International Review of Cytology: A Survey of Cell Biology, Volume 106

INTERN ATlON AL

REVIEW OF CYTOLOGY A SURVEY OF CELLBIOLOGY

ADVISORY EDITORS H. W. BEAMS HOWARD A. BERN GARY G. BORISY PIET BORST BHARAT B. CHATTOO STANLEY COHEN RENE COUTEAUX MARIE A. DIBERARDINO CHARLES J. FLICKINGER OLUF GAMBORG M. NELLY GOLARZ DE BOURNE YUKIO HIRAMOTO YUKINORI HIROTA K. KUROSUMI GIUSEPPE MILLONIG ARNOLD MITTELMAN AUDREY MUGGLETON-HARRIS DONALD G. MURPHY

ROBERT G. E. MURRAY RICHARD NOVICK ANDREAS OKSCHE MURIEL J. ORD VLADIMIR R. PANTIC W. J. PEACOCK DARRYL C. REANNEY LIONEL I. REBHUN JEAN-PAUL REVEL L. EVANS ROTH JOAN SMITH-SONNEBORN WILFRED STEIN HEWSON SWIFT K. TANAKA DENNIS L. TAYLOR TADASHI UTAKOJI ROY WIDDUS ALEXANDER YUDIN

INTERNATIONAL

Review of Cytology A SURVEY OF CELLBIOLOGY

Editor-in-Chief

G. H. BOURNE St. George's University School of Medicine St. George's, Crenadri West Indies

Editors

K. W. JEON

M. FRIEDLANDER

Department of Zoology University of Tennessee Knoxville, Tennessee

Jules Stein Eye Institute U C L A School of Medicine Los Angeles, California

VOLUME106

I987

ACADEMIC PRESS, INC. Harcourt Brace Jovanovich, Publishers

Orlando San Diego New York Austin Boston London Sydney Tokyo Toronto

COPYRIGHT 0 1987 BY ACADEMIC PRESS. INC. ALL RIGHTS RESERVED. NO PART OF THIS PUBLICATION MAY BE REPRODUCED OR TRANSMITTED IN ANY FORM OR BY ANY MEANS, ELEflRONIC OR MECHANICAL, INCLUDING PHOTOCOPY. RECORDING, OR ANY INFORMATION STORAGE AND RETRIEVAL SYSTEM. WITHOUT PERMISSION IN WRITING FROM THE PUBLISHER.

ACADEMIC PRESS, INC. Orlando, Florida 32887

United Kingdom Edition published by

ACADEMIC PRESS INC. (LONDON) LTD. 24-28 Oval Road, London NW I 7DX

LIBRARY OF CONGRESS CATALOG CARD NUMBER:52-5203 ISBN 0-12-364506-9

(alk. paper)

PRINTED IN ME UNITED STATES OF AMERICA 87 88 89 90

9 8 1 6 5 4 3 2 1

Contents

Biochemical Transmitters Regulating the Arrest and Resumption of Meiosis in Oocytes EIMEISATOA N D S. S. KOIDE

I . Introduction ........................................................... I1 . Factors Sustaining Meiotic Arrest ....................................... I11 . Factors Inducing Resumption of Meiosis ................................. IV . Mechanism of Meiotic Resumption ...................................... References ............................................................

i

2 13 19 28

Morphology and Cytochemistry of the Endocrine Epithelial System in the Lung D . w . SCHEUERMANN 1. Introduction ........................................................... I1 . Light Microscopic Aspects ............................................. 111. Argentaffinity and Argyrophilia ......................................... IV . Cholinesterase Activity ................................................. V . Neuron-Specific Enolase ............................................... VI . Aspects of Induced Fluorescence ........................................ VI1 . Immunocytochemistry for Regulatory Peptides ............................ VIII . Electron Microscopic Aspects ........................................... I X . Location .............................................................. X . Innervation ........................................................... XI . Concluding Remarks . . . . . . . . ....................................... References ............................................................

35 39 43 45 46 48 53 55 71 73 79 80

Intrinsic Nerve Plexus of Mammalian Heart: Morphological Basis of Cardiac Rhythmical Activity? JOSEF MORAVECA N D

MlRElLLs MORAVEC

I . Introduction ........................................................... I1 . Autonomic Innervation of the Heart ..................................... V

89 91

vi

CONTENTS

111. Intracardiac Ganglionic Cells ............................................ IV . Terminal Nerve Plexus ................................................. V . New Developments in Studies of the Autonomic Nervous System .......... V1 . Morphological Basis of the Rhythmical Activity of the Heart: A Working Hypothesis ............................................................ VII . Conclusion ............................................................ References ............................................................

96 119 132 135 139 139

Structural and Functional Evolution of Gonadotropin-Releasing Hormone ROBERT P. MILLARAND JUDYA . KING I. I1 . 111. IV .

Introduction ........................................................... Structure and Distribution of GnRH and Related Molecular Forms .......... Biological Activity of GnRH ............................................ Conclusions ........................................................... References ............................................................

149 150 163 174 171

Excitons and Solitons in Molecular Systems

.

A . S DAVYDOV 1. Introduction ........................................................... I1 . The New Concept of Energy Transport along Protein Molecules ............ 111. History of Observation of Solitary Waves ................................ IV . Nonlinear Phenomena in Biology ........................................ V . Solitons in Real a-Helical Protein Molecules .............................. VI . Solitons in Discrete Models: Numerical Calculations ...................... VII . Solitons and the Molecular Mechanism of Muscle Contraction .............. VIII . Intracellular Dynamics and Solitons ..................................... IX . The Laser Raman Scattering by Metabolically Active Cells ................. X . Possible Mechanism for Anesthesia ...................................... XI . Electron Transfer along Protein Molecules ............................... XI1 . Electrosolitons Pairing in Soft Molecular Chains .......................... References ............................................................

138 187 189 192 199 201 204 207 213 214 216 221 223

The Centrosome and Its Role in the Organization of Microtubules I . A . VOROBJEV AND E . S . NADEZHDINA I . Introduction ........................................................... 11. Ultrastructure of Centrioles and Basal Bodies .............................

I11 . The Ontogenesis of Basal Bodies and Centrioles .......................... IV The Organization of the Centrosome and Its Behavior in a Cell Cycle ....... V . The Biochemistry of Centrioles and Basal Bodies .........................

.

227 229 239 244 249

CONTENTS

vii

V1 . Assembly of Microtubules on Microtubule-OrganizingCenters (MTOCs) in Virro

.................................................................

VII . Assembly of Microtubules on Microtubule-OrganizingCenters in Vivo ....... VIII . The Centrosome and the Cell ........................................... IX . Localization and Orientation of Centrioles in Cells ........................ X . Conclusion ............................................................ References ............................................................

INDEX

......................................................................

257 265 272 276 280 284 295

This Page Intentionally Left Blank

INTERNATIONAL REVIEW OF CYTOLOGY. VOI.. IIK

Biochemical Transmitters Regulating the Arrest and Resumption of Meiosis in Oocytes' EIMEISATO* A N D

s. s. KOIDEt

'Depurtment of' Animul Sciences, Fuculty c d Agricultrrre, Kyoto University, Kyoto 606, Jupun, and 'Center for Biomedicrrl Reseurch, The Poprrlution Coimcil. N e w York, New York 10021

I. Introduction

Germ cells migrate to the genital ridge from the yolk sac region during early embryonic development. In the genital ridge, the female germ cells start to divide and differentiate into oogonia. They enter meiosis and become primary oocytes. Nuclear division progresses to the diplotene stage of the first meiotic prophase and is arrested. The chromosomes decondense and are distributed diffusely throughout the oocyte nucleus. Progression of meiosis to the diplotene stage occurs before or shortly after birth. The oocytes may remain arrested at the dictyate stage for a prolonged period. Subsequently, a follicle develops enclosing the oocyte which contains a large clear nucleus designated as the germinal vesicle. It is generally accepted that the surge of luteinizing hormone (LH) during each ovarian cycle triggers the resumption of meiosis of the mature oocyte enclosed within Graafian follicles (Channing ef al., 1980, 1982a,b; Tsafriri, 1978b,c). The resumption of meiosis follows a sequence of programmed events while the oocytes are situated within the preovulatory follicles. The process is designated as oocyte maturation and is characterized by a series of biochemical, morphological, and functional changes that take place within the nucleus, highlighted by the following events: ( I ) dissolution of the nuclear membranes manifested as germinal vesicle breakdown (GVBD), (2) chromatin condensation and the formation of distinct chromosomes, (3) formation of the first meiotic spindle, (4)translocation of the spindle to the peripheral region, (5) formation and extrusion of t h e first polar body, (6) formation and positioning of the second meiotic division, (7) rearrest at the second metaphase. To better understand the biochemical mechanisms involved in oocyte maturation, in v i t r o culture systems have been developed. Pincus and Enzmann (1935) were the first to demonstrate that rabbit oocytes removed 'The authors dedicate this paper to Dr. Haruo Kanatani, pathfinder of starfish oocyte maturation and discoverer of I-methyladenine.

I Copyright B ' IYX7 hy Ac;idrmic h e \ \ . Inc. All right.. ol' repriiductiun in tiny limn rewrved.

2

EIMEI S A T 0 A N D S. S. KOIDE

from follicles resume meiosis spontaneously and mature under in vitro culture without the addition of hormones. This phenomenon of spontaneous maturation has been observed in all mammalian species examined (Biggers, 1973). Oocytes with adhering cumulus cell complexes or denuded from cumulus cells are widely used as models to study oocyte maturation. Studies with preovulatory follicles cultured in vitro demonstrate the interdependency of the various cells and fluid of the follicles and yield pertinent information on the resumption of meiosis triggered by LH added to the culture medium. The contrasting results obtained using denuded oocytes and follicle-enclosed oocytes suggest that maturation of mammalian oocytes is prevented by the follicular cells or factors in follicular fluid. By removing these factor(s) resumption of meiosis proceeds spontaneously. The involvement of the follicular cell-oocyte complex in the regulation of meiotic arrest was further investigated by coculturing isolated oocytes with follicular cells. Several review articles are available on various aspects of mammalian oocyte maturation. General and historical accounts have been covered by Donahue (1972), Tsafriri (1978b,c, 1984), and Masui and Clarke (1979). Technical problems relating to the in vitro culture of oocytes were discussed by Biggers (1973) and McGaughey (1978). Biochemical events involved in mammalian oocyte maturation have been presented by Mangia and Canipari (1977) and Wassarman et al. (1978). Morphological and ultrastructural changes were described by Albertini (1984). Informative review articles on the hormonal control and factors regulating oocyte maturation have been published (Lindner et ul., 1974, 1977, 1983; Schuetz, 1974; Channing and Tsafriri, 1977; Baker, 1979; Thibault, 1977; Channing et al., 1978, 1980, 1981, 1982a,b; Tsafriri and Bar-Ami, 1982; Tsafriri et al., 1982a; McGaughey, 1983; Eppig, 1980a). Cell-to-cell communication of cumulus-oocyte complexes has been discussed by Schuetz (19781, Moor (1983), and Dekel (1984). In the present article factors sustaining meiotic arrest and regulating resumption of meiosis in mammalian oocytes are discussed. A hypothesis of the sequence of events during oocyte maturation is proposed, based on our recent results. 11. Factors Sustaining Meiotic Arrest

Oogonia undergo the initial stages of the first meiotic division to reach the dictyate stage of prophase. The oocyte may remain in meiotic arrest for a prolonged period until activated shortly before ovulation or may undergo atretic degeneration. This suspended metabolic state of oocytes is an unusual phenomenon and has attracted the attention of many in-

ARREST A N D RESUMPTION OF MEIOSIS IN OOCYTES

3

vestigators. To identify the factors that sustain meiotic arrest, studies have been conducted with fully grown oocytes obtained from untreated follicles. Although the in vitro results may not fully reflect the physiological state, the findings are relevant and significant. Isolated oocytes will resume meiosis spontaneously when placed in hormone-free media (Pincus and Enzmann, 1937; Edwards, 1965), while follicle-enclosed oocytes remain in the dictyate stage (Tsafriri, 1978b; Lindner rt al., 1983). These findings suggest that the follicular microenvironment plays a dominant role in the physiological stability of oocytes at the dictyate stage. To clarify the role of various structural elements sustaining meiotic arrest, it has been found that oocytes in contact with granulosa cells remain arrested in the dictyate stage (Foote and Thibault, 1969; Sat0 rt al., 1982). Also follicular fluid and extracts of granulosa cells suppressed the occurrence of spontaneous maturation (Tsafriri and Channing, 1975a,b; Tsafriri et al., 1976, 1977; Tsafriri, 1978a; Hillensjo et al., 1978; Stone et al., 1978; Channing et ul., 1983; Eppig and Downs, 1984; Downs r t ul., 1985). These preliminary studies indicate that granulosa cells produce factors that sustain meiotic arrest. The meiotic-arresting factors to be discussed are cyclic nucleotides, cyclic nucleotide-potentiating factor, and maturation inhibitory peptides or meiosis-arresting peptides. A. CYCLIC ADENOSINE 3',5'-MONOPHOSPHATE (CAMP) There are substantial number of reports supporting the hypothesis that cAMP maintains meiotic arrest in oocytes. This contention is based on the finding that the derivatized CAMP,dibutyryl cAMP (db CAMP),blocks the spontaneous resumption of meiosis of isolated cumulus-enclosed and cumulus-free oocytes cultured in vitro (Cho et al., 1974; Magnusson and Hillensjo, 1977; Dekel and Beers, 1978, 1980; Nekola and Smith, 1975; Ahren et al., 1978). Also activators of adenylate cyclase and inhibitors of phosphodiesterase elevate intraoocyte cAMP and prevent GVBD (Nekola and Smith, 1975; Hillensjo, 1977; Hillensjo et al., 1978; Dekel and Beers, 1978; Ekholm rt al., 1984; Hubbard and Terranova, 1982; Dekel rt ul., 1984; Powers and Paleos, 1982; Olsiewski and Beers, 1983; Sat0 and Koide, 1984a). Unmodified cAMP added to the suspending medium of oocytes did not influence the occurrence of GVBD. The finding that unmodified cAMP does not influence GVBD while the derivatized cAMP is an effective inhibitor is attributed to the low uptake of the unmodified cAMP by the oocytes, its low plasma membrane permeability, instability, and rapid metabolism (Hillensjo et al., 1978). To render credence to the hypothesis that cAMP is the regulator of

4

ElMEl SAT0 A N D S. S. KOIDE

meiotic arrest, the level of this nucleotide in oocytes during the resting stage and following resumption of meiosis was determined. The cAMP content of resting oocytes was estimated to be 6.3 5 0.7 fmoVoocyte (Moor and Heslop, 1981). Treatment with gonadotropin to induce maturation did not affect the cAMP level of oocytes. The addition of 3-isobutyl-1-methylxanthine (IBMX), an inhibitor of phosphodiesterase activity, blocked the occurrence of spontaneous GVBD and induced a rise in the intracellular cAMP level of cumulus-free oocytes (Vivarelli et al., 1983). In further support of this hypothesis, Schultz et al. (1983a,b) found that the level of oocyte cAMP decreased significantly during the period when the oocyte resumed meiosis. This fall in cAMP can be inhibited with IBMX at the same time preventing the occurrence of GVBD. This decrease in the oocyte’s cAMP level precedes GVBD and occurs concomitantly with a paradoxical rise in cAMP of the follicular fluid and cumulus cells. These findings suggest that the resumption of meiosis in mammalian oocytes is triggered by a fall in oocyte cAMP level, similar to that observed with amphibian oocytes (Masui and Clarke, 1979). In the Xenopus oocyte progesterone probably acts by inhibiting adenylate cyclase (Sadler and Maller, 1985). Cholera toxin, an activator of adenylate cyclase, inhibited spontaneous GVBD of cumulus-enclosed oocytes but not of cumulus-free oocytes (Dekel and Beers, 1980). The inability of cholera toxin to block GVBD of denuded oocytes is not clear in view of the fact that the oocytes are able to synthesize cAMP and that zona-free oocytes possess adenylate cyclase activity (Schultz et al., 1983a.b; Urner et al., 1983; Sat0 and Koide, 1984a; Bornslaeger and Schultz, 1985). The lack of response of denuded oocytes may be due to the lag period between the time of toxin exposure and the increase in cAMP level (Moss and Vaughan, 1979). The occurrence of a lag period before the resumption of meiosis is further supported by the observation that the fall in cAMP takes place earlier with denuded oocytes compared to cumulus-enclosed oocytes (Dekel and Beers, 1980). We have demonstrated that forskolin, an activator of adenylate cyclase, blocked GVBD of cumulus-free oocytes (Sato and Koide, 1984a). These findings using cholera toxin and forskolin suggest that oocytes do possess adenylate cyclase that lacks the stimulatory GTP-binding regulatory subunit. In this case the enzyme will not be affected by cholera toxin since adenosine diphosphoribosylation of the regulatory subunit will not take place (Gill, 1982). During the initial period of oocyte maturation, protein synthesis takes place leading to GVBD. Richter and McGaughey (1981) reported that the oocytes synthesized stage-specific polypeptides during meiotic maturation and that db cAMP blocked the synthesis of some of these polypeptides.

5

A R R E S T A N D RESUMPTION OF MEIOSIS IN OOCYTES

The relationship of cAMP to protein synthesis is not clear. There are reports indicating that db cAMP did not alter the rate of protein synthesis nor the spectrum of proteins synthesized (Stern and Wassarman, 1974). Nonetheless the clearest evidence indicates that cAMP is involved in meiotic arrest. Determination of the level of cAMP in oocyte at the dictyate stage is a critical factor to test the validity of this hypothesis.

B. CAMP-POTENTIATING FACTORS lntraoocyte level of cAMP may be the physiological factor sustaining meiotic arrest. In addition follicular fluid contains a CAMP-potentiating factor(s) (Eppig et d . , 1983; Eppig and Downs, 1984; Freter and Schultz, 1984; Downs et al., 1985; Racowsky, 1983; Sat0 et ul., 1985). A factor was identified in porcine follicular fluid that blocks mouse oocyte maturation in vitro when combined with cAMP (Eppig and Downs, 1984). The substance was identified to be hypoxanthine (Fig. I ) (Downs et d . , 1985). Based on these findings it was proposed that the active factor is produced by a CAMP-dependent process in the granulosa-cumulus cells. The factor is transported to the oocyte through cytoplasmic channels that couple the cumulus cells to the oocyte. Alternatively, an inactive factor is taken up directly by the oocyte and activated by a CAMP-dependent process. The biochemical steps involved in the activation of the factor and the mech-

Hypoxant hine

Cyclic adenosine 3: 5cpyrophosphate ( C A P P I

0

II

HO-

p- 0- CHz

I 0

I HO-

P

H ’

H d

-0

H OH

II

0

FIG.I .

Diagram of follicular substances that sustain meiotic arrest in mammalian oocytes.

6

ElMEI SAT0 AND S. S. KOIDE

anism of its action are not clear. The presumptive control of oocyte maturation is that there is a decrease in the levels of both the follicular fluid factor and CAMP. A reduction of these factors may result from a fall in their production by the cumulus cells or alternatively a block in their transport from the cumulus cells to the oocyte. These events will trigger the resumption of meiosis. We have recently purified a factor from bovine follicular fluid that inhibits mouse oocyte maturation in combination with cAMP (Sato et ul., 1985). The factor was purified by extraction with 70% ethanol, chromatography on a Dowex 1-X8 column, and reversed-phase high-performance liquid chromatography. The physicochemical properties of the follicular fluid substance are similar to that of cyclic adenosine 3’3’-pyrophosphate (cAPP) (Fig. 1). Both the follicular fluid factor and cAPP in combination with db cAMP blocked mouse oocyte maturation in combination with cAMP and inhibited protein kinase activity. 17P-Estradiol inhibits maturation of denuded porcine oocytes (McGaughey, 1977). This steroid is effective when used in a chemically defined medium containing bovine serum albumin (BSA) or dextran (Richter and McGaughey, 1979), but not in a BSA-free medium (Racowsky and McGaughey, 1982). The inhibition is reversible (Richter and McGaughey, 1979). Testosterone can potentiate the maturation-arresting activity of db cAMP (Richter and McGaughey, 1981; Racowsky, 1983). Androgens, however, may modulate CAMP-induced meiotic arrest in vitro by being converted enzymatically to 17P-estradiol via the aromatase system (Racowsky, 1983). This conversion to 17P-estradiol as the mediator of the meiotic arrest is supported by the observations that follicle-stimulating hormone (FSH) and cAMP stimulate aromatase activity (Lacroix et ul., 1974; Moon et al., 1975; Armstrong et al., 1979; Lindsey and Channing, 1979; Anderson et al., 1979).

C. OOCYTEMATURATION INHIBITOR (OMI) Inhibition of spontaneous maturation of isolated rabbit oocytes by follicular fluid was first described by Chang (1955). Tsafriri and Channing (1975a,b) demonstrated a similar factor in porcine follicular fluid designated as oocyte maturation inhibitor (OMI). Other investigators claim that follicular fluid does not influence oocyte maturation (Liebfried and First, 1980a,b; Racowsky and McGaughey, 1982; Fleming et al., 1983). This controversy has not been resolved. An explanation was offered by Channing et al. (1982a) for the apparent contradictory results. They suggested

ARREST AND RESUMPTION OF MEIOSIS IN OOCYTES

7

that follicular fluid contains an inhibitory factor and a maturation-inducing factor, and varying content of these two factors can account for the different results reported. The inhibitor and inducer were separated by chromatography on CM-Sephadex column (Channing et al., 1982a). The inducer has not been characterized. Several factors with maturation inhibitory activity are present in the follicular fluid. OM1 is a peptide with an estimated molecular weight of 2000 (Tsafriri et al., 1976; Stone et al., 1978). A similar factor was extracted from granulosa cells (Centola et al., 198I). The granulosa cell factor when added to the culture medium prevented oocyte maturation (Tsafriri, 1978b). suggesting that granulosa cells produced OMI. Also it was found that an extract prepared from granulosa cells of small follicles was more potent than those obtained from large follicles, indicating that its content decreases as the follicles mature, paralleling the physiological state of the follicles (Channing et al., 1982a; Tsafriri et al., 1982a,b; Tsafriri and BarAmi, 1982). Another oocyte maturation inhibitory factor was discovered in follicular fluid. Its properties differ from that of OM1 (Chari et al., 1983). It apparently potentiates the inhibitory potency of CAMP (Eppig and Downs, 1984) and is identified as hypoxanthine (Downs et al., 1985). A third inhibitory factor is an immunoreactive prolactin-like substance that cross-reacts with anti-prolactin antiserum (Baker and Hunter, 1978; Channing et al., 1982a). When anti-prolactin antiserum is added to the medium containing follicle-enclosed porcine oocytes, maturation is accelerated (Baker and Hunter, 1978), suggesting that prolactin might induce oocyte maturation. It has been further suggested that this hormone may act indirectly on the oocyte by stimulating the granulosa cells to synthesize OM1 (Channing et al., 1982a,b). The production of OM1 is blocked by testosterone and dihydrotestosterone (Channing et al., 1982a). Since androgens or estrogens fail to influence oocyte maturation directly, follicular androgens probably act by decreasing OM1 production by the granulosa cells (Channing et al., 1982a). OM1 acts on cumulus cells instead of directly on the oocytes since denuded oocytes will undergo spontaneous maturation in the presence of OM1 (Hillensjo et al., 1979). There are multiple effects attributed to OM1 on the cumulus cells. It inhibits the spontaneous maturation of cumulus-enclosed porcine oocytes, prevents morphological differentiation of cumulus cells, and blocks progesterone secretion by cumulus cells (Schaerf et d., 1982). These data suggest that cumulus cells take up OM1 and transport it to the oocyte where it can sustain meiotic arrest. An alternative possibility is that OM1 promotes the production of yet another inhibitor, e.g., CAMP. It may also act by preventing the formation of an oocyte maturation inducer.

8

EIMEI S A T 0 AND S. S. KOIDE

D. GRANULOSA CELLFACTOR(GCF) When isolated porcine oocytes are in contact with porcine granulosa cells, maturation is inhibited (Foote and Thibault, 1969; Sat0 et nl., 1977). Isolated oocytes will remain in the dictyate stage when juxtaposed to a layer of granulosa cells. The oocytes will undergo maturation when detached from these cells. Granulosa cells obtained from small follicles were more potent in preventing the spontaneous maturation of isolated oocytes than cells from Graafian follicles (Tsafri and Channing, 1975a), suggesting that GCF is the active agent within the follicles. Other investigators claim that granulosa cells did not affect the occurrence of spontaneous maturation of isolated oocytes (Liebfried and First, 1980a,b). Nonetheless they found that segments of follicular wall attached to the oocytes can prevent the occurrence of GVBD. The addition of LH to hemisectioned follicles induced resumption of meiosis of the oocytes. The inhibitory potency of granulosa cells can be demonstrated providing the cells are in contact with each other (Sato et al., 1977, 1980, 1982, 1984b, 1986). The mere coculturing of oocytes with a granulosa cell layer (about lo7 cells) obtained from medium-sized (2-5 mm) follicles did not prevent the occurrence of spontaneous GVBD. Inhibition was observed only when the oocytes were in direct contact with the granulosa cells. Maturation block can be induced with cumulus-enclosed oocytes by having a portion of the granulosa cell layer in contact with the cumulus cells, whereas to induce maturation block of denuded oocytes the entire surface has to be enclosed by the granulosa cells (Sato et al., 1982), indicating that meiotic arrest is dependent upon cell-to-cell communication between the cumulus-oocyte complex and the granulosa cells. These findings further suggest that the inhibitory factor is located on the surface of the granulosa cells or may be a component of the extracellular matrix of the granulosa cell layer. The inhibitory factor can be extracted from the surface of granulosa cells with a buffer containing I M urea and 5 mM ethylenediaminetetraacetate (EDTA) (Sato and Koide, 1984b; Sat0 et al., 1986). This buffer was used to dissociate sea urchin embryo (Kondo and Sakai, 1971) to extract surface components from cultured fibroblasts (Igarashi and Yaoi, 1975) and sperm-aggregating factor from Spisula oocytes (Sato et al., 1983). Cells treated with the urea-EDTA solution recover without any deleterious effect. The maturation inhibitory factor was extracted from bovine granulosa cells with a buffer containing 1 M urea and 5 mM EDTA and purified by gel filtration on Sephadex G-25. Two protein peaks were obtained (Fig. 2). The material in the second (minor) peak at a concentration of 400 pg (dry weight)/ml of culture medium completely prevented spontaneous maturation of isolated mouse oocytes. At a lower concentration (50 pg/ml),

ARREST AND RESUMPTION OF MEIOSIS IN OOCYTES

9

O.81

Fraction no.

FIG.2. Gel filtration of peptides extracted from bovine granulosa cells on Sephadex (3-25 column. Column size: 1.5 x 75 cm. Arrows indicate position of reference markers: insulin. 6 kDa: bradykinin, 1.2 kDa. Fraction I (tube numbers 24-32) and fraction 2 (tube numbers 33-49) were pooled. Fraction 2 possessed oocyte maturation-preventing activity. Effective concentration was 400 pg/ml.

it blocked GVBD by 58% (Fig. 3). Peak I (major) possessed slight inhibitory activity. Inhibition was 58 and 6% at concentrations of 4000 and 500 pg/ml, respectively. The factor in peak 2, designated as granulosa cell factor (GCF), was further purified by affinity chromatography on Con ASepharose 4B column. The unabsorbed fraction contained the maturationpreventing activity showing that the factor is probably devoid of sugar moieties. The inhibitory effect of GCF was found to be reversible at lower concentrations. At a high concentration (400 pg/ml), the oocytes remain in meiotic arrest even after washing and transfer to the control medium. At a concentration of 200 pg/ml, approximately 10% of the oocytes have undergone GVBD. The remainder of the oocytes resumes meiosis after washing. GCF at a concentration of 50 pg/ml permits GVBD in 18% of the oocytes after 2 hours of incubation which increased gradually to 35% by the end of 6 hours. During this period, 86% of control oocytes have undergone GVBD. When oocytes are cultured in medium containing GCF for 3 hours and transferred to control medium, 76% of oocytes undergo GVBD compared to 35% of unwashed oocytes. The possibility of contaminating EDTA or urea to account for the inhibitory effect was excluded by experimental design (Sato et ul., 1984b, 1986). GCF is a peptide since it is destroyed by Pronase but not by DNase, RNase, or glycosidase. Its estimated molecular weight has been determined to be less than 6000 by gel filtration on Sephadex G-25. Thus, GCF and OM1 are related compounds possessing common properties.

10

EIMEl SAT0 A N D S. S. KOIDE 100,

’0°1

B

,

100

50 (3

1 2 3 4 5 6 Incubation time (hrs)

FIG.3. Effect of bovine granulosa cell factor (GCF) on the time course of spontaneous GVBD of isolated mouse oocytes. Fraction purified by gel filtration on Sephadex (3-25 was Control medium; -0, GCF in the medium throughout the experiment: used. -, 0-0, GCF in medium for 3 hours, oocytes washed three times, and resuspended in control medium. Concentrations tested were 400 p,g of GCF/ml of medium (A), 200 pg/ml (B). 50 p,g/ml (C). Values are mean 2 SD ( n = 5 ) .

E. CALCIUM Plasma membranes of many cells are usually impermeable to Ca’+. When the cells are stimulated or activated they become sensitive to exogenous Ca” . These findings suggest that the mechanism of activation of mammalian oocytes and other cells may involve influx of exogenous calcium. External Ca2+is essential in maintaining mouse oocytes viable in the culture medium (Paleos and Powers, 1981; De Felici and Siracusa, 1982). Small meiotically incompetent oocytes and early embryos do not require exogenous Ca” in the medium for survival, indicating that the

AKKESI‘ AND RESUMPTION OF MEIOSIS IN OOCYTES

II

Ca” requirement is restricted to specific stages in the growth and development of oocytes (De Felici and Siracusa, 1982). Various hypotheses have been proposed to account for the Ca” requirement of the dictyate oocytes. One possibility is that there is an activation of the membrane calcium pumps and internal calcium buffering systems in the oocytes upon release from the ovary. In the absence of extracellular calcium, the intraoocyte calcium level will fall below that required to sustain metabolic activities. Another reason is that calcium may be necessary for the repair of membrane injury sustained during the mechanical release of the oocytes from the follicles (Okamoto et al., 1977). The effect of calcium was studied by using the calcium ionophore A21 387. Although the ionophore does not influence the spontaneous maturation of isolated rat oocytes, it can induce GVBD in follicle-enclosed rat oocytes (Tsafriri and Bar-Ami, 1978), suggesting that calcium may trigger the resumption of meiosis. This thesis is supported by the report that the total calcium concentration of cumulus-enclosed rat oocytes increases paralleling the serum LH level (Batta and Knudsen, 1980). The rise in oocyte calcium and serum LH levels occurs during the time maturation is initiated. Unfortunately the oocyte calcium level was not determined with the onset of GVBD. Although lowering calcium or magnesium content of the medium appears not to influence the resumption of meiosis of isolated bovine oocytes with adherent cumulus cells, oocyte maturation was blocked when cultured in calcium- and magnesium-free medium (Liebfried and First, 1979). We have demonstrated that there is a dramatic decrease in the incidence of GVBD of oocytes incubated in a Ca”- and Mg”-free medium (Sato er al., 1980). This finding supports the thesis that calcium and magnesium ions are essential ingredients for the occurrence of GVBD. It is interesting that db CAMP-induced meiotic arrest in mouse oocytes can be overcomed by elevating the extracellular calcium level: although at a higher concentration of db cAMP (0.2 mM), the block cannot be reversed with calcium (Paleos and Powers, 1981). Moreover, the ionophore is able to induced GVBD in oocytes treated with db cAMP (0.1 mM) (Powers and Paleos, 1982).These findings suggest that calcium and cAMP may regulate oocyte maturation by influencing a common mechanism. The proposed mechanism is based on the premise that the intracellular reservoir of calcium is sufficient to promote spontaneous GVBD in virro. The addition of db cAMP to the medium may create a need for exogenous calcium (Powers and Paleos, 1982). It is postulated that db cAMP reduces cytoplasmic Ca’+ level by stimulating the calcium pumps of the membrane (Berridge, 1975) as found in other cell systems. To elucidate the role of intracellular Ca2+in the resumption of meiosis of oocytes, further technical

12

ElMEI SAT0 AND S . S . KOlDE

advancement in the method of measuring intracellular Ca” translocation needs to be developed. It is clear that the metabolism and action of calcium ions in eukaryotic cells are regulated by calmodulin. This protein and Caz+are involved in cyclic nucleotide metabolism, protein phosphorylation, microtubule assembly, and calcium flux. To demonstrate the participation of calmodulin in the resumption of meiosis in mouse oocytes, the effect of calmodulin antagonists on GVBD in isolated mouse oocytes was examined. W7 [ N ( 16-aminohexyl)-5-chloroI -naphthalene-sulfonamidehydrochloride] at concentration of 5 x lo-’ M or greater inhibited GVBD (E. Sato and S. S. Koide, unpublished data). The block exerted by W7 is partially reversible. W5 [N-(6-aminohexyl)-1 -naphthalene-sulfonamide hydrochloride], a calmodulin antagonist with less specificity than W7, did not inhibit maturation of isolated oocytes at a concentration of I x M. The meiotic block in cumulus-free and cumulus-enclosed oocytes induced with W7 was not reversed by estrogen, progesterone, or a combination of estrogen and progesterone. These findings suggest that calmodulin may be involved in the resumption of meiosis. Sperm, ethanol, and phorbol ester activate cellular processes associated with a sustained oscillation of [Caz’Ii in mouse oocytes (Cuthbertson and Cobbold, 1985). Although the [Ca”li responses are significantly different with each inducer, activation of the mouse eggs resulted. We hypothesized that oocyte membranes contain a meiotic-arresting component (Sato et al., 1984a). When oocytes are recovered from the ovaries of Spisula, they are arrested in the dictyate stage and possess a large germinal vesicle. Oocyte maturation is signaled by the dissolution of the germinal vesicle which can be induced in Spisula oocytes by sperm, KCI, or serotonin (Allen, 1953; Hirai et al., 1984). Our results show that trypsin induces GVBD of Spisula oocytes only in the presence of Ca” . The maturation-promoting activity of this protease can be blocked by a membrane component(s) (Sato et al., 1984a). These observations suggest that the oocyte membrane component may sustain meiotic arrest and that the hydrolysis of this component may be an early step in the induction of oocyte maturation in Spisula. Since the induction of GVBD with trypsin is dependent upon Ca”, the membrane component may sustain meiotic arrest by preventing Ca” influx. This point needs to be clarified. Proteolytic enzymes may mediate oocyte maturation induced by sperm and may account for the subsequent events associated with fertilization. The proposed hypothesis is as follows: upon sperm-egg interaction, proteolytic enzymes are liberated, hydrolyzing the meiotic-arresting component of the membrane, promoting calcium influx, and triggering the resumption

A R R E S T A N D R E S U M P T I O N OF M E I O S I S I N O O C Y T E S

13

of meiosis. The existence of the meiotic-arresting factor in mammalian oocyte membranes has to be verified.

111. Factors Inducing Resumption of Meiosis

A. GONADOTROPINS LH is considered to be the physiological agent that induces maturation of mammalian oocytes. The hormone acts on follicular cells and not directly on the oocyte. This thesis is based o n the finding that oocytes undergo GVBD when whole ovaries or isolated follicles in organ culture are treated with LH or HCG (human chorionic gonadotropin) (Baker and Neal, 1972; Lindner et a/., 1974; Tsafriri er al., 1972). However, FSH is equally effective and its action is not due to contaminating LH (Lindner et a / . , 1974; Neal and Baker, 1975). The minimal effective dose of FSH required to induce resumption of meiosis in follicle-enclosed oocyte is lower than that of LH (Neal and Baker, 1975). Furthermore, anti-LH antiserum raised against the p-subunit of LH abolishes LH action on the follicular cells, but not that of FSH (Lindner er d.,1974). The capability of LH and FSH to induce oocyte maturation varies with the species. With isolated follicles obtained from swine (Baker et al., 1975) or women (Baker and Neal, 1974),neither LH nor FSH was effective, while both hormones induce maturation of sheep follicle-enclosed oocytes in organ culture. The resulting eggs can be fertilized and develop into viable young embryos when transplanted into suitable recipients (Moor and Trounson. 1977; Staigmiller and Moor, 1984). Cultures of follicle-enclosed oocytes of rat (Lindner rt a / . , 1974), mouse (Baker and Neal, 1972), rabbit (Thibault et d.,1975). and sheep (Hay and Moor, 1975) were used to elucidate the mechanism of LH induction. An early action of LH on follicles is the stimulation of adenylate cyclase activity. Stimulation of the cyclase will increase intracellular cAMP level (Tsafriri et a / . , 1972; Marsh et al., 1973; Nilsson el a / . , 1974) which undoubtedly plays an important role in the resumption of meiosis (Marsh, 1976; Tsafriri et al., 1972). cAMP and its derivatives may promote meiosis under certain conditions. For example, elevating follicular cAMP levels by microinjection of this nucleotide into the follicle (Tsafnn et a / . , 1972) or preincubation of follicles with db cAMP (Hillensjo et al., 1978) promotes resumption of meiosis. The involvement of cAMP in oocyte maturation is further supported by the findings that cAMP levels in isolated follicles are elevated within 5 minutes after exposure to forskolin. reaching a plateau at 15 minutes (Dekel

14

EIMEI SAT0 AND S. S. KOIDE

and Scherizly, 1983). Forskolin mimics LH by stimulating cAMP production in rat ovarian follicles and inducing GVBD in follicle-enclosed oocytes. Derivatized cAMP or cyclic nucleotide phosphodiesterase inhibitors blocks LH-induced maturation of follicle-enclosed oocytes (Hillensjo et al., 1979; Dekel et al., 1981). LH failed to induce GVBD in follicleenclosed rat oocytes exposed to db cAMP (Tsafriri et al., 1972; Hillensjo et al., 1978). Moreover, db cAMP and related compounds in vitro prevent the occurrence of spontaneous GVBD of isolated oocytes. Thus elevating cAMP level in follicular cells will promote oocyte maturation, while cAMP acting directly on the oocyte prevents GVBD. To explain this paradoxical action of cAMP on oocyte maturation, it is hypothesized that, at the time when cumulus-oocyte complexes are isolated from follicles, cAMP level is sufficient to sustain meiotic arrest. Because oocytes contain an active phosphodiesterase, the intraoocyte cAMP level falls rapidly, triggering oocyte maturation. It is proposed that the oocyte phosphodiesterase is activated following the isolation procedure whereby the oocyte is removed from the action of an inhibitor of the enzyme present in the follicular fluid. LH-induced oocyte maturation is dependent on the stimulation of adenylate cyclase by the hormone and to the subsequent rise in cAMP levels within the follicular cells. The apparent contrasting findings to reconcile is the high cAMP level in granulosa cells accompanied by a fall of cAMP level in the oocyte (Channing and Tsafriri, 1977). A postulated mechanism is that follicular cells are interlinked by gap junctions. As a consequence of the rise in the cAMP level in the granulosa cells, the cell-to-cell communication system is disrupted and the transport of metabolites from the cumulus cells to the oocyte is blocked. It has been proposed that LH in some indeterminate manner terminates cell-to-cell communication in the cumulus-oocyte complex and that nucleotide transport to the oocyte ceases promoting the resumption of meiosis (Dekel and Kracier, 1978; Dekel ef al., 1981, 1984). This hypothesis that the uncoupling of the oocyte and cumulus cells is a prerequisite for the resumption of meiosis has been questioned by the findings that oocytes and cumulus cells remain interlinked to one another even after the resumption of meiosis (Moor et ul., 1980; Eppig, 1982) and that transport of CAMP from the cumulus cells to the oocyte may not take place under physiological condition. An alternative hypothesis is that LH may prevent the action of a maturation-arresting factor or a CAMP-potentiating factor produced by the granulosa cells on the oocyte. This hypothesis is supported by the findings that gonadotropins stimulate the production of the extracellular matrix and promote the accumulation of glycosaminoglycans in the interstitial spaces between the cumulus cells. The glycosaminoglycan components can interact with the granulosa cell factor and neutralize its meiotic-

A R R E S T A N D RESUMPTION OF MEIOSIS IN OOCYTES

15

arresting activity, thereby promoting resumption of meiosis (Sato et ul., 1984b, 1986). Steroids have been implicated as regulators of oocyte maturation. After the LH surge, the concentrations of estradiol in follicular fluid decline, initially followed by a marked rise in the synthesis of progesterone (Thibault, 1977; Eiler and Nalbandov, 1977; Dorrington and Armstrong, 1980). The relative concentrations of progesterone to estradiol in the follicular fluid is elevated at the onset of oocyte maturation (Thibault, 1977; Gerard et a / . , 1979). Numerous steroids were added to the culture medium of isolated oocytes to determine their ability to influence oocyte maturation. Conflicting results were obtained (McGaughey, 1983). Bae and Foote (1975) demonstrated that progesterone added to the culture medium stimulated maturation of bovine and rabbit oocytes. Other investigators were unable to demonstrate any effect of progesterone on the maturation of isolated porcine oocytes with or without adherant cumulus cells in vitro (Richter and McGaughey, 1979). The LH induction of oocyte maturation in cultured Graafian follicles obtained from rat ovaries was not impaired when hormone-induced steroidogenesis was completely suppressed with 17P-ol-3-one)or cyanoketone (2a-cyano-4,4,17a-trirnethylandrost-5-enaminoglutethimide (Tsafriri et al., 1972). Also GVBD was not induced in cultured rat follicles treated with progesterone, 20a-dihydroprogesterone, or 17P-estradiol. None of the steroids tested inhibited LH-induced GVBD. Thus they concluded that the resumption of meiosis induced by LH is not dependent upon its ability to influence the rate and pattern of follicular steroidogenesis (Lieberman et al., 1976). B. GONADOTROPIN-RELEASING HORMONE (GnRH) The principal function of GnRH is to promote LH and FSH release from the pituitary gland. An additional response attributed to GnRH agonist is to mimic LH in hypophysectomized rats by inducing resumption of meiosis and dispersion and mucification of cumulus cells (Ekholm et a / . , 19811. GnRH and its agonists stimulate maturation of follicle-enclosed oocytes in an in vitro culture system in a dose-dependent manner (Hillensjo and LeMaire, 1980). However, these hormones did not influence the spontaneous GVBD of isolated oocytes (Anderson and Hillensjo, 1982). Also the GnRH antagonist effectively abolished the stimulating effect of a GnRH agonist to induce GVBD of oocytes in isolated preovulatory rat follicles and yet did not influence the inducing action of LH. These findings suggest that GnRH acts via the follicular cells and does not directly affect the oocytes, and its action appears to be independent of LH effect on meiosis.

16

EIMEl SAT0 A N D S. S. KOlDE

c. REGULATORSOF MEIOTICCOMPETENCE As oocytes develop they mature and acquire the ability to resume meiosis beyond the dictyate stage. This property is acquired during specific stages of development. Oocytes recovered from mice younger than 15 days postpartum are unable to undergo GVBD in vitro (Szybek, 1972). Similarly the proportion of oocytes recovered from prepubertal rabbits (70-90 days of age) that had matured to the second metaphase in vitro was lower than adult rabbits (Thibault, 1977). We have found that about 50% of the oocytes recovered from prepubertal swine undergoes spontaneous GVBD (Sato et al., 1977), indicating that the capacity to resume meiosis is acquired during the late stages of oocyte growth and development. Although the exact developmental age of the animal when oocyte competence is acquired has not been determined, it probably corresponds to the time period when oocytes attain full growth. One of the criteria used is the size of the oocytes. Incompetent oocytes recovered from small follicles were significantly smaller than competent oocytes recovered from larger antral follicles, indicating that there is a correlation between maturation competence and fertilizability (Sorensen and Wassarman, 1976). However, the average diameter of competent rat oocytes explanted on day 20 was 61.8 & 1.2 pm compared to 76.5 ? 0.8 pm for incompetent oocytes (Bar-Ami and Tsafriri, 1981), suggesting a lack of correlation between oocyte size and competence. The acquisition of competence occurs at a defined period of growth and development when some essential components are being produced. Immaturity may result from a deficiency of specific components required for maturation, for example, in the assembly of spindle proteins as found in immature amphibian oocytes (Brachet, 1977). One of the essential components may be RNA (Iwamatsu and Yanagimachi, 1975). RNA synthesis increases gradually during growth and development in mouse oocytes (Moore et d.,1974) and decreases when they reach their maximal size at the Graafian follicle stage shortly before ovulation. Investigations to determine the precise period when the developing oocytes attain competence to resume meiosis have been carried out. In the mouse, oocytes acquire the capacity to resume meiosis shortly before the appearance of the antrum in the primary follicle (Pincus and Enzmann, 1937). This belief is based on the findings that oocytes attain their maximal size at this time, and it is the earliest stage when meiotic maturation figures can be identified in oocytes of atretic follicles (Engle, 1927).The strongest support of this thesis is the finding that 83-91% of the oocytes removed from antral follicles (300-600 pm in diameter) progressed to the first or

ARREST AND RESUMPTION OF MEIOSIS IN OOCYTES

17

second metaphase, whereas 98% of those isolated from preantral follicles failed to undergo spontaneous maturation (Erickson and Sorensen, 1974). There is species variability in the potential to undergo maturation. In swine and cattle, 90-100% of the oocytes isolated from Graafian follicles (1015-mm diameter) undergoes GVBD, while 50% of the oocytes removed from small follicles (2-5 mm in diameter) resumes meiosis. This indicates that the potential to undergo maturation is acquired during the maturation of follicles (Sato et al.. 1977). The acquisition of meiotic competence, i.e., the ability to resume meiosis, may be influenced by hormones that regulate follicular development. This thesis is based on the findings that oocytes from hypophysectomized rats, performed on 15 days postpartum, fail to undergo maturation (Bar-Ami rt ul., 1983). The ability of the oocytes to resume meiosis can be reversed on administering FSH to the hypophysectomized rats, but not with LH. Furthermore, 17P-estradiol administered as an implant to hypophysectomized rats partially restored meiotic competence to the oocytes within 24 hours, while progesterone and androstenedione were not effective. Also coadministration of inhibitors of steroidogenesis with FSH to the hypophysectomized rats did not restore meiotic competence to the oocytes. These findings suggest that FSH is involved in the induction of meiotic competence and that its action is partially mediated by estrogens produced within the follicles. D.

MATURATION PROMOTORS in

Vitro VERSUS in v i v o

The observed spontaneous maturation of isolated oocytes in vitro may or may not reflect the natural occurrence of maturation under in vivo conditions (Motlik and Fulka, 1976).The changes associated with maturation appear earlier in isolated oocytes cultured in vifro compared to those induced to mature within the follicles after HCG injection (Thibault, 1977). The observed time lag in vivo corresponds to the duration when HCG released from the injection site reaches a critical level to initiate its action. In the swine, this time lag is quite prolonged (Motlik and Fulka, 1976; Sat0 et al.. 1978a,b). Nonetheless, the rate of in vitro maturation is affected by the culture conditions and hormones added to the media (Sato et d., 1978a,b). For example, the pH of the culture medium influences the maturation time of porcine oocytes. At pH 6.8-7.0, the time required to progress to the second metaphase was delayed by several hours, compared to pH 7.2-7.4. The delay occurs at the early stage of the process leading to GVBD, while the time from the first to the second metaphase was not appreciably changed.

18

ElMEl S A T 0 A N D S. S. KOIDE

Denuded and cumulus-enclosed oocytes resume meiosis spontaneously at the same rate when cultured in vitro. FSH added to the medium caused a delay in the occurrence of meiotic resumption of cumulus-enclosed oocytes, but not with denuded oocytes. The temporal sequence of events with cumulus-enclosed oocytes paralleled the process of oocyte maturation in vivo (Salustri and Siracusa, 1983). LH added to the culture medium of rat oocytes accelerated the progress of GVBD (Lopata et al., 1977; Kaplan et al., 1978). The period of dissolution of the nuclear membrane ranged from 105 to 130 minutes after isolation determined by cinemicrography. This period is shortened to 45-80 minutes after the addition of LH to the medium (Lopata et al., 1977).

E. MATURATION-PROMOTING FACTOR(MPF) The formation of MPF during maturation of mammalian, amphibian, and invertebrate oocytes has been well documented (Masui and Clarke, 1979). Hybrid cells formed by fusing dictyate-arrested oocytes obtained from sexually immature mice and maturing oocytes with Sendai virus will undergo germinal vesicle dissolution and proceed to the first metaphase (Balakier, 1978). Using a similar method, fused porcine and rabbit oocytes were prepared. The presence of MPF in the maturing oocytes was demonstrated (Fulka, 1983). It should be pointed out that the capability to produce MPF is acquired on reaching maturity because small oocytes from sexually immature mice do not undergo GVBD in vitro (Szybek, 1972; Sorensen and Wassarman, 1976). The direct demonstration of MPF production in mouse oocytes undergoing spontaneous maturation was carried out by Kishimoto et al. (1984), who showed that the cytoplasm from maturing mouse oocytes induced GVBD when microinjected into starfish oocytes. MPF in the Xenopus oocyte cell-free system induces phosphorylation of lamins A and C followed by a gradual depolymerization of the nuclear lamina and finally to the dissolution of the nuclear envelope (Miake-Lye and Kirschner, 1985). It should be pointed out that MPF from oocytes (meiosis) and somatic cells (mitosis) is interchangeable and is active over the broad evolutionary range of species (Sunkara et al., 1979; Nelkin et al., 1980; Kishimoto et al., 1982, 1984; Miake-Lye and Kirschner, 1985). The oncogene product, ras protein, microinjected into Xenopus oocyte, induces GVBD (Birchmeier et a/., 1985), simulating MPF action, while the protooncogene product of ras was less effective. It is noteworthy that the intraoocyte CAMP level did not change and yet cholera toxin blocked ras action. To clarify how MPF and ras protein induce GVBD is a fertile field of study.

ARREST A N D RESUMPTION OF MEIOSIS IN OOCYTES

19

IV. Mechanism of Meiotic Resumption A. PERMEABILITY OF CUMULUS-OOCYTE COMPLEXES The major cellular elements within the follicles are the granulosa cells. The granulosa cells are heterogeneous, showing structural variation and differences in hormone binding. They are found in association with a variety of tissue structures and cellular components, i.e., juxtaposed with the follicular wall, interconnected to other granulosa cells, and immersed in antral fluid (Amsterdam et al., 1976; Gilula et al., 1978). The innermost layer of the cumulus, the corona radiata, is composed of granulosa cells surrounding the oocyte. These cells possess cytoplasmic processes that pierce through the intervening zona pellucida and are in contact with the oolemma (Bjorkman, 1962; Odor, 1960; Zamboni, 1974). There are gap junctions in the regions of contact between the cumulus cells and the oocyte (Amsterdam et al., 1976; Lindner ef al., 1977; Anderson and Albertini, 1976).Gap junctions have been implicated as structural pathways for cellto-cell communication (Gilula et al., 1972). Through these junctions, the exchange of nutrients between the cumulus cells and the oocyte takes place during follicular development. These specialized membrane structures contain channels that permit intercellular exchange of metabolites with molecular weights less than 1000 (Flagg-Newton et al., 1979; Anderson and Albertini, 1976). The cumulus cells may serve an important function in transporting essential substances from granulosa cells to the surface of the oocytes. The entire cumulus layer is connected by gap junctions (Gilula et al., 1978). The cumulus-oocyte complex acts as a single unit in the transport of small essential nutrients. Whether or not the entire cellular components of the follicle are interconnected by gap junctions has not been verified. It is unlikely that they are interconnected, since there is a marked regional variation in the response of the granulosa cell population to gonadotropins. The involvement of the cumulus cells on oocyte maturation and in the resumption of meiosis has not been clarified. Removal of cumulus cells from hamster and rat oocytes accelerated the process of spontaneous maturation (Gwatkin and Andersen, 1976; Dekel and Beers, 1980). The maturation rate was slower in cumulus-deprived mouse oocytes (Cross and Brinster, 1970). Differences in the maturation rate was not detected by other investigators who studied cumulus-enclosed and cumulus-free mammalian oocytes (Cross, 1973; Binor and Wolf, 1979). The morphological features of the cumulus cells suggest that these cells may influence the permeability of the oocyte or the transport of essential

20

ElMEl S A T 0 AND S . S. KOlDE

nutrients to the oocyte (Heller et a / . , 1981). That is to say cumulus cells act cooperatively in the transfer of metabolites from granulosa cells or follicular fluid to the oocytes. This thesis is supported by demonstrating significantly higher uptake of radiolabeled leucine, uridine, and ribonucleosides by cumulus-enclosed oocytes than by denuded oocytes (Cross and Brinster, 1974; Wassarman and Letourneau, 1976; Heller and Schultz, 1980). Cumulus cells can facilitate the entry of tritium-labeled choline, uridine, and inositol into oocytes via the gap junctions (Moor et ul., 1980). Active transport of metabolites by the cumulus cells may take place by the following processes: the metabolities may ( I ) undergo structural changes that would facilitate their passage through the membranes, (2) increase the permeability capacity of the oolemma, (3) promote formation of direct communicatingjunctions between cumulus cells and the oocyte. The most probable cause is due to an increase in membrane permeability, since direct oocyte-cumulus cells contact is a prerequisite for the facilitated transfer of metabolites (Moor et al., 1980). FSH effectively decreases the flow of small molecules from the cumulus cells to the oocytes (Moor et a/., 1980). Steroids may be involved as mediators of gonadotropin suppression and in the maintenance of junctional competence between cumulus cells and oocytes. Several studies revealed that the degree of metabolic dependency between cumulus cells and oocytes decreased during meiotic maturation (Cross and Brinster, 1974; Moor et ul., 1980; Heller and Schultz, 19801, suggesting that the permeability of metabolites from the cumulus cells to the oocyte decreased with the resumption of meiosis. In antral follicles prior to the preovulatory stage, the oocyte is surrounded by tightly packed cumulus cells. The cumulus cells in contact with the oocyte become elongated and send fine processes radiating toward the oocyte to form the corona radiata (Eppig, 1982; Dekel and Phillips, 1979; Gilula et ul., 1978). Following the preovulatory LH surge, the follicles mature, accompanied by the disintegration of the cumulus structure, resulting from the accumulation of glycosaminoglycans in the intercellular spaces (Dekel et ul., 1979) and a decrease in the permeability of the cumulus-oocyte complexes. In the follicles with fully expanded cumulus cell masses isolated during the late pre- and postovulatory periods, the oocytes have undergone nuclear membrane dissolution. These findings indicate that the following three events are linked, i.e., alterations in the cumulus cell mass, decreased permeability of cumulus-oocyte complexes, and resumption of meiosis. It should be pointed out that the granulosa cell layer from large Graafian follicles containing the expanded cumulusoocyte complex possesses meiotic-arresting activity. This effect was

ARREST A N D RESUMPTION OF MEIOSIS IN OOCYTES

21

demonstrated by attaching oocytes to the granulosa cell layer from medium-sized follicles (Sato e t ul., 1977). This finding supports the thesis that the expanded cumulus contains substances that prevent the action of the meiotic-arresting factor or that the oocyte is committed to undergo maturation. Various components of the extracellular matrix were tested for their ability to block the action of meiosis-arresting factor. Hyaluronic acid, chondroitin sulfate, heparin, heparan sulfate, and dextran sulfate at concentrations of 500 pg/ml or less did not influence the spontaneous maturation of isolated mouse oocytes in vitro. Heparin and heparan sulfate at concentrations exceeding 100 & n l . however, blocked the inhibitory action of purified GCF (Sato et al., 1984b, 1986). GCF was purified by gel filtration on Sephadex (3-25 and by reversed-phase high-performance liquid chromatography. Heparin did not influence the maturation inhibitory activity of db CAMP, forskolin, calmodulin antagonist (W7), or IBMX (Sato ct al., 1984b, 1986). When purified GCF was applied to a heparinagartose column, the factor was eluted in the adsorbed fraction, suggesting that GCF interacts with heparin directly. This direct interaction of GCF and heparin may account for the loss of meiotic-arresting activity after incubation with heparin (Sato et al., 1984b. 1986). It should be pointed out that the predominant glycosaminoglycan in the cumulus-expanded intracellular matrix is hyaluronic acid. The physiological relevance of heparin-GCF interaction and the high content of hyaluronic acid in cumulusexpanded follicles need to be clarified. The present results indicate that the cumulus cells may play a key role in meiotic arrest and in the resumption of meiosis. During the expansion of the cumulus mass, the glycosaminoglycans of the extracellular matrix may nullify the activity of the meiotic-arresting factors produced by the granulosa cells and/or present in the follicular fluid, triggering the resumption of meiosis.

B. CUMULUS DIFFERENTIATION At the time of cumulus expansion, meiotic maturation of oocyte takes place within the follicle (Dekel et al., 1979; Schuetz and Swartz, 1979). Other studies suggest that the dissociation of the cumulus-oocyte junctions may follow rather than precede GVBD in the rat (Hillensjo et ul., 1979; Dekel and Kraicer, 1978) and rabbit (Szollosi et al., 1978). The administration of pregnant mare serum gonadotropins (PMSG) may initiate GVBD without inducing changes in the cumulus masses (Vermeiden and Zeilmaker, 1974). In mice, 3 hours after the injection of HCG, more than 90% of the oocytes isolated from large Graafian follicles had undergone

22

ElMEI SAT0 AND S. S. KOIDE

GVBD, although cumulus expansion and changes in intercellular communication were not observed (Eppig, 1982). In growing follicles, gap junctional contacts are detected in the cumulusoocyte complex; while in the Graffian follicles, the numbers are reduced and the cumulus layer appears to undergo disorganization. At the time of ovulation, cell-to-cell communication is disrupted, and the remaining cumulus cells become loosely arranged with bullous cytoplasmic extensions (Moor et al., 1980), suggesting that the reduction in intercellular communication may be a consequence of cumulus expansion which continues after oocyte maturation. Dekel et al. (1979) pointed out that cumulus expansion is a gradual process in that the accumulation of glycosaminoglycans takes place initially at the periphery and gradually involves the central portion of the cumulus cell mass. The final stage is the disruption of the cumulus attachment to the oocyte (Dekel et al., 1979), while the disintegration of cell-to-cell communication between cumulus-granulosa cells may occur at the beginning of cumulus dispersion. The end result is that the oocyte escapes from the maturation inhibitory effect of the granulosa cells due to the binding of the meiotic-arresting factor and from the reduced flow of cAMP to the oocyte (Sato et ul., 1984b, 1986). Cumulus expansion of isolated rat follicles cultured in a chemically defined medium is stimulated by gonadotropins (Hillensjo ef al., 1976). Both LH and FSH can induce cumulus expansion in isolated cumulus-oocyte complexes. However, other investigators reported that FSH and not LH induced cumulus expansion in isolated mouse cumulus-oocyte complexes (Dekel and Kraicer, 1978; Dekel et al., 1979; Eppig, 1979a,b, 1980a,b). Although there is a general consensus that LH is the only gonadotropin capable of stimulating follicular events associated with ovulation, FSH appears to be responsible for cumulus expansion (Eppig, 1979a,b). Also follicular fluid obtained from PMSG-primed mice was very active in stimulating cumulus expansion of isolated cumulus-oocyte complexes, demonstrating that follicular fluid contains a FSH-like activity capable of stimulating extracellular matrix synthesis. The mediating factor may be cAMP because db CAMP,phosphodiesterase inhibitors, and cholera toxin stimulate cumulus expansion in vitro (Hillensjo, 1977; Dekel and Beers, 1978; Eppig, 1979a,b). The granulosa and cumulus cells differ in their capacities to produce glycosaminoglycans in response to gonadotropins. The cumulus and granulosa cells originate from a single layer of follicular cells surrounding the primordial oocyte (Franchi et al., 1962). Gonadotropins stimulate mucification of the cumulus and not of the granulosa cells (Dekel er al., 1979). It is noteworthy that the ability of FSH to induce cumulus expansion in vitro is dependent upon the addition of fetal bovine serum (FBS) to the

ARREST A N D RESUMPTION OF MEIOSIS IN OOCYTES

23

cultured medium (Eppig, 1980a,b). FSH stimulates the synthesis of glycosaminoglycans in oocyte-cumulus cell complexes in the presence or absence of FBS. In the presence of FBS the glycosaminoglycans are retained within the complexes, while in the absence of FBS they are released into the culture medium. Containment of glycosaminoglycans within the complex corresponds to cumulus expansion or mucification. The active factor in FBS, promoting cumulus expansion by inducing retention of the glycosaminoglycans within the complex (Eppig, 1979a,b, 1980a,b), has been found to be a protein with a molecular weight exceeding 10,000. Follicular fluid from preovulatory follicles contain components similar to that of FBS (Hadek, 1963; Edwards, 1974) and potentiate the action of FSH in promoting cumulus mucification. Based on these findings, it was hypothesized that the active factor(s) in follicular fluid is a core component of the extracellular matrix of cumulus cells which binds the glycosaminoglycans produced to form the mucified matrix. Since follicular fluid contains an FSH-like factor that stimulates cumulus expansion, why is it that cumulus expansion does not occur spontaneously and is dependent upon a LH surge or on HCG administration? It is postulated that there may be components in Graafian follicles that block the activity of the FSH-like factor (Eppig, 1980a,b, 1981a,b; Eppig and WardBailey, 1984). Sulfated glycosaminoglycans may be involved in suppressing the activity of the FSH-like substance, since these compounds block the incorporation of radiolabeled glucosamine into the extracellular matrix components of cultured cumulus-oocyte complexes from mice (Eppig, 1981a,b). The order of potency of sulfated glycosaminoglycans to inhibit FSH-induced cumulus expansion correlates with the degree of sulfation: heparin > heparan sulfate > chondroitin sulfate B > chondroitin sulfate C > chondroitin sulfate A (Eppig, 1981a). The sulfated glycosaminoglycans may inhibit the synthesis of the extracellular matrix components in response to FSH. Eppig and Ward-Bailey (1984) showed that sulfated glycosaminoglycans suppress the synthesis of glycosaminoglycans after the development of partial cumulus expansion. We proposed that heparin and heparan sulfate accumulate in the extracellular spaces between cumulus cells at the beginning of the expansion and interact with the granulosa cell factor with maturation-preventing activity (Sato ef af., 1984b, 1986). C. FOLLICULAR GLYCOSAMINOGLYCANS

The contents of glycosaminoglycans in follicular fluid is about 0.2-0.3% ( w h ) (Jensen and Zachariae, 1958; Yanagishita et al., 1979). The glycosaminoglycans of follicular fluid are composed of a core protein with an estimated molecular weight of 400,000 with an average of 20 dermatan

24

EIMEI SAT0 AND S . S. KOlDE

sulfate chains and 350 sialic acid-containing oligosaccharides (Yanagishita et ul., 1979). The glycosaminoglycans of rat and porcine granulosa cells are chemically similar (Yanagishita and Hascall, 1979; Yanagishita et al., 1979). The follicular fluid glycosaminoglycans are produced by the granulosa cells. It is known that rat granulosa cells cultured in vitro synthesize and secrete glycosaminoglycans into the medium at a linear rate (Yanagishita et al., 1979). The mechanism of how the glycosaminoglycans accumulate in the intercellular spaces is not clear. Yanagishita and Hascall ( 1979) reported that 90% of the glycosaminoglycans secreted by rat granulosa cells are susceptible to hydrolysis by chondroitinase A, B, and C, indicating that the major constituent produced is chondroitin sulfate. Heparan sulfate was not detected in porcine follicular fluid (Yanagishita and Hascall, 1979). Heparin-like substances, chondroitin sulfate, and dermatan sulfate were identified in rat ovarian tissue and follicles (Gebauer et al., 1978; Ax and Ryan, 1979a,b), indicating that these substances were synthesized in the follicles and may block the maturation inhibiting activity of GCF. (Sato et ul., 1984b, 1986). It was also observed that the production of glycosaminoglycans varied with the size of the follicles. Granulosa cells from large porcine follicles produced slightly less glycosaminoglycans compared to the cells from small follicles (Schweitzer et al., 1981). As the follicles mature, the content of glycosaminoglycans decreased markedly in porcine and bovine follicular fluids (Ax and Ryan, 1979a,b; Grimek and Ax, 1982). The production of glycosaminoglycans was studied by determining the rate of incorporation of radiolabeled precursors. FSH stimulated the incorporation of precursors into rat and porcine follicular glycosaminoglycans (Mueller et ul., 1978; Ax et al., 1978; Ax and Ryan, 1979a,b). In contrast, LH decreased glycosaminoglycans production in rabbit follicles in vitro (Zachariae, 1957) and in rat ovarian slices in vitro (Gebauer et al., 1978). As described in an earlier section, db CAMP can mimic FSH stimulation of glycosaminoglycans production by porcine and rat granulosa cells (Ax and Ryan, 1979a; Schweitzer et al., 1981). Progesterone, in contrast, decreased glycosaminoglycans synthesis by rat ovarian slices (Gebauer et al., 1978) and porcine granulosa cells (Schweitzer et al., 1981). During the occurrence of cumulus expansion, glycosaminoglycans accumulate in the intercellular spaces between the cumulus cells. The major component produced by the mouse cumulus-oocyte complexes in response to gonadotropin is hyaluronic acid (Eppig, 1979a,b). Also, Ball et al. (1982) reported that, during the expansion of bovine cumulus-oocyte complexes induced with FSH or CAMP, massive amounts of glycosaminoglycans accumulated in the matrix enveloping the cumulus cells. The intercellular matrix material was isolated and subjected to electrophoretic analysis.

ARREST A N D RESUMPTION OF MEIOSIS IN OOCYTES

25

The radiolabeled glycosaminoglycans comigrated with reference hyaluronic acid. The substance was resistant to chondroitinase A, B, and C and nitrous acid degradation and hydrolyzed by hyaluronidase. It was concluded that the glycosaminoglycans of the intercellular matrix produced by bovine cumulus-oocyte complexes are rich in hyaluronic acid (Ball et al., 1982). A portion of the glycosaminoglycans of the extracellular matrix of the cumulus should contain heparin-like substance, since ovarian tissues are known to produce this substance (Gebauer et al., 1978; Ax and Ryan, 1979a).

D. A HYPOTHETICAL

S C H E M E OF O O C Y T E

MATURATION

The pertinent facts relating to oocyte maturation can be summarized as follows: isolated follicle-enclosed oocytes remain arrested in the dictyate stage (Tsafriri, 1978a; Lindner et al., 1983). When released from the follicles, the oocytes undergo spontaneous maturation (Pincus and Enzmann, 1935; Biggers, 1973), indicating that meiotic-arresting factors within the follicles maintain the oocyte in the dictyate stage. Several maturationpreventing factors have been identified in follicular fluid and granulosa cells. Resumption of meiosis of cumulus-enclosed oocytes can be prevented by being in contact with the granulosa cell layer (Sato et al., 1982, 1984b).The meiotic-arresting activity of the granulosa cells is demonstrable only when they are adherent to the cumulus-oocyte complex (Sato et a l . , 1982). Based on these studies, a substance with meiotic-arresting activity from the intercellular matrix and the external surfaces of bovine granulosa cells has been isolated (Sato et al., 1984b, 1986) and designated as the granulosa cell factor (GCF). Other maturation inhibitory factors found in follicular fluid are OM1 and hypoxanthine. The latter substance acts only in combination with cAMP (Tsafriri, 1984; Downs e / al., 1985). Follicular fluid may contain another nucleotide with meiotic-arresting activity in combination with cAMP (Sato e/ ul., 1985). Among these arresting factors, GCF is the most potent agent. GCF and OM1 possess similar properties and may possibly be identical substances. Nonetheless, at least two types of factors appear to sustain meiotic arrest: a peptide and a base or nucleotide. During cumulus expansion, cells in the periphery of the cumulus separate and begin to disperse, while the oocyte undergoes maturation (Dekel et al., 1979). The resumption of meiosis in oocytes surrounded by dispersed cumulus is not inhibited even when the cumulus-oocyte complexes are in contact with a layer of adherent granulosa cells (Sato et d., 1982).Thus the dispersion of the cumulus cells, resulting from the accumulation of glycosaminoglycans in the intercellular matrix, will promote maturation.

26

ElMEl S A T 0 AND S. S. KOIDE RESTING OOCYTE c o o c y t e cytoplasm

Follicular flLid inhibitors

FIG.4. Hypothetical scheme of follicular factors inducing meiotic arrest and promoting maturation of mammalian oocytes. A factor, GCF, located on the surfaces and in the intercellular spaces of granulosa cells sustains meiotic arrest by acting on the cumulus-oocyte complex by direct contact, increases CAMP level within the cumulus-oocyte complex, and inhibits calcium transport. Additional associated meiotic-arresting substances are the oocyte

ARREST A N D RESUMPTION OF MEIOSIS IN OOCYTES

27

LH stimulates the preovulatory follicle to produce glycosaminoglycans which accumulate on the surface and in the intercellular spaces of the cumulus cells (Eppig, 1979a,b). Following cumulus expansion, the oocytes undergo maturation. We have demonstrated that among the glycosaminoglycan, heparin and heparan sulfate interact with GCF and nullify its maturation inhibitory activity (Sato et al., 1984b, 1986). A consequence of cumulus expansion is a decrease in the permeability of the oolemma and the transport of essential nutrients from cumulus cells to the oocyte (Heller and Schultz, 1980; Moor et af., 1980). Based on the above findings, the following hypothesis is proposed to explain the mechanism of meiotic arrest and the resumption of meiosis (Fig. 4). GCF is located on the external surface and in the intercellular spaces of the granulosa cells. By direct contact between the granulosa cells and the cumulus-oocyte complex, GCF can sustain meiotic arrest. GCF acts on the cumulus cells and stimulates an increase in the intraoocyte level of CAMPand suppresses calcium transport. Gonadotropins stimulate the production and accumulation of glycosaminoglycans in the intercellular spaces of the cumulus cells. The net result is a disruption of the gap junctions between cumulus cells and between the cumulus cells and oocytes, impeding the transport of metabolites within the cumulus-oocytes complexes. The glycosaminoglycans can prevent the action of GCF on the oocyte by binding the factor. Cumulus expansion results in preventing the action of the maturation inhibitors on the oocyte by binding these factors, reducing the flow of essential nutrients to the oocyte, and disrupting the cell-to-cell communication system within the follicles. In this manner, the oocyte is protected from the arresting influence of the maturation inhibitory factors and resumes meiosis. Protein phosphorylation and dephosphorylation may be involved in meiotic arrest and the resumption of meiosis. In mouse oocytes, phorbol esters prevent spontaneous GVBD (Urner and Schoderet-Slatkine, 1984). Stith and Maller (1985)reported that phorbol esters can induce maturation of Xenopus oocytes, providing they are primed initially with inositol I ,4,5trisphosphate (IP,) which facilitates Ca” release (Berridge, 1984). Microinjection of IP, alone into Xenopus oocytes induced Ca2+release without GVBD (Busa et al., 1985). Tumor-promoting phorbol ester, phospholipase C. and diacylglycerol, l-oleoyl-2-acetylglycerol,induced GVBD maturation inhibitor and CAMP-potentiating factors including hypoxanthine and a novel nucleotide. Luteinizing hormone stimulates the production and accumulation of glycosaminoglycans. specifically of heparin-like substances, that surround the cumulus4ocyte complex. The intercellular matrices prevent the action of CGF on the oocyte and interfere with the flow of metabolites to the oocyte. promoting the resumption of meiosis.

28

EIMEI S A T 0 AND S. S. KOIDE

in follicle-enclosed rat oocytes (Aberdam and Dekel, 1985). These findings suggest that Ca" release in combination with diacyglycerol promotes GVBD. Ca" and diacylglycerol are known to activate protein kinase C (Nishizuka, 1984), which mediates protein phosphorylation. Since diacylglycerol and IP, are hydrolytic products of phospholipase C action on phosphatidylinositol (Berridge, 1984), progesterone may stimulate this membrane enzyme on interaction with its receptor.

ACKNOWLEDGMENTS This study was supported in part by grant numbers HD 13184 from NICHD. NIH. GA PS 8506 ( S . S . K.), and 8429 (E. S . ) from the Rockefeller Foundation and grant numbers 60304036 and 60760207 from the Ministry of Education, Science and Culture. Japan ( E . S.).

REFEKENCES Aberdam, E., and Dekel, N. (1985). Biochem. Biciphvs. Res. Coinmrrn. 132, 570-574. Ahren, K . . Dekel, N.. Hamberger, L., Hillensjo, T., Hultborn. R., Magnusson. C . . and Tsafriri, A. (1978). Ann. Biol. Anirn. Biochim. Bioip/iv.s. 18, 409-417. Albertini, D. F. (1984). B i d . Reprod. 30, 13-28. Allen, R. D. (1953). Biol. Bull. (Woods Hole. Moss.) 105, 213-239. Amsterdam. A., Josephs, R.. Lieberman. M. E., and Lindner, H. R. (1976). J. C P / /S1.i. 21, 93-105. Anderson, E.. and Albertini. D. F. (1976). J . Ci4l Biol. 71, 680-686. Anderson, L. D.. and Hillensjo, T. (1982). Aciu P/iy.siol. Sectrid. 114, 623-625. Anderson, L. D.. Schaerf, F. W., and Channing, C. P. (1979). I n "Ovarian Follicular and Corpus Luteum Function" (C. P. Channing, J. Marsh, and W. A. Sadler, eds.). pp. 187195. Plenum, New York. Armstrong, D. T., Goff, A. F., and Dorrington, J . H. (1979). I n "Ovarian Follicular Development and Function" (A. R. Midgley and w. A. Sadler, eds.). pp. 169-181. Raven. New York. Ax. R. L., and Ryan, R. J. (1979a). Biol. Reprod. 20, 1123-1132. Ax. R. L., and Ryan, R. J. (1979b). J. Clin. Endocrinol. Meicrb. 49, 646-648. Ax, R. L., LaBarbera, A. R.. and Ryan, R. J. (1978). In "Ovarian Follicular and Corpus Luteum Function" (C. P. Channing. J . Marsh, and W. A. Sadler. eds.) pp. 77-82. Plenum, New York. Bae, I. H., and Foote. R. H. (1975). Exp. Ce// Res. 90, 432-436. Baker. T . G. (1979). I n "Ovarian Follicular Development and Function" (A. R. Midgley and W. A. Sadler, eds.) pp. 353-364. Raven, New York. . 419Baker, T. G., and Hunter. R. H. F. (1978). Ann. B i d . Aniin. Biochim. B i o p l i y . ~ 18, 426. Baker, T. G . , and Neal, P. (1972). In "Oogenesis" (J. D. Biggers and A. W. Schuetz. cds.). pp. 377-396. Univ. Park Press, Baltimore, Maryland. Baker, T. G., and Neal, P. (1974). J. A n d . 117, 361-371. Baker, T. G . , Hunter, R. H. F.. and Neal, P. (1975). E.vprrieniicr 31, 133-135. Balakier. H. (1978). Exp. Ce// Res. 112, 137-141.

AKKESI AND RESUMPTION OF MEIOSIS IN OOCYTES

29

Ball. C;. D.. Bellin, M. E.. Ax, K. L.. and First. N. L. (1982). Mol. C i 4 . E ~ I ~ / o u ~ M28, J/. 113-122. friri. A. (1981). Gurwrc. Res. 4, 463-472. Bar-Ami, S.. Nimrod. A.. Brodie. A. M. H.. and Tsafriri. A. (1983). J . Srcroid Eiochcwi. 19, 965-97 I . Batta. S. K.. and Knudsen. J. F. (1980). Biol. RiTprod. 22, 243-246. Berridge. M. J. (1975). Adi'. Cvclic Nircleoridi. R r s . 6, 1-98. Berridge. M. J. (1984). Biochern. J . 220, 345-360. Biggers. J . D. (1973). In "The Regulation of Mammalian Reproduction" (S. J . Segal. R. Crozier. P. A. Corfman. and P. G . Condliffe. eds.). pp. 273-283. Thomas. Springfield. Illinois. Binor. Z. and Wolf. D. P. (1979). J . Reprod. Fertil. 56, 309-314. Birchmeier. C.. Broek. C . . and Wigler, M. (19x5). Cc4/ 43, 615-621. Bjorkman. N . (1962). Acro Aniit. 51, 125-147. Bornslaeger. E. A , , and Schultz. R. M. (1985). ESP. Cell. R1J.s.156, 277-281. Brachet, J. (1977). Crfrr. 7iJp. &it.E i ~ l 11, . 133-186. Busa. W. B., Ferguson, J . E.. Joseph. S. K., Williamson, J. R . . and Nuccitelli. R. (1985). J . Ct4l Biol. 101, 677-682. Centola. G. M.. Anderson. L. D., and Channing. C . P. (1981). Gcrrne/e Rcs. 4, 451-461. Chang. M. C. (1955). J . ESP.ZOO/. 128, 378-399. Channing. C . P., and Tsafriri, A. (1977). Mc,/crho/i.srf?26, 413-468. Channing, C. P., Hillensjo. T.. and Schaerf, F. W. (1978). Clin. Endocrinol. Mcrtrh. 7, 601624. Channing. C. P.. Schaerf. F. W.. Anderson. L. D.. and Tsafriri, A. ( 1980). Inr. Reit. P l i g ~ i ~ l . 22, 117-1-01. Channing. C . P.. Anderson. L. D., Stone. S. L., and Batta, S. (1981). In "Reproductive Processes and Contraception" (K. W. McKerns. ed.). pp. 619-645. Plenum. New York. Channing. C. P., Pomerantz, S. H.. Bae. 1. H., Evans, V. W., and Atlas, S. J . (1982a). In "lntraovarian Control Mechanisms" (C. P. Channing and S. J. Segal, eds.), pp. 189-210. Plenum. New York. Channing. C . P.. Anderson, L. D., Hoover, D. J., Kolena, J.. Osteen, K. G., Pomerantz, S. H.. and Tdnahe, K. (1982b). Recent Prog. Horm. Res. 38, 331-408. Channing. C. P., Liu. C . Q., Jones, G. S., and Jones, H. (1983). Proc. Nurl. Acud. Sci. U . S . A . 80, 418441238, Chari. S., Hillensj6, T.. Magnusson, C., Sturm, G., and Daume. E. (1983). Arch. Gvnecol. 233, 155-164.

Cho, W. K., Stern. S.. and Biggers. J . D. (1974). J . Exp. Zoo/. 187, 383-386. Cross, P. C. (1973). J . Reprod. Ferril. 34, 241-245. Cross, P. C., and Brinster, K. L. (1970). Eiol. Reprod. 3, 298-307. Cross, P. C.. and Brinster, R. L. (1974). Exp. Cell Res. 86, 43-46. Cuthbertson. K. S. R., and Cobbold, P. H. (1985). Nutrrre (London) 316, 541-542. De Felici. M . . and Siracusa, G . (1982). Dev. Eiol. 92, 539-543. Dekel, N. ( 1984). In "Hormonal Control of the Hypothalamo-Pituitary-Gonadal Axis" (K. W. McKerns and N. Z. Naor, eds.), pp. 325-336. Plenum, New York. Dekel. N.. and Beers, W. H. (1978). Proc. Nut/. Acud. Sri. U.S.A. 75, 4369-4373. Dekel, N . , and Beers, W. H. (1980). Drv. Biol. 75, 247-254. Dekel. N.. and Kraicer. P. F. (1978). Endocrinology (Bnlfirnore) 102, 1797-1802. Dekel, N., and Phillips, D. M. (1979). B i d . Reprod. 21, 9-18. Dekel, N.. and Scherizly, I. (1983). FEES Lerr. 151, 153-155. Dekel, N., Hillensjo, T.. and Kraicer, P. F. (1979). Eiol. Reprod. 20, 191-197.

30

ElMEI S A T 0 AND S. S. KOIDE

Dekel, N., Lawrence, T. S., Gilula, N. B.. and Beers, W. H. (1981). Dev. Biol. 86, 356362. Dekel, N., Aberdam, E., and Sherizly, I. (1984). B i d . Reprod. 31, 244-250. Donahue, R. P. (1972). I n “Oogenesis” (J. D. Biggers and A. W. Schuetz, eds.), pp. 413438. Univ. Park Press, Baltimore, Maryland. Dorrington, J. H., and Armstrong. D. T. (1980). Recent Prog. Horm. Res. 35, 301-373. Downs, S. M., Coleman, D. L., Ward-Bailey, P. F., and Eppig, J. J. (1985). Proc. Nut/. Acad. Sci. U.S.A. 82, 454458. Edwards, R. G. (1965). Nature (London) 208, 349-351. Edwards, R. G. (1974). J . Reprod. Fertil. 37, 189-219. Eiler. H., and Nalbandov, A. V. (1977). Endocrinology (Baltimore) 100, 331-338. Ekholm, C., Hillensjo, T., and Isaksson, 0. (1981). Endocrinology (Baltimore) 108, 20222023. Ekholm. C., Hillensjo, T., Magnusson, C., and Rosberg, S. (1984). B i d . Reprod. 30, 537543. Engle. E. T. (1927). Am. J . Anat. 39, 187-203. Eppig, J. J. (1979a). Nature (London) 281, 483484. Eppig, J. J. (1979b). J . Exp. Zool. 208, 111-120. Eppig, J. J. (1980a). Biol. Reprod. 22, 629-633. Eppig, J. J. (1980b). Biol. Reprod. 23, 545-552. Eppig, J. J. (1981a). Endocrinology (Baltimore) 108, 1992-1994. Eppig, J. J. (1981b). Biol. Reprod. 25, 599-608. Eppig, J. J. (1982). Dev. Biol. 89, 268-272. Eppig. J. J., and Downs, S. M . (1984). B i d . Reprod. 30, 1-11. Eppig, J . J., and Ward-Bailey, P. F. (1984). Exp. Cell Res. 150, 459-465. Eppig, J. J., Freter, R. R., Ward-Bailey, P. F.. and Schultz, R. M. (1983). Dev. B i d . 100, 39-49. Erickson. G. F.. and Sorensen. R. A. (1974). J . Exp. Zod. 190, 123-127. Flagg-Newton, J., Simpson. I.. and Loewenstein. W. R. (1979). Science 205, 404407. Fleming, A. D., Khalil, W.. and Armstrong, D. T. (1983). J . Reprod. Fertil. 69, 665-670. Foote, W. D., and Thibault. C. (1969). Ann. B i d . Anim. Biochim. Biophys. 9, 329-349. Franchi, L. L., Mandle. A. M.. and Zuckerman, S. (1962). I n “The Ovary” (S. Zuckerman, ed.), Vol. I , pp. 1-71. Academic Press, New York. Freter, R. R., and Schultz, R. M. (1984). J. CeIlBiol. 98, 1119-1128. Fulka, J., Jr. (1983). Exp. Cell Res. 146, 212-218. Gebauer, H.. Lindner, H. R., and Amsterdam, A. (1978). B i d . Reprod. 18, 350-358. Gerard, M.. Menezo, Y.. Rombauts. P., Szollosi, D., and Thibault, C. (1979). Ann. Biol. Anim. Biochim. Biophys. 19, 1521-1535. Gill, D. M. (1982). I n “ADP-Ribosylation Reactions” (0.Hayaishi and K. Ueda, eds.). pp. 594-621. Academic Press, New York. Gilula, N. B.. Reeves, 0. R., and Steinbach, A. (1972). Nature (London) 235, 262-265. Gilula. N. B., Epstein, M. L.. and Beers, W. H. (1978). J . Cell Biol. 78, 58-75. Grimek, H. J., and Ax, R. L. (1982). Biochem. Biophys. Res. Comrnun. 104, 1401-1406. Gwatkin. R. B. L., and Andersen, 0. F. (1976). Life Sci. 19, 527-536. Hadek, R. (1963). J . Ultrastruct. Res. 9, 4 4 5 4 5 8 . Hay, M. F., and Moor, R. M. (1975). J . Reprod. Fertil. 45, 583-593. Heller, D. T., and Schultz, R. M. (1980). J . Exp. Zoo/. 214, 355-364. Heller, D. T., Cahill, D. M., and Schultz, R. M. (1981). Dev. Biol. 84, 455-464. Hillensjo. T. (1977). Acta Physiol. Scand. 100, 261-263. Hillensjo, T.. and LeMaire, W. J. (1980). Nature (London) 287, 145-146.

ARREST AND RESUMPTION OF MEIOSIS IN OOCYTES

31

Hillensjo. T., Dekel, N.. and Ahren, K. (1976). Acta Physiol. Scand. 96, 558-568. Hillensjo, T., Ekholm, C., and Ahren, K. (1978). Acta Endocrinol. (Copenhagen) 87,377388. Hillensjo, T., Channing, C. P., Pomerantz, S. H., and Schwartz-Kripner, A. (1979). In Vitro 15, 32-39. Hirai, S., Kishimoto. T., Koide, S. S., and Kanatani, H. (1984). B i d . Bull. (Woods Hole, Muss.)167, 518. Hubbard, C. J., and Terranova, P. F. (1982). B i d . Reprod. 26, 628-632. Igarashi, Y.. and Yaoi, Y. (1975). Nature (London) 254, 248-250. Iwamatsu, T., and Yanagimachi, R. (1975). J. Reprod. Fertil. 45, 83-90. Jensen, C. E., and Zachariae, F. (1958). Acta Endocrinol. (Copenhagen) 27, 356-368. Kaplan, R., Dekel, N., and Kraicer, P. F. (1978). Gamete Res. I , 59-63. Kishimoto, T., Kuriyama, R., Kondo, H., and Kanatani, H. (1982). Exp. Cell Res. 137, I 2 I - I 26. Kishimoto, T., Yamazaki, K., Kato. Y.. Koide, S. S., and Kanatani, H. (1984). J. Exp. ZOO^. 231, 293-295. Kondo. K., and Sakai, H. (1971). Dev. Growth Differ. 13, 1-14. Lacroix, E.. Eechante, W., and Leusen, I. (1974). Steroids 23, 337-356. Lieberman. M. E., Tsafriri, A., Bauminger, S.. Collins, W. P., Ahren, K., and Lindner, H. R. (1976). Acra Endocrinol. (Copenhagen) 83, 151-157. Liebfried, L., and First, N. L. (1979). J. Exp. Zool. 210, 575-580. Liebfried, L., and First, N. L. (1980a). Biol. Reprod. 23, 699-704. Liebfried, L.. and First, N. L. (1980b). B i d . Reprod. 23, 705-709. Lindner. H. R., Tsafriri, A.. Lieberman, M. E., Zor, U., Koch, Y., Bauminger. S.. and Barnea, A. (1974). Recent Prog. Norm. Res. 30,79-138. Lindner, H. R., Amsterdam, A., Salomon, Y., Tsafriri, A,, Nimrod, A., Lamprecht, S. A.. Zor, U.. and Koch, Y. (1977). J . Reprod. Fertil. 51, 215-235. Lindner, H. R., Bar-Ami, S., and Tsafriri, A. (1983). In “The Ovary” ( G . B. Serra. ed.), pp. 83-94. Raven, New York. Lindsey, A. M., and Channing. C. P. (1979). Biol. Reprod. 20, 473482. Lopata, A., Fonseca, J. R., and Szego, C. M. (1977). J. Reprod. Fertil. 50, 21 1-216. McGaughey, R. W. (1977). Endocrinology (Baltimore) 100, 3 9 4 5 . McGaughey. R. W . (1978). In “Methods in Mammalian Reproduction” (J. C. Daniel. Jr., ed.). pp. 1-19. Academic Press, New York. McGaughey, R. W. (1983). Oxford Rev. Reprod. B i d . 5, 106-130. Magnusson, C., and Hillensjo, T. (1977). J . Exp. Zool. 201, 139-147. Mangia, F.. and Canipari, R. (1977). In “Development in Mammals” (M. H. Johnson, ed.), Vol. 2, pp. 1-29. Elsevier, Amsterdam. Marsh, J. M. (1976). B i d . Reprod. 19, 3049. Marsh, J . M.. Mills, T. M., and LeMarie, W. J. (1973). Biochim. Eiophys. Actrc 304, 197202. Masui, Y., and Clarke, H. J. (1979). Int. Rev. Cytol. 57, 185-282. Miake-Lye, R.. and Kirschner, M. W. (1985). Cell41, 165-175. Moon, Y. S . , Dorrington, J. H.. and Armstrong, D. T. (1975). Endocrinologv (Baltimore) 97, 244-247. Moor, R. M. (1983). In “Current Problems in Germ Cell Differentiation” (A. McLaren and C. C. Wylie, eds.), pp. 307-326. Cambridge Univ. Press, London and New York. Moor, R. M., and Heslop, J. P. (1981). J . Exp. Zool. 216, 205-209. Moor, R. M., and Trounson, A. 0. (1977). J . Reprod. Fertil. 49, 101-109. Moor, R . M.. Smith, M. W., and Dawson, R. M. C. (1980). Exp. Cell Res. 126, 15-29.

32

ElMEl S A T 0 A N D S. S. KOIDE

Moore, G. P. M., Lintern-Moore, S., Peters, H., and Faber, M. (1974). J. Cell Biol. 60, 416-422. Moss, J., and Vaughan, M. (1979). Annu. Rev. Biochem. 48, 581-600. Motlik, J., and Fulka. J. (1976). J. Exp. Zool. 198, 155-162. Mueller. P. L., Schreiber, J. R.. Lucky, A. W., Schulman, J. D., Rodbard. D.. and Ross, G. T.(1978). Endocrinologv (Bultimore) 102, 824-831. Neal. P., and Baker, T.G. (1975). J . Endocrinol. 65, 27-32. Nekola. M. V . , and Smith, D. M. (1975). J. Exp. Zool. 194, 529-534. Nelkin, B.. Nichols, C.. and Vogelstein. B. (1980). FEES L c t f . 109, 233-238. Nilsson, L., Rosberg, S.. and Ahren, K. (1974). Arta Endocvinol. (Copenhugen) 77, 559574. Nishizuka, Y. (1984). Nuture (London) 308, 693-698. Odor, L. D. (1960). J. Biophvs. Biochem. Cvtol. 7, 567-574. Okamoto. H . , Takahashi. K., and Yamashita. N. (1977). J . Physiol. (London) 267,465495. Olsiewski, P. J . , and Beers, W. H. (1983). Dev. Biol. 100, 287-293. Paleos. G . A..and Powers, R. D. (1981). J . Exp. Zoo/. 217, 409-416. Pincus, G., and Enzmann, E. V. (1935). J . Exp. Med. 62, 665-675. Pincus, G.. and Enzmann. E. V. (1937). J. Morphol. 61, 351-384. Powers, R. D., and Paleos, G. A. (1982). J. Reprod. Fertil. 66, 1-8. Racowsky, C. (1983). B i d . Reprod. 28, 774-787. Racowsky, C.. and McGaughey, R. W. (1982). J. Reprod. Fertil. 66, 505-512. Richter, J . D., and McGaughey. R. W. (1979). J. Exp. Zool. 209, 81-90. Richter. J. D., and McGaughey, R. W. (1981). DPV.Biol. 83, 188-192. Sadler, S. E., and Maller, J. L. (1985). Adv. Cyclic Nucleotide Protein Phosphorylotion Res. 19, 179-194. Salustri, A., and Siracusa, G. (1983). J. Reprod. Ferril. 68, 335-341. Sato, E.. and Ishibashi, T.(1977). J p n . J . Zootech. Sci. 48, 22-26. Sato. E., and Koide, S. S. (1984a). J . Exp. Zool. 230, 125-129. Sato, E.. and Koide, S. S. (1984b). Dgferenfiution (Berlin) 26, 59-62. Sato, E., Iritani, A.. and Nishikawa, Y. (1977). Kuchiku Hunshokugukrr Zasshi 23, 12-18. Sato, E., Iritani, A..and Nishikawa, Y. (1978a). Nippon Chikusan Gukkui H o 49, 236-242. Sato, E., Iritani, A., and Nishikawa, Y. (1978b). Nippon Chikusan Gakkui H o 49,400405. Sato, E., Iritani, A., and Nishikawa, Y. (1980). Nippon Funin Gakkai Zasshi 25, 229-233. Sato, E.. Ishibashi. T.. and Iritani, A. (1982). I n “lntraovarian Control Mechanisms” (C. P. Channing and S. J. Segal. eds.), pp. 161-173. Plenum, New York. Sato, E., Segal. S. J.. and Koide. S. S. (1983). Gamete Res. 8, 119-127. Sato, E.. Segal. S. J . , and Koide, S. S. (1984a). Biol. Bull. (Woods Hole, M a s s . ) 167, 520. Sato, E., Ueno. H.,and Koide, S. S. (1984b). J . Steroid Biochem. 20, 15%. Sato. E.. Wood. H. N.,Lynn, D. G.,and Koide, S. S. (1985). Cell Differ. 17, 169-174. Sato, E., Ueno, H., and Koide, S. S. (1986). Gumete Res. 13, 115-124. Schaerf, F. W.. Anderson, L. D., and Channing. C. P. (1982). Gumete Res. 5, 207-215. Schuetz, A. W. (1974). B i d . Reprod. 10, 150-178. Schuetz, A. W. (1978). In “Ovarian Follicular and Corpus Luteum Function” (C. P. Channing. J. Marsh, and W. A. Sadler, eds.). pp. 307-314. Plenum, New York. Schuetz. A. W., and Swartz. W. J. (1979). J. Exp. Zool. 207, 399406. Schultz, R. M., Montgomery, R. R., and Belanoff, J. R. (1983a).Dev. B i d . 97, 264-273. Schultz, R. M., Montgomery, R. R., Ward-Bailey, P. F., and Eppig. J. J. (1983b). DeIB. Biol. 95, 294-304. Schweitzer, M., Jackson, J. C.. and Ryan, R. J. (1981). B i d . Reprod. 24, 332-340. Sorensen. R. A.. and Wassarman, P. M. (1976). Dev. B i d . 50, 531-536.

ARREST AND RESUMPTION OF MEIOSIS IN OOCYTES

33

Staigmiller, R. B.. and Moor. R. M. (1984). Gamefe Res. 9, 221-229. Stern. S., and Wassarman. P. M. (1974). J. Exp. Zoo/. 189, 275-281. Stith, B. J., and Maller, J. L. (1985). J . C d Biol. 101, 4Wa. Stone. S. L.. Pomerantz, S. H.. Schwartz-Kripner, A.. and Channing, C. P. (1978). B i d . R i p w d . 19, 585-592. Sunkard. P. S., Wright, D. A., and Rao. P. N. (1979). Proc. Nrifl. A c i d . Sci. U.S.A. 76, 2799-2802.

Szollosi. D., Gerard. M., Mknezo. Y.. andThibault, C. (1978). Ann. B i d . Biochim. Biophys. 18, 511-521.

Syzbek. K. (1972). J. Endocrinol. 54, 527-528. Thibault, C. (1977). J. Ri.prod. Ferfil. 51, I-IS. Thibault. C.. Gerard. M.. and Menezo. Y. (1975). Anti. B i d . Anim. Bioclzirn. Biopliys. 15, 705-714.

Tsafriri. A. ( 1978a). Ann. Bird. Atibn. Biochirn. Biopliys. 18, 523-528. Tsafriri. A. (1978b). In “Ovarian Follicular and Corpus Luteum Function” (C. P. Channing. J. Marsh, and W. A. Sadler. eds.). pp. 269-281. Plenum, New York. Tsafriri. A. (1978~).I n “The Vertebrate Ovary” (R. E. Jones. ed.), pp. 409442. Plenum, New York. Tsafriri. A. (1984). Biol. Ferr. I , 221-252. Tsafriri. A., and Bar-Ami, S. (1978). J. Exp. Z o o / . 205, 293-300. Tsafriri, A.. and Bar-Ami, S. (1982). In “lntraovarian Control Mechanisms” (C. P. Channing and S. J. Segal, eds.), pp. 145-159. Plenum, New York. Tsafriri. A., and Channing, C. P. ( 1975a). Endocrinology (Brtltitnore) 96, 922-927. Tsafriri. A., and Channing, C. P. (1975b). J. Reprod. F u f i l . 43, 149-152. Tsafriri, A.. Lindner, H. R.. Zor, U.. and Lamprecht, S. A. (1972). J. Reprod. Fiwil. 31, 39-50.

Tsafriri, A.. Pomerantz. S . H., and Channing, C. P. (1976). B i d . Reprod. 14, 51 I-Sl6. Tsafriri. A., Channing. C. P., POmerAntz. S . H., and Lindner, H. R. (1977). J . Endocrinol. 75, 285-29 I.

Tsafriri. A.. Bar-Ami. S.. and Dekel. N . (1982a). In “Non-Steroidal Regulators in Reproductive Biology and Medicine” (T. Fuji and C. P. Channing, eds.), pp. 147-156. Pergamon. Oxford. Tsafriri. A.. Dekel, N.. and Bar-Ami, S. (1982b). J. Reprod. Ferfil. 64, 541-551. Umer, F.. and Schoderet-Slatkine, S. (1984). Exp. Cell Res. 154, 600-605. Urner. F., Herrmann, W. L., Baulieu, E. E., and Schorderet-Slatkine. S. (1983). Etidocrit i o / o g v (Biilfirnorr) 113, 1170-1 172. Vermeiden. J. P. W., and Zeilmaker. G. H. (1974). Endocrinologv (Brilfimore)95, 341-351. Vivarelli. E., Conti. M., De Felici, M.. and Siracusa. G. (1983). Ce// Diffrr. 12, 271-276. Wassarman, P. M.. and Letourneau, G. E. (1976). Nuti4re (London) 261, 73-74. Wassarman. P. M., Schultz, R. M., Letourneau, G . E., LaMarca, M. J . , Josefowicz. W. J.. and Bleil, J . D. (1978). In “Ovarian Follicular and Corpus Luteum Function” (C. P. Channing. J. Marsh, and W. A. Sadler. eds.), pp. 251-267. Plenum, New York. Yanagishita. M.. and Hascall. V. C. (1979). J. Biol. Chrm. 254, 12355-12364. Yanagishita. M., Rodbard. D., and Hascall, V. C. (1979). J. B i d . Chctn. 254, 91 1-920. Zachariae. F. ( 1957). Acfu Endocrinol. (Copvr/irigetiJ26, 215-233. Zdmboni. 1.. (1974). B i d . Rcprod. 10, 125-149.

This Page Intentionally Left Blank

INTEKNATIONAL REVIEW OF CYTOLOGY, VOL. lob

Morphology and Cytochemistry of the Endocrine Epithelial System in the Lung D.W.SCHEUEKMANN Institrrtc>of Histology and Micwscwpic Anritomy, University of’Antwerp, 2020 Antwerp, Belgium

I. Introduction In his light microscopic study of 1938, Feyrter first described what he called “Helle Zellen” (clear cells), due to their almost transparent cytoplasm in hematoxylin- and eosin-stained sections, lying dispersed throughout the epithelial tissue of various organs. He assumed them to be endocrine-like cells belonging to a widespread endocrine epithelial system with paracrine, neurocrine, and hemocrine functions, located in various organs including the respiratory system. Later, in an extensive study, Frohlich (1949) provided a precise description of these clear cells in the epithelium of the tracheobronchial tract of several mammals (rabbit, cat, guinea pig, dog, wether, and monkey) including man (executed persons) by means of different conventional staining methods. Strikingly, in mammals, these clear cells-particularly those in man-were found to occur not only as solitary elements; indeed, Frohlich also outlined and illustrated the existence of distinctive groups of clear cells forming round or oval corpuscular structures. According to the same author, these clear cells are situated in the epithelial tissue, close to the basement membrane, the apical cytoplasm contacting the airway lumen only occasionally. He attributed to these cells a neurosensory function and realized a relationship with the dispersed endocrine epithelial system originally outlined by Feyrter ( 1938). Moreover, Frohlich provided the first demonstration of nerve endings in close contact with the basal cytoplasm of both solitary and groups of pulmonary clear cells, an observation later confirmed by several authors (e.g., Glorieux, 1963; Shul’ga, 1965; Lauweryns ef al., 1972, 1974; Hung et al., 1973; Lauweryns and Cokelaere, 1973b; Hung and Loosli, 1974; Lauweryns and Goddeeris, 1975; Wasano, 1977; Hung, 1980; Scheuermann et al., 1983a,b; Scheuermann, 1984; Stahlman and Gray, 1984). The early investigators demonstrated the reactivity of pulmonary clear cells to argentaffin and/or argyrophilic silver techniques and argued that they might possibly have a chemoreceptive and neurosecretory function, acting primarily at the pulmonary level (Frohlich, 1949; Feyrter, 1953, 35 Copyright l a 19x7 by Academic Pre% Inc. All rightr of reproduclion in any Iorm r r w v e d .

36

D. W. SCHEUERMANN

1954, 1958). Since these first histological observations, the presence of endocrine-like clear cells among seemingly darker epithelial cells has been revealed by both light and electron microscopy in the extra- and intrapulmonary airways of several mammals, including man (Glorieux, 1963; Bensch et al., 1965; Lauweryns and Peuskens, 1969; Lauweryns et al., 1970, 1972; Cutz and Conen, 1972; Ericson et al., 1972; Hage, 1972, 1973a,b, 1974, 1980; Hung et a l . , 1973, 1979; Cutz et al., 1974, 1975, 1978a,b; Hung and Loosli, 1974; Jeffery and Reid, 1975; Lauweryns and Goddeeris, 1975; Hung, 1976, 1980; Hage et al., 1977; Hernandez-Vasquez et al., 1977, 1978a,b; Wasano, 1977; Sorokin and Hoyt, 1978; Edmondson and Lewis, 1980; Palisano and Kleinerman, 1980; Dey et al., 1981, 1983; Keith et al., 1981, 1982; Wasano and Yamamoto, 1981; Carabba et al., 1982; Sarikas et al., 1982; Pearsall et al., 1985), birds (Cook and King, 1969; Walsh and McLelland, 1974; Wasano and Yamamoto, 1979), amphibians (Rogers and Haller, 1978, 1980; Wasano and Yamamoto, 1978; Goniakowska-Witalinska, 1980a,b, 1981), and a reptile (Scheuermann et al., 1983a,b). These endocrine-like cells occur isolated or in distinctive groups of two or three cells, as well as in large clusters of more than 100 cells (Hoyt er al., 1982a,b) within the epithelium at every level of the bronchoalveolar tract. Combined histochemical, fluorescence microscopic, and ultrastructural investigations have shown the clear cells of the pulmonary epithelium to contain intracytoplasmic chemical mediators, such as 5-hydroxytryptamin (5-HT) (Lauweryns et al., 1974, 1982; Rogers and Haller, 1978; Keith et al., 1982; Scheuermann et al., 1983a) and neuropeptide hormones (bombesin, Wharton et al., 1978; Dayer et a l . , 1985; gastrin-releasing peptide; Iwanaga, 1983; Tsutsumi et al., 1983a,b; calcitonin, Becker er al., 1980; leu-enkephalin, Cutz et al., 1981), thereby displaying in many aspects a similarity to the elements of the APUD (amine precursor uptake and decarboxylation) endocrine system conceived by Pearse (1969, 1977). The structural resemblance of these cells to some known receptor cells (Cook and King, 1969; Lauweryns et al., 1972; Lauweryns and Peuskens, 1972; Hung et d., 1973; Wasano, 1977; Cutz et al., 1978a,b; Scheuermann et al., 1983a) has led to their classification in the paraneuronic system of Fujita (1977), indicating their functional relation to neurons. In the course of recent years, a plethora of names has been coined on the basis of morphological and presumed functional features to designate these cells. For instance, they have been referred to as Feyrter cells (Moosavi et al., 1973; Hernandez-Vasquez et al., 1977, 1978a,b; Taylor, 1982) after the pathologist who first described them. But they were also called enterochromaffin-like cells (Ericson et al., 1972) and Kultschitzkylike cells (or K cells) (Bensch et al., 1965; Cutz et al., 1974, 1975). since

T H E ENDOCRINE EPITHELIAL SYSTEM IN THE LUNG

37

they display many features similar to those found in their counterparts in the gastrointestinal tract (Bensch et al., 1965; Cutz and Conen, 1970, 1972; Lauweryns et al.. 1970; Terzakis et al., 1972; Jeffery and Reid, 1975; Breeze and Wheeldon, 1977; Capella et al., 1978; Sorokin et ul., 1983). Another term suggested was argyrophil cells (Lauweryns and Peuskens, 1969; Taylor, 1977) or even AFG k e . , the initial letters of argyrophilic, fluorescent, and granulated) cells (Lauweryns et al., 1970). Other terms which have been proposed are chromafln-type cells (Basset et al., 1971), endocrine cells (Hage, 1972, 1973a,c, 1974), endocrine-like cells (Hage, 1976; Cutz and Conen, 1972; Ewen et al., 1972; Hage et al., 1977), neurosecretory cells (Becci et al., 1978) or neurosecretory-appearing cells (Terzakis et al., 1972). small granulated cells (McDowell et al., 1976), small granule endocrine cells (Sorokin and Hoyt, 1978), biogenic aminecontaining cells (Eaton and Fedde, 1978), or neuroendocrine cells (Keith et al., 1981). Their amino acid uptake characteristics have inspired some authors to call them APUD cells (Hage, 1973a; Sidhu, 1979), since they fulfill the principal criterium of this system (Pearse, 1969, 1977). However, except for the Feyrter and Kultschitzky terms, none of these names specify whether they are situated intraepithelially or in the pulmonary connective tissue. Since it was demonstrated that both solitary and groups of neuroendocrine cells containing biogenic amines may belong to intrapulmonary ganglia (McLean and Burnstock, 1967a,b; Bliimcke, 1968; Bock, 1970; Mann, 1971; Jacobowitz et al., 1973; Knight, 1980; Scheuermann and De Groodt-Lasseel, 1983; Scheuermann et al., 1983b, 1984a,b,c, 198% two populations of endocrine cells of the respiratory system should be included in the APUD series. Accordingly, it seems necessary to maintain a terminology which refers to the characteristic location of the cells in the epithelial tissue. This shortcoming applies equally to the term dense-core granulated cells introduced by Jeffery and Corrin (1984) in a recent study on the structural analysis of the respiratory tract. Indeed, the intraepithelial, granule-containing cells and the small intensely fluorescent cells occurring in the ganglia of the pulmonary interstitium are characterized by an abundance of dense-cored vesicles and are both assumed to have a chemoreceptor and endocrine or paracrine function (Bock, 1970; Knight, 1980; Scheuermann et al., 1983b, 1984a,b,c). Since the present work deals with neuroendocrine epithelial cells-and not with endocrine cells of the intrapulmonary ganglia-it seems justified to use the term neuroepithelial endocrine (NEE) cells. The well-demarcated epithelial organoid structures, composed of aggregated NEE cells-for the first time extensively described by Glorieux (1963), who called them “corpuscule kpithelial” (i.e., epithelial corpuscle)-will be referred to in the course of this review as neuroepithelial

38

D.

W.

SCHEUERMANN

bodies (NEBs), a term proposed by Lauweryns et al., (1972) because of their prominent nerve supply. The endocrine clear cells grouped in NEBs have been compared to the solitary endocrine cells, lying dispersed in the epithelium of the respiratory system and described in human fetuses (Cutz and Conen, 1970, 1972; Hage, 1973a,b; Cutz et al., 1973, in children (Lauweryns and Peuskens, 1969; Cutz and Conen, 1970; Lauweryns et al., 1970; Rosan and Lauweryns, 1971; Cutz et al., 1975; Lauweryns and Goddeeris, 1975), in the adult human lung (Cutz and Conen, 1970; Terzakis et al., 1972; Lauweryns and Goddeeris, 1975; Hage et al., 1977) as well as in various mammalian species (Jeffery and Reid, 1973, 1975; Cutz et al., 1974; King et al., 1974; Hernandez-Vasquez et al., 1977; Sorokin and Hoyt, 1978; Edmondson and Lewis, 1980; Palisano and Kleinerman, 1980; Lauweryns et al., 1985), in birds (Cook and King, 1969), and in amphibians (Wasano and Yarnamoto, 1978; Goniakowska-Witalinska, 1980a). The corpuscular appearance of groups of NEE cells in NEBs and their conspicuous innervation are considered by some authors to be a separate neuroendocrine cell system, distinct from solitary NEE cells (Lauweryns and Cokelaere, 1973a; Lauweryns et d., 1974, 1978, 1985; Lauweryns and Goddeeris, 1975; Lauweryns and Liebens, 1977; Loosli and Hung, 1977; Hung et d., 1979; Sonstegard et al., 1979; Foliguet and Cordonnier, 1981). However, in developing rabbit lungs, it was demonstrated by electron microscopy (Sorokin et al., 1982) that scattered solitary clear cells appear very early during gestation as the first population in the pulmonary epithelium undergoing differentiation into NEE cells and that mature NEBs are derived from them. According to these authors, embryonal undifferentiated precursor cells, situated in the primary pulmonary epithelium, are observed in various stages of transformation to NEE cells, which subsequently appear in groups of two to three cells that will finally mature, at least partly, to NEBs. Similarly, Stahlman and Gray (l984), investigating electron microscopic preparations of the fetal human lung, describe putative neuroendocrine cells which, during development, differentiate into either singly occurring neuroendocrine cells or into NEBs. In a combined immunohistochemical and ultrastructural investigation of the development of NEE cells in the human lung from the early fetal to the perinatal period, Cutz et al. (1984) demonstrated the presence, in the canalicular stage, of NEE cells occurring either solitarily or packed in NEBs. As observed by these authors, the NEE cells grouped in NEBs share identical ultrastructural features with the isolated NEE cells. Both light and electron microscopic investigations seem to indicate that single NEE cells and NEBs represent stages of differentiation of one and the same embryonal precursor cell. In agreement with these findings, studies have shown that in the res-

T H E ENDOCRINE EPITHELIAL SYSTEM IN T H E LUNG

39

piratory system both the diffuse endocrine cells and the NEBs, which may be involved in the production of amines and/or polypeptide hormones, share similar histochemical properties (Cutz et al., 1984). Moreover, the innervation of NEE cells is not restricted to NEBs. Nerve terminals with synaptic contact have been described at the base of single NEE cells in the infant bronchial epithelium (Lauweryns et al., 1970). Unmyelinated axons were also found in close association with individual NEE cells of the pulmonary system in the rabbit (Hung, 1980), hamster (Edmondson and Lewis, 1980), rat (Jeffery and Reid, 1973), and guinea pig (DiAugustine et al., 1984). Since functions of these endocrine cells remain still unknown, the pulmonary single NEE cells and those arranged in NEBs will be treated as a single neuroepithelial endocrine system. 11. Light Microscopic Aspects

Epithelial cells with transparent cytoplasm are observed in routine light microscopic preparations of the entire respiratory tract, in particular, when using, after fixation with formaldehyde, hematoxylin-eosin, the trichrome method of Masson, or a modification of the Goldner-Masson staining method (Frohlich, 1949; Feyrter, 1958). However, since the translucent appearance is not a specific morphological feature, the solitary NEE cells remain in this way relatively inconspicuous. Conversely, NEBs can be readily detected by conventional staining techniques, forming clearly demarcated epithelial corpuscles. In some species, they protrude slightly into the airway lumen (Hung et al., 1973, 1979; Cutz et al., 1978b), but in others they are enveloped in the epithelium, indenting the underlying connective tissue (Pearsall et al., 1985). They can also be observed in a pitlike depression of the pulmonary epithelium. In most species, the more centrally located cells of the NEBs are characterized by a well-ordered appearance consisting of nonciliated cells, joined side-by-side, with their longitudinal axes more or less at right angles to the basal lamina, albeit slightly inclined to the center of the luminal surface. The shape of these corpuscular cells is almost columnar with a more or less oval nucleus. Sometimes, at the branching region of the airways, the NEE cells appear stratified or form ajumble of cells, often with a pyramidal form, the apex of which is directed to the airway lumen and the broad face situated against the basement membrane. Some profiles of NEBs do not contact the lumen of the airways, but seem isolated from the surface by dark nonciliated cells stretched over the luminal and lateral cytoplasm as it expands (Fig. 1). They are assumed to be modified Clara cells (Cutz et ul., 1978b; Hung et al., 1979; Pearsall e f al., 1985). From serially cut sections, it is apparent that, in most animal species,

FIG. I . In some sections, the NEB of the red-eared turtle appears elongated, with granulecontaining cells in a palissade-like row lying on the basal lamina. Most of the apical surfaces are covered with flattened Clara-like cells. The NEB is separated from capillaries by a narrow subendothelial space containing collagen fibers. To the left, Clara-like cells abut on ciliated epithelial cells. Silver method applied to semithin section of Epon-embedded material according to Lopez e/ d.(1983). Light microscopy. x 1100. FIG.2. NEB in the epithelial lining of the bronchiolus of a neonatal rabbit composed of yellow fluorescent, elongated cells, as revealed by formaldehyde-induced fluorescence. The contours at the base of the individual cells are hardly visible, because of their close apposition and very intense fluorescence. The emission and excitation spectra of the fluorophore of this NEB is rendered in Fig. 5 . Fluorescence microscopy. x 800. FIG.3. The same NEB of the neonatal rabbit as in Fig. 2, revealed by the argyrophilic method according to Grimelius. Light microscopy. x 800. FIG. 4. Whole-mount stretch preparation of the red-eared turtle lung treated for formaldehyde-induced fluorescence. An extensive group of intensely yellow-fluorescent neuroendocrine epithelial cells, with, in their neighborhood, a few solitary and grouped neuroendocrine epithelial cells. Green-fluorescent nerve fibers running to the yellow-fluorescent NEB. Fluorescence microscopy. x 90.

THE ENDOCRINE EPITHELIAL SYSTEM IN THE LUNG

41

a small portion of the apical cytoplasm often reaches the airway lumen (Fig. I). In most species, there is a striking similarity in NEB architecture and morphology; their size is highly variable (for review, see Foliguet and Cordonnier, 1981). However, in the toad lung, a NEB appears covered by a dark apical cell, provided with a single cilium protruding into the airway lumen (Rogers and Haller, 1978, 1980). The basement membrane rests on an usually thin lamina propria, which envelops one or more capillaries close to the NEE cells (e.g., Lauweryns and Goddeeris, 1975; Scheuermann er al., 1983a; Hung, 1984; Pearsall et ul., 1985). Fascicles of smooth muscle may closely approach the base of the NEE cells (Pearsall er al., 1985). In order to detect the NEE cells, besides conventional light microscopy, various convenient staining methods have been used, including the use of masked metachromasia (Solcia et ul., 1968) and such methods as lead hematoxylin (Solcia et ul., 1969) or periodic acid-Schiff and lead hematoxylin (Sorokin and Hoyt, 1978). Acid hydrolysis which precedes staining enhances the metachromasia to basic dyes of secretory granules in endocrine cells attributed to sidechain carboxyl or carboxamide groups of granule proteins (Pearse, 1969), whereas the diffuse basophilia, due to RNA, DNA, and acid polysaccharides, is not realized by extraction of these acid substances (Solcia et ul., 1968). A consecutive treatment with basic dyes after HCI hydrolysis stained the endocrine cells in pancreatic islets, in thyroid and parathyroid endocrine cells, in the gastroenteric endocrine cells, and in the adenohypophysis (Solcia et ul., 1968). It was believed that probably all cells from the APUD series contain metachromatic substances in their secretory granules in a “masked” form which can be unmasked by HCI hydrolysis (Bussolati et d.,1969; Fujita and Kobayashi, 1974). This technique was therefore applied to demonstrate the presence of endocrine cells in the lung. According to Hage ( 1972, 1976, 1980), HCI-toluidine blue-positive cells are distributed throughout the pulmonary epithelium of human fetuses, whereas in the human adult lung, these have not been found (Hage et d.,1977; Hage, 1980), nor, for that matter, in the rabbit, guinea pig, mouse (Hage, 19741, and rat (Cutz et al., 1974). The lead hematoxylin method is frequently used in light microscopy for staining secretory granules in endocrine cells known to produce polypeptide hormones (Solcia et al.. 1969). Some authors report a positive lead hematoxylin reaction in NEE cells of the lung of human fetuses (Hage, 1980) and in carcinoid lung tumors (Hage, 1976), but not in the normal adult human lung (Hage er al.. 1977). Lauweryns and Cokelaere (1973a) demonstrated weak lead hematoxylin-positive NEBS in the lung of neonatal animals (rabbit and mouse). However, Sorokin and Hoyt (1978) furnished

42

I

D. W. SCHEUERMANN

evidence that, in lungs from young mice, rats, hamsters, kittens, as well as in late fetal and neonatal rabbits, lead hematoxylin alone produces little if any staining of the NEE cells. Conversely, a staining method which combines the periodic acid-Schiff (PAS) method to lead hematoxylin and which is applied to plastic-embedded sections seems particularly useful for granule-containing cell populations in the lung (Sorokin and Hoyt, 1978). The cells are recognizable by the magenta coloration of the cytoplasm, frequently heavier toward the cell base. In plastic-embedded material of control lungs and S h y droxytryptophan pretreated animals examined for the formaldehyde-induced fluorescence (FIF) technique and consecutively stained with the PAS-lead hematoxylin method, the NEE cells revealed a magenta staining corresponding precisely to sites of 5-HT fluorescence (Sorokin and Hoyt, 1978). In a systematic study on the infracardiac lobe of the hamster lung, five different types of NEE cells have been identified by the PAS-lead hematoxylin method (Hoyt et al., 1982a). Types I, 11, and V bear granules of about 0.2 pm in diameter, whereas types 111 and IV contain larger granules. Types I and I1 can be readily segregated, since only the granules of the first type stain reddish-pink without affinity for lead hematoxylin. Of the coarse-grained PAS-positive types 111 and IV, only the latter show affinity for lead hematoxylin. Type V cells appear reddish-purple with PAS and the small granules of these cells are stained by lead hematoxylin. Whereas types I, 11, 111, and 1V encompass columnar cells whose granules tend to accumulate at the basal pole of the cell, type V cells are apparent by their spheroidal shape. Types I, 11,111, and V may occur both solitarily and in organized clusters, whereas it was demonstrated that the PASpositive and lead hematoxylin-positive coarse-grained type IV cells, which never occur as single cells, are usually present in large NEBS. These structural differences of NEE cell subtypes are signifcant in view of the preferential relationship which each of them displays with nonendocrine cells and tissues, whether occurring solitarily or in organized clusters. For instance, NEE cell types I and I11 can be found in relation to the capillary network of the pulmonary artery; types 11, IV, and V may be linked with the capillaries of the pulmonary circuit; types I1 and IV can be related predominantly to smooth muscle cells. From these data, it might be concluded that, at least in the hamster lung, the NEE cell subtypes may have different functions, independent of their single or clustered appearance. The neuronal-like cytochemical features of NEE cells, which allow their detection in light microscopy, include silver techniques, cholinesteraseand neuron-specific enolase cytochemistry, FIF, and immunocytochemical detection of 5-HT and peptide hormones.

THE ENDOCRINE EPITHELIAL SYSTEM IN THE L U N G

43

111. Argentaffinity and Argyrophilia

The argentaffin method to reveal silver-reducing cells was first developed by Masson (l914), demonstrating that after formalin fixation the Kultschitzky cells contain an endogenous substance which reduces alkaline solutions of silver salts, resulting in these cells being impregnated by metallic silver; ever since, the term argentaffinity points to impregnation of the cells by silver after reduction of an alkaline silver solution without additional treatment with an extraneous reducing agent. After it was demonstrated that Kultschitzky cells contain 5-HT (Erspamer and Asero, 1952), Barter and Pearse (1953, 1955) showed that Kultschitzky cells in formalin-fixed tissue and synthetic 5-HT models reacted in a similar way with the usual argentaffin method, demonstrating that the silver-reducing power of these cells after formalin fixation is due to 5-HT. It was shown that, after subsequent treatment with alkaline or neutral solutions of silver salts and a weak extraneous reducing agent, not only were the argentaffin cells stained but also a wide range of other cells and tissue components (Hamperl, 1932). This other way to reveal tissue structures with silver salts is usually called the argyrophilic method, which is most frequently used for the demonstration of bronchopulmonary NEE cells; however, the NEE cells of the respiratory tract, being the pulmonary counterparts of the Kultschitzky cells, is the case in which it seems more obvious to apply the argentaffin reaction. Cytoplasmic argyrophilia was revealed in some clearly isolated NEE cells and in NEBS of paraffin-processed lung tissues from human fetuses and children (Hamperl, 1952; Feyrter, 1954, 1958; Hage, 1972, 1973b, 1976; Lauweryns and Peuskens, 1972; Cutz et a/., 1975; Lauweryns and Goddeeris, 1975), of adult human lungs (Tateishi, 1973; Lauweryns and Goddeeris, 1975; Hage et al., 1977; Hage, 1980), of rabbits, rats, sheep fetuses, and neonatal rats (Lauweryns et al., 1972, 1973, 1974; Moosavi et al., 1973; Hage. 1974, 1976; Cutz et al., 1975; Hung, 1980; Sorokin et al., 1982), of adult rabbits, dogs, and cats (Frohlich, 1949; Lauweryns et al., 1973; Taylor, 1977), and of a young armadillo (Cutz et al., 1975). This reaction yields fine, dark-brown argyrophilic granules either diffusely scattered throughout the cell or concentrated in its basal portion (Fig. 3). In this way, the solitary, pulmonary NEE cells of most species usually appear fusiform, pyramidal, or flask shaped, resting on the basement membrane, with a dark-stained, narrow, cytoplasmic, apical process pointed toward the airway lumen. In adult man, most NEE cells, observed only by Tateishi (1973) as extending to the airway lumen, lack luminal contacts. Some cells revealed dark, long, basolateral cytoplasmic processes extending along the basement membrane or between other epithelial

44

D. W. SCHEUERMANN

cells. The NEBs exhibiting a strong and massive argyrophilia can be readily identified at the light microscopic level throughout the entire respiratory system. The argentaffin reaction on paraffin sections of formalin-fixed material in Bouin’s solution and immersed in Fontana’s ammoniacal silver nitrate solution (often with omission of gold toning and counterstaining) revealed, in fetal and neonatal rabbit NEE cells (Lauweryns et al., 1972, 1973; Sonstegard et al., 1982) and in NEBs of the turtle lung (Scheuerrnann et al., 1983a), dark-brown deposits in the cytoplasm, preferentially in the basal portion of the cell (Fig. 8). Consequently, these NEE cells possess silverreducing properties after formalin fixation. Nevertheless, several investigators observed little if any argentaffin reaction in NEE cells within the wall of the intrapulmonary airways of man and different animals (Lauweryns and Peuskens, 1969; Hage, 1972, 1974, 1976, 1980, 1984; Cutz et al., 1975). In the lung of Polypterus, NEE cells appeared to reduce the ammoniacal silver salt, some of them strongly, while in others this was only barely visible; hence, they may be assumed to be argentaffin. As shown by consecutive serial sections, the argyrophilic technique applied to the same cells impregnated them with silver as well (D. W. Scheuerrnann and M. H. A. De Groodt-Lasseel, unpublished work). Since the argyrophilic reaction seems more likely to occur than the argentaffin reaction, the former will also be positive whenever argentaffin cells are demonstrated. It therefore appears that argentaffin NEE cells are argyrophilic, whereas some argyrophilic cells do not seem to be argentah. This finding is consistent with reports by Tateishi (1973) and Hage (1980), who demonstrated several argyrophilic NEE cells in the human adult lung to be nonargentaffin. Hage (1980, 1984) attributed a negative argentaffin reaction to the very low concentration of the biogenic amine, as shown by the FIF technique. In this interpretation, it is not necessary to consider argentaffin and argyrophilic cells as distinct cell types. In fact, they have been thought of as representing different stages in the secretory cycle of a single cell (Hamperl, 1952), a hypothesis supported by electron microscopic investigations (Ratzenhofer and Leb, 1965; Ratzenhofer, 1966a.b) and by cytochemical studies on 5-HT fluorescence and argyrophilia (Penttila, 1966, 1967). It is known that the argentaffin silver technique requires a critical level of silver reduction, i.e., a minimal cytoplasm-reducing ability to effectively demonstrate amine-containing cells by light microscopy. As we have seen, the argentaffinity reflects the endogenous capability of the formalin-fixed cytoplasm to reduce silver salts; therefore, this reaction has a histochemical value. Although the argyrophilic reaction is more likely to occur than the argentah reaction, it will obviously become

T H E ENDOCRINE EPITHELIAL SYSTEM I N THE L U N G

45

positive whenever the argentaffin reaction is demonstrated. With the argyrophilic technique, however, the silver ions, which are mainly caused by addition of an extra reducer, deposit metallic silver on different tissue constituents. This staining characteristic mainly reflects the physical properties and not a chemical composition; it is therefore histochemically unspecific. As a result, the argyrophilic cells are by far more numerous than the argentaffin cells, but they are also much less precisely defined. In summary, a positive argentaffin reaction is indicative of the presence of reducing substances, whereas a negative reaction could imply that the reducing capacity is minimal. The situation is quite different for the argyrophilic reaction caused by treatment with an extraneous reducing agent. Although it is believed that here too the reducing substance of the cytoplasm first partly reduces the silver salt, whereafter additional silver is deposited on top, the latter deposit also occurs on a range of other tissue constituents that seem unrelated to the argentaffin components: it is not yet clear how and why this process occurs. It might be explained by the fact that 5-HT is closely linked to other nonamine components, which may interfere with its reactivity. Hence, as far as the NEE cells of the respiratory system are concerned, it is not surprising that different authors argue that the argyrophilic reaction is not entirely reliable (Cutz et af.. 1974; Sorokin et uf., 1982), probably due to species differences or to the age of the animals, as well as the development and maturation of the endocrine cells. Certainly these possibilities and restrictions should be taken into account when dealing with a comparative distribution of argentaffin and argyrophilic cells. Both silver methods can be applied to semithin sections of Aralditeembedded material according to the method described by Lopez ut af. (1983). Hence, the solitary NEE cells as well as the NEBs are readily identified in the lung of the adult monkey, pig, and red-eared turtle throughout the length of the airways by their low cytoplasmic density. The silver-impregnated granules are most numerous in the cytoplasm facing the basal lamina (Fig. 1). IV. Cholinesterase Activity

Cholinesterase activity was observed by some authors throughout the cytoplasm of NEE cells in the fetal lung of the rabbit (Lauweryns and Cokelaere, 1973a; Hung, 1980. 1984; Sonstegard et af., 1982)and rat (Morikawa rt d.,197th). Moreover, it has been demonstrated that NEBs of the embryonic rat lung, differentiating in virro and segregated from the central nervous system, revealed acetylcholinesterase-containing granules

46

D. W. SCHEUERMANN

(Morikawa et al., 1978b). These observations are in agreement with the original view of Pearse (1969) that high levels of cholinesterase may characterize the cells of the APUD series. Indeed, monoamine-containing cells have been reported to contain cholinesterase activity, such as the chief cells of the carotid body (Koelle, 1950, 1951; Ballard and Jones, 1971; Korkala and Waris, I977), the cardiac (aorticopulmonary) glomus bodies (Papka, 1975, 1980), the catecholamine-containing cells in the pulmonary ganglia of the calf, goat, and fetal sheep (Mann, 1971), in the pelvic paraganglia (Thompson and Gosling, 1976), in adrenomedullary cells (Palkama, 1967), and in some adrenergic neurons (Jacobowitz and Koelle, 1965). As yet, the functional significance of this enzyme in monoaminecontaining cells is not clear. Since most of these cells receive a synaptic input from cholinergic nerve terminals, they are sensitive to acetylcholine. Acetylcholinesterase hydrolyzes and thereby inactivates the neurotransmitter acetylcholine. Hence, this cholinesterase, synthetized in NEE cells, transported on the outer surface of the cellular envelope, and released from this site, is likely to participate in modifying the microenvironment of the NEE cells, regulating the responsiveness to acetylcholine after its release from presynaptic axon terminals. It might also be that the presence of acetylcholinestemse is related to that of active acetylcholine metabolism in NEE cells, since, in some biogenic amine-containing cells, choline acetyltransferase (Ballard and Jones, 1972) as well as the uptake of [3H]choline was demonstrated (Fidone et al., 1976). Indeed, acetylcholine probably plays a role in modulating chemosensory discharge (Eyzaguirre and Zapata, 1968). Moreover, it was demonstrated that cultures of cells from pheochromocytoma may synthetize, store, and release acetylcholine (Greene and Rein, 1977). In adrenergic tissue (Burn and Rand 1959, 1965)and in parathyroid C cells (Welsch and Pearse, 1969), the release of catecholamines is mediated or facilitated by acetylcholine, so that cholinesterase activity may be correlated with physiological states of these cells.

V. Neuron-Specific Enolase

Neuron-specific enolase is a protein which, in light microscopic immunocytochemistry, was found distributed in neuronal perikarya, dendrites, and axons (Pickel et al., 1976; Schmechel et ul., 1978a). It was shown to be a major neuronal protein correlated with neuronal development and differentiation; observations have apparently established it to be essential to the neuronal function (Marangos et al., 1978; Marangos

THE ENDOCRlNE EPITHELIAL SYSTEM IN THE LUNG

47

and Schmechel, 1980). Since immunocytochemical investigations have revealed neuron-specific enolase to be present in endocrine cells of the APUD series (Schmechel et a/., 1978b) and in the larger cell group called paraneurons (Fujita et al., 1983), this enzyme seems to be a functional marker for the diffuse neuroendocrine system (Marangos et al., 1981; Tapia et a / . , 1981). Consistent with this finding, immunostaining using antibodies to neuron-specific enolase was found in both single NEE cells and in NEBS of the respiratory tract (Cole et a / . , 1980; Tapia et al., 1981; Wharton et al., 1981; Polak and Bloom, 1982, 1984; Sheppard et al., 1984). Although in the adult human lung distinctly immunostained NEE cells can be found, a considerably larger number is present in lungs from human neonates and perinates (Polak and Bloom, 1984). In serial sections of the human fetal lung, at least three different types of NEE cells on the basis of their immunoreactive content were identified containing ( 1) neuron-specific enolase, 5-HT, and bombesin; (2) neuronspecific enolase and 5-HT; and (3) neuron-specific enolase only (Wharton et a/., 1981). This enzyme was for the first time observed in lungs of 16week-old human fetuses; they assumed it to be a useful marker for the immunocytochemical detection of NEE cells in the lungs at any age and regarded its presence as indicative of the starting functional activity of these cells, since, in neuronal tissue, its appearance coincides with the initiation of synaptic contacts (Marangos et al., 1979). Some authors reported the first immunoreactive cells for neuron-specific enolase in the human fetal lung at about 8 weeks gestation, i.e., before neuropeptide could be detected (Sheppard et al., 1984), an observation strengthened by immunostaining for 5-HT and the electron microscopic findings of cytoplasmic secretory granules (Cutz et a / . , 1984). In contrast herewith, is a report by Takahashi and Yui (1983), who detected the first immunoreactive cells for 5-HT in the human fetal lung at 12 weeks. In rats chronically exposed to asbestos fibers, an increased number of large and irregular clusters of NEE cells may be demonstrated in the bronchopulmonary tree, using antibodies to neuron-specific enolase (Cole et a / . , 1982; Sheppard et a/., 1982; Polak and Bloom, 1984). According to these and other authors, using this antibody, neuroendocrine tumors can be identified in the lung (Tapia et a/., 1981; Polak, 1983). Although immunocytochemistry has detected neuron-specific enolase mostly in neurons and paraneurons (Marangos et a / . , 1981), it appeared that immunostaining for this neuronal protein in tissues and cells includes the reaction with a hybrid form of enolase (Marangos et al., 1980; Kato et al., 1982; Haimoto et al., 19851, warning that this way to detect NEE cells must be used with care.

48

D. W. SCHEUERMANN

VI. Aspects of Induced Fluorescence

Recent microscopic studies with the FIF method (Falck and Owman, 1965) have made it clear that many NEE cells and NEBs in the wall of the respiratory system contain biogenic amines (Lauweryns and Peuskens, 1969, 1972; Lauweryns et al., 1970, 1973; Hage, 1972, 1976; Cutz et al., 1974; Cutz, 1982; Scheuermann et al., 1983a, 1984a; Scheuermann, 1984). After treatment with formaldehyde vapor, these cells exhibit a fluorescence of variable intensity (Fig. 4). The ring-closing condensation reaction, occumng between formaldehyde and indolylethylamines or catecholamines, yields heterocyclic compounds. The subsequent dehydrogenation, in the presence of proteins, produces strongly fluorescent dihydro-P-carbolines (Corrodi and Jonsson, 1965) and dihydroisoquinolines(Corrodi and Hillarp, 1964), respectively. There are authors who make use of a recently developed glyoxylic acidinduced fluorescent method for the demonstration of biogenic amines in NEE cells (Rogers and Haller, 1978; Hung, 1980). Because of the low endogenous amine content in APUD cells of some animals, it is difficult to reveal NEE cells directly by fluorescence microscopy without treatment with an amine precursor. Others, without treatment, are apparently devoid of a cellular store of biogenic amines. They can be revealed by in vivo or in vitro treatment of the specimens with an amine precursor, (3,4-dihydroxypheny1)-L-alanine or ~-5-hydroxytryptophan,which demonstrates their amine-handling properties (Ericson er al., 1972; Hage, 1973a, 1974, 1980,1984;Cutz etal., 1974,1975;Walsh and McLelland, 1974;Sonstegard et al., 1976; Hage et al., 1977; Lauweryns et al., 1977; Palisano and Kleinerman, 1980; Dey et al., 1981). The fluorescence appears predominantly in the basal, paranuclear cytoplasm; the nucleus is free of fluorescence (Figs. 2 and 7). After treatment with sodium borohydride solution, the fluorescence is nullified and then reestablished by further exposure to formaldehyde vapor, demonstrating the specificity of the histochemical reaction for monoamines (Corrodi et al., 1964). As determined in intrapulmonary NEBs of the rabbit (Lauweryns et al., 1973, 1974, 1977) and in single NEE cells of the same animal (Dey et al., 1981), as well as in human NEE cells (Keith er al., 1981), where the maximum intensity of the fluorescence emission is situated between 520 and 530 nm, it seems likely that the amine involved is 5-HT. Special caution, however, is required by the fact that the maximal emission intensity of catecholamines, usually about 480 nm, may under certain conditions be situated in the wavelength range from 500 to 540 nm, thus appearing yellowish. This spectral shift of catecholamines may occur when high concentrations are present, provided the catecholamine-protein ratio

THE ENDOCRINE EPITHELIAL SYSTEM IN T H E LUNG

49

is very high (Caspersson et d . ,1966; Corrodi and Jonsson, 1967;Jonsson, 1967a. 1971a,b; Bjorklund et d . , 1968; Bjorklund and Falck, 1973; Laszlo, 1975). An exact differentiation between the fluorophores of indolylethylamine derivatives and catecholamines can only be achieved by microspectrofluorometric readings of excitation spectra. Indeed, at neutral pH, the maximal excitation intensity for catecholamine fluorophores is usually higher (i.e., 410 nm) as compared with that of fluorophores from indolylethylamine compounds (around 385 nm). Furthermore, upon acidification. the excitation maxima of catecholamine fluorophores change from 410 to 370 nm, with additional excitation peaks at 320 nm (Bjorklund c’t al.. 1972a.b) and, in an extended excitation range from 240 to 450 nm. a peak at about 260 nm (Reinhold and Hartwig, 1982; Scheuermann et d . , 1984b). On the other hand, it was found that, in some tissues, 5-HT and its metabolic precursor, 5-hydroxytryptophan, may be present simultaneously (Hartwig and Reinhold, 1981). In addition to the excitation peak at 385 nm, excitation recordings conducted in an extended wavelength range yield a distinct clear excitation peak at 310 nm. Moreover, the relative height of the peak at 385 nm, as compared to the excitation peak at 310 nm, appears much higher for 5-HT than for 5-hydroxytryptophan (Reinhold and Hartwig. 1982). Thus, microspectrofluorometrically in an extended excitation range, it is possible to distinguish clearly the fluorophores of 5-HT from those of 5-hydroxytryptophan. Additionally, the differentiation between fluorophores of indolylethylamine derivatives and catecholamines can be performed by studying the rate of fading of the fluorescence: the decrease in fluorescence intensity of the 5-HT fluorophore upon irradiation with the most effective wavelength is more rapid than that of the catecholamine fluorophores (Jonsson, 1967b). Moreover, the final fading rate of the 5-hydroxytryptophan fluorophore is much less pronounced than that of the 5-HT fluorophore (Reinhold and Hartwig, 1982). After formaldehyde condensation, the microspectrofluorometric measurements revealed, in NEBS of several vertebrates (rabbit, pig, and turtle), a maximal emission from 520 to 530 nm and an excitation maximum at 385 nm (Fig. 5). Furthermore, the characteristic spectral shift of catecholamine fluorophores after acid treatment do not materialize. These results allow a classification of the monoamine content in the NEE cells of the lung as an indolylethylamine derivative. Also, the excitation spectra revealed a much higher ratio of the 385-310 peak as compared to S h y droxytryptophan (Fig. 6). After irradiation at the most effective wavelength, the photodecomposition is very rapid, with a loss of 50% of the original fluorescence intensity during the first minute, followed by a much

50

D. W. SCHEUERMANN

Qamar 1

0.5

c.

200

300

400

500 nm

FIG.5 . Excitation (left) and emission (right) spectra recorded by formaldehyde-induced fluorescent, neuroendocrine epithelial cells of the neonatal rabbit. FIG. 6. Microspectrofluorometric excitation spectrum from formaldehyde-induced fluorescent, neuroendocrine epithelial cells of the neonatal rabbit. The extended excitation range from 240 to 460 nm with the characteristic height of the peaks at 385 and 310 nm makes a differentiation possible between 5-hydroxytryptamine and 5-hydroxytryptophan.

THE ENDOCRINE EPITHELIAL SYSTEM IN THE LUNG

51

slower fading rate and ending in a final intensity decrease of about 70% after 5 minutes (Fig. 9). Hence, these microspectrofluorometric recordings point to the presence of 5-HT. Upon staining the same section with the argyrophilic reaction, the shape, size, location, and distribution of fluorescent cells correspond to that of the argyrophilic cells (Figs. 2 and 3) (Cutz et al., 1975; Dey et ul., 1981), although the argyrophilic reaction is assumed to be less sensitive than the FIF (Palisano and Kleinerman, 1980). When staining was camed out using the argentaffin reaction, those cells, which show a basal yellow fluorescence with spectral characteristics of 5-HT, revealed dark-brown deposits in the basal cytoplasm, indicating that the yellow fluorescent cells are identical to the argentaffin cells (Dey et al., 1981; Scheuermann et al., 19834. On the basis of their fluorescence color, following formaldehyde condensation, it has been suggested (Eaton and Fedde, 1977) that the NEBs of the mouse lung may occur in two different groups, one comprising yellow, quickly fading fluorescent cells, claimed to produce and store an indolylethylamine compound, whereas the other consists of blue-green fluorescent cells producing catecholamines. Consistent with this finding, in whole mounts of the lizard lung, yellow fluorescent cells and cells which fluoresce a faint green have been observed (McLean and Burnstock, 1967b). As mentioned before, the formaldehyde-induced, typically yellowish fluorescence, is in itself, i.e., without objective microspectrofluorometric measurements, inconclusive for the critical identification of a biogenic amine. As we have reported in the turtle lung, the yellow fluorescent cells show the microspectrofluorometric characteristics of 5-HT (Scheuermann et al., 1983a; Scheuermann, 1984). They are situated in the epithelial tissue and belong to the NEBs. The blue-green fluorescent cell groups, which appear as small intensely fluorescent cells, belong to pulmonary ganglia, located in the intraparenchymal connective tissue (Scheuermann et ul., 1984b,c). After formaldehyde condensation and microspectrofluorometry, the latter cells show excitation and emission maxima at about 415 and 480 nm, respectively, corresponding to the occurrence of catecholamines (Bjorklund et al., 1975). Moreover, the fading reaction follows the typical course for catecholamines. After acidification and analysis of the relative intensities of the excitation maxima, the identification of the catecholamine at the cellular level was performed (Scheuermann et al., 1984b). In the lungs of the monkey, pig, rat, and rabbit, these two types of intensely fluorescent cells, i.e., 5-HT-containing NEBs and intraganglionic, catecholamine-containing, small intensely fluorescent cells with the characteristic microspectrofluorometric recordings, can be demonstrated as

9

0

1

2

3

4

5 min

FIGS.7 A N D 8. 5-HT fluorescence and argentaffinity demonstrated consecutively in the same tissue section of the basal portion of a NEB in the lung of the red-eared turtle. (Fig. 7) Cluster of yellow-fluorescent cells occurring in the trabecular epithelium. Note the nerve fibers. located in juxtaposition to the base of the yellow-fluorescent cells and emitting a blue-green fluorescence (arrow). x 1050. (Fig. 8) The same section as in Fig. 7 after subsequent exposure to the Masson-Hamper1 silver technique, showing argentaffinity of granular material in the basal portion of the cells (arrow). Cells displaying a basal formaldehyde-induced fluorescence are identical to those which reveal argentaffin granules in the basal portion of the cell. x 1050.

THE ENDOCRINE EPITHELIAL SYSTEM I N THE LUNG

53

well (D. W. Scheuermann. unpublished observations). Immunocytochemistry carried out at the light microscopic level, using a specific antiserum against 5-HT, confirmed the presence of 5-HT in both NEBS and solitary NEE cells (Cutz et a/., 1982; Lauweryns e l a / . , 1982; Sonstegard et al., 1982; Memoli el al., 1983). It should be made clear that the bluegreen fluorescent cells described in the mouse lung (Eaton and Fedde, 1977) contain catecholamines and do not belong to the NEBs, being intraganglionic, small intensely fluorescent cells located in the lung interstitium. It can be assumed that both populations of APUD cells, i.e., the NEE cells as well as the small intensely fluorescent cells of the intrapulmonary ganglia, might each serve as an origin for lung tumors belonging as they do to elements of the diffuse peripheral endocrine system (Scheuermann et a / . , 1983b).

VII. Immunocytochemistry for Regulatory Peptides Since NEE cells and NEBs of the respiratory system have many features in common with cells from the APUD series initially described by Pearse (1966, 1968, 1969), they may be expected to produce polypeptide hormones. Indeed, as revealed by recent immunohistochemical studies in paraffin sections, pulmonary NEE cells may contain a variety of regulatory peptides. Bombesin-like immunoreactivity was detected in NEE cells of the bronchial and bronchiolar epithelium of the fetal and newborn human lung, both in single cells and in groups of cells (Wharton et a/., 1978; Johnson et a / . , 1982; Stahlman e l al., 1982, 1985); it was shown to be particularly abundant in the second trimester to term (Track and Cutz, 1982). Furthermore, bombesin-like immunoreactivity was detected in NEE cells at all levels of the tracheobronchiolar tract of the adult human lung (Cutz et d.,1981, 1984; Polak and Bloom, 1984) and in NEE cells of the airways of adult rats (Marchevsky and Kleinerman, 1982). Calcitonin-like immunoreactivity was revealed in single cells as well as in some NEBs of human fetuses and neonates (Becker et al., 1980; Stahlman et d.,1982, 1985), in fetal and adult human lungs (Cutz et a / . . 1981), FIG.9. Two fading curves from NEB cells and adjacent varicose nerve fibers. demonstrating the different photodecomposition rate of both fluorophores. The lower curve shows a rapid decomposition of the fluorescence in NEB cells after irradiation with the most effective wavelength. which is especially pronounced in the first minute and finally results in a total fluorescence decrease of about 70% after 5 minutes, typical of formaldehyde-induced 5-HT fluorophore. The upper curve renders the slow and gradual decrease (-20%) in fluorescence intensity of the nerve fibers under identical conditions. i.e., typical of formaldehyde-induced catecholamine fluorophores. [Fig. 7-9 modified from Scheuermann c’t ul. ( 1983a). Reprinted with permission of publisher.]

54

D. W. SCHEUERMANN

and in NEBs of adult rats (Marchevsky and Kleinerman. 1982). However, calcitonin-positive NEE cells appeared most numerous in neonates compared with fetal and adult lungs. Leu-enkephalin antiserum revealed immunostaining in the peripheral airways of the fetal, neonatal, and adult human lung, but only in a few single NEE cells (Cutz et al., 1981). This was also true, to the same extent, during hyperplasia of pulmonary NEE cells in patients with bronchiectasis and bronchial epithelial neoplasms (Memoli et al., 1983). Immunoreactivity to gastrin-releasing peptide, the mammalian counterpart of amphibian bombesin (McDonald et al., 1979; Iwanaga, 1983; Tsutsumi et al., 1983b), was found in the fetal and adult human lung, both in single NEE cells and in NEBs (Takahashi and Yui, 1983; Tsutsumi et al., 1983a,b). However, bombesin immunoreactivity should be attributed to gastrin-releasing peptidelike molecules, which are present in the human lung (Price et al., 1983; Yoshizaki et al., 1984). Consecutive sections of the fetal human lung, alternatively immunostained for 5-HT and gastrinreleasing peptide, revealed the coexistence, within the same NEE cell, of a biogenic amine and a peptide (Takahashi and Yui, 1983). Chemical and immunocytochemical studies confirmed the simultaneous occurrence of neuropeptides and amines in NEE cells and lead to the concept that the coexistence of different bioactive substances in neuroendocrine cells is the rule rather than the exception (for reviews, see Owman et al., 1973; Fujita and Kobayashi, 1974; Pearse, 1976; Hokfelt et al., 1980; Sundler et al., 1980; Fujita, 1983): Using serial sections, immunoreactivity for 5-HT, bombesin, and somatostatin have been detected in the same NEBs of the fetal monkey lung (Dayer et al., 1985). According to their immunoreactivity, four groups of NEBS can be distinguished, i.e., NEBs containing ( I ) 5-HT, bombesin, as well as somatostatin; (2) 5-HT and somatostatin; (3) 5-HT and bombesin; (4) 5-HT only. Will et al. (1985) showed that, in the monkey, NEBs may also yield cholecystokinin immunoreactivity. Whether these peptides are present in different or in the same cells remains to be clarified. Although the simultaneous localization of more than one peptidergic antigen in a single cell has not frequently been reported, the coexistence of gastrin-releasing peptide and calcitonin was demonstrated within a subpopulation of NEE cells of the human bronchial tree using a serial section technique (Tsutsumi et al., 1983b). Whether these different peptides coexist in the same secretory granules of the NEE cells is unknown, although a coexistence in the same granules was recently described for met-enkephalin and oxytocin within nerve terminals of the posterior pituitary gland (Adachi et al., 1985). Species differences exist in relation to the immunoreactivity to regulatory peptides. For instance, in cats, bombesin

THE ENDOCRINE EPITHELIAL SYSTEM I N T H E L U N G

55

is located in .scattered NEE cells of the upper respiratory tract and only occasionally in the bronchiolar epithelium, whereas immunoreactivity to bombesin could not be confirmed in rats and guinea pigs (Ghatei et al., 1982).Apparently, the NEE cell neoplasms of the respiratory system may express a larger spectrum of neuropeptides than has been found so far in normal NEE cells, e.g., they may demonstrate ectopic ACTH immunoreactivity (Gould et al., 1983a,b; Tsutsumi et al., 1983a; Polak and Bloom, 1984; Said, 1984). Since the above-mentioned peptides revealed in the NEE cells of the respiratory tract are known to be widely distributed in the central and peripheral nervous system (Becker et al., 1979; Yanaihara et al., 1981; Yui et al., 1981), it may be that they discharge peptides either into intercellular spaces, acting at least locally as neurotransmitters and/or neuromodulators, or as circulating hormones into adjacent blood capillaries. VIII. Electron Microscopic Aspects

The NEE cells, whether occurring solitarily or clustered, may have various aspects and can be clearly recognized as intraepithelial elements (Fig. 10). The shape of single NEE cells is variable; but triangular, flask-shaped, or pear-shaped cells were regularly encountered, extending from the basement membrane to the airway surface, where they may terminate in a narrow tuft of microvilli (Fig. 18). Others, resting with a broad, basal pole on the basement membrane, prolong their narrow apical portion toward the luminal surface, without actually reaching the airway lumen; it is not unlikely that the latter transforms into the former. Grouped cells appear mostly in a palisade-like arrangement spanning the height of the epithelium between the airway lumen and the underlying connective tissue. Sometimes, the NEE cells revealed a stratification. The basal cells are oval or polygonal, resting on the basement membrane as well as interdigitating with the overlying rather pear-shaped cells with a major portion of the cytoplasm at their base. Some cells of the superficial layer contact the airway lumen by means of a narrow process (Fig. 11) (Ericson et d., 1972; Lauweryns et al., 1972; Terzakis ef al., 1972; Hung et al., 1973, 1979; Moosavi et al., 1973; Tateishi, 1973; Cutz et d., 1974; Hage, 1974; Walsh and McLelland, 1974; Lauweryns and Goddeeris, 1975; McDowell et ul., 1976; Taira and Shibasaki, 1978; Edmondson and Lewis, 1980; Johnson et al., 1980; Goniakowska-Witalinska, 1981; Scheuermann et al., 1983a,b, 1984a). As a result of the irregular course of the apical portion of these cells, it is not always possible in the same section to observe both the cell body and the apical

THE ENDOCRINE EPITHELIAL SYSTEM IN THE L U N G

57

process contacting the airway lumen. However, microvillous projections can be found arising from the apical surfaces. In some species, NEE cells of a NEB arising from a basal body may extend a modified or primary solitary cilium into the intercellular space or into the airway lumen (Rogers and Hailer, 1980; Scheuermann rt ul., 1983a). In transverse sections of the proximal portion of the cilium, the axonemal pattern appears as 9 + 0, representing only nine peripheral doublets without a central pair. Somewhat more distally, one of the peripheral paired tubules may be observed to be displaced gradually toward the center of the cilium, while eight pairs are arranged evenly around the periphery, resulting in an 8 + 1 axonemal configuration. An associated centriole connected to striated rootlets can be observed as a diplosomal basal structure. A similar kind of cilium has been reported in some neurons and sensory cells with a well-known receptor function (Barnes, 1961; Meyer and Bencosme, 1965; Dubois and Girod, 1970; Munger, 1971; Vigh and Vigh-Teichmann, 1973; Afzelius, 1975; Vigh-Teichmann et ul., 1976a,b. 1980; Kataoka, 1974). Although it should be borne in mind that the function of the single cilium remains unknown, the occurrence of such a modified cilium in some sensory receptor cells is suggestive of the detection of environmental conditions, such as chemical, osmotic, or local mechanical changes. Moreover, receptors on the microvillous apical cell membrane might recognize stimuli from the airway lumen and transduce them by an intracellular mechanism which triggers off or arrests the release of secretory granules. The morphological features of the apical portion of the NEE cells of the respiratory system are similar to those of the gastroenteric endocrine cells; hence, the hypothetical term of “taste cells,” proposed by Fujita and Kobayashi (1974) in their study on the gut, might be borrowed to designate similar cells in the lung. In lower vertebrates, as reported in the toad lung, when discussing light microscopic aspects, the apical surface of the NEE cells of a NEB can be observed in electron micrographs to be completely covered by an apical cell, the luminal pole of which is comparable with that of well-known receptor cells (Rogers and Hailer, 1980). This apical cell, provided with microvilli and a primary cilium, also contains microtubules, bundles of microfilaments, rough endoplasmic reticulum, and dense-cored granules.

FIG. 10. NEB in the trabecular epithelium of the red-eared turtle. The neuroendocrine epithelial cells form an organoid structure between the basement membrane and the flattened Clara-like cells (arrow) covering most of the apical surface. Some neuroendocrine epithelial cells ( * I extend from the basement membrane to the airway lumen. The capillary lumen (C) is surrounded by a thin-walled endothelium and a narrow, subendothelial space. At lower left, the pulmonary ciliated epithelium. x 5600.

FIG.I I . Superficial neuroendocrine epithelial cell of a NEB of the red-eared turtle covered by cytoplasm from Clara-like cells. The remainder opens onto the air space and bears microvilli. Intercellular junctions (arrow). Note the concentration of dense-cored vesicles in the basal portion of the cell. x 8500.

FIG. 12. Putative neuroepithelial endocrine cell in the lung of a developing red-eared turtle, observed in mitotic division. The cytoplasm contains a moderate number of characteristic dense-cored vesicles. x 10,500.

‘THE ENDOCRINE EPITHELIAL SYSTEM IN T H E L U N G

59

It was proposed that the apical cell functions as a receptor-transducer cell and that the underlying NEE cells serve as an additional source of peptides-biogenic amines to be released on stimulation of the apical cell. The combination of NEE cells buried in the respiratory epithelium with a specialized epithelial cell that could serve as a chemoreceptor stimulating the secretion by contacting NEE cells was also observed in the fetal human lung (Stahlman and Gray, 1984). In the adult human tracheobronchial tract, NEE cells occur singly (Bensch r t d . , 1965; Gmelich et a / . , 1967; Basset el ul., 1971; Terzakis c’t ul., 1972; Tateishi, 1973; Bensch et ul., 1968), adjacent to the basement membrane, rarely extending to the airway surface (McDowell et ul., 1976). Conversely, in the fetal and neonatal human lung, numerous NEE cells extend from the basement membrane to the luminal surface, terminating in a microvillous border (Hage, 1973b; Stahlman and Gray, 1984). In scanning electron microscopy, the luminal surface of NEBs of the fetal rabbit (Cutz et id., 1978b) and rat (Carabba et d . , 1985) appears partially covered with nonciliated cells showing dome-shaped protrusions, which make them similar in appearance to Clara cells (Fig. 14) (Kuhn et a / . , 1974; Kuhn, 1976). Correlated scanning and transmission electron microscopy revealed that, in neonatal mouse lungs, the Clara-like cells failed to display both the large amounts of endoplasmic reticulum and the secretory granules described as characteristic for mature Clara cells (Hung et r i l . , 1979). However, numerous large mitochondria and accumulations of particulate glycogen did occur (Sonstegard et ul., 1982). The differences in ultrastructural features between Clara cells of neonatal and adult mammals may reflect the immaturity of the former (Smith et al., 1974). Yet, in the red-eared turtle, the boundary of NEBs was outlined by flattened nonciliated cells containing numerous mitochondria as well as membranebound secretory granules and conspicuous cisterns of smooth and rough endoplasmic reticulum, which are all cytoplasmic features of Clara cells (Scheuermann et ul., 1983a). Interdigitating basolateral cytoplasmic processes of NEE cells may extend between other epithelial cells (Bensch et d . , 1965; Lauweryns and Peuskens, 1969; Tateishi, 1973; Dey et ul., 1981). Along the lateral interfaces next to the lumen of adjacent epithelial cells, tight junctions have been observed in the lung of the adult rat, the adult hamster (Edmondson and Lewis, 1980), and the adult mouse (Hung et ul., 1973; Hung and Loosli, 1974). In the adult human lung, the lateral cell membranes of NEE cells are linked to adjoining cells by desmosomal structures (Hage et ul., 1977). Junctional complexes, composed of zonulae occludens, zonulae. and maculae adherens, linking NEE cells to adjacent nonendocrine cells, are reported in the toad lung (Rogers and Haller, 1978). In the red-eared turtle,

60

D. W. SCHEUERMANN

NEE cells are interconnected and linked to adjacent Clara-like cells by small desmosomes (Scheuermann er a / . , 1983a). As yet, there is no physiological evidence that would indicate a possible coupling between NEE cells. In large NEE cell clusters of the rat and hamster lung, narrow interstitial spaces reveal expansions, forming channels in which microvilli are seen to extend (Moosavi et al., 1973; Edmondson and Lewis, 1980). These intercellular canaliculi-like spaces resemble the channels between other endocrine cells, such as the hypophysis (Rennels, 1964) and the adrenomedullary cells (Wetzstein, 1957; Coupland, 1965; Elfvin, 1965; Coupland and Weakley, 1970; Grynszpan-Winograd, 1975). Hematogenous and interstitial substances may be transported through the channels in order to reach the receptor site of the basolaterdl cell membrane of the NEE cells, to which these may respond by triggering off or arresting a hormone release, in analogy to the D-glucose activation process described in the islets of Langerhans of the rat (Niki P r ul., 1974). The nucleus, usually spherical or ovoid with some small indentations, is located in the bottle-shaped cells at the entrance to the narrowed portion as well as in the polygonal cells almost at the center. This nucleus contains patches of dense chromatin, situated along the nuclear membrane and surrounding a prominent nucleolus. The nuclear envelope shows numerous pores. In the human fetal lung (Stahlman and Gray, 1984) and in the lung of the developing red-eared turtle, a putative NEE cell was observed in mitotic division (Fig. 12). A well-developed Golgi complex is often composed of greatly distended sacculi and different kinds of vesicles, partly filled with an electron-dense content (Fig. 16). It is usually present in a supranuclear position, although sometimes a lateral Golgi complex can be observed as well. This pattern, more evident in NEE cells abutting on the luminal endings, suggests some double functional polarity. There are investigations providing morphological indications for the involvement of the Golgi complex in the formation of dense granules (Gmelich et ul., 1967; Lauweryns and Cokelaere, 1973a; Hage, 1974; Taira and Shibasaki, 1978; Scheuermann et al., 1983a). FIG.13. Long strands of rough endoplasmic reticulum with attached and free polyribosomes visualized in the lateral region of a granule-containing cell in the red-eared turtle. x 38,000. FIG.14. NEB in the bronchiolus of a neonatal rabbit is largely covered by protruding nonciliated Clara-like cells. The uncovered part of neuroepithelial endocrine cells can be recognized in a craterlike depression between the Clara-like cells, where the stubby microvillous projections of the narrow tips of these cells reach the airway lumen. Arrow indicates the exposed surface of the NEE cells. Adjacent bronchiolar epithelium is composed of ciliated cells. Scanning electron microscopy. x 3200.

T H E ENDOCRINE EPITHELIAL SYSTEM IN T H E L U N G

61

62

D. W. SCHEUERMANN

In most species, the cytoplasm contains rather small mitochondria (Ericson et a/., 1972; Cutz et a/., 1974, 1975). Multivesicular bodies, lysosomes, pinocytotic vesicles, microtubules, and a few glycogen particles are present in variable numbers. The granular endoplasmic reticulum is usually dispersed in the cytoplasm, but sometimes a configuration with attached and free polyribosomes may resemble, to some degree, the Nissl substance of neurons (Fig. 13). Bundles of branching microfilaments, frequently packed in sheaves (Fig. 15) (Cook and King, 1969; Cutz and Conen, 1972; Ericson e f a / . , 1972; Lauweryns et a / . , 1972; Terzakis rf d . , 1972; Hung et ul., 1973; Hung and Loosli, 1974; Cutz et a / . , 1975; Taira and Shibasaki, 1978; Johnson et a / . , 1980; Scheuermann et a / . , 1983a; Hung, 1984; Pack and Widdicombe, 1984; Stahlman and Gray, 1984; Pearsall ef a / . , 1985), are considered by many authors as a major distinguishing feature of these cells. Although the role of microfilaments is as yet unknown, their presence is of special importance in the pathology-oriented classification of carcinoids of the bronchopulmonary tumors (Gould et a / ., 1983b). In electron micrographs, the most striking feature of the NEE cells is the presence of numerous vesicles with granular cores, referred to as dense-cored vesicles (DCV), which vary in number, sometimes scattered throughout the cytoplasm, but frequently tending to accumulate in the broad basal portion of the cell (Figs. 17 and 19). The shape of the cells, with their deep broad pole, the location of the nucleus, and the basal position of the secretory granules indicate that the pole of discharge is directed toward the connective tissue, in which capillaries are closely adjacent to the basis of the NEE cells. These features, in addition to the fact that the capillary endothelium sometimes appears fenestrated (Lauweryns ef d . , 1974; Scheuermann et d . , 1983a), strongly suggest that the NEE cells may function as endocrine glands. The DCV are usually spherical, but can also be irregular in shape. A limiting smooth-surfaced membrane encloses the osmiophilic content entirely. Usually there is a clear halo between the dense core and the membrane, ranging in width from a few nanometers to 20 nm. In larger vesicles, the dense core may be at the center of the vesicle or adhering eccentrically to the limiting membrane. The external diameter usually ranges from 60 to 200 nm in diameter (Bensch et a / . , 1965; Cook and King, 1969; Ericson et a / . , 1972; Lauweryns et a / . , 1972; Terzakis et d . , 1972; Hung ef a / . , 1973; Moosavi et a/., 1973; Cutz et a / . , 1974; Hung and Loosli. 1974; Walsh and McLelland, 1974; Cutz and Orange, 1977; Hage et u/., 1977; Becci et a / . , 1978; Taira and Shibasaki, 1978; Johnson et d . , 1980; Hung, 1984; DiAugustine and Sonstegard, 1984; Stahlman and Gray, 1984). The size of the DCV is sometimes described as specific for each animal species.

T H E ENDOCRINE EPITHELIAL SYSTEM IN T H E L U N G

63

F a . 15. Sheaves of microtilaments can be observed in the cytoplasm between the nucleus ( N ) and the Golgi complex ( G ) of a neuroepithelial endocrine cell of the red-eared turtle. x 40,000.

FIG. 16. Golgi complexes of a neuroendocrine epithelial cell of the red-eared turtle with electron-dense materials in some Golgi vesicles, indicated the involvement of this organelle in the formation of the specific granules. x 2 6 . 0 0 .

64

D. W. SCHEUERMANN

FIG. 17. Base of a NEB, showing numerous secretory granules concentrated near the basement membrane. Some granules (arrow) contact the plasma membrane. In the connective tissue, numerous nonmyelinated nerve fibers occur immediately below the NEB. x 14,000.

THE ENDOCRINE EPITHELIAL SYSTEM IN THE LUNG

65

e.g., mouse, 107 nm, and rabbit, 142 nm (Hage, 1974). However, both the size and electron density of the DCV vary greatly not only from cell to cell, but even within the same cell (Lauweryns et ul., 1972, 1974; Taira and Shibasaki, 1978; Goniakowska-Witalinska, 1981). When the core is less dense, it may show a faintly granular substructure. In some DCV, a central constriction is seen, which appears to divide the granule into two portions, containing two dense cores as in a hourglass. They possess the same fine structure as the neurosecretory granules found as a normal component in the adrenomedullary cells. Different fixation and staining procedures may affect the ultrastructural appearance of the membrane-bound granules (Chen et a l . , 1969; Matthiessen et al., 1973; Schafer et ul., 1973). The dense core of the DCV in adult human bronchial NEE cells, stained intensely with phosphotungstic acid at low pH. suggests the presence of a glycoprotein (McDowell et d., 1976). Since glycosylation of proteins in order to form glycoproteins appears to be one of the functions of the Golgi complex (Dauwalder et ul., 1972), the close relationship between maturating DCV and the Golgi complex does not seem surprising. The material contained in the halo of the vesicle may be variously extracted by the fixation fluids or by other substances used for dehydration and embedding, thus accounting for the different electron density of the granule halos (Schafer et ul., 1973). but it is hard to accept that the procedures used for tissue processing can produce different effects in the same sample on identical cell components. Ultrastructural studies have revealed as many as three types of NEE cells in fetal human lungs (Cutz and Conen, 1972; Hage, 1972, 1973b, 1980; Capella et ul., 19781, mainly on the basis of the fine-structural morphology of their cytoplasmic secretory granules. The first and most frequent cell type, called P, cells, bears small secretory granules with a mean diameter of about I10 nm. Two varieties of granules may be present: ( I ) membrane-bound, spherical granules displaying a thin clear halo interposed between the central dense core and the membrane; (2) spherical to ovoid vesicles containing a small, eccentrically situated core of variable electron density. These P, cells are provided with an extensive rough endoplasmic reticulum, an expanded Golgi complex, scattered small vesicles, and microtubules. P, cells are located in all parts of the bronchial tree of the developing lung and give a positive argyrophilic reaction, whereas only

FIG. 18. Enlargement of the luminal pole of a N E B , showing the apical portion of a neuroepithelial endocrine cell bearing a tuft of rnicrovilli. Junctional complexes (arrow).The Golgi complexes are supranuclear. The apical cytoplasm contains a moderate number of characteristic dense-cored vesicles and lysosomes. X 17,000.

66

D. W. SCHEUERMANN

T H E EN1)OCKINE EPITHELIAL SYSTEM IN THE LUNG

67

the cells with vesiculated granules are argentaffin. They can be observed forming clusters similar to the corpuscular arrangement of NEBS. The second type, called P2 cells, contains slightly larger, spherical granules with a mean diameter of about 140 nm and provided with a moderately electron-dense core surrounded by a thin transparent rim. Likewise, PI cells are found in all parts of the bronchial tree of the human fetus. The third type, P, endocrine cells, possess large, spherical, membranebound granules with a mean diameter of 190 nm, displaying a homogeneous electron-dense content. In the normal fetal human lung, these P, cells are restricted to the larger bronchial tubes, where they are present in small numbers. According to Hage (1973b), P, and P2 cells may be merely different functional stages of the same cell. In the normal adult human lung, she observed only a single type of endocrine cell, which she called Pa cells; they are provided with spherical, dense-cored. secretory granules characterized by their uniform size and homogeneous appearance. The diameter of the vesicle ranges from 110 to 140 nm (Hage et ul., 1977). There is a narrow clear rim between the core and the surrounding membrane, similar to the secretory granules of the PI cells in the fetal human lung. The P, cells of the fetal human lung and the P, cells from the adult human lung resemble certain endocrine cells in the stomach and pancreas (Solcia e t d.,1975; Capella et ul., 1978). A similar ultrastructural identification of NEE cells in the fetal human lung, which contain distinctive secretory granules, was described by Stahlman and Gray (1984). Some authors correlate these distinct types of NEE cells, defined by the features of vesicle appearance, with the three types of NEE cells distinguished according to their immunoreactivity to neuron-specific enolase, whether or not combined to 5-HT and bombesin (Polak and Bloom, 1982). However, it should be pointed out that biochemical data (Scrutton and Utter, 1968) as well as ultrastructural immunocytochemical observations on neuroendocrine cells (Zabel and Schafer, 1985) indicate that neuron-specific enolase apparently is not associated with secretory granules. Nevertheless, in electron micrographs of the fetal human lung, gastrin-releasing peptide immunoreactivity was found in cytoplasmic granules (Iwanaga, 1983) similar to the DCV of the cell type classified as P, (Hage, 1973b). FIG.19. Enlargement ofthe basal part o f a NEB. showing secretory granules of various shapes and electron density. At the upper right, a capillary separated from the neuroendocrine epithelial cell by a subendothelial space containing collagen fibers. x 28,000. FIG. 20. Part of ii NEB of the red-eared turtle stained with the Masson-Hamper1 argentaftin reaction on ii grid. Intense deposition of silver grains is shown in the granules. A slight background precipitation of silver is seen over the cytoplasm and nucleus. x 15.000.

68

D. W. SCHEUERMANN

Some authors feel inclined to assume that every ultrastructural type of NEE cell produces one specific peptide, i.e., that each peptide is located in a different type of NEE cell (Cutz et al., 1981). However, this finding is difficult to reconcile with the observation, in serial sections of immunostained material, that the same NEE cell may contain different peptides (Tsutsurni et al., 1983b; Zabel, 1984). These peptides may even be present in the same secretory granule, as was demonstrated for met-enkephalin and oxytocin within nerve terminals of the neurohypophysis (Adachi et al., 1985). Conclusive correlation of the ultrastructurally defined types of NEE cells with the presence of specific peptides requires the application of a double immunocytochemical staining technique for the simultaneous demonstration of coexistent neuropeptides at the electron microscopic level. In fetal and adult human trachea as well as in the trachea of adult rabbits, Cutz et al. (1975) found NEE cells with only one type of spherical granules, measuring about 100 nm in diameter and displaying a homogeneous dense core surrounded by a clear space of 16-18 nm. DiAugustine et al. (1984) also described one kind of NEE cells in the trachea of the guinea pig. Conversely, in the tracheal mucosa of lamb and armadillo, Cutz et ul. (1975) demonstrated two distinct types of NEE cells whose DCV differ not only in their mean diameter, i.e., 168 versus 112 nm for the lamb and 175 versus 125 nm for the armadillo, but also in their configuration and electron density. It seems likely that NEE cells in the lung of man and various animals have ultrastructural differences particularly with respect to their DCV. This diversity could reflect species variation (Hage, 1974). but the differences might just as well be due to variations in the physiological and/or pathological state. Ultrastructural studies have revealed that a number of human bronchial carcinoid tumors and oat cell carcinomas of the lung (e.g., Bensch et ul., 1965, 1968; Toker, 1966; Gmelich et al., 1967; Hachmeister and Okorie. 1971; Hattori et al., 1972; Gould et a l , , 1983a,b, 1984) is composed of DCV-containing cells. Compared to the normal adult human bronchial epithelium, the secretory granules in these cells may display a wider range in size and shape, as well as a greater variation in both electron density and ultrastructural configuration (for reviews, see Hage et al., 1977; Taira and Shibasaki, 1978; Gould et ul., 1983a,b, 1984; Hage, 1984). This heterogeneity of the granules points to a similarity with some enterochrornafin cells of the human gastric mucosa (Pearse et al., 1970; Hage, 1973d; Solcia at ul., 1975),which reflects the common entodermal origin of the gut wall and the bronchoalveolar tree. It has been reported that morphometrical analysis, after glutaraldehyde fixation, of fetal rabbit NEBS might provide evidence for the existence

T H E ENDOCRINE EPITHELIAL SYSTEM IN THE LUNG

69

of two types of DCV within the same cell (Lauweryns et al., 1972, 1974; Sonstegard et al., 1979). Type I DCV are wedge or ovoid, with a diameter of about I34 nm, containing an electron-dense amorphous core, usually with a narrow clear halo subjacent to the limiting membrane. In this halo, acetylcholinesterase was detected (Lauweryns and Cokelaere, 1973a). Near the center, the electron-opaque core may contain a compact deposit surrounded by a more grayish periphery extending up to the limiting membrane. Type I1 DCV have a more spherical shape with a diameter of about 112 nm and a less electron-dense core surrounded by a distinct, large, perigranular, clear halo of about 15-20 nm. Formalin pretreatment blocks the active sites of catecholamines but does not prevent the indolamines from reacting with glutaraldehyde in order to form a SchifT monobase. Since this is necessary for the subsequent reaction with potassium dichromate in order to obtain electron-opaque deposits, the formalin-glutaraldehyde-dichromate method is considered to be specific for indolamines (Wood, 1967; Jaim Etcheverry and Zieher, 1968). The latter method was applied to NEE cells, demonstrating that only type 1 DCV yield dense reactive granules, in contrast to type 11 DCV, which can be seen after uranyl acetate staining only (Lauweryns et d., 1972, 1977). This observation therefore suggests that only type I DCV contain 5-HT. In addition. the latter authors consider the possibility of a direct conversion of DCV I1 into DCV I, which may represent their mature form (Lauweryns et a / . , 1977). Thus, the differences in size, shape, and characteristics of the DCV within the same NEE cell are assumed to represent stages of granular genesis (Lauweryns et a / . , 1977; Sonstegard et al., 1979). These results might parallel the different maturation stages during the development of some granular vesicles in the adrenergic neuron (Machado, 1971). Their formation in adrenergic fibers seems to be initiated by an agranular vesicle in which the development takes the place of an eccentric small core attached to the vesicle membrane. The size of the core increases after further accumulation of dense or semidense material, finally resulting in a vesicle with a large dense core, apparently forming a mature DCV. This comparison is supported by the fact that secretory granules of NEE cells display transitional stages with regard to the density of their content. Thus, besides vesicles which are almost empty, there are granular vesicles whose dense core area is extremely small, i.e., with a large clear halo up to the limiting membrane, as well as others with a large dense core filling the vesicle nearly completely, i.e., up to the membrane. In controlled studies of hypoxia on neonatal rats, DCV reveal ultrastructural changes indicating almost the reverse phenomenon of what was suggested in connection with the development (Moosavi et al., 1973). A

70

D. W. SCHEUEKMANN

widening of the clear halo between the osmiophilic core and the limiting membrane goes hand-in-hand with a decrease in size and electron density of the core, which often becomes not only minute but eccentric, lying attached to the limiting membrane. Moreover, many empty vesicles are formed. Hence, one gains the impression that type 1 and type 11 DCV are the extremes of a continuous series of different degrees of amine and/or polypeptide filling. Thus, the type 11 DCV would contain little if any biogenic amines, which are hardly detectable by the poorly sensitive aldehyde-osmium tetroxide method. The morphological variations of DCV observed in chronic hypoxia might be considered as an enhancement of variations observed in DCV of normal controls. Quantitative fluctuations between these different types of granular vesicles occurring under normal conditions may reflect physiologic changes in the oxygen content of the inhaled air. Using Fontana’s ammoniacal silver technique at the fine-structural level (Hgkanson et al., 1971), the membrane-bound granules of the bronchopulmonary NEE cells, at least those of the frog (Rogers and Haller, 1978) and the turtle (Scheuermann et ul., 1983a, 1984a), stained selectively by silver deposits over the dense core of the vesicles, demonstrating the argentaffin reaction of the DCV (Fig. 20). This is in agreement with electron microscopic studies of Ericson and co-workers (l972), who have shown and 5-hydroxytryptophan are that tritiated 3,4-dihydroxyphenyl-~-alanine taken up and incorporated into DCV of NEE cells of the mouse lung (trachea). This indicates the presence in the granules of a synthetized amine from exogenous precursors. Moreover, electron microscopic immunocytochemical studies on NEE cells of the respiratory system of human fetuses revealed that gastrin-releasing peptide immunoreactivity is located in DCV (Iwanaga, 1983). Thus, the results obtained so far by light and electron microscopy seem to indicate that, as has also been assumed for endocrine cells in other organs, the DCV of the NEE cells belonging to the respiratory system may be considered as a storage site of biogenic amines and polypeptide hormones (Owman et ul., 1973). There is evidence suggesting that a biogenic amine, whether in combination with coexisting substances, is liberated from the NEE cells by vesicular exocytosis, i.e., direct extrusion of the entire vesicular content to the extracellular space after fusion of the vesicular limiting membrane and the basolateral plasma membrane. Invaginations of the plasma membrane, sometimes containing an amorphous material, similar in size to the DCV content, are indicative of this phenomenon (Lauweryns and Cokelaere, 1973a; Taira and Shibasaki, 1978; Scheuermann et al., 1983a). This suggests that the secretion from the NEE cells could be directed to struc-

T H E ENDOCRINE EPITHELIAL SYSTEM IN T H E LUNG

71

tures below the basement membrane, such as capillaries, smooth muscle cells, and mucosal glands. Granular release at the luminal surface has never been observed. In bronchopulmonary NEE cells of the rabbit, during hypoxia or after intake of certain drugs, such as nicotin or reserpin, the DCV are clearly shifted to the basal pole of the cell. Eventually their bounding membranes are in contact with the basal cytoplasmic membrane, a feature which is rather exceptional in normal animals (Lauweryns et d . , 1977). The content of the granular material appears to empty into the intercellular space at an increased rate by emiocytosis. This phenomenon is correlated with a decrease of 5-HT, as shown by FIF (Lauweryns et d . , 1977). IX. Location

The presence of NEE cells in the epithelium of the respiratory system of man and every vertebrate species examined is well established, although they are not evenly distributed. As early as 1949, Frohlich reported argyrophilic cells to occur mainly at the bifurcations of large and small bronchi as well as at the sites of transition from the bronchioli terminalis to the bronchioli respiratorii. This was later confirmed by several investigators (e.g., Lauweryns et d.,1972; Lauweryns and Goddeeris, 1975; Hage, 1976; Hung et al., 1979; Foliguet and Cordonnier, 1981; Cutz et (11.. 1984; Sarikas et ul., 1985a.b). Single NEE cells were found distributed over almost the entire respiratory system [e.g., larynx (Ewen et a / . , 1972; Kirkeby and Rgmert, 19771, trachea (Ericson et al., 1972; Cutz et al., 1975; Dey et al., 19x1, 19831, and bronchi and bronchioli terminalis (Lauweryns and Peuskens, 1969; Terzakis et d.,1972; Moosavi et d.,1973; Hernandez-Vasquez et al., 1977; Stahlman and Gray, 1984; Stahlman et d . , 1985)l. whereas NEBS seemed to be restricted to the intrapulmonary airways (Cutz et d.,1975). Frohlich (1949) and various later investigators reported that the number of these NEE cells increases in a distal direction up to the smallest bronchi and that the distance between NEE cells increases from the bifurcations of the bronchioli respiratorii onward. As for the upper airways, the ventral mucosa of the trachea revealed more NEE cells than the dorsal mucosa. predominantly in the cranial segment (Dey et d . , 1981),a finding confirmed in the guinea pig (Kirkeby and Rgmert. 1977; DiAugustine et a/., 1984) and rat (Kleinerman et al., 1981). In the lungs of fetal and newborn rabbits (Lauweryns e t al., 1972) and mice (Hung et al., 1979), the number of NEE cells appears high. Hemandez-Vasquez and co-workers ( 1977, 1978a) have shown the number of identifiable NEE cells in fetal rabbit to decrease from 26 to 29 days,

72

D. W. SCHEUERMANN

followed by an increase between 29 days of gestation and I day of extrauterine life, and finally by an initial decrease after birth. The decrease observed in the rat pulmonary NEE cells after the fourth postnatal day is consistent with this pattern (Moosavi et al., 1973). In the adult rat, these cells occur predominantly in the trachea, gradually decreasing in the smaller airway branchings (Kleinerman et al., 1981). Postnatally, the NEB density and average diameter in rabbits were found to decrease in conjunction with the increase in lung volume (Redick and Hung, 1984). Systematic studies have revealed large numbers of NEE cells in fetal human lungs in the early canalicular period (Cutz and Orange, 1977). These were more numerous in proximally differentiated bronchial tubes than in terminal buds. Thereafter, a gradual decrease in the number of NEE cells was observed, although the number per airway remained unchanged (Cutz and Orange, 1977). In the lungs of infants and adult man, NEE cells were found to be more numerous in small bronchi and proximal bronchioli, as compared with major bronchi and bronchioli terminalis (Lauweryns and Peuskens, 1969; Tateishi, 1973). The number of NEE cells is generally reported to decrease with age (for review, see Cutz, 1982; Keith and Will, 1982; DiAugustine and Sonstegard, 1984; Hung, 1984; Pack and Widdicombe, 1984). In some mammalian species, after an initial increase in NEE cells demonstrable in the fetal lung, an apparent decrease close to term was reported, followed by a considerable increase at birth. Since the apparent decrease close to term may be the result of a depletion of cytoplasmic secretory material, some authors ascribed an important role in the respiratory adaptation at birth to the activity of these cells (Lauweryns et al., 1982; Hernandez-Vasquez et al., 1978a; Cutz et al., 1984; Redick and Hung, 1984). However, in contrast to other mammalian species studied, maturation of NEE cells in hamsters does not appear to have been completed at birth (Sarikas et al., 1985a). The significance of these findings is strengthened by quantitative studies in this animal, demonstrating that, I day before birth, most peripheral bronchioles were devoid of NEE cells (Sarikas et al., 1985b). Obviously, it is not inconceivable that the apparently higher frequency of NEE cells in developing lungs of some animal species might be due to their early differentiation, or possibly, also to the smaller dimensions of the fetal lung, i.e., it could be argued that, with peri- and postnatal development of alveoli and growth of the airways, the NEE cells are distributed over an enlarged surface, as a result of which they cannot readily be detected. It appears that the number, as well as the presence, of NEE cells in the different parts of the airways varies not only according to the age of an animal, but also with respect to the species involved. Therefore, in the

THE ENDOCRINE EPITHELIAL SYSTEM IN THE LUNG

73

author's opinion, it is not yet possible to draw a final conclusion with regard to the actual significance of NEE cells at birth.

X. Innervation Using a modified silver impregnation technique for the staining of nervous tissue, Frohlich (1949) revealed, in the bronchial epithelium of rabbits and cats, fine nerve terminals to the very surface of the NEE cells and entering NEBs. He therefore suggested that these cells constitute an afferent' chemosensitive system comparable to the specific cells in the carotid and aortic bodies. Following Frohlich, a number of investigators has observed both light and electron microscopically a distinct innervation of the single and grouped NEE cells in the pulmonary tree of various animal species and in man (Cook and King, 1969; Lauweryns et al., 1970; Hung et al., 1973; Jeffery and Reid, 1975; Hung, 1976, 1980, 1984; Taira and Shibasaki, 1978; Goniakowska-Witalinska, 1980a, I98 I ; Al-Ugaily et a / . , 1983; Stahlman and Gray, 1984). Some authors described nerve endings on NEBs only (Lauweryns et al., 1972, 1974, 1985; Lauweryns and Cokelaere. 1973a; Cutz et al., 1974; Hung and Loosli, 1974; Rogers and Haller, 1978, 1980; Scheuermann et al., 1983a); others have not observed them on solitary NEE cells (Bensch et al., 1965; Terzakis et al., 1972; Cutz et al., 1974, 1975; Hage, 1974; Hage et al., 1977; Hage, 1980), which argues in favor of a subclassification in multicellular NEBs and solitary NEE cells. Different methods have been used to differentiate the nerves associated with NEE cells in the respiratory system. In silver-impregnated sections, unmyelinated nerve endings are described near the basement membrane and surrounding the epithelial cells of NEBs in newborn infants (Lauweryns and Peuskens, 1972) and various vertebrate species (Lauweryns et ul., 1972. 1973, 1974; Hage, 1976; Hung, 1980; Scheuermann et a / . , 1983a). They apparently originate from bundles of nerve processes running in the subepithelial connective tissue, where they are ensheathed by Schwann cells. Although some nerve endings may be traced in the immediate vicinity of argyrophil NEE cells, the pronounced argyrophilia of the latter often impedes a clear recognition of nerve terminals. Some authors reported a dense network of acetylcholinesterase-positive fibers in apposition to NEE cells [lamb (Cutz and Orange, 19771, neonatal rabbit and mouse (Lauweryns and Cokelaere, 1973a), fetal rabbit (Hung, 'The author uses the terms afferent and efferent for terminals on NEE cells of nerve fibers that conduct to and from the central nervous system, respectively.

74

D.

W.

SCHEUERMANN

1980), fetal rat (Morikawa et al., 1978a,b), and fetal human lungs (Taylor and Smith, 1971)l. In some species, a blue-green fluorescent nerve plexus with microspectrofluorometric characteristics of catecholamines was demonstrated by the FIF method in the subepithelial connective tissue, with presumed nerve terminals contacting the yellow fluorescent NEBS (Lauweryns et al., 1972; Hung, 1980; Scheuermann et al., 1983a; Redick and Hung, 1984). The recorded peak of maximum emission is situated at 480 nm, with an excitation maximum at 4 10 nm, characteristic for catecholamines (Bjorklund r t al., 1975). After irradiation at the most effective wavelength, the photodecomposition showed a slow, almost linear decrease in fluorescence intensity, with a loss of less than 20% of the original intensity (Fig. 9). This fading characteristic is in accordance with the results of excitation and emission maxima, arguing for a catecholamine-dependent FIF (Ritzen, 1966; Bjorklund rt al., 1972a,b), and is in contrast with those results obtained for 5-HT-containing cells. In electron microscopic studies, isolated or small groups of nerve fibers and presumed nerve terminals are reported invaginating between NEE cells of a NEB (Hung et al., 1973; Lauweryns and Cokelaere, 1973a; Hung and Loosli, 1974; Lauweryns et al., 1974; Rogers and Haller, 1978, 1980; Goniakowska-Witalinska, 1981; Scheuermann et al., 1983a; Hung, 1984; Stahlman and Gray, 1984) or in close apposition to the basolateral plasma membrane of a single NEE cell (Lauweryns ef al., 1970; Hung et ul., 1973; Jeffery and Reid, 1973; Hung, 1976; Goniakowska-Witalinska, 1980a; Stahlman and Gray, 1984). From the examination of serial sections, it appears that the same nerve fiber may innervate, after branching, several NEE cells by means of bulbous and basket endings or fusiform “en passant” dilatations (Bensch et al., 1965). As a result, the course of one nerve terminal may display a range of appearances. The axoplasm is characterized by the presence of neurotubules. neurofilaments, and small mitochondria (e.g., Hung, 1984). Sometimes, glycogen particles are condensed in larger amounts. Many nerve endings feature an accumulation of various types of vesicles. Clusters of densely packed agranular vesicles of about 60 nm in diameter are almost always present in the approximately oval nerve endings (Figs. 21 and 23) (Cutz et al., 1974; Rogers and Haller, 1978; GoniakowskaWitalinska, 1980a; Scheuermann et al., 1983a). Between the clear vesicles, a few large granular vesicles can usually be observed, ranging in diameter from 90 to 110 nm. These nerve terminals and NEE cells are separated by an extracellular space, about 20 nm wide. Some authors reported nerve processes forming junctions with NEE cells, quite specific for synapses. Features characteristic of these synapses are the presence of cytoplasmic

THE ENDOCRINE EPITHELIAL SYSTEM IN T H E L U N G

75

densities attached asymmetrically to apposed plasma membranes of both the nerve fiber and the NEE cell, as well as numerous small clear vesicles forming clusters close to the junctional material (Lauweryns ef u / . , 1970, 1972. 1974; Lauweryns and Cokelaere. 1973a; Rogers and Haller, 1978; Scheuermann et ( I / . , 1983a; Hung, 1984; Stahlman and Gray, 1984).These structures display the characteristics of efferent cholinergic nerve endings. Furthermore, electron microscopic observations have shown that some adrenergic nerve fibers appear related not only to blood vessels, but also to NEE cells, forming distinct synaptic contacts (Rogers and Haller, 1978). In addition to clear vesicles, these are characterized by small granular vesicles of about 60 nm in diameter, typical of adrenergic nerve varicosities (Rogers and Haller. 1978; Scheuermann et uI., 1983a; Stahlman and Gray, 1984).They can be correlated with the nerve endings observed using the FIF method, since this kind of vesicle is shown to store noradrenaline (Bisby and Fillenz. 1971). Rogers and Haller (1978) argue that the function of the adrenergic nerve fibers might be efferent, since there is no indication of transmission from the NEE cells to the nerve varicosities. In certain vertebrates. two kinds of synaptic regions may be recognized on the same afferent nerve terminal, arranged side-by-side. One region is postsynaptic' and the other presynaptic' to the NEE cell. In the former, an accumulation of small, dense-cored, and clear vesicles occurs along the surface of the presynaptic membrane thickenings in the cytoplasm of the NEE cell. In the latter synaptic component, clusters of small, agranular vesicles (25-50 nm in diameter) are aggregated near dense presynaptic projections on the surface membrane of the NEE cell, without accumulation of dense-cored vesicles in the NEE cell (Rogers and Haller, 1978). These complex paired synaptic contacts were described for the first time as reciproccd synupses in the central nervous system (Reese and Shepherd, 1972). In the peripheral nervous system, they were encountered in the carotid body (King et a / . , 1975; McDonald and Mitchell, 1975; Osborne and Butler, 1975) and cardiac ganglia (Yamauchi el d.,I975a,b). It has been suggested that the regions of reciprocal synaptic junctions, where the nerve terminal is postsynaptic to the NEE cell, are involved in a synaptic mechanism from NEE cell to nerve terminal, while the adjacent synaptic component interacts in the reverse direction. Nerve profiles containing another type of filled vesicle with a wider range in size (80-225 nm in diameter) and a moderately electron-dense content appear in apposition to single NEE cells and to NEBS (Stahlman and Gray, 1984), without yielding synaptic structures. These large granular 'The author uses the tcrms presynaptic and postsynaptic to designate the direction of synaptic transmission.

76

D. W. SCHEUERMANN

THE ENDOCRINE EPITHELIAL SYSTEM IN THE LUNG

77

vesicles may be compared with similar granules observed in peptidergic neurosecretory systems (Rinne, 1966; Bargmann et af., 1967; Krisch, 1974; Gibbins and Haller, 1979; Helen and Hervonen, 1981), which are presumed to belong to nonadrenergic, noncholinergic pathways (Baumgarten et a l . , 1970; Burnstock, 1982). Other enlarged nerve endings on NEE cells are crowded with slender mitochondria (Fig. 22). These terminals have been observed close (6-20 nm) to the plasma membrane of NEE cells of the respiratory system in different mammalian species, e.g., in the mouse (Hung et af., 1973) and rabbit (Lauweryns and Cokelaere, 1973a), and in nonmammalian species, e.g., in birds (Cook and King, 1969; King et al., 1974) and reptiles (Scheuermann et ul., 1983a). Some authors considered these nerve endings as characteristic for sensory (afferent) nerve fibers (Cook and King, 1969; Lauweryns and Cokelaere, 1973a; King et al., 1974; Rogers and Haller, 1978; Hung, 1980; Stahlman and Gray, 1984), with mitochondria serving as a source of energy for the transformation of the stimulus into a nerve impulse. As revealed by serial sections of the turtle lung, the number of mitochondria appeared to vary greatly from one region of a given nerve ending to another (Scheuermann et al., 1983a), confirming observations on nerve terminals in the carotid body (Verna, 1973). In comparison with the innervation of other tissues, the nerve terminals, tightly packed with mitochondria, are mostly considered as sensory (Rees, 1967; Bock et al., 1970; Chiba and Yamauchi, 1970; Kobayashi, 1971; Kondo, 1971; Munger. 1971; Chiba, 1972; Verna, 1973; King et al., 1974). Other authors attributed the accumulation of mitochondria to a degenerative change associated with a process of aging (Seitelberger, 1971; Leonhardt, 1976). However, in the latter case, mitochondria display various stages of disintegration and transitional forms between mitochondria

FIG.21. Cross section through a bundle of subendothelial nerve fibers in the lung of a red-eared turtle near neuroendocrine epithelial cells. Some nerve fibers, partially encased by Schwann cell cytoplasm, contain ( I ) peptidergic granules (PI,(2) cholinergic granules (arrow), (3) adrenergic nerve varicosities (arrowhead). x 19,000. FIG.22. A nerve ending associated with a neuroendocrine epithelial cell in the red-eared turtle lung. No synapse is present. but the nerve ending is densely packed with numerous slender mitochondria. x 47,000. FIG.23. Nerve terminal on the perikaryonal region of a neuroendocrine epithelial cell in the red-eared turtle lung filled with small clear vesicles. Part of a synapse is visible between the arrows. The basal lamina of the neuroendocrine epithelial cell runs at the outer side of the nerve terminal. x 56.000.

78

D. W. SCHEUERMANN

and dense bodies. Nevertheless, in the lateral vestibular nucleus of the rat, dendritic growth cones packed with slender mitochondria are suggestive of regeneration (Sotelo and Palay, 1968). Alternating degenerative and regenerative processes might be considered to occur in nerve fibers abutting the NEE cells as suggested for nerve terminals in the carotid sinus (Knoche and Addicks, 1976). Although no strictly morphological criteria exist for the establishment of the afferent or efferent nature of the NEE cell innervation, selective nerve degeneration experiments are of considerable importance. It was shown that, after unilateral cervical infranodose vagotomy, degenerating intraepithelial axons appeared in the trachea of both the rat (Hoyes and Barber, 1981) and the cat, as well as in the bronchi (Das et a / . , 1979). This is in agreement with the concept that there exist intraepithelial afferent nerve fibers in the tracheobronchial tree. Recently it was shown that most axon endings in the NEBS rapidly degenerate after unilateral sectioning of the homolateral vagus nerve below the nodose ganglion, a process which does not take place after homolateral supranodose vagotomy. Hence, it appears that the cell bodies of these nerve endings are located in the nodose ganglion (Lauweryns and Van Lommel, 1983; Lauweryns et a / . , 1985). Selective labeling with tritiated amino acids of the nodose ganglia proved that the wall of the respiratory system possesses an afferent innervation, at least in the adult hen (Bower et al., 1978), where labeled fibers are observed to enter groups of NEE cells, suggesting a role as receptor. Although great differences in the innervation of the lung may exist among animal species (Richardson, 1979), it appears that the NEE cells are provided with a cholinergic, adrenergic, and nonadrenergic, noncholinergic innervation. The cholinergic and adrenergic nerve terminals are generally considered to be stimulatory, the nonadrenergic, noncholinergic nerve terminals playing an inhibitory role. Although the unequivocal identification of afferent nerve terminals remains difficult, the presumed function of NEE cells, i.e., subserving as chemoreceptors of the airways, leads to assume the existence of an afferent innervation. Finally, it should be mentioned that the morphology and histochemistry of the pulmonary NEE cells were compared to the structure of type 1 cells of the carotid body. The dual innervation with afferent and efferent pathways as well as the morphological features that are produced after hypoxic conditions have both cell types in common (e.g., Keith and Will, 1982; Gould et d.,1983b; Becker, 1984). Nonetheless, the NEE cells differ from principal cells of the carotid body in several respects. For instance, the former contain 5-HT, while type I cells of the carotid body store catecholamines (for review, see Verna, 1979). Moreover, the latter are situated

THE ENDOCRINE EPITHELIAL SYSTEM IN THE LUNG

79

in the carotid body among ganglionic cells, which were never observed between NEE cells. In addition, NEE cells seem to react directly to hypoxia of inhaled air, whereas, in contrast to cells of the carotid body, these do not respond to hypoxemic conditions (Lauweryns et al., 1977; Cutz et a / . , 1982). In consequence, NEE cells do not appear to be vascular chemoreceptors. Conversely, granule-containing cells of the intrapulmonary ganglia have many structural features in common with both type I cells of the carotid body and glomus cells of the aorticopulmonary bodies (Verna, 1979; Papka, 1980; Bock, 1982; Scheuermann et a / . , 1984b). The latter three types of cells contain catecholamines in their secretory granules and, being associated with ganglionic cells, are thought to have a receptor-secretory function. It should be given thought whether the reaction of the lung to hypoxemic conditions may be effectuated through stimulation by vascular chemoreceptors, including the granule-containing cells of the pulmonary ganglia and their release of catecholamines, since the latter substances may produce vasoconstriction in the lung (Bergofsky, 1980; Becker, 1984). The functional relations between the NEE cells, sensing oxygen levels in the pulmonary airways, and the vascular chemoreceptors to oxygen levels in the blood, e.g., the carotid body, the aortic bodies, and perhaps the granule-containing cells of the pulmonary ganglia, are not yet fully understood. A further investigation of the efferent and afferent innervation pattern of the receptors located near gas and blood in the lung is thus required. XI. Concluding Remarks

To date, the typical histological, histochemical, and ultrastructural characteristics as well as the bioactive substances of NEE cells of the lung are fairly well-known. The function of these cells, however, is at present far from elucidation and therefore remains subject to speculative thought. The shape of the NEE cells, whether solitary or grouped into clusters, with a narrow apical portion bearing villous projections into the airway lumen, is indicative of a receptor function. The basal portion is found adjacent to capillaries and may be synaptically connected with varicosities of subepithelial nerve fibers. Ultrastructurally, the nervous connections are suggestive of both afferent and efferent innervation. Furthermore, most of these cells are located in strategic positions at the bifurcations of the bronchial tree. It seems likely that these structures perceive changes in

80

D. W. SCHEUERMANN

the intraluminal environment of the lung, upon which they respond by releasing their secretory products. This assumption is supported by the fact that the NEE cells are degranulated by hypoxia releasing 5-HT. Possibly 5-HT, presumably released in association with polypeptides, could then influence their specific target cells via synaptic structures, local action, or in a vascular way. The central nervous system may modulate, by means of efferent cholinergic nerve fibers, the release from NEE cells of 5-HT and/or polypeptides in response to intraluminal stimuli, e.g., changes in the airway gases. This assumption is supported by the fact that synaptic contact between nerve terminals and NEE cells has been demonstrated with certainty. Thus, the secretory products released from NEE cells may activate afferent nerve terminals, evoking local or systemic reflex changes. The release and diffusion of neurally active substances into nonsynaptic intercellular spaces may provide a morphological basis for a paracrine function, e.g., a local response of the bronchial and vascular smooth muscle and perhaps of the intrapulmonary neuronal plexuses. Moreover, the release of 5-HT and/or polypeptides may be transported by the blood stream, either systemic or specific, from the NEE cells to remote targets. The regularity with which capillaries, at times fenestrated, are observed in proximity to the enlarged basal foot of NEE cells provides a strong indication that bioactive substances secreted by the NEE cells diffuse into the blood circulation, making them widely accessible. Much work remains to be done in the field of lung endocrinology. An improved knowledge of the secretory activities of the NEE cells and a clarification of the functional relationships between these cells and the nervous system represent challenges demanding further research. REFERENCES Adachi. T., Hisano, S.. and Daikoku, S. (1985). J . Hisrochem. Cytochem. 33, 891-899. Afzelius, B. A. (1975). In "Handbook of Molecular Cytology" (A. Lima-de-Faria. ed.), pp. 1219-1242. North-Holland Publ., Amsterdam. Al-Ugaily, L. H., Pack, R. J . . and Widdicombe, J . G . (1983). J . Physiol. (London) 340,54P. Ballard, K. J . , and Jones. J . V. (1971). J . Physiol. (London) 219, 747-753. Ballard, K . J . . and Jones, J . V. (1972). J . Physiol. (London) 227, 87-94. Bargmann, W., Lindner, E., and Andres, K. H . (1967). Z . ZelUorsch. 77, 282-298. Barnes, B. G . (1961). J . Ulrrustruct. Res. 5, 453-467. Barter, R . , and Pearse. A. G. E. (1953). Nutitre (London) 172, 810. Barter, R . , and Pearse, A. G . E. (1955). J . Purhol. Bucreriol. 69, 25-31. Basset, F., Poirier, J . , Le Crom, M., and Turiaf, J . (1971). Z . Zellforsch. 116, 425-442. Baumgarten, H. G . , Holstein, A.-F., and Owman, C. (1970). Z . Zellforsch. 106, 376-397. Becci, P. J . , McDowell, E. M., and Trump, B. F. (1978). J . Null. Cuncer Inst. U.S.61, 551-561.

THE ENDOCRINE EPITHELIAL SYSTEM IN T H E LUNG

81

Becker, K. L. (1984).In "The Endocrine Lung in Health and Disease" ( K . L. Becke\and A. F. Gazdar, eds.). pp. 3-46. Saunders. Philadelphia, Pennsylvania. Becker. K. L., Snider, R. H.. Moore, C. F.. Monaghan. K. G.. and Silva. 0. L. (1979). Actcr Endocrinol. (Copi,n/rcrgc,n)92, 746-75 I . Becker. K. L.. Monaghan. K. G., and Silva. 0. L. (1980). Arch. Ptrrlrol. Lcrh. Med. 104, 196- 198.

Bensch. K. 0 . .Gordon, G . B., and Miller. L. R. (1965). J. Ul/rtts/rrcct. Res. 12, 668-686. Bensch. K . G.. Corrin, B., Pariente. R., and Spencer. H. (1968). Ccrncer (Anrsterdirwr) 22, 1163-1 172. Bergofsky, E. H. (1980). Annrr. Reif. Phvsiol. 42, 221-233. Bisby. M. A , . and Fillenz, M. (1971). 1.Physiol. (London) 215, 163-179. Bjorklund, A.. and Falck, B. (1973).In "Fluorescence Techniques in Cell Biology" ( A . A. Thaer and M. Sernetz, eds.). pp. 171-181. Springer-Verlag. Berlin and New York. Bjorklund. A.. Falck. B., and HPkanson, R. (1968).Actrr Physiol. Sccrnd. Srrppl. 318, 3-31. Bjorklund. A , , Ehinger, B.. and Falck. B. (1972a).J. Hi.stochi~m.C v t o c h n ? . 20, 56-64. Bjorklund, A. Falck. B.. and Owrnan, C. (1972b). In "The Thyroid and Biogenic Arnines" (J. E. Rall and 1. J. Kopin, eds.), pp. 318-368. North-Holland Publ., Amsterdam. Bjorklund. A,, Falck, B . . and Lindvall, 0. (1975).I n "Methods in Brain Research" (P. B. Bradley. ed.), pp. 249-294. Wiley, New York. Blumcke, S. ( 1968). Bcitr. Pir//io/.Anut. Allg. Pcrr/iol. 137, 239-286. Bock. P. ( 1970).Z.Zr//f;wscVr.110, 243-257. Bock. P. ( 1982). "The Paraganglia." Springer-Verlag. Berlin and New York. B&k, P.. Stockinger. L., and Vyslonzil, E. (1970).Z. Zel/fiwsch.105, 543-568. Bower. A . J.. Parker. S., and Molony. V. (1978).J . Anot. 126, 169-180. Breeze. K. G . , and Wheeldon. E. B. (1977). Am. Rev. Ri..spir. D i s . 116, 705-777. Burn. J. H.. and Kand. M . J. (1959). Ntrirrrc~(London) 184, 163-165. Burn. J. H . , and Kand, M . J. (1965). Annr,. Re\*. Plirrnnucd. 5, 163-182. Burnstock. G . (1982).Sccrnd. J . Gir.stroc~n/evd.Srrppl. 71, 135-138. Bussolati. G . . Rost. F. W. D.. and Pearse. A. G . E. (1969). Hi.stoihern. J. I , 517-530. Capella. C . , Hage, E.. Solcia, E.. and Usellini. L. (1978).Cc4 Ti.ssrre Res. 186, 25-37. Carabba. V. H.. Hoyt. R. F.. Jr.. and Sorokin. S. P. (1982). Antrt. RCT. 202, 25A. Carabba. V. H., Sorokin, S. P.. and Hoyt. R. F., Jr. (1985). Am. J . Antrr. 173, 1-27. Caspersson. T.,Hillarp, N.-A.. and Ritzen, M. (1966). Exp. Cell Rrs. 42, 415-428. Chen. 1 . -L.. Yates. R. D., and Duncan, D. (1969).J. CellBiol. 42, 804-816. Chiba, T. ( 1972).J. Electron Microsc. 21, 139-148. Chiba, T.. and Yamauchi, A. (1970).Z. Zellforsch. 108, 324-388. Cole. G . A,. Polak, J. M.. Wharton, J., Marangos, P., and Pearse, A. G . E. (1980). J. Puthol. 132, 35 1-35?, Cole, G . A.. Polak. J. M.. Johnson, N. F., Marangos, P., and Pearse, A. G . E. (1982). J . P ( / / h d . 137, 92-93. Cook, R. D., and King, A. S.jl969).Experientia 25, 1162-1 164. Corrodi, H.. and Hillarp, N.-A. (1964). Hc41,.Chim. A c f a 47, 91 1-918. Corrodi, H., and Jonsson. G. (1965). Actrr Histoc/ic,m. 22, 247-258. Corrodi, H.. and Jonsson. G . (1967).J . Histoclrem. Cvtochcvn. 15, 65-78. Corrodi. H.. Hillarp. N.-A,. and Jonsson. G. (1964).J . His/oc.kc,/u.Cytodicni. 12, 582-586. Coupland. R. E. ( 1965). J . Anu/. 99, 23 1-254. Coupland, R. E., and Weakley, B. S. (1970).J . A n ( / / . 106, 213-231. Cutz, E. (1982). Ekp. Lrrng Res. 3, 185-208. Cutz. E.. and Conen. P. E. (1970).Fed. Proc. F d . Am. SOC.Exp. B i d . 29, 361. Cutz. E., and Conen. P. E. (1972). Anirt. Rec. 173, I 15-122.

82

D. W. SCHEUERMANN

lmmunopharmacology and Cutz. E., and Orange, R. P. (1977). In "Asthma-Physiology, Treatment" (L. M. Lichtenstein and K. F. Austen. eds.), pp. 51-76. Academic Presa, New York. Cutz. E., Chan, W., Wong, V., and Conen, P. E. (1974). Lab. Invest. 30, 458-464. Cutz. E., Chan, W., Wong, V., and Conen. P. E. (1975). Cell Tissue Res. 158, 425-437. Cutz. E.. Sonstegard, K. S., and Chan, W. (1978a). In "Electron Microscopy II.Biology" (J. M. Sturgess, ed.), pp. 490-491. Microsc. SOC.Canada, Toronto. Cutz, E., Chan, W., and Sonstegard, K. S. (l978b). Anat. Rec. 192, 459-466. Cutz, E., Chan, W., and Track, N. S. (1981). Experientici 37, 765-767. Cutz. E., Chan. W., Kay, J. M., and Chamberlain, D. W. (1982). Lab. I n v e s f . 46, 16A. Cutz. E., Gillan, J. E., and Track, N. S. (1984). In "The Endocrine Lung in Health and Disease" (K. L. Becker and A. F. Gazdar. eds.). pp. 210-231. Saunders, Philadelphia, Pennsylvania. Das. R. M., Jeffery. P. K., and Widdicombe, J. G. (1979). J. Anat. 128, 259-267. Dauwalder, M., Whaley, W. G.. and Kephart. J. E. (1972). Subcell. Biochem. 1, 225-275. Dayer, A. M., De Mey, J.. and Will, J . A. (1985). Cell Tissue Res. 239, 621-625. Dey, R. D., Echt, R., and Dinerstein. R. J. (1981). Anat. Rec. 199, 23-31. Dey, R. D., Shannon, W. A., Jr., Hagler, H. K., and Said, S. I. (1983). J. Histochcwi. Cvtochem. 31, 501-508. DiAugustine, R. P., and Sonstegard, K. S. (1984). Environ. Health Perspect. 55, 271-295. DiAugustine. R. P., Jahnke. G. D., and Talley, F. (1984). In "The Endocrine Lung in Health and Disease" (K. L. Becker and A. F. Gazdar, eds.), pp. 232-248. Saunders. Philadelphia. Pennsylvania. Dubois, P., and Girod, C. (1970). Z . Zellforsch. 103, 502-517. Eaton, J. A.. Jr.. and Fedde. M. R. (1977). Cell Tissue Res. 176, 243-249. Eaton, J. A.. Jr., and Fedde, M. R. (1978). Poult. Sci. 57, 793-797. Edmondson, N . A., and Lewis, D. J. (1980). Thorax 35, 371-374. Elfvin, L. -G. (1965). J . Ultrustruct. Res. 12, 263-286. Ericson. L. E., Hikanson, R., Larson. B., Owman, C., and Sundler, F. (1972). Z . Zel/for.sch. 124, 532-545.

Erspamer, V., and Asero, B. (1952). Nature (London) 169, 800-801. Ewen, S. W. B., Bussolati, G., and Pearse, A. G. E. (1972). Historhem. J . 4, 103-110. Eyzaguirre, C., and Zapata, P. (1968). J. fhysiol. (London) 195, 589-607. Falck, B.. and Owman, C. (1965). Acra Univ. Lund. Sect. 2 7, 1-23. Feyrter, F. (1938). "Uber Diffuse Endokrine Epitheliale Organe." Barth, Leipzig. Feyrter, F. (1953). "Uber Die Peripheren Endokrinen (Parakrinen) Driisen des Menschen." Maudrich, Wein. Feyrter, F. (1954). Z . Mikrosk. Anat. Forsch. 61, 73-81. Feyrter, F. (1958). Dtsch. Med. Wochenschr. 83, 958-963. Fidone, S. J., Weintraub, S. T., and Stavinoha, W. B. (1976). J. Neurochem. 26, 10471049.

Foliguet, B., and Cordonnier, J. L. (1981). Bull. Eur. fhysiopathol. Respir. 17, 113-134. Frohlich, F. (1949). Frunkf. Z. fathol. 60,517-559. Fujita, T. (1977). Arch. Histol. Jpn. Suppl. 40, 1-12. Fujita, T. (1983). Biorned. Res. 4, 239-256. Fujita, T., and Kobayashi, S. (1974). In "Gastro-Entero-Pancreatic Endocrine System" (T. Fujita, ed.), pp. 1-16. Thieme, Stuttgart. Fujita, T., Iwanaga. T., and Nakajima, T. (1983). In "Gut Peptides and Ulcer" (A. Miyoshi, ed.), pp. 81-88. Biomed. Res. Found., Tokyo. Ghatei, M. A.. Sheppard, M. N.. O'Shaughnessy. D. J., Adrian, T. E., McGregor, G. P., Polak. J. M.. and Bloom, S . R. (1982). Endocrinology (Baltimore) 111, 1248-1254.

THE ENDOCRINE EPITHELIAL SYSTEM IN THE LUNG

83

Gibbins. I.L.. and Haller, C. J. (1979). Cell Tissrie Res. 200, 257-271. Glorieux. R. (1963). Arch. Biol. 74, 377-390. Gmelich, J. T., Bensch. K. G., and Liebow, A. A. (1967). Lob. I n w s t . 17, 88-98. Goniakowska-Witalinska, L. (1980a). Cell Tissrie Res. 210, 521-524. Goniakowska-Witalinska, L. (1980b). Cell Tissue Res. 210, 479-484. Goniakowska-Witalinska, L. (1981). Cell Tissrre Res. 217, 435-441. Could. V. E.. Linnoila, R. I.,Memoli, V. A.. and Warren, W. H. (1983a). fathol. Annrr. 18, 287-330. Could. V. E.. Linnoila. R. I., Memoli, V. A.. and Warren, W. H. (1983b). Lab. I n i v s t . 49, 5 I 9-537. Could, V. E.. Warren. W. H., and Memoli, V. A. (1984).In “The Endocrine Lung in Health and Disease” (K. L. Becker and A. F. Gazdar. eds.). pp. 406445. Saunders, Philadelphia, Pennsylvania. Greene, L. A.. and Rein, G. (1977). Nutrrre (London) 268, 349-351. Grynszpan-Winograd, 0.(1975). I n “Handbook of Physiology and Endocrinology” ( H . Blaschko, G. Sayers, and A. D. Smith, eds.), Vol. 6, pp. 295-308. Amer. Physiol. Soc., Washington, D.C. Hachmeister. U.. and Okorie, 0. (1971). Verh. Dtsch. G e s . Pathol. 55, 716-721. N Microbial. Scand. Sect. A 80, 225-234. Hage, E. (1972). A ~ I Puthol. Hage, E. (1973a). Actu fathol. Microhiol. Scand. Sect. A 81, 64-70. Hage. E. (1973b). 2. Zellforsch. 141, 401-412. Hage, E. (1973~).2. Anut. Entn~icklringsgesch.140, 271-279. Hage, E. (1973d). Virchows Arch. A : fathol. Anat. 361, 121-128. Hage, E. (1974). Cell Tissrie Res. 149, 513-524. Hage. E. (1976). In “Chromafin, Enterochromaffin and Related Cells” (R. E. Coupland and T. Fujita. eds.), pp. 317-332. Elsevier, Amsterdam. Hage. E. (1980). Invest. Cell Puthol. 3, 345-351. Hage. E. (1984). In “The Endocrine Lung in Health and Disease” ( K . L. Becker and A. F. Gazdar, eds.), pp. 193-209. Saunders, Philadelphia, Pennsylvania. Hage, E.. Hage. J., and Juel, G. (1977). Cell Tissue Res. 178, 39-48. Haimoto, H., Takahashi. Y., Koshikawa, T., Nagura, H., and Kato, K. (1985). Lrrb. Invest. 52, 257-263. Hikanson, R.. Owman. C.. Sporrong, B., and Sundler, F. (1971). Histochemie 27,226-242. Hamperl, H. (1932). Virchows Arch. fathol. Anal. fhysiol. 286, 81 1-833. Hamperl. H. (1952). Virchows Arch. fathol. Anat. fhysiol. 321, 482-507. Hartwiy. H . G . . and Reinhold, C. (1981). I n “The Pineal Organ: Photobiology-Biochronometry-Endocrinology” (A. Oksche and P. Pevet, eds.), pp. 237-246. North-Holland Publ.. Amsterdam. Hattori. S.. Matsuda. M., Tateishi, R., Nishihara. H.. and Horai, T. (1972). Crrncer (Amsterdum) 30, 1014-1024. Helen. P.. and Hervonen, A. (1981). A m . J . Anat. 162, 119-130. Hernandez-Vasquez, A.. Will, J. A., and Quay, W. B. (1977). Thorax 32, 449-456. Hernandez-Vasquez, A., Will, J. A., and Quay, W. B. (1978a). Cell Tissrte Res. 189, 179186. Hernandez-Vasquez. A., Will, J. A., and Quay. W. B. (1978b). CeIl Ti.s.srie Res. 186, 203207. Hokfelt. T., Rehfeld, J. F.. Skirboll. L., Ivemark, B.. Goldstein, M., and Markey, K. (1980). Nature (London) 285, 476-478. Hoyes, A. D., and Barber, P. (1981). J. Annt. 132, 331-339. Hoyt, R. F., Jr., Sorokin. S. P., and Feldman, H. (1982a). Exp. Lung Res. 3, 273-298. Hoyt. R . F.. Jr.. Feldman. H., and Sorokin, S. P. (1982b). Exp. Lung Res. 3, 299-31 I .

84

D. W. SCHEUERMANN

Hung, K.-S. (1976). Anut. Rec. 185, 85-92. Hung, K.-S. (1980). Am. J . Anat. 159, 73-83. Hung, K.-S. (1984). I n “The Endocrine Lung in Health and Disease” (K. L. Becker and A. F. Gazdar, eds.), pp. 162-192. Saunders, Philadelphia, Pennsylvania. Hung, K . 4 . and Loosli. C. G. (1974). Am. J . Anat. 140, 191-200. Hung, K.-S., Hertweck, M. S.. Hardy, J. D., and Loosli, C. G. (1973). J. Ultrustrrtct. Res. 43, 426-437. Hung, K.-S., Chapman, A. L., and Mestemacher, M. A. (1979). Anat. Rec. 193, 913-926. Iwanaga. T. (1983). Biomed. Res. 4, 93-104. Jacobowitz, D., and Koelle, G. B. (1965). J. Pharmucol. Exp. Ther. 148, 225-237. Jacobowitz, D., Kent, K. M., Fleisch, J. H., and Cooper, T. (1973). Proc. Soc. Exp. Biol. Med. 144,464-466. Jaim Etcheverry, G., and Zieher. L. M. (1968). J . Histochem. Cytochem. 16, 162-171. Jeffery, P. K., and Corrin, B. (1984). I n “Immunology of the Lung and Upper Respiratory Tract” (J. Bienenstock, ed.), pp. 1-27. McGraw-Hill, New York. Jeffery, P., and Reid, L. (1973). J . Anuf. 114, 35-45. Jeffery, P. K., and Reid, L. (1975). J . Anur. 120, 295-320. Johnson. D. E.. Lock, J. E., Elde. R. P., and Thompson, T. R. (1982). Pediutr. Res. 16, 446-454. Johnson, N. F., Wagner, J. C., and Wills, H. A. (1980). Lung 158, 221-228. Jonsson, G. (1967a). Actu Histochem. 26, 379-390. Jonsson, G. (l967b). Histochemie 8, 288-296. Jonsson, G. (1971a). f r o g . Bruin Res. 34, 53-61. Jonsson, G. (1971b). Prog. Histochem. Cvtochem. 2, 299-334. Kataoka, K. (1974).I n “Gastro-Entero-Pancreatic Endocrine System” (T. Fujita, ed.). pp. 39-48. Thieme, Stuttgart. Kato, K., Ishiguro, Y.,Suzuki, F., Ito, A., and Semba, R. (1982). Bruin Res. 237,441-448. Keith, 1. M., and Will. J. A. (1982). Exp. Lung Res. 3, 387-402. Keith, 1. M., Wiley, L. A., and Will, J. A. (1981). Cell Tissue Res. 214, 201-205. Keith, 1. M., Wiley, L. A., and Will, J. A. (1982). Histochemistry 75, 251-258. King, A. S.. McLelland, J., Cook, R. D., King, D. Z., and Walsh. C. (1974). Respir. Phvsiol. 22, 21-40. King, A. S., King, D. Z., Hodges, R. D., and Henry, J. (1975). Cell Tissue Res. 162,459473. Kirkeby, S., and Rfimert. P. (1977). J . Anat. 123, 87-92. Kleinerman, J., Marchevsky, A. M., and Thornton, J. (1981). Am. Rev. Respir. Dis. 124, 458-462. Knight, D. S. (1980). J . Anat. 131, 413-428. Knoche, H.. and Addicks, K. (1976). Cell Tissue Res. 173, 77-94. Kobayashi, S. (1971). Arch. Histol. J p n . 33, 397-420. Koelle. G. B. (1950). J . Pharmacol. Exp. Ther. 100, 158-179. Koelle, G. B. (1951). J. Pharrnacol. Exp. Ther. 103, 153-171. Kondo, H.(1971). J. Ultrustruct. Res. 37, 544-562. Korkala, 0.. and Waris, T. (1977). Experientia 33, 1363-1367. Krisch, B. (1974). Cell Tissue Res. 151, 117-140. Kuhn, C. (1976). In “Lung Cells in Disease” (A. Bouhuys, ed.), pp. 91-108. Elsevier. Amsterdam. Kuhn, C., Callaway. L. A., and Askin, F. B. (1974). J . Ultrustrucf.Res. 49, 387-400. Laszlo, I. (1975). Lab. Pract. 24, 146-150. Lauweryns, J. M., and Cokelaere, M. (1973a). Z . Zellforsch. 145, 521-540. Lauweryns, J. M., and Cokelaere, M. (1973b). Experientia 29, 1384-1386.

THE ENDOCRINE EPIIHELIAL SYSTEM IN T H E LUNG Lauweryns. J . Lauweryns, J. Lauweryns, J. Lauweryns, J. Lauweryns, J. Lauweryns. J. Lauweryns. J. Lauweryns. J. Lauweryns. J.

85

M., and Goddeeris, P. (1975). Am. Rev. Respir. Dis. 111, 469-476. M.. and Liebens, M. (1977). Experientiu 33, 1510-1511. M.. and Peuskens. J. C. (1969). Life Sci. 8, 577-585. M., and Peuskens. J. C. (1972). Anur. Rec. 172, 471482. M.. and Van Lommel, A. (1983). Experientiu 39, 1123-1 124. M., Peuskens, J . C., and Cokelaere, M. (1970). Life Sci. 9, 1417-1429. M.. Cokelaere, M., and Theunynck, P. (1972). Z. ZellJorsch. 135, 569-592. M.. Cokelaere, M., and Theunynck, P. (1973). Science 180, 410-413. M.. Cokelaere, M., Theunynck, P., and Deleersnyder, M. (1974). Chest 65,

22s-29s.

Lauweryns. J . M., Cokelaere, M., Deleersnyder, M., and Liebens, M. (1977). Cell Tissrie Res. 182, 425-440. Lauweryns. J. M.. Cokelaere, M., Lerut, T., and Theunynck, P. (1978). Cell Tissue Ros. 193, 373-386.

Lauweryns, J. M.. De Bock. V.. Verhofstad, A. A. J., and Steinbusch, H. W. M. (1982). Cell T ~ S S URes. C . 226, 215-223. Lauweryns. J. M.. Van Lommel. A. T., and Dom. R. J. (1985). J . Neurol. Sci. 67, 81-92. Leonhardt,.H. (1976). Cell Tissue Res. 174, 99-108. Loosli. C. G., and Hung. K.-S. (1977). In “Development of the Lung” (W. A. Hodson, ed.). pp. 269-306. Dekker, New York. Lopez. J., Diaz De Rada, 0.. Sesma. P., and Vazquez, J. J. (1983). Anur. Rec. 205, 465470.

McDonald. D. M.. and Mitchell, R. A. (1975). J. Nerirocvtol. 4, 177-230. McDonald, T. J.. Jornvall. H., Nilsson, G . , Vagne, M., Ghatei, M., Bloom, S. R., and Mutt, V. (1979). Biocliem. Biophys. Res. Commun. 90, 227-233. McDowell. E. M.. Barrett, L. A., and Trump, B. F. (1976). Luh. Invesr. 34, 202-206. Machado. A. B. M. (1971). Prog. Bruin Res. 34, 171-185. McLean. J. R., and Burnstock. G. (1967a). Comp. Biochem. Physiol. 22, 767-773. McLean, J. R., and Burnstock, G. (1967b). Cornp. Biochem. Physiol. 22, 809-813. Mann. S. P. (1971). Hisrochem. J. 3, 319-331. Marangos. P. J.. and Schmechel. D. (1980). In “Essays in Neurochemistry and Neuropharmacology” (M. B. H. Youdim. W. Lovenberg, D. F. Sharman, and J . R. Lagnado, eds.), Vol. 4, pp. 21 1-247. Wiley. New York. Marangos. P. J., Goodwin. F. K.. Parma, A.. Lauter. C., and Trams. E. (1978). Bruin Res. 145, 49-58.

Marangos, P. J., Schmechel, D.. Zis, A. P., and Goodwin, F. K. (1979). B i d . Psvcl~iutn, 14, 563-579.

Marangos, P. J.. Campbell, 1. C.. Schmechel, D. E., Murphy. D. L., and Goodwin. F. K. (1980). 1.Neurochem. 34, 1254-1258. Marangos. P. J., Schmechel. D. E.. and Oertel, W. H. (1981). In “Gut Hormones” (S. R. Bloom and J. M. Polak, eds.). pp. 101-106. Livingstone, Edinburgh. Marchevsky. A. M., and Kleinerman. J. (1982). Arch. Parho/. Luh. Med. 106, 400-403. Masson, P. (19141. C. R. Acud. Sci. (Puris) 158, 59-61. Matthiessen. D., Mollmann. H.. and Knoche. H. (1973). 2. Zel/for.sch. 138, 133-154. Memoli. V. A.. Linnoila. I.. Warren, W. H.. Rios-Dalenz. J.. and Gould, V. E. (1983). Lob. Invi~st.48, 57A.

Meyer. J.. and Bencosme. S. A. (1965). Rev. Cun. B i d . 24, 179-205. Moosavi. H.. Smith, P.. and Heath, D. (1973). Thorux 28, 729-741. Morikawa, Y.. Donahoe. P. K.. and Hendren. W. H. (1978a). Dev. B i d . 65, 541-546. Morikawa. Y.. Donahoe. P. K.. and Hendren, W. H. (1978b). J . Pediutr. Surg. 13, 65366 I .

86

D. W. SCHEUERMANN

Munger, B. L. (1971). In "Handbook of Sensory Physiology" (W. R. Loewenstein, ed.), Vol. I , pp. 523-557. Springer-Verlag. Berlin and New York. Niki, A., Niki. H.. Miwa, I., and Okuda, J. (1974). Science 186, 150-151. Osborne, M. P., and Butler, P. J. (1975). Nature (London) 254, 701-703. Owman, C., Hlkanson, R.. and Sundler, F. (1973). Fed. Proc. Fed. A m . Soc. Exp. B i d . 32, 1785-1791.

Pack, R. J.. and Widdicombe, J. G. (1984). Eur. J . Respir. Dis. 65, 559-578. Palisano, J. R., and Kleinerman, J. (1980). Thorax 35, 363-370. Palkama, A. (1967). Ann. Med. Exp. Biol. Fenn. 45, 295-306. Papka, R. E. (1975). Cell Tissue Res. 162, 185-194. Papka, R. E. (1980). Adv. Eiochem. Psvchopharmocol. 25, 241-247. Pearsall. A. D., Hoyt, R. F., Jr.. and Sorokin, S. P. (1985). Anrit. R e c . 212, 132-142. Pearse, A. G. E. (1966). Vet. Rec. 79, 587-590. Pearse, A. G. E. (1968). Proc. R . Soc. London Ser. E 170, 71-80. Pearse. A. G. E. (1969). J. Historhem. Cytochem. 17, 303-313. Pearse. A. G. E. (1976). In "Chromaffin, Enterochromaffin and Related Cells" (R. E. Coupland and T. Fujita, eds.), pp. 147-154. Elsevier, Amsterdam. Pearse, A. G. E. (1977). Med. B i d . 55, 115-125. Pearse, A. G. E., Coulling, I., Weavers, B., and Friesen. S. (1970). Gut 11, 649-658. Penttila, A. (1966). Acta Physiol. Scctnd. Suppl. 281, 1-77. Penttill, A. (1967). Histochemie I I , 185-194. Pickel, V. M., Reis, D. J., Marangos, P. J., and Zomzely-Neurath, C. (1976). Brcrin Rev. 105, 184-187.

Polak, J. M. (1983). Actu Endocrinol. (Copenhugen) Suppl. 256, 42. Polak, J. M.,and Bloom, S. R. (1982). Exp. Lung Res. 3, 313-328. Polak, J. M., and Bloom, S. R. (1984). In "The Endocrine Lung in Health and Disease" (K. L. Becker and A. F. Gazdar, eds.), pp. 300-327. Saunders. Philadelphia, Pennsylvania. Price, J., Penman, E.. Bourne, G. L., and Rees, L. H. (1983). Regid. Pept. 7, 315-322. Ratzenhofer, M. (1966a). Mikroskopie 21, 57-64. Ratzenhofer, M. (1966b). Klin. Wochenschr. 44, 109-1 15. Ratzenhofer, M., and Leb, D. (1965). Z. Zel/forsch. 67, 113-150. Redick, M. L., and Hung, K.-S. (1984). CelI Tissue R e s . 238, 583-587. Rees, P. M. (1967). J. C o m p . Neirrol. 131, 517-548. Reese, T. S., and Shepherd, G . M. (1972). In "Structure and Function of Synapses" ( G . D. Pappas and D. P. Purpura, eds.), pp. 121-136. Raven, New York. Reinhold, C., and Hartwig, H . 4 . (1982). Erain Res. Bull. 9, 97-105. Rennels, E. G. (1964). A m . J . Anal. 114, 71-91. Richardson, J. B. ( 1979). A m . Rev. Respir. Dis. 119, 785-802. Rinne. U. K. (1966). Z. Zellforsch. 74, 98-122. Ritzen, M. (1966). Exp. Cell R e s . 45, 178-194. Rogers, D. C., and Haller, C. J. (1978). CelI Tissue R e s . 195, 395410. Rogers, D. C., and Haller, C. J. (1980). Cell Tissue Res. 209, 485498. Rosan, R. C., and Lauweryns, J. M. (1971). Nature (London) 232, 60-61. Said, S. I. (1984). In "The Endocrine Lung in Health and Disease" (K. L. Becker and A. F. Gazdar. eds.), pp. 267-276. Saunders. Philadelphia, Pennsylvania. Sarikas, S. N., Hoyt. R. F., Jr.. and Sorokin, S. P. (1982). Anat. R e c . 202, 167A-168A. Sarikas, S. N., Hoyt, R. F.. Jr.. and Sorokin, S. P. (1985a). Anat. R e c . 213, 396-409. Sarikas. S. N., Hoyt, R. F., Jr., and Sorokin, S. P. (1985b). Anut. Rec. 213, 410-420. Schafer, D.. Bock, P. Acker, H., and Fisher, M. (1973). Arzneim. F o r d . 23, 1613-1614. Scheuermann, D. W. (1984). Front. Horm. Res. 12, 108-1 17.

THE ENDOCRlNE EPITHELIAL SYSTEM IN T H E LUNG

87

Scheuermann, D. W.. and De Groodt-Lasseel. M. H. A. (1983). Anut. Rec. 205, 175A. Scheuermann. D. W.. De Groodt-Lasseel, M. H. A., Stilman, C., and Meisters, M.-L. (1983a). Cell Tissue Res. 234, 249-269. Scheuermann. D. W., Meisters, M.-L., and De Groodt-Lasseel, M. H. A. (1983b). Verh. Anrit. Ges. 77, 671-673. Scheuermann. D. W.. De Groodt-Lasseel. M. H. A,, Meisters. M.-L., and Stilman. C. (1984a). Verli. Anat. Ges. 78, 475-476. Scheuermann. D. W.. De Groodt-Lasseel, M. H. A., and Stilman, C. (1984b). A m . J . Anot. 171, 377-399. Scheuerrnann, D. W.. De Groodt-Lasseel, M. H. A,, Stilman. C., and Meisters, M.-L. (1984~). Verli. Anur. Ges. 78, 473-474. Scheuermann, D. W.. Stilman, C., and De Groodt-Lasseel, M. H. A. (1985). Verb. Anut. Ges. 79, 299-30 I. Schmechel. D.. Marangos. P. J., Zis, A. P., Brightman, M., and Goodwin, F. K. (1978a). Science 199, 313-315. Schmechel. D.. Marangos, P. J., and Brightman, M. (197%). Nature (London) 276, 834836. Scrutton, M . C.. and Utter, M. F. (1968). Annu. Rev. Eiochem. 37, 249-302. Seitelberger. F. ( 1971). Actu Neuropu/hol. Sicppl. 5, 17-29. Sheppard. M. N.. Johnson. N. F.. Cole, G. A,. Bloom, S. R., Marangos, P. J.. and Polak, J. M. (1982). His/ochemistr?, 74, 505-513. Sheppard. M. N.. Marangas. P. J., Bloom, S. R.. and Polak, J. M. (1984). Life Sci. 34, 265271. Shul'ga, M. I. (1965). Fed. Pro(.. Fed. A m . Soc. Exp. Eiol. 24 (Transl. Suppl.), T605-T607. Sidhu, G. S. (1979). Am. J . Put/iol. 96, 5-17. Smith, P.. Heath, D.. and Moosavi. H. (1974). Thorax 29, 147-163. Solcia. E.. Vassallo. G., and Capella, C. (1968). Stuin Teclinol. 43, 257-263. Solcia. E., Capella, C., and Vassallo, G. (1969). Histochemie 20, 116-126. Solcia. E.. Capella. C., Vassallo. G.. and Buffa, R. (1975). In/. Rev. Cytol. 42, 223-286. Sonstegard, K. S . . Cutz. E., and Wong, V. (1976). Am. J. Anat. 147, 357-374. Sonstegard, K.. Wong. V., and Cutz, E. (1979). Cell Tissue Res. 199, 159-170. Sonstegard, K. S., Mailman, R. B.. Cheek, J. M., Tomlin, T. E., and DiAugustine. R. P. (1982). Exp. Lung Res. 3, 349-377. Sorokin. S. P.. and Hoyt, R . F., Jr. (1978). Anut. Rec. 192, 245-260. Sorokin. S. P., Hoyt, R. F., Jr., and Grant, M. M. (1982). Exp. Licng Res. 3, 237-259. Sorokin. S. P., Hoyt. R. F., Jr., and Pearsall, A. D. (1983). A m . Rev. Respir. Dis. 128, S 2 6 S 3I . Sotelo. C.. and Palay. S. L. (1968). J . Cell Eiol. 36, 151-179. Stahlman, M. T., and Gray, M. E. (1984). Lab. Invest. 51, 449-463. Stahlman, M . T., Gray, M. E.. Kasselberg. A. G., and Orth, D. N . (1982). Pediutr. Res. 16, 362A. Stahlman. M. T.. Kasselberg, A. G., Orth, D. N . , and Gray, M. E. (1985). Lab. Inves/. 52, 52-60. Sundler. F., Hlkanson, R.,Loren, I.. and Lundquist, 1. (1980). Invest. Cell Pothol. 3, 87103. Taira. K., and Shibasaki, S. (1978). Arch. Histol. Jpn. 41, 351-366. Takahashi, S . , and Yui, R. (1983). Eiomed. Res. 4, 315-320. Tapia. F. J . , Polak. J. M., Barbosa, A . J. A,, Bloom, S. R., Marangos, P. J.. Dermody. C.. and Pearse. A. G. E. (1981). Lancet 1, 808-811. Tateishi. R. (1973). Arch. Parhol. 96, 198-202.

88

D. W. SCHEUERMANN

Taylor, 1. M., and Smith, R. B. (1971). Biol. Neonate 18, 193-202. Taylor, W. (1977). J. Pmhol. 122, 137-144. Taylor. W. (1982). In “Cellular Biology of the Lung” (G. Cumming and G. Bonsignore. eds.). pp. 381404. Plenum. New York. Terzakis, J . A,. Sommers, S . C., and Andersson, B. (1972). Lab. Invest. 26, 127-132. Thompson. S . A.. and Gosling, J . A. (1976). Cell Tissrte R e s . 170, 539-548. Toker. C. (1966). CcinciJr(Amsterd(rrri) 19, 1943-1948. Track, N . S., and Cutz, E. (1982). LiJe Sci. 30, 1.553-1556. Tsutsumi, Y., Osamura, R. Y ., Watanabe, K., and Yanaihara, N. (1983a). Lcrh. f n i v s r . 48, 623-632. Tsutsumi, Y.. Osamura. R. Y.. Watanabe, K., and Yanaihara. N. (1983b). Virchows Arch. A : Pathol. A n a l . 400, 163-171. Verna. A. (1973). J. Microsc. (Pciris) 16, 299-308. Verna, A. (1979). I n / . Rail. Cylol. 60, 271-330. Vigh, B., and Vigh-Teichmann. I. (1973). I n / . R e v . Cylol. 35, 189-251. Vigh-Teichmann, I . , Vigh, B.. and Aros, B. (1976a). Cell Tissrte R e s . 165, 397-413. Vigh-Teichmann, 1.. Vigh. B.. and Aros, B. (1976b). Cell Tissue R e s . 174, 139-160. Vigh-Teichmann, I..Vigh. B.. Rohlich, P., and Olsson, R. (1980). In “Circulatory and Developmental Aspects of Brain Metabolism” (M. Spalz. B. B. Mrsulja, L. M. Rakic. and W. D. Lust, eds.), pp. 415428. Plenum, New York. Walsh. C.. and McLelland, J . (1974). Cell Tissrte Res. 153, 269-276. Wasano. K . (1977). Arch. Hisrol. Jpn. 40 (Suppl.), 207-219. Wasano, K., and Yamamoto, T. (1978). Cell Ti.s.site Res. 193, 201-209. Wasano, K . , and Yamamoto, T. (1979). Cell Tissue Res. 201, 197-205. Wasano, K . , and Yamamoto, T. (1981). C i 4 Tissue Res. 216, 481490. Welsch. U.. and Pearse. A. G. E. (1969). Hislochemir 17, 1-10. Wetzstein, R. (1957). Z. Zel/fi)rsrh. 46, 517-576. Wharton, J., Polak, J. M.. Bloom. S. R., Ghatei. M. A.. Solcia. E., Brown, M. R.. and Pearse, A. G. E. (1978). Nu/irri* (London) 273, 769-770. Wharton. J., Polak. J. M., Cole, G. A., Marangos, P. J., and Pearse, A. G. E. (1981). J. Hislochmi. Cyrochcvn. 29, 1359-1364. Will, J. A.. Rademakers, A., and Dayer. A. M. (1985). Fed. Proc. Fed. A m . So(..Exp. Biol. 44. 917. Wood. J . G. (1967). Ancit. R e ( , . 157, 343-344. Yamauchi, A., Fujimaki, Y.. and Yokota, R. (1975a). J. Ulfrustritcz. Res. 50, 47-57. Yamauchi, A.. Yokota, R.. and Fujimaki, Y. (1975b). Anol. R w . 181, 195-210. Yanaihara, N.. Yanaihara, C.. Mochizuki. T., Iwahara, K., Fujita. T., and Iwanaga. T. (1981). Pi.piidr.s N . Y . 2 (Suppl.), 185-191. Yoshizaki. K., de Bock. V.. and Solomon. S. (1984). Life Sci. 34, 835-843. Yui. R.. Yamada, Y., Kayamori, R.. and Fujita. T. (1981). Biomed. R r s . 2, 208-216. Zabel. M. (1984). Hi.sIoi~liiwi.J. 16, 1265-1272. Zabel. M.. and SchBfer. H. (1985). Hi.stoc~hc~mistrv 83, 77-80.

INTEKNATIONAL KI;.VIEW 0 1 : CYTOLOGY. VOI. IM

Intrinsic Nerve Plexus of Mammalian Heart: Morphological Basis of Cardiac Rhythmical Activity?' JOSEF MORAVECA N D MIREILLEMORAVEC Utiitk dc~Patliologie Ctrrtliovtrsc.rrluirt~tie I'ltistitrct Nationcil de In Stititc; ot tte Itr RechivchtJ MPdicwle ( I N S E R M ) , Hfipittrl LPon Bertiurd, 94456 LirneilBrPvmnes CPdex, Frcinw

I. Introduction

The purpose of the rhythmical beats of the heart remained unexplained until the early seventeenth century, when Harvey (1628) made his important discovery that the heart beats in order to circulate the blood (cf. Noble, 1979). However, even after this decisive step, the mechanism of the initiation of cardiac beats remained obscure. For several centuries, the inherent rhythmicalactivityof the heart has beenconsidered as amajor example of vital forces. At the same time, different mechanistic conceptions have been postulated. Each investigator tried to use the knowledge of physics and chemistry available to him in order to assess the process of heartbeating. At present, the heartbeat is considered an electromechanical process initiated by the nodal tissue (Noble, 1979) and distributed through the ventricular myocardium via the conduction system. The microanatomy of the latter has been progressively elaborated since the original descriptions of Purkyne (1843, Aschoff (1910). Tawara (1906). and Keith and Flack (1907). More recently, various components of the intracardiac conduction tissue have been thoroughly examined by modern electrophysiological methods (Weidman, 1967; Noble and Tsien, 1968; Rougier ~t d., 1969; Winegrad. 1979; Coraboeuf, 1982). This allowed the identification of different transmembrane currents, which are now considered as responsible for the repetitive depolarizations of pacemaker cells and for the normal intracardiac conduction. The fact that some of the specialized cells can continue to fire even upon their isolation (Cranefield, 1978; Noble, 1979) led to the contention that most of their rhythmical activity depends on kinetic properties of different ionic channels, which were shown to coexist in their cell membranes (Noble and Tsien, 1968; Rougier et d., 1969; Coraboeuf, 1982). 'This work i\ a tribute t o J. E. Purkyne on the occasion of the forthcoming bicentennial of his hirth (December 17. 1787).

89 C'opyrlght (11 19x7 hy Av;idmuc Pru\\. Inc. All right\ of reprodoclion in any form rewrved.

90

JOSEF MORAVEC AND MlRElLLE MORAVEC

However, as Cranefield (1983) pointed out, it is not certain that the electrical behavior of disaggregated cells, which are used for the voltageclamp and voltage-patch studies, is wholly comparable to those that can be observed in situ. The cells of the conduction system of the heart were shown to constitute complicated networks in which operate not only several types of specialized cells (James, 1973; Cranefield, 1983), but also connective elements (Thornell et al., 1976), nerve fibers (Bojsen-Mgiller and Tranum-Jensen, 1972), and ganglionic cells (Moravec et af., 1985; Moravec and Moravec, 1984). The existence of multiple intercellular interactions can thus be expected at different levels of the conduction system. This would explain why the resting potential, source impedance, and intercellular coupling vary from one part of the heart to the other (Cranefield, 1983) and why some of the electrical properties of one single cell can change upon its isolation from the neighboring structures (Mendez et al., 1969). It became evident that one part of the above pluricellular network can impose a decisive load on its other parts. The nature of the latter might be very variable. Apart from direct electrical stimulation via the electrical synapses or through local membrane currents, the electrical properties of specialized cells can be modulated by mechanical stretch (Brooks and Lu, 1972; Irisawa, 1978) and by neurotransmitter release from intracardiac nerves (Sarnoff and Mitchell, 1962; Jacobowitz, 1967; Randall, 1976) as well as from nonneuronal storage sites (Lignon and Le Douarin, 1978; Pollack, 1978). The appropriate interaction between different components of the conduction system and a permanent feedback from the surrounding structures seems to be necessary for the initiation and propagation of the excitation wave to proceed optimally (Cranefield, 1983). The contribution of the nervous system to the control of cardiac pacemaker is believed to vary from one animal species to another. The crustacean hearts seem to be entirely neurogenic (Irisawa, 1978). In this case, the triggering signal is provided by a cell-driven oscillator of the cardiac ganglion (Selverston and Moulins, 1985). A similar situation was also shown to occur in the leech, where rhythmical depolarizations of the heart are under the control of an oscillatory network composed of a group of ganglionic cells interconnected by reciprocal inhibitory synapses (Stent et al., 1979). In contrast, the generation of rhythmical heartbeats in mammals has been considered essentially myogenic (Irisawa, 1978), resulting from slow diastolic depolarizations of pacemaker cells (Noble and Tsien, 1968) or from subthreshold oscillations of their membrane potentials (Irisawa, 1978; Cranefield, 1983; Goto, 1986). The role of the nervous system is often restricted to an external modulation of the inherent muscular pacemaker

INTRINSIC N E R V E P L E X U S OF MAMMALIAN H E A R T

91

(Randall, 1976; Levy and Martin, 1979). This situation results from the fact that current knowledge of cardiac innervation of the mammalian heart has been derived from classical, physiological, and microanatomical observations (Dogiel, 1882; Gaskell, 1886; Langley and Orbelli, 1910; Nonidez, 1939; Kuntz, 1953; Randall, 1976). According to these studies, the pacemaker area of the heart receives abundant cholinergic and adrenergic postganglionics originating from the extracardiac ganglia (Randall, 1976; Levy and Martin, 1979; Armour and Hopkins, 1984). However, apart from this extrinsic component, abundant intramural ganglionic cells can be identified in light and electronic microscopy, mainly at the level of the interatrial septum and all along the sulcus terminalis (Kuntz, 1953; Jacobowitz et al., 1967; Abraham, 1969; Ellison and Hibbs, 1976; Papka, 1976; Rossi, 1978). In rat and in other rodents, some of these ganglionic cells seem to be structurally associated with different portions of the intracardiac conduction system (Nielsen and Owman, 1968; Weihe et al., 1984; Moravec and Moravec, 1984). This finding may suggest that, also in mammals, some of the electrophysiological properties of the intrdcardiac conductive tissue may result from the intimate cooperation between specialized cells and the intrinsic nervous components. In other words, the conduction system of the heart should no longer be regarded as a specialized muscular tissue, but rather as a highly differentiated neuromuscular organ (Oppenheimer and Oppenheimer, 1912; Wensing, 1965; Bojsen-MZller and Tranum-Jensen, 1972; Irisawa, 1978; Moravec and Moravec, 1984).

11. Autonomic Innervation of the Heart

A. GENERAL ORGANIZATION, ANATOMICAL, A N D PHYSIOLOGICAL EVIDENCES FOR DUALINNERVATION According to classical physiological (Sarnoff and Mitchell, 1962; Randall, 1976; Armour and Hopkins, 1984) and histochemical studies (Jacobowitz et a / . , 1967; Ehinger et al., 1968; Abraham, 19691, the autonomic nerve supply to the heart of different vertebrates is still considered in terms of Dale’s principle, i.e., in terms of a dual (adrenergic and cholinergic) innervation (Dale, 1953; Yamauchi, 1973; Levy and Martin, 1979; Armour and Hopkins, 1984). This separation of the autonomic nervous system into its two major divisions is based on the original observations of Gaskell and Langley (Gaskell, 1886; Langley. 1921) concerning the anatomy and pharmacology of the autonomic nervous system. However, the possibility of the existence of a third category of the autonomic system has been

92

JOSEF MORAVEC A N D MIREILLE MORAVEC

suggested as early as in 1898, when Langley introduced the term “intrinsic nerve system.” He also suggested that peripheral nerves act on different tissue “receptors” (cf. Brooks, 1981). Today this field has been reopened; new transmitters and new receptors have been identified (Burnstock, 1969; Gershon, 1977; Lundberg et al., 1984), and the role of the intramural plexuses is now being reexamined (Burnstock and Bell, 1974; Brooks, 1978; Leranth and Unguary, 1980; Nozdrachev and Vataev, 1981; Wood, 1981; Weihe et al., 1984). It seems to be accepted that sympathetic and parasympathetic fibers feed into the intramural plexuses of peripheral organs which, per se, can mediate reflexlike reactions and exert a local control (Hillarp, 1959; Yamauchi, 1%9; Selverston et al., 1976; Brooks, 1981; Wood, 1981). Some authors do not hesitate to extend these new concepts also to the field of regulatory mechanisms of cardiovascular function (Brooks and Lange, 1977; Priola et al., 1977; Brooks, 1981; Drake-Holland et al., 1982; Weihe et al., 1984). B.

SYMPATHETIC

INNERVATION

OF THE

MAMMALIAN HEART

I . Efferent Sympathetic Pathways to the Heart The cell bodies of the preganglionic sympathetic neurons are located in the intermediolateral columns of the upper eight thoracic segments of the spinal cord (Henri and Calaresu, 1972). The preganglionic fibers emerge from the spinal cord through the white rami communicates of the first six thoracic segments (Gaskell, 1886; Randall et al., 1957; Seagdrd et al., 1978; Levy and Martin, 1979) and enter the paravertebral chain of sympathetic ganglia. In some animals, the preganglionic sympathetic fibers can also arise from the spinal cord via the ventral roots of the last two cervical segments (Levy and Martin, 1979). Most of these sympathetic preganglionics transit throughout the ipsilateral stellate ganglion and pass through the ventral or dorsal limbs of the ansa subclavia to the inferior cervical ganglion. The synapses between the preganglionic and postganglionic neurons are believed to take place in the cervical and upper thoracic ganglia including the stellate ganglion (Kuntz, 1953; Wacksman el al., 1969; Wechsler et al., 1969; Tollack et al., 1971; Levy and Martin, 1979). Only few sympathetic spinal fibers reach the heart directly without interruption (Kunz, 1953; Brown, 1967). The presence of these sporadic preganglionic fibers may have a serious impact on the control of cardiac function: they can be expected to innervate a distinct population of the intracardiac ganglion cells. A bulk of these noninterrupted presynaptic fibers (large myelinated fibers of ACYcategory) has been identified in the ventrolateral cardiac nerve which supplies selectively the atrioventricular

I N T R I N S I C N E R V E PLEXUS OF MAMMALIAN H E A R T

93

junction of the dog heart (Seagard et ul., 1978). It is not known whether a similar situation also prevails in other species, but it should be noted that a recent electron microscopic analysis of serial sections of the interatrial septum of the rat heart (Moravec and Moravec, 1984; Moravec et ul., unpublished data) has revealed the presence of adrenergic neurons in this area. The traditional affirmations concerning the absence of sympathetic ganglia in the terminal nerve plexus of the mammalian heart (Hirsch et al., 1963; Jacobowitz, 1967; Yamauchi, 1973; Randall, 1976; Levy and Martin, 1979; Armour and Hopkins, 1984) will possibly need a detailed verification. 2 . Afferent Symputhetic Fibers from the Heart In addition to the above efferent nerve fibers, the sympathetic nerves of the heart also contain afferent nerve fibers which, like the corresponding preganglionics, are the components of the first to the sixth thoracic nerves (Kuntz, 1953). At least two types of afferent fibers can be distinguished by means of studies of conduction velocities and according to fiber diameters (Armour rt al., 1975; Seagard et al., 1978): myelinated A6 fibers (2-5 k m in diameter) and unmyelinated C fibers (below 1 km). Reflexes carried by these fibers may play a role in the fine regulation of cardiac performance (Uchida, 1979; Malliani, 1979) and in the control of cardiac rhythm (Shepherd, 1985). The possible involvement of these fibers in the transmission of anginal pain has been suggested by curative effects of the sympathectomy or stellate ganglionectomy (Uchida et al., 1971). The majority of the sympathetic reflexes is centrally coordinated. However, the existence of short reflex loops, mediated by either mediastinal (Armour and Hopkins, 1984) or stellate ganglia (Bosnjak et al., 1982), is not excluded. In addition, the existence of the intramural axonal reflexes should also be considered (Leranth and Unguary, 1980; Brooks, 1981).

c. PARASYMPATHETIC I N N E R V A T I O N OF THE MAMMALIANHEART I. Efferent Purusymputhetic Pathways to the Heart The parasympathetic efferent innervation of vertebrate heart involves two cholinergic synapses arranged in series. The presynaptic nerve fibers are provided by the right and left vagi (Randall, 1976). These fibers originate from the ipsilateral regions of the brain stem, e.g., from nucleus ambiguus and dorsal motor nucleus of the vagus (Weiss and Priola, 1972; Levy and Martin, 1979). Some cardioinhibitory neurons can also be detected in the intermediate zone of the dog and some other species so far examined (Cabot and Cohen, 1980; Armour and Hopkins, 1984). They

94

JOSEF MORAVEC A N D MIREILLE MORAVEC

synapse on intracardiac ganglionic cells that are mostly associated with the subepicardial layer of the posterior aspect of the atria (Kuntz, 1953; Loffelholz, 1982; Moravec and Moravec, 1984). The cardiac postganglionic nerve fibers differ from motor nerves supplying skeletal muscles in that they terminate by extensive varicose end fibers, ramifying in the effector organ, from which acetylcholine is released by action potentials invading en passant one varicosity after the other (Loffelholz, 1982). As to the neuromuscular cholinergic synapses, they have been considered as quite rudimentary. Some authors (Boeke, 1936; Abraham, 1969) have even concluded that they are absent. Only recently, small cholinergic synapses similar to those of skeletal muscle spindles could be demonstrated in the electron microscope material, at least at the level of the atrioventricular junction (Moravec-Mochet et al., 1977). In some cases, the terminal cholinergic fibers run parallel with the preterminal adrenergic fibers of the autonomic ground plexus (Hillarp, 1959; Ehinger et ul., 19701, and, occasionally, cholinergic and adrenergic fibers can share the same Schwann cell sheath. This led to the contention that some of the vagosympathetic interactions (Levy and Martin, 1979) may result from this particular structural configuration (Napolitano et al., 1965; Ehinger et al., 1970; Brooks, 1981). However, direct proof that the two fibers are really functionally interacting is lacking. 2 . Afierent Parasympathetic Fibers from the Heart The presence of afferent fibers in nearly all vagal branches to the heart was suggested by the electrophysiological studies (Nettleship, 1936; Sleight and Widdicombe, 1965; Oberg and Thoren, 1972; Coleridge et NI., 1973). They could be subdivided into two categories: medullated (myelinated) fibers and unmedullated (unmyelinated) C fibers. The receptors associated with the myelinated fibers predominate in the atria (Shepherd, 1985). They have been also described in the right and left ventricles and in coronary vessels (Sleight, 1979). However, they are rather sparse and their reflex effects unknown. The majority of ventricular receptors with unmyelinated C fibers is confined to the left ventricle (Coleridge et ul., 1964). They are involved in depressor reflexes such as the pronounced bradycardia elicited by left ventricular distension (Thoren, 1979) and in body fluid volume regulation (Oberg, 1979). It has been also suggested that these receptors are essential for the circulatory response to acute exercise and to other stressful conditions. Their stimulation may prevent an excessive tachycardia normally occurring at high levels of circulating catecholamines. In contrast to the profusion of data concerning the physiological role of vagal cardiac mechanoreceptors, relatively little is known about their structure. Their identification in the light microscope (methylene blue or

I N T R I N S I C N E R V E PLEXUS OF MAMMALIAN HEART

95

silver staining) proved not to be easy. Most of the structures described (i.e., compact encapsulated endings and diffuse unencapsulated endings) are derived from myelinated axons (Nettleship, 1936; Nonidez, 1939; Khabarova, 1961; Abraham, 1969; Floyd, 1979). The contribution, if any, from unmyelinated fibers is not known. The latter are believed to subserve the diffusely distributed end-nets (Floyd, 1979; Shepherd, 1985).

D. FINESTRUCTURE OF INTKACARDIAC

MECHANORECEPTORS

Our knowledge of the fine structure of cardiac receptors supplied by either sympathetic or parasympathetic fibers is also rather sparse (TranumJensen, 1975; Yamauchi, 1979; Moravec and Moravec, 1982). As pointed out by Tranum-Jensen (l979), the uncertainties and limitations relating to the ultrastructural identification of cardiac receptors are determined by the arbitrary choice of tissue samples, selected on the basis of the preliminary light microscopic examination (Floyd, 1979). The detailed description of the intracardiac sensory corpuscles and, mainly, the understanding of their relations to the surrounding structures are not possible without the systematic use of serial semithin and thin sections (Thaemert, 1970; Moravec-Mochet et al., 1977; Moravec and Moravec, 1982). Until present, the existence of two types of intracardiac mechanoreceptors has been suggested by electron microscopy studies. ( I ) Unencapsulated end organs (baroreceptors), associated with 5- and 10-pm-thick myelinated fibers, have been described in the subendocardium of the right and left atria of the pig (Tranum-Jensen, 1979) and rat hearts (Moravec and Moravec, 1982). These terminals were covered by a basement membrane and partly devoid of Schwann cell sheath. Their cytoplasm contained small filiform mitochondria (0.2 x 1.0 pm), bundles of microfilaments, and numerous vesicles of the endoplasmic reticulum associated with abundant glycogen granules. Similar structures were also described in the rat aorta (Yamauchi, 1979) and in human atria (Chiba and Yamauchi, 1970). (2) Muscular spindlelike structures, previously suggested by Lawrentjew and Gurritsch-Lasowskaja (1930), have been described in the rat atrioventricular junction (Moravec-Mochet et al., 1977; Moravec and Moravec, 1982). In this case, agglomerates of nodal cells, surrounded by a common basal lamina and elaborated connective sheath, were innervated by efferent cholinergic fibers terminated by small en grappe synapses. Other types of spiral fibers wrapped the same nodal cell corpuscles and interpenetrated into deep invaginations of muscular cell membranes, in a manner which was similar to the afferent fibers of skeletal muscle spindles (Landon, 1972; Ovalle, 1972). The absence of any basal lamina in the intermembrane space (less than 20 nm) as well as the characteristic cytoplasmic content

96

JOSEF MORAVEC AND MIREILLE MORAVEC

of these relatively large nerve endings (numerous small mitochondria, abundant glycogen granules, and few small vesicles), strongly suggested that the above spiral fibers were sensory in nature. Similar coiled endings were formerly described using optical microscopy with samples of right and left ventricular myocardium (King, 1939; Plechkova, 1948; Khabarova, 1961).According to a more recent study of silverstained serial sections of the entire rat heart (Moravec et ul., 1985). these spindlelike structures are selectively located in the upper portion of the interventricular septum and all along the specialized tissue of the atrioventricular junction, namely, in the reticular portion of the atrioventricular node and in proximal portions of the right and left branches (Moravec et a / . , 1985).This association of the sensory innervation with the specialized tissue may have a functional significance (Brooks and Lu. 1972; Pollack, 1974), which we shall discuss in the following sections.

111. Intracardiac Ganglionic Cells

As the vagus nerves enter the chest, they pass in the vicinity of the cervical sympathetic ganglia. The postganglionic adrenergic and preganglionic cholinergic fibers then course together to the heart, sharing some of their satellite cells (Nonidez, 1939; Kuntz, 1953; Levy and Martin, 1979). They ramify in the pericardiac nerve plexus located between the aortic arch and pulmonary veins and supply both the heart and great vessels (Nonidez, 1939; Kuntz, 1953; Randall et al., 1972; Goldberg and Randall, 1973; Levy and Martin, 1979). The microanatomical description of these mixed cardiac nerves was considerably improved by the recent physiological studies of Randall and colleagues (Hageman et al., 1975; Hondenghem et al., 1975; Goldberg and Randall, 1973; Randall, 19761, who studied the respective contributions of sympathetic and parasympathetic fibers to different cardiac nerves of the dog heart. They also established the cartography of peripheral projections of the individual cardiac branches to the epicardial layers of the left and right cardiac cavities. According to these studies, the proportions of sympathetic and parasympathetic efferents are variable in different left and right branches. So are the respective contributions of the left- and right-side nerves to the innervation of different cardiac structures, such as the sinoatrial node and the atrioventricular junction. It has been demonstrated that, for embryological reasons (Taylor, 19771, the pacemaker area receives the branches from the right vagosympathetic trunk, while the atrioventricular junction is innervated quasi-entirely from the left side (West and Toda, 1967; lrisawa et al., 1971; Hondenghem et al., 1975). However, species differences in the overlap and in the specificity of the projection areas of cardiac nerves

INTRINSIC N E R V E PLEXUS OF M A M M A L I A N H E A R T

97

(Burnstock, 1969; Yamauchi, 1969; Cabot and Cohen, 1980; O’Shea and Evans, 1985) are not negligible. A. EVIDENCE FOR INTRINSIC GANGLIONIC CELLS One of the most significant contributions of the above physiological studies was the confirmation of the earlier morphological observations (Lawrentjew and Gurwitsch-Lasowskaja, 1930; Vitali, 1937; Field, I95 I ; Hirsch. 1962; Abraham, 1969) concerning the selective innervation of the intracardiac conductive tissue and the adjacent portions of the right atrium. In contrast to the ventricular myocardium which is innervated by sparse individualized nerve fibers, often directed to the coronary vessels (Abraham, 1969; Somlyo, 19731, the sinus node and, mainly, the structures of the atrioventricular junction are literally enmeshed by abundant terminal fibers of the intracardiac nerve plexus (Akkeringa, 1949; Wensing, 1965; Bojsen-MGller and Tranum-Jensen, 1972; Moravec and Moravec, 1984). In addition to the extrinsic nerves, intracardiac ganglionic cells provide the bulk of terminal nerve fibers to both the conduction system and the ventricles. The existence of the latter was recognized during the early decades of the last century (Purkynb, 1845; Ludwig, 1848). Later on, they have been reported in different parts of the heart in both lower species 1963; Mc Mahan and Kuffler, 1970; Taxi, 1976) (Dogiel, 1882; Falck ef d., and mammals, including man (Kuntz, 1953; Rossi, 1955; Khabarova, 1961; Hirsch, 1963, 1970; Abraham, 1969; Papka, 1976). A majority of these ganglionic cells is concentrated in the subepicardium of the base of the heart, around the ostia of great vessels and in the coronary sulcus. where they form more or less delimitated subepicardial ganglia. Isolated ganglionic cells can also be found scattered in the meshes of the terminal nerve plexus located in the wall of the right atrium and in the interatrial septum all along the sulcus terminalis (Nielsen and Owman, 1968; Ehinger et a/., 1968; Ellison and Hibbs, 1976; Papka, 1976). The presence of ganglionic cells in the ventricular myocardium. except for the upper segment of the interventricular septum, has been reported less frequently (Smith, 1971; Bolton, 1976; Loffelholz and Pappano, 1985). According to denervation studies, some of the intracardiac ganglionic cells can survive a long time after cardiac denervation. This can be obtained by surgical interventions, such as mediastinal neural ablation (Jacobowitz r t al., 1967) and cardiac autotransplantation (Napolitano ef al., 1965; Potter et u / . , 1965) or immunologically (Zaimis et a / . , 1970; Gabella, 1976). The selective destruction of the sympathetic component can be also obtained by chronical administrations of guanethidine (Burnstock et [ I / . , 1971) or by a single injection of 6-OH-dopamine (Tranzer and Thoenen, 1967). After most of these interventions, functionally active intrinsic nerve components

98

JOSEF MORAVEC A N D MIREILLE MORAVEC

could be demonstrated both at the level of the atrioventricular junction (Ljima et al., 1974) and in the ventricular myocardium of different species (Priola et al., 1977; Drake-Holland et al., 1982). The persistence of more than one-half of cholinergic and adrenergic postganglionic fibers in the atrial and ventricular walls was also confirmed morphologically (Napolitano et al., 1965; Potter et al., 1965). These data suggest that the majority of postganglionic fibers is derived from the intracardiac ganglion cells, which constitute an intrinsic nervous component similar to the intramural nervous system of Langley (1921). It is now widely accepted that this special division of the autonomic nervous system takes part in the innervation of most of peripheral tissues supplied by the autonomic nerves. According to recent studies, the intramural plexus is composed of a network of terminal nerve fibers and abundant ganglionic cells with surprisingly high synaptic plasticity (Brooks. 1981; Gershon and Erde, 1981; Wood, 1981). Apart from the two classical neurotransmitters (i.e., acetylcholine and noradrenaline). a number of neuromodulators (substance P, vasoactive intestinal peptide, neurotensin, etc.) as well as a series of putative transmitters (ATP, serotonin, etc.) were shown to contribute to its function (Burnstock, 1969; Hokfelt, 1979; Gershon et al., 1981). Some of the latter compounds were recently identified in the ganglion cells of the intrinsic nerve plexus of the guinea pig and rabbit hearts (Crowe and Burnstock, 1982; Weihe et al., 1984). This confirms the sporadic suggestions concerning the functional autonomy of the intracardiac nerve plexus. Experimental evidences in favor of such a hypothesis will be presented in the forthcoming sections. B. REGIONALSPECIALIZATIONS OF THE INTRINSIC NERVEPLEXUS OF THE INTERATRIAL SEPTUM At the first sight, the distribution of the atrial ganglionic cells might seem rather arbitrary. However, after having examined series of silverstained serial sections from a large number of animals, Moravec et al. (1984, 1986) suggested that, at least in the rat, the distribution of the intracardiac ganglia respects a reproducible pattern. In agreement with other authors (Abraham, 1%9; Bojsen-Mgller and Tranum-Jensen, 1972; Papka, FIG. I . (A) Light microscopic view of a small epicardial ganglion adjacent to the insertions of the ascending aorta and superior vena cava (arrow). An agglomerate of chromaftin cells surrounding a capillary can also be seen on the same picture (arrowhead). (From the Am. J . Anal.. 1984, 171, 307-320.) (8) Electron microscopic view of a small unipolar neuron of the epicardial ganglion. Note abundant membrane-bound ribosomes in its cytoplasma as well as the axonal profiles surrounding the cell body. (Arrow) Axonal hillock.

INTRINSIC NERVE PLEXUS OF MAMMALIAN HEART

99

100

JOSEF MORAVEC AND MIREILLE MORAVEC

FIG.2. The epicardial ganglion as seen on a silver-stained paraffin section of a formaldehyde-fixed heart. Note presence of both unipolar and multipolar neurons (arrows).

1976), they regularly found a large epicardial ganglion between the root of the aorta and superior vena cava, i.e., close to the sinoatrial node. Apart from the unipolar neurons, which might be the second ganglion cells of cholinergic paths to the heart (Papka, 1976; Ellison and Hibbs, 1976), sporadic multipolar neurons could also be demonstrated (Figs. 1 and 2). In electron microscopy, most of the neuronal bodies presented abundant Nissl bodies (Fig. I B). Occasionally, neurosecretory-like profiles could also be seen on the same sections. Abundant ganglion cell bodies were also present all along the posterior branch of the sinoatrial ring bundle (SARB) of Bojsen-MZller and Tranum-Jensen ( I 972), which courses parallel to the sulcus terminalis and interconnects the sinus node area with the coronary sinus and posterior part of the atrioventricular node. Some of these nerve cells constitute a discontinuous intraseptal ganglionic lamina (Fig. 3), which sends several nerve branches in a forward direction to the structures of the atrioventricular junction, i.e., the atrioventricular node and bundle of His. Some of the peripheral extensions of these nerve fibers join the nerve plexus surrounding the right and left branches (Fig. 4). These nerve fibers follow a similar course as the previously described middle internodal pathway of James (1963).

FIG.3. ( A ) The section through the ganglionic lamina, which runs parallel to the sulcus terminalis in the interatrial septum. Note paucity of the periganglionic capsule. (From the Am. J. A n u f . . 1984, 171, 307-320.) (B) A ganglion cell found in the vicinity of a large nerve trunk running from the intraseptal ganglionic lamina in a forward direction to the specialized structures of the atrioventricular junction. ( C )An isolated unipolar ganglion cell as revealed by the silver staining in the interventricular septum. (From the “Advances in Myocardiology.” Vol. 6. pp. 13-23. Plenum, 1985.)

102

JOSEF MORAVEC AND MIREILLE MORAVEC

FIG.4. Schema illustrating the relationships between the atrioventricular (AV) specialized tissue and different components of the intrinsic nerve plexus of the interatrial septum; reconstruction from serial semithin sections. IAS, Interatrial septum; SVC, superior vena cava; HB, bundle of His; MV, mitral valve; CS,coronary sinus; TV, tricuspid valve; IVS. muscular interventricular septum; AV, aortic valve; MS, membraneous septum; LB. left branch; RB. right branch; AVN. atrioventricular node; ---,retroaortic ganglion; @) , ganglionic lamina; A, chromaffm cells; @) , glomeruli; 0 , neurosecretory cells: ‘)rr , coiled endings.

In contrast, the anterior branch of the SARB, as previously identified by cholinesterase reaction (Bojsen-Mdller and Tranum-Jensen, 1971, 1972), does not contain any ganglionic cells (Moravec et al., 1985). It is composed of a strand of small muscle cells surrounded by a fine plexus of terminal nerve fibers in a manner which is similar to that encountered in the sinus node and in the accessory atrioventricular node (Anderson, 1972). It would seem that the anterior branch of the SARB, located in the upper portion of the interatrial septum, represents a neuromuscular link between these two structures. From the topographical point of view, the posterior and anterior branches of the SARB follow the courses of the third and first “specialized pathways” for the intraatrial conduction (James, 1963; Viragh and Challice, 1973; Hiraoka and Sano, 1976; Anderson et a / . , 1978), which were found to have their electrophysiological counterparts in, respectively, the posterior and anterior inputs to the atrioventricular node (Paes de Carvalho et al., 1959; Emberson and Challice, 1970; Spach et a / . , 1971; Janse et al., 1978). All of the above interatrial nerve pathways converge to the specialized

INTKINSIC NERVE PLEXUS OF MAMMALIAN HEART

I03

FIG.5 . Small ganglion cells as found in the nodal interstitium. Note abundant microtubules, microvesicles, and small mitochondria in their cytoplasma. The glycogen granules are sometimes encircled by a single membrane (arrow). The cell bodies are surrounded by a discontinuous Schwann cell sheath (Sw), which they share with numerous neurites (Ne). N, Nucleus; arrowhead, membrane-bound ribosomes; Nc, nodal cell.

I04

JOSEF MORAVEC A N D MIREILLE MORAVEC

tissue of the atrioventricularjunction, where they ramify and feed in the terminal nerve plexus constituted by the intrinsic ganglionic cells (Fig. 5 ) . These small ganglion cells were shown to invest the atrioventricularjunction of different species (Rossi, 1955; Abraham, 1969). At several occasions, it was suggested that presence of these cells might be of particular importance for the function of the intracardiac conductive tissue (Wensing, 1965; Bojsen-Mpller and Tranum-Jensen, 1972; Moravec and Moravec, 1984).

c. CHOLINERCIC OR ADRENERGIC CHOICE: PRESENT STATEOF THE QUESTION The ganglionic cells of mammalian hearts have been generally described as unipolar cholinergic neurons responsible for the second ganglionic relay of parasympathetic fibers to the heart (Jacobowitz, 1%7; Yamauchi, 1973; Pdpka, 1976; Ellison and Hibbs, 1976). The existence of intracardiac adrenergic neurons has been considered excluded for embryological reasons; the cardiac mesenchyme of birds and mammals is believed to induce invariably the expression of cholinergic phenotype in all ganglionic precursors migrating to the heart (Yamauchi, 1973; Viragh and Challice, 1977; Cabot and Cohen, 1980; Le Douarin ef al., 1981; Smith, 1983). The fact that most of the intracardiac neurons are intensely AChE-positive and, at the same time, devoid of specific fluorescence induced by formaldehyde vapors (Falck ef al., 1963; Jacobowitz, 1967; Ehinger ef al., 1968; Yamauchi, 1973) was often considered a confirmation of this rule. However, it should be noted that a small fraction of intracardiac ganglionic cells may be derived from the aortic plexus which is an adrenergic structure, similar to sympathetic ganglia and adrenal medulla (Le Douarin and Cochard, 1983; Le Douarin and Smith, 1983). It is not excluded that, during the folding of the primitive cardiac tube (Anderson ef al., 1978; Viragh and Challice, 1977), some of these adrenergic cells invest the ostia of neoformed great vessels and the intracardiac septum. The fact that they exhibit only moderate specific fluorescence should be considered in the light of the following data. Neurons devoid of specific fluorescence are present in the sympathetic ganglia of many species (Norberg, 1967; Gabella, 1976). The injection of nialamide [an inhibitor of the monoarnine oxidase (MAO)] produces a general increase in their fluorescence (Hamberger and Norberg, 1963). A similar observation was also reported by Jacobowitz (1967). who, after the administration of another inhibitor of the MA0 (i.e., tranylcypromine), could reveal fluorescent neurons similar to those of sympathetic ganglia in rat and guinea pig cardiac nerve plexus. Their yellow-green fluorescence

INTRINSIC NERVE PLEXUS OF MAMMALIAN HEART

I05

was further enhanced by a consecutive administration of catecholamine precursors, such as DL- or L-dopa (Jacobowitz, 1967). All of these observations may illustrate the heterogeneous distribution of the neurotransmitter between different compartments of adrenergic neurons (Gabella, 1976). It has been estimated that each neuronal body of the superior cervical ganglion of the cat contains on average 0.4 pg of norepinephrine, while the terminal varicosities of the same neuron totalize up to 150 pg of norepinephrine (Dahlstrom and Haggendal, 1966). It would seem that some of the sympathetic ganglionic cells may operate with low transmitter concentrations, except for their peripheral ramifications (Ehinger et ul., 1970). A similar situation also occurs in the intramural nerve plexus of the gut (Read and Burnstock, l%9). Also in this case, drastic interventions, such as the use of MA0 inhibitors and simultaneous administration of catecholamine precursors, are necessary in order to reveal the presence of adrenergic neurons (Burnstock and Bell, 1974). The autonomic neurons devoid of specific fluorescence have often an intense acetylcholine esterase activity (Hamberger et al., 1965), which was frequently considered as a histochemical manifestation of their cholinergic or, more exactly, parasympathetic nature (Jacobowitz, 1967; Yamauchi and Lever, 1971; Yamauchi, 1973). However, it has been shown that acetylcholinesteraseactivity can sometimes be associated with a weak, moderate, or strong fluorescence (Eranko, 1966). In the hedgehog, all nerve cells of sympathetic ganglia exhibit regularly an intense acetylcholinesterase reaction (Cauna et ul., 1961). This acetylcholinesterase of sympathetic neurons has been separated into external and internal pools. The former is believed to be involved in cholinergic ganglionic transmission (Mc Isaac and Koelle, 1959; Gabella, I976), the latter is possibly implicated in catecholamine synthesis, since it is associated with the Nissl bodies (Koelle r t ul., 1974; Black ef ul., 1979; Smith, 1983). I n vitro, adrenergic neurons synthetize both acetylcholine and norepinephrine, and it was shown that acetylcholine secreted by a sympathetic neuron triggers the release of norepinephrine from the same cell via its own nicotinic receptors (Burn. 1975). For all these reasons, the acetylcholinesterase activity cannot be any more regarded as the best marker of cholinergic (parasympathetic) neurons (Johnson et al., 1981; Potter et ul., 1981). Therefore, some of the nerve cells of mammalian hearts could, in fact, be sympathetic ganglionic cells, despite of the fact that they are frequently acetylcholinesterase-positive. A similar situation was previously described in the frog heart (Falck ct d., 1963; Taxi, 1976). In this latter species, an adrenergic component was described in the intramural nerve plexus of the atrioventricular junction. A neurosecretory component was also described in Bidder's ganglia, which

106

JOSEF MORAVEC AND MIREILLE MORAVEC

are located in the interatrial septum close to the atrioventricular valves (Bidder, 1868) and which do not seem to receive any vagal fibers (Marceau, 1946; Taxi, 1976).

D. EVIDENCE FOR INTRINSIC ADRENERGIC NEURONSIN THE NERVEPLEXUS OF THE ATRIOVENTRICULAR JUNCTION TERMINAL The existence of sympathetic preganglionic fibers in the heart of higher vertebrates still remains rather controversial. Most of the acetylcholinesterase-containing fibers are considered as vagal preganglionics and/or postganglionics (Jacobowitz, 1%7; Randall, 1976; Levy and Martin, 1979). However, it has been suggested that some of the sympathetic preganglionic fibers can transit through the stellate ganglia without interruption (Gabella, 1976). This has been confirmed, at least for the dog heart, by the denervation studies of Seagard et al. (1978). These authors have performed dorsal root ganglionectomy on left thoracic segments T,-T4. The peripheral cardiac nerves were examined 3 weeks later. At that time, totality of thin myelinated afferent fibers (up to 15% of total fibers) was degenerated in most of the left cardiac branches. The only exception was the ventrolateral cardiac nerve, which contained the least amount of degeneration (less than 5% of total fibers). In this nerve, intact large myelinated fibers (Aa category) persisted for a long time after the surgery. The authors of the above study concluded that these thick A a fibers are sympathetic preganglionics. The presence of myelin sheath and absence of any electrical activity during the stimulation of reflectogenic areas of the heart (Armour et al., 1975) would argue in favor of that possibility. Another significant point is the selectivity of peripheral projections of the ventrolateral cardiac nerve; this nerve was found to innervate selectively the atrioventricular junction where, also in the dog (Abraham, 1%9), abundant ganglionic cells are present. In this connection, it should be noted that large neurosecretory cells, closely associated with the specialized structures of the atrioventricular junction of the rat heart, have been revealed by a recent electron microscopic study (Moravec and Moravec, 1984; Moravec et al.;1986). In that FIG.6. An electron microscopic view of the perihissian ganglion of the rate pretreated by a single injection of 5-OH-dopamine.Note the presence of a large neurosecretory neuron with an eccentric position of its nucleus (N), separated from the rest of the cell body by a profound infolding of the neurolemrna (arrow). The central element is accompanied by several cellular processes which have the same cytoplasmic content, i.e.,abundant small mitochondria and two kinds of electron-dense microvesicles (35 and 60 nm in diameter) (insert). The size of these vesicles, as well as their disposition in clusters, reminds one of the cytological descriptions of adrenergic neurons. Nc, Nodal cells: Co, collagen: arrowheads, emergence process; asterisks, vacuoles.

INTRINSIC NERVE PLEXUS OF MAMMALIAN HEART

I07

108

JOSEF MORAVEC AND MIREILLE MORAVEC

work, the adrenergic components of cardiac innervation were prelabeled by a false precursor of catecholamines, i.e., 5-OH-dopamine(Tranzer and Thoenen, 1967; Chiba, 1973). The size of these neurosecretory neurons (up to 35 km in diameter) and the presence of two kinds of electron-dense vesicles (sporadic large ones and clusters of small vesicles with the respective diameters of 60 and 30 nm) (Fig. 6) resembled those of other sympathetic neurons (Matthews, 1974; Gabella, 1976; Taxi, 1976). However, the true nature of the transmitter, which could be released under physiological conditions [i.e., noradrenaline, adrenaline, or dopamine (Gabella, 1976; Drake-Holland et al., 1982; Geffen and Jarrott, 1977)], was not elucidated. These cells had one or two thick peripheral extensions, containing the same granular material as their cell bodies (Fig. 7). which resembled the dendritic collaterals of the adrenergic ganglion cells (Taxi and Droz, 1969; Jacobowitz, 1974) and, in particular, the 5-OH-dopaminecumulating dendritic specializationsof the guinea pig atrium (Chiba, 1973). One of the most interesting features of these perinodal ganglion cells is apparent in their immunohistochemical properties (Fig. 8a, b). According to our recent study (Moravec et al., unpublished data), cell bodies and peripheral extensions of these cells contain both neuropeptide Y (NPY) and C-terminal-flankingpeptide of neuropeptide Y (C-PON) (Fig. 8d, e). They also react with the anti-TH sera (Fig. 8c). This seems to suggest that catecholamines of the electron-dense vesicles, as observed in electron microscopy, can be synthetized in situ together with other prospective cotransmitters of adrenergic neurons (Lundberg et al., 1985; Van Noorden et al., 1985). In contrast, the perinodal ganglion cells do not react with anti-SP nor with anti-VIP sera which, in turn, gave positive results in other portions of the intracardiac nerve plexus, i.e., in the epicardial ganglia and in the intraseptal ganglionic lamina (Moravec et al., unpublished data). Most of these intramural, “short adrenergic neurons” (Sjostrand, 1%5), were surrounded by several thin layers of Schwann cells, which constituted a discontinuous sheath of loose myelin. Except for their dendrites, which frequently shared their Schwann cell sheaths with small cholinergic neurites, they were only sparsely innervated. Their nuclei were often eccentric, having been separated from the rest of the cell body by profound infoldings of the cell membranes (Fig. 6). This may explain the absence of nuclear profiles on some of the thin sections examined. FIG.7. (A) A dendritic-like adrenergic varicosity as seen in the nodal interstitium of the rat pretreated by a single injection of 5-OH-dopamine. Note two sizes of electron-dense vesicles and the presence of a continuous Schwann cell sheath (Sw). (From the Am. J . Anut.. 1984, 171, 307-320.) (B) A dendritic-like varicosity contrasted by 5-OH-dopamine, this time uncompletely coated only by the neighboring satellite cell (Sc). Note two enlarged axons synapting on the surface of this neurosecretory element of the perinodal interstitium (arrows). (From the Am. J . Anut.. 1984. 171, 307-320.)

INTRINSIC NERVE PLEXUS OF MAMMALIAN HEART

I 09

110

JOSEF MORAVEC AND MIREILLE MORAVEC

INTRINSIC NERVE PLEXUS OF MAMMALIAN HEART

111

The distinction between axons and thick dendritic collaterals was not easy without the use of 5-OH-dopamine (Moravec et al., 1985). In hearts of rats pretreated with this false precursor of catecholamines, the axons could be distinguished from the rest of neuronal body by their clear appearance. They contained microtubules and microfilaments, but only a few mitochondria and sparse vesicles (Fig. 9). Glycogen granules were sometimes surrounded by a single membrane (Fig. 9) in a manner which was similar to that described in the postganglionic fibers of sympathetic neurons of Rana esculenta by Taxi (1976) and in the bull frog by Berthold (1966). The initial segment of the axon was not myelinated. This fits well with the contention that this part of the nerve cell should be considered as its functional axon hillock (Palay et al., 1968; Kuffler and Nichols, 1976; Taxi, 1976). On some sections, even the Schwann cell sheath seemed to be discontinuous and, at these locations, several whorls of axonal profiles approached the ganglion cell body. However, the peripheral portion of the axon rapidly acquired a sheath of compact myelin. It was remarkable that this myelin developed in the external layers of Schwann cells, having been separated from the axon by several other layers of Schwann cell cytoplasm (Fig. 9). These sleeved fibers (Pick, 1962) resembled those described as typical for the postganglionic emergence processes of frog sympathetic neurons (Taxi, 1976). Another noteworthy feature of these large adrenergic neurons was their fragility face-to-face the isotonic solutions used for the perfusion fixation (2.3% glutaraldehyde in 45 mM sodium cacodylate) and for washing of samples (I50 mM sodium cacodylate buffer). Large exploding vacuoles frequently found in the perikarya of these cells would suggest that their intracellular milieu was strongly hyperosmolar. A similar situation was

FIG.8. Serial cryostat sections through the atrioventricular junction of the rat heart counterstained with pontamine sky blue. The nerve elements of the specialized tissue and the interventricular septum were visualized by indirect immunocytochemistry with fluorescein-conjugated secondary antibody. (a) A dense protein gene product (PGP)-positive nerve network interpenetrating His' bundle (HB) and the arising right branch (RB). An agglomerate of PGP-immunoreactive ganglion cells (arrow) supplying nerve bundles (arrowhead) into the interventricular septum (IVS). x 120. (b) A detail of the PGP-immunoreactive nerve cells (arrows)with their thick dendrites (arrowheads) interpenetrating the interventricular septum (IVS). x 300. (c) Tyrosine hydroxylase-immunoreactive nerve cell (arrow) at the junction of interventricular septum (IVS) and His' bundle (HB). ~ 4 0 0 (d) . Neuropeptide tyrosine (NPY)-immunoreactive nerve cell (arrow) of the atrioventricular junction with its thick dendritic projection interpenetrating the interventricular septum (IVS). HB, His' bundle. x 400. (e) C-PON-immunoreactive nerve cell (arrow) in the connective tissue separating His' bundle (HB) from the interventricular septum (IVS). Note a thin (axonic) projection of this cell (arrowhead). ~ 4 0 0 The . authors would like to thank Professor J. Polak who supplied the samples of sera used in this work and Professor J. Taxi who allowed us to use his fluorescence microscope.

112

JOSEF MORAVEC AND MIREILLE MORAVEC

INTKINSIC NEKVE PLEXUS OF MAMMALIAN HEART

1 I3

previously described at the level of terminal nerve fibers innervating the specialized cells of the atrioventricular node of the rat (Moravec-Mochet et al., 1977).

E.

CYTOLOGICAL ANALOGIES BETWEEN SENSORY NEURONS A N D THE NEUROSECRETOKY CELLS OF THE ATRIOVENTRICULAR JUNCTION

The irregular shape of the above cells, as well as the presence of an irregular satellite cell sheath, giving rise to several lamellae of loose or compact myelin, remind one of the cytological descriptions of autonomic ganglia (Forssmann, 1964; Hess, 1965; Gabella, 1976; Taxi, 1976) and those of sensory neurons (Hess et al., 1969; Peters et al., 1976; Pannese, 1981). The myelinated nerve cell bodies are numerous in the acoustic and vestibular ganglia in all classes of vertebrates, from elasmobranchs to man. They occur occasionally in other sensory ganglia, especially those of fish, amphibians, and reptiles. Most of these myelinated nerve cells are bipolar, some pseudounipolar. The myelin sheaths surrounding the nerve cell bodies extend along the initial portions of their axons up to the first nodes of Ranvier. Under the polarized light, the perikaryal myelin presents the same characteristics as those surrounding the nerve fibers (Scharf, 1958; Pannese, 1981). According to electron microscopic studies (Rosenbluth and Palay, 1961; Rosenbluth, 1967; Merck et al., 1975; Perre et al., 1977; Pannese. 1981). several types of perykaryal myelin can be distinguished. In some cases, the myelin sheath is built of a varying number of lamellae, each consisting of a layer of satellite cell cytoplasm bounded by the plasma membranes of satellite cells and separated from the nearby ones by a narrow space (loose myelin). In other cases, the perikaryal myelin displays a highly regular pattern. The ctyoplasm between the plasma membranes has disappeared, and the space between the lamellae is obliterated. In such cases, the perikaryal myelin is very comparable to the compact myelin surrounding the nerve fibers (Pannese, 1981). Between these two extremes of loose and compact myelin, a wide range of perikaryal myelin sheaths, such as single satellite cell layers, multiple layers of satellite cells, pseudomyelin FIG.9. (A) A large nerve process as seen in the perihissian glomerule, which might be either the emerging axon of an intrinsic neuron acquiring progressively its myelin sheath or a peripheral myelinated fiber loosing it myelin. Note the glycogen inclusions in the cytoplasma (arrowheads). SW. Schwann cells. (B) Section through a perinodal glomerule presenting a peripheral nerve fiber surrounded by an irregular layer of Schwann cell cytoplasm developing an eccentric sheath of compact myelin. Similar sleeved fibers characterize the postganglionic emergence processes of sympathetic ganglia. (Arrows) Loose myelin, (arrowheads) compact myelin. (asterisk)glycogen.

114

JOSEF MORAVEC AND MIREILLE MORAVEC

and, finally, compact myelin, can be found in sensory (Rosenbluth, 1967; Merck et al., 1975; Stensaas and Fidone, 1977; Pannese, 1981) and autonomic (Hess, 1965; Taxi, 1965; Takahashi and Hama, 1965) ganglia. The irregularity of their cell shape is another common feature of some sensory and autonomic neurons (Gabella, 1976; Peters et u l . , 1976; Pannese, 1981). Their cell bodies are often folded, which results in the eccentric position of their nuclei (Takahashi and Hama, 1965; Cantino and Mugnani, 1975). The axon is first directed proximally then, very close to the preganglionic fibers, it bends and courses distally where it becomes myelinated (Pannese, 1981). Several thick and short intracapsular dendrites (unipolar or pseudounipolar neurons) arise from these cells (Gabella, 1976). The cells of this type were described in dorsal root ganglia (Pannese, 1981) and in two types of peripheral sensory ganglia, i.e., the acoustic and vestibular ganglia (Peters et al., 1976). Some authors also pointed out the presence of irregular myelinated nerve fibers in autonomic ganglia, such as the paravertebral ganglia of frog (Taxi, 1976) and the ciliary ganglion of birds (Hess, 1965; Hess et al., 1%9). In this latter case, it was suggested that the time at which the myelin lamellae occur (before hatching in the chick and during the second posthatching period in the pigeon) correlates well with the appearance of electrical coupling and the bidirectional conduction through the developing ciliary ganglion (Hess et a!., 1%9; Manvitt ef al., 1971; Gabella, 1976). According to these data, the myelin sheath of the above ganglionic cells might be essential for the perception and transmission of local electrical phenomena. The presence of ganglionic cells of similar cytology at the level of the atrioventricularjunction might therefore be of considerable significance for the integration of local electrical phenomena and for the feedback regulation of cardiac activity (James, 1973; Pollack, 1974).

F. EVIDENCE FOR SENSORY NEURONSIN THE TERMINAL NERVE PLEXUS OF THE ATRIOVENTRICULAR JUNCTION Some of the ganglionic cells of the terminal nerve plexus of the atrioventricular junction could also be involved in the perception of mechanical phenomena related to ventricular systoles (Brooks and Lu, 1972; Irisawa, 1978). The existence, at the level of the intracardiac specialized tissue, of a nonidentified self-governed device involved in the control of the spontaneous activity of the heart has been suggested by several authors (James, 1973; Irisawa, 1978; Pollack, 1978; Brooks, 1981). According to our data (Moravec-Mochetet al., 1977; Moravec et al., 198% the density of sensory coiled endings in and around the atrioventricular junction is surprisingly high. Recently, we studied the atrioventricular junctional area of the rat heart, using thick serial silver-stained sections (Moravec, 1985).

INTRINSIC NERVE PLEXUS OF MAMMALIAN HEART

1 I5

The results of that study were quite significant. The specialized tissue of the atrioventricular junction was embedded in a connective tissue of the membraneous septum, which separates the left ventricular cavity from the right atrium, i.e., the arterial portion of the systemic circulation from its venous return (Anderson el a/., 1978; Moravec and Moravec, 1982). All along the course of the atrioventricular node, the bundle of His and the upper portions of right and left branches, agglomerates of large granular ganglion cells, could be identified (Fig. 10). Some of these cells had thick dendritic projections, which interpenetrated the upper portion of the interventricular septum and the reticular portion of the atrioventricular node (Moravec et a/., 1986). They terminated by large coiled endings enrolled around the respective muscle cells (Fig. I I), which, in electron microscopy, could be identified as sensory in nature (Moravec-Mochet ef al., 1977; Moravec and Moravec, 1982). The size of these ganglion cells (up to 35 pm), as well as the presence of a granular material in their cytoplasm, strongly suggested that these sensory neurons were identical to the abovedescribed adrenergic cells of the atrioventricular junction. Their axons encountered the enlarged terminal nerve fibers of the intraatrial neuromuscular pathways described in one of the preceding sections. At this level, abundant axodendritic and dendrodendritic glomeruli (De Castro, 1932; Gabella, 1976) could be identified by means of a combined phasecontrast and electron microscopic examination of the alternate thin and semithin sections (Moravec and Moravec, 1984). Further work, i.e., the use of serial-thin sections, will be necessary in order to follow the course of axons of these cells and to study the ultrastructure of their active zones as well as their relationships to the surrounding myelin. The association of an adrenergic component with the function of different sensory receptors is not new. Already Bernard (1851) had suggested that sympathetic nerves modulate the sensory input to the brain (cf. Gabella, 1976), and Bechterev (1896) presented the sympathetic ganglia as an offshoot of the central nervous system, which conducts impulses both centripetally and centrifugally from one functional unit to another (cf. Khabarova, 1961). However, direct proof for this integrative function of the sympathetic system has been lacking until recently. At present, the possibility that the sympathetic pathways may be involved in sensory functions has been open anew (Gabella, 1976; Leranth and Unguary, 1980; Brooks, 1981). For a while it seemed forgotten that, in many organs, autonomic nerves terminate in intramural plexuses (Meissner, 1857; Auerbach, 1864; Langley, 19211, which possess a high level of intrinsic activity and responsiveness (Langley, 1921; Brooks, 1981; Gershon and Erde, 1981; Wood, 1981). These latters may affect, or even elicit, the function of surrounding structures. Today, there are multiple examples of the role of an intrinsic adrenergic component in the local control of different myogenic

116

JOSEF MORAVEC AND MIREILLE MORAVEC

FIG. 10. A large sensory neuron of the perihissian ganglion as revealed by silver impregnation. Note several short dendrites terminated by large coiled endings surrounding a strand of muscle cells (arrowheads) and several large whorls of the axonal process (arrow). In contrast to the unipolar neurons of cardiac ganglia. this cell has a dark and granular cytoplasm similar to the above neurosecretory neurons. found in the electron micrograph at the same location. VS, Interventricular septum.

systems (Pollack, 1978; Leranth and Unguary, 1980; Cranefield, 1983; Weihe et al., 1984). As concerns the function of sensory organs, catecholaminergic components were already found in different types of mechanoreceptors, such as skeletal muscle spindles (Paintal, 1973; Barker, 1974), acoustic receptors (Spoendlin and Lichtensteiger, 1966), baroreceptors (Aars, 197 1 ; Belrnonte et al., 1972; Chiba, 1973), and in cardiac chemoreceptors (Knoche et al., 1970; Eyzaguirre et al., 1977). The association of large neurosecretory elements with myelinated nerve fibers was also described in frog hearts, namely, at the level of Bidder's ganglia (Bidder, 1868; Taxi, 19761.' 'Quite recently it has been suggested that the primary sensory neurons of cranial nerves, as well as some of the sensory neurons of dorsal root ganglia of adult rat, may express an adrenergic phenotype (Jonakait e/ a / . , 1984; Katz ct d..1983). This later disappears when they were disconnected from their respective projection areas which, in most cases, belong to the cardiovascular system (Katz and Black, 1986). The centripetal transport of trophic factors from the target tissues seems to be necessary for the catecholaminergic traits of cranial sensory neurons to be maintained (Katz and Black, 1986).

INTRINSIC NERVE PLEXUS O F MAMMALIAN HEART

I17

FIG.I I . ( A ) A detail of the coiled ending (arrow) surrounding a strand of specialized cells of the reticular segment of the AVN (atrioventricular node) as revealed by the silver impregnation technique. (B) An electron micrograph view of a section through the presumptive sensory corpuscle. Several profiles of the same nerve fiber (arrows) can be found in the vicinity and within an agglomerate of nodal cells. Note the predominance of very small mitochondria and small empty vesicles (30 nm in diameter). The nerve elements are accompanied by a discontinuous satellite cell sheath (Sw), which is lost at the site of the terminal nerve-muscle interactions (arrows). (From the "Advances in Myocardiology." Vol. 6, pp. 13-23. Plenum, 1985.)

1 I8

JOSEF MORAVEC AND MIREILLE MORAVEC

INTRINSIC NERVE P L E X U S OF MAMMALIAN HEART

1 I9

Another finding, which seems to be of particular significance for our discussion, is the existence of peripheral mechanosensory neurons, inserting their dendrites into the fine strands of skeletal muscle fibers that were found operating in crayfish and other crustaceans (Alexandrowicz, 1951; Eyzaguirre and Kuffler, 1955; Bullock et al., 1977). They were shown to provide a powerful proprioceptive feedback to the neural networks involved in the generation of motor patterns responsible for rhythmic swimmeret movements (KuMer and Nichols, 1976). The fact that, in the heart, the perception of electrical and mechanical phenomena, affecting the atrioventricular junction during each contraction cycle, seems to be in the hands of the intrinsic adrenergic neurons (Figs. 6, I I and 12) (Moravec and Moravec, 1984) might thus be of considerable significance. The cyclic catecholamine release from the above neurosecretory cells can be expected to affect both specialized tissue and cholinergic components of the cardiac nerve plexus (Jacobowitz, 1967; Yamauchi et al., 1973; Pollack, 1978). In this way, the catecholamine (and neuropeptide) recycling may contribute to the intrinsic autoregulation of the pacemaker activity (Pollack, 1977; Cranefield, 1983) and intracardiac conduction (Pollack, 1974; Irisawa, 1978). This can also explain some still badly understood aspects of vagosympathetic (Levy and Martin, 1979) and vagus-cardiac pacemaker (Spear et al., 1979; Michaels et al., 1983) interactions. IV. Terminal Nerve Plexus According to light microscopic studies, the intramural cardiac ganglia contain unipolar, bipolar, and multipolar neurons (Davies et al., 1952; Abraham, 1969; Ellison and Hibbs, 1976). They also contain many nerve fibers in transit, including a number of sensory fibers and fibers which originate from, or are directed toward, other ganglia (Gabella, 1976). Most of the preganglionic fibers were believed to be preganglionic inhibitory branches from the vagus nerves. However, with fluorescence microscopy, numerous adrenergic fibers running through the cardiac plexus could also be identified (Ehinger et al., 1968; Nielsen and Owman, 1968; Forsgren, 1985). Most of these adrenergic fibers originated from the inferior cervical FIG.12. (A) A section through a preterminal nerve fiber of the autonomic ground plexus of the atrioventricular junction in a 5-OH-dopamine-pretreated rat. Note the coexistence of an adrenergic (ADR) and a cholinergic (ACH) nerve fibers sharing the same Schwann cell (Sw) sheath. Nc, Nodal cell. (B) A free adrenergic varicosity (arrow) as seen in the A V nodal interstitium in a 5-OH-dopamine-treated rat. (C) An adrenergic neuromuscular junction encountered in the compact zone of the A-V node of a rat pretreated with 5-OH-dopamine. Note the absence of basement membrane in the intermembrane space, the densification of synaptic membranes, and an equivocal postsynaptic apparatus (arrows).

I20

JOSEF MORAVEC AND MIREILLE MORAVEC

ganglion (Wacksman et al., 1%9) or from the thoracic paravertebral ganglia (Gabella, 1976). Only some of them persisted after surgical sympathetic denervation (Potter et al., 1965). The contribution of an intrinsic adrenergic component to cardiac innervation is therefore not excluded. The persistence of the tyrosine hydroxylase activity as well as that of dopaminebinding sites in chronically denervated dog hearts (Drake-Holland et al., 1982) would argue in this sense. Terminal varicosities of these adrenergic fibers impinge on ganglion cells forming very small terminals whose detection in both optical and electron microscopy requires optimal conditions (Nielsen and Owman, 1%8; Ehinger et al., 1970). Some of these adrenergic varicosities can form axosomatic synapses with the intramural ganglion cells and, sometimes, they can approach cholinergic postganglionic axons (Jacobowitz, 1967; Ehinger et al., 1970; Yamauchi, 1973). The existence of adrenergic neuromuscular junctions was also suggested. In our work, we found them in the rat atrium (Fig. 12).

A. AUTONOMIC GROUNDPLEXUS From the morphological point of view, this terminal portion of cardiac innervation resembles Hillarp’s description of the autonomic “ground plexus” (Hillarp, 1946, 19591, which has been identified in most of the peripheral organs supplied by autonomic nerves (Burnstock and Bell, 1974; Gabella, 1976; Taxi, 1976; Brooks, 1981; Gershon and Erde, 1981). According to this concept, the terminal autonomic axons invest a finely meshed network of Schwann cells. Each Schwann cell sheath contains several axons derived from different neurons (Yamauchi, 1973), enabling a convergence of nerve fibers to the same effector cell. The terminal nerve fibers course within the plexus for longer distances and innervate en passant several effector cells. According to some authors, the adrenergic and cholinergic axons can also form reciprocal synapses (Ehinger et ul., 1970) (Fig. 12A). However, there is no proof that the appositions of adrenergic and cholinergic fibers do have a functional significance. At the electron microscopic level, the autonomic ground plexus has a certain number of morphological features, which are common to all organs presenting an intrinsic innervation (Taxi, 1965; Rogers and Burnstock, 1966; Yamauchi, 1973; Taxi, 1976). It consists of varicose branching fibers enveloped by Schwann cells which interpenetrate the interstitial spaces. Close to the neuroeffector junctional area, the Schwann cell sheaths become discontinuous and, through these exposed foci, the transmitter substance can be released (Fig. 12B). In some cases, rudimentary neuromuscular synapses could also be identified (Richardson, 1964; Yamauchi, 1973; Moravec-Mochet et d . , 1977).

INTRINSIC NERVE PLEXUS OF MAMMALIAN HEART

121

I . Sympathetic Postganglionic Fibers The density of adrenergic nerves is higher in the atria than in the ventricles in almost all mammalian species, except for the cat (Nielsen and Owman, 1968; Yamauchi, 1973; Levy and Martin, 1979) and the bat heart (O'Shea and Evans, 1985). In the case of this hibernating animal, the density of adrenergic innervation is quite uniform in both the atria and the ventricles. In other species, the adrenergic axons are the most numerous in the auricular appendages and around the structures of the conductive tissue, namely, at the level of the sinus node (Ehinger et al., 1968; Yamauchi. 1973). The atrioventricular node, mainly its reticular portion, also receives the adrenergic innervation. Its compact zone, on the other hand, is supplied essentially by abundant cholinergic axons (Thaemert, 1970; Bojsen-Mgller and Tranum-Jensen, 197 I ; Moravec-Mochet et ul., 1977). There seems to be a controversy as to the adrenergic innervation of the ventricular segments of the conductive tissue, i.e., that of the bundle of His, Purkyne cells, and moderator bands (Yamauchi, 1973). These structures were shown not to have any adrenergic terminals at all, at least in the dog and pig hearts (Dahlstrom et al.. 1965; Bojsen-Mgller and TranumJensen, 1971). However, using the glyoxylic acid technique (Axelsson e t d . , 1973), which gives better results than the classical Falck and Hillarp method, Forsgren (1985) succeeded to demonstrate numerous adrenergic varicosities in the ventricular conduction system of bovine and human hearts. He suggested that catecholamine release from these intraventricular adrenergic endings might be responsible for the development of postischemic or postinfarction ventricular tachyarrhythmias. As to the adrenergic innervation of both ventricles, there seems to be a general consensus that it is less extensive than that of atria. Although adrenergic axons can be found distributed within the ventricular myocardium (Jacobowitz et al., 1967; Winckler, 1969; Yamauchi, 1973), the adrenergic terminals are almost exclusively concentrated to the perivascular spaces (Nielsen and Owman, 1968). Adrenergic neuromuscular junctions were found only occasionally (Ehinger et al., 1970) (Fig. 12C). There is now experimental evidence that some of the intracardiac adrenergic fibers do not degenerate after chronical cardiac denervation (Jacobowitz, 1967; Potter et al., 1965). This may suggest that they are, at least some of them, derived from the intrinsic adrenergic cells. The existence of the latter was suggested at many occasions by fluorescence and electron microscopic studies (Truex, 1950; Jacobowitz e t al., 1967; Yamauchi, 1973; Ellison and Hibbs, 1976; Moravec and Moravec, 1984). According to the biochemical analysis (Gabella, 1976), up to 50% of norepinephrine of the pacemaker area may be provided by the intracardiac chromafin cells and by the above-mentioned intramural adrenergic neu-

122

JOSEF MORAVEC AND MIREILLE MORAVEC

rons, which can be supposed rather resistant to the effects of both surgical and chemical denervation (Potter et al., 1965; Jacobowitz, 1967; Tranzer and Thoenen, 1967; Burnstock and Bell, 1974; Gabella, 1976; Drake-Holland et al., 1982).

2. PARASYMPATHETIC POSTGANGLIONIC FIBERS In comparison with the efferent synapses of skeletal muscles, the autonomic nerve endings on smooth and cardiac muscle cells are small, and their ultrastructure resembles that of embryonic skeletal muscle end plates and that of “en grappe” endings of the intrafusal fibers (Yamauchi, 1973; Uehara et al., 1976). The pre- and postsynaptic specializations, typical for neuromuscular synapses (Couteaux, 1978), are often rudimentary or missing (Couteaux, 1961 ; Yamauchi, 1973; Moravec-Mochet et al., 1977). Only occasionally, close membrane-to-membrane appositions of presumptive cholinergic terminals and specialized muscle cells have been observed in the cardiac nerve plexus of different species (Couteaux and Laurent, 1958; Viragh and Porte, 1961; Thaemert, 1970; Taxi, 1976). However, the interpretation of these structures was questioned by Yamauchi (1969). Some authors (Moravec-Mochet et al., 1977) found these endings, leaving a gap less than 20 nm and devoid of any basal lamina, associated with terminal arborizations of large coiled fibers, which were similar to sensory innervation of skeletal muscle spindles (Uehara et al., 1976). Some of these sensory specializations seemed to be supplied by large intramural sensory neurons (see above), at least at the level of the atrioventricularjunction. As concerns the parasympathetic postganglionics, characterized by empty vesicles (30-50 nm in diameter), even after the administration of 5-OH-dopamine (Chiba, 1973; Yamauchi, 1973; Moravec and Moravec, 1984), they were suggested to be derived from cholinergic neurons of the parasympathetic cardiac ganglia (Hirsch et al., 1963; Yamauchi, 1973; Ellison and Hibbs, 1976). In optical and electron microscopy, these fibers were strongly AChE-positive and so were the plasma membranes of the adjacent muscle cells (Hirano and Ogawa, 1967; Jacobowitz et al., 1967; Yamauchi, 1973). Up to 70% of these cholinergic axons remain intact after the cardiac autotransplantation (Napolitano et al., 1%5; Potter et al., 1965) and mediastinal neural ablation (Jacobowitz et al., 1967). This proves that FIG.13. ( A ) A small en grappe efferent ending with empty vesicles (50 nm in diameter) presumably cholinergic. bl, Basal lamina; arrow, agglomerate of synaptic vesicles: Nc, nodal cell. (From the J . Ultruslrucr. Res.. 1977,58, 196-209.) (B) A small presumptive cholinergic ending with scarce vesicles and small mitochondria. Note the persistence of the basal lamina in the synaptic cleft and rudimentary postsynaptic folds (arrows). Nc, Nodal cell. (From the J . Ulrrustrucr. Res., 1982, 81, 47-65.

INTRINSIC NERVE PLEXUS OF MAMMALIAN HEART

I23

I24

JOSEF MORAVEC AND MIREILLE MORAVEC

INTRINSIC NERVE PLEXUS OF MAMMALIAN HEART

125

a majority of cholinergic postganglionic fibers is derived from the intrinsic ganglion cell of the cardiac nerve plexus (Jacobowitz, 1967; Loffelholz and Pappano, 1985). The distribution of the cholinergic fibers and, mainly, the density of cholinergic neuromuscular synapses vary throughout different cardiac compartments. Only in hibernating animals, a diffuse distribution of vagal postganglionic fibers was recently reported (O’Shea and Evans, 1985). Apart from this exception, both atria were comparatively more innervated than the ventricles in most mammalian species examined (Jacobowitz et al., 1967; Yamauchi, 1973; Levy and Martin, 1979). After surgical denervation, the number of cholinergic nerves in the left atrium was considerably reduced, while only minor reduction was observed in right atria and in the ventricles (Jacobowitz et al., 1967). As concerns the conduction system of the heart, an accumulation of cholinergic synapses (small en grappe endings with empty vesicles and minor pre- and postsynaptic specializations)was mentioned in the compact zone of the atrioventricular node (Yamauchi, 1973) and in the bundle of His (Moravec-Mochet et a / . , 1977). More elaborated synapses (very small en plaque terminals), similar to those of polar regions of skeletal muscle spindles (Uehara et al., 1976), predominated in the superficial, reticular layers of the atrioventricular node of the rat (Moravec and Moravec, 1982) (Fig. 13A and B). According to Thaemert (1970), who studied the tail of the mouse atrioventricular node on serial sections, every nodal cell in that portion of the mouse heart receives at least one cholinergic terminal. In addition to these efferent synapses, the presence of afferent nerve varicosities was also described in this part of the atrioventricular junction (Fig. 14) (Moravec-Mochet et al., 1977),as well as at the level of the mole sinoatrial node (Kikuchi, 1976). This double (efferent and afferent) innervation of different segments of the intracardiac conduction system of mammals may suggest that it should be considered as an intrinsic spindlelike organ, similar to the in situ mechanoreceptors of lower species (Kuffler and Nichols, 1976). B. CHROMAFFIN CELLS The first description of clusters of small granular cells within, or close to cardiac ganglion of the dog, is that of Truex (1950). In 1961, Viragh FIG. 14. A muscular spindlelike structure of the atrioventricular node of the rat. Note double [sensory (arrow) and motor (arrowheads)] innervation of a group of nodal cells surrounded by a common basal lamina. (Inset) Detail of the sensory ending. Note the small mitochondria with concentric cristae, accumulation of glycogen, and scarce empty vesicles. The intermembrane space (less than 20 nm) is devoid of basal lamina. Abundant pinocytic vesicles surround the nerve ending. (From the J . Ulrrustrucr. Res., 1977, 58, 1%-209.)

126

JOSEF MORAVEC AND MIREILLE MORAVEC

and Porte (1961) published their ultrastructural observations concerning the “cellules particulihres,” which they found in rat cardiac ganglia. There is no doubt that these cells are identical with the chromaffin cells, which have been consequently identified by histochemical (Jacobowitz, 1967; Nielsen and Owman, 1968; Winckler, 1969) and electron microscopic studies (Ellison and Hibbs, 1974; Yamauchi et al., 1975; Moravec and Moravec, 1984). These cells occur grouped in small clusters surrounding a capillary (Fig. 15). They are either isolated in the wall of the interatrial septum or associated with the intracardiac ganglia. In this case, they can share their satellite cell sheath with the adjacent nerve structures. They contain large secretory granules (75-120 nm in diameter), each with an eccentric dense core surrounded by an electron-translucent zone. The density of their secretory granules can be enhanced by the preliminary administration of 5-OH-dopamine (Moravec and Moravec, 1984). On the other hand, in contrast to the terminal adrenergic varicosities, the chromaffin cells are less sensitive to chemical and immunological denervation, as well as to the effects of reserpine (Iversen er al., 1966; Burnstock and Bell, 1974). Some authors also reported their persistence in surgically denervated hearts (Jacobowitz, 1967). The chromaffin cells were found associated with cardiac ganglion of the turtle heart (Yamauchi, 1973), in which they formed reciprocal inhibitory synapses with cholinergic neurons. A similar situation also occurs in rat and guinea pig hearts (Ellison and Hibbs, 1974; Yamauchi et al., 1975). These observations strengthened the view of Jacobowitz (1967), according to which, the chromaffin cells of the heart should be considered adrenergic interneurons modulating ganglionic transmissions in the terminal nerve plexus. In fact, some authors suggested that the chromaffin cells, as seen in electron microscopy, are morphologically similar to small intensely fluorescent (SIF) cells of Norberg et al. (1966), which were found associated with sympathetic ganglia. According to Ellison and Hibbs (1974), these two types of catecholaminecontaining cells are one and the same. Phylogenetically, they might be derived from specific granule-containing cells, which are evenly distributed in the heart of the hagfish and concentrated to the atrioventricularjunction in lampreys and higher vertebrates (Lignon and Le Douarin, 1978). These cells probably recycle catecholamines during each cardiac cycle (Pollack, 1977) and are thought to take part in adrenergic control of the heart via their dual action on cardiac muscle cells and intracardiac cholinergic neurons. A different point of view was formulated by Burnstock and Bell (1974). These authors considered the SIF cells as an intermediate cytological class between short adrenergic neurons of Sjostrand (1%5) and chromaffin cells of the adrenal medulla. As to their function, Burnstock and Bell were in

INTRINSIC NERVE PLEXUS OF MAMMALIAN HEART

127

FIG.15. Cluster of small chromafin cells (5 kum in diameter) found in the interstitial space of the atrioventricular conductive tissue. Note the irregularity of the Schwann cell sheath (Sw), the presence of several neurites (arrows), and large vesicles with eccentric electron-dense cores. G, Golgi apparatus. (From the Am. J . Anal.. 1984. 171, 307-319.)

I28

JOSEF MORAVEC A N D MIREILLE MORAVEC

favor of the possibility that the SIF cells, which are most abundant in the autonomic ganglia, can play the role of the interneurons. On the other hand, the chromaffin cells, which are often associated with blood vessels, may function as chemoreceptors involved in the control of the systemic and coronary circulation. The same opinion was also shared by James et al. (1972) and by Ellison and Hibbs (1974). However, apart from SIF and chromafin cells, another adrenergic component of the cardiac plexus (i.e., the above-described large sensory neurons of the atrioventricularjunction) could take an active part in the intrinsic parasympathetic-sympathetic interactions (Levy and Martin, 1979). This would considerably enhance the number of synaptic patterns available and improve the function of the intracardiac oscillatory networks, providing both a focal neuromodulation and a phasic sensory input into the system. These two mechanisms are believed to be essential for the control of oscillatory discharges produced by different neuromuscular motor pattern generators (Selverston and Moulins, 1985). It has been also suggested that the fluorescent cells of the atria may contain, not only catecholamines, but also serotonin and other related substances (Angelakos et al., 1969; Nee1 and Parsons, 1986). In this connection, it should be noted that serotonin might act as a neurotransmitter, at least at the level of the myenteric plexus (Gershon, 1977). Several other studies have indicated that adrenergic axons are responsible neither for the synthesis nor for the high affinity uptake of this compound; according to embryological studies (Gershon and Thompson, 1973), the development of serotoninergic components of the myenteric plexus precedes the ingrowth of adrenergic axons. This would indicate that the uptake of serotonin must be accounted for by intramural neurons which can survive in organotypic cultures. A similar situation may also occur in the cardiac nerve plexus, since serotoninergic neurons have been demonstrated in the hamster (Sole et al., 1979) and in the rat (Votavova et a / . , 1971) hearts. Another prospective source of serotonin and of other vasoactive substances was identified at the level of the interatrial septum of the rat. In this species, a richly innervated lacunar body containing numerous mast cells was invariably present in the proximity of the reticular portion of the atrioventricular node (Fig. 16). A similar structure was also described close to the accessory atrioventricular node (Anderson, 1972), and it was suggested that these secretory cells might be involved in different cardiogenic reflexes (James et al., 1972). A local action of the 5-OH-tryptamine and histamine on the neural structures of the atrioventricularjunction is also not excluded, if we take into account the recent demonstrations of serotoninergic receptors encountered in other divisions of the autonomic and cerebrospinal nervous system (Gershon, 1977). One of the structures which might be particularly sensitive to 5-HT action is the sensory network

INTRINSIC NERVE PLEXUS OF MAMMALIAN HEART

I29

FIG. 16. Detail of the paranodal lacunar body showing the presence of mast cells with large secretory granules of variable density. These cells are supposed to synthetize and store the 5-OH-tryptamine (serotonin).

composed of substance P (SP)-immunoreactive nerve fibers and SP-containing nerve cells, which was recently described in the rat and guinea pig atria (Weihe et al., 1986; Moravec et al., unpublished data).

C. INTERSTITIALCELLS Several species of supporting cells can be distinguished in the intracardiac nerve plexus of different vertebrates. Typical Schwann cells and small satellite cells with fibrous cytoplasm and few organelles tightly surround the large nerve profiles and accompany small terminal neurites into the intracardiac interstitium. This perineural sheath is lost as the nerve fibers join the terminal nerve plexus, for example, at the level of the atrioventricular node. Here the Schwann cells are substituted by large irregular cells which seem to be interconnected by their peripheral processes (Yamauchi et d.,1973; Moravec and Moravec, 1984). The cytoplasms of these cells contain abundant ribosomes arranged in rosettes and associated with the endoplasmic reticulum. Their principal process contains few microtubules, and, often, a large collagen bundle seems to be inserted into its base. The cell membrane of these cells is devoid of basal

130

JOSEF MORAVEC AND MIREILLE MORAVEC

lamina and presents particular specializations,such as microvilli and cilia with their basal bodies (Fig. 17). Their principal extensions are closely associated with the muscle cell membranes in which attachment plaques could be identified (Yamauchi et al., 1973). Sometimes they also receive adrenergic endings. This association resembles axosomatic synapses, at least in the fish heart (Yamauchi, 1973) and in the atrioventricularjunction of the rat (Moravec and Moravec, 1984). In this case, the above intercalary cells were found in nodal interstitium and in the interstitial spaces of the adjacent interventricular septum close to the sensory specializations of the terminal nerve plexus. Some of these glial cells can undergo cell divisions, since typical mitotic figures can be found in the optical and electronic microscopes (Moravec and Moravec, 1984). The above cytological description resembles that of the autonomic interstitial cells of Cajal (1894), which are structurally and functionally interposed between the postganglionic axons and the effector cells of the intramural plexus of the gut and other tissues receiving autonomic innervation (Thuneberg, 1982). According to histological studies, the interstitial cells can share some of their staining properties with the autonomic neurons themselves. However, an exhaustive embryological study will be necessary, for it is possible to state whether they have diverged from the same precursors as ganglion cells or whether the ingrowing autonomic nerve fibers have attracted the mesenchymal elements from the target tissues (Taxi, 1965; Rogers and Burnstock, 1966; Yamauchi et al., 1973). Most of the authors interested in the physiological role of the interstitial cells came to the conclusion that these cells represent a special type of glial cells characteristic for the intramural division of the autonomic nervous system. Their presence in the heart confirms that this intrinsic nervous component is also present in hearts of different vertebrates. However, an interesting hypothesis was recently postulated by Thuneberg (1982), who, after an exhaustive morphological and physiological study, considered the interstitial cells of the gut as the “intestinal pacemaker” cells. He found them structurally associated with the electrophysiologically identified foci, which initiated peristaltic sequences of isolated intestinal segments under study. Whether a similar situation also holds in the heart is not known. However, the presence of interstitial cells in the sinus venosus of the fish heart (Yamauchi et al., 1973) as well as their association with the atrioventricular junction of the rat heart (Moravec and Moravec, 1984) are quite significant in the light of these data. One may conclude, together with Yamauchi et al. (1973), that the interstitial cells of the nodal tissue represent a specialized cell population playing a role in the modulation of the autonomic nerve input directed to the effector myocardial cells.

INTRINSIC NERVE PLEXUS OF MAMMALIAN HEART

131

FIG. 17. Interstitial cell (Ic) found in the upper portion of the interventricular septum of the rat heart. Note its irregular shape and its close relationships with the adjacent myocytes. The cytoplasm contains rough endoplasmic reticulum and abundant free ribosomes. Co. Collagen fiber; arrows, microvilli. (From the Am. J . Anat., 1984, 171, 307-319.)

132

JOSEF MORAVEC A N D MIREILLE MORAVEC

V. New Developments in Studies of the Autonomic Nervous System

In the above sections, the nerve supply to the heart is still considered in terms of the dual (adrenergicand cholinergic) innervation (Dale, 1953). According to this classical principle, each nerve cell makes and releases only one transmitter. However, morphological and neurochemical studies performed is recent years strongly suggest that single neurons may contain and release more than one active compound (Furshpan et al., 1976; Burnstock, 1978; Potter et al., 1981; Lundberg et al., 1983; Smith, 1983). This concept of cotransmission holds for both sympathetic and parasympathetic innervation of the heart (Forssmann et al., 1982; Reinecke et al., 1982; Gu et al., 1983; Lundberg et al., 1983) as well as for the intrinsic cardiac neurons (Gu et al., 1984; Hassall and Burnstock, 1984; Weihe et al., 1984). Generally speaking, the coexistence of the classical neurotransmitters, i.e., norepinephrine or acetylcholine, with a series of putative neurotransmitters (ATP and serotonin) and neuromodulators [substance P, neuropeptide Y (NPY), vasoactive intestinal peptide (VIP), neurotensin, etc.)] could be demonstrated in all divisions of the autonomic nervous system, i.e., in the autonomic ganglia, postganglionic nerve fibers, and intramural neurons (Hokfelt et al., 1977; Leranth and Feher, 1983; Cummings et al., 1984; Lundberg et al., 1984; Elfvin, 1983). In some cases, the two traditionally antagonistic transmitters, i.e., norepinephrine and acetylcholine, were shown to coexist in a single neuron (Patterson, 1978; Burnstock, 1978). It would seem that each neuron, being a cell body supplied with a complete set of genes, possesses the potential ability to synthetize the entire set of enzymes for all transmitter substances. Its actual phenotypic expression is determined by the environmentalfactors and the type of its interactions with the target tissue (Potter et al., 1974; Furshpan et al., 1976; Patterson, 1978; Smith, 1983). It has been suggested that, during the organogenesis, the migrating precursors of nerve cells, having been exposed to different microenvironmental conditions, develop distinct phenotypes leading to different transmitter choices (Patterson, 1978; Le Douarin, 1980, 1984). According to Gershon (Gershon et al., 1981), the microenvironmental signals provided by the target tissues at critical points of their development can explain the sequential changes occurring in the phenotypic expression of peripheral nerve cells during their ontogeny (Le Douarin, 1984) and in conditioned cell cultures (Patterson, 1978). This plasticity of nerve cell precursors can account for the multiplication of the neuronal types and for the diversity of neurotransmitters encountered in some portions of the autonomic nervous system, such as the enteric nervous plexus (Gershon and Erde, 1981; Cummings et al., 1984; Wood, 1984) and the terminal nervous plexus of the heart (Forssmann et a / . ,

INTRINSIC NERVE PLEXUS OF MAMMALIAN HEART

I33

1982; Gu e t a / . , 1983; Dalsgaard et a/. 1985; Hassall and Burnstock, 1984; Moravec and Moravec, 1984; Weihe et a/., 1984). Although it is not easy to reconcile all available data, it would seem that VIP-like immunoreactivity predominates in the perivascular nerves of mainly parasympathetic origin. VIP and PHI (peptide histidine isoleucine) (Lundberg and Tatemoto, 1982) are considered as strong candidates for the atropin-resistant vasodilatation seen upon the parasympathetic stimulation. In the heart, VIP has been shown to be responsible for coronary vasodilatation and positive inotropic effects, which have been reported in dog and cat heart preparations (Said et al., 1972). VIP-immunoreactive cell bodies were demonstrated in the epicardial ganglia, and VIP-immunoreactive fibers predominated in the conduction system. Some of these VIP-containing nerve fibers were in close contact with the cells of the atrioventricular node (Weihe et al., 1984). On the other hand, the NPY and the C-terminal-flankingpeptide of neuropeptide Y (C-PON) have been shown to coexist with norepinephrine in sympathetic nerves around blood vessels. NPY itself induces local vasoconstriction and hypertension upon the intravenous administration (Lundberg and Tatemoto, 1982). The vasoconstrictor response to NPY is long lasting as compared to norepinephrine. The response to NPY is not associated with any postinfusion hyperemia which is seen after the administration of norepinephrine. Finally, the effect is resistant to a-adrenoreceptor-blocking agents and persists after the sympathectomy (Lundberg et a/., 1984). In the heart, the NPY-immunoreactive fibers were detected in the atria of different species including man (Gu et a / . , 1983). Lundberg et a / . (1983) suggested that NPY coexists with both dopamine (3-hydroxylase and norepinephrine. NPY-immunoreactive cell bodies with weak tyrosine hydroxylase-irnmunoreactivity were also found in the rat and mouse hearts (Gu et d.,1984; Moravec et al., unpublished data). According to Hassall and Burnstock (19841, NPY is not exclusively associated with the extrinsic nerves to the heart; it can be synthetized by cultures of the intrinsic cardiac neurons obtained from newborn guinea pigs, the hearts of which are not yet innervated by the extrinsic nerves. In this case, NPY does not coexist with norepinephrine, but it is not excluded that this would be the case in sitir where these intrinsic cardiac neurons still contain dopamine P-hydroxylase and tyrosine hydroxylase (Gu et al., 1984). Another peptide which has been found associated with the intrinsic ganglion cells, at least in the gut (Costa et al., 1982), is substance P. This compound has been identified as an atropine-resistant excitatory neuromodulator secreted by the network of multipolar intestinal interneurons (Costa rt a/., 1981; Cummings et a/., 1984; Wood, 1984). The activation of the entire network is triggered by serotonin release from neurons with

134

JOSEF MORAVEC AND MIREILLE MORAVEC

Dogie1 I morphology (unipolar neurons); the activity within the network is terminated by an inhibitory action of the optoid peptides. These are known to hyperpolarize and inhibit the intramural neurons by increasing their GK (cf. Wood, 1984). Another substance P-mediated action is the antidromic stimulation of afferent C fibers involved in local vasodilatation (Hokfelt et al., 1975; Lundberg et al., 1984). The association of substance P with sensory nerve fibers investing various peripheral tissues, including the heart, is now rather well established (Lundberg et al., 1984; Weihe et al., 1984; Nee1 and Parson, 1986). In this respect, it should be noted that a small proportion of substance P containing ganglion cells could be identified in sympathetic ganglia (Hokfelt et al., 1975; Leeman, 1980). In particular, the superior and the middle plus inferior cervical ganglia as well as the upper thoracic (T3-T,) ganglia of the rat contain the substance P-immunoreactive primary nerve fibers sensitive to capsaicin treatment (Tsunoo et al., 1982; Papka et al., 1981). According to Forssmann et al. (Weihe et al., 1984), various receptor structures of the heart (coronary chemoreceptors, baroreceptors, etc.) could be interconnected, via axonal collaterals of substance P-positive afferents, with the efferent sympathetic and parasympathetic pathways far below the level of their respective central integration centers. Abundant substance P-immunoreactive nerve fibers were found branching on the intracardiac cholinergic and VIP-ergic nerve bodies. These authors suggested that the cardiac substance P-containing afferents may modulate the actual postganglionic nervous input to the heart. They can be also involved in short, intrinsic, substance P- and VIP-mediated feedback loops necessary for the fast beat-to-beat autoregulation of cardiac electrical activity (Covell et al., 1981; Weihe et al., 1984).

The above distinction between the VIP, NPY, and substance P-ergic nerve fibers should be considered as rather didactic; it is not excluded that some of the neuropeptides coexist in a single nerve cell (White et al., 1985). It has been demonstrated that different neuropeptides can be derived from the same precursor synthetized by a unique gene. The choice of a given amino acid sequence, that is a given phenotypic expression, takes place at the transcriptional level, different mRNA being used for the synthesis of different neuropeptides (White et al., 1985). This situation may contribute to the multiplicity of neuronal phenotypes as well as to the diversity of interneuronal connections encountered in the terminal nerve plexuses. This, in turn, seems to argue against the idea that the intramural plexuses should be considered as a diffuse parasympathetic ganglia in which direct synaptic connections between preganglionic fibers and postganglionic cholinergic neurons are the unique basis of organ function (Wood, 1984).

INTRINSIC NERVE PLEXUS OF MAMMALIAN HEART

I35

Some of the ancestral conceptions concerning the exclusivity of both adrenergic and cholinergic centrifugal transmission in the autonomic nervous system (Dale, 1953) do not hold any longer. At present, the peripheral extensions of the autonomic nervous system are more and more considered in terms of pluricellular networks with a high degree of histotopographical specialization providing a nonnegligible functional autonomy (Leranth and Unguary, 1980; Brooks, 1981; Weihe et al., 1984; Wood, 1984). This evolution is nothing else but the rehabilitation of Langley’s concept of the functionally independent intrinsic nervous system of peripheral organs (Langley. 1921). At the same time, the recent progress in the electron microscopic and immunohistochemical techniques gave rise to the contention that this special category of the autonomic nervous system is present, not only in the intestinal wall (Selverston et a / . , 1976; Gershon and Erde, 1981; Wood, 1984), but also in other organs receiving the autonomic innervation including the heart (Crowe and Burnstock, 1982; Reinecke et a / . , 1982; Lundberg e t a / . , 1983; Gu et al., 1984; Hassall and Burnstock, 1984; Moravec and Moravec, 1984; Weihe et al., 1984). According to the electrophysiological and pharmacological studies (Hirst e t a / . , 1973; Erde at a / . , 1980; Nozdrachev and Vataev, 1981; Wood, 1981; North, 1982), the enteric nervous system is considered as the best example of an independent integrative system composed of subsets of interneuronal circuits that control the respective intestinal segments and ensure the appropriate function of the whole organ (Wood, 1984). Each segment is composed of a network of interneurons that processes the sensory information and controls the activity of the respective motor neurons. The stereotyped reflex behavior is mediated by the integrative circuits which also generate the motor patterns of the individual intestinal segments. The synaptic connections between the adjacent neuronal subsets ensure their reciprocal coordination that is necessary for peristaltic movements. The role of the central nervous system is restricted to the modulation of these preprogrammed circuits. The intestinal tract can thus be considered as a self-governed fluid-propelling system possessing its own little brain (Weems, 1981). VI. Morphological Basis of the Rhythmical Activity of the Heart: A Working Hypothesis

A similar situation may also prevail in the mammalian heart, the complex structure of which is also derived from a primitive cardiac tube (Anderson et d.,1978). The development of the intracardiac septum and the separation of the outflow trunk into the ascending aorta and pulmonary artery

136

JOSEF MORAVEC A N D MIREILLE MORAVEC

are initiated by the migration of the ectomesenchymal elements derived from the cranial neural crest (Kirby, 1985). The presence of ganglion cells within the interatrial septum also depends on the migration of the ectodermic elements (Kirby and Stewart, 1984). It has been suggested that neural crest cells from both cranial (S,-S,) and thoracic segments 6,"Szo)may contribute to the embryogenesis of the intracardiac conduction system (Taylor, 1977; Stewart and Kirby, 1985). In postnatal life, they may constitute a heterogeneous nerve network interconnecting the upper segments of the left ventricle with the specialized structures of the right atrium (Wensing, 1965; Bojsen-M#ller and Tranum-Jensen, 1972; Weihe et al., 1984; Forsgren, 1985). Such a neuromuscular network may provide for the anatomical substratum of the intracardiac servomechanism involved in the local control of the electrical activation of the heart and the propulsion of the blood through the cardiac cavities (Fig. 18). The strategic position of richly innervated structures of the atrioventricular junction between the venous return and arterial portion of the systemic circulation is, in this respect, quite significant. In other words, the intracardiac conduction system of higher vertebrates resembles a neuromuscular oscillatory system similar to the oscillatory networks responsible for rhythmical movements of some of lower species (Fields and Kennedy, 1965; Bullock et al., 1977; Connor, 1985). The following observations seem to argue in the sense of such a possibility. ( I ) Presence of sensory specializations (mechanoreceptors)which is necessary for beat-to-beat modulation of these oscillatory systems (Weihe et a/., 1984; Selverston and Moulins, 1985). In some cases (Altman, 1982), a phasic sensory input is responsible for the generation of the oscillatory patterns to be observed. As concerns the rhythmical depolarizations of the nodal tissue, it has been experimentally demonstrated that they are under the control of mechanical stretch. It has been shown that a phasic sensory input can both entrain the pacemaker as well as reset it to a new rhythm (Brooks and Lu, 1972; Pollack, 1977; Irisawa, 1978). (2) Existence of the intracardiac neurosecretory (adrenergic and peptidergic) neurons which, together with the highly specific histotopography of the intrinsic multitarget nerves, seems to suggest that multiple and complex transmitter interactions can occur at the level of the interatrial septum. This local neuromodulation may explain some of the physiological properties of the cardiac pacemakers (existence of true pacemaker cells and follower cells) and intracardiac conduction tissue [dual atrioventricular node (AVN) entry, slow and fast intranodal pathways, AVN-NH conduction delay, etc.] (Janse et al., 1978). In particular, it has been suggested that a sudden catecholamine release from tissue-binding sites may trigger the sinus node (SN) depolarizations (Cranefield, 1978; Pollack, 1978). Catecholamine re-

I37

IN'IKINSIC NERVE PLEXUS OF MAMMALIAN HEART CNS

'

I

' ''

\ \

\

\

\ \

MECHANO-

VENTRICLES

RECEPTOR

SA N O D E

I

Fic;. 18. Model of the functional arrangement of the neuromuscular network of the interatrial septum responsible for the intrinsic generation of rhythmical coordinated contractions of cardiac cavities. The sinus node can be considered a motor pattern generator. The rate of its triggered depolarizations as well as the intracardiac conduction are supposed under the control of the intrinsic nerve plexus. which constitutes a local integrative component. The latter can process the efferent stimuli from the C N S as well as the sensory feedback from intracardiac mechanoreceptors necessary for a fast beat-to-beat modulation of cardiac excitability and contraction. SA. Sinoatrial node.

cycling between the intracardiac adrenergic cells and tissue receptors could thus be responsible for the sustained rhythmical activity of different cardiac preparations (Pollack, 1977; Cranefield, 1978; Zipes, 1981). According to our data, large sensory neurons of the atrioventricular junction may contribute to catecholamine traffic at the level of the interatrial septum (Fig. 19). The existence of a positive sensory feedback between the upper segment of the interventricular septum and the S N area is therefore not excluded. The adrenergic nature of this sensory feedback could be particularly adapted for the control of the heart rate: the same neurotransmitter (norepinephrine or dopamine?) may stimulate directly the nodal cells (Cranefield, 1978; Irisawa, 1978). having at the same time, an inhibitory effect on cholinergic nerves supplying the pacemaker area (Levy and Martin, 1979; Sole clt d.,1982; Loffelholz and Pappano, 1985). This multiplicity of norepinephrine action may considerably amplify its overall effect on

138

JOSEF MORAVEC AND MIREILLE MORAVEC

HlSSlAN

FIG. 19. Sinus node and atrioventricular junction as relaxation oscillators. Schema of the extrinsic innervation and of the local adrenergic feedback between the interventricular septum and cardiac pacemaker. Phasic stimulations of the sensory neurons of the AV junction may be coresponsible for cyclic catecholamine release and reuptake resulting in triggering and synchronization of the sinus node. It can be also involved in the modulation of vagal input to the heart, which would amplify indirectly the effects of sympathetic stimulation. PG. Epicardial cholinergic ganglion; ISG, intraseptal ganglion; SG, stellate ganglion; C,-T,, cervical and upper thoracic sympathetic ganglia; SN, sinus node; AVN. atrioventricular node; RB, right branch; LB, left branch.

the electrical properties of the pacemaker cells. The above catecholamine release is believed to be phasic, since it is under the control of mechanical events accompanying cardiac contractions (Brooks and Lu, 1972; Pollack, 1977; Weihe et al., 1984). This may explain the dynamic nature of vagusSN interactions (Jalife et al., 1983; Michaels et a/., 1983) and some badly understood aspects of sympathetic-parasympathetic interactions (Levy and Martin, 1979; Levy, 1984). These latter observations suggest that different segments of the intracardiac conduction system should not be considered functionally isolated. It would seem that specialized cells together with the respective nervous components constitute a self-governed nonlinear dynamic system with variable transit functions (Yates, 1983; Van Rossum et al., 1984). Such a highly organized network would not only react to input stimuli, but it should be able to process the input informations and control its own environment (Aplevich, 1968). These are the obvious prerequisites for the efficient and well-integrated cardiac function (Wurster, 1984).

INTRINSIC NERVE PLEXUS OF MAMMALIAN HEART

I39

VII. Conclusion

The above hypothesis will need a thorough verification. The morphological data presented in this review have to be completed by new electrophysical and pharmacological evidence. However, the upward axonal projections of the adrenergic neurons of the atrioventricular junction may provide for the retrograde limb of the connecting loop transforming the two cardiac pacemakers into a system of coupled relaxation oscillators (Van der Pol, 1940). In such a system, the atrioventricularjunction should not be any more considered as functionally subordinated to the sinus node and simply transmitting the excitation values from the right atrium to the left and right ventricles. The electromechanical events that accompany any heartbeat and which are perceived at this level may, after local processing influence the timing of the forthcoming sinus node depolarization and modulate the execution of the next mechanical systole (Covell et d., 19811. The need for such an arrangement has been predicted by James in his original "Brief Review" (James, 1973). At that time, the internodal communications were not completely understood, but according to James, the closing limb of the internodal loop should combine both a neural (reflex) factor and mechanical (hemodynamic)events. According to our data, both of these parameters might be interacting at the level of a single cell of the terminal nerve plexus of the rat atrioventricular junction (Moravec and Moravec, 1984; Moravec et a / . , 1986).

ACKNOWLEWMENTS

Professor R. Couteaux and Professor J. Taxi are thanked for their encouragement and for their critical advice. We also would like to thank Mrs. M.-T. Dronne for her patience in typing our article and Mrs. A. Courtalon for her skillful technical assistance.

REFERENCES Aars. M. (1971). Ac/ci P/iv.siol. Scwnd. 83, 133-138. Abraham. A. ( 1969). "Microscopic Innervation of the Heart and Blood Vessels in Vertebrates Including Man." Pergamon, Oxford. Akkeringa. L. J. (1949). Ac/ii Nrerl. Morpliol. Norin. Poihol. 6, 1-1 I . Alexandrowicz. J . S. (1951). Q. J . Micwsc. Sci. 92, 163-199. Altman. J. (1982). Trcvds NrrrroSci. (Peru. Ed.) 5, 257-261. Anderson. R. H. (1972).J. Ancir. 113, 197-211. Anderson. R. H.. Becker. A. E., Wenink. A. C., and Janse. M.J. (1978). In "The Conduction System of the Heart." (M. D. Wellens. I. Liek, and M. J. Janse, eds.). pp. 3-28. Nijhoff. The Hague. Angelakos. E. T.. Fuxe. K.. and Torchina, M. L. (1963). A m i Pliysiol. Sccind. 59, 184192.

140

JOSEF MORAVEC AND MlRElLLE MORAVEC

Aplevich. J . D. (1968). In "Neural Networks." pp. 110-1 IS. Springer-Verlag, Berlin and New York. Armour. J. A., and Hopkins, D. (1984). I n "Nervous Control of Cardiovascular Function" (W.C. Randall. ed.). pp. 20-45. Oxford Univ. Press, London and New York. Armour. J . A.. Randall, W. C.. and Sinha. S. (1975). Am. J. Physiol. 228, 141-148. Aschoff, L. (1910). Verh. D/sch. f u f h . Ges. 14, 3-31. Auerbach, L. (1864). Virchows Arch. Pufh. Anuf. Physiol. 30, 457-460. Axelsson. S.. Bjorklund, A., Falck, B., Lindvall, 0.. and Svensson. L. A. (1973). Aero Physiol. Sciind. 81, 57-62. Barker, D. (1974). I n "Handbook of Sensory Physiology" (C. C. Hunt, ed.), pp. 1-190. Springer-Verlag. Berlin and New York. Bechterev. V. M. (1896). "The conductivity pathways of the spinal cord and brain." St. Peters burg . Belmonte. C.. Simon, J., Gallego, R., and Baron, M. (1972). Bruin Res. 43, 25-35. Bernard, C. (1851). "Oevies." Baillitre, Paris. Berthold, C. H. (1966). J . Ulfrcisfrucf. Res. 14, 254-267. Bidder, C. H. (1868). Pflitcgers Arch. Gesutnre fhysiol. Menschen Tiere 1, 1-50. Black. I. B.. Coughlin, M. D.. and Cochard. P. (1979). Soc. Neiirosci. Svmp. 4, 187-207. Boeke. J. (1936). Z. Mikrosk. Anur. Forsch. 34, 330-342. Bojsen-Mdller. F.. and Tranum-Jensen. J. (1971). J . Anut. 108, 375-386. Bojsen-Moiler. F.. and Tranum-Jensen, J. (1972). J. Anof. 112, 367-382. Bolton. T. B. (1976). I n "Avian Physiology" (P. D. Sturkie. ed.). pp. 1-28. Springer-Verlag. Berlin and New York. Bosnjak. I. J.. Seagard. J. L.. and Kampine. J. P. (1982). Am. J . fhysiol. 242, R237-243. Brooks. Ch. Mc C.. and Lu, H. H. (1972). "Sinoatrial Pacemaker of the Heart.'' Thomas. Springfield, Illinois. Brooks. Ch. Mc C.. and Lange. G . (1977). Proc. Niirl. A i d . Sci. U.S.A. 76, 1761-1762. Brooks. Ch. Mc C. (1978). "Integrative Function of the Autonomic Nervous System." Univ. of Tokyo Press, Tokyo. Brooks, Ch. Mc. C. (1981). J. Airton. Nerv. Sysf. 4, 115-120. . (London) 191, 271-288. Bullock. T. H.. Orkand. R., and Grinnell, A. (1977). "Introduction to Nervous System." Freeman, San Francisco. California. Burn, J . M. (1975). "The Autonomic Nervous System." Blackwell. Oxford. Burnstock. A. ( 1969). Phurmcicol. Rev. 21, 247-324. Burnstock. A.. Evans, B.. Cannon, B. J., Health. J. W.. and Jones, V. (1971). Br. J . Phtrrmticol. 43, 295-301. Burnstock, G . (1972). Phcir~niicol.Rev. 24, 509-581. Burnstock. G. (1976). Niwroscience 1, 239-248. Burnstock. G. ( 1978). frog. Neitrobiol. (OxfbrdJ 11, 205-222. w York. Burnstock. G . . H6kfelt. T., Gershon. M. D.. Iversen. L. L., Kosterlitz, W.. and Szurszewski, J. H. (1980). Ncwrosci. Res. Progrum Birll. 17, 378-519. Cabot. J. B., and Cohen. D. H. (1980). In "Heart and Like Organs" ( G . H. Bourne, ed.). Vol. I. pp. 199-259. Academic Press, New York. Cajal. S. R. (1894). C . R . Sore. Biol. Ses Fil. 5, 217-223. Cantino. D.. and Mugnani. E. (1975). J . Neurocy/ol. 5, 505-536. Cauna, N., Naik, N. T.. Leahing, D. B., and Alberti, P. (1961). Bib/. Anur. 2, 90-96. Chiba. T. (1973). Ancrf. Rac. 176, 3 5 4 8 . Chiba. T.. and Yamauchi. A. (1970). Z . Ze/~or.sch.Mikrosk. Ancti. 108, 324-338.

INTRINSIC NERVE PLEXUS O F MAMMALIAN HEART

141

Coleridge, H. M.. Coleridge. J. C. C., and Kidd, C. (1964). J. Physiol. (London) 174, 323329. Coleridge, H . M.. Coleridge, J. C. C., Dangel. A., Kidd, C.. Luck, J. C., and Sleight, P. 33,. 87-97. (1973). Girl'. R ~ J X Connor. J. A. (1985). Annrr. Rev. Physiol. 47, 17-28. Coraboeuf. E. (1982). I n "Excitation and Neural Control of the Heart" (M. N. Levy and M . Vassale. eds.). pp. 1-35. Am. Physiol. SOC.,Bethesda, Maryland. Costa. M.. Furness. J. B., Franco. R.. Llewellyn-Smith, I., Murphy, R., and Beardsley. A. M. (1982). Cihu Found. Svmp., 83, 129-144. Couteaux, R.. and Laurent, P. (1957) C . R. A i ~ r d Sci. . 245, 2097-2100. Couteaux, R.. and Laurent. P. (1958). C. R. Assoc. Ancrf. 97, 230-234. Couteaux. R. (1961). Acfrriil. Nertroplivsiol. 3, 143-173. Couteaux. R. (1978). "Recherches Morphologiques et Cytochimiques sur I'Organisation des Tissus Excitables." Robin and Mareuge, Pans. (19%). C . R . Assoc. A w / . 97, 230-234. Cranefield, P. F. (1978). I n "The Sinus Node" (F. I. M. Bronke, ed.), pp. 348-356. Nijhoff, The Hague. Cranefield, P. F. (1983). Am. J. Physiol. 245, H90LH910. Crowe, R.. and Burnstock. G. (1982). Crrrdiovusc. Res. 16, 384-390. Covell, J. W., Lab, M. J.. and Palavec, R. (1981). J. Physiol. (London) 320, 34P. Cuello, A. C. (1981). Nertroscience 6, 41 1 4 2 2 . Cuello. A. C. (1983). Fed. Pro 1.74 pH units, reflecting a heterogeneous population of His residues in the poorly active analogs, and presumably a greater heterogeneity of GnRH conformers (Milton et al., 1983). cGnRH I falls into this group and has a relative potency of 1-5% in mammalian systems (Sandow et al., 1978; Millar and King, 1983a; Milton et al., 1983). In the bird, these structural requirements of the receptor clearly do not pertain, as cGnRH I is equipotent with mGnRH in chickens (Johnson et al., 1984; Sterling and Sharp, 1984) and quail (Chan et al., 1983) in vivo and in chicken pituitary cells in vitro (Millar and King, 1983a; Milton et al., 1983). Using chicken pituitary membranes, these biological effects were shown to be directly related to receptor-binding affinity (Millar and King, 1983a). We have now defined the requirements of the chicken pituitary GnRH receptor in more detail by comparing the ability of a range of position eight-substituted GnRH analogs to release LH from dispersed chicken pituitary cells. Relative to cGnRH I, Arg' and Phe8 analogs have full LH-releasing activity. Met', His', and Leu' analogs exhibit about 30% activity. Ser', Trp', Cit', and Ile' analogs have 1&20% activity; and only the acidic residue Glu' had low LH-releasing activity (R. P. Millar, J. A. King, and R. W. Roeske, unpublished). Thus, accepting the validity of the structural conformer stabilization model for the mammal, one must conclude that the avian receptor is promiscuous and binds a number of GnRH conformers present in the unstabilized analogs which have substitutions for Arg'. However, this proposal does not exclude the possibility that the importance of a basic amino acid in position eight of GnRH for biological activity is simply related to a charge interaction with a negative charge at the binding site of the mammalian GnRH receptor. Studies on the influence of the position eight amino acid on gonadotropin-releasing activity in reptiles, amphibians, and fish are much less extensive. As indicated above, cGnRH I is active in amphibians and fish (Licht et al., 1984; Peter et al., 1985), cGnRH I1 in certain reptiles (Phillips et al., 1983, and sGnRH is active in fish (King et al., 1984b; MacKenzie et al., 1984; Peter et al., 1985) (Table 11). Although these findings are not sufficiently extensive to draw definite

168

ROBERT P. MILLAR AND JUDY A . KING

conclusions regarding the requirements for the amino acid in position eight in these nonmammalian vertebrates, they do indicate that the pituitary GnRH receptor in these classes is similar to that of the bird in tolerating considerable amino acid variation in this position.

2. Effects of Conformational Constraint on GnRH Bioactivity In support of the postulate that Args plays a role in stabilizing GnRH conformation and that this is important for receptor binding and biological activity in mammals, Freidinger et al. (1980) reported that a GnRH analog containing a y-lactam bridge (between the C of Gly6 and N of Leu7) which stabilizes the (3-turn of residues five to eight of GnRH, was 9 times more potent than native GnRH in stimulating LH release from dispersed rat pituitary cells. This analog is also 9-10 times more active in sheep pituitary cells (R. P. Millar, J. A. King, R. W. Roeske, unpublished). On the basis of the hypothesis that the avian receptor does not require Arg' for conformational stabilization and will bind a number of GnRH conformers, it follows that the y-lactam conformational restraint would not enhance bioactivity in the chicken bioassay. In a recent study, we observed that the stimulation of LH release from dispersed chicken pituitary cells by native mGnRH and the y-lactam analog over the dose range of 10-11-10-6M was indistinguishable (R. P. Millar, J. A. King, and R. W. Roeske, unpublished). These findings, therefore, support the concept that conformational stabilization is less important for GnRH interaction with its receptor in the chicken and quail, and probably in birds in general. The effect of this conformational constraint on GnRH activity in other nonmammalian vertebrates has not been established. 3. Influence of Positions Five and Seven on Biological Activity Studies in mammalian bioassays established that some modifications such as N-methyl-Leu7 actually enhance activity (Ling and Vale, 1979, presumably through stabilizing the (3-turn of GnRH (Chandrasekaren et al., 1973). Ile7-, Nle7-, Ser7-, and (Boc)Lys7-substituted mGnRH analogs retain significant activity in mammalian systems (Sandow et al., 1978), suggesting a degree of nonspecificity of the mammalian receptor for the amino acid in this position. Trp7-mGnRH has high activity when compared with the parent compound (R. P. Millar, J. A. King, and R. C. deL. Milton, unpublished) and the presence of Trp7 in cGnRH I1 and sGnRH appears to enhance binding to the mammalian receptor when compared with cGnRH 1. The mammalian receptor is, however, not totally indiscriminate in binding position seven-substituted analogs, as Gly7, Ala7, Val7, Lys7,

EVOLUTION OF GONADOTROPIN-RELEASING H O R M O N E

169

Arg7, D-L~u'.and Pro7 substitution leads to a marked decline in activity (Sandow et al., 1978). The mammalian GnRH receptor appears to differ from the chicken GnRH receptor in its requirements for the amino acid in position five. His5enhances biological activity in the chicken and mammalian bioassays when incorporated in cGnRH 11. This substitution in mGnRH results in an increase in activity in the mammal, while a marked decrease in activity ensues in the chicken (R. P. Millar, J. A. King, and R. C. deL. Milton, unpublished). Ala' and Pro' substitution results in reduced gonadotropinreleasing activity in the mammal, while Phe'-mGnRH retains substantial activity (Sandow et ul., 1978). Our observation that sGnRH (Trp7,L e d mGnRH) is 2.5-fold more active than cGnRH I (Gln8-mGnRH) in stimulating LH release from chicken pituitary cells indicates that the avian receptor is also tolerant of alterations in position seven of GnRH. However, it is clear that the combination of substitutions in positions seven and eight can be important as Gln7, Leu8-mGnRH has only 4% relative potency in the chicken pituitary cell system (King et al., 1983). As LeunmGnRH has 30% relative activity, it appears that Gln in position seven is largely contributory to the decline in activity. cGnRH 11 (His', Trp7, Tyr8-mGnRH) is more active than cGnRH I and mGnRH in stimulating gonadotropin release from dispersed chicken pituitary cells (Millar et al., 1986). Since sGnRH, which shares Trp7 in common with cGnRH 11, was also more active than cGnRH I , this residue appears to be responsible for the enhancement in activity. Indeed, Trp7mGnRH is 2.8 times more active than mGnRH in the chicken system (R. P. Millar, J. A. King, and R. C. deL. Milton, unpublished). It also appears that His substitution for Tyr' is acceptable to the chicken receptor and may even contribute to the enhanced activity of cGnRH 11. However, the nature of other residues in positions seven and eight is clearly important since His'-mGnRH was actually much less active than any of the natural GnRHs, and His'. Trp7-mGnRH was less active than cGnRH 11 (R. P. Millar, J . A. King, and R. C. deL. Milton, unpublished). Comprehensive data on the influence of amino acids five and seven for GnRH activity in other nonmammalian vertebrates are lacking. The high biological activity of all the naturally occurring vertebrate GnRHs suggests a considerable tolerance of amino acid substitutions in these positions.

4. Structural Requirements for Superactive GnRH Agonists It is now well established that substitution of D-amino acids for Gly", N-methyl-Leu for Leu7, and N-ethylamide for Gly"' - NH2 in GnRH en1978) hances gonadotropin-releasing activity of the peptide (Sandow et d.,

I70

ROBERT P. MILLAR A N D JUDY A. KING

(Fig. 5). In view of the less basic nature and different hydrophobicity of both forms of chicken GnRH and sGnRH and the different requirements of the nonmammalian vertebrates’ pituitary GnRH receptors, the same principles for producing agonists with enhanced activity need not necessarily apply. a. Analogs of Mammalian GnRH. D-Trp6-mGnRH exhibits an enhancement of 26-fold in stimulating L H release from chicken pituitary cells, which is similar to the 36-fold increased activity in sheep pituitary cells (Millar and King, 1983b). o-Ser6(Bu‘),Prog-NHEt-mGnRH,which has enhanced activity in mammals, exhibits increased activity in chickens in vivo (Sterling and Sharp, 1984). On the other hand, ~-His~(Bzl),Pro’-NHEtmGnRH, which has a relative potency of about 50 in the mammal, exhibits only a 4-fold enhancement in the chicken system (R. P. Millar and J. A. King, unpublished). Similarly D-Leu6-mGnRH, which is 27 times more active than mGnRH in releasing L H from rat pituitary cells, has no enhanced activity in chicken pituitary cells (Hasegawa et d . , 1984). A new (J. E. Rivier and H. GnRH agonist, ~-Glu~(anisole),Pro~-NHEt-mGnRH Anderson, unpublished), also displayed no enhancement in activity in chicken pituitary cells in contrast to a 4-fold increase in activity in the sheep pituitary cell bioassay. Similarly D - ~ A ~ ~ ~ ( E ~ , ) , P I - o ’ - N H E ~ - ~ G ~ R H , which is a very active analog in the rat, is only twice a s active as mGnRH in the chicken pituitary cell bioassay (J. J. Nestor and R. P. Millar, unpublished). It is clear, therefore, that the enhanced activity of mGnRH agonists in mammalian systems is Frequently not paralleled by their activity in the chicken. This is likely to be due to the differences in GnRH receptors described earlier and may be related to the fact that conformational constraint of GnRH does not affect biological activity in the chicken system to the same extent as in the mammal. Fewer data on comparative potencies of mGnRH agonists in other nonmammalian vertebrates are available. ~-His~(Im-Bzl),Pro’-NHEt-rnGnRH is 45 times more potent than mGnRH in the bullfrog (McCreery et a f . ,

AGONIST

pGlul

ANTAGONIST

.\“:“i 1

D-AMINO ACID (ESPECIALLY AROMATICS)

-

His2

-

Trp3

1 1

D-AMINO ACID (ESPECIALLY BULKY AROMATICS)

-

Ser4

-

Tyr5

-

Gly

-

Leu7

ETHYLAMIDE

- Arg8 -

D-AMINO ACID (ESPECIALLY BASICS)

Fa. 5. Structure of GnRH agonists and antagonists.

Pro9

-

1

Glylo*NH2

.c

D-Ala

EVOLUTION OF GONADOTROPIN-RELEASING HORMONE

171

1982) in accordance with mammalian data. This analog is approximately 4 times more active than mGnRH in the goldfish (Peter et al., 1985), which is similar to our observations in the chicken (R. P. Millar and J. A. King, unpublished). In turtles, this analog does not increase plasma gonadotropins or steroid levels (Licht et al., 1982). In the frog, D-Ser"(Bu'),Pro9NHET-mGnRH has been shown to stimulate testicular steroidogenesis both in vivo and in vitro (Pierantoni et a l . , 1984). ~-Ala",pro'-NHEtmGnRH, which is 14 times more active than mGnRH in the rat, is active in the African catfish (de Leeuw et al., 1985),but does not exhibit enhanced activity in the goldfish (Peter et al., 1985). Other information on the actions of GnRH analogs in lower vertebrates has been reviewed (L. W. Crim, 1984). b. Analogs ofChicken GnRH I . D-T~~"-cG~RH I is 26 times more potent than the parent peptide in the chicken pituitary cell bioassay (Millar and King, 1983b), which is identical to the increased activity when ~ - T r p "is incorporated in mGnRH. This analog is 100 times more potent than cGnRH I in the sheep pituitary cell bioassay (R. P. Millar and J. A. King, unpublished). In goldfish, D-Trp"-cGnRH I exhibited only a 5-fold increase in activity (Peter et al., 1985), suggesting a difference in the GnRH receptor in birds and fish. However, the analog was approximately 20 times more active than the parent peptide in trout (L. W. Crim, personal communication). D-hArg"(Et,)-cGnRH I exhibited a 5-fold increase in activity in the chicken pituitary cell bioassay (J. J. Nestor and R. P. Millar, unpublished). c. Analogs of Chicken GnRH ZZ. D-Arg" incorporation in cGnRH I1 led to a 4-fold increase in activity in the chicken pituitary cell bioassay (Millar rt al., 1986). D-hArg"(Et,) resulted in no enhancement in the chicken system, but increased the activity 46-fold in the sheep pituitary system when compared with the poor activity of the parent peptide (J. J. Nestor and R. P. Millar, unpublished). The D-Arg" analog appears to exhibit good activity in the iguana (Phillips et al., 1985). d. Analogs of Salmon GnRH. D-Arg", Pro9-NHEt-sGnRH had an 8fold increase in activity in chicken pituitary cells (R. P. Millar, J. E. Rivier, and J. A. King, unpublished) and a 12-fold increase in the goldfish (Peter rt ul., 1985). D-His" (Im-Bzl), Pro9-NHEt-sGnRH was also 8 times more active in the chicken system, but exhibited no significant increase in activity in the goldfish as did a D-Ala" analog (Peter et al., 1985). In summary, it is apparent that the majority of GnRH analogs which are superactive agonists in mammals exhibits relatively little or no increase in activity in the chicken and probably also the other nonmammalian vertebrates. This emphasizes the differences in structural requirements of the respective receptors. Differences in metabolic clearance may con-

I72

ROBERT P. MILLAR A N D JUDY A. KING

tribute in the in vivo studies, but are unlikely to be of much significance in the in vitro bioassays. Although there is a much closer parallelism of activity in birds and fish, there may also be differences in these receptors, as certain analogs (e.g., DHis6 (Im-Bzl), Pro9-NHEt-sGnRH)have marked differences in activity in the chicken and goldfish bioassays. The studies on GnRH analog activity in the chicken pituitary cell bioassay provide the most comprehensive data and demonstrate that, in general, substitution of a particular D-amino acid for Gly6 results in a similar increase in activity when substituted in any of the naturally occurring vertebrate GnRHs. An exception to this is D-hArg6(Et,) substitution in sGnRH, which is considerably more active than the corresponding mGnRH, cGnRH I, and cGnRH 11 analogs. With the limited information available, it is not possible to formulate definite rules concerning the types of GnRH analogs likely to exhibit enhanced activity in nonmammalian vertebrates, as has emerged for analogs in mammals. However, the sequence of activity of the D-hArg6(Et,)analogs (sGnRH-A > cGnRH 11-A > cGnRH I-A > mGnRH-A) relates to the relative hydrophobicity of these analogs, as has been demonstrated for GnRH analogs in mammals (Nestor et al., 1984).

5 . Structural Requirements for GnRH Antagonists GnRH analogs which have antagonist activity in mammals (Fig. 5 ) were tested for their ability to inhibit the LH responses of sheep and chicken pituitary cells to cGnRH (J. A. King and R. P. Millar, unpublished). The analogs were also tested alone for their intrinsic LH-releasing activity. The most potent antagonist in the chicken pituitary bioassay was Ac~-Phe',~-pCl-Phe~,D-Trp-'.~,D-Ala'~-mGnRH with an IC,, of 5 x lo-' M and no intrinsic LH-releasing activity. In the sheep pituitary cell bioassay, this analog was intermediate in activity with an IC,, of 2.5 x M. ~-pGlu',~-Phe',~-Trp'.~-mGnRH had slightly lower antagonist activity in the chicken system while, it was a very weak antagonist in the sheep bioassay. Ac-~-pCl-Phe'~~,~-Trp~~~,~-Ala'~-mGnRH had intermediate anhad h e similar ~ tagonist activity in both bioassays, and ~ - P h e * , ~ - T r p ' , ~ - P activity in the chicken bioassay, but was a weak antagonist in the sheep. The most active antagonist in the sheep bioassay was Ac-D-pCI-Phe'.',DTrp',~-Phe~,~-Ala''-mGnRH (IC,, lo-'' M ) . This antagonist had the M ) and weakest antagonist activity in the chicken bioassay (IC,,, displayed agonistic activity when administered alone, with an ED,,, of M. There are, therefore, marked differences in the antagonistic and agonistic activities of these analogs in the chicken and sheep bioassays, which fur-

EVOLUTION OF GONADOTROPIN-RELEASING HORMONE

I73

ther indicates differences in the receptors with respect to their GnRH structural requirements for interaction with the NH2-terminal part of the molecule. Differences in this regard are perhaps unexpected, as the sequence of the NH,-terminal region of GnRH (amino acids 1-4) has been completely conserved in vertebrate GnRHs, and a similar conservation of the receptor in its interaction with this region might have been anticipated. An antagonist incorporating the natural Trp7of cGnRH I1 and sGnRH rather than the Leu7 of rnGnRH has been shown to have increased potency in rats (Folkers et al., 1984). Among lower vertebrates, a mGnRH antagonist (Ac-dehydro-Pro',pCI~-Phe',~-Trp',~,NaMeLeu~-mGnRH) has been shown to block mGnRHinduced gonadotropin release in the bullfrog (McCreery et al., 1982). DPhe2,Phe3,D-Pheb-mGnRHinhibits mGnRH-stimulated gonadotropin release in trout (Crim et al., 1981). 6. Pituitary Desensitization Since pituitary desensitization to prolonged and/or high doses of GnRH is a receptor-mediated event in the rat (Smith and Conn, 1984), it was of interest to determine the influence of the differences in the nonrnammalian receptor on expression of desensitization. When dispersed chicken pituitary cells suspended in a Biogel column were stimulated with 2-minute M cGnRH I every 30 minutes for 3% hours, a LH response pulses of was associated with every pulse. In contrast, a definite desensitization of the pituitary cells occurred when they were continuously perifused with M cGnRH I or 10-7M D-Trp6-cGnRHI (Millar and King, 1984). After 100 minutes of perifusion, LH release declined to basal levels. In view of our difficulty in satisfactorily quantitating GnRH receptors in chicken pituitary cells or membranes (Millar and King, 1983a),we have been unable to determine whether the receptor "down-regulation" plays a major part in pituitary desensitization, as in the rat (Zilberstein et al., 1983). In chickens in vivo, daily injections of D - S ~ ~ ~ ( B U ' ) , P ~ ~ ~ - N H E ~ - ~ G ~ did not reduce pituitary responsiveness to the analog (Sterling and Sharp, 1984). Prolonged doses of D - S ~ ~ ~ ( B U ' ) , P ~ O ~ - N H E have ~ - ~been ~ G ~reRH ported to induce desensitization in lizards in vivo (Ciarcia et al., 1983). while in turtles this desensitization phenomenon has not been observed (Licht et al., 1982). The perifused bullfrog pituitary continues responding to sustained high doses of GnRH and lacks the phenomenon of desensitization (McCreery and Licht, 1983). The goldfish pituitary is reported to exhibit desensitization when exposed to superactive GnRH analogs (Peter, 1980).

I74

ROBERT P. MILLAR AND JUDY A. KING

IV. Conclusions

It is evident that heterogeneous molecular forms of GnRH are present within vertebrates. GnRH structure also varies in different tissues within the same vertebrate species and even within the same tissue, as occurs in the brains of several nonmammalian vertebrate species. At present, four forms of GnRH have been structurally characterized: the original GnRH from porcine and ovine hypothalamus; Gln8-GnRH and His’,Trp7,Tyr8-GnRHin chicken hypothalamus; and Trp7,Leu8-GnRHin salmon brain. Immunological, chromatographic, and biological properties indicate the presence of other forms of GnRH in mammalian and nonmammalian tissues which await characterization. All of the GnRHs exhibit gonadotropin-releasing activity, especially in nonmammalian vertebrates in which the pituitary receptor is relatively nondiscriminatory in binding GnRHs with substitutions in positions five, seven, and eight. However, it is conceivable that the GnRHs have other biological activities, such as is demonstrated by the stimulation of growth hormone secretion in fish, actions on the placenta, ovary, testis, and adrenal gland in certain mammals, and effects in the central and peripheral nervous systems. It appears, therefore, that the basic GnRH structure has been recruited through evolutionary selective processes to serve diverse functions. The specificity of these actions is achieved in a variety of ways: (1) through the evolution of different molecular forms of GnRH, which interact with tissue-specific receptors; (2) through paracrine regulation by anatomically closely localized cells, which does not allow the peptide to enter the general circulation in concentrations sufficient to bind other GnRH receptors. In this system, which appears to pertain in the gonads, the GnRH and receptors in different tissues can be identical; (3) through the existence of a “private” conducting system, as is found in the hypothalamohypophyseal portal system; and (4) through intimate contact between the secretory and target cells, as occurs in neuronal communication. In addition to the heterogeneity of GnRHs within vertebrates and in tissues of the same species, it is apparent that at least two forms of the peptide are generally present in the hypothalamus (or brain) of nonmammalian vertebrates. In mammals, only a single molecular form has been isolated, and only a single GnRH sequence has been detected in the mammalian genome (Seeburg and Adelman, 1984). However, the presence of a different form of GnRH in the mammalian pineal gland suggests that two or more forms are present in the mammalian brain (and possibly hypothalamus). Conceivably, the detection of only one GnRH-coding sequence in the mammalian genome may be due to significant structural

EVOLUTION OF GONADOTROPIN-RELEASING HORMONE

I75

differences in other form(s) which do not hybridize well with the mGnRH probe. The functional significance of two hypothalamic GnRHs in vertebrates has not been established. In the chicken, both forms of hypothalamic GnRH stimulate pituitary secretion of LH and FSH at concentrations appropriate for physiological regulation ( I),the chain deformation has not enough time to follow the intrapeptide excitation. In this case, the nonlinearity parameter (9) is negative and Eq. (6) in an infinite chain has no stationary solutions normalized by the condition (6a). In a finite chain of sufficiently large length (Nu >> a), such solutions have the form ’Note that in the recent monograph of Davydov (1984) the value x is used which is twice of that utilized in (7) and in his pioneer works on solitons.

EXCITONS A N D SOLITONS IN MOLECULAR SYSTEMS

@(c) = N - ” : ,

0

6

I95

N >> 1

uN,

In this case, the excitation is described by the function in Eq. (3) taking the form of a plane wave +(x,

r)

=

N-”2 exp[i(kx

-

Er/h)l

(12)

Hence, according to Eq. (7), all interpeptide distances u change by the same value p,,([)

=

-2X/NK(S’

- 1)

(13)

which tends to zero at N + x. In the same approximation, A = W = 0. The excitation described by the functions in Eqs. (12) and (13) is called an exciton. The excitons transport only the intrapeptide excitation energy. In this case at small velocities, this energy is determined by the expression E,,(v) = E,,(o)

+ +mv2 = EJO) + h2F/2m

(14)

where

E,,(O) =

t,, - h’/mu2 = to - 25

(15)

The energy levels corresponding to different values form the exciton band. The E,,(O) characterizes the position of the exciton band “bottom,” m is the effective exciton mass. At ? < I , the plane waves of Eq. (12) are also the solutions of Eq. (6). However, the slow excitons described by the above solutions are metastable, since in this case there are more stable localized excitationssolitons which are considered below. According to Eq. (12), the exciton states with the definite value of k (energy) are distributed uniformly along the whole length of the chain [ I +cx(.x,t) 1 = constant]. Therefore, they transfer no energy and information. Usually, in the chain, the states described by a superposition of plane waves in Eq. (12)-a wave packet-are excited. Since the exciton states in Eq. (12) are dispersive (the phase velocity of these waves hk/2m depends on the wavelength), wave packets, made up of them, smear. The exciton states, corresponding to the excitation of the amide I vibrations in real protein molecules characterized by a more complex distribution of peptide groups, are also described by plane waves. They are excited when the proteins absorb infrared radiation. The absorption spectra of infrared absorption by proteins were studied in detail by Chirgadze and Nevskaya (1976). They have shown that, using the exciton theory developed by Davydov (1951, 1968) for the interpretation of such spectra, one can obtain very important information about the secondary structure of protein molecules. In particular, studying the infrared absorption spec-

196

A. S. DAVYDOV

tra, they determined the magnitude of the dipole moment of the amide I vibrations (0.35 D) and the energies of the resonance interactions between peptide groups which occupy different positions in proteins with the configurations: a-helix. (3-sheets, and globular conformation. The exciton excitation by radiation of frequency w takes place on condition that the conservation law of energy Aw = E(k) and quasi-momentum hk = hw/c is fulfilled and if the projection of electric field intensity of the wave in the direction of the transition dipole moment (d)is different from zero. In this case, the wave packet, whose spatial extension is about the radiation wavelength, i.e., exceeds by thousands of times the distance between PGs, is excited. Instability of excitons formed is due to the wave packet smearing, the energy loss for the phonon radiation, and the possible transition to a more stable state-the soliton of lower energy. The intensive energy loss by excitons for the phonon radiation is caused by the great value of the coupling coefficient ( x ) of amide 1 vibrations with displacements of equilibrium positions of PGs. Owing to this strong coupling, the exciton, when moving, excites significantly the vibrations of PGs relative to their equilibrium positions. €4.

SOLITONS

At s’