolajsen, L., Hansen, C.L., Nielsen, J., Keller, J., ArendtNielsen, L. and Jensen, T.S. (1996) The effect of ketamine on phantom pain: a central neuropathic disorder maintained by peripheral input. Pain, 67: 69-77. Nikolajsen, L., Ilkjaer, S., Kroner, K., Christensen, J.H. and Jensen, T.S. (1997a) The influence of preamputation pain on postamputation stump and phantom pain. Pain, 72: 393-405. Nikolajsen, L., Ilkjaer, S., Kroner, K., Christensen, J.H. and Jensen, T.S. (1997b) Randomised trial of epidural bupivacaine and morphine in prevention of stump and phantom pain in lower-limb amputation. L.ancet b, 350: 348-354. Nikolajsen, L., Ilkjmr, S. and Jensen, T.S. (1998) Effect of preoperative extradural bupivacaine and morphine on stump sensation in lower limb amputees. BI: J. Anaesth., 81: 348354. Nikolajsen, L., Got&up, H., Kristensen, A.G.D. and Jensen, T.S. (2000a) Memantine (a N-methyl-D-aspartate receptor antagonist) in the treatment of neuropathic pain following amputa-

tion or surgery: a randomised, double-blind, cross-over study. Anesth. Analg., in press. Nikolajsen, L., Ilkjaer, S. and Jensen, T.S. (2OOOb) Relationship between mechanical sensitivity and postamputation pain: A prospective study. Eur: J. Pain, in press. Novakovic, S.D., Tzoumaka, E., McGivem, J.G., Haraguchi, M., Sangameswaran, L., Gogas, K.R., Eglen, R.M. and Hunter, J.C. (1998) Distribution of the tetrodotoxin-resistant sodium channel PN3 in rat sensory neurons in normal and neuropathic pain conditions. J. Neurosci., 18: 2174-2187. Nystriim, B. and Hagbarth, K.E. (1981) Microelectrode recordings from transected nerves in amputees with phantom limb pain. Neurosci. L&t., 27: 21 l-216. Pinzur, M.S., Garla, P.G.N., Pluth, T. and Vrbos, I. (1996) Continuous postoperative infusion of a regional anaesthetic after an amputation of the lower extremity. J. Bone Joint Surg., 78: 1501-1505. Puke, M.J.C. and Wiesenfeld-Hallin, Z. (1993) The differential effects of morphine and the ct2-adrenoceptor agonists clonidine and dexmedotomidine on the prevention and treatment of experimental neuropathic pain. Anesth. A&g., 77: 104-109. Rajbhandari, S.M., Jarett, J.A., Griffiths, P.D. and Ward, J.D. (1999) Diabetic neuropathic pain in a leg amputated 44 years previously. Pain, 83: 627-629. Sandkiihler, J., Benrath, J., Brechtel, C., Ruscheweyh, R. and Heinke, B. (2000) Synaptic mechanisms of hyperalgesia. In: J. Sandktihler, B. Bromm and G.F. Gebhart (Eds.), Nervous System Plasticity and Chronic Pain. Progress in Brain Research, Vol. 129. Elsevier, Amsterdam, pp. 81-100. Sherman, R.A. and Sherman, C.J. (1983) Prevalence and characteristics of chronic phantom limb pain among American veterans. Am. J. Phys. Med., 62: 227-238. Schug, S.A., Burell, R., Payne, J. and Tester, P. (1995) Pre-emptive epidural anaesthesia may prevent phantom limb pain. Reg. Anesth., 20: 256. Svendsen, F., Hole, K. and Tjolsen, A. (2000) Long-term potentiation in single WDR neurons induced by noxious stimulation in intact and spinalized rats. In: J. Sandkiihler, B. Bromm and G.F. Gebhart (Eds.), Nervous System Plasticity and Chronic Pain. Progress in Brain Research, Vol. 129. Elsevier, Amsterdam, pp. 153-161. Wall, P.D. (1988) The prevention of postoperative pain. Pain, 33: 289-290. Wartan, S.W., Hamann, W., Wedley, J.R. and McCall, I. (1997) Phantom pain and sensation among British veteran amputees. BI: J. Anaesth., 78: 652-659. Waxman, S.G. (1999) The molecular pathophysiology of pain: abnormal expression of sodium channel genes and its contrbutions to hyperexcitability of primary sensory neurons. Pain Sappl., 6: 133-140. Wilder-Smith, O.H.G. (2000) Pre-emptive analgesia and surgical pain. In: J. Sandktihler, B. Bromm and G.F. Gebhart (Eds.), Nervous System Plasticity and Chronic Pain. Progress in Brain Research, Vol. 129. Elsevier, Amsterdam, pp. 505-524. Wilkins, K.L., McGrath, P.J., Finley, G.A. and Katz, J. (1998) Phantom limb sensations and phantom limb pain in child and adolescent amputees. P&n, 78: 7-12.

503 Woolf, C.J. (1992) Excitability changes in central neurons following peripheral damage. Role of central sensitization in the pathogenesis of pain. In: W. Willis (Ed.), Hyperulgesia and Allodynia. Raven Press, New York, pp. 221-243. Woolf, C.J. and Chong, M. (1993) Pre-emptive analgesia: treating postoperative pain by preventing the establishment of central sensitization. Anesth. Analg., 71: 362-379. Woolf, C.J. and Salter, M.W. (2000) Neuronal plasticity: Increasing the gain in pain. Science, 288: 1765-1768. Woolf, C.J. and Thompson, S.W.N. (1991) The induction

and maintenance of central sensitisation is dependent on &‘-methyl-D-aspartic acid receptor activation; implications for the treatment of post-injury pain hypersensitivity states. Pain, 44: 293-299.

Woolf, C.J., Shortland, P. and Coggeshall, R.E. (1992) Peripheral nerve injury triggers central sprouting of myelinated afferents. Nature,

355: 75-78.

Yaksh, T.L. (1993) The spinal pharmacology of facilitation of afferent processing evoked by high threshold afferent input of the postinjury pain state. Cur< Opin. Neural. Neurosurg., 6: 250-256.

I. Sandkiihler, B. Bromm and G.F. Gebbart (Eds.) Progress in Brain Research, Vol. 129 0 2000 Elsevier Science B.V. All rights reserved

CHAPTER 37

Pre-emptive analgesia and surgical pain Oliver H.G. Wilder-Smith * Nociception

Research

Group

Beme

University

Pre-emptive analgesia: concepts and background The effective and consistent management of pain after surgery continues to be a therapeutic challenge (Zenz, 1997). In a recent survey of United Kingdom hospitals, significant pain was experienced after surgery by over 80% of the patients questioned (Bruster et al., 1994). About one third of these patients experienced postoperative pain that was continuously or almost continuously present for long periods of time. A United States survey of patients’ concerns before surgery found postoperative pain to be the primary concern in almost 60% of those questioned, with approximately three-quarters of the respondents having experienced significant postoperative pain after previous surgery (Warfield and Kahn, 1995). Bearing the hope of achieving significant improvements in the management of postoperative pain, the concept of pre-emptive analgesia was introduced in the late 1980s. The concept was founded upon an increasing body of animal research demonstrating central nervous system plasticity and sensitisation after nociception (Woolf, 1983; Woolf and Wall, 1986a; Woolf and Thompson, 1991), and was rapidly popularised by a number of pertinent editorials (Wall, 1988; McQuay and Dickenson, 1990; McQuay, 1992; Dahl and Kehlet, 1993; Bridenbaugh, 1994) and review articles (Woolf, 1989; Coderre et

* Corresponding author: O.H.G. Wilder-Smith, Nociception Research Group, Beme University, Bubenbergplatz 11, CH-3011 Beme, Switzerland. Tel.: +41-31-3123737; Fax: +41-31-3123770; E-mail: [email protected]

Bubenbergplatz

11, CH-3011

Berne,

Switzerland

al., 1993; Woolf and Chong, 1993; McQuay, 1995) in the medical literature. In its original form, pre-emptive analgesia comprised two main postulates: firstly, that an analgesic intervention started before nociception would be more effective than the same intervention commenced afterwards; and secondly, that this advantageous effect would outlast the pharmacological duration of action of the analgesic concerned (see Jensen and Nikolajsen, 2000, this volume). The presence of neuroplasticity is fundamental to the concept of pre-emptive analgesia. It is based upon findings in animal models (see: Gerber et al., 2000, this volume; Moore et al., 2000, this volume; Sandktihler et al., 2000, this volume; Svendsen et al., 2000, this volume) showing that nociceptive input to the central nervous system alters its subsequent function. These changes were initially shown to affect neurones in the dorsal horn of the spinal cord, but similar changes have now also been demonstrated to occur further up the synaptic chain of the central nervous system, e.g. in the thalamus (Gautron and Guilbaud, 1982; Guilbaud et al., 1989; see: Dostrovsky, 2000, this volume; Lenz et al., 2000, this volume) and cortex (see: Bromm et al., 2000, this volume; Casey, 2000, this volume; Flor, 2000, this volume). Early studies of nociceptive neuroplasticity, usually in non-intact (i.e. decerebrate or spinalised) preparations, elicited mainly excitatory changes (sensitisation) in central neuronal function that were most easily (but not exclusively) produced via C-fibre input. Signal summation was found to play a major role: temporal summation for noninflamed tissue, spatial summation for inflamed or traumatised tissues. In the case of sensitisation, such summated input leads to long-lasting depolarisation

506 which can significantly outlast the original nociceptive signal by seconds up to minutes and which can even start to occur spontaneously (wind-up, spontaneous pain). The altered electrophysiological behaviour of the neurones concerned (i.e. synaptic long-term potentiation; Gerber et al., 2000, this volume; Sandktihler et al., 2000, this volume) spreads to adjacent neurones and results in reduced thresholds, increased responses to stimulation and after-discharging or spontaneous discharge. This central sensitisation expresses itself in the clinical symptoms of allodynia (previously non-painful stimuli are perceived as painful), hyperalgesia (increased pain sensation with a given nociceptive stimulus), wind-up (prolonged or spontaneous pain after stimulation) and increased size of the hypersensitive neuronal sensory fields (secondary hyperalgesia). Central sensitisation is the result of both posterior horn neuronal input facilitation as well as loss of inhibitory inputs, and must be distinguished from the primary hypersensitivity and hyperalgesia around a lesion due to sensitisation of peripheral nociceptors (Treede, 1995). The phenomenon of central sensitisation after nociception as seen in animal models is considered to play an important role in explaining the clinical manifestations of postoperative pain (cf. reviews cited above). Biochemically, these changes in central neuronal function are mediated by the synaptic release of excitatory amino acid (e.g. glutamate, aspartate) - and also neuropeptide (e.g. substance P, neurokinin A) - neurotransmitters and the subsequent binding to their membrane receptors (e.g. N-methyl-D-aspartate [NMDA] and tachykinin receptors) (e.g. Dickenson, 1995; Sandktihler et al., 2000, this volume). Activation of the NMDA receptor, dependent upon continuing NO production and the subsequent release of soluble GMP-cyclase (Meller and Gebhart, 1993), is necessary for the elicitation of central sensitisation (Woolf and Thompson, 1991; see, however, Hoheisel and Mense, 2000, this volume). Both NMDA and tachykinin receptor antagonists interfere with the electrophysiological consequences of nociceptive input at dorsal horn neurones as well as blocking the behavioural/clinical consequences of nociception. The release of excitatory amino acids (and/or tachykinins) results in slow synaptic dorsal horn potentials, the prerequisite for the electrophysiological

neuronal membrane changes of central sensitisation, described above. Excitatory amino acid or tachykinin receptor activation (ligand gating) as well as membrane depolarisation (voltage gating) is accompanied by increased calcium entry into the neuronal cell bodies via calcium ionophores, thus raising cellular second-messenger (e.g. cGMP) and protein kinase C activity. The end result of these biochemical changes is not only positive feedback on the NMDA receptors but also the activated expression of early intermediate genes (e.g. c-fos or B-&z) regulating the production of modulatory substances such as the hyperalgesic dynorphin. An early observation in animal models was that it is easier to prevent the establishment of central sensitisation (by providing analgesia, e.g. by morphine, before nociception occurred; pre-emptive analgesia) than to treat (suppress) it once established (Dickenson and Sullivan, 1986; Woolf and Wall, 1986b). In the animal models used, the dose of an analgesic - typically morphine - necessary to prevent electrophysiological central neuronal sensitisation was found to be much smaller than the dose necessary to suppress established sensitisation after nociceptive input. The electrophysiological advantages conferred by analgesic pre-emption were shown to outlast the pharmacological duration of action of the analgesic. In the context of the important role central sensitisation is considered to play in postoperative pain, it is logical that the discovery of pre-emptive analgesia was rapidly followed by attempts to extend the concept into clinical - particularly surgical - practice. Extrapolating from the animal data at the time, the hypothesis for the human surgical pre-emptive analgesia studies undertaken in the early 1990s was that performing an analgesic intervention before surgery would result in a clearly better clinical pain outcome postoperatively than the same analgesic intervention initiated after surgery had started. It was further expected that the improvement in postoperative pain outcome would clearly outlast the pharmacological duration of action of the analgesic intervention used. Pre-emptive analgesia and surgical pain: the evidence of clinical studies A large number of investigations of pre-emptive analgesia in the surgical context were undertaken in

507 the early 1990s. Unfortunately, by the mid-1990s it was becoming clear that this first wave of studies was either unable to show any clinical effects of preemptive analgesia at all, or that the improvements in clinical pain outcomes were clinically disappointing, with only modest effects on postoperative pain measures or analgesic consumption. These modest effects were visible mainly using opioids (systemically and particularly neuraxially) and local anaesthetics. As discussed in a number of editorials and reviews published at the time (e.g. Woolf and Chong, 1993; Kehlet, 1994; McQuay, 1995; Wilder-Smith, 1995), problems in demonstrating clinically convincing effects of pre-emptive analgesia were initially considered in large part due to faults in clinical study design. Apart from basic points such as blinding, randomisation, prospectiveness and group sizes, these faults included issues such as contamination of the anaesthetic technique by other analgesic substances, absence of equality between the pre-emptive and post-emptive analgesic intervention, inadequacies in control states and lacking sensitivity of the clinical postoperative pain measures used. The question of how to match the analgesic intervention to the extent and duration of the nociception occurring during and after surgery was increasingly raised. Since then a second series of clinical studies has been published, many with better designs. Some of the studies have addressed the question of adequately matching analgesia and nociception by studying genuine perioperative analgesia (i.e. for the entire duration of nociception) as opposed to only bolus analgesic interventions (e.g. Gottschalk et al., 1998; Likar et al., 1998). The results of a literature survey for this ‘second wave’ of pre-emptive analgesia studies are summarised in Tables 1 and 2. The survey was performed using MEDLINE and the keyword ‘pre-emptive analgesia and surgery’, and includes only randomised, controlled and prospective studies with valid design (i.e. pre- and post-nociceptive interventions). As can be seen from the studies detailed in Tables 1 and 2, improvements in study design have been followed by more success in demonstrating effects of pre-emptive analgesia for opioids, ketamine (a non-competitive NMDA receptor antagonist) and local anaesthesia. It should be noted that in the positive studies (Table l), the improvements in clinical

pain outcomes achieved are generally small and of short duration. Studies with adequate design but not finding any effect also continue to be published. Due to the well-known publication bias against negative studies, it remains difficult to establish a true final weight of evidence for or against clinically relevant pre-emptive analgesia. On the present balance of evidence, we would suggest that while there is evidence that opioids, NMDA-antagonists and local anaesthesia have pre-emptive analgesia effects, these effects are modest and of limited clinical significance. There is a suggestion that the clinical efficacy of analgesic pre-emption may improve with better matching between analgesic intervention and nociceptive input. Redefining the pre-emptive analgesia problem Why is there such a discrepancy between the success of pre-emptive analgesia in the experimental animal model and in the clinical surgery patient? In looking more closely at the two models and at their commonalties and similarities we would suggest that three closely linked problems are operating: (1) the problem of extrapolation from experimental to clinical; (2) the problem of clinical study design; (3) the problem of how to measure pain outcomes. The problem of extrapolation rests upon the fundamental differences between experimental and surgical pain models. Firstly, the duration and magnitude of, and number of modalities involved in surgical nociception are far greater than for experimental nociception models. In surgery, nociceptive input continues in the presence of extensively chemically sensitised traumatised tissues, which is often not the case in animal models, particularly the early ones. Secondly, a large proportion of animal models involve non-intact (i.e. spinalised or decerebrate) preparations - again, particularly earlier models while surgical models are intact with regard to their central nervous system. Non-intact animal models will therefore neither reflect the integral effects of nociception on a whole central nervous system, nor will they be able to demonstrate the responses with which an intact central nervous system defends itself against nociceptive inputs. These defences will include both neuronally (e.g. descending inhibitory controls; see Sandktihler et al., 2000, this volume, and Svendsen et al., 2000, this volume) and hormon-

1

Richmond et al., 1993 Katz et al., 1992

Romej et al., 1996 Pasqualucci et al., 1996 Rockernan” et al., 1996

Chia et al., 1999 wu et al., 1999 Ke et al., 1998 Kundra et al., 1998 Liar et al., 1998 Gottschalk et al., 1998 Griffin et al., 1997 Kundra et al., 1997 F” et al., 1997 Choe et al., 1997

Study

Yes Yes jW

Yes Yes Yes Yes

Yes Yes Yes yes periop vs. postop periop vs. postop

Yes Yes

Yes

Yes Yes pre > post

IV ED

ml co”trol small groups (2 x 15)

intervention

PR IP balanced analgesia

IV IM II caudal ED IV ED IV ED IV ED

Type

Analgesic

cant: placebo, pre + post placebo control

small groups (2 x 19) small groups (2 x 15)

placebo control

also: placebo control also: placebo control

Comme”ts

analgesia effect

Analgesia: pre = post’?

showing pre-emptive

Analgesia: pre + post present?

Design

Summary of clinical studies 1992-1999

TABLE

short infusion bolus single single infusion infusion short infusion bolus

dextromethorphan 5 mg/kg dextromethorphan 40 mg 0.5% bupivacaine bupivacaine, morph 0.02 mg/kg ketoprofen 100 mg, then 12 mg/h bupivacaine, fentanyl alfentanil70 Kg/kg morphine 3 mg ketamine ketamine 60 mg morphine 2 mg paracetamol20 mg/kg bupivacaine diclofenac 75 mg metamizol I g ED mepivacaine ED morphine I mg morphine 10 mg fentanyl4 Kg/kg

abdominal hysterectomy thoracotomy

tonsillectomy laparoscopic cholecystect major abdominal surgery

major abdo surgery laparoscopic cholecystect diagn gynae laparoscopy hemiorraphy gynae surgery radical prostatectomy abdo hysterectomy lumbar Iaminectomy abdo surgery mixed surgery

Surgery

VAS (rest) VAS (exert) VAS

Likar et al., 1998

VAS VAS VAS VAS

Pasqualucci et al., 1996 Rockemann et al., 1996 Richmond et al., 1993 Katz et al., 1992 46h

J30 “Ii” $240 min $24 h 4b-5 d ?

Oh ++dl,2

&24 h &24 h Jhosp, 9.5 wk ++3.5,5.5 wk i-+&72 h -0-72 h

J24h JO.% 4, 8 h

J

suppl. analgesia PCA morphine PCA morphine PCA morphine

PCA morphine PCA morphine no. receiving sup. morphine “use demand morphine

PCA morphine

ED!

PCA pi&amide

suppl. analgesia “se nurse demand morphine

IM pethidine on demand

h

&24 h (total) /d2, d5 (total) J ?

$24 h (total) Jdl2 4. $24 h (total)

$48-72

ff

ff $24 h (total)

$48 h (total)

Cd12

Improved? When?

;t3h 424 h

blood glucose, co&o1 relative pain thresholds

children

Tl-2A RMPO 01 (time, total)

wk

pre c placebo

PONV

PCA pi&ram:

children

worst: placebo

Isedation,

f3.5 wk -5.5.9.5

r

t

ff post = placebo

t

.A cf

Improved? When?

TFA

activity

TFA

TFA TFA side effects

TFA bed rest

pat sat&fact side effects

Others

CO”Ulle”tS

Abbreviations: IV = intravenous; IM = intramuscular; I1 = incisional infiltration; ED = epidural: PR = per rectum; VAS = visual analogue score; MPPIS = McGill present pain intensity score; OPS = objective pain score; FLACC = faces, legs, activity, cry, consolability; PCA = patient-controlled analgesia; TFA = time to first analgesia demand; Tl-2A = time between first and second analgesia demand: RMPO = rescue morphine directly postoperatively: 01 = oral intake.

FLACC

et al., 1997

et al., 1998

Kundra et al., 1997 Fu et al., 1997 Choe et al., 1997 Romej et al., 1996

Griffin

Got&chalk

VAS (rest) VAS (exert) VAS VAS

MPPIS OPS

Ke et al., 1998 Kundm et al., 1998

wu et al., 1999

++I%3 d ++&3 d

PCA morphine

Analgesic consumotion

Pai” intensitv

Improved? When?

pre vs. post

Pain measures and their improvement

VAS (rest) VAS (exert) worst VAS

1 (continued)

Chia et al., 1999

Study

TABLE

1999 Yes

pre > post

Analgesia: pre = post?

no preemptive

VRS VAS FPS VAS (rest) VAS (cough) VAS (move) VAS VRS

LAB IV IV ED II 1v caudal ED

one session

ED pethidine IM morphine nurse demand piritramide analgesic supplement PCEA bupiv 0.125% fentanyl6 Kg/ml nurse demand opioids

IM opioids -

Analgesic consumption

pre vs. post

placebo control

N = 51

~24

h (total)

ttdL2 (total) ++3d -24 h tf -48 h

tf -

Improved’? When?

TFA

TFA

Others

single bolus

bolus single short infusion single

bolus f infusion

intervention

thoracic ED IV

IV

placebo control

power calculation! N= 112

Type

Analgesic Comments

analgesia effect

tt3d ++24 h tl postop, home cr48 h ++48 h -48 h f30,120 min 1-30, 120 min

tf ++6h,l.?xhd tfl d

VAS VAS

MPQ

Improved? When?

Pai” intensity

Pain measures and their improvement

Analgesia: pre + post present?

Design

showing

c-f

tj

Improved? When?

min-

pain questionnaire;

0.5% bupivacaine fentanyl 10 kg/kg sufentanil 1 pg/kg

ketamine 0.5 mg/kg i IO pg kg-’ bupivacaine 0.5% tenoxicam alfentanil ketamine 60 mg bupivacaine 0.25% ketoprofen 2 mg/kg 0.25% bupivacaine

Abbreviations: IV = Intravenous; LAB = local anaesthetic block; ED = epidural; II = incisional infiltration; VAS = visual analogue score: MPS = McGill FPS = faces pain scale: IM = intramuscular: PCEA = patient-controlled epidural analgesia: TFA = time to first analgesia demand.

Fassoulaki et al., 1995

Kucuk et al., 1998 Bourget et aI., 1997 Liiar et al., 1997 Ho et al., 1997 Aguilar et al.. 1996

Heinke and Grimm, 1999 Campbell et al.. 1998

Study

Aguilar et al., 1996 Fassoulaki et al., 1995

Kucuk et al., 1998 Bourget et al., 1997 Likar et al., 1997 Hoet al.. 1997

Campbell et al., 1998

Heinke and Grimm.

Study

Summary of clinical studies 1992-1999

TABLE 2

VRS = verbal rating score;

children alfentanil intraop, ED fentanyl at operation end

alfentanil intraop

Comments

upper abdominal elective abdo laparotomy gynae procedures hernia, circumcision, orchidopexy thoracotomy abdominal hysterectomy

bilateral lower 3rd molar extraction

gynae laparotomy

Surgery

511

ally (e.g. stress-induced analgesia) mediated central nervous system responses to nociceptive input (Le Bars et al., 1979, 1992; Kelly, 1986; Termann et al., 1986). It is of interest that intact animal models using longer-lasting nociceptive inputs show pre-emptive analgesia results much closer to the modest results achieved in human surgical pre-emptive analgesia studies (e.g. Jayaram et al., 1995; Fletcher et al., 1996) than those involving non-intact preparations. The main difficulties with clinical study design involve questions of standardisation. Standardising surgery is notoriously difficult, but should be attempted, both by the use of agreed defined surgical protocols and by limiting the number of surgeons as far as possible (one surgeon is ideal!). Standardisation of the patients is far more difficult, as the variability of patient’s responses to pain and nociception is generally underestimated and - so far - not amenable to selection or prediction. Animal studies of the genetically inherited component in the responses to pain (e.g. sensitivity, inhibitory controls) have shown a large variability in this area (Lutfy et al., 1994; Mogil et al., 1995; Stemberg and Liebeskind, 1995; Kest et al., 1999). Thus study size must always be adequate to cope with this variability of pain sensitivity and responses, albeit tempered by the desire to demonstrate truly clinically relevant effects. Other important design features include the equality of the pre- and post-nociceptive analgesic interventions as well as an adequate match between nociception and analgesic intervention regarding intensity, duration and sensory modalities involved. The latter feature was a particular problem with earlier pre-emptive analgesia studies: a single bolus of morphine may well adequately cover the nociception of a brief electrical C-fibre stimulus, but it is unlikely to cover the nociception associated with an abdominal laparotomy lasting several hours. Included in this problem is the question of when nociception ends; again, for a brief electrical C-fibre nociceptive stimulus this is easy to define, but when does surgical nociception really end? This problem will obviously also affect the definition of the post-emptive analgesia comparison state. The final problem concerns the question of measuring pain and its outcomes. A major difference between experimental animal and clinical human pain studies is the difference in pain measures used.

Clinical, e.g. post-surgical, pain is a subjective phenomenon, directly accessible only to the conscious person experiencing it. This experience of pain is influenced by a multitude of factors apart from the nociception causing it and therefore bears no direct, linear relationship to the nociceptive event causing it. In intact animal pain studies, we have to rely on indirect behavioural correlates of the pain experience (e.g. tail flick latency, hot plate latencies, etc.) as the animal cannot communicate the pain in a more direct fashion. It must be clear that these behavioural pain measures are only surrogate measures which neither measure the pain experience nor the characteristics of the nociceptive event directly. The electrophysiological measures used in non-intact animal pain studies are even further removed from the subjective pain experience of clinical, post-surgical pain. They may, however, encode more information about the original nociceptive event and the damage or stress it is causing to the body than behavioural measures. Thus the comparison of behavioural or electrophysiological results from animal research with the results of clinical pain measures in humans will always be difficult. The above-mentioned difficulties pertain to human pain research, too. Subjective clinical pain measures such as pain intensity scales or postoperative analgesia use tell us different things than the objective measures of psychophysical (e.g. pain thresholds) or electrophysiological (e.g. nociceptive flexion reflexes) testing do. Despite the fact that clinical pain measures such as visual analogue pain intensity scales (VAS) or postoperative patient-controlled analgesia (PCA) consumption have been used for quite some time in pain research, we are still unsure as to what they really tell us and how they behave under different circumstances. This is illustrated, e.g. by the quanta1 - not linear - doseeffect relationship between titrated dose of alfentanil given intravenously and postoperative pain relief (Fig. 1) (Tverskoy et al., 1996). The non-linear relationship between dose of opioid given and degree of (subjective) pain relief obtained is a reflection of the multifactorial nature of the subjective pain experience already discussed, and specified in the original IASP definition of pain. In view of the multi-aetiological nature of the subjective, clinical pain experience, it is unrealistic to expect clinical pain to correlate closely with the lower-order func-

512

alfentanil (w/kg)

Fig. 1. Figure modified after Tverskoy et al. (1996) showing the cumulative frequency distribution curve for the complete relief of spontaneous postoperative pain by 3 kg/kg intravenous increments of alfentanil, given at 5-min intervals.

tional changes in the central nervous system function described in the original animal studies of neuroplasticity after nociception and pre-emptive analgesia. If we are to investigate the human correlates of the preemptive analgesia found in animal studies, we must study similar-order phenomena in humans, namely changes in objective psychophysical or electrophysiological measures reflecting altered central nervous system function after nociception. Indeed, if changes in the analgesic management of surgery are to have effects on long-term medical outcome, it should be of particular interest to study these objective measure of altered sensory processing, as they may reflect the damage done to the body by nociception - and its modulation - much more closely than subjective clinical pain measures could. Towards a relevant human study design for pre-emptive analgesia The pre-emptive effect of analgesia in animal studies is primarily defined in terms of altered central nervous system function. These alterations have mainly been described as affecting sensory and lower-order (spinal, thalamic) central nervous system processing. In order to effectively study pre-emptive analgesia in the human context, it therefore appears logical to study alterations in central nervous system sensory processing as a result of surgical nociception, and how this can be modulated by analgesic intervention. It is at this level that the reality or not of preemptive analgesia should jirst be established, The relationship between altered central nervous system processing and subjective, clinical pain measures or

long-term medical outcomes can then be established in a second step. If studies of human pre-emptive analgesia and surgery are to be performed in the clinical context, the techniques involved need to be simple, valid, easily performed and not too time-consuming. Psychophysical measures such as pain or sensation thresholds offer such a technique. With adequate training of experimenter and subject together with good protocol design and standardisation, they are simple, reproducible, reliable and not too time-consuming in use, and have been well-validated as reflecting central nervous system sensory processing and its alterations (Rollmann and Harris, 1987; Lautenbacher and Rollman, 1993; Arendt-Nielsen et al., 1995; Wilder-Smith et al., 1996, 1998). In order to obtain comprehensive answers about alterations in sensory processing after surgery, thresholds need to be measured at multiple sites. Ideally, thresholds should be measured on the wound (primary hyperalgesia), near the wound (ca. lo-15 cm from the wound; secondary hyperalgesia), as well distant to the site of surgery (to detect generalised effects such as descending inhibition or stress-induced analgesia). Measuring multiple sensory modalities (i.e. sensation, pain detection, pain tolerance thresholds) and multiple stimulation modalities (i.e. mechanical, electric, thermal) increases the understanding of the sensory changes involved (see also Treede and Magerl, 2000, this volume). It should also be remembered that direct nerve stimulation will give different answers than dermatomal stimulation. All the psychophysical tests mentioned so far will only involve superficial, cutaneous structures; methods for testing deep (e.g. muscle (see Hoheisel and Mense, 2000, this volume)) or visceral structures remain experimental and difficult to transfer to the clinical context for the time being (Cervero, 1995; Gavrilov et al., 1996; Bajaj et al., 1999). The psychophysical measures under discussion need to be embedded in a well-conceived and well-standardised design. The analgesic intervention should be well-matched to the nociception concerned to make the results relevant. Thought should be given to making the study feasible in the clinical context, otherwise too many data points will be missed. In order to enable objective, psychophysical and subjective clinical pain measures to be compared or

513

correlated, they should be measured simultaneously. If several clinical pain measures are measured (e.g. pain VAS and PCA morphine use), which is the primary and which is the secondary endpoint must be determined (i.e. if PCA morphine use is to be the measure, aim for similar VAS scores targets in all patients). It is desirable to extend the period of study for as long as feasible, particularly if effects on medical outcome are to be investigated. Sensory processing after surgery: defining the changes In the last decade, a number of studies have been performed to investigate changes in sensory processing due to surgical nociception and how these can be modulated by analgesic interventions. The findings of these selected studies, both clinical human and animal, are summarised in Table 3. From the results of the studies we will try to answer the following pertinent questions: (1) what are the changes in sensory processing after surgery? (2) can central sensitisation be detected after surgery? (3) how are the changes in sensory processing affected by analgesic intervention? (4) what other factors affect post-surgical sensory change? Mechanical pain thresholds measured distant to the surgical wound are either unaltered or moderately decreased in the studies concerned (Lascelles et al., 1995, 1997, 1998; Welsh and Nolan, 1995; Moiniche et al., 1997). In none of the studies cited were nonnociceptive mechanical thresholds measured distant to surgery. When present, distant hyperalgesia is visible within the first 24 h postoperatively and is suppressed by opioid pre-emptive analgesia. Mechanical hyperalgesia close to the wound (secondary hyperalgesia), considered to reflect central sensitisation (see Treede and Magerl, 2000, this volume), is more pronounced and of longer duration, appearing to be present up to 4-5 days postoperatively (Richmond et al., 1993; Moiniche et al., 1997; Stubhaug, 1997) and gone by 8 days (Moiniche et al., 1997). The area of secondary hypersensitivity has been shown to be reduced by ketamine pre-emptive analgesia (Stubhaug, 1997). Wound hyperalgesia (primary hyperalgesia) is considered to reflect both peripheral nociceptor and central nervous sensitisation. As expected, mechanical pain thresholds at the surgical incision are

lower postoperatively, with such hyperalgesia being reduced - but not abolished - by opioid pre-emption (Dahl et al., 1990; Lascelles et al., 1997, 1998; Moiniche et al., 1997). One study has shown wound hyperalgesia to be gone at 8 days postoperatively (Moiniche et al., 1997). Only one clinical animal study has investigated thermal hyperalgesia after surgery (Welsh and Nolan, 1995) demonstrating early (< 1 h postoperatively) distant hyperalgesia, reduced by opioid pre-emption, after laparotomy. Electrical sensory thresholds are particularly easy to use in the clinical context and have been extensively validated experimentally (Rollmann and Harris, 1987; Lautenbacher and Rollman, 1993). Electrical stimulation has the additional advantage of producing multimodal sensory stimulation, stimulating both large and small nerve fibres, which may be particularly relevant to the multimodal sensory input resulting from surgery (Arendt-Nielsen et al., 1994). We have used electrical stimulation to study postoperative sensory change both in back surgery (prolapsed intervertebral discs) and in abdominal surgery (hysterectomies) (Wilder-Smith et al., 1996, 1998). Compared to preoperative values, dermatomal electric stimulation distant to the surgical incision shows early (~24 h postoperatively) increases in both nociceptive and non-nociceptive thresholds, increased by pre-emptive analgesia (Figs. 2 and 3). Closer to the wound (secondary hyperalgesia) the absolute nociceptive and non-nociceptive thresholds also show early increases (l-24 h postoperatively), with a tendency to be higher than at sites distant to surgery (Figs. 2 and 3). Again, this hypoalgesia and hyposensitivity is augmented by opioid pre-emption. It is unlikely to be explained solely by postoperative morphine analgesia, as it is also visible for non-nociceptive thresholds which are not directly affected by opioids (Van der Burght et al., 1994). If, however, thresholds close to the wound relative to the distant thresholds are calculated (i.e. for a given time: threshold close to wound divided by threshold distant from wound) to remove any generalised inhibitory effects, these relative thresholds close to the wound show an early (l-24 h postoperatively) reduction compared to preoperatively in the absence of opioid pre-emption (Fig. 4), suggesting central neuronal sensitisation. This sensitisation is visible in

et al., 1998

et al., 1999~3

Wilder-Smith

Wilder-Smith

I, 2,4,6

h

I h-5 d 4h,id I h-5 d 2,4 d, I m

4, 6 h, 5 d

1,5d

-

ABS: ABS: REL: REL: ABS: REL: .3

t(NPE) tt(PE) tt(NPE/PE) $(NPE) tt(PE) tt ff

-

4,6h, 1.4d 8d 1,3,7d

24,48 h

Dahl et al., 1990 Dahl et al., 1992 Dahl et al., 1990 Wilder-Smith et al., 1996

$(PE)

$(PE: J.area)

*

REL: J(NPE)

Stubhaug, 1997

HUMAN Richmond et al., 1993 Tverskoy et al., 1994 Moiniche et al., 1997

postop

when

et al., 1999a

Wilder-Smith

Near wound (2”) (post vs. preop)

E CD)

hysterectomy

et al., 1998

Wilder-Smith

Study

E CD) E 6’0 M (P) E CD)

hysterectomy gynae lap hernia back (herniated disc)

Lund et al., 1990 D&l et al., 1992 Dahl et al., 1990 Wilder-Smith et al., 1996

E CD)

M (+I

renal

Stubhaug, 1997

hysterectomy

M (0

M (vF) M

hysterectomy hysterectomy hysterectomy

Type of stimulation

sensory change after surgery to date

HUMAN Richmond et al., 1993 Tverskoy et al., 1994 Moiniche et al., 1997

Study

Summary of studies investigating

TABLE 3

ABS: t REL: N/A J

ABS: T(PE) ‘/J(NPE)’ ABS: t) (PE/NPE) REL: N/A

ff ff

Distant to wound (post vs. preop)

s, IT

S. PD, PT

S PIPD S, PD, PT

S, PD, TS

S, PD PD PD

Stimulation endpoint

h

4d

4h

l&4h’/5d’rest

-

48 h 48-96

8d

4,6 h, I,4 d

-

postop

When

ABS ABS ABS ABS REL to am ABS RELtoaml ABS

ABS

REL to forearm ABS ABS

Absolute or I&&X

fentanyl ketamine magnes

fentanyl

k&amine

morphine fentanyl ketamine

Preempt analgesia

Yes

Yes Yes yes Yes

Yes

J(PE)

PCA ma/tram

PCA morphine

IM morphine IM morphine IM motphine PCA morphine

PCA morphine

PCA morphine IV pethidine IM morphine

Postop analgesia

-

J -

-

LJ c, -

JUNPE)

Yes

-

“0

Y

4,6h, 8d

24.48

postop

no

When

On wound (1”)

(postYS.preop)

h 1,4d

rat ovario-hysterectomy dog ovario-hysterectomy dog ovario-hysterectomy

Near wound (2”) (post vs. preop)

Lascelles et al., 1995 Lascelles et al., 1997 Lascelles et al., 1998

Study

(P) (P) (P)

(0

postop

When

M T M M M

Type of stimulation

M: e T: IJWW LOW IWE) ++(W UNPW ++@‘E) IWE) +G’E)

Distant to wound (post vs. preop)

PD PD PD

PT

Stimulation endpoint

or

O-120 min 45,60 tin 2.5-6.5 h 12,20h 8, 12, 20 h 12,20h 8, 12, 20 h

When postop

ABS ABS ABS

ABS

Absolute relative

carprofen

pethidine pethidine

cxprofen

Preempt analgesia

Yes Yes yes

Yes

measures?

Fk0p

-

-

Postop analgesia

C&WE) JUNPE)

UW J@‘W

On wound (1”) (post vs. preop)

8, 12.20 h 8, 12, 20 h

When postop

Abbreviations: M = mechanical; P = pressure; vF = van Frey hair; E = electrical; N = direct nerve stimulation; D = dennatomal stimulation; T = thermal; S = sensation threshold; PD = pain detection threshold; PT = pain tolerance threshold: ABS = absolute: REL = relative threshold; NPE = no pre-emptive analgesia; PE = pre-emptive analgesia; N/A = not applicable; h = hours; d = days; m = months; IM = intramuscular; PCA = patient-controlled analgesia; magnes. = magnesium sulphate. Notes: ’ on arm ‘in leg demmtome most affected by herniated disc prolapse.

Lascelles et al., 1998

Lascelles et al., 1995 Lascelles et al., 1997

ANIMAL Welsh and Nolan, 1995

sheep laparotomy

Type of S”rgC5l.y

ANIh4AL Welsh and Nolan, 1995

Study

TABLE 3 (continued)

516

l

fentanyl

BLlh2h4h6hldW

arm

q

placebo

BLlh2h4h6hld5d

contralateral

BLlh2h4h6hld5d

ipsilateral

BLIh2h4h6hld5d

affected

dermatome

Fig. 2. Absolute thresholds by electric skin stimulation in mA (means, standard deviations) at preoperative baseline (BL) and 1, 2, 4, 6 and 24 h and 5 days after surgery for herniated intervertebral discs (after Wilder-Smith et al., 1996). Anaesthesia: isoflurane/nitrous oxide/oxygen % 3 kg/kg fentanyl i.v. before intubation. Sites of measure: arm, contralateral and ipsilateral to the back incision, and the dermatome of the nerve most affected by disc prolapse. Thresholds measured: sensation (,Sn, pain detection (PM’) and pain tolerance (Pm. Significant overall statistics (repeated measures ANOVA for drug, threshold site, threshold type and time): fentanyl 1 placebo; PTT > PDT 1 ST; arm -C contralateral = ipsilateral = affected dermatome; 4 h > BL. Significances (p cc 0.05) for specific times, sites and measures are marked on the graph: * = significant vs. BL, with values for placebo marked above and for fentanyl marked below the curves.

non-nociceptive thresholds not directly affected by opioid analgesia. Central sensitisation is no longer visible in the presence of analgesic supplementation by opioids or NMDA antagonists (Figs. 4 and 5). Using another technique, namely directly electrically stimulating a nerve distant to the site of surgery, but innervated by spinal cord segments convergently involved in the surgery, another group (Dahl et al., 1992) has also been able to demonstrate hyperalgesia and central neuronal sensitisation 48-96 h postoperatively. Studies of the process whereby pain becomes chronic (see Jensen and Nikolajsen, 2000, this volume) suggest that a central nervous system already

sensitised by preceding nociception, e.g. ischemic pain before limb amputation (Bach et al., 198X), is more vulnerable to further neuroplastic change and its chronification. An advantage of studying sensory change after surgery for prolapsed intervertebral discs is that many of these patients suffer pain preoperatively and may thus show central nervous system sensitisation preoperatively. In a study of preoperative sensory change in patients scheduled for hemiated disc surgery (Wilder-Smith et al., 1999b), we were able to demonstrate that patients with moderate to severe preoperative pain (VAS > 5) showed pain thresholds significantly different from those without preoperative pain (Fig. 6). Of note is the fact that

517

o ketamine

BL

lh

4h arm

Id

q magnesium

5d

BL

lh

4h thorax

l fentanyl

Id

5d

BL

lh

4h

Id

5d

BL

lh

4h

Id

5d

opsite

leg Fig. 3. Absolute thresholds by electric skin stimulation in mA (means, standard deviations) at preoperative baseline and 1, 4 and 24 h and 5 days after surgery for abdominal hysterectomy (after Wilder-Smith et al., 1998). Anaesthesia: isoflurane/nitrous oxide/oxygen supplemented by either fentanyl, magnesium or ketamine. Sites of measure: arm, operation site, thorax and upper thigh. Thresholds measured: sensation (ST), pain detection (PD7J and pain tolerance (P77’). Significant overall statistics (repeated measures ANOVA for drug, threshold site, threshold type and time): PTT z PDT > ST; l-24 h > BL. There is no significant effect by drug group. Significances (p < 0.05) for specific times, sites and measures are marked on the graph: * = significant vs. BL, with values for fentanyl marked above and for ketamine marked below the curves.

while somatic pain (i.e. pain in the back) was associated with lowered pain tolerance thresholds to electrical stimulation - congruent with central sensitisation - pain of neuropathic quality (i.e. radiating down the leg) was associated with increased pain thresholds, suggesting either inhibitory processes or nerve dysfunction. Interestingly, in the study of back surgery mentioned above (Wilder-Smith et al., 1996), the patients not receiving analgesic supplementation for anaesthesia not only continued showing lowered relative thresholds compared to preoperatively in the dermatome most affected by disc prolapse 5 days postoperatively, they also had lower absolute electric pain tolerance thresholds on day 5 (Figs. 2 and 4). It should be noted in this context that different sensory modalities of measurement may give somewhat differing results, e.g. mechanical thresholds

may be less sensitive to descending inhibition than thermal or electrical measures (cf. also, e.g. Lautenbather and Rollman, 1993). In summary, to date there is evidence for the following central changes in sensory processing following surgical nociception: (1) distant from the wound: early hypoalgesia (up to ca. 12-24 h postoperatively), increased by preemptive analgesia, affecting both nociceptive and non-nociceptive sensory processing; later modest hyperalgesia possible (up to 4-7 d postoperatively), decreased or abolished by pre-emptive analgesia, probably increased by preoperative sensory sensitisation (preoperative pain); (2) close to the wound: early absolute hypoalgesia (up to 12-24 h postoperatively), increased by pre-emptive analgesia; modest relative/mechanical

518 l

fentanyl

BL lh

0 placebo

2h 4h 6h Id ipsilateral

5d

BL lh

2h 4h 6h Id contralateral

5d

BL lh 2h 4h 6h Id affected dermatome

5d

Fig. 4. Relative thresholds by electric skin stimulation (ratio; current threshold/arm threshold) (means, standard deviations) at preoperative baseline (BL) and 1, 2, 4, 6 and 24 h and 5 days after surgery for herniated intervertebral discs (after Wilder-Smith et al., 1996). Anaesthesia: isoflurane/nitrous oxide/oxygen I!C 3 Kg/kg fentanyl i.v. before intubation. Sites of measure: contralateral and ipsilateral to the back incision, and the dermatome of the nerve most affected by disc prolapse. Thresholds determined: sensation (So, pain detection (PDT) and pain tolerance (Pm. Significant overall statistics (repeated measures ANOVA for drug, threshold site, threshold type and time): PTT r PDT > ST. For ST: placebo c fentanyl. Significances (p < 0.05) for specific times, sites and measures are marked on the graph: * = significant vs. BL, with values for placebo marked above and for fentanyl marked below the curves.

hyperalgesia (up to 5-7 d postoperatively) decreased or abolished by pre-emptive analgesia with opioid agonists or NMDA receptor antagonists, probably increased by preoperative sensory sensitisation (preoperative pain); (3) on the wound: marked hyperalgesia (up to 4-5 d postoperatively), decreased but not abolished by analgesic pre-emption. Thus psychophysical measures of sensory processing in the context of human surgery provide evidence that central nervous system neuroplastic change - both inhibitory and excitatory - takes place after surgical nociception and that this is positively influenced by relatively modest pre-emptive analgesic intervention. The presence of inhibitory central neuroplastic change is not one which would be predicted by or detectable in non-intact animal models of nociception, hence providing at least one

partial explanation of the discrepancies between experimental animal and clinical human models of pre-emptive analgesia. In the studies cited, wound hyperalgesia tends to be more marked than secondary hyperalgesia and not completely abolished by pre-emptive analgesic intervention, suggesting that peripheral primary nociceptor sensitisation (see Reeh and Petho, 20, this volume) after nociception plays an important role in acute postoperative pain. Altered sensory processing and clinical pain after surgery

Finally, we must turn to the question of whether there is a simple relationship between sensory change and clinical pain measures after surgical nociception. Of the selection of studies surveyed above, only one has demonstrated a formal correlation between

519

0 ketamine

0 magnesium

0 fentanyl

$.-~~~I

‘E

;iGjl

ikjj:__.I. BL

lh

4h

Id

5d

Fig. 5. Relative thresholds by electric skin stimnlation (ratio; current threshold/arm threshold) (means, standard deviations) at preoperative baseline (BL) and 1, 4 and 24 h and 5 days after surgery for abdominal hysterectomy (after Wilder-Smith et al., 1998), measured close to site of surgical incision. Anaesthesia: isoflurane/nitrous oxide/oxygen supplemented by either fentanyl, magnesium or ketamine. Thresholds measured: sensation (Si”), pain detection (PDT) and pain tolerance (PIU). Significant overall statistics (repeated measures ANOVA for drug, threshold site, threshold type and time): for fentanyl, thresholds overall l-4 h > BL. There were no significant differences for specific times, sites and measures.

altered sensory processing and clinical pain measures. Moiniche et al. (1997), studying renal surgery patients, showed a modest correlation (r = -0.4) between secondary mechanical wound hyperalgesia and pain VAS at rest or on coughing. Some studies have found differences in primary (Tverskoy et al., 1994) or secondary wound hyperalgesia (Richmond et al., 1993; Stubhaug, 1997) due to analgesic pre-emption to be reflected by modest, mainly early and relatively short-lasting differences in pain intensity VAS or postoperative analgesia consumption. In our studies of back surgery and hysterectomy, the clear differences in sensory processing after surgery allied to differences in perioperative analgesic management were not reflected in differences in clinical pain measures such as pain VAS or morphine PCA consumption (Wilder-Smith et al., 1996, 1998). We have found no other studies establishing formal correlations between changes in sensory processing and clinical pain measures.

On present evidence we must conclude that while a link between altered sensory processing and clinical pain measures after surgery may well be present, it is likely to be weak. Such an outcome is to be expected on the basis of the multifactorial and multiaetiological nature of clinical pain, as explained above. Clearly, further investigations into the nature of the relationship between objective alterations in central sensory processing and subjective clinical pain measures are needed. This is even more the case for long-term pain and medical outcomes after surgery and their relation to alterations in sensory processing, as no studies of this type have been published so far. Conclusions The concept of pre-emptive analgesia purports that an analgesic intervention commenced before a nociceptive event will be more effective than the same

520

25

20

15

Cl

Cl

z

Cl

; 10

f..

-

!

5

t

is

0 n

b

I arm

bl

n

b I contralateral

bl

n

b I ipsilateral

bl

n

b affected

I bl derm atom e

Fig. 6. Absolute thresholds by electric skin stimulation (means, standard deviations) preoperatively, before surgery for herniated intervertebral discs (after Wilder-Smith et al., 1999b). Sites of measure: arm, contralateral and ipsilateral to the planned back incision, and the derxnatome of the nerve most affected by disc prolapse. Thresholds measured: sensation (ST), pain detection (PDT) and pain tolerance (PZT). Significant overall statistics (ANOVA for drug, threshold site, threshold type and pain presence): highly significant (p < 0.000001) effect of presence of moderate to severe pain (VAS z 5) (n = no pain; b = only pain in back; 1 = only pain radiating into leg; bl = both back and leg pain present) on thresholds. For PTT overall: 1 > n > b (p c 0.05).

analgesic intervention practised afterwards. Originally postulated on the basis of animal studies demonstrating central nervous system plasticity after nociception, this idea was introduced to clinical medicine with’the hope of achieving substantial improvements in postoperative pain management. Unfortunately, such substantial improvements have not been forthcoming in the clinical context. In this chapter we have outlined why it has proven difficult to achieve the scale of improvement suggested to be possible by animal experimental models in the clinical arena. These difficulties involve obstacles to extrapolating from experimental models to the clinical situation, the challenges of achieving adequate clinical study designs, as well as problems and confusion regarding the choice of study endpoints relevant to pain outcomes in the clinical context. Regarding pain endpoints for clinical studies, these need to be closer to the objective electrophysiological measures of altered central nervous system sensory processing used in animal studies of

pre-emptive analgesia and neuroplasticity after nociception. We suggest that psychophysical testing, e.g. by sensory thresholds, provides such an objective measure suitable for clinical use. The altered central sensory processing reflected by psychophysical testing is much more likely to give a strong measure of neuroplasticity and pre-emptive analgesia than multifactorial and multi-aetiological subjective clinical pain measures such as pain intensity scales or postoperative analgesic use. In summarising the data available on acute changes in central sensory processing after surgery to date we do in fact find clear evidence of acute neuroplastic change after surgery in humans (and animals). This neuroplasticity involves both inhibitory (e.g. descending inhibition) and excitatory (e.g. spinal sensitisation) components, whose net manifestation depends both on the time and place of measure. Preemptive analgesia with opioid agonists or NMDA receptor antagonists has a positive effect on both inhibitory (reinforcement) and excitatory (suppres-

521

sion) neuroplastic change after surgery. Evidence is further discussed that preoperative pain is also associated with altered central sensory processing, and thus perhaps with increased vulnerability to further post-nociceptive neuroplastic change, particularly in the absence of analgesia during surgical nociception. At present these neuroplastic changes in central nervous system sensory processing have been demonstrated acutely, i.e. for up to 7 days postoperatively. There appears to be a weak relationship between changes in psychophysical measures and clinical subjective pain measures (pain VAS, analgesia use), but any such correlations are little and poorly defined at present. No studies have investigated the relationship between psychophysical measures of neuroplasticity and longer-term, chronic pain or medical outcomes after surgery to date. More studies are needed to better define the relationship between surgical nociception, neuroplastic change in central nervous system sensory processing, subjective clinical pain measures, and long-term chronic pain and medical outcomes of surgery. References Aguilar, J.L., Rincon, R., Domingo, V., Espachs, P., Preciado, M.J. and Vidal, F. (1996) Absence of an early pre-emptive effect after thoracic extradural bupivacaine in thoracic surgery. BI: J. Anaesth., 16: 12-76. Arendt-Nielsen, L., Brennum, J., Sindmp, S. and Bak, P (1994) Electrophysiological and psychophysical quantification of temporal summation in the human nociceptive system. Eur: J. Appl. Physiol., 68: 266-213. Arendt-Nielsen, L., Petersen-Felix, S., Fischer, M., Bak, P., Bjerring, P. and Zbinden, A.M. (1995) The effect of NMDA-antagonist (ketamine) on single and repeated nociceptive stimuli: a placebo-controlled human study. Anesth. Analg., 81: 63-68. Bach, S., Noreng, M.F. and Tjellden, N.U. (1988) Phantom limb pain in amputees during the first 12 months following limb amputation, after preoperative lumbar epidural blockade. Pain, 33:297-301. Bajaj, P., Graven-Nielsen, T., Wright, A., Davies, 1.1. and ArendtNielsen, L. (1999) Ultrasonic stimulation for assessment of muscle hyperalgesia and temporal summation of muscle pain. In: Abstract Band, World Congress of Pain. IASP Press, Seattle, WA, p. 515. Bourget, J.L., Clark, J. and Joy, N. (1997) Comparing preincisional with postincisional bupivacaine infiltration in the management of postoperative pain. Arch. Surg., 132: 766-769. Bridenbaugh, P.O. (1994) Preemptive analgesia - is it clinically relevant?. Anesth. Analg., 78: 203-204. Bromm, B., Scharein, E. and Vahle-Hinz, C. (2000) Cortex areas

involved in the processing of normal and altered pain. In: J. Sandkiihler, B. Bromm and G.F. Gebhart (Eds.), Nervous System Plasticity and Chronic Pain. Progress in Brain Research, Vol. 129. Elsevier, Amsterdam, pp. 289-302. Bruster, S., Jarmaun, B., Bosanquet, N., Weston, D., Erens, R. and Delbanco, T.L. (1994) National survey of hospital patients. B,: Med. .I., 309: 1542-1546. Campbell, WI., Kendrick, R.W. and Fee, J.P. (1998) Balanced pre-emptive analgesia: does it work? A double-blind, controlled study of bilaterally symmetrical oral surgery. BK J. Anaesth., 8 1: 727-730. Casey, K.L. (2000) Concepts of pain mechanisms: The contribution of functional imaging of the human brain. In: J. Sandktihler, B. Bromm and G.F. Gebhart (Eds.), Nervous System Plasticity and Chronic Pain. Progress in Brain Research, Vol. 129. Elsevier, Amsterdam, pp. 277-287. Cervero, F. (1995) Visceral pain: mechanisms of peripheral and central sensitisation. Ann. Med., 27: 235-239. Chia, Y.Y., Liu, K., Chow, L.H. and Lee, T. (1999) The preoperative administration of intravenous dextromethorphan reduces postoperative morphine consumption. An&h. Analg., 89: 748752. Choe, H., Choi, Y.S., Kim, Y.H., Ko, S.H., Choi, H.G., Han, Y.J. and Song, H.S. (1997) Epidural morphine plus ketamine for upper abdominal surgery: improved analgesia from preincisional versus postincisional administration. Anesth. Analg., 84:560-563. Coderre, T.J., Katz, J., Vaccarino, A.L. and Melzack, R. (1993) Contribution of central neuroplasticity to pathological pain: review of clinical and experimental evidence. Pain, 52: 259285. Dahl, J.B. and Kehlet, H. (1993) The value of preemptive analgesia in the treatment of postoperative pain. BI: J. Anaesth., 70: 434-439. Dahl, J.B., Rosenberg, J., Molke-Jensen, F. and Kehlet, H. (1990) Pressure pain thresholds in volunteers and hemiorrhaphy patients. Acta Anaesthesiol. Stand., 34: 673-676. Dahl, J.B., Erichsen, C.J., Fuglsang-Frederiksen, A. and Kehlet, H. (1992) Pain sensation and nociceptive reflex excitability in surgical patients and human volunteers. BK J. Anaesth., 69: 117-121. Dickenson, A.H. (1995) Central acute pain mechanisms. Ann. Med., 21: 223-227. Dickenson, A.H. and Sullivan, A.F. (1986) Electrophysiological studies on the effect of intrathecal morphine on nociceptive neurons in the rat dorsal horn. Pain, 42: 21 l-222. Dostrovsky, J.O. (2000) Role of thalamus in pain. In: J. Sandktihler, B. Bromm and G.F. Gebhart (Eds.), Nervous System Plasticity and Chronic Pain. Progress in Brain Research, Vol. 129. Elsevier, Amsterdam, pp. 245-257. Fassoulaki, A., Sarantopoulos, C., Zotou, M. and Papoulia, D. (1995) Preemptive opioid analgesia does not influence pain after abdominal hysterectomy. Can. .I. Anaesth., 42: 109-l 13. Fletcher, D., Kayser, V. and Guilbaud, G. (1996) Influence of timing of administration on the analgesic effect of bupivacaine infiltration in carrageenin-injected rats. Anesthesiology, 84: 1129-1137.

522

Flor, H. (2000) The functional organization of the brain in chronic pain. In: J. Sandkiihler, B. Bromm and GE Gebhart (Eds.), Nervous System Plasticity and Chronic Pain. Progress in Brain Research, Vol. 129. Elsevier, Amsterdam, pp. 313322. Fu, E.S., Miguel, R. and Scharf, J.E. (1997) Preemptive ketamine decreases postoperative narcotic requirements in patients undergoing abdominal surgery. Anesth. Analg., 84: 1086-1090. Gautron, M. and Guilbaud, G. (1982) Somatic responses of ventrobasal thalamic neurones in polyarthritic rats. Brain Rex, 231: 459-47 1. Gavrilov, L.R., Tsirulnikov, E.M. and Davies, LA. (1996) Application of focused ultrasound for the stimulation of neural structures. Ultrasound Med. Biol., 22: 179-192. Gerber, G., Youn, D.-H., Hsu, C.H., Isaev, D. and Randic, M. (2000) Spinal dorsal horn synaptic plasticity: Involvement of group I metobotropic glutamate receptors. In: J. Sandktihler, B. Bromm and G.F. Gebhart (Eds.), Nervous System Plasticiiy and Chronic Pain. Progress in Brain Research, Vol. 129. Elsevier, Amsterdam, pp. 115-134. Gottschalk, A., Smith, D.S., Jobes, D.R., Kennedy, S.K., Lally, S.E., Noble, V.E., Grugan, K.F., Seifert, H.A., Cheung, A., Malkowicz, S.B., Gutsche, B.B. and Wein, A.J. (1998) Preemptive epidural analgesia and recovery from radical prostatectomy: a randomized controlled trial. JAMA, 279: 1076-1082. Griffin, M.J., Hughes, D., Knaggs, A., Donnelly, M.B. and Boylan, J.F. (1997) Late-onset preemptive analgesia associated with preincisional large-dose alfentanil. Anesth. Analg., 85: 1317-1321. Guilbaud, G., Benoist, J.M., Eschalier, A., Gautron, M. and Kayser, V. (1989) Evidence for peripheral serotonergic mechanisms in the early sensitization after carrageenin-induced inflammation: electrophysiological studies in the ventrobasal complex of the rat thalamus using a potent specific antagonist of peripheral 5HT receptors. Brain Res., 502: 187-197. Heinke, W. and Grimm, D. (1999) Preemptive effects caused by co-analgesia with ketamine in gynecological laparotomies?. Anaesthesiol. Reanim., 24: 60-64. Ho, J.W., Khambatta, H.J., Pang, L.M., Siegfried, R.N. and Sun, L.S. (1997) Preemptive analgesia in children. Does it exist?. Reg. Anesth., 22: 125-130. Hoheisel, U. and Mense, S. (2000) The role of spinal nitric oxide in the control of spontaneous pain following nociceptive input. In: J. Sandktihler, B. Bromm and G.F. Gebhart (Eds.), Nervous System Plasticity and Chronic Pain. Progress in Brain Research, Vol. 129. Elsevier, Amsterdam, pp. 163-172. Jayaram, A., Singh, P. and Carp, H.M. (1995) An enkephalinase inhibitor, SC 32615 augments analgesia induced by surgery in mice. Anesthesiology, 82: 1283-1287. Jensen, S.T. and Nikolajsen, L. (2000) Pre-emptive analgesia in postamputation pain: an update. In: J. Sandkiihler, B. Bromm and G.F. Gebhart (Eds.), Nervous System Plasticity and Chronic Pain. Progress in Brain Research, Vol. 129. Elsevier, Amsterdam, pp. 493-503. Katz, J., Kavanagh, B.P., Sandier, A.N., Nierenberg, H., Boylan, J.F., Friedlander, M. and Shaw, B.F. (1992) Preemptive

analgesia. Clinical evidence of neuroplasticity contributing to postoperative pain. Anesthesiology, 77: 439-446. Ke, R.W., Portera, S.G., Bagous, W. and Lincoln, S.R. (1998) A randomized, double-blinded trial of preemptive analgesia in laparoscopy. Obstet. Gynecol., 92: 972-975. Kehlet, H. (1994) Postoperative pain relief - what is the issue?. BI: J. Anaesth., 72: 315-378. Kelly, D.D. (1986) Stress-induced analgesia. Ann. NY Acud. Sci., 467: l-449. Kest, B., Jenab, S., Brodsky, M., Sadowski, B., Belknap, J.K., Mogil, J.S. and Inturrisi, C.E. (1999) Mu and delta opioid receptor analgesia, binding density, and mRNA levels in mice selectively bred for high and low analgesia. Brain Res., 816: 381-389. Kucuk, N., Kizilkaya, M. and Tokdemir, M. (1998) Preoperative epidural ketamine does not have a postoperative opioid sparing effect. Anesth. Analg., 87: 103-106. Kundra, I?, Gumani, A. and Bhattacharya, A. (1997) Preemptive epidural morphine for postoperative pain relief after lumbar laminectomy. Anesth. Analg., 85: I35- 138. Kundra, I?, Deepalakshmi, K. and Ravishankar, M. (1998) Preemptive caudal bupivacaine and morphine for postoperative analgesia in children. Anesth. A&g., 87: 52-56. Lascelles, B.D., Waterman, A.E., Cripps, P.J., Livingston, A. and Henderson, G. (1995) Central sensitization as a result of surgical pain: investigation of the pre-emptive value of pethidine for ovariohysterectomy in the rat. Pain, 62: 201212. Lascelles, B.D., Cripps, P.J., Jones, A. and Waterman, A.E. (1997) Post-operative central hypersensitivity and pain: the pre-emptive value of pethidine for ovariohysterectomy. Pain, 73: 461-471. Lascelles, B.D., Cripps, P.J., Jones, A. and Waterman-Pearson, A.E. (1998) Efficacy and kinetics of carprofen, administered preoperatively or postoperatively, for the prevention of pain in dogs undergoing ovariohysterectomy. Vet.Surg., 27: 568-582. Lautenbacher, S. and Rollman, G.B. (1993) Sex differences in responsiveness to painful and non-painful stimuli are dependent upon stimulation method. Pain, 53: 255-264. Le Bars, D., Dickenson, A.H. and Besson, J.C. (1979) Diffuse noxious inhibitory controls (DNIC), I. Effects on dorsal horn convergent neurones in the rat. Pain, 10: 283-304. Le Bars, D., Willer, J.C. and De Broucker, T. (1992) Morphine blocks descending pain inhibitory controls in humans. Pain. 48: 13-20. Lenz, EA., Lee, J.-I., Garonzik, I.-M., Rowland L.H., Dougherty, PM. and Hua, S.E. (2000) Plasticity of pain-related neuronal activity in the human thalamus. In: J. Sandktihler, B. Bromm and G.F. Gebhart (Eds.), Nervous System Plasticity and Chmnic Pain. Progress in Brain Research, Vol. 129. Elsevier, Amsterdam, pp. 259-273. Likar, R., Krumpholz, R., Mathiaschitz, K., Pipam, W., Burtscher, M., Ozegovic, G., Breschan, C., Bematzky, G. and Sittl, R. (1997) The preemptive action of ketoprofen. Randomized, double-blind study with gynecologic operations. Anaesthesist, 46: 186-I 90. Likar, R., Krumpholz, R., Pipam, W., Sadjak, A., Kapral, S.,

523 Forsthuber, E., Bematzky, G. and List, F.W. (1998) Randomized, double-blind study with ketoprofen in gynecologic patients. Preemptive analgesia study following the BrevikStubhaug design. Anaesthesist, 47: 303-3 10. Lund, C., Hansen, O.B. and Kehlet, H. (1990) Effect of surgery on sensory threshold and somatosensory evoked potentials after skin stimulation. BK J. Anaesth., 65: 173-176. Lutfy, K., Sadowski, B., Kwon, I.S. and Weber, E. (1994) Morphine analgesia and tolerance in mice selectively bred for divergent swim stress-induced analgesia. Eur .I. Pharmacol., 265: 171-174. McQuay, H.J. (1992) Pre-emptive analgesia. Br J. Anaesth., 69: l-3. McQuay, H.J. (1995) Pre-emptive analgesia: a systematic review of clinical studies. Ann. Med., 27: 249-256. McQuay, H.J. and Dickenson, A.H. (1990) Implications of nervous system plasticity for pain management. Anaesthesia, 45: 101-102. Meller, S.T. and Gebhart, G.F. (1993) Nitric oxide (NO) and nociceptive processing in the spinal cord. Pain, 52: 127-136. Mogil, J.S., Flodman, P., Spence, M.A., Sternberg, W.F., Kest, B., Sadowski, B., Liebeskind, J.C. and Belknap, J.K. (1995) Oligogenic determination of morphine analgesic magnitude: a genetic analysis of selectively bred mouse lines. Behav. Genet., 25: 397-406. Moiniche, S., Dahl, J.B., Erichsen, C.J., Jensen, L.M. and Kehlet, H. (1997) Time course of subjective pain ratings, and wound and leg tenderness after hysterectomy. Acta Anaesthesiol. &and., 41: 785-789. Moore, K.A., Baba, H. and Woolf, C.J. (2000) Synaptic transmission and plasticity in the superficial dorsal horn. In: J. Sandktlhler, B. Bromm and G.F. Gebhart (Eds.), Nervous System Plasticity and Chronic Pain. Progress in Brain Research, Vol. 129. Elsevier, Amsterdam, pp. 63-80. Pasqualucci, A., De Angelis, V., Contardo, R., Colo, F., Terrosu, G., Donini, A., Pasetto, A. and Bresadola, E (1996) Preemptive analgesia: intraperitoneal local anesthetic in laparoscopic cholecystectomy. A randomized, double-blind, placebo-controlled study. Anesthesiology, 85: 1 l-20. Reeh, P. and Petho, G. (2000) Nociceptor excitation by thermal sensitization - a hypothesis. In: J. Sandkiihler, B. Bromm and G.F. Gebhart (Eds.), Nervous System Plasticity and Chronic Pain. Progress in Brain Research, Vol. 129. Elsevier, Amsterdam, pp. 39-50. Richmond, C.E., Bromley, L.M. and Woolf, C.J. (1993) Preoperative morphine pre-empts postoperative pain. Lancet, 342: 73-75. Rockemann, M.G., Seeling, W., Bischof, C., Borstinghaus, D., Steffen, P. and Georgieff, M. (1996) Prophylactic use of epidural mepivacaine/morphine, systemic diclofenac, and metamizole reduces postoperative morphine consumption after major abdominal surgery. Anesthesiology, 84: 1027-1034. Rollmann, G.B. and Harris, H. (1987) The detectability, discriminability, and perceived magnitude of painful electric shock. Percept. Psychophys., 42: 247-268. Romej, M., Voepel-Lewis, T., Merkel, S.I., Reynolds, P.I. and Quinn, P (1996) Effect of preemptive acetaminophen on post-

operative pain scores and oral fluid intake in pediatric tonsillectomy patients. AANA J., 64: 535-540. Sandktihler, I., Benrath, J., Brechtel, C., Ruscheweyh, R. and Heinke, B. (2000) Synaptic mechanisms of hyperalgesia. In: J. Sandkiihler, B. Bromm and G.F. Gebhart (Eds.), Nervous System Plasticity and Chronic Pain. Progress in Brain Research, Vol. 129. Elsevier, Amsterdam, pp. 81-100. Stemberg, W.F. and Liebeskind, J.C. (1995) The analgesic response to stress: genetic and gender considerations. Eur J. Anaesthesiol., 10 (Suppl.): 14-17. Stubhaug, A. (1997) A new method to evaluate central sensitization to pain following surgery. Effect of ketamine. Acta Anaesthesiol. Stand. (Suppl.), 110: 154-155. Svendsen, F., Hole, K. and Tjolsen, A. (2000) Long-term potentiation in single WDR neurons induced by noxious stimulation in intact and spinalized rats. In: J. Sandktlhler, B. Bromm and G.F. Gebhart (Eds.), Nervous System Plasticity and Chronic Pain. Progress in Brain Research, Vol. 129. Elsevier, Amsterdam, pp. 153-161. Termann, G.W., Penner, E.R. and Liebeskind, J.C. (1986) Stimulation-produced and stress-induced analgesia: cross-tolerance between opioid forms. Brain Res., 372: 167-17 1. Treede, R.D. (1995) Peripheral acute pain mechanisms. Ann. Med., 27: 213-216. Treede, R.-D. and Magerl, W. (2000) Multiple mechanisms of secondary hyperalgesia. In: J. Sandkiihler, B. Bromm and G.F. Gebhart (Eds.), Nervous System Plasticity and Chronic Pain. Progress in Brain Research, Vol. 129. Elsevier, Amsterdam, pp. 331-341. Tverskoy, M., Oz, Y., Isakson, A., Finger, J., Bradley, E.L. and Kissin, I. (1994) Preemptive effect of fentanyl and ketamine on postoperative pain and wound hyperalgesia. Anesth. Analg., 78: 205-209. Tverskoy, M., Oren, M., Dashkovsky, I. and Kissin, I. (1996) Alfentanil dose-response relationships for relief of postoperative pain. Anesth. Analg., 83: 387-393. Van der Burght, M., Rasmussen, S.E., Arendt-Nielsen, L. and Bjerring, P (1994) Morphine does not affect laser-induced warmth and pin prick thresholds. Acta Anaesthesiol. Stand., 38: 161-164. Wall, P.D. (1988) The prevention of postoperative pain. Pain, 33: 289-290. Warfield, C.A. and Kahn, C.H. (1995) Acute pain management. Programs in US hospitals and experiences and attitudes among US adults. Anesthesiology, 83: 1090-1094. Welsh, E.M. and Nolan, A.M. (1995) The effect of abdominal surgery on thresholds to thermal and mechanical stimulation in sheep. Pain, 60: 159-166. Wilder-Smith, O.H. (1995) Preemptive analgesia. Anaesthesist, 44(Suppl. 3): 529-534. Wilder-Smith, O.H., Tassonyi, E., Senly, C., Otten, P. and Arendt-Nielsen, L. (1996) Surgical pain is followed not only by spinal sensitization but also by supraspinal antinociception. Br J. Anaesth., 76: 816-821. Wilder-Smith, O.H., Arendt-Nielsen, L., Gaumann, D., Tassonyi, E. and Rifat, K.R. (1998) Sensory changes and pain after abdominal hysterectomy: a comparison of anesthetic supplemen-

524 tation with fentanyl versus magnesium or ketamine. Anesth. Analg., 86: 95-101. Wilder-Smith, C.H., Hill, L., Wilkins, J. and Denny, L. (1999a) Effects of morphine and tramadol on somatic and visceral sensory function and gastrointestinal motility after abdominal surgery. Anesthesiology, 91: 639-647. Wilder-Smith, O.H., Tassonyi, E. and Arendt-Nielsen, L. (1999b) Somatic and neuropathic pain have opposing effects on pain thresholds. Dolor, 14(Suppl. III): 11. Woolf, C.J. (1983) Evidence for a central component of postinjury pain hypersensitivity. Nurure, 306: 686-688. Woolf, C.J. (1989) Recent advances in the pathophysiology of acute pain. BK J. Annesth., 63: 139-146. Woolf, Cl. and Chong, MS. (1993) Preemptive analgesia treating postoperative pain by preventing the establishment of central sensitisation. Anesth. At&g., 77: 362-379. Woolf, C.J. and Thompson, S.W.N. (1991) The induction

and maintenance of central sensitisation is dependent on N-methyl-D-aspartic acid receptor activation: implications for post-injury pain hypersensitivity states.Pain, 44: 293-299. Woolf, C.J. and Wall, P.D. (1986a) Relative effectiveness of C primary afferent fibres of different origins in evoking a prolonged facilitation of the flexion reflex in the rat. J. Neurosci., 6: 1433-1442. Woolf, C.J. and Wall, P.D. (1986b) Morphine-sensitive and morphine-insensitive actions of C-fibre input on the rat spinal cord. Neurosci. Lett., 64: 221-225. Wu, CT., Yu, J.C., Yeh, CC., Liu, S.T., Li, C.Y., Ho, S.T. and Wong, C.S. (1999) Preincisional dextromethorphan treatment decreases postoperative pain and opioid requirement after laparoscopic cholecystectomy. Anesth. Analg., 88: 133 I-1334. Zenz, M. (1997) Editorial comment: pain therapy. Cum @in. Anuesthesiol., 10: 367-368.

J. Sandkiihler, B. Bromm and G.F. Gebhart (Eds.) Progress in Brain Research, Vol. 129 0 2000 Elsevier Science B.V. All rights reserved

CHAPTER 38

Attentional control of pain and the process of chronification Monika Hasenbring * Abteilung

Medizinische

Psychologie,

Ruhr-Universitiit Bochum, Gebiiude 44780 Bochum, Germany

Introduction It is now widely acknowledged that cognitive coping strategies can affect the individual perception and tolerance of pain sensations (Eccleston, 1995) and that these mechanisms play an important role in acute laboratory pain as well as in chronic pain states. In contrast, little is known about the impact of specific pain-related cognitions on the development of chronic pain in clinically relevant pain syndromes. For example, epidemiological studies regarding the sequelae of acute low back pain have shown that nearly 35% of the patients who experience an acute low back injury will develop chronic recurrent or persistent pain (Macfarlane et al., 1999). Does the plasticity of somatic interpretation, which is mainly influenced by the individual attentional focus (Cioffi, 1991), also play a role in the chronification of pain? This paper will discuss cognitiveperceptual approaches in laboratory pain research as well as in clinical pain studies. The findings on attentional processes in several research settings are briefly reviewed and are followed by the formulation of hypotheses regarding possible pathogenetic pathways leading into the development of chronic pain. * Corresponding author: M. Hasenbring, Abteilung Medizinische Psychologie, Ruhr-Universitlt Bochum, Geblude MA-O/145 Universitltsstrasse 150, 44780 Bochum, Germany. Tel.: +49 234-322-5439; Fax: +49 234-321-4203; E-mail: [email protected]

MA-O/145,

Universitiitsstrasse

Attentional control of experimental healthy subjects

150,

pain in

Among several attentional strategies for coping with experimental pain (e.g. cold pressor pain stimulation) distraction is the best investigated and seems most effective. According to McCaul and Mallot (1984), distraction can be defined as directing one’s attention away from the sensations or emotional reactions produced by a noxious stimulus. The major theoretical underpinning of the analgesic effect of distraction relies on the limited-capacities model of attention (Kahnemann, 1973). The model holds that controlled information processing is capacity-bound. The allocation of attentional resources to one input or task limits the resources available to other inputs. Thus an attentional strategy may be effective in reducing pain perception as it will compete with nociceptive sensations. Through several different modes of distraction - e.g. to focus attention away from pain by watching slides of landscapes (McCaul and Haugtvedt, 1982), by engaging in various numerical tasks (e.g. Beers and Karoly, 1979; Devine and Spanos, 1990; Hodes et al., 1990) or by pleasant imageries (e.g. Avia and Kanfer, 1980; Rosenbaum, 1980) - these strategies led to a better reduction of pain intensity and distress and to an increase in the length of time that the stimulus was endured (tolerance) in comparison to uninstructed or placebo control conditions. However, numerous studies have shown that strategies of willful monitoring of the somatic sensations of experimental pain stimulation will lead to the

526

same effects as distraction (Leventhal et al., 1979; Ahles et al., 1983; Suls and Fletcher, 1985). Suls and Fletcher (1985) pointed out that these strategies will lessen distress especially when they focus on the concrete characteristics of the physical sensation rather than on diffuse or potentially harmful physical states, such as fatigue or tension. At a first glance, distraction and sensory monitoring seem to be opposite modes of attention direction, therefore the effects were equivocal (Eccleston, 1995). In their intriguing ‘parallel processing’ model of pain, Leventhal and colleagues have suggested that sensory monitoring works to the extent that sensation-distress associations are disrupted, thus allowing the formation of more objective schemata. Focusing on the objective qualities of a physical stimulus may act as a form of distraction from harmful interpretations and distressing emotions. These theoretical considerations make it plausible that directing one’s attention away from the emotional reactions produced by a noxious stimulus would be the essential mechanism of both strategies, distraction and sensory monitoring. This hypothesis was supported by numerous studies which have shown that emotion-focused monitoring (catastrophizing) led to higher pain intensity, higher distress, and a lower level of tolerance (Leventhal et al., 1979; Ahles et al., 1983). In this case attention is directed to potentially threatening consequences of an experimental pain stimulation and to their frequent concomitants, such as unpleasant emotional reactions or diffuse body sensations. At the beginnings of the nineties Cioffi and Holloway (1993) investigated the effects of a further type of attention direction. Using the cold pressor pain stimulus, they compared the effectiveness of distraction, sensory monitoring, and suppression as the new technique. Whereas subjects in the distraction condition were told to form a vivid mental picture of their rooms at home, subjects in the suppression condition only were told to not think about their hand sensations, to eliminate awareness of them. In contrast to former experiments the authors studied short- and long-term outcomes in the laboratory by assessing pain intensity every 20 s after pain stimulation over a period of 2 n-tin and by confirming a further experimental task. 30 min after the pain test a mild vibration stimulus which was experienced as relatively neutral by a pretest sample was applied at

the subject’s neck. Results indicated that subjects in the suppression condition initially showed less intense pain, but later during the 2-min interval they showed more intense pain than the subjects being either distracted or monitored. They also rated the vibration stimulus as being less pleasant compared to distraction and sensory monitoring. Furthermore, the suppressors tended to higher skin conductance levels as an indicator of physiological arousal in both the cold pressure test and the vibration stimulus. Cioffi and Holloway (1993) assumed that especially the instruction to suppression produces a rebound efect, which is well investigated in research on cognition and mental control (e.g. Wegner et al., 1987). This research has shown that when people were asked to suppress awareness of a thought, they were in fact able to do so but this suppression produced a rebound effect - that is, subsequently more frequent occurrences of this thought than if one had not initially suppressed it. Wegner and colleagues attributed this phenomenon to the different goal structures of distraction, sensory monitoring, and suppression. Whereas the goal of distraction is to replace one thought with another, that is, distraction specifies what to do, the goal of suppression is to remove a thought from mind or it specifies what not to do. If people try to get goal-relevant feedback about the success of their mental strategy, in the case of suppression they ask themselves repeatedly how empty their mind will be of the thought X. As a consequence, in the case of pain stimulation, they focus to this pain stimulus again and again. In contrast, distraction and also sensory monitoring instructions give a concrete and vivid mental event, success is evaluated by an awareness of sensations devoid of unpleasant features of the pain stimulus. Furthermore, Wegner and colleagues have found that the instruction to suppress often induces a search for distracters which was disorganized and nonfocused and which has often failed (Wegner et al., 1991). This search for distracters requires more effort than receiving a specific distractor from the experimenter. Furthermore, frequent failure leads to a decreased perceived self-efficacy and to an increase in emotional distress. Summarizing, the short-term effect of suppressive cognitions is a reduced awareness of pain, the long-term effects are repeatedly focusing on the pain experience and increased emotional distress.

527 Attentional control of pain in patients with clinical pain Research on attentional control of pain in clinically relevant chronic pain states has tried to answer several questions. How often are the experimentally investigated attentional strategies used in patients with clinically relevant pain? Which strategies do patients prefer? Is the efficiency of the different attentional strategies in experimental pain comparable to the effects in chronic pain patients? When persons are asked to evaluate coping strategies, distraction is rated as highly effective and is preferred over alternative techniques (McCaul and Haugtvedt, 1982; Ahles et al., 1983). Field studies on clinical pain, for example during medical procedures or childbirth, or distress over chemotherapy or radiotherapy, suggest that deploying sensory monitoring techniques decreases pain and distress more than distraction (Johnson et al., 1973; Love et al., 1983; Nerenz et al., 1984). Nevertheless, the patients preferred distraction techniques. In a laboratory experiment, McCaul and Haugtvedt (1982) have seen that 80% of their subjects would have preferred distraction over attending sensations for coping with cold pressor pain, whereas 92% of the subjects who were instructed to attend to the concrete sensations of the pain stimulus reported less distress than subjects in the distraction condition. A number of studies investigating the correlation of several attentional strategies with measures of adjustment have discovered that distraction was commonly used by chronic pain patients in spite of questionable utility (Jensen et al., 1991; Kleinke, 1992). For example, Rosenstiel and Keefe (1983) and Turner and Clancy (1986) found that chronic pain patients who had higher scores on the factor ‘attention diversion, hoping, and praying’ of the Coping Strategies Questionnaire (CSQ) of Rosenstiel and Keefe showed higher average pain. Also Keefe and Williams (1990) found a positive correlation between the intensity of pain and the attention diversion score. In a laboratory experiment with chronic low back pain (CLBP) patients and healthy controls, Johnson and Petrie (1997) have investigated the influence of a word shadowing distraction task on pain intensity and tolerance during two experimental conditions: a cold pressor test and a brief step-up

exercise that temporarily increased the pain in CLBP patients. Whereas distraction increased cold pressor tolerance for the healthy control group, pain tolerance for the CLBP group showed no change. The CLBP group also reported higher overall pain in the hand than the control group. In contrast, the distraction task increased the step-up tolerance with regard to the time spent exercising and the number of repetitions in the CLBP group. In a group of patients with acute sciatic pain and a lumbar disc prolapse, Hasenbring (1992) found that cognitions of minimizing the pain or cognitions of suppression showed the highest frequency among several other coping strategies, assessed by the Kiel Pain Inventory KPI (Hasenbring, 1994), whereas distraction by vivid positive phantasies was extremely seldom. In a prospective study Hasenbring et al. (1994) have seen that minimizing and suppression belonged to the high-risk factors of the development of chronic pain following conservative or operative treatment, whereas distraction by positive phantasies was a predictor of a good long-term therapy outcome. We can summarize that distraction as well as suppression are less or not effective compared to sensory monitoring with respect to the reduction of pain or distress in clinical medical pain or in chronic pain states. In chronic pain patients it only seems effective in the case of increasing pain tolerance during physical exercises. Nevertheless, distraction and suppression are much more popular among chronic pain patients than sensory monitoring. Several questions remain from this research: what are the reasons for the popularity of distraction and suppression techniques in coping with pain even in the face of contradicting evidence and why do patients not show sensory monitoring spontaneously? What is the reason for the low efficacy of distraction in chronic pain patients compared to the proven efficacy in healthy controls? Cioffi (1991) assumed that willfully focusing on sensory information would rarely be a spontaneous strategy for dealing with clinical pain. We can add that willfully focusing on pain-irrelevant positive phantasies was seldom used spontaneously. One reason could be that minimizing or suppressing pain sensations is easier to deploy than the development of a concrete distress-irrelevant picture. Another pos-

528

sible reason could be that when patients experience pain, for example low back or sciatic pain, the goal of suppression is to avoid the disruption of current activities. Hasenbring and her colleagues (Hasenbring, 1993; Hasenbring et al., 1994) found that strategies of suppression were accompanied by the effort to focus on activities patients were still running in order to finish all activities, often in spite of severe pain. If the hypothesis formulated by Wegner et al. (1991) and Cioffi and Holloway (1993) was true, patients with acute sciatic pain caused by a physical injury must experience a rebound effect with an ongoing switching between their pain and other activities. This process is accompanied by increased distress due to the inefficacy of their coping strategy. We can assume that in the case of repeatedly ineffective coping these patients will experience fear of helplessness and hopelessness in the long term. Cognitions of catastrophizing must be one of these consequences. This hypothesis finds support in the data from Hasenbring and colleagues: cognitions of catastrophizing and help-/hopelessness were nearly as common in patients with sciatic pain as cognitions of suppression. Some patients showed increased scores in both cognitions, helplessness and suppression (Hasenbring, 1993). We can speculate that these patients suffer from a conflict between both types of attentional strategies or it is an expression of the above-described process of switching between suppression and the unpleasant feelings due to pain (see also Eccleston and Crombez, 1999). The reasons for the low efficacy of distraction in chronic pain patients are also not yet clear. Eccleston and colleagues found in a series of experiments that distraction tasks must demand attention to a high degree in order to have an impact on pain (e.g. Eccleston, 1995). They argue that if these strategies in fact demand sufficient attention to be effective, they are themselves effortful and fatiguing. This fatigue may impair the ability to engage in other behaviors which are helpful for patients with chronic pain. Johnson and Petrie (1997) focus on another problem. They argue that an individual who uses distraction tasks for coping with pain might turn off warning signals and without this feedback may exacerbate the injury accompanied by increasing pain. A further reason could be that the instruction to attention diversion increases the conflict between the sponta-

neously preferred suppression and the repeated focus on pain. The newly learned distraction technique competes with the well-known strategy of suppression, and if these two are not effective they compete with the recurrent pain experience. The process of switching is further increased and causes more distress, catastrophizing, and helplessness. Information-processing patients

biases in chronic pain

There is now strong evidence to suggest that patients with chronic long-lasting pain show increased scores on catastrophizing, helplessness, and fear of pain. With the hypervigilance model of pain perception this phenomenon is attributed to a perceptual style of amplification with increased sensitivity to pain (McDermid et al., 1996). Amplification is defined as the tendency to focus on distressing and potentially harmful aspects of a sensoric or visceral sensation (Barsky, 1992, 1998). Attentional biases are therefore implicated as vulnerability factors in the process of chronicity of pain. There is now increasing experimental research investigating attentional biases in chronic pain patients. These studies focus (a) on the recall of past pain experiences, (b) on the processing bias in actual pain stimulation, and (c) on an attentional bias regarding future pain. Recall of past pain

Clinical research has shown that patients who underwent a medical treatment and who show high levels of anxiety tend to an overestimation of treatment-related pain some weeks later (e.g. Kent, 1985, 1989). This phenomenon was also described in chronic pain patients (e.g. Linton and Melin, 1982; Linton, 1991). For example, Linton assessed intensity of pain on a VAS scale three times per day during a one-week baseline period as a part of routine diagnostics. Eighteen months later the patients had to remember the mean intensity of pain during this baseline period. 70% of the participants recalled more intense pain than their baseline levels. This overestimation showed a significant positive correlation with pain-related helplessness but not with depression in general. There is further evidence from experimental research to suggest that chronic pain

patients exhibit a recall bias towards pain-related stimuli (e.g. Pincus et al., 1993; Edwards et al., 1995). Edwards et al. (1995) presented four mixed lists of sensory, affective, neutral, and gardening pain-related and non-pain words to chronic pelvic pain patients undergoing hysterectomy. The patients were investigated prior to the intervention, eight weeks post-surgery, and at a 6-month follow-up. The results indicated a better recall for pain-related words prior to treatment when the intensity of pain was high, but an opposite effect with a better recall of non-pain-related words at the 6-month follow-up when the intensity of pain was significantly reduced. The authors suggested that selective memory for pain-related words is more likely a secondary consequence of the long-term experience of pain than a stable cognitive vulnerability factor. Some methodologically oriented, well-controlled studies have further shown that chronic pain patients in general show a high percentage of correct pain ratings (Salovey et al., 1993; Jensen et al., 1996). For example, Jensen and colleagues instructed patients with chronic low back pain during an inpatient medical treatment to fill in a diary with pain ratings for 30 days. Four weeks later they had to recall the mean, lowest and highest pain of that time. The authors found correlations between r = 0.75 and r = 0.83, which they interpreted as very high. We can summarize and suppose that a biased memory for pain-related stimuli in chronic pain patients is more a consequence of long-lasting pain and pain-related anxiety or helplessness than a predisposing vulnerability factor for the process of chronicity. Attentional bias regarding actual pain stimuli

The hypervigilance model of pain perception suggests further that there is an attentional bias towards actual pain-related sensations. A laboratory experiment conducted by Pearce and Morley (1989) provided evidence for such an attentional bias. The authors used the modified Stroop color naming task in which categories of sensory pain, affective pain, and emotionally negative and neutral words are presented in different colors, and response times to name the color of each word are measured. This task was chosen because its successful performance requires the employment of central attentional processes (Logan,

1985). The authors predicted that chronic pain patients will show longer response latencies for naming pain-related stimuli than neutral stimuli compared to healthy controls. The results of this study clearly indicated that pain patients showed greater interference on pain-related sensory and affective words. These data were in line with attentional biases that were found in patients with anxiety disorders (e.g. Williams et al., 1996) or with depression (e.g. Edwards et al., 1992). Based on several unpublished studies, Pincus et al. (1998) hypothesized that emotional disturbances in pain patients like increased anxiety or depression could be the main cause of greater interference, irrespective of pain status. They pointed out that Pearce and Morley (1989) did not control for pain-related anxiety or depression. In a series of experiments Pinkus and colleagues used the classical Stroop with color and name incongruent and added a control condition with color and name congruent. Further they assessed anxiety by the Spielberger State and Trait Anxiety Inventory (Spielberger et al., 1970) and depression by the Beck Depression Inventory (Beck et al., 1961). As in several former unpublished studies they found no difference between chronic pain patients in general and healthy controls with regard to response time to pain-related and neutral stimuli. The only significant difference was between the two word categories with higher response times in the classical Stroop condition. Supporting their hypothesis, response time to pain stimuli was highly correlated with trait anxiety (r = 0.63) and with depression (r = 0.61), whereas pain scores did not correlate significantly with color naming times. The authors concluded that the presence of an attentional bias in chronic pain patients can best be accounted for as arising from mood state rather than pain. These data were supported by a study conducted by Eccleston et al. (1997), who have shown that those chronic pain patients reporting high-intensity pain and high scores on somatic awareness, assessed by the Modified Somatic Perception Questionnaire (MSPQ, Main, 1983) have shown more behavioral disruption in a modified Stroop Test. Patients who show more somatic awareness also displayed higher scores in pain-related anxiety. There was no relationship between the intensity of pain and the results of Stroop testing. Thus, attentional disruption in chronic pain cannot

530 be accounted for by pain intensity alone. The authors concluded that in patients with high pain intensity and also high somatic awareness cognitive strategies based on distraction may be difficult to apply. Estimation offuture pain With the model of amplification of pain sensations it was also hypothesized that chronic pain patients would show a pattern of overestimation of future pain. For example, Rachman and Arntz (1991) have suggested that chronic pain patients may expect an activity to be more painful than is actually the case. McCracken et al. (1993) studied this phenomenon in 43 CLBP patients during the Straight Leg Raising test (SLR), which was a routine part of the physical examination. Patients were exposed to six SLR trials which generally produce pain in the low back, hip, or leg. Prior to each trial the patients were asked to predict the maximum intensity of pain that they expected during the SLR procedure. Following each trial they were asked to report the maximum intensity of pain they had actually experienced. Surprisingly the authors found that 38% of the predictions were underpredictions, 42% were correct, and only 20% were overpredictions. Testing the effect of false predictions on subsequent predictions they found that underpredictions were followed by increases in prediction on the next trial in 64% of the cases, overpredictions were followed by decreases 69% of the time, and correct predictions were followed by no change (59%). Furthermore, the results have shown that patients with low pain-related anxiety, assessed by the Pain Anxiety Symptoms Scale (PASS; McCracken et al., 1992), revealed mainly underpredictions whereas the high-anxiety group showed more overpredictions. During the six trials the low-anxiety patients became more accurate relatively late in the sequence of exposures, whereas high-anxiety patients became accurate earlier. Using a finger pressure stimulator for pain stimulation in a series of six stimulations, Arntz and Peters (1994) compared the prediction of future pain in 20 male CLBP patients and 20 age-matched healthy controls. The authors found a significant group difference with a general underprediction of pain in the CLBP patients and a nonsignificant tendency to overprediction in the healthy controls. The trends throughout the tri-

als have shown that both groups started with an underprediction, but CLBP patients made a larger underprediction than the controls. After the first trial the control group tended to overpredict pain and became quite accurate at the end of the series. The CLBP patients also adjusted their prediction to more correct levels, but showed underprediction again at the end of the series. In a study with higher ecological validity Murphy et al. (1997) investigated the inter-relationship between predictions of pain, anxiety, and avoidance of physical activity in a group of CLBP patients using repeated pain-related physical activities (e.g. 5 min walking, 2 min standing up from sitting, 1 min climbing stairs) as daily living pain stimuli. In this study 60% of the subjects underpredicted the experienced pain for the first exercise. Furthermore, results have shown that the greater the discrepancy between predicted and experienced pain, the greater the increase in pain predicted for the next exercise, the greater the anxiety, and the less the subjects were willing to perform the next exercise. Thus, the patients who under-predicted pain showed an increased avoidance on the subsequent activity. In general there was a significant increase in predicted pain as well as in anxiety from the first to the third exercise. However, there was no correlation between the experienced level of pain and anxiety or avoidance of physical exercises. Summarizing the results of these experimental studies, the data only partly provide support for the model of Rachman and Amtz (1991). CLBP patients revealed a strong tendency to be inaccurate in their predictions of future pain, but the direction of this inaccuracy was opposite to that hypothesized. Nevertheless, the consequences of underprediction of pain fit the model of amplification or the Rachman and Amtz model. Underprediction of pain led to an increase in anxiety and in avoidance of physical activities. Individual differences in pain patients due to pain-related attentional strategies The differences between patients with high and low pain-related anxiety with regard to over- and underprediction of pain, McCracken et al. (1993) have found, suggest that there are marked individual differences in their daily living attentional strategies to cope with pain. Hasenbring and coworkers (Hasen-

531 LR patients

Intensity of pain

CT patients

7 6

1

*MT

ST patients patients

5432l0

(Grebner et al., 1999; Hasenbring et al., unpublished data). The prospective evaluation of the development of pain, inability, and work status revealed that according to the hypotheses all three high-risk groups showed chronic persistent or recurrent low back pain, whereas the low-risk group displayed a good recovery following the medical treatment (see Fig. 1). Furthermore, the high-risk groups showed a significantly lower ability to return to work compared to the low-risk group (Grebner et al., 1999; Hasenbring et al., unpublished data).

I

The potential role of attentional process of chronicity of pain Fig. 1. The development of pain following conservative medical therapy in three high-risk and one low-risk group of patients with acute sciatic pain at TO. LR = Low Risk, CT = Catastrophizing Thoughts, ST = Suppressive Thoughts, MT = Minimizing Thoughts.

bring, 1993; Hasenbring et al., unpublished data) conducted a series of studies where they classified patients with acute sciatic pain into three groups of patients with maladaptive attentional strategies. Using the Kiel Pain Questionnaire (KPI, Hasenbring, 1994) they built a group ‘Catastrophizing Thoughts’ (CT) that showed primarily cognitions of catastrophizing and help-/hopelessness, a group ‘Suppressive Thoughts’ (ST) that showed high scores on the scale ‘Suppression’ of the KPI, and a group ‘Minimizing Thoughts’ (MT). CT patients also showed increased pain-related anxiety, depression, and avoidance behavior. ST patients showed increased depression and also increased avoidance behavior according to physical and social activities. Group MT displayed increased positive mood and suppressive behavior with regard to physical and social activities. A fourth group showed low or mean scores on these scales; we called them patients with ‘Low Risk of chronicity’ (LR) due to the hypothesis that patients with marked catastrophizing, suppression or minimizing cognitions will have a high probability to develop chronic pain over time, whereas patients with low scores will have a low risk for chronicity. The former clinical classification was recently validated by empirical methods of cluster analysis

cognitions in the

There is strong evidence from experimental and clinical research that patients with chronic pain deploy mainly attentional cognitions of distraction, suppression, and minimizing for coping with their pain. Only patients with high pain-related anxiety display cognitions of distress-focusing attention and catastrophizing. Low-anxiety patients also tend to underpredict future pain and do not show any attentional bias to pain-related material in tests like the Stroop task. Although experimental as well as clinical research has shown that distraction, suppression, and minimizing pain experiences in patients with natural pain are counterproductive in coping with pain, patients prefer these strategies to sensory monitoring. Several hypotheses about this preference of maladaptive attentional coping and about the low efficacy of these strategies were discussed above. With the avoidance-endurance model of chronic&y of pain (Fig. 2), we formulated a conceptualization towards the role of attentional processes in the process of coping with acute pain and development of chronic pain states (Hasenbring, 1993; Hasenbring et al., 1994). The model suggests that there are at least three different maladaptive types of attentional coping and one adaptive one. Patients who tend to focus on the harmful and distressing aspects of an acute pain state, for example acute low back pain, show increased catastrophizing cognitions. These are directly related to fear of pain and fear-related avoidance behavior. Avoidance behavior is often accompanied by long-lasting one-sided postures of relieve (sitting or lying) that lead to a deconditioning syndrome with muscular insufficiency

532

Sensory

Catastrophizing 1 Fear/Anxiety

Suppressive

thoughts

1 Irritated/depressive

mood

1 Avoiding behavior

Avoiding behavior

1 Immobility/depression

Immobility/depression

1 Muscular insufficiency

1 Muscular Insufficiency

L

1

Chronicity

of pain

Fig. 2. The avoidance-endurance

Minimizing thoughts

monitoring and Distraction by positive phantasies

1 Positive mood

Gus “1” ‘8* J& -^q”‘i‘

1

Flexible change (active/passive behavior)

1

Chronicity of pain

model

of chronicity

i*; I x3 iq a+&

Chronicity of pain

of pain developed

and swollen disks (Nachemson, 1987). Under the condition of insufficient muscular strength, even a normal physical strain will cause back pain due to processes of neurophysiological sensitization (McQuade et al., 1988). Long-term consequences due to the avoidance of social activities are increased social withdrawal and depression. The tendency to focus on harmful aspects of pain may be influenced by a more general personality mode of anxiety sensitivity (Asmundsonn et al., 1999). Patients who in contrast try to suppress the experience of low back or sciatic pain may produce the above-described rebound phenomenon (Wegner et al., 1991) with a repeated switching between pain and concentration on current activities of daily living. These patients feel increasingly distressed because of recurrent failures in effectively coping with pain. Long-term consequences are feelings of help- and hopelessness and depression. Although there are no data until now, it can be assumed that these patients will underprediet future pain and develop distress and increased avoidance behavior due to their false predictions. The third group of high-risk patients tries to ignore their pain and concentrate fully on their daily activities. These patients show no depression but in contrast increased scores in positive mood. It is as-

Reduction of pain

by Hasenbring.

sumed that they will deploy physical overexertion with long-lasting one-sided postures of strain (standing, walking) that lead to a physical overload on the muscles, joints, and discs with increasing pain as one consequence. It can be speculated that also this group of patients will underpredict future pain but without the aversive consequences of distress and avoidance. Thus, these patients seem to be more successful in minimizing pain sensations, but this efficiency is counterproductive with regard to the chronicity of pain. The reason for the higher efficacy in minimizing pain than in suppression remains unclear. The low-risk group showed low scores on suppression, minimizing, and catastrophizing. In their cognitions they deployed more types of distraction that were accompanied by positive phantasies and relaxation. It is hypothesized that this group displays an adaptive and flexible change between exertion and relaxation due to their pain (Nachemson, 1987) and that they are able to calibrate their pain experience against the nature of the sensory-discriminative stimulus (see also Lethem et al., 1983). It can be speculated that these patients also deploy types of sensory monitoring spontaneously.

533

References Ahles, T., Blanchard, E. and Leventhal, H. (1983) Cognitive control of pain: attention to the sensory aspects of the cold pressor stimulus. Cogn. The,: Res., 7: 159-177. Arntz, A. and Peters, M. (1994) Chronic low back pain and inaccurate predictions of pain: is being too tough a risk factor for the development and maintenance of chronic pain?. Behav. Res. Ther, 33: 49-53. Asmundsonn, G.J., Norton, PJ. and Norton, G.R. (1999) Beyond pain: the role of fear and avoidance in chronicity. Clin. Psychol. Rev., 19: 97-l 19. Avia, M.D. and Kanfer, F.H. (1980) Coping with aversive stimulation: the effects of training in a self-management context. Cogn. Ther Res., 4: 73-8 I. Barsky, A.J. (1992) Amplification, somatization, and the somatoform disorders. The Academy of Psychosomatic Medicine: Festschrif fir Thomas B. Hackett. Barsky, A.J. (1998) A comprehensive approach to the chronically somatizing patient. J. Psychosom. Res., 45: 301-306. Beck, A.T., Ward, C.H., Mendelson, M., Mock, N. and Erbaugh, J. (1961) An inventory for measuring depression. Arch. Gen. Psychiatry, 4: 561-571. Beers, T.M. and Karoly, P. (1979) Cognitive strategies, expectancy, and coping style in the control of pain. J. Consult. Clin. Psychol., 47: 179-180. Cioffi, D. (1991) Beyond attentional strategies: a cognitiveperceptual model of somatic interpretation. Psychol. Bull., 109: 25-41. Cioffi, D. and Holloway, J. (1993) Delayed costs of suppressed pain. J. Pers. Clin. Psychol., 64: 274-282. Devine, D.P and Spanos, N.P. (1990) Effectiveness of maximally different cognitive strategies and expectancy in attenuation of reported pain. J. Pets. Sot. Psychol., 58: 472-478. Eccleston, C. (1995) The attentional control of pain: methodological and theoretical concerns. Pain, 63: 3-10. Eccleston, C. and Crombez, G. (1999) Pain demands attention: a cognitive-affective model of the interruptive function of pain, Psychol. Bull., 125: 356-366. Eccleston, C., Crombez, G., Aldrich, S. and Stannard, C. (1997) Attention and somatic awareness in chronic pain, Pain, 72: 209-2 15. Edwards, L., Pearce, S., Collett, B.J. and Pugh, R. (1992) Selective memory for sensory and affective information in chronic pain and depression. Br J. Clin. Psychol., 31: 239-248. Edwards, L.C., Pearce, S.A. and Beard, R.W. (1995) Remediation of pain-related memory bias as a result of recovery from chronic pain. .I. Psychosom. Res., 39: 175-I 81. Grebner, M., Breme, K., Rothoerl, R., Woertgen, C., Hartmann, A. and Thorn&, C. (1999) Coping und Genesungsverlauf nach lumbaler Bandscheibenoperation. Schmerz, 13: 19-30. Hasenbring, M. (1992) Chroni$zierung bandscheibenbedingter Schmerzen. Risikofaktoren und gesundheitsforderndes Verhalten. Schattauer Verlag, Stuttgart. Hasenbring, M. (1993) Durchhaltestrategien - ein in Schmerzforschung und therapie vemachllssigtes Phanomen?. Schmerz, 7(4): 304-313.

Hasenbring, M. (1994) Das Kieler Schmerzinventar KU. 3 Fragebogenskalen zur Er&assung kognitiver und emotionaler Reaktionen sowie Copingreaktionen in Schmerzsituationen. Huber Verlag, Bern. Hasenbring, M., Marienfeld, G., Kuhlendahl, D. and Soyka, D. (1994) Factors of chronicity in lumbar disc patients. Spine, 19: 2759-2765. Hodes, R.L., Howland, E.W., Lightfoot, N. and Cleeland, C.S. (I 990) The effects of distraction on responses to cold pressor pain. Pain, 41: 109-I 14. Jensen, M.P., Turner, J.A., Romano, J.M. and Karoly, P (1991) Coping with chronic pain: a critical review of the litterature. Pain, 47: 249-283. Jensen, M.P., Turner, L.R., Turner, J.A. and Romano, J.M. (1996) The use of multiple-item scales for pain intensity measurement in chronic pain patients. Pain, 67: 35-40. Johnson, J., Morrissey, J. and Leventhal, H. (1973) Psychological preparation for an endoscopic examination. Gastrointest. Endosc., 19: 180-182. Johnson, M.H. and Petrie, SM. (1997) The effects of distraction on exercise and cold pressure tolerance for chronic low back pain sufferers. Pain, 69: 43-48. Kahnemann, D. (1973) Attention and Eflort. Prentice-Hall, Englewood Cliffs, NJ. Keefe, F. and Williams, D.A. (1990) A comparison of coping strategies in chronic pain patients in different age groups. J. Gerontol., 45: 161-165. Kent, G. (1985) Memory and dental pain. Pain, 21: 187-194. Kent, G. (1989) Memory and dental experiences as related to naturally occurring changes in state anxiety. Cogn. Emotion, 3: 45-53. Kleinke, C.L. (1992) How chronic pain patients cope with pain: relation to treatment outcome in a multidisciplinary pain clinic. Cogn. Ther Res., 16: 208-209. Lethem, J., Slade, P.D., Troup, J.D.G. and Bentley, G. (1983) Outline of a fear-avoidance model of exaggerated pain perception - I. Behav. Res. Ther, 21: 401-408. Leventhal, H., Brown, D., Shacham, S. and Engquist, G. (1979) Effects of preparatory information about sensations, threat of pain, and attention on cold pressor distress, .I Pets. Sot. Psychol., 37: 688-714. Linton, S.J. (1991) Memory for chronic pain intensity: correlates of accuracy. Perceptual and Motor Skills, 72: 1091-1095. Linton, S.J. and Melin, C. (1982) The accuracy of remembering chronic pain. Pain, 13: 281-285. Logan, G.D. (1985) Executive control of thought and action. Acta Psychol., 60: 193-210. Love, R., Nerenz, D. and Leventhal, H. (1983) Anticipatory nausea with cancer chemotherapy: development through two mechanisms. Proc. Am. Sot. Clin. Oncol., 2: 62. Macfarlane, G.J., Thomas, E., Croft, P.R., Papageorgiou, A.C., Jayson, M.I.V. and Silman, A.J. (1999) Predictors of early improvement in low back pain amongst consulters to general practice: the influence of pre-morbid and episode-related factors. Pain, 80: 113-I 19. Main, C.J. (1983) The modified somatic perceptions questionnaire. J. Psychosom. Res., 27: 503-514.

534 McCaul, K.D. and Haugtvedt, C. (1982) Attention, distraction, and cold pressor pain. J Pers. Sot. PsychoZ., 43: 154-162. McCaul, K.D. and Mallot, J.M. (1984) Distraction and coping with pain. Psychol. Bull., 95: 516-533. McCracken, L.M., Zayfert, C. and Gross, R.T. (1992) The Pain Anxiety Symptoms Scale: development and validation of a scale to measure fear of pain. Pain, 50: 67-73. McCracken, L.M., Gross, R.T., Sorg, P.J. and Edmands, T.A. (1993) Prediction of pain in patients with chronic low back pain: effects of inaccurate prediction and pain-related anxiety. Behav. Res. TheK, 31: 647-652. McDermid, A.J., Rollman, G.B. and McCain, G.A. (1996) Generalized hypervigilance in fibromyalgia: evidence of perceptual amplification. Pain, 66: 133-144. McQuade, K.J., Turner, J.A. and Buchner, D.M. (1988) Physical fitness and chronic low back pain. Clin. Orthop., 233: 198204. Murphy, D., Lindsay, S. and De C. Williams, A.C. (1997) Chronic low back pain: predictions of pain and relationship to anxiety and avoidance. Behav. Rex Ther., 35: 231-238. Nachemson, A. (1987) Lumbar intradiscal pressure. In: M.I.V. Jayson (Ed.), The Lumbar Spine and Back Pain. Churchill Livingstone, Englewood Cliffs, NJ, pp. 191-203. Nerenz, D., Leventhal, H., Love, R. and Ringler, K. (1984) Psychological aspects of cancer chemotherapy. Int. Rev. Appl. Psychol., 22: 521-529. Pearce, .I. and Morley, S. (1989) An experimental investigation of the construct validity of the McGill Pain Questionnaire. Pain, 39: 115-121. Pincus, T., Pearce, S., McLelland, A. and Turner-Stokes, L. (1993) Self-referential selective memory in pain patients. BI: J. Clin. Psychol., 32: 365-374. Pincus, T., Fraser, L. and Pearce, S. (1998) Do chronic pain

patients ‘Stroop’ on pain stimuli?. BK J. Clin. Psychol., 37: 49-58. Rachman, S. and Amtz, A. (1991) The overprediction and the underprediction of pain. Clin. Psychol. Rev., 11: 339-355. Rosenbaum, M. (1980) Individual differences in self-control behaviors and tolerance of painful stimulation. J. Abnorm. Psychol., 89: 581-590. Rosenstiel, A.K. and Keefe, F.J. (1983) The use of coping strategies in chronic low back pain patients: relationship to patient characteristics and current adjustment. Pain, 17: 3344. Salovey, I?, Smith, A.F., Turk, D.C., Jobe, J.B. and Willis, G.B. (1993) The accuracy of memory for pain: not so bad most of the time. Am. Pain Sot. J., 2: 184-191. Spielberger, CD., Gorsuch, R. and Lushene, R. (1970) The State-Trait Anxiety Inventory (STAI). Manual. Consulting Psychologists Press, Palo Alto, CA. Suls, 3. and Fletcher, B. (1985) The relative efficacy of avoidant and non-avoidant coping strategies: a meta-analysis. Health Psychol., 4: 249-288. Turner, J.A. and Clancy, S. (1986) Strategies for coping with chronic low back pain: relationship to pain and disability. Pain, 24: 355-364. Wegner, D.M., Schneider, D., Carter, S.R. and White, T.L. (1987) Paradoxical effects of thought suppression. J. Pers. Sot. Psychol., 53: 5-13. Wegner, D.M., Schneider, D., Knutson, B. and McMahon, S. (1991) On pollution the stream of consciousness: the effect of thought suppression on the mind’s environment. Cogn. Ther: Res., 15: 121-152. Williams, J.M.G., Mathews, A. and MacLeod, C. (1996) The emotional Stroop task and psychopathology. Psychol. Bull., 120: 3-24.

535

Subject Index

wunino-3-hydroxy-5methyl-4-isoxazolepropionic acid (AMPA), 115 A-fiber, 167,238 high-threshold mechanoreceptors, 335 A-fiber nociceptor, 332 ~12receptor, 70 postsynaptic, 7 1 cwz-adrenoceptor spinal nerve ligation, 235 As/g-fiber, 167 Ap-fiber, 66,73,75,82,84,116,119,298,319,359 sprouting, 495 acetylcholine tissue injury, 28 acid-sensing ion channel, 21,34 action potential pathological bursting, 3 spontaneous, 3 acupuncture, 92 A6 thermoreceptors, 146 A&fiber, 30, 66, 84, 90, 119, 138, 139, 167, 289,318,378 adenosine, 27 ischemic pain, 3 1 adenosine triphosphate (ATP), 22,102 as a neurotransmitter, 103 tissue damage, 21 afferent fiber high-threshold, 378 low-threshold, 378 mechanosensitive pelvic nerve, 378 sensitization, 378 silent, 380 2-AG, 471 biosynthesis, 472 agmatine, 438 allodynia, 51, 73, 81, 108, 116, 141, 154, 173, 199, 205, 231, 284, 298, 308, 313, 326, 335, 357, 399, 415, 422, 430, 496,506 bilateral mechanical, 406 mechanical, 56, 399,433,434,453

224,

192, 315, 455,

mirror image mechanical, 400 newborn, 370 Nociceptin/OFQ-induced, 207 supraspinal control, 23 1 tactile, 117, 232 AMPA receptor, 65,101,157 calcium permeable, 65 amputation, 267 amygdala, 138 analgesia cannabinoid, 474 epidural or spinaI,498 opioids, 5 13 opioids and NSAIDs synergistic, 488 pre-emptive, 158,493,496, 505 stimulation-produced, 483 systemic, 499 anandamide, 471,472 formalin-evoked nociception, 475 anesthesia dolorosa, 254 angina pectoris, 29 arachidonic acid, 486 astrocyte hypertrophy cytokines, 392 glutamate/aspartate transporters, 392 growth factor, 392 attention, 526 autotomy, 412,497 social isolation, 4 12 guanethidine, 453 axonal transection TTX-resistant sodium current, 10 TTX-sensitive sodium current, 10 axotomy, 365 action potential generation, 11 peripheral, 2 19 bone destruction, 390 bradykinin (BK), 31,40 excitatory effect, 45 heat threshold, 46 novel theory, 45

536

sensitizing effect, 45 brain source analysis, 313 brainstem periaqueductal (PAG), 483 C-fiber, 26, 30, 40, 53, 63, 66, 68, 72, 74, 83, 90, 116, 119, 138, 139, 146, 157, 167, 213, 224, 238, 318, 331, 358, 378, 441, 495, 505,511 cardiac sympathetic, 33 polymodal, 47 synapses, 82 C-fiber nociceptor, 332 c-Fos protein, 195 expression in lamina I neurons, 392 calcitonin-gene related peptide (CGRP), 41 calcium channel, 174 high threshold, 176 calcium infhtx peripheral nociceptive terminals, 43 calcium ion, 54, 87, 125, 126 concentration in acutely dissociated SG neurons, 126 protein kinases, 116 protein phosphatases, 116 calcium ion concentration, 83,90 in spinal dorsal horn neurons, 87 calcium-calmodulin-dependent protein kinase II nociceptive transmission, 5 1 LPT, 116 s. calcium-dependent enzymes calpains, 442 cannabinoid, 471 CBl and CB2,474 endogenous, 47 1 receptor subtypes, 474 capsaicin, 332 neonatal, 74 receptor, 40 ruthenium red, 42 SP receptor internalization, 358 cardiac ischemia, 21 carrageenan, 14,15,176,357,370 SP receptor internalization, 358 central sensitization, 72, 81, 116, 173, 210, 361, 375,407 induced by C-fiber nociceptors, 332 membrane excitability, 8 1 morphological reorganization, 81

phenotypical changes, 8 1 synaptic mechanisms, 8 1 chronic constrictive nerve injury sodium channel, 12 cingulate cortex, 279,294 anterior, 147, 248, 249,282, 297 posterior, 297 cingulate sulcus, 141 anterior, 142 clonidine, 293 cold pain, 280 colorectai distension (CRD), 377 complete Freund’s adjuvant, 75,192,357,372 Complex Regional Pain Syndrome (CRPS), 308 cordotomy anterolateral, 412 bleeding into the surgical cavity, 418 hemotoxicity, 4 19 cortex associative, 3 13 frontal, 308 primary somatosensory, 3 14 somatosensory, 95,308,313 somatotopic organization, 3 13 activated by painful stimuli, 306 cortical reorganization, 317 counter-irritation, 92 cutaneous laser, 280 cyclic AMP-dependent protein kinase nociceptive transmission, 5 1 cyclic GMP-dependent protein kinase nociceptive transmission, 5 1 hi-opioid receptor agonists, 70 deafferentation, 326 2-deoxyglucose technique metabolic activity, 191 descending facilitation, 231 descending inhibition, 84,92,233, tonic, 85 descending modulation, 138 discharge pattern synchronized, 84 distraction analgesic effect, 525 dorsal horn

,507,517

537

Afi-fiber, 73 lamina I, 137, 138,246 lamina II, 73 lamina IV-V, 247 superficial, 434,441 dorsal horn neuron, 102 GABAergic, 69 glycinergic, 69 hyperexcitability, 185 dorsal root compression, 399 fiber, 402 inflammation, 399 injury, 404 dorsal root ganglion, 399 axonal transection, 10 CB 1 receptors, 475 electrical properties, 6 persistent TTX-resistant sodium currents, 8 sodium channel a-111, 9 sodium channels, 3,5 sprouting of sympathetic fibers in the, 458 TTX-resistant sodium currents, 11, 13, 15 voltage-gated sodium currents, 4 electrical stimulation, 280 electroencephalogram (EEG), 289 some technical prerequisites, 290 endocannabinoid, 471 2-arachidonylglycero1(2-AG), 47 1 anandamide, 47 1 formation and inactivation, 47 1 endogenous pain control, 239 evoked magnetic fields, 313 excitatory postsynaptic current (EPSC), 102 monosynaptic, 102 polysynaptic, 102 excitotoxicity, 431 fibromyaigia, 327,328 forebrain, 277 formalin, 357 SP receptor internalization, 358 formalin test PKA RIB mutant mice, 53 functional magnetic resonance imaging (fMRI), 142,143,277,303 GABA* receptor, 69

presynaptic, 69 GABAs receptor, 69 presynaptic, 69 galanin, 219 dorsal horn, 222 dorsal root ganglion, 220 intrathecal, 220 knock-out mice, 220, 222 receptor subtypes, 222 up-regulation. 224 gate control theory, 92 GDNF, 14 gene expression, 191, 196 glutamate, 102 tissue injury, 28 glutamate receptor co-activation, 123 LTP of synaptic strength in C-fibers, 86 metabotropic, 86, 87, 117 spinal, 86 substantia gelatinosa, 119 grooming behavior, 434 habituation, 325 headache, 328 heart attack, 29 heat pain, 280 helplessness, 528 5-HT spinal nociceptive transmission, 105 recruitment of silent glutamatergic synapses, 105 5HTi receptor, 70 5-HTz receptor, 70,72 hyperalgesia, 51, 73, 81, 84, 108, 116, 154, 173, 192, 199, 231, 308, 313, 315, 326, 328, 336, 357, 365, 375, 399, 415, 422, 430, 486,496,506 afferent-induced, 82, 92 bilateral thermal, 406 cannabinoid antagonists, 474 during peripheral inflammation, 461 generated by nerve growth factor (NGF), 462 LTP-like, 96 maintenance of, 384 mechanical, 127, 308, 333,453,455 newborn, 370

538

primary, 185,233,326,331,512 rectal, 348 related to referred muscle pain, 346 secondary, 141, 173, 233, 235, 326, 331, 333, 345,365,489,506,512,513 sympatho-adrenal (SA) system, 463 thermal, 127, 164, 205,400,434,453,455 visceral, 347, 350, 376 wound, 518 hyperexcitability primary sensory neuron, 15 sodium channels, 15 spinal cord neuron, 178 hyperinnervation NGF, 368 wounds, 367 hypersensitivity, 367 hypothalamus, 138 infant, 365 inflammation carrageenan, 125 complete Freund’s adjuvant, 74 heat sensitivity, 39 knee, 181 neurogenic, 53, 56 peripheral, 125 peripheral tissue, 125 injury axonal, 8 central nervous system, 269 early, 366 hyperinnervation, 367 nerve, 85, 116 peripheral nerve, 126,267 tissue, 116 innocuous heat stimuli, 280 cold stimuli, 280 insula, 141 insular region, 307 interleukin-10 (IL-lo), 438 interleukind (IL-6), 196,197 mFWA increase, 199 neurons of the dorsal horn, 197 ischemia myocardial, 28

K-opioid agonist, 71 kainate receptor, 101,103 in lamina II neurons, 65 kaolin, 176 knee joint, 176 lactic acid, 32 long-term depression (LTD), 83,325 synaptic strength, 92, 115 long-term potentiation (LTP), 53,72,83,163,325, 506 depotentiation, 92 excitability and firing of neurons, 154 maintenance, 89, 158 natural noxious stimulus, 156 nociception in humans, 95 pain perception, 96 responses in deep dorsal horn, 93 responses in thalamus and cortex, 95 responses of motoneurons, 94 reversibility, 92 role of tachykinin receptors, 89 slow ventral root potential (sVFW), 94 spinalized rats, 156 synaptic strength, 82, 83, 115, 153 p.-opioid receptor, 70,71 magnetic resonance image (MRI), 281 magnetoencephalogram (MEG), 289 some technical prerequisites, 290 metabolic activity bilateral rise, 195 in the spinal cord, 193 monoarthritic rat, 193 metabotropic glutamate receptor (mGluR) group I, 87, 117 group II, 87 group III, 87 monoarthritis, 192 morphine spinal nerve ligation, 234 motor cortex, 306 mRNA differential display technique, 192 muscle pain algogenic substances, 344 electrical stimulation, 344 hyperalgesia, 346 referred. 345

539 mustard oil, 178 myositis, 164 acute, 164 subacute. 164 N-methyl-D-aspartate (NMDA), 82,94,115 receptor, 495 receptor antagonist, 157 N-methyl-D-aspartate (NMDA) receptor, 65, 72, 83,101,104,234,361,376,407 co-activation of Group I mGluRs and NMDARs, 123 hyperexcitability of central neurons, 127 spinal, 234 supraspinal, 234 nerve transection, 75 nerve growth factor (NGF), 13 hyperalgesia, 462 hyperinnervation, 368 induced sensitization of nociceptors, 462 overexpression, 368 neurokinin A, 6890 neurokinin-1 (NK-1) receptor, 68,89,407 neuron bursting firing pattern in, 252 nociceptive-specific, 139 spinothalamic, 406 thermoreceptive-specific, 139 wide dynamic range (WDR), 405 neuronal reorganization, 298 neuropathic pain sympathetic nervous system contributes, 453 Neuropeptide FF spinal nerve ligation, 236 neuroprotection calpain, 442 neurotransmitter release modulation by Nociceptin/OFQ and NST, 2 12 PKA-mediated regulation, 53 neurotrophin, 15 sodium channel expression, 12 nitric oxide (NO), 163,376,443 nitric oxide synthase (NOS), 163 nociceptin, 205 nociceptin-orphanin FQ (Noc/OFQ), 207 antisense oligonucleotide, 210 anxiolytic-like effects, 208

formalin-induced pain, 2 10 gene, 206 hyperalgesic activity, 207 induced allodynia, 2 13 induced hyperalgesia, 213 induced pain, 2 12 knockout mice, 208 localization of, 214 neurotransmitter release, 212 nociception, 209 substance P release, 214 tactile allodynia, 211 thermal hyperalgesia, 211 nociception descending control, 483 downstream to the first synaptic relay, 92 long-term changes, 92 nociceptive-specific neuron, 248 nociceptor, 138,389 C-mechano-heat-sensitive, 39 polymodal, 43 sensitization, 47 transduction mechanisms, 42 nocistatin (NST), 205 formalin-induced pain, 210 non-steroidal anti-inflammatory drugs (NSAIDs), 486 noxious heat stimuli, 280 nuclear factor-kappa B, 432 opioid analgesia, 483 dependence, 488 disinhibition, 484 inhibition of adenylyl cyclase, 485 inhibition of Ca2+ channels, 485 inhibition of presynaptic neurotransmitter lease, 485 membrane hyperpolarization, 485 molecular mechanisms, 483 NSAID synergy, 487 peptides, 483 receptor homologue (ORL,), 205 receptors, 483 supersensitivity, 489 tolerance, 488 orphanin FQ, 205

re-

540

P2X receptor, 21,22,66,103 presynaptic, 103 sensory-specific, 22 P2X3 receptor, 26 Selectivity in nociceptors, 26 pain acute inflammatory, 357 affective and unpleasant components, 303 affective motivational aspects, 137, 248, 252 after surgery, 505.5 18 anxiety, 529 attentional control, 527,529, 53 1 autotomy, 4 11 avoidance-endurance model, 53 1 bone cancer, 390 cancer-related, 389 cardiac, 30 cardiac sympathetic afferents, 29 catastrophizing, 526, 528, 531 central, 144,245, 248, 249, 298 central deafferentation, 441 central dysesthetic, 429 central modulation, 474 chronic back, 3 13 chronicity, 53 1 cognitive coping strategies, 525 cognitive-perceptual approaches, 525 cognitive-evaluative, 327 cold pressor, 526 cold stimuli, 280 conditioning model, 327 cortical reorganization, 3 13 deafferentation zone, 411 depression, 529 discriminative sensory, 137 disinhibition, 146 emotional influences, 315 estimation of future pain, 530 following spinal cord injury, 411 formalin-induced, 210 gender influences, 281 generator, 265 hypervigilance model, 528 imprint, 497 indeterminate, 327 inflammatory, 12,210, 357,460 intensity, 282 ischemic, 28

learning, 3 15 long-term inflammatory, 359 lumbar disc prolapse, 527 maladaptive attentional coping, 53 1 memories, 3 15, 3 16 minimizing thoughts, 53 1 mirror image, 4 18 moment-to-moment variations, 424 motivational-affective, 298, 327 muscles, 343 myoelectric or sensorimotor prosthesis, 320 nerve injury, 493 nervous system origin, 424 neuropathic, 8, 12, 56, 126, 219, 226, 239, 265, 284,326,399,407,442,452,496 non-stroke, 249 normal, 277 pathological, 137, 277, 308 patients and their spouses, 315 pattern generators, 441 peripheral modulation, 475 phantom, 254,493 phantom limb, 3 13, 3 17 phasic, 308 post-stroke, 249 postamputation, 493 postoperative, 116 preamputation, 3 17 preoperatively, 5 16 psychogenic, 327 radicular, 399 recall, 528 referred, 343 reinforcement, 3 15 related words, 529 sciatic, 527 sensory-discriminative aspects, 294, 298, 303, 315, 327 short-term inflammatory, 358 somatosensory, 3 15 somatosensory memories, 3 18 spinal cord injury, 429,440 spontaneous, 164, 399 suppressive thoughts, 53 1 thalamic, 249 thermosensory inhibition, 146 tissue injury, 53 tonic, 308

541

treatment, 320 unpleasantness, 282 verbal reinforcement, 3 16 visceral, 248, 343,346 willful monitoring, 525 pain generator spinal, 437 palmitylethanolamide as a peripheral analgesic, 476 CB2 receptor, 477 PDZ domain, 101,106 phantom limb sensation, 494 phantom pain, 493,494 blind and placebo-controlled trial, 499 character and location, 494 cortical reorganisation, 496 incidence, 494 peripheral generator, 494 preamputation pain, 497 time course, 494 polymodal afferents high-threshold C-mechanoreceptors, 46 positron emission tomography (PET), 142, 144, 277,282 technical and analytical issues, 278 postsynaptic calcium ions, 115 potassium tissue damage, 21 pre-emptive analgesia, 493 clinical studies, 500, 506, 511 differences between experimental and surgical pain, 507 experimental, 500 ketamine, 507 local anaesthesia, 507 measuring pain, 5 11 neuropathic pain, 496 opioids, 507 postamputation pain, 498 primary hyperalgesia, 5 13 relevant human study design, 5 12 secondary hyperalgesia, 5 13 pressure innocuous, 176 noxious, 176 primary afferent depolarization (PAD), 69 primary afferent fiber, 102 glutamate, 117

mechano-heat-sensitive, 39 primary somatosensory cortex (SI), 248 projected field (PF) maps, 259 propriospinal connections, 434 prostaglandin Ez, 44,47,54,74 protein kinase A (PKA), 485 isoforms of, 53 protein kinase C (PKC), 40,54,72,191,361,443, 485 isoenzymes, 55 LTP, 116 nociceptive transmission, 5 1 protein kinase C (PKCy) inflammation deficits, 56 lamina II of the spinal dorsal horn, 56 mutant mice, 55 protein kinase C (PKCE) mutant mice, 56 protein phosphatases 2 and 2A LTD, 116 proton-sensing molecule, 33 quisqualic acid intraspinal microinjection, 432 receptive field (RF) maps of, 259 mechanosensitive, 404 receptor antagonist capsazepine, 42 referred pain, 343 central mechanisms, 345 modulation, 35 1 neurophysiological mechanisms, 352 rostro-ventral medulla, 483 neutral cells, 237 off-cells, 237 on-cells, 237 sciatic nerve autotomy, 226 chronic constriction injury, 226,454 ligation and transection, 226,453 partial lesion, 453 secondary hyperalgesia, 141, 173, 233, 235, 326, 331, 332, 333, 345, 365, 489, 506, 512, 513

542

secondary somatosensory cortex (MI), 248 sensitization, 315,325,489 C-MH fibers, 40 central, 51, 81, 84, 96, 124, 153, 154, 173, 331, 357,495,506 central neuron, 326 heat, 44 lamina I neurons, 141 nociceptor, 47, 125 peripheral, 51, 357,495 wide dynamic range neurons, 214 sensory neuron axonal injury, 8 silent synapse, 72,86,101, 125 5-HT, 104 functional implications, 107 intracellular mechanisms of recruiting, 106 superficial dorsal horn, 104 skin wound, 368 sleep, 326 sodium channel genes, 5, 8, 9 inflammatory pain, 12 knockout mutants, 7 neurotrophins, 12 pharmacological manipulation, 16 sensory neuron-specific, 3 voltage-gated, 3 sodium current ‘ITX-resistant, 6 TTX-sensitive, 5 somatosensory cortex, 306 primary, 291 secondary, 29 1 somatotopic organization, 291 spinal cord, 63,326,365,399 astrocyte hypertrophy, 392 dorsal horn neuron, 402 expression of the pro-hyperalgesic peptide, 392 lamina II, 63 NSAIDs, 488 opioids, 488 slices, 64 spinal cord injury (SCI), 75,429 excitotoxic, 43 1,433, 437 loss of inhibitory tone, 440 NMDA receptor activation, 441 pain, 440

spatial profile, 437 synaptic plasticity, 442 spinal cord slice, 119 spinal cord transection reorganization of Vc, 262 spinal dorsal horn, 82,176,245 descending control of plasticity, 90 spinal nerve lesion, 455 ligation-induced neuropathy, 127, 233 spinal neuron afterdischarge, 437 bursting discharge, 437 increased excitability, 437 wind-up phenomenon, 153 spinal transection, 85,91,233,269 stimulation A& or C-fiber, 170 C-fiber, 334 electrical, 249, 313,413 heat, 41, 42 high-threshold mechanical, 402 long-lasting nociceptive, 327 natural noxious, 84,91 sciatic nerve, 167 tetanic, 155 thalamic, 249,265 touch, 308 Vc, 261 stimuli chemical, 138 cold, 139, 147 electric, 3 15 formalin, 84 innocuous cool, 248 low-threshold mechanical, 248,406 mechanical, 305 mechanical (distension), 348 mechanical trauma, 84 metabolic, 138 noxious electrical, 147 noxious heat, 248 noxious mechanical, 139, 248 noxious thermal, 404,406 skin heating, 84 sympathetic, 457, 458 tactile cutaneous, 248 thermal, 434, 347

543 tonic pressure, 303 warm, 139 vibration, 526 stump pain, 494 substance P, 68,82,407 Afi-fibers, 68 receptor internalization in lamina I, 357, 392 receptor up-regulation of, 360 substance P receptor (SPR) expression and internalization, 357 sympathectomy, 454 sympathetic nervous system, 452 sympathetic-afferent coupling, 457-459 sympathetically maintained pain (SMP) behavioral animal models, 452 reduced animal models in vitro, 459 reduced animal models in vivo, 456 synaptic efficacy, 63 synaptic facilitation, 107 synaptic plasticity, 72,442 brain derived neurotrophic factor (BDNF), 74 in PAG, 489 inflammation-induced synaptic plasticity, 73 long-term depression (LTD), 72, 8 1 long-term potentiation (LTP), 72, 8 1 nerve injury-induced, 7.5 prostanoids, 73 spinal, 115 unmasking of silent synapses, 72 windup, 72 synaptic transmission acetylcholine, 70 adenosine, 7 1 Adenosine Triphosphate (ATP), 66 fast excitatory transmission, 65 GABA, 69 glutamate, 65, 68 inhibitory transmission, 69 norepinephrine, 70 norepinephrine (NE), 71 opioids, 70, 71 postsynaptic inhibition, 7 1 presynaptic inhibition, 69 serotonin (5-HT), 70, 71 slow excitatory transmission, 68 sodium channel expression, 16

techniques, 63 tachykinin receptor, 86 tachykinins, 89 temperature coefficient, 41 tetrodotoxin (TTX), 4 thalamus, 95, 138,245,326 dystonia, 267 human, 259 intralaminar nuclei, 248 lateral, 249 medial, 248 plasticity, 259 posterolateral, 138 reorganization of, 262 ventral caudal (Vc), 259 ventral posterior lateral, 284 ventral posterior nucleus (VP), 246 tissue damage, 21 transcription factors inducible, 19 1 transcutaneous electrical nerve stimulation (TENS), 92 vanilloid receptor capsaicin, 41 heat, 41 in cardiac afferent neurons, 34 proton, 41 vigilance, 293 visceral pain, 346 chemical stimuli, 347 electrical stimuli, 348 hyperalgesia, 350 referred, 348 thermal stimuli, 347 voltage-dependent calcium channel, 485 L-type, 174 N- and P/Q-type, 174 Wallenberg syndrome, 284 Wallenberg’s disease, 298 Wallenberg’s infarcts, 144 wide dynamic range (WDR) neuron, 75, 93,153, 176,236,2&l