Tumors of the Central Nervous System, Volume 2: Gliomas: Glioblastoma (Part 2)

Tumors of the Central Nervous System

Tumors of the Central Nervous System Volume 2

For other titles published in this series, go to www.springer.com/series/8812

Tumors of the Central Nervous System Volume 2

Tumors of the Central Nervous System Gliomas: Glioblastoma (Part 2) Edited by

M.A. Hayat Distinguished Professor Department of Biological Sciences, Kean University, Union, NJ, USA

123

Editor M.A. Hayat Department of Biological Sciences Kean University Union, NJ, USA [email protected]

ISBN 978-94-007-0617-0 e-ISBN 978-94-007-0618-7 DOI 10.1007/978-94-007-0618-7 Springer Dordrecht Heidelberg London New York Library of Congress Control Number: 2011923069 © Springer Science+Business Media B.V. 2011 No part of this work may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, microfilming, recording or otherwise, without written permission from the Publisher, with the exception of any material supplied specifically for the purpose of being entered and executed on a computer system, for exclusive use by the purchaser of the work. Printed on acid-free paper Springer is part of Springer Science+Business Media (www.springer.com)

“Although touched by technology, surgical pathology always has been, and remains, an art. Surgical pathologists, like all artists, depict in their artwork (surgical pathology reports) their interactions with nature: emotions, observations, and knowledge are all integrated. The resulting artwork is a poor record of complex phenomena.” Richard J. Reed MD

Preface

The primary objective of this series, Tumors of the Central Nervous System, is to present the readers with the most up-to-date information on the initiation, progression, recurrence, metastasis, and treatment of the CNS tumors. As in volume 1, volume 2 has discussed in detail biomarkers and diagnosis of gliomas, especially glioblastoma. The role of a large number of biomarkers in the diagnosis of glioblastoma is included. Advantages and limitations of the use of biomarkers for diagnosis are presented. The role of TP53 gene mutation in the initiation and progression of glioblastoma is presented as well as germline mutations of this gene. Role of oncogenes and tumor suppressor genes is also discussed. Also, is discussed the role of specific genes in the resistance to drug therapy. The importance of the use of imaging modalities (e.g., PET, CT, MRI, and SPECT) in clinical diagnosis, treatment assessment, and recurrence determination is pointed out. It is well established that early diagnosis is the key to cancer “cure”. Prognosis is highly dependent on the stage of the disease. Thus, a simple and reliable screening method would be of tremendous advantage. Imaging techniques in clinical practice are used for the staging of tumors, detection of tumor recurrence, monitoring of efficacy of therapy, and differentiation between malignant and benign tissues. In this volume, use of PET in diagnosing glioma and in assessment of biological target volume in high-grade glioma patients is explained. Also is discussed the use of MRI in glioma surgery. Present and future therapeutic drugs for malignant gliomas are described. The efficacy of several drugs, such as cyclosporine, interferon, heparin, and cannabinoids in treating glioblastoma is explained. Effectiveness of therapies, such as resection, radiation, chemotherapy, and immunotherapy, against high-grade gliomas is detailed. Therapy for recurrent high-grade glioma with bevacizumab and irinotecan is presented. Use of dendritic cell therapy and adenoviral vectors for glioblastoma is discussed. Brainstem gliomas are also described, so is tumor-associated epilepsy. This work consists of 37 chapters that were contributed by 101 authors representing 16 countries. The high quality of each manuscript made my work as the editor an easy one. Strictly uniform style of manuscript writing has been accomplished. The results are presented in the form of both black-and-white and color images and diagrams. I am indebted to the contributors for their promptness in accepting my suggestions, and appreciate their dedication and hard work in sharing their knowledge and expertise with the readers. Each chapter provides unique individual, practical knowledge based on the expertise of a large number of researches and physicians. A vast medical vii

viii

Preface

field such as tumors of the CNS can be discussed adequately only by a large number of experts. It is my hope that this volume will be published expediously. I am thankful to Dr. Dawood Farahi, Dr. Kristie Reilly, and Mr. Philip Connelly for recognizing the importance of scholarship in an institution of higher education, and providing resources for completing this project. Union, New Jersey September 2010

M.A. Hayat

Contents

1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . M.A. Hayat Part I

1

Biomarkers and Diagnosis . . . . . . . . . . . . . . . . . . . .

9

2 Gliomagenesis: Advantages and Limitations of Biomarkers . . . . . Michel Wager, Lucie Karayan-Tapon, and Christian-Jacques Larsen

11

3 Molecular Subtypes of Gliomas . . . . . . . . . . . . . . . . . . . . . Lonneke A.M. Gravendeel and Pim J. French

25

4 Glioblastoma: Germline Mutation of TP53 . . . . . . . . . . . . . . . Haruhiko Sugimura, Hidetaka Yamada, Shinji Kageyama, Yasuhiro Yamamura, Naoki Yokota, Hiroki Mori, Moriya Iwaizumi, Kazuya Shinmura, Kiyotaka Kurachi, Toshio Nakamura, Masaru Tsuboi, Masato Maekawa, and Tomoaki Kahyo

31

5 Familial Gliomas: Role of TP53 Gene . . . . . . . . . . . . . . . . . . Soufiane El Hallani and Ilham Ratbi

39

6 The Role of IDH1 and IDH2 Mutations in Malignant Gliomas . . . . Yukihiko Sonoda, Ichiyo Shibahara, Ryuta Saito, Toshihiro Kumabe, and Teiji Tominaga

47

7 Malignant Glioma: Isocitrate Dehydrogenases 1 and 2 Mutations . . Zachary J. Reitman and Hai Yan

53

8 Metabolic Differences in Different Regions of Glioma Samples . . . . Francisca M. Santandreu, Jordi Oliver, and Pilar Roca

63

9

73

Glioblastoma Patients: Role of Methylated MGMT . . . . . . . . . . Giulio Metro and Alessandra Fabi

10 Brain Tumor Angiogenesis and Glioma Grading: Role of Tumor Blood Volume and Permeability Estimates Using Perfusion CT . . . . . . . . . . . . . . . . . . . . . . . . . . . . Rajan Jain

81

ix

x

Contents

11

Vasculogenic Mimicry in Glioma . . . . . . . . . . . . . . . . . . . . Zhong-Ping Chen and Yin-Sheng Chen

12

Newly Diagnosed Glioma: Diagnosis Using Positron Emission Tomography with Methionine and Fluorothymidine . . . . Nobuyuki Kawai, Yoshihiro Nishiyama, and Takashi Tamiya

103

Role of Diffusion Tensor Imaging in Differentiation of Glioblastomas from Solitary Brain Metastases . . . . . . . . . . . Sumei Wang, Harish Poptani, Elias R. Melhem, and Sungheon Kim

113

13

93

14

131 I-TM-601

SPECT imaging of Human Glioma . . . . . . . . . . . . Adam N. Mamelak and David Hockaday

123

15

Assessment of Biological Target Volume Using Positron Emission Tomography in High-Grade Glioma Patients . . . . . . . . Habib Zaidi and Srinivasan Senthamizhchelvan

131

16

Skin Metastases of Glioblastoma . . . . . . . . . . . . . . . . . . . . Rebecca Senetta and Paola Cassoni

Part II 17

18

19

20

143

Therapy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

151

Diffuse Low-Grade Gliomas: What Does “Complete Resection” Mean? . . . . . . . . . . . . . . . . . . . . . . . . . . . . Johan Pallud

153

Quantitative Approach of the Natural Course of Diffuse Low-Grade Gliomas . . . . . . . . . . . . . . . . . . . . . . . . . . . Johan Pallud and Emmanuel Mandonnet

163

Impact of Extent of Resection on Outcomes in Patients with High-Grade Gliomas . . . . . . . . . . . . . . . . . . . . . . . . Debraj Mukherjee and Alfredo Quinones-Hinojosa

173

Glioma Surgery: Intraoperative Low Field Magnetic Resonance Imaging . . . . . . . . . . . . . . . . . . . . . . . . . . . . Christian Senft

181

21

Low-Grade Gliomas: Intraoperative Electrical Stimulations . . . . . Hugues Duffau

189

22

Malignant Gliomas: Present and Future Therapeutic Drugs . . . . . Linda Coate and Warren Mason

207

23

Recurrent Malignant Glioma Patients: Treatment with Conformal Radiotherapy and Systemic Therapy . . . . . . . . . Abhirami Hallock and Lauren VanderSpek

24

Glioblastoma: Boron Neutron Capture Therapy . . . . . . . . . . . . Tetsuya Yamamoto, Kei Nakai, and Hiroaki Kumada

25

Glioblastoma: Anti-tumor Action of Cyclosporin A and Functionally Related Drugs . . . . . . . . . . . . . . . . . . . . . Bozena Kaminska, Magdalena Tyburczy, Konrad Gabrusiewicz, and Malgorzata Sielska

215 229

241

Contents

xi

26 Glioblastoma Patients: Chemotherapy with Cisplatin, Temozolomide and Thalidomide . . . . . . . . . . . . . . . . . . . . . Fable Zustovich and Giuseppe Lombardi

255

27 Glioblastoma: Role of Galectin-1 in Chemoresistance . . . . . . . . . Florence Lefranc and Robert Kiss

261

28 Glioma-Initiating Cells: Interferon Treatment . . . . . . . . . . . . . Atsushi Natsume, Masasuke Ohno, Kanako Yuki, Kazuya Motomura, and Toshihiko Wakabayashi

269

29 Glioblastoma: Anti-tumor Action of Natural and Synthetic Cannabinoids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Aleksandra Ellert-Miklaszewska, Iwona Ciechomska, and Bozena Kaminska

277

30 Patients with Recurrent High-Grade Glioma: Therapy with Combination of Bevacizumab and Irinotecan . . . . . . . . . . Benedikte Hasselbalch, Ulrik Lassen, and Hans Skovgaard Poulsen

289

31 Monitoring Gliomas In Vivo Using Diffusion-Weighted MRI During Gene Therapy-Induced Apoptosis . . . . . . . . . . . . Timo Liimatainen, Olli Gröhn, and Kimmo Lehtimäki

301

32 High-Grade Gliomas: Dendritic Cell Therapy . . . . . . . . . . . . . Hilko Ardon, Steven De Vleeschouwer, Frank Van Calenbergh, and Stefaan W. Van Gool

313

33 Glioblastoma Multiforme: Use of Adenoviral Vectors . . . . . . . . . Thomas Wirth, Haritha Samaranayake, and Seppo Ylä-Herttuala

335

34 Fischer/F98 Glioma Model: Methodology . . . . . . . . . . . . . . . David Fortin

349

35 Cellular and Molecular Characterization of Anti-VEGF and IL-6 Therapy in Experimental Glioma . . . . . . . . . . . . . . Sophie Javerzat and Martin Hagedorn 36 Adult Brainstem Gliomas: Diagnosis and Treatment . . . . . . . . . Florence Laigle-Donadey and Jean-Yves Delattre

361 371

37 The Use of Low Molecular Weight Heparin in the Treatment and Prevention of Thromboembolic Disease in Glioma Patients . . . Bo H. Chao and H. Ian Robins

379

Part III

Prognosis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

385

38 Brainstem Gliomas: An Overview . . . . . . . . . . . . . . . . . . . . Marco Antonio Lima

387

39 Tumor-Associated Epilepsy in Patients with Glioma . . . . . . . . . Anja Smits and Anette Storstein

397

40 Brain Tumors Arising in the Setting of Chronic Epilepsy . . . . . . . Richard A. Prayson

407

xii

41

42

Contents

Low-Grade Gliomas: Role of Relative Cerebral Blood Volume in Malignant Transformation . . . . . . . . . . . . . . . . . . Neil Upadhyay and Adam D. Waldman

417

Angiocentric Glioma-Induced Seizures: Lesionectomy . . . . . . . . Dave F. Clarke and Timothy M. George

425

Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

435

Contents of Volume 1

1 Introduction 2 Molecular Classification of Gliomas 3 Glioblastoma: Endosialin Marker for Pericytes 4 Glioma Grading Using Cerebral Blood Volume Heterogeneity 5 The Role of Ectonucleotidases in Glioma Cell Proliferation 6 Gliomas: Role of Monoamine Oxidase B in Diagnosis 7 Glioma: Role of Integrin in Pathogenesis and Therapy 8 Proton Magnetic Resonance Spectroscopy in Intracranial Gliomas 9 Infiltration Zone in Glioma: Proton Magnetic Resonance Spectroscopic Imaging 10 Malignant Gliomas: Role of E2F1 Transcription Factor 11 The Role of Glucose Transporter-1 (GLUT-1) in Malignant Gliomas 12 Malignant Gliomas: Role of Platelet-Derived Growth Factor Receptor A (PDGFRA) 13 Molecular Methods for Detection of Tumor Markers in Glioblastomas 14 Role of MGMT in Glioblastomas 15 Glioblastomas: Role of CXCL12 Chemokine 16 Cell Death Signaling in Glioblastoma Multiforme: Role of the Bcl2L12 Oncoprotein 17 Glioblastoma Multiforme: Role of Polycomb Group Proteins 18 Glioblastoma Multiforme: Role of Cell Cycle-Related Kinase Protein (Method) 19 Markers of Stem Cells in Gliomas 20 Efficient Derivation and Propagation of Glioblastoma Stem-Like Cells Under Serum-Free Conditions Using the Cambridge Protocol

xiii

xiv

Contents of Volume 1

21

Glioma Cell Lines: Role of Cancer Stem Cells

22

Glioblastoma Cancer Stem Cells: Response to Epidermal Growth Factor Receptor Kinase Inhibitors

23

Low- and High-Grade Gliomas: Extensive Surgical Resection

24

Brainstem Gangliogliomas: Total Resection and Close Follow-Up

25

Glioblastoma: Temozolomide-Based Chemotherapy

26

Drug-Resistant Glioma: Treatment with Imatinib Mesylate and Chlorimipramine

27

Glioblastoma Multiforme: Molecular Basis of Resistance to Erlotinib

28

Enhanced Glioma Chemosensitivity

29

Malignant Glioma Patients: Anti-Vascular Endothelial Growth Factor Monoclonal Antibody, Bevacizumab

30

Aggravating Endoplasmic Reticulum Stress by Combined Application of Bortezomib and Celecoxib as a Novel Therapeutic Strategy for Glioblastoma

31

Targeted Therapy for Malignant Gliomas

32

Glioblastomas: HER1/EGFR-Targeted Therapeutics

33

Epidermal Growth Factor Receptor Inhibition as a Therapeutic Strategy for Glioblastoma Multiforme

34

Role of Acyl-CoA Synthetases in Glioma Cell Survival and Its Therapeutic Implication

35

Malignant Glioma Patients: Combined Treatment with Radiation and Fotemustine

36

Malignant Glioma Immunotherapy: A Peptide Vaccine from Bench to Bedside

37

Malignant Glioma: Chemovirotherapy

38

Intracranial Glioma: Delivery of an Oncolytic Adenovirus

39

Use of Magnetic Resonance Spectroscopy Imaging (MRSI) in the Treatment Planning of Gliomas

40

Malignant Glioma Cells: Role of Trail-Induced Apoptosis

41

Long-Term Survivors of Glioblastoma

42

Glioblastoma Patients: p15 Methylation as a Prognostic Factor

Contributors

Hilko Ardon Department of Neurosurgery, Laboratory of Experimental Immunology, Catholic University Leuven, Leuven, Belgium, [email protected] Paola Cassoni Department of Biomedical Sciences and Human Oncology, University of Turin, Turin, Italy, [email protected] Bo H. Chao Department of Medicine, University of Wisconsin School of Medicine and Public Health, Madison, WI, USA, [email protected] Yin-Sheng Chen Department of Neurosurgery, Cancer Center, Sun Yat-sen University, Guangzhou, China, [email protected] Zhong-Ping Chen Department of Neurosurgery, Cancer Center, Sun Yat-sen University, Guangzhou, China, [email protected] Iwona Ciechomska Department of Cell Biology, Nencki Institute of Experimental Biology, Warsaw, Poland, [email protected] Dave F. Clarke Division of Neurology, Department of Pediatrics, Dell Children’s Medical Center, Austin, TX, USA, [email protected] Linda Coate Department of Medical Oncology, Princess Margaret Hospital, Toronto, ON, Canada, [email protected] Steven De Vleeschouwer Department of Neurosurgery, Laboratory of Experimental Immunology, Catholic University Leuven, Leuven, Belgium, [email protected] Jean-Yves Delattre Service de Neurologie Mazarin, Hôpital de la Pitié-Salpêtrière – APHP, 75651 Paris Cedex 13, France, [email protected] Hugues Duffau Department of Neurosurgery and INSERM U1051, Institute for Neurosciences of Montpellier, Montpellier University Medical Center, Hôpital Gui de Chauliac, CHU Montpellier 34295, Montpellier Cedex 5, France, [email protected] Soufiane El Hallani Cancer Imaging Unit, Integrative Oncology Department, British Columbia Cancer Research Centre, Vancouver, BC, Canada V5Z 1L3, [email protected]

xv

xvi

Aleksandra Ellert-Miklaszewska Department of Cell Biology, Nencki Institute of Experimental Biology, Warsaw, Poland, [email protected] Alessandra Fabi Division of Medical Oncology, Regina Elena National Cancer Center Institute, Rome, Italy, [email protected] David Fortin Division of Neurosurgery, Department of Surgery, Sherbrooke University, Sherbrooke, QC, Canada, [email protected] Pim J. French Department of Neurology, Josephine Nefkens Institute, Erasmus Medical Centre, Rotterdam, The Netherlands, [email protected] Konrad Gabrusiewicz Laboratory of Transcription Regulation, Department of Cell Biology, Nencki Institute of Experimental Biology, Warsaw, Poland, [email protected] Timothy M. George Division of Neurosurgery, Department of Pediatrics, Dell Children’s Medical Center, Austin, TX, USA, [email protected] Lonneke A.M. Gravendeel Department of Neurology, Josephine Nefkens Institute, Erasmus Medical Centre, Rotterdam, The Netherlands, [email protected] Olli Gröhn Department of Neurobiology, University of Eastern Finland, Kuopio, Finland, [email protected] Martin Hagedorn INSERM U1029, Université Bordeaux 1, Talence, Cedex, France, [email protected] Abhirami Hallock Department of Radiation Oncology, London Regional Cancer Program, University of Western Ontario, E. London, ON, Canada, [email protected] Benedikte Hasselbalch Department of Radiation Biology, The Finsen Center, Sec 6321, Copenhagen University Hospital, Copenhagen, Denmark, [email protected] M.A. Hayat Department of Biological Sciences, Kean University, Union, NJ, USA, [email protected] David Hockaday Department of Neurosurgery, Cedars-Sinai Medical Center, Los Angeles, CA, USA Moriya Iwaizumi Department of Investigative Pathology I, Hamamatsu University School of Medicine, Higashi-ward, Hamamatsu, Japan, [email protected] Rajan Jain Division of Neuroradiology, Department of Radiology and Department of Neurosurgery, Henry Ford Health System, Detroit, MI, USA, [email protected] Sophie Javerzat INSERM U1029, Université Bordeaux 1, Talence, Cedex, France, [email protected] Shinji Kageyama Department of Investigative Pathology I, Hamamatsu University School of Medicine, Higashi-ward, Hamamatsu, Japan, [email protected]

Contributors

Contributors

xvii

Tomoaki Kahyo Department of Investigative Pathology I, Hamamatsu University School of Medicine, Higashi-ward, Hamamatsu, Japan, [email protected] Bozena Kaminska Laboratory of Transcription Regulation, Department of Cell Biology, Nencki Institute of Experimental Biology, Warsaw, Poland, [email protected] Lucie Karayan-Tapon Department of Molecular Biology, Poitiers University Hospital, University of Poitiers Medical School, Poitiers, France, [email protected] Nobuyuki Kawai Department of Neurological Surgery, Faculty of Medicine, Kagawa University, Kita-gun, Kagawa, Japan, [email protected] Sungheon Kim Department of Radiology, Center for Biomedical Imaging, New York University School of Medicine, New York, NY, USA, [email protected] Robert Kiss Laboratoire de Toxicologie, Faculté de Pharmacie, Université Libre de Bruxelles (ULB), Brussels, Belgium, [email protected] Toshihiro Kumabe Department of Neurosurgery, Tohoku University Graduate School of Medicine, Aoba-ku, Sendai-shi, Miyagi, Japan, [email protected] Hiroaki Kumada Proton Medical Research Center, University of Tsukuba, Tsukuba City, Ibaraki, Japan, [email protected] Kiyotaka Kurachi Department of Investigative Pathology I, Hamamatsu University School of Medicine, Higashi-ward, Hamamatsu, Japan, [email protected] Florence Laigle-Donadey Service de Neurologie Mazarin, Hôpital de la Pitié-Salpêtrière – APHP, Paris Cedex 13, France, [email protected] Christian-Jacques Larsen Poitiers University Hospital, University of Poitiers Medical School, Poitiers, France, [email protected] Ulrik Lassen Department of Radiation Biology, The Finsen Center, Sec 6321, Copenhagen University Hospital, Copenhagen, Denmark, [email protected] Florence Lefranc Laboratoire de Toxicologie, Faculté de Pharmacie, Université Libre de Bruxelles (ULB), Brussels, Belgium, [email protected] Kimmo Lehtimäki Department of Biotechnology and Molecular Medicine, University of Eastern Finland, Kuopio, Finland, [email protected] Timo Liimatainen Department of Biotechnology and Molecular Medicine, University of Eastern Finland, Kuopio, Finland, [email protected] Marco Antonio Lima Department of Neurosurgery, Instituto Nacional de Câncer-INCa, Centro, Rio de Janeiro, Brazil, [email protected] Giuseppe Lombardi Oncologia Medica 1, I.O.V. – IRCCS, Ospedale Busonera, Padova, Italy, [email protected]

xviii

Masato Maekawa Department of Investigative Pathology I, Hamamatsu University School of Medicine, Higashi-ward, Hamamatsu, Japan, [email protected] Adam N. Mamelak Department of Neurosurgery, Cedars-Sinai Medical Center, Los Angeles, CA, USA, [email protected] Emmanuel Mandonnet Laboratoire Kastler Brossel de l’ens, Paris Cedex 5, France, [email protected] Warren Mason Department of Medical Oncology, Princess Margaret Hospital, Toronto, ON, Canada, [email protected] Elias R. Melhem Division of Neuroradiology, Department of Radiology, Hospital of the University of Pennsylvania, Philadelphia, PA, USA, [email protected] Giulio Metro Division of Medical Oncology, Regina Elena National Cancer Center Institute, Rome, Italy, [email protected] Hiroki Mori Department of Investigative Pathology I, Hamamatsu University School of Medicine, Higashi-ward, Hamamatsu, Japan, [email protected] Kazuya Motomura Department of Neurosurgery, Nagoya University School of Medicine, 65 Tsurumai-cho, Showa-ku, Nagoya 466-8550, Japan, [email protected] Debraj Mukherjee Maxine Dunitz Neurosurgical Institute, Cedars-Sinai Medical Center, Los Angeles, California 90048, [email protected] Kei Nakai Department of Neurosurgery, Institute of Clinical Medicine, University of Tsukuba, Tsukuba City, Ibaraki, Japan, [email protected] Toshio Nakamura Department of Investigative Pathology I, Hamamatsu University School of Medicine, Higashi-ward, Hamamatsu, Japan, [email protected] Atsushi Natsume Department of Neurosurgery, Nagoya University School of Medicine, 65 Tsurumai-cho, Showa-ku, Nagoya 466-8550, Japan, [email protected] Yoshihiro Nishiyama Department of Radiology, Faculty of Medicine, Kagawa University, Kita-gun, Kagawa, Japan, [email protected] Masasuke Ohno Department of Neurosurgery, Nagoya University School of Medicine, 65 Tsurumai-cho, Showa-ku, Nagoya 466-8550, Japan, [email protected] Jordi Oliver Universitat de les Illes Balears, Palma de Mallorca, Spain, [email protected] Johan Pallud Service de Neurochirurgie, Hôpital Sainte-Anne, Paris Cedex 14, France, [email protected] Harish Poptani Division of Neuroradiology, Department of Radiology, Hospital of the University of Pennsylvania, Philadelphia, PA, USA, [email protected]

Contributors

Contributors

xix

Hans Skovgaard Poulsen Department of Radiation Biology, The Finsen Center, Sec 6321, Copenhagen University Hospital, Copenhagen, Denmark, [email protected] Richard A. Prayson Section of Neuropathology, Department of Anatomic Pathology, Cleveland Clinic, Cleveland, OH, USA, [email protected] Alfredo Quinones-Hinojosa Department of Neurosurgery, The Johns Hopkins School of Medicine, Baltimore, MD, USA, [email protected] Ilham Ratbi Department of Medical Genetics, Human Genomic Center, Faculty of Medicine and Pharmacy, University Mohamed V Souissi, National Institute of Health, Rabat, Marocco, [email protected] Zachary J. Reitman The Pediatric Brain Tumor Foundation Institute, The Preston Robert Tisch Brain Tumor Center, Duke University Medical Center, Durham, NC, USA; The Department of Pathology, Duke University Medical Center, Durham, NC, USA, [email protected] H. Ian Robins Department of Medicine, Human Oncology, and Neurology, University of Wisconsin School of Medicine and Public Health, Madison, WI, USA, [email protected] Pilar Roca Department de Biologia Fonamental i Ciències de la Salut. Ed. Guillem Colom, Universitat de les Illes Balears, Carretera Valldemossa Km 7.5, Palma de Mallorca, 07122, Balearic Islands, Spain, [email protected] Ryuta Saito Department of Neurosurgery, Tohoku University Graduate School of Medicine, Aoba-ku, Sendai-shi, Miyagi, Japan, [email protected] Haritha Samaranayake Department of Biotechnology and Molecular Medicine, A. I. Virtanen Institute, University of Kuopio, Kuopio, Finland, [email protected] Francisca M. Santandreu Universitat de les Illes Balears, Palma de Mallorca, Spain, [email protected] Rebecca Senetta Department of Biomedical Sciences and Human Oncology, University of Turin, Turin, Italy, [email protected] Christian Senft Department of Neurosurgery, Johann Wolfgang Goethe-University, Frankfurt, Germany, [email protected] Srinivasan Senthamizhchelvan Division of Nuclear Medicine, The Russell H. Morgan Department of Radiology and Radiological Science, Johns Hopkins University School of Medicine, Baltimore, MD, USA, [email protected] Ichiyo Shibahara Department of Neurosurgery, Tohoku University Graduate School of Medicine, Aoba-ku, Sendai-shi, Miyagi, Japan, [email protected] Kazuya Shinmura Department of Investigative Pathology I, Hamamatsu University School of Medicine, Higashi-ward, Hamamatsu, Japan, [email protected] Malgorzata Sielska Laboratory of Transcription Regulation, Department of Cell Biology, Nencki Institute of Experimental Biology, Warsaw, Poland, [email protected]

xx

Anja Smits Department of Neuroscience, Neurology, University Hospital Uppsala, Uppsala, Sweden, [email protected] Yukihiko Sonoda Department of Neurosurgery, Tohoku University Graduate School of Medicine, Aoba-ku, Sendai-shi, Miyagi, Japan, [email protected] Anette Storstein Department of Neuroscience, Haukeland University Hospital, N5021 Bergen, Norway, [email protected] Haruhiko Sugimura Department of Investigative Pathology I, Hamamatsu University School of Medicine, Higashi-ward, Hamamatsu, Japan, [email protected] Takashi Tamiya Department of Neurological Surgery, Faculty of Medicine, Kagawa University, Kita-gun, Kagawa, Japan, [email protected] Teiji Tominaga Department of Neurosurgery, Tohoku University Graduate School of Medicine, Aoba-ku, Sendai-shi, Miyagi, Japan, [email protected] Masaru Tsuboi Department of Investigative Pathology I, Hamamatsu University School of Medicine, Higashi-ward, Hamamatsu, Japan, [email protected] Magdalena Tyburczy Laboratory of Transcription Regulation, Department of Cell Biology, Nencki Institute of Experimental Biology, Warsaw, Poland, [email protected] Neil Upadhyay Institute of Clinical Science, Imperial College London, Charing Cross Hospital, W68RF, London, UK, [email protected] Frank Van Calenbergh Department of Neurosurgery, Catholic University Leuven, Leuven, Belgium, [email protected] Lauren VanderSpek Department of Radiation Oncology, London Regional Cancer Program, University of Western Ontario, E. London, ON, Canada, [email protected] Stefaan W. Van Gool Department of Pediatrics and Laboratory of Experimental Immunology, Catholic University Leuven, Leuven, Belgium, [email protected] Michel Wager Poitiers University Hospital, University of Poitiers Medical School, Poitiers, France, [email protected] Toshihiko Wakabayashi Department of Neurosurgery, Nagoya University School of Medicine, 65 Tsurumai-cho, Showa-ku, Nagoya 466-8550, Japan, [email protected] Adam D. Waldman Institute of Clinical Science, Imperial College London, Charing Cross Hospital, W68RF, London, UK, [email protected] Sumei Wang Division of Neuroradiology, Department of Radiology, Hospital of the University of Pennsylvania, Philadelphia, PA, USA, [email protected] Thomas Wirth Department of Biotechnology and Molecular Medicine, A. I. Virtanen Institute, University of Kuopio, Kuopio, Finland, [email protected]

Contributors

Contributors

xxi

Hidetaka Yamada Department of Investigative Pathology I, Hamamatsu University School of Medicine, Higashi-ward, Hamamatsu, Japan, [email protected] Tetsuya Yamamoto Department of Neurosurgery and Radiation Oncology, Institute of Clinical Medicine, University of Tsukuba, Tsukuba City, Ibaraki, Japan, [email protected] Yasuhiro Yamamura Department of Investigative Pathology I, Hamamatsu University School of Medicine, Higashi-ward, Hamamatsu, Japan, [email protected] Hai Yan The Pediatric Brain Tumor Foundation Institute, The Preston Robert Tisch Brain Tumor Center, Duke University Medical Center, Durham, NC, USA; The Department of Pathology, Duke University Medical Center, Durham, NC, USA, [email protected] Seppo Ylä-Herttuala Department of Biotechnology and Molecular Medicine, A. I. Virtanen Institute, University of Kuopio, Kuopio, Finland, [email protected] Naoki Yokota Department of Investigative Pathology I, Hamamatsu University School of Medicine, Higashi-ward, Hamamatsu, Japan, [email protected] Kanako Yuki Department of Neurosurgery, Nagoya University School of Medicine, 65 Tsurumai-cho, Showa-ku, Nagoya 466-8550, Japan, [email protected] Habib Zaidi Division of Nuclear Medicine, Geneva University Hospital, Geneva 4, Switzerland, [email protected] Fable Zustovich Oncologia Medica 1, I.O.V. – IRCCS, Ospedale Busonera, Padova, Italy, [email protected]

Chapter 1

Introduction M.A. Hayat

In developed countries cancer is the second leading cause of death exceeded only by cardiovascular diseases. There are more than 100 types of cancers that can inflict any part of the body. In 2005, 7.6 million people died of cancer, which constitutes 13% of the 58 million deaths worldwide. In the global population exceeding 6 billion in the year 2002, there were approximately 10.9 million new cancer cases, 6.7 million cancer deaths, and 22.4 million surviving from cancer diagnosed in the previous 5 years. In 2020, it is expected that the world’s population will increase to 7.5 billion, with 15 million new cancer cases and 12 million cancer deaths. Approximately, 1.4 million new cases of cancer and 550,000 cancer deaths were reported in the United States in 2008 (Am. Cancer Soc.), These data amount to ∼1500 deaths caused by cancer every day in the United States. In 2006 an estimated 19,000 new cases of brain tumors and 13,000 deaths were reported in the United States. This figure accounts for ∼1.4% of all cancer cases and 2.3% of all cancer cases that cause death. More than 10,000 Americans die annually from glioblastoma. Survival for this disease has not changed much in three decades. Since 1970, the number of cancer survivors has increased four-fold, with cancer survivors representing ∼3.5% of the United States population and 5-years survival rates increasing into the 60% range. These raises invite issues related to long-term and late effects of cancer treatment and the realization that cancer survivors represent ∼16% of all new primary cancers.

M.A. Hayat () Department of Biological Sciences, Kean University, Union, NJ 07083, USA e-mail: [email protected]

Glioblastoma Gliomas can arise either spontaneously (primary glioma) or can progress from a lower-grade to a highergrade (glioblastoma) of tumor. Malignant glioma is the most common tumor of the CNS, and glioblastoma is the most malignant form. Glioblastoma is characterized by rapid, highly invasive growth, extensive neovascularisation, and high mortality. The key reason for the lack of successful therapy is the infiltration of single tumor cells into the surrounding brain parenchyma cells, preventing complete glioblastoma resection. This process is facilitated by two related processes: (1) angiogenesis, the sprouting of new blood vessels from preexisting vasculature in response to external chemical stimulation, and (2) vasculogenesis, the reorganization of randomly distributed cells into a blood vessel network. Tumor cells can also acquire blood supply through other ways to escape conventional antiangiogenesis. In other words, blood vessels are formed by tumor cells instead of endothelial cells. This novel concept in tumor vascularization is termed as vasculogenic mimicry, which is the ability of aggressive tumor cells to express endotheliumassociated genes and form extracellular matrix-rich vasculogenic-like networks in three-dimensional culture. Such networks recapitulate embryonic vasculogenesis, and have been observed in human aggressive tumors such as glioblastoma (El Hallani et al., 2010, and Chapter 11, in this volume). Glioblastoma can be divided into two subtypes based on amplication and mutation of different genes, and characterization of molecular pathways has opened new venues to targeted therapies based on the individual genetic signature of the tumor (Ohgaki and

M.A. Hayat (ed.), Tumors of the Central Nervous System, Volume 2, DOI 10.1007/978-94-007-0618-7_1, © Springer Science+Business Media B.V. 2011

1

2

Kleihues, 2007). Two main subtypes on the basis of genetic differences are: (1) Primary glioblastomas typically occur in patients alder than 50 years of age, and are characterized by epidermal growth factor receptor (EGFR) amplification and mutations, loss of heterozygosity of chromosome log, deletion of the phosphatase and tension homologue on chromosome 10 (PTEN), and p16 deletion. (2) Secondary glioblastomas occur in younger patients as low-grade or anaplastic astrocytomas that are transformed over a period of several years into glioblastoma multiforme. Secondary glioblastomas are much less common than primary glioblastoma, and are characterized by mutations in the TP53 tumor suppressor gene, overexpression of the platelet-derived growth factor receptor (PDGFR), abnormalities in the p16 and retinoblastoma (Rb) pathways, and loss of heterozygosity of chromosome loq. Although glioblastoma is the most common and aggressive type of glioma and has the poorest survival, a small percentage of patients survive longer than the established median. Identification of genetic variants that influence long term survival of such patients may provide insight into tumor biology and treatment. A recent study by Liu et al. (2010) indicates that polymorphisms in the LIG4, BTBD2, HMGA2, and RTEL1 genes are associated with the survival of glioblastoma patients. LIG4 and BTBD2 are predictors of short-term survival, while CCDC26, HMGA2, and RTEL1 are predictors of long-term survival. In general, older patients (>50 years) have the worst survival (1.2 years), while younger patients may show a median survival-term of 7.8 years. These genes are known to be involved in the double strand break repair pathway. It is known that glioblastoma responds poorly to conventional therapies. Glioblastoma cells thrive despite an irregular blood supply and a frequently hypoxic microenvironment. Compensatory mechanisms, including glucose uptake and glycolytic activity, are enhanced in these tumors. In other words, glioblastoma cells, posses sufficient glycolytic capacity (more than that in normal brain tissue), supporting their migration even without mitochondrial involvement (Beckner et al., 1990) A recent study has identified ATP citrate lyase (ACLY) as a positive regulator of glycolysis in glioblastomas (Beckner et al., 2010). ACLY can be targeted to suppress hypoxic cell migration and invasion, and restore glycolytic inhibition. This inhibition of hypotoxic tumor cells potentially

M.A. Hayat

complement antiangiogenesis therapies that compromise the blood supply to tumor cells.

Treatment The three most common treatments are resection, radiation, and chemotherapy, or a combination of these methods. According to Sandmair et al. (2000), when radiation is utilized following surgical resectioning of the tumor, the median survival time may increase from 14 to 40 weeks. Maximal resection of brain glioma is usually the first or second treatment choice. Radiation is the alternative treatment. The goal is to maximize the effectiveness of resection, while minimize the operative risk. To accomplish this goal is not easy. Active migration of glioblastoma cells through the narrow extracellular spaces in the brain makes them elusive targets for surgical management. Glioma cells are “self-propelled”, and are able to adjust their shape and volume rapidly as they invade the brain parenchyma. The infiltrative nature of malignant gliomas results in poor demarcation of malignant boundaries. Another reason is the frequent location of supratentorial gliomas near or within eloquent areas. Consequently, the advantage of maximal resection in all cases is controversial. Image-guided surgery utilizing fluorescence with 5-aminolevulinic acid, neuronavigation, and intraoperative MRI has enabled more complete resectioning of contrast-enhancing tumors (Stummer et al., 2006; Nimsky et al., 2006) Chemotherapy in conjunction with surgery and radiation can increase the estimated survival of patients by 10.1% at 1 year and 8.6% at 2 years, but these rates apply to lower-grade glioma tumors. The current recommended chemotherapeutic agent is alkylating drug temozolomide. The 2-year survival rate of patients with newly diagnosed glioblastoma treated with radiotherapy and temozolomide is 26.5%, compared with 10.4% for radiotherapy alone (Stupp et al., 2005). Telozolomide also exerts antitumor affects by impairing angiogenic process. In vitro and in vivo studies have shown antiangiogenic activity by this drug even when it is used alone (Mathieu et al., 2008). The efficacy can be further enhanced by combining this treatment with bevacizumab; the later also has an antiangiogenic effect, although with a different mechanism of action. Antiangiogenic compounds

1

Introduction

also increase the therapeutic benefits of radiotherapy (Nieder et al., 2006). A phase 2 pilot study of bevacizumab in combination with telozolomide and regional radiotherapy for the treatment of patients with newly diagnosed glioblastoma recently reported that toxicities were acceptable to continue enrollment and a preliminary analysis of efficacy showed encouraging mean progression-free survival (Lai et al., 2008). It is concluded that a range of side-effects, including post-therapeutic neurological deterioration, can commonly or uncommonly are experienced by patients undergoing chemotherapy. To overcome some of the limitations mentioned above, intraoperative eletrostimulation can be used (Duffau, 2007, also, see Chapter 22, in this volume). The purpose is to understand the interindividual anatomical-functional variability in the case of glioma patients. In order to tailor the resection for each patient, it is mandatory to study the cortical functional organization, the affective connectivity, and the potentiality for brain plasticity. Another limitation is the innate inter individual prognostic variability encountered among malignant glioma patients. This limitation can be overcome by carrying out analysis of prognostic factors, which can predict the outcome of a therapy among diagnosed malignant glioma patients. Such an information can affect the design and conduct of clinical trials in the case of patients with recurrent glioma. Phase II trials play a critical role in the assessment of novel therapeutic approaches. Factors associated with an increased risk of death are old age (50 years or older), lower karnofsky performance score (60% of tumors) with abnormalities of TP53 gene itself or one member of the pathway (MDM2, MDM4, p14ARF); (b) the pRB1 pathway (68% of tumors) that includes RB1, CD4, CDKN2A (p16INK4A); 3) the PI3 K-PTEN pathway that involves PIK3CA, PIK3R1, PTEN, and IRS1. As a rule one abnormality of one of the genes belonging to a given pathway excludes aberrations of other genes belonging to this pathway. Even if this huge number of data has largely contributed to a better understanding of the mechanisms underlying the biology of gliomas, none of them presents a real practical interest on a therapeutic point of view. In contrast, a few abnormalities not implicated in the above-mentioned pathways and more or less specific to gliomas are currently used in the clinical field. Loss of portions of chromosomes 1p36 and 19q13.3, and methylation status of the MGMT gene

– Amplification of the BRAF gene on chromosome band 7q34 in pilocytic astrocytpomas. A tandem duplication of BRAF activates the gene by generating a fusion gene between the kinase domain of BRAF and the adjacent locus KIAA1549 (Jones et al., 2009). This rearrangement is present in 60–80% of pylocytic astrocytomas and infrequent in diffuse low grade astrocytomas thus providing clues for differential diagnosis. Constitutive activation of BRAF activates the MEK/ERK pathway leading to increased proliferation of tumor cells. In view of these data, inhibition of the BRAF/MEK/ERK cascade might constitute a therapeutic target for the treatment of low grade pediatric astrocytomas. However, a serious caveat has occurred with a new report showing that the use of some BRAF inhibitors in the control of tumor growth actually resulted in stimulatory effects on proliferation rather than stopping it. According to this work, BRAF-specific inhibitors (for instance 885-A) facilitate the association to its partner CRAF that is linked to Ras and induce an over activation of the MEK/ERK pathway. In contrast, pan-RAF inhibitors (sorafenib) while they do not block the association of BRAF to CRAF, inhibit the two partners and abolish the stimulation of the MAP kinase pathway (Heidorn et al., 2010). – Mutations of IDH1 and IDH2 gene (chromosome band 2q33) as a differential diagnosis tool. The Isocitrate dehydrogenase (IDH) gene encodes a metabolic enzyme that converts isocitrate to α-ketoglutarate with the concurrent reduction of NAD(P)+ into NAD(P)H. The two IDH1 and IDH2 isoforms are cytosolic and mitochondrial respectively. Whereas IDH3, as a part of the tricarboxylic acide (TCA) cycle, generates NADH for energy production, IDH1 and 2 participate to shuttling of electrons between mitochondria and cytosol. In 2008, it was found that mutations in IDH1 all located on the R132 residue and less frequent mutations on R172 in IDH2 were present in 12% of gliomas analysed in the series. The IDH1 mutated tumors were most common to young

18

M. Wager et al.

patients and secondary glioblastomas and associated to longer survival (Parsons et al., 2008). A subsequent study (Balss et al., 2008) explored the presence of IDH1 mutations on a series of 685 tumors comprised of most if not all major glioma subtypes. The mutations were nearly absent in primary glioblastoma and present in 88% of secondary glioblastomas. The high frequency of IDH1 R132 mutations in oligodendrogiomas (69%) and mixed oligoastrocytic tumors (78%) were consistent with an arising of secondary glioblasoma from lower grade tumors. No mutation was found in pilocytic astrocytomas making possible a molecular discrimination between this tumor type and infiltrating low-grade gliomas. Analysis of another series (Watanabe et al., 2009) provided arguments suggesting that IDH mutations are early events in gliomagenesis. The biochemical consequences of the IDH1 mutation most likely rely on the substitution of R132/R172 in IDH1/IDH2 because the two arginine residues contract hydrophilic interactions that allow the binding of isocitrate to the enzymes. Because of the large range of substituting residues, it is probable that R replacement supports tumorigenesis by impairing isocitrate binding. This leads to a loss of function and qualifies IDH1 and 2 as tumor suppressor genes (Zhao et al., 2009). In fact new recent data have been reported indicating that R132 mutant IDH1 not only does not catalyze the conversion of isocitrate to α-ketoglutarate but instead reduces α-ketoglutarate to the D-2 enantiomer form of 2-hydroxyglutarate (Dang et al., 2009). Consistent with this gain of function by the mutant enzyme, human gliomas with the R132 mutation in IDH1 contain 100-fold more 2-hydroxyglutarate than in tumors with WT IDH1. One possible explanation for an oncogenic function of mutant IDH1 could result from its inhibitory capacity of the oxygen-sensing enzymes hypoxia-inducible prolylhydroxylases (PHDs). Indeed, PHDs are inhibited in cells carrying mutant IDH1. PHD inhibition leads to the activation of the HIF transcription factor but currently, this does not give a clue as to know the real function of the mutant IDH1 in an oncogenic process. Pathological increase in the D2-hydroxyglutarate levels are found is associated with encephalopathy and cardiomyopathy but not with tumor risk, whereas the L-2-hydroxyglutarate

enantiomer also associated with progressive neural defects is linked to increased risk of glioma (Aghili et al., 2009). – A new mechanism relevant to the occurrence of resistance to alkylating agents has emerged from data showing an increased frequency of inactivating mutations in the MSH6 mismatch repair gene in patients who had recurrence following temozolomide treatment. This suggested that MSH6 deficiency may contribute to the emergence of resistance. Thus, MSH6 mutations are likely to be selected during temozolomide therapy. Most recently, studies on significant cohorts of annotated tumors have resulted in the detection of new genetic abnormalities: the NF1 gene was found mutated in 14% of glioblastomas; mutations of PIK3 R1 and ERBB2 genes were also detected respectively in 10 and 8% of analysed glioblastomas. However, the practical interest has not yet been proven and further studies will be necessary before concluding on their utility for clinicians.

Brain Tumours and Stem Cells A profusion of data in the literature feeds the notion that, as for other tumor types, brain tumors and particularly glioblastoma multiforme contain a fraction of tumor cells that share a number of features characteristic of neural stem cells and immature precursors (progenitors) more or less engaged in a differentiated lineage. Briefly, these cells are defined by two main properties: ability to perpetuate themselves (self-renewal capacity for true stem cells) and capacity to give rise to populations of postmitotic cells of the cerebral tissues through differentiation. While the existence of SC has been strongly established in hematopoietic tissue since the nineties, the presence of NSC in brain has been delayed until the demonstration of neural stem cells in subventricular zones of the adult human brain was reported (Sanai et al., 2004). This immediately reinforces the hypothesis that recurrence of GBM is likely to occur through cancer-initiating precursors with stem-cell properties, involving principally self-renewal in vitro and in vivo. For different reasons, this model postulates that only this small subset of cells is able to proliferate extensively and has

2

Gliomagenesis: Advantages and Limitations of Biomarkers

19

a true tumorigenic potential whereas most if not all of the cells composing the bulk of the tumor have lost this potential (for a review see Vescovi et al., 2006). The model immediately raises one obvious question about the status of NSC as targets for transforming events. There are at least two reasons that this may be the case: first, fewer events are required to impair the phenotype of cells that have self-activating mechanisms already operating. In addition, the persistence for long periods of time of self-renewing SC implies that the opportunity for mutations to accumulate in individual stem cells is higher than in non proliferating dying cells. This is of particular interest to explain the resistance to radiotherapy and chemotherapy acquired by TSC (Eramo et al., 2006; Liu et al., 2006; Bao et al., 2006). Two important consequences emerge from the tumor stem cell paradigm: first, distinct signaling pathways are activated in self-renewing stem cells and progenitors more advanced in differentiation. This opens the way to the definition of new targets. Second, the impact of therapeutics on a tiny subset of tumorigenic stem cells could reduce the useful targets for efficient treatment of tumors. In addition, the new data provided by integrated genomic analysis (see next section) could be useful for defining with accuracy molecular signatures for each category of TSC or progenitors.

and pathways involved in gliomas physiopathology. Second they should provide the clinician with new tools for improving diagnosis of tumors that are not easily classified by conventional histopathological analysis. In this respect, the definition of gene signatures specific of tumors types and grades is likely to introduce new classification schemes that will go further than those resulting from histopathological and imaging analysis alone. Such large studies require the establishment of publicly available banks of annotated tumors. By definition, these banks should contain sufficient numbers of each tumor subtype as to allow robust statistical analysis. Furthermore, use of efficient algorithms for achieving successfully specific signatures predictive of a given cellular phenotype. These conditions have been fulfilled by the creation of large networks, the Cancer Genome Atlas Research Network (TCGA, URL:http://cancer genome.nih.gov) being the most representative. These approaches confirmed some of the molecular pathways that had been previously involved in gliomagenesis and, most interestingly, identified new subtypes of GBM. Among the series of recent publications on the subject, a few deserve particular attention as they open the way to real progress in terms of diagnosis, prognosis and treatment. The work published by the Cancer Genome Atlas Network in 2008 is the first integrated genomic analysis implicating several parameters: DNA copy number, gene expression (trancriptome) and DNA methylation profiling in 206 human glioblastomas (an almost simultaneously published study led to the same general conclusions). This study univocally showed that most of the GBM carry abnormalities in TP53, RB1 and tyrosine kinase receptor pathways. One major conclusion was that these abnormalities implicating significant signalization pathways were a “core requirement for GBM pathogenesis”. The most recent data published by the TCGA start from the idea that different tumor subtypes may originate from different causes or different cells of origin, or both (Verhaak et al., 2010). By combining gene expression classification data, discarding several possible bias and by demonstrating their reproducibility on an independent validation set, the authors defined a molecular classification into four primary GBM subtypes, namely Proneural, Neural, Classical and Mesenchymal subtypes. For each of them, the 210 genes (per subtype) signature that was established was integrated with genomic aberrations. Principal characteristics for each

Perspectives Resulting from Integrated Genomic Analysis to Define a Framework for a Molecular Classification of Tumours and for Investigation of Targeted Therapies The New Tools of Genomic Analysis Molecular biology of cancer is more and more dominated by the development of large scale-array-based profiling methods at the DNA, RNA and protein levels. New technologies that are now commonly used involve array-based comparative genomic hybridization (array-CGH), SNP (single nucleotide polymorphism) arrays, transcriptomic arrays for gene expression, new sequencing techniques. The interest of such approaches is double: first from a fundamental point of view, the findings resulting from these technologies should culminate in an overview of unregulated genes

20

M. Wager et al.

CLASSICAL SUBTYPE: Chromosome 7 amplification and chromosome 10 loss Four-fold increase EGFR amplification (97%) EGFR point mutation or vII EGFR mutation (12/22 tumors) Relative absence of TP53 mutations Frequent deletion of chromosome band 9p21.3 (p16INK4A and p14ARF) accompanying EGFR alterations (94%). Mutual exclusion with alterations of the RB pathway indicating the implication of this way by the only CDKN2A deletions. Overexpression of components of stem cell marker NES, NOTCH (NOTCH3, JAG1, LFNG), Sonic Hedgehog (SMO, GAS1, GLI2) signaling pathways.

MESENCHYMAL SUBTYPE:

Focal hemizygous deletions of a 17q11.2 chromosome region containing NF1 (53%°

lower

NF1 expression levels. No methylation in the NF1 locus Frequency of NF1 mutations (14/20 tumors classified as Mesenchymal) Expression of Mesenchymal markers: CHI3L1/YKL40, MET High expression of members of the tumor necrosis factor superfamily pathway and NF-kb pathway: TRADD, RELB, TNFRS1A

PRONEURAL:

Focal amplification at 4q12 GBM than other subtypes)

alterations of PDGFRA most frequently found in this subtype of PDGFRA overexpression almost excusively found in this subtype

Point mutations (R132) in IDH1 (11/12 tumor) Frequent TP53 mutations and loss of heterozygosity (LOH) High expression of oligodendrocyitic development genes: PDGFRA, NKX2-2, OLIG2 Mutations of PI3KCA/PI3KR1 (10/16 tumors) Expression of several neural development genes: DCX, DLL3, ASCL1, TCF4

NEURAL:

Expression of neuron markers: NEFL, GABRA1, SYT1, SLC12A5. Two normal brain tissues were classified as Neural

Fig. 2.2 Properties of the four glioblastoma subtypes as reported by Verhaak et al. (2010)

subtype are presented in Fig. 2.2. Characteristically, the four subtypes exhibited molecular expression features reminiscent of distinct cell type gene sets associated with neurons, oligodendrocytes, astrocytes and cultured astroglial cells. For instance, the Proneural subtype was associated to an oligodendrocytic signature but not the astrocytic signature.

A clinical correlation could also be established with each subtype. The most consistent one was age, with younger patients overrepresented in the Proneural subtype. This association survived the multicentric origin of the patients. Furthermore, a comparison of intensive treatments (concurrent chemotherapy and radiotherapy or more than three subsequent cycles

2

Gliomagenesis: Advantages and Limitations of Biomarkers

21

of chemotherapy) on survival was done. It showed a significant reduction in mortality in Classical and Mesenchymal subtypes but not in Proneural subtype. Finally, the methylation status of MGMT promoter gene, previously associated to response to therapy, was not found associated to subtype.

is that the identification of the small number of MR generating and maintaining a specific tumor signature is synonymous of efficient tools for diagnostic and pharmacological interventions. Integrative genome analysis has also been performed to the rapidly progressing microRNA (miRs) field. The importance of the miRs in the oncogenic process and more precisely the metastatic progression has been particularly explored in various tumor pathologies. In gliomasa number of papers have confirmed the stimulatory role of certain miRs such as miR-21 in the progression of metastasis. Down-regulation of miR-21 inhibits the EGFR pathway thus confirming the stimulating activity of this miR on oncongenic genes as shown previously in other tumors (Zhou et al., 2010). More recently, Johnson’s group in Harvard reported the characterization of miR-26a as a cooperating component of an amplicon the presence of which in human glioblastomas is associated with a markedly decreased survival (Kim et al., 2010). This amplicon contains CDK4 and CENTG1, two oncogenes, regulating the RB1 and PI3K/AKT pathways respectively. miR-26a alone induces in vivo decrease of PTEN, RB1 and MAP3K2/MEKK2 pathways, thereby promoting AKT activation, proliferation and decreasing c-Jun terminal kinase-dependent apoptosis. As a whole, these data define a functional oncomiR/oncogene DNA cluster that promotes aggressiveness of glioblastomas by targeting the RB1, PI3K/AKT and JNK pathways. While these data should be completed by functional studies that will definitely characterize the pathways targeted by the different miRs, these results open the way to a molecular follow-up of the tumors by use of miR microarrays. The next approach would consist of designing inhibitors of miRs in order to block some of the steps in which their tumor stimulatory function is involved.

Perspectives for Transfer to Patients Care As a whole, this amazing set of data is going to provide the clinician with a frame of relevant informations that might be exploited for diagnosis and treatment. However, it is clear that at the present level, only clinical departments with an access to specialized laboratories will acquire the capacity to treat their patients in function of the recent data. Simplification of analysis is therefore needed in terms of tests and cost. A recent publication suggests that this could be achieved (Carro et al., 2010). The authors postulated that specific transcriptional signatures of tumors are placed under the regulatory effects of master gene regulators (MR). In other words, the model of gene expression profile-based cancer analysis which is currently studied in most cancer pathologies could be advantageously replaced by a cellular network analysis of the MR controlling the signatures and the corresponding phenotypes (subtypes). Such an approach was applied to the search of a transcriptional module of MRs regulating the molecular signature of the mesenchymal subtype. This analysis identified two transcriptional genes, C/EBPβ and STAT3, whose simultaneous over-expression oriented cell neural stem cells toward mesenchymal transformation. Most significantly, co-expression of the two factors was necessary and sufficient to reprogram the neural lineage along the malignant mesenchymal lineage. In contrast, their simultaneous elimination led to reduction of the mesenchymal signature and diminution of tumor aggressiveness. These results were corroborated by data showing a direct relationship between the expression levels of STAT3 and C/EBPβ in human GBM and poor clinical outcome. The biological significance of these results is presently unknown as STAT3 induces astrocytic differentiation and counteracts neuronal differentiation while C/EBPβ induces neurogenesis and inhibits gliomagenesis. The principal conclusion regarding the clinical utility of these results

Neurosphere Cultures: Biomarkers Themselves, or a Tool for Analysing Molecular Biomarkers? On the border of the biomarkers field, we have to mention the ability for some gliomas to generate neurospheres. It has been shown not only that only a subset of gliomas can generate neurospheres, but also that this ability is predictive of overall survival of patients

22

(Laks et al., 2009). On a clinical point of view, several factors preclude clinical use of neurosphere. A number of papers have shown that glioblastoma contain cells that can generate neurospheres when placed under stem cell culture conditions. Because these cells derived from tumors exhibit representative markers of stemness (usually CD133) and because of their tumorigenic potential in vivo, it is generally admitted that neurospheres might reflect the tumor cell origin. In this respect, molecular analysis of individual neurospheres could be of obvious interest to establish the tumor phenotype of each tumor and predict tumor response to treatment and outcome. However, several points must be clarified before achieving this goal. One of the most recent papers on the subject (Chen et al., 2010) reports a hierarchy of self-renewing tumor-initiating cell types in glioblastoma. It has shown that individual glioblastoma multiforme carry a CD133+ /CD133– heterogeneous population of self-renewing cells with tumor initiation capacity. These points are important as they raise the crucial question of the representativity of cultures of neurospheres. Furthermore, while the growth of neurospheres has already been shown to be limited to a subset of glioblastoma, the new data establish that PTEN deficiency of the tumor cells is a necessary requirement for successful neurospheres propagation. In other words, on the basis of these two parameters – cell heterogeneity and PTEN insufficiency – the possibility exist that cell types initially present in the tumor are not conserved in the cultures. For these reasons and in spite of the obvious interest for a better understanding of the biology of glioblastoma, other studies will have to be performed before neurospheres acquire a biomarker status. In conclusion, recent advances in molecular biology have permitted a better understanding of the complex mechanisms underlying gliomagenesis, and numerous biomarkers appeared. Some of them – LOH 1p/19q, methylation status of the MGMT promoter gene – have proven such an association to overall survival as far as groups of patients are concerned that stratifying analysis of clinical trials on these parameters has become a standard. But uncertainties remain regarding what pertains to natural history of gliomas, and what pertains to response to treatments, as assessed by biomarkers. It is also the reason why their importance in the choice between standard treatment and innovative therapies during clinical trials on this basis should remain cautious. Due to these uncertainties, and other technical

M. Wager et al.

and standardization reasons, time has not come – yet?for biomarkers based personalized medicine. Many other biomarkers are currently under study, and some of them – IDH1 and IDH2 – seem to be on the verge of reaching the clinical field. Recent data of integrated genomic analysis show that we are in the edge of new approaches which will improve tools given to clinicians for the diagnosis, decision making, and follow-up of adult gliomas. The molecular characterization of glioblastoma main subcategories will probably allow innovations in their classification. It should benefit from data to come on micro-RNAs (MiRs), as several preliminary results showed that their status of expression could contribute to tumor signatures. One can conceive a personalized approach to each tumor generating neurospheres, beginning with culture. Neurospheres might be valuable for their capacity at maintaining the original tumor, assessment of response and study of resistance to treatments.

References Aghili M, Zahedi F, Rafiee E (2009) Hydroxyglutaric aciduria and malignant brain tumor: a case report and literature review. J Neurooncol 91:233–236 Balss J, Meyer J, Mueller W, Korshunov A, Hartmann C, Von Deimling A (2008) Analysis of the IDH1 codon 132 mutation in brain tumors. Acta Neuropathol 116:597–602 Bao S, Wu Q, McLendon RE, Hao Y, Shi Q, Hjelmeland AB, Dewhirst MW, Bigner DD, Rich JN (2006) Glioma stem cells promote radioresistance by preferential activation of the DNA damage response. Nature 444:756–760 Berger MS (1995) Role of surgery n diagnosis and management of low grade gliomas. In: Appuzzo MLJ (eds) Benign cerebral glioma, vol II, Chapter 15. Neurosurgical topics. AANS Publications, Park Ridge, IL, pp 293–307 Cairncross JG, Ueki K, Zlatescu MC, Lisle DK, Finkelstein DM, Hammond RR, Silver JS, Stark PC, Macdonald DR, Ino Y, Ramsay DA, Louis DN (1998) Specific predictors of chemotherapeutic response and survival in patients with anaplastic oligodendroglioma. J Natl Cancer Inst 90: 1473–1479 Carro MS, Lim WK, Alvarez MJ, Bollo RJ, Zhao X, Snyder EY, Sulman EP, Anne SL, Doetsch F, Colman H, Lasorella A, Aldape K, Califano A, Iavarone A (2010) The transcriptional network for mesenchymal transformation of brain tumours. Nature 463:318–325 Chen R, Nishimura MC, Bumbaca SM, Kharbanda S, Forrest WF, Kasman IM, Greve JM, Soriano RH, Gilmour LL, Rivers CS, Modrusan Z, Nacu S, Guerrero S, Edgar KA, Wallin JJ, Lamszus K, Westphal M, Heim S, James CD, VandenBerg SR, Costello JF, Moorefield S, Cowdrey CJ, Prados M, Phillips HS (2010) A hierarchy of self-renewing

2

Gliomagenesis: Advantages and Limitations of Biomarkers

23

tumor-initiating cell types in glioblastoma. Cancer Cell 17:362–375 Dang L, White DW, Gross S, Bennett BD, Bittinger MA, Driggers EM, Fantin VR, Jang HG, Jin S, Keenan MC, Marks KM, Prins RM, Ward PS, En KE, Liau LM, Rabinowitz JD, Cantley LC, Thompson CB, Vander Heiden MG, Su SM (2009) Cancer- associated IDH1 mutations produce 2-hydroxyglutarate. Nature 462:739–744 Daumas-Duport C, Beuvon F, Varlet P, Fallet-Bianco C (2000) Gliomas: WHO and Sainte-Anne hospital classifications. Ann Pathol 5:413–428 Eramo A, Ricci-Vitiani L, Zeuner A, Pallini R, Lotti F, Sette G, Pilozzi E, Larocca LM, Peschle C, De Maria R (2006) Chemotherapy resistance of glioblastoma stem cells. Cell Death Diff 17:1238–1241 Hegi ME, Diserens AC, Gorlia T, Hamou MF, DeTribolet N, Weller M, Kros JMJM, Hainfellner JA, Mason W, Mariani L, Bromberg JEC, Hau P, Mirimanoff RO, Cairncross JG, Janzer RC, Stupp R (2005) MGMT gene silencing and benefit from temozolomide in glioblastoma. N Engl J Med 352:997–1003 Heidorn SJ, Milagre C, Whittaker S, Nourry A, NiculescuDuvas I, Dhomen N, Hussain J, Reis-Filho JS, Springer CJ, Pritchard C, Marais R (2010) Kinase-dead BRAF and oncogenic RAS cooperate to drive tumor progression through CRAF. Cell 140:209–221 Hormigo A, Gu B, Karimi S, Riedel E, Panageas KS, Edgar MA, Tanwar MK, Rao JS, Fleisher M, DeAngelis LM, Holland EC (2006) YKL-40 and matrix metalloproteinase-9 as potential serum biomarkers for patients with high grade gliomas. Clin Cancer Res 12:5698–5704 Idbaih A, Omurao A, Ducray F, Hoang-Xuan K (2007) Molecular genetic markers as predictors of response to chemotherapy in gliomas. Curr Opin Oncol 19:606–611 Jenkins RB, Blair H, Ballman KV, Giannini C, Arusell RM, Law M, Flynn H, Passé S, Felten S, Brown PD, Shaw EG, Buckner JC (2006) A t(1;19)(p10;q10) mediates the combined deletions of 1 p and 19q and predicts a better prognosis of patients with oligodendroglioma. Cancer Res 66:9852–9861 Jones DT, Kocialkowski S, Liu L, Pearson DM, Ichimura K, Collins VP (2009) Oncogenic RAF1 rearrangement and a novel BRAF mutation as alternatives to KIAA1549:BRAF fusionin activating the MAPK pathway in pilocytic astrocytoma. Oncogene 28:2119–2123 Karayan-Tapon L, Quillien V, Guilhot J, Wager M, Fromont G, Saikali S, Etchevery A, Hamlat A, Loussouarn D, Campion L, Campone M, Vallette FM, Gratas-Rabbia-Ré C (2010) Prognostic value of O6-methylguanine-DNA methyltransferase status in glioblastoma patients, assessed by five different methods. J Neurooncol 97:311–322 Kim H, Huang W, Jiang X, Pennicooke B, Par PJ, Johnson MD (2010) Integrative genome analysis reveals an oncomir/oncogene cluster regulating glioblastoma survivorship. Proc Natl Acad Sci USA 107:2183–2188 Laks DR, Masterman-Smiyh M, Visnyel K, Angenieux B, Orozco N, Foran I, Young WH, Vinters HV, Liau LM, Lazareff JA, Mischel PS, Cloughesy TF, Horvath S, Kornblum HL (2009) Neurosphere formation is an independent predictor of clinical outcome in malignant glioma. Stem Cells 27:980–987

Liu G, Yuan X, Zeng Z, Tunici P, Ng H, Abdulkadir IR, Lu L, Irvin D, Black KL, Yu JS (2006) Analysis of gene expression and chemoresistance of CD133+ cancer stem cells in glioblastoma. Mol Cancer 5:67 Louis DN, Ohgaki H, Wiestler OD, Cavenee WK (eds) (2007) WHO Classification of tumours of the central nervous system, 3rd ed. IARC Press, Lyon Mittler MA, Walters BC, Stopa EG (1996) Observer reliability in histological grading of astrocytoma stereotactic biopsies. J Neurosurg 85:1091–1094 Pallud J, Mandonnet E, Duffau H, Kujas M, Guillevin R, Galanaud D, Taillandier L, Capelle L (2006) Prognostic value of initial magnetic resonance imaging growth rates for world health organization grade II gliomas. Ann Neurol 60:380–383 Parsons DW, Jones S, Zhang X, Lin JC, Leary RJ, Angenendt P, Mankoo P, Carter H, Siu IM, Gallia GL, Olivi A, McLendon R, Rasheed BA, Keir S, Nikolskaya T, Nikolsky Y, Busam DA, Tekleab H, Diaz LA Jr, Hartigan J, Smith DR, Strausberg RL, Marie SK, Shinjo SM, Yan H, Riggins GJ, Bigner DD, Karchin R, Papadopoulos N, Parmigiani G, Vogelstein B, Velculescu VE, Kinzler KW (2008) An integrated genomic analysis of human glioblastoma multiforme. Science 321:1807–1812 Reifenberger J, Reifenberger G, Liu L, James CD, Wechsler W, Collins VP (1994) Molecular genetic analysis of oligodendroglial tumors shows preferential allelic deletions on 19q and 1p. Am J Pathol 145:1175–1190 Ricard D, Kaloshi G, Amiel-Benouaich A, Lejeune J, Marie Y, Mandonnet E, Kujas M, Mokhtari K, Taillibert S, LaigleDonadey F, Carpentier AF, Omuro A, Capelle L, Duffau H, Cornu P, Guillevin R, Canson M, Hoang-Xuan K, Delattre JY (2007) Dynamic history of low-grade gliomas before and after temozolomide treatment. Ann Neurol 61:484–490 Sanai N, Berger MS (2008) Glioma extend of resection and its impact on patient outcome. Neurosurg 62(4):753–4766 Sanai N, Tramontin AD, Quinones-Hinojosa A, Barbaro NM, Gupta N, Kunwar S, Lawton MT, McDermott MW, Parsa AT, Manuel-Garcia Verdugo J, Berger MS, Alvarez-Buylla A (2004) Unique astrocyte ribbon in adult human brain contains neural stem cells but lacks chain migration. Nature 427:740–744 Stupp R, Mason WP, Van den Bent MJ, Weller M, Fisher B, Taphoorn MJ, Belanger K, Brandes AA, Marosi C, Bogdahn U, Curschmann J, Janzer RC, Ludwin SK, Gorlia T, Allgeier A, Lacombe D, Cairncross JG, Eisenhauer E, Mirimanoff ROand the European Organisation for Research and Treatment of Cancer Brain Tumor and Radiotherapy Groups; National Cancer Institute of Canada Clinial Trials Group (2005) Radiotherapy plus concomitant and adjuvant temozolomide for glioblastoma. N Engl J Med 352:987–996 Tamber MS, Bansal K, Liang M-L, Mainprize TG, Salhia B, Northcott P, Taylor M, Rutka JT (2006) Current concepts in the molecular genetics of pediatric brain tumors: implication for emerging therapies. Child Nerv Syst 22:1379–1394 Varlet P, Jouvet A, Miquel C, Saint-Pierre G, Beuvon F, DaumasDuport C (2005) Criteria of diagnosis and grading of oligodendrogliomas or oligoastrocytomas according to the WHO and Sainte-Anne classifications. Neurochirurgie 51:239–246 Varlet P, Soni D, Miquel C, Roux FX, Meder JF, Chneiweiss H, Daumas- Duport C (2004) New variants of malignant

24 glioneuronal tumors: a clinicopathological study of 40 cases. Neurosurg 55:1377–1391 Verhaak RGW, Hoadley KA, Purdom E, Wang V, Qi Y, Wilkerson MD, Miller CR, Ding L, Golub T, Mesirov JP, Alexe G, Lawrence M, O’Kelly M, Tamayo P, Weir BA, Gabriel S, Winckler W, Gupta S, Jakkula L, Feiler HS, Hogson JG, James CD, Sarkaria JN, Brennan C, Kahn A, Spellman PT, Wilson RK, Speed TP, Gray JW, Meyerson M, Getz G, Perou CM, Hayes DN, and Cancer Genome Atlas Research Network (2010) Integrated genomic analysis identifies clinically relevant subtypes of glioblastoma characterized by abnormalities in PDGFRA, IDH1, EGFR, and NF1. Cancer Cell 17:98–110 Vescovi AL, Galli R, Reynolds BA (2006) Brain tumour stem cells. Nat Rev Cancer 6:425–436 Wager M, Guilhot J, Blanc JL, Ferrand S, Milin S, Bataille B, Lapierre F, Denis S, Chantereau T, Larsen CJ, KarayanTapon L (2006) Prognostic value of increase in transcript levels of Tp73 DeltaEx2-3isoforms in low-grade glioma patients. Br J Cancer 95:1062–1069 Wager M, Menei P, Guilhot J, Levillain P, Michalak S, Bataille B, Blanc JL, Lapierre F, Rigoard P, Milin S, Duthe F,

M. Wager et al. Bonneau D, Larsen CJ, Karayan-Tapon L (2008) Prognostic molecular markers with no impact on decision-making: the paradox of glioma based on a prospective study. Br J Cancer 98:1830–1838 Watanabe T, Nobusawa S, Kleihues P, Ohgaki H (2009) IDH1 mutations are early events in the development of astrocytomas and oligodendrogliomas. Am J Pathol 174: 1149–1153 Weller M, Stupp R, Reifenberger G, Brandes AA, Van Den Bent MJ, Wick W, Hegi ME (2010) MGMT promoter methylation in malignant gliomas: ready for personalized medicine? Nat Rev Neurol 6:39–51 Zhao S, Lin Y, Xu W, Jiang W, Zha Z, Wang P, Yu W, Li Z, Gong L, Peng Y, Ding J, Lei Q, Guan KL, Xiong Y (2009) Glioma-derived mutations in IDH1 dominantly inhibit IDH1 activity and induce HIF-1α. Science 32: 261–265 Zhou X, Ren Y, Moore L, Mei M, You Y, Xu P, Wang G, Jia Z, Pu P, Zhang W, Kang C (2010) Downregulation of miR-21 inhibits EGFR pathway and suppresses the growth of human glioblastoma cells independent of PTEN status. Lab Invest 90:144–155

Chapter 3

Molecular Subtypes of Gliomas Lonneke A.M. Gravendeel and Pim J. French

Abstract Gliomas are the most common primary brain tumors with heterogeneous morphology and variable prognosis. However, differences between histological subclasses and grades are subtle, and classifying gliomas is subject to a large inter-observer variability. This variability can result in misdiagnosis of gliomas. As treatment decisions in patients rely mainly on histological classification and clinical parameters, there is an urgent need of developing a more accurate and objective classification model. The identification of specific molecular markers (LOH of 1p19q, IDH1 mutation, MGMT methylation status), as well as molecular clusters based on gene expression have been studied extensively. These specific molecular features within gliomas can help diagnosis, can give a more accurate prognosis, and may also be used to develop personalized targeted therapy in the future. Keywords Gliomas · Molecular markers · Molecular cluster · LOH 1p19q · IDH1 · MGMT

Introduction Gliomas are the most common type of primary brain tumors in adults with an incidence rate of 5.27 per 100,000 patients every year (Ohgaki et al., 2004; Louis et al., 2007). In 1926, Bailey and Cushing suggested a classification model based on distinct

P.J. French () Department of Neurology, Josephine Nefkens Institute, Erasmus Medical Centre, 3000 DR Rotterdam, The Netherlands e-mail: [email protected]

histological morphologies (Bailey and Cushing, 1926), which forms the basis of the currently used WHO classification (Louis et al., 2007). Two major subtypes are recognized: Astrocytic (A) and oligodendrocytic (OD) tumors, the latter including pure OD tumors and mixed oligoastrocytic (MOA) tumors. Astrocytic tumors are further separated into grades I (pilocytic astrocytomas [PA]), II (low grade), III (anaplastic), and IV (glioblastoma [GBM]). Oligodendrocytic tumors are further separated into grades II (low grade) and III (anaplastic). Patient survival, time to tumor progression, and response to therapy are all associated with subtype and grade of the tumor (Louis et al., 2007). This classification model, combined with the patients’ prognostic features (e.g. age and Karnofsky Performance Score [KPS]), guides treatment decisions. Differences between histological subtypes are very subtle, and classifying gliomas is subject to a large interobserver variability (Scott et al., 1995; Murphy et al., 2002; Kros et al., 2007; Hildebrand et al., 2008). Clearly, this variability can result in misdiagnosis of gliomas, for example by assigning a prognostically favorable lower-grade glioma into a poor prognostic glioma (e.g. “false-positive GBM”), and in assigning a prognostically less favorable higher-grade glioma into a good prognostic glioma (e.g. “false-negative GBM”). Since treatment protocols often depend on the diagnosed histological subtype, accuracy in diagnosis is very important for patients in order to get optimal treatment (Murphy et al., 2002). Therefore, more accurate methods to diagnose gliomas are urgently required. The molecular characteristics of gliomas have been studied extensively over the last years, in order to provide more objective and accurate methods of identifying distinct molecular tumor subgroups, and to identify specific molecular tumor markers that can help

M.A. Hayat (ed.), Tumors of the Central Nervous System, Volume 2, DOI 10.1007/978-94-007-0618-7_3, © Springer Science+Business Media B.V. 2011

25

26

diagnosis. In the future, these molecular features may also be used to develop personalized targeted therapy.

Single Molecular Markers in Gliomas LOH 1p19q Loss of heterozygosity (LOH) of 1p19q is a chromosomal aberration that is strongly associated with classical ODs (Griffin et al., 2006; Jenkins et al., 2006; Cairncross and Jenkins, 2008). Determination of the 1p19 status in gliomas is clinically relevant because of two important features. First, gliomas with LOH of 1p19q generally grow more slowly than most other gliomas and therefore have a better prognosis. Second, the presence of this mutation is predictive for response to treatment with alkylating agents such as PCV (procarbazine, lomustine, and vincristine) (Cairncross et al., 1998; Kouwenhoven et al., 2006). However, many glial tumors benefit from alkylating agents at some level. Nevertheless, 1p19q status is a strong prognostic factor, and the presence or absence of 1p and 19q is currently tested for in patients with a tumor containing oligodendroglial features.

IDH1 Recently, somatic mutations in the gene encoding isocitrate dehydrogenase 1 (IDH1) have been identified in gliomas (Parsons et al., 2008). IDH1 mutations occur mainly in lower grade gliomas and in secondary GBMs and are therefore thought to be early events in glioma genesis (Hartmann et al., 2009). Interestingly, glioma-specific mutations in IDH1 always affect the amino acid arginine in position 132 of the amino acid sequence, which belongs to an evolutionary highly conserved region located at the binding site for isocitrate (Parsons et al., 2008; Bleeker et al., 2009). The most frequent mutation occurring in this region is the R132H mutation (>90%), but other variants also have been found (R132S, R132C, R132G, and R132L) (Bleeker et al., 2009; Gravendeel et al., 2010). Interestingly, non-R132H mutations segregate in distinct histological and molecular subtypes of glioma (Hartmann et al., 2009; Gravendeel et al., 2010).

L.A.M. Gravendeel and P.J. French

Histologically, non-R132H mutations occur sporadically in classic oligodendrogliomas and at significantly higher frequency in other grade II and III gliomas. Genetically, non-p.R132H mutations occur in tumors with TP53 mutation, are virtually absent in tumors with LOH on 1p and 19q and accumulate in distinct (gene-expression profiling based) intrinsic molecular subtypes (Gravendeel et al., 2009; 2010). Importantly, IDH1 mutations are associated with improved prognosis (Parsons et al., 2008; Kloosterhof et al., 2010). Therefore IDH1 mutation status is likely to be used as a prognostic molecular marker in the near future. Additionally, two studies have examined whether IDH1 mutation status can predict response to treatment in gliomas. In a group of patients with dedifferentiated low-grade astrocytomas progressive after radiotherapy response to TMZ did not differ between IDH1 mutant and wild-type tumors (Dubbink et al., 2009). In patients with anaplastic oligodendroglial tumors treated with radiotherapy alone or radiotherapy with adjuvant PCV, IDH1 mutations reported had no predictive value for response (van den Bent et al., 2010).

MGMT The O6 -methylguanine-DNA methyltransferase (MGMT) gene encodes for the nuclear repair enzyme alkyltransferase, which removes alkylating adducts from the O6 position of thymine. By doing this, the enzyme is involved in maintaining the integrity of the DNA. More specifically, the product of the MGMT gene protects the cells from being damaged by alkylating and methylating agents (e.g. BCNU [N,N[prime]-bis(2-chloroethyl)-N-nitrosourea], procarbazine and TMZ). In gliomas, the CpG islands located in the promotor of the MGMT gene are frequently methylated, causing “epigenetic” silencing of this gene. Theoretically, this methylation would result in a greater susceptibility for alkylating and methylating agents. In daily practice, the meaning and implications of the MGMT status are more difficult to interpret. Several studies showed that the epigenetic silencing of the MGMT gene is of clinical importance, because its association with increased survival and better response to combined chemoirradiation in GBMs (Gerson, 2004;

3

Molecular Subtypes of Gliomas

Hegi et al., 2005). It was shown that the effect of the MGMT silencing was especially present when TMZ was used as chemotherapy (Hegi et al., 2005). Other studies showed that the time of administering the TMZ seemed to be crucial. One study showed that a positive effect of the methylated MGMT gene on survival was only seen if TMZ was administered at the time of radiation, while other studies showed positive effects when administering TMZ before or after irradiation (Chakravarti et al., 2006; Everhard et al., 2006; Criniere et al., 2007). Interestingly, a recent study showed that MGMT promotor methylation in GBMs was a predictor for better response to radiotherapy, and a prognostic marker even in patients not receiving adjuvant alkylating chemotherapy (Rivera et al., 2010). This study suggests that MGMT might be a general favorable prognostic factor in GBMs, instead of being a predictive marker for response to alkylating chemotherapy. Another issue that makes it difficult to interpret the role of MGMT, is that there are several different approaches in the assays for measuring the MGMT promoter methylation status. These different approaches cause variable results, which may be difficult to compare. MGMT activity can be measured using immunohistochemistry, but also using Q-pcr, or promotor methylation assays. Also, differences in outcome arise using assays on frozen tissue as well as on paraffin samples, as well as by possible interobservervariability. Nevertheless, MGMT status is thought to be an important biomarker in gliomas.

Genome Wide Molecular Markers in Gliomas Several techniques have been developed to perform genome wide analysis of the tumors’ epigenome, genome or transcriptome. Whilst markers for glioma have been identified on all levels (TCGARN, 2008), this review will discuss the molecular markers identified by the tumor’s transcriptome, that are currently used (or show promise to aid) in clinical decision making. Gene expression profiling involves the measurement of the activity (the expression) of thousands of genes at once. In general, gene expression profiling is performed using microarrays; chips that contain the complementary sequence to thousands of target mRNA

27

sequences. mRNA isolated from tumor samples is processed and labeled and subsequently hybridized to the microarray. The signal extracted from the microarray is a measure of gene expression levels, which is visualized using fluorescence. Expression profiling can be used to identify molecular subtypes of tumors roughly by two methods: Supervised and unsupervised. Supervised clustering uses external information to separate tumors into predefined subgroups (e.g. responders vs. non-responders; long vs. short-survivors), and then specifically screens for genes that are differentially expressed between these groups. In contrast, unsupervised clustering does not use external information and thus classifies tumors based on homologies in gene expression profiles. One of the first large studies that used supervised clustering (on survival) has identified three large subtypes of glioma with distinct prognosis. Subtypes were named according to the signature of the genes they predominantly expressed: Proneural, Mesenchymal and Proliferative subtypes (Phillips et al., 2006). These subtypes have also been identified in GBM using unsupervised methods (Verhaak et al., 2010). A different supervised study identified genes associated with response to treatment (French et al., 2005). Most often however, supervised clustering has been used to define gene expression signatures based on histological subtypes (Nutt et al., 2003). Thus far, only three groups have performed unsupervised analysis to define “intrinsic” molecular subgroups of gliomas (Gravendeel et al., 2009; Li et al., 2009; Verhaak et al., 2010). In all cases, the unsupervised clusters identified more subtypes of gliomas than histology. The molecular clusters correlate better with survival than histology (Nutt et al., 2003; Gravendeel et al., 2009; Li et al., 2009). Therefore, molecular clustering provides an objective and more accurate method to classify gliomas, and may even be used to predict patients’ prognosis. The molecular clusters all contained a wide variety of histological subtypes. The fact that different histological subtypes were assigned to the same molecular cluster means that these phenotypically different tumors have a similar genetic composition. Indeed, two independent studies have demonstrated that genetic changes segregate into distinct molecular subtypes indicating that causal genetic change drives a distinct pattern of gene expression (Gravendeel et al., 2009; 2010; Verhaak et al., 2010).

28

For example, gliomas of different histological subtypes with LOH of 1p19q are accumulating within one distinct molecular profile, regardless of their histological appearance, showing significant longer survival times than other molecular subgroups (Gravendeel et al., 2009). These findings imply that 1p19q status should be determined in all histological subtypes of gliomas, instead of testing this mutation in oligodendroglial-like tumors only. In the future, the specific genetic features of molecular glioma subgroups can be used to improve diagnosis, to give a more accurate prognosis, as well as to develop personalized therapies. It is likely that each molecular glioma subgroup will benefit from its own specific treatment based on the specific underlying molecular pathways and markers. Novel randomized controlled trials should take these molecular clusters into account when comparing different therapy regimens in gliomas.

References Bailey P, Cushing HW (1926) A Classification of the tumors of the glioma group on a histogenetic basis with a correlated study of prognosis. J. B. Lippincott, Philadelphia, PA Bleeker FE, Lamba S, Leenstra S, Troost D, Hulsebos T, Vandertop WP, Frattini M, Molinari F, Knowles M, Cerrato A, Rodolfo M, Scarpa A, Felicioni L, Buttitta F, Malatesta S, Marchetti A, Bardelli A (2009) IDH1 mutations at residue p.R132 (IDH1(R132)) occur frequently in high-grade gliomas but not in other solid tumors. Hum Mutat 301:7–11 Cairncross JG, Ueki K, Zlatescu MC, Lisle DK, Finkelstein DM, Hammond RR, Silver JS, Stark PC, Macdonald DR, Ino Y, Ramsay DA, Louis DN (1998) Specific genetic predictors of chemotherapeutic response and survival in patients with anaplastic oligodendrogliomas. J Natl Cancer Inst 90(19):1473–1479 Cairncross G, Jenkins R (2008) Gliomas with 1p/19q codeletion: a.k.a. oligodendroglioma. Cancer J 14(6):352–357 Chakravarti A, Erkkinen MG, Nestler U, Stupp R, Mehta M, Aldape K, Gilbert MR, Black PM, Loeffler JS (2006) Temozolomide-mediated radiation enhancement in glioblastoma: a report on underlying mechanisms. Clin Cancer Res 12(15):4738–4746 Criniere E, Kaloshi G, Laigle-Donadey F, Lejeune J, Auger N, Benouaich-Amiel A, Everhard S, Mokhtari K, Polivka M, Delattre JY, Hoang-Xuan K, Thillet J, Sanson M (2007) MGMT prognostic impact on glioblastoma is dependent on therapeutic modalities. J Neuro-Oncol 83(2): 173–179 Dubbink HJ, Taal W, van Marion R, Kros JM, van Heuvel I, Bromberg JE, Zonnenberg BA, Zonnenberg CB, Postma TJ, Gijtenbeek JM, Boogerd W, Groenendijk FH, Smitt PA, Dinjens WN, van den Bent MJ (2009) IDH1 mutations in

L.A.M. Gravendeel and P.J. French low-grade astrocytomas predict survival but not response to temozolomide. Neurology 73(21):1792–1795 Everhard S, Kaloshi G, Criniere E, Benouaich-Amiel A, Lejeune J, Marie Y, Sanson M, Kujas M, Mokhtari K, Hoang-Xuan K, Delattre JY, Thillet J (2006) MGMT methylation: a marker of response to temozolomide in low-grade gliomas. Ann Neurol 60(6):740–743 French PJ, Swagemakers SM, Nagel JH, Kouwenhoven MC, Brouwer E, van der Spek P, Luider TM, Kros JM, van den Bent MJ, Sillevis Smitt PA (2005) Gene expression profiles associated with treatment response in oligodendrogliomas. Cancer Res 65(24):11335–11344 Gerson SL (2004) MGMT: its role in cancer aetiology and cancer therapeutics. Nat Rev Cancer 4(4):296–307 Gravendeel LA, Kouwenhoven MC, Gevaert O, de Rooi JJ, Stubbs AP, Duijm JE, Daemen A, Bleeker FE, Bralten LB, Kloosterhof NK, De Moor B, Eilers PH, van der Spek PJ, Kros JM, Sillevis Smitt PA, van den Bent MJ, French PJ (2009) Intrinsic gene expression profiles of gliomas are a better predictor of survival than histology. Cancer Res 69(23):9065–9072 Gravendeel LA, Kloosterhof NK, Bralten LB, van Marion R, Dubbink HJ, Dinjens W, Bleeker FE, Hoogenraad CC, Michiels E, Kros JM, van den Bent M, Smitt PA, French PJ (2010) Segregation of non-p.R132H mutations in IDH1 in distinct molecular subtypes of glioma. Hum Mutat 31(3):E1186–E1199 Griffin CA, Burger P, Morsberger L, Yonescu R, Swierczynski S, Weingart JD, Murphy KM (2006) Identification of der(1;19)(q10;p10) in five oligodendrogliomas suggests mechanism of concurrent 1p and 19q loss. J Neuropathol Exp Neurol 65(10):988–994 Hartmann C, Meyer J, Balss J, Capper D, Mueller W, Christians A, Felsberg J, Wolter M, Mawrin C, Wick W, Weller M, Herold-Mende C, Unterberg A, Jeuken JW, Wesseling P, Reifenberger G, von Deimling A (2009) Type and frequency of IDH1 and IDH2 mutations are related to astrocytic and oligodendroglial differentiation and age: a study of 1,010 diffuse gliomas. Acta Neuropathol 118(4):469–474 Hegi ME, Diserens AC, Gorlia T, Hamou MF, de Tribolet N, Weller M, Kros JM, Hainfellner JA, Mason W, Mariani L, Bromberg JE, Hau P, Mirimanoff RO, Cairncross JG, Janzer RC, Stupp R (2005) MGMT gene silencing and benefit from temozolomide in glioblastoma. N Engl J Med 352(10): 997–1003 Hildebrand J, Gorlia T, Kros JM, Afra D, Frenay M, Omuro A, Stupp R, Lacombe D, Allgeier A, van den Bent MJ, Investigators, E. B. T. G (2008) Adjuvant dibromodulcitol and BCNU chemotherapy in anaplastic astrocytoma: results of a randomised European organisation for research and treatment of cancer phase III study (EORTC study 26882). Eur J Cancer 44(9):1210–1216 Jenkins RB, Blair H, Ballman KV, Giannini C, Arusell RM, Law M, Flynn H, Passe S, Felten S, Brown PD, Shaw EG, Buckner JC (2006) A t(1;19)(q10;p10) mediates the combined deletions of 1p and 19q and predicts a better prognosis of patients with oligodendroglioma. Cancer Res 66(20):9852–9861 Kloosterhof NK, Bralten LB, Dubbink HJ, French PJ, van den Bent MJ (2010) Isocitrate dehydrogenase-1 mutations: a fundamentally new understanding of diffuse glioma? Lancet Oncol. Epub ahead of print

3

Molecular Subtypes of Gliomas

Kouwenhoven MC, Kros JM, French PJ, Biemond-ter Stege EM, Graveland WJ, Taphoorn MJ, Brandes AA, van den Bent MJ (2006) 1p/19q loss within oligodendroglioma is predictive for response to first line temozolomide but not to salvage treatment. Eur J Cancer 42(15):2499–2503 Kros JM, Gorlia T, Kouwenhoven MC, Zheng PP, Collins VP, Figarella-Branger D, Giangaspero F, Giannini C, Mokhtari K, Mork SJ, Paetau A, Reifenberger G, van den Bent MJ (2007) Panel review of anaplastic oligodendroglioma from European Organization For Research and Treatment of Cancer Trial 26951: assessment of consensus in diagnosis, influence of 1p/19q loss, and correlations with outcome. J Neuropathol Exp Neurol 66(6):545–551 Li A, Walling J, Ahn S, Kotliarov Y, Su Q, Quezado M, Oberholtzer JC, Park J, Zenklusen JC, Fine HA 2009 Unsupervised analysis of transcriptomic profiles reveals six glioma subtypes. Cancer Res 69(5):2091–2099 Louis DN, Ohgaki H, Wiestler OD, Cavenee WK (eds) (2007) WHO Classification of tumours of the central nervous system, Lyon Murphy M, Loosemore A, Ferrer I, Wesseling P, Wilkins PR, Bell BA (2002) Neuropathological diagnostic accuracy. Br J Neurosurg 16(5):461–464 Nutt CL, Mani DR, Betensky RA, Tamayo P, Cairncross JG, Ladd C, Pohl U, Hartmann C, McLaughlin ME, Batchelor TT, Black PM, von Deimling A, Pomeroy SL, Golub TR, Louis DN (2003) Gene expression-based classification of malignant gliomas correlates better with survival than histological classification. Cancer Res 63(7):1602–1607 Ohgaki H, Dessen P, Jourde B, Horstmann S, Nishikawa T, Di Patre PL, Burkhard C, Schuler D, Probst-Hensch NM, Maiorka PC, Baeza N, Pisani P, Yonekawa Y, Yasargil MG, Lutolf UM, Kleihues P (2004) Genetic pathways to glioblastoma: a population-based study. Cancer Res 64(19): 6892–6899 Parsons DW, Jones S, Zhang X, Lin JC, Leary RJ, Angenendt P, Mankoo P, Carter H, Siu IM, Gallia GL, Olivi A, McLendon R, Rasheed BA, Keir S, Nikolskaya T, Nikolsky Y, Busam DA, Tekleab H, Diaz LA Jr., Hartigan J, Smith DR, Strausberg RL, Marie SK, Shinjo SM, Yan H, Riggins GJ, Bigner DD, Karchin R, Papadopoulos N, Parmigiani G,

29 Vogelstein B, Velculescu VE, Kinzler KW (2008) An integrated genomic analysis of human glioblastoma multiforme. Science 321(5897):1807–1812 Phillips HS, Kharbanda S, Chen R, Forrest WF, Soriano RH, Wu TD, Misra A, Nigro JM, Colman H, Soroceanu L, Williams PM, Modrusan Z, Feuerstein BG, Aldape K (2006) Molecular subclasses of high-grade glioma predict prognosis, delineate a pattern of disease progression, and resemble stages in neurogenesis. Cancer Cell 9(3):157–173 Rivera AL, Pelloski CE, Gilbert MR, Colman H, De La Cruz C, Sulman EP, Bekele BN, Aldape KD (2010) MGMT promoter methylation is predictive of response to radiotherapy and prognostic in the absence of adjuvant alkylating chemotherapy for glioblastoma. Neurooncology 12(2):116–121 Scott CB, Nelson JS, Farnan NC, Curran WJ Jr., Murray KJ, Fischbach AJ, Gaspar LE, Nelson DF (1995) Central pathology review in clinical trials for patients with malignant glioma. A Report of Radiation Therapy Oncology Group 83-02. Cancer 76(2):307–313 TCGARN (2008) Comprehensive genomic characterization defines human glioblastoma genes and core pathways. Nature 455(7216):1061–1068 van den Bent MJ, Dubbink HJ, Marie Y, Brandes AA, Taphoorn MJ, Wesseling P, Frenay M, Tijssen CC, Lacombe D, Idbaih A, van Marion R, Kros JM, Dinjens WN, Gorlia T, Sanson M (2010) IDH1 and IDH2 mutations are prognostic but not predictive for outcome in anaplastic oligodendroglial tumors: a report of the European Organization for Research and Treatment of Cancer Brain Tumor Group. Clin Cancer Res 16(5):1597–1604 Verhaak RG, Hoadley KA, Purdom E, Wang V, Qi Y, Wilkerson MD, Miller CR, Ding L, Golub T, Mesirov JP, Alexe G, Lawrence M, O’Kelly M, Tamayo P, Weir BA, Gabriel S, Winckler W, Gupta S, Jakkula L, Feiler HS, Hodgson JG, James CD, Sarkaria JN, Brennan C, Kahn A, Spellman PT, Wilson RK, Speed TP, Gray JW, Meyerson M, Getz G, Perou CM, Hayes DN Cancer Genome Atlas Research, N (2010) Integrated genomic analysis identifies clinically relevant subtypes of glioblastoma characterized by abnormalities in PDGFRA, IDH1, EGFR, and NF1. Cancer Cell 17(1):98–110

Chapter 4

Glioblastoma: Germline Mutation of TP53 Haruhiko Sugimura, Hidetaka Yamada, Shinji Kageyama, Yasuhiro Yamamura, Naoki Yokota, Hiroki Mori, Moriya Iwaizumi, Kazuya Shinmura, Kiyotaka Kurachi, Toshio Nakamura, Masaru Tsuboi, Masato Maekawa, and Tomoaki Kahyo

Abstract Various brain tumors are components of familial cancer syndromes, and Li-Fraumeni syndrome and Li-Fraumeni-like syndromes are the most famous entities. A rare, sporadic occurrence of brain tumor in peculiarly young subjects, however, sometimes provides a clue to understanding of carcinogenesis due to germline mutations. In this section, two cases of glial tumors in subjects having germline mutation of TP53 are presented, and genetic etiology of brain tumor is discussed. The first case is a 41 year-old father of a fatal adrenocortical carcinoma case of a 4-years-old daughter, 16 years prior to the occurrence of astrocytoma of himself. The search for family history and previous clinical records in the four different community hospitals for the last three decades and the followups of his family members disclosed cancer of urinary bladder, cancer of pancreas, hepatoblastoma, and thymoma in the relatives. Germline mutation E286A of TP53 was identified in the affected members. Another case is a 21-year-old male, without any family history of cancer, who suffered from brain tumor and colorectal cancer. An attending physician was sticky to find an etiology in this unusual occurrence of cancer. After several attempts in vain, an I195T germline mutation was identified and functional analysis was performed. These anecdotes highlight importance of genetic analysis in case of glial tumors in relatively young adult or in adolescence whether or not they have family history of cancer. The problems and strategies to find

H. Sugimura () Department of Investigative Pathology I, Hamamatsu University School of Medicine, Higashi-ward, Hamamatsu 431-3192, Japan e-mail: [email protected]

TP53 mutation carriers and to prevent or to delay the occurrence of the tumors in them are discussed. Keywords Glioblastoma · Germline mutation · TP53 · Mismatch deficiencies · Inhibitors · Tumors

Introduction Traditional epidemiological studies have identified few environmental risk factors for malignant brain tumors (Osborne et al., 2001), but genetic components of the etiology of brain tumors, although rare, have been relatively well-defined for glial tumors (Schwartzbaum et al., 2006). Brain tumors are often accompanied by a genetic cancer syndrome such as Li-Fraumeni syndrome (LFS1, OMIM accession 151623), CHEK2related syndrome (LFS2, OMIM accession 609265), Li-Fraumeni-like syndrome (LFLS), Maffuci syndrome (OMIM accession 166000), Olier syndrome (OMIM accession 166000), tuberous sclerosis (OMIM accession 191100), or von Hippel-Lindau syndrome (OMIM accession 193300). The last two of these syndromes are usually related to tumors having specific histopathology, subependymal giant astrocytoma and hemangioblastoma, respectively, but the other syndromes accompany a variety of glial tumors and choroid plexus tumors. TP53 is undoubtedly the most influential genetic factor related to the occurrence of human brain tumors. In this review, we report two cases in which a glial tumors was caused by a germline mutation of TP53. The first case one was diagnosed in a family in which several members had suffered from various malignancies for decades and consulted different community hospitals in the county where they lived, which had a population of 700,000.

M.A. Hayat (ed.), Tumors of the Central Nervous System, Volume 2, DOI 10.1007/978-94-007-0618-7_4, © Springer Science+Business Media B.V. 2011

31

32

A definitive diagnosis of Li-Fraumeni syndrome was finally made by identifying the germline mutation of TP53 (Sameshima et al., 1992), 20 years after the first cancer occurred in the family. The second case was a sporadic case of synchronous colon cancer and glioblastoma multiforme. TP53 sequencing is not routine in such cases, and subsequent identification of the mutation was a very unexpected positive finding. In this chapter we review the TP53 germline mutations identified in glial tumor cases that have been reported in the literature, and we discuss the importance of functional assessment of the variants. We also address the issue of genetic testing for TP53 mutations in patients with glial tumors in situations that would arouse suspicion in most experienced clinicians, such as their occurrence in particularly young persons or clustered in a family, especially in routine clinical practice, and we offer perspectives in regard to the management of TP53 mutant carriers, that may seem bold at the moment.

Case Reports Case 1 A 41-year-old male was diagnosed with a gemistocytic astrocytoma (Fig. 4.1a), and a retrospective and follow-up study of his family over a 40-year period revealed cases of hepatoblastoma, adrenocortical carcinoma, thymoma, pancreatic cancer, and stomach cancer (Fig. 4.1b). A germline TP53 E286A mutation was identified (Sameshima et al., 1992), and there was marked TP53 overexpression in the astrocytoma (Fig. 4.1c). The functional significance of the E286A mutation is not definitively recapitulated in vitro, but the DNA-damage-associated dysregulation of the cell cycle has been investigated in cells derived from carriers (Goi et al., 1997). The spectrum of cancers in this family is typical of Li-Fraumeni syndrome.

Case 2 A 21-year-old male with no family history of cancer simultaneously developed symptoms of colorectal cancer and a brain tumor at the same time. The brain tumor was diagnosed as a glioblastoma multiforme

H. Sugimura et al.

(Fig. 4.1d, e), and marked TP53 overexpression was detected in the glioblastoma cells. The histological diagnosis of the colon cancer was well-differentiated adenocarcinoma (Fig. 4.1f), and TP53 was detected in the tumor cells (Fig. 4.1g). The curiosity of the attending physician was aroused by the clinical phenotype of this patient, and the physician suspected one of the genetic syndromes. A tentative diagnosis of Turcot syndrome was made, but the colon cancer did not exhibit microsatellite instability. The patient and his family were very cooperative in regard to further attempts to determine the etiology of his disease. A genetic counselor explained several possible etiologies of the patient’s clinical phenotype and the importance of conducting a genetic analysis. DNA from the blood cells of the patient and his healthy parents was used to sequence the entire germline TP53, including all of the exons, and an I195T mutation, which has been known as a hot spot for somatic mutations, was found as de novo germline mutation in the patient’s gene but not in his parents’. This missense variant TP53 is a wellknown somatic mutation but has never been recorded in Li-Fraumeni syndrome. The locus at the codon 195 has been found to alter transcriptional activity in a yeast system (Petitjean et al., 2007), and deficient transactivation capacity has been demonstrated in a mammalian cell system in vitro by co-transfection with p53 target molecules (Fig. 4.1h) (Yamada et al., 2009).

TP53 and Glial Tumors Glial tumors of varying biological grades are often encountered in both Li-Fraumeni syndrome and Li-Fraumeni-like syndrome (Pearson et al., 1982; Santibanez-Koref et al., 1991). Li’s early collection included 14 brain tumors, 9 of which were gliomas. All of the patients were under age 45 years old, and the mode was 15–29 years old (Li et al., 1988). Germline mutations of TP53 have been one of the best-known genetic etiologies of human cancers, including brain tumors. Cases of human cancer consistent with and not consistent with Li-Fraumeni syndrome have been investigated for germline mutations of TP53. The latest database of TP53 germline mutations (IARC TP53 database, an R14 release, November, 2009) indicates that 138 of the 1053 recorded germline mutations of TP53 have occurred

4

Glioblastoma: Germline Mutation of TP53

1

2

4

33

a

e

b

f

c

g

3

5

6

Relative luciferase activity

*

d Fig. 4.1 (a) Astrocytoma, grade 2, gemistocytic, of Case 1. Hematoxylin-eosin stain; (b) A family pedigree of the Case 1. An arrow indicates a case of astrocytoma. 1, pancreas cancer at age 36. 2, bladder cancer at age 20. 3, astrocytoma at 41. 4, hepatoblastoma at 2. 5, hepatoblastoma at 2. 6, thymoma at 17. Members 3, 5 and 6 were revealed to be carriers of the TP53 mutation (Sameshima et al., 1992). Closed circles and rectangles are carriers identified. The other members were not tested for TP53 mutation; (c) TP53 immunostaining of astrocytoma of the Case 1. Nuclear staining is prominent; (d) Glioblasoma multiforme of the Case 2. Hematoxylin eosin stain; (e) Glioblastoma multiforme, anti-glial fibrillary acidic protein

*

*

*

*

*

1.0

WT Thr195 Empty

0.5

0.0 p21

BAX

MDM2

h

(GFAP) (f) Adenocarcinoma of the colon in the Case 2; (g) TP53 immunostaining of colon cancer in the Case 2; (h) Transcription repression by mutant TP53 in the Case 2. Evaluation of the transcriptional activation function of p53 by luciferase assay. p21-, BAX-, and MDM2-luciferase activities were measured in p53-null H1299 cells transiently transfected with a p53 expression vector, the firefly luciferase reporter vector pGL4.10, and the transfection control vector pGL4.74. Values for luciferase activity are means ± standard deviation of three independent experiments. For each p53-responsive gene, the luciferase activity of cells transfected with wild-type (wt) p53 was set at 1.0

34

in patients with a brain tumor. Even in the absence of characteristic tumors, such as breast cancer, soft tissue sarcoma, and adrenocortical cancers, which are the characteristic tumors of Li-Fraumeni syndrome, there are the cases of familial aggregation, particularly of brain tumors alone (Kyritsis et al., 1994) surveyed immortalized lymphocytes derived from the gliomas of 51 patients and found TP53 mutations in 6 of them. They proposed that multiplicity (in the brain and other organs) and a familial history implies the presence of a TP53 germline mutation, and they pointed out that 6 subjects in 9 pedigrees were in their 30’s when their glial tumors were diagnosed. These reports are instructive when we encounter brain tumor cases in adolescents and young adults in our practice. More than 20 different germline missense mutations have been reported in the germline of patients with glial tumors of various biological grades, ranging from oligodendroglioma to astrocytoma and to glioblastoma multiforme.

Is Any Specific Mutation Spectrum Associated with Brain Tumors? Examination of the updated IARC database, suggests two questions in regard to the relation between germline TP53 mutations and brain tumors. The first question concerns whether there are germline TP53 mutations that specify the occurrence of a tumor in a particular organ. Lubbe et al. reported a familial brain tumor syndrome associated with codon 236 (Lubbe et al., 1995), and Vital et al. (1998) reported two families with a cluster of astrocytomas and choroid plexus tumors caused by codon 248 mutations. The specific significance of this mutation in these families seems to have drawn a great deal of attention since then, but these mutations have never been particularly questioned as brain tumor-causing germline mutations (Kleihues et al., 1997) reviewed 91 families with germline mutations and brain tumors and found an earlier age of onset of their brain tumors than in families with brain tumors but no germline mutations, but they were not able to identify any particular spectrum of germline mutations that tended to be more common in patients with brain tumors than in patients with tumors at other sites. The spectrum of somatic TP53 mutations has been extensively investigated and widely accepted

H. Sugimura et al.

as carcinogen fingerprints of the tumors (Hussain et al., 2000), but there has been little assessment of the spectrum of germline TP53 mutations in terms of a possible relationship to the specific phenotype of the proband and the proband’s descendants. Varley summarized the germline mutations in Li-Fraumeni syndrome and other syndromes, and claimed to have detected a mutation specificity preferentially associated with adrenocortical carcinoma (Varley, 2003). That claim requires validation by thorough testing of family members, but the concept of specific germline mutations deserves further corroboration. In any event, no spectrum of germline mutations has ever been identified in brain tumors patients. The second question concerns whether the brain tumors that develop in persons with a germline TP53 mutation have any specific characteristics in common. Bogler (Bogler et al., 1995) reviewed the spectrum of TP53 mutations in brain tumors and found that loss of a remaining wild-type allele was the most important second hit in tumorigenesis and loss of the remaining wild-type allele was detected in our case 2. His review focused on the nature of somatic genetic changes in brain tumors. Glioblastoma multiforme is one of the targets of The Atlas of Cancer Genome, prepared in the United States (Ledford, 2010). Watanabe (Watanabe et al., 2009) recently reported identifying an R132C mutation of the IDH1 gene in an astrocytoma in a patient with the Li-Fraumeni syndrome. The recent “next” generation sequencing has generated tremendous information on genetic changes in glioblastoma (Verhaak et al., 2010). IDH gene mutations are one of the most remarkable discoveries of high-throughput sequencing projects in regard to glioblastoma in general (Yan et al., 2009), and since 12% of glioblastomas have been found to harbor a IDH1 mutation (Parsons et al., 2008), the IDH1 mutations discovered by Watanabe are not specific for glioblastoma multiforme associated with Li-Fraumeni syndrome. The TP53 mutations often coincide found in germline cells are often the same as those found in somatic cells. With few exceptions, the genotype-phenotype relationship in genetic cancer diseases has not been extensively studied. The landscape of somatic genetic changes, including mutations and copy number amplifications, in tumors that occur in persons with a particular genetic background will soon become available, and it will be necessary to wait until enough cases accumulate to draw conclusions as

4

Glioblastoma: Germline Mutation of TP53

to whether any correlation exists between the spectrum of germline mutations and spectrum of phenotypic (somatic) mutations in patients with the tumors.

When and How Should Requests Be Made to Perform Genetic Testing and How and When Should Testing Be Performed? It is not easy to obtain genomic information, especially for practicing physicians, when there is no or little family history of cancer or when the patient’s relatives are not very familiar with recent genetic knowledge, including the significance of DNA. Even after overcoming the hurdles of the difficulty of obtaining DNA and conducting a genomic analysis, the odds of obtaining a result that can completely explain the patient’s condition are seldom high. Several studies, including the study by Paunu et al. (Paunu et al., 2001), did not reveal any germline TP53 mutations in familial glial tumors. Portwine also reported negative findings in a search for germline TP53 mutations in children with sporadic brain tumors (Portwine et al., 2001). The prevalence of germline TP53 mutations that have biological significance in asymptomatic populations appears to be extremely low, perhaps only 0.3% (Palmero et al., 2008), and it is difficult to persuade the healthy relatives to undergo genetic testing. We consider the likelihood of detection of a germline TP53 mutation in the adolescent cancer patient in Case 2 to have been very small. Most brain tumors are not thought of as genetic or hereditary diseases, and brain tumors are still basically classified according to their morphology and somatic genetic information (Burger and Scheithauer, 2007). Despite anecdotal reports of brain tumors as genetic manifestations, the overall attempt to identify constitutional genetic changes in brain tumor patients has been tedious until recently. Li et al. (1995) investigated TP53 germline mutations in 80 unselected glioma patients and found a TP53 germline mutation in only 1 of the 65 cases, that did not have a peculiar family history, as opposed in 3 of the 15 cases with a familial or “unusual” personal history. As stated in the introduction, taking a complete family history is often difficult, and cases in family members who would have developed cancer but died at an early age of some other cause would be missed. When the parents have few siblings, it is very difficult to conclude

35

that the occurrence of a disease in a particular case is sporadic. Then, how about “unusual” cases? In many of the cases in which we persuaded young patients and family members to undergo DNA testing and obtained their consent the results were negative, and the reason the patient suffered this particular disease, which was unusual for his or her age remained a mystery to the parents or relatives. There are still no valid scientific criteria for performing or recommending genetic tests in clinical settings. The criteria are often subjective and depend on the feeling of the attending physician or his or her enthusiasm regarding how far to pursue the investigation of the cause of the disease in cases that arouse suspicion in attending clinicians. Unusual clusterings of malignant tumors in a family attract general practitioners attention, but the absence of a family history of cancer or difficulty in obtaining it often impede further investigation by genetic analysis. The recent worldwide trend toward small families further reduces the possibility of identifying a peculiar pedigree to analyze for genetic disease. Then, can luck alone be expected to result in the identification of significant mutations? In compensation the era of small family sizes and the rigorous procedures required to obtain patient consent, a wide variety of genetic test, technologies are now available and are easier and friendlier than ever before. A genome-wide, personal whole genome sequence strategy is now available, and it will become less costly in the near future. The introduction of cytogenetic arrays in recent years has facilitated the discovery of unexpected changes in copy numbers, small deletion/insertion polymorphisms, amplifications, and rather rare (0.1–5%) single nucleotide polymorphisms (SNPs) when we attempt to determine the etiology of a common disease in an unusual situation (Rieber et al., 2009; Schwarzbraun et al., 2009). The era of personal sequencing is coming, and we, as one of the part of medical systems these days, i.e., as a patient, a family member of the patient, counselor, physician, insurer, or scientist, may have to prepare for the time when germline tests will be routinely performed in every patient with an “unusual” brain tumor.

Syndromes Expanding? Mismatch deficiencies are one of the well known genetic causes of brain tumors. Paraf proposed the

36

existence of two types of brain tumor-polyposis (BTP) syndrome (Paraf et al., 1997). The BTP syndrome type 1 is characterized by non-polypotic (thus, the name is paradoxical) colorectal tumors and microsatellite instability, the same as found in hereditary non-polypotic colorectal cancer (HNPCC). BTP1 syndrome corresponds to Turcot syndrome (MIM accession 270630) and/or Muir-Torre syndrome (MIM accession 158320), which can also have glioblastoma multiforme as a component of the syndrome (Park et al., 2009). BTP syndrome type 2 is characterized by brain tumors associated with adenomatous polyposis coli. We are not sure whether our Case 2 is a case of BTP syndrome type 1, because of the absence of microsatellite instability in the colorectal cancer in Case 2. It would be inappropriate to call cases like Case 2 “brain tumor-polyposis syndrome”. In conclusion, we are compelled to expand the concept of brain tumor-colorectal tumor syndrome to include what we refer to here as “TP53 disorders”.

Problems in Screenings and Preventions The unusual occurrence of common cancers in subjects harboring a germline mutation of TP53 prompted us to consider the feasibility of detecting or mass screening for such carriers in the general population. An R337H mutation of TP53 was found in two of 750 healthy subjects in a study conducted in southern Brazil (Palmero et al., 2008). A thorough characterization of this germline mutation (variant) in terms of its penetrance and the cancer spectrum of the subjects’ families is under way, and justification of neonatal mass screening for this variant has been debated (Achatz et al., 2009). To justify this kind of attempts of neonatal mass screening, we must be ready and feasible for detection of the early occurrence of tumors in carriers, and we also have to have the tools to treat the tumors or to reduce the hazard caused by tumors in the carriers. Endoscopic surveillance of the gastrointestinal (GI) tract for cancer and surveillance for breast cancer by mammography would be an acceptable choice for many subjects instead of more invasive measures. Some GI tumors can be treated endoscopically, and some breast tumors can be cured by minimally invasive surgery. The situation in regard to brain tumors is different. CT, MRI, and positron emission

H. Sugimura et al.

tomography (PET) are powerful diagnostic imaging methods for detecting small tumors, but is it acceptable for the carriers of germline TP53 mutations to undergo craniotomy or stereotactic radioneurosurgery? Since carriers of germline TP53 mutations are thought to be more sensitive to radiation it is not certain that the gamma knife is the best choice of treatment. Many investigators are attempting to find agents that will reduce the risk associated with a germline heterogeneous TP53 mutation. One bold, promising strategy would be to use p53-activating drugs to treat patients with germline heterozygosity of p53. Deacetylase inhibitors are promising candidates for such drugs that would reduce the risk associated with germline heterogeneous TP53 mutation, because acetylation of the C-terminal region of TP53 protein up-regulates its transcriptional and transactivational activity. Treatment with a deacetylase inhibitor is expected to compensate for the depleted function of mutant p53 in the p53-heterozygous cells by increasing the acetylation level of wild-type p53. One of the deacetylase inhibitors, the sirtuin (histone deacetylase type [HDAC] III) -targeting inhibitor, is particularly noteworthy, because the sirtuin (HDAC III) is catalytically different from the other histone deacetylases. The members of the sirtuin deacetylase family require NAD to exert their deacetylase activity, and these characteristics may lead to novel efficacy against cancer cells and cancer-predisposing cells. There is a mouse model with haploinsufficiency of p53 that is prone to develop various tumors (Jacks et al., 1994), and tailless, a nuclear receptor gene has recently been shown to contribute to brain tumor initiation in the p53 haploinsufficient mouse (Liu et al., 2010). Such models will provide monitoring systems to test the preventive efficacy of drugs against tumorigenesis in persons with greater genetic susceptibility to cancers. Various sirtuin inhibitors, both natural and synthetic, have been evaluated in vitro (Kahyo et al., 2008), and application of these drugs to an animal model and then clinically must be assessed in the near future.

References Achatz MI, Hainaut P, Ashton-Prolla P (2009) Highly prevalent TP53 mutation predisposing to many cancers in the Brazilian population: a case for newborn screening? Lancet Oncol 10:920–925

4

Glioblastoma: Germline Mutation of TP53

Bogler O, Huang HJ, Kleihues P, Cavenee WK (1995) The p53 gene and its role in human brain tumors. Gila 15:308–327 Burger PC, Scheithauer BW (2007) Tumor of the central nervous system, vol 7. American Registry of Pathology and Armed Forces Institute of Pathology, Washington, DC Goi K, Takagi M, Iwata S, Delia D, Asada M, Donghi R, Tsunematsu Y, Nakazawa S, Yamamoto H, Yokota J, Tamura K, Saeki Y, Utsunomiya J, Takahashi T, Ueda R, Ishioka C, Eguchi M, Kamata N, Mizutani S (1997) DNA damageassociated dysregulation of the cell cycle and apoptosis control in cells with germ-line p53 mutation. Cancer Res 57:1895–1902 Hussain SP, Hollstein MH, Harris CC (2000) p53 tumor suppressor gene: at the crossroads of molecular carcinogenesis, molecular epidemiology, and human risk assessment. Ann NY Acad Sci 919:79–85 Jacks T, Remington L, Williams BO, Schmitt EM, Halachmi S, Bronson RT, Weinberg RA (1994) Tumor spectrum analysis in p53-mutant mice. Curr Biol 4:1–7 Jeannot E, Mellottee L, Bioulac-Sage P, Balabaud C, Scoazec JY, Van Nhieu JT, Bacq Y, Michalak S, Buob D, Genthep (Inserm network), Laurent-Puig P, Rusyn I, Zucman-Rossi J (2010) Spectrum of HNF1A somatic mutations in hepatocellular adenoma differs from that in MODY3 patients and suggests genotoxic damage. Diabetes. 59:1836–1844 Kahyo T, Ichikawa S, Hatanaka T, Yamada MK, Setou M (2008) A novel chalcone polyphenol inhibits the deacetylase activity of SIRT1 and cell growth in HEK293T cells. J Pharmacol Sci 108:364–371 Kleihues P, Schauble B, zur Hausen A, Esteve J, Ohgaki H (1997) Tumors associated with p53 germline mutations: a synopsis of 91 families. Am J Pathol 150:1–13 Kyritsis AP, Bondy ML, Xiao M, Berman EL, Cunningham JE, Lee PS, Levin VA, Saya H (1994) Germline p53 gene mutations in subsets of glioma patients. J Natl Cancer Inst 86:344–349 Ledford H (2010) Big science: the cancer genome challenge. Nature 464:972–974 Li FP, Fraumeni JF Jr, Mulvihill JJ, Blattner WA, Dreyfus MG, Tucker MA, Miller RA (1988) A cancer family syndrome in twenty-four kindreds. Cancer Res 48:5358–5362 Li YJ, Sanson M, Hoang-Xuan K, Delattre JY, Poisson M, Thomas G, Hamellin R (1995) Incidence of germ-line p53 mutations in patients with gliomas. Int J Cancer 64:383–387 Liu HK, Wang Y, Belz T, Bock D, Takacs A, Radlwimmer B, Barbus S, Reifenberger G, Lichter P, Schütz G (2010) The nuclear receptor tailless induces long-term neural stem cell expansion and brain tumor initiation. Genes Dev 24:683–695 Lubbe J, von Ammon K, Watanabe K, Hegi ME, Kleihues P (1995) Familial brain tumour syndrome associated with a p53 germline deletion of codon 236. Brain Pathol 5:15–23 Osborne RH, Houben MP, Tijssen CC, Coebergh JW, van Duijn CM (2001) The genetic epidemiology of glioma. Neurol 57:1751–1755 Palmero EI, Schuler-Faccini L, Caleffi M, Achatz MI, Olivier M, Martel-Planche G, Marcel V, Aguiar E, Giacomazzi J, Ewald IP, Giugliani R, Hainaut P, Ashton-Prolla P (2008) Detection of R337H, a germline TP53 mutation predisposing to multiple cancers, in asymptomatic women participating in a breast cancer screening program in Southern Brazil. Cancer Lett 261:21–25

37 Paraf F, Jothy S, Van Meir EG (1997) Brain tumor-polyposis syndrome: two genetic diseases? J Clin Oncol 15:2744–2758 Park DM, Yeaney GA, Hamilton RL, Mabold J, Urban N, Appleman L, Flickinger J, Lieberman F, Mintz A (2009) Identifying Muir-Torre syndrome in a patient with glioblastoma multiforme. Neuro Oncol 11:452–455 Parsons DW, Jones S, Zhang X, Lin JC, Leary RJ, Angenendt P, Mankoo P, Carter H, Kamiyama H, Jimeno A, Hong SM, Fu B, Lin MT, Calhoun ES, Kamiyama M, Walter K, Nikolskaya T, Nikolsky Y, Hartigan J, Smith DR, Hidalgo M, Leach SD, Klein AP, Jaffee EM, Goggins M, Maitra A, Iacobuzio-Donahue C, Eshleman JR, Kern SE, Hruban RH, Karchin R, Papadopoulos N, Parmigiani G, Vogelstein B, Velculescu VE, Kinzler KW (2008) An integrated genomic analysis of human glioblastoma multiforme. Science 321: 1807–1812 Paunu N, Syrjakoski K, Sankila R, Simola KO, Helen P, Niemela M, Matikainen M, Isola J, Haapasalo H (2001) Analysis of p53 tumor suppressor gene in families with multiple glioma patients. J Neurooncol 55:159–165 Pearson AD, Craft AW, Ratcliffe JM, Birch JM, Morris-Jones P, Roberts DF (1982) Two families with the Li-Fraumeni cancer family syndrome. J Med Genet 19:362–365 Petitjean A, Mathe E, Kato S, Ishioka C, Tavtigian SV, Hainaut P, Olivier M (2007) Impact of mutant p53 functional properties on TP53 mutation patterns and tumor phenotype: lessons from recent developments in the IARC TP53 database. Hum Mutat 28:622–629 Portwine C, Chilton-MacNeill S, Brown C, Sexsmith E, McLaughlin J, Malkin D (2001) Absence of germline and somatic p53 alterations in children with sporadic brain tumors. J Neurooncol 52:227–235 Rieber J, Remke M, Hartmann C, Korshunov A, Burkhardt B, Sturm D, Mechtersheimer G, Wittmann A, Greil J, Blattmann C, Witt O, Behnisch W, Halatsch ME, Orakcioglu B, von Deimling A, Lichter P, Kulozik A, Pfister S (2009) Novel oncogene amplifications in tumors from a family with Li-Fraumeni syndrome. Genes Chromosomes Cancer 48: 558–568 Sameshima Y, Tsunematsu Y, Watanabe S, Tsukamoto T, Kawaha K, Hirata Y, Mizoguchi H, Sugimura T, Terada M, Yokota J (1992) Detection of novel germ-line p53 mutations in diverse-cancer-prone families identified by selecting patients with childhood adrenocortical carcinoma. J Natl Cancer Inst 84:703–707 Santibanez-Koref MF, Birch JM, Hartley AL, Jones PH, Craft AW, Eden T, Crowther D, Kelsey AM, Harris M (1991) p53 germline mutations in Li-Fraumeni syndrome. Lancet 338:1490–1491 Schwartzbaum JA, Fisher JL, Aldape KD, Wrensch M (2006) Epidemiology and molecular pathology of glioma. Nat Clin Pract Neurol 2:494–503, quiz 491 p following 516 Schwarzbraun T, Obenauf AC, Langmann A, Gruber-Sedlmayr U, Wagner K, Speicher MR, Kroisel PM (2009) Predictive diagnosis of the cancer prone Li-Fraumeni syndrome by accident: new challenges through whole genome array testing. J Med Genet 46:341–344 Varley JM (2003) Germline TP53 mutations and Li-Fraumeni syndrome. Hum Mutat 21:313–320 Verhaak RG, Hoadley KA, Purdom E, Wang V, Qi Y, Wilkerson MD, Miller CR, Ding L, Golub T, Mesirov JP, Alexe G,

38 Lawrence M, O’Kelly M, Tamayo P, Weir BA, Gabriel S, Winckler W, Gupta S, Jakkula L, Feiler HS, Hodgson JG, James CD, Sarkaria JN, Brennan C, Kahn A, Spellman PT, Wilson RK, Speed TP, Gray JW, Meyerson M, Getz G, Perou CM, Hayes DN, and Cancer Genome Atlas Research Network (2010) Integrated genomic analysis identifies clinically relevant subtypes of glioblastoma characterized by abnormalities in PDGFRA, IDH1, EGFR, and NF1. Cancer Cell 17:98–110 Vital A, Bringuier PP, Huang H, San Galli F, Rivel J, Ansoborlo S, Cazauran JM, Taillandier L, Kleihues P, Ohgaki H (1998) Astrocytomas and choroid plexus tumors in two families with identical p53 germline mutations. J Neuropathol Exp Neurol 57:1061–1069

H. Sugimura et al. Watanabe T, Vital A, Nobusawa S, Kleihues P, Ohgaki H (2009) Selective acquisition of IDH1 R132C mutations in astrocytomas associated with Li-Fraumeni syndrome. Acta Neuropathol 117:653–656 Yamada H, Shinmura K, Yamamura Y, Kurachi K, Nakamura T, Tsuneyoshi T, Yokota N, Maekawa M, Sugimura H (2009) Identification and characterization of a novel germline p53 mutation in a patient with glioblastoma and colon cancer. Int J Cancer 125:973–976 Yan H, Parsons DW, Jin G, McLendon R, Rasheed BA, Yuan W, Kos I, Batinic-Haberle I, Jones S, Riggins GJ, Friedman H, Friedman A, Reardon D, Herndon J, Kinzler KW, Velculescu VE, Vogelstein B, Bigner DD (2009) IDH1 and IDH2 mutations in gliomas. N Engl J Med 360:765–773

Chapter 5

Familial Gliomas: Role of TP53 Gene Soufiane El Hallani and Ilham Ratbi

Abstract The presence of a personnel or familial history of cancer in a patient with primary brain tumor should prompt the search for genetic predisposition. Familial aggregation of brain tumors and especially gliomas has been reported in 5–7% of the cases. Rarely, familial gliomas can be attributed to hereditary multisystem syndromes. In majority of the cases, a hereditary syndrome cannot be identified and genetic alterations predisposing to familial gliomas are not clearly identified. Studies investigating candidate loci in glioma families have frequently examined TP53 gene which encodes p53, a checkpoint protein that plays a crucial role in DNA damage repair and apoptosis. Although occasional germline TP53 mutations were reported in glioma families, the frequency remained low (2.5–6.7%). However, germline TP53 mutations are more incriminated in glioma patients with multifocality or secondary malignancy, particularly if these factors are combined. Moreover, a functional single nucleotide polymorphism at codon 72 of TP53 gene was found to be associated with earlier onset of sporadic glioblastomas, opening new insights into the role of low-risk variants as genetic susceptibility to develop brain tumors. This chapter reviews the genetic predisposition in brain tumors and highlights the role of TP53 gene in familial gliomas, and in other forms of genetic predisposition or susceptibility to gliomas.

S. El Hallani () Cancer Imaging Unit, Integrative Oncology Department, British Columbia Cancer Research Centre, Vancouver, BC, Canada V5Z 1L3 e-mail: [email protected]

Keywords Familial gliomas · TP53 · Candidate loci · p53 · Apoptosis · Codon

Introduction Primary brain tumors are a mixture of different histopathologies which include both benign and malignant forms. Gliomas are the major subtype accounting for 90% of primary brain tumors. Astrocytomas, oligodendrogliomas, ependymomas, and tumors of the choroid plexus all arise from the glial origin of the central nervous system. In most cases, glioma etiology is not well understood. Ionizing radiation is one of the few well-recognized exogen factors associated with gliomas and meningiomas. Other environmental agents have been suggested (electromagnetic field, cellular phone, professional exposures) but epidemiologic studies have not been conclusive. Familial aggregation of gliomas is reported in 5–7% of the cases, suggesting a genetic etiology or common familial exposure to environmental agents. A small number of familial gliomas can be attributed to hereditary multisystem syndromes such as Neurofibromatose 1 and 2, Tuberous Sclerosis, Turcot syndrome and Li-Fraumeni syndrome. However, in many other cases, a hereditary syndrome cannot be identified and genetic alterations predisposing to familial gliomas are not yet clarified. The study of inherited predisposition to cancer has been one of the most attractive research areas in the past two decades. Indeed, the identification of susceptibility genes provides a better understanding of molecular oncogenesis, offering potential targets for therapeutic interventions. Furthermore, the ability to identify those at increased risk is of immediate clinical relevance in terms of primary and secondary interventions.

M.A. Hayat (ed.), Tumors of the Central Nervous System, Volume 2, DOI 10.1007/978-94-007-0618-7_5, © Springer Science+Business Media B.V. 2011

39

40

Genetic linkage analysis has led to locate highly penetrant genes for several common cancers, including breast and ovarian cancers (BRCA1 and BRCA2), colon cancer with adenomatous polyposis coli (APC), hereditary non-polyposis colon cancer (MSH2 and MLH1), melanoma (CDNK2A) and testicular cancer (TCG1). However, reports on familial gliomas remained few. Segregation analysis ruled out a single gene explanation and strongly incriminated a multifactorial inheritance model with a gene-gene or gene-environmental interactions. The difficulty to obtain a sufficient number of constitutional DNA from affected members of informative glioma families limited the approach of genetic linkage analysis. Studies investigating candidate loci in glioma families have frequently examined TP53 gene which encodes p53, a checkpoint protein that plays a crucial role in DNA damage repair and apoptosis. Germline TP53 mutations are found in 71–77% of classic Li-Fraumeni syndrome that predisposes individuals to a wide spectrum of cancers including brain tumors. Interestingly, previous studies identified germline TP53 mutations in several families with multiple glioma patients but without ascertained Li-Fraumeni syndrome. This chapter reviews the genetic predisposition in brain tumors and highlights the role of TP53 gene in familial gliomas, and in other forms of genetic predisposition or susceptibility to gliomas.

Genetic Predisposition to Familial Cancer Cancers arise by accumulating structural or functional alteration in DNA. Majority of the DNA alterations involve mutations of proto-oncogenes, tumor suppressor genes and DNA repair genes. Cancers may be sporadic or as a part of hereditary predisposition syndrome. Sporadic cancers are generally characterized by late onset, unilateral and isolated tumor. Mendelian cancer syndromes account for 5% of the common cancer cases. They are characterized by early onset, multifocality and secondary malignancy. They occur in multiple generations and their inheritance fits an autosomal dominant or recessive model. Different strategies have been used to identify predisposing genes in familial cancers including gliomas. Linkage analysis consists to identify an eventual locus of predisposition in one or more informative families.

S. El Hallani and I. Ratbi

Based on polymorphic DNA probes, linkage studies are to locate the predisposing gene by analyzing the cosegregation of polymorphic markers with the disease. Two linkage analysis studies have been conducted in familial gliomas and were able to identify an autosomal dominant locus at 15q23-q26 (Paunu et al., 2002), and an autosomal recessive locus at 1q21-q25 (Malmer et al., 2005). The limitation of this approach lies in the difficulty of obtaining a sufficient number of informative families for such rare cancer syndromes. A second strategy is to identify a possible locus of predisposition using fine genetic mapping of the tumors. By comparing tumor DNA with constitutional DNA of familial glioma patients, molecular techniques known as allelic loss (or loss of heterozygosity) have led to define frequent deletions of several chromosomal regions: 1p, 6q, 9p, 10q, 11p, 13q, 17p, 19q and 22q. Recurrent deletions of these chromosomal segments may reflect the existence within them of one (or more) tumor suppressor gene, whose one copy could be inactivated by this mechanism, and the other one by a germline mutation. More recently, analysis of familial glioma by CGH-array allowed to bring new fields of exploration of locus predisposition (on 6q, 12p, 16p, 22q) (Paunu et al., 2000). Testing known candidate genes is alternative when only the blood sample from the proband is available. Several candidate genes have been explored in familial gliomas. Germline TP53 tumor suppressor gene mutations were reported in some cases, whereas no mutations were identified in the PTEN, P16, CDK4 and CHEK2 genes (Tachibana et al., 2000; El Hallani et al., 2009a).

Brain Tumors in Predisposition Syndromes Hereditary tumor syndromes of the nervous system are varied and their recognition is critical to provide optimal clinical care and genetic counselling. In the last decade, many of the genes responsible for these typically autosomal dominant familial tumor syndromes have been identified.

Neurofibromatosis Type 1 The phenotype of the disease varies widely. Whereas some patients remain asymptomatic, others develop

5

Familial Gliomas: Role of TP53 Gene

multiple tumors of the nervous system (neurofibroma, optic glioma, plexiform neurofibroma, neurofibrosarcoma, astrocytoma, and meningioma), tumors of other organs (pheochromocytoma, hypothalamic tumor, rhabdomyosarcoma, duodenal carcinoid), cafe-au-lait spots, axillary or inguinal freckling, hamartomas of the iris, vascular diseases, cognitive impairment, and behavioral problems. It is caused by mutations in the NF1 tumor suppressor gene located on chromosome 17 at q11.2.

Neurofibromatosis Type 2 Bilateral vestibular schwannomas are pathognomonic of the disorder. Patients often develop schwannomas of other cranial, spinal, and peripheral nerves, as well as intracranial and intraspinal meningiomas. Less frequently, they may develop low grade gliomas and ependymomas. Other features associated with neurofibromatosis type 2 are ocular abnormalities, specifically juvenile posterior subcapsular lenticular opacities (60% to 80% of patients), retinal hamartomas, hearing loss, and occassional cafe-au-lait spots. It is caused by mutations in the NF2 tumor suppressor gene located on chromosome 22q12.

Tuberous Sclerosis Complex (TSC) This condition predisposes to hamartomas and benign tumors in the brain, heart, and kidney. Nervous system lesions of TSC include tubers in the cerebral cortex, subependymal nodules, and subependymal giant cell astrocytomas. Two tumor suppressor genes have been identified: TSC1 at 9q34 (coding for hamartin) and TSC2 at 16p13 (coding for tubulin).

Von Hippel-Lindau Syndrome (VHL) Individuals are predisposed to develop a variety of malignant and benign neoplasms, most frequently retinal, cerebellar, and spinal hemangioblastomas, renal cell carcinomas, pheochromocytomas, and pancreatic tumors. The finding of multiple hemangioblastomas at young age is highly suggestive of VHL. It is caused

41

by mutations in the VHL gatekeeper tumor suppressor gene on chromosome 3p25.

Gorlin Syndrome Also called nevoid basal cell carcinoma syndrome, it is characterized by the association of developmental abnormalities and an increased incidence of malignancy. Neoplastic manifestations are mainly basal cell carcinomas, medulloblastoma, and, occasionally, meningiomas. Gorlin syndrome is caused by abnormalities in the patched homolog 1 gene (PTCH1) on chromosome 9q31 in about 85% of cases.

Turcot Syndrome (TS) Occurrence of primary brain tumor and multiple colorectal adenomas and/or colorectal adenocarcinoma are the hallmarks of this syndrome. The most frequent brain tumors are astrocytomas, glioblastomas, medulloblastomas, ependymoma, spongioblastoma, gliosarcoma, and oligodendroglioma. TS type 1 includes patients with glioblastoma, often associated with hereditary nonpolyposis colorectal carcinoma and corresponding germline mutations in the DNA mismatch repair genes (MMR): postmeiotic segregation increased 2 (PMS2), MutL homolog 1 (MLH1), and MutS homolog 2 (MSH2). TS type 2 includes patients with medulloblastoma and familial adenomatous polyposis that result from mutations in the adenomatous polyposis of the colon gene (APC).

Cowden Syndrome and Lhermitte-Duclos Disease Cowden syndrome is a rare autosomal dominant disorder, associated with thyroid, breast, and uterine cancer as well as tricholemmomas, hamartomas, skin tumors, and meningiomas. Germline mutations in the phosphatase and tensin homolog (PTEN) gene located on chromosome 10q22-23 were described. LhermitteDuclos disease, also known as dysplastic cerebellar gangliocytoma, is strongly associated with Cowden syndrome.

42

Melanoma-Astrocytoma Syndrome People with this rare syndrome have an increased risk of developing malignant cutaneous melanoma and nervous system tumors such as astrocytoma, neurofibroma, schwannoma, and meningioma. To date all published families documented with melanomaastrocytoma syndrome are linked to the CDKN2 locus on 9p21.

Li-Fraumeni Syndrome Li-Fraumeni syndrome (LFS) is a rare but highly penetrant autosomal cancer predisposition syndrome that is characterized by a familial clustering of early onset tumors including sarcomas, breast cancers, brain tumors and adrenocortical carcinomas (Li et al., 1988; Malkin et al., 1990). Initially considered as a rare syndrome, LFS and its variants are increasingly recognized as one of the most frequent and diverse forms of predisposition to cancer. To fulfill the classical LFS definition, the family history must include a proband with sarcoma before 45 years of age; a first degree relative of the proband with any cancer below 45 years and another first or second degree relative in the same lineage with any cancer before 45 years or sarcoma at any age. The most common childhood and adolescent cancers are soft-tissue sarcomas and osteosarcomas. Leukemia and brain tumors, including choroid plexus tumors, are almost exclusively confined to infants and early childhood, whereas adrenocortical carcinomas occur from infancy through late. In young adults, breast cancer is by far the most common malignancy. Other cancers including early onset melanoma, lung, gastric, pancreatic, prostate and colorectal cancer were also described in excess in some families. Germline TP53 gene mutations are the underlying cause of LFS in 71–77% of the cases (Evans et al., 2002). The TP53 tumor suppressor gene located on chromosome 17p13 encodes a protein involved in many overlapping cellular pathways that control cell proliferation and homeostasis, such as cell cycle, apoptosis, and DNA repair. The p53 protein is a transcription factor constitutively expressed in most cell types and activated in response to various stress signals. Loss of p53 function is thought to suppress a mechanism of protection against accumulation of genetic alterations.

S. El Hallani and I. Ratbi

Somatic TP53 genetic alterations are frequent in a variety of human sporadic cancers, with frequencies varying from 10 to 60%, depending on the tumor type or population group. Germline TP53 mutations may contribute up to 17% of all familial cancer cases. Both somatic and germline mutations are compiled in a worldwide database at the International Agency for Research on Cancer (www-p53.iarc.fr). Most mutations result in missense substitutions that are scattered throughout the gene but are particularly dense in exons 5–8, encoding the DNA binding domain. Analysis of tumor patterns in carriers of a TP53 germline mutation has demonstrated some genotype/phenotype correlations. Brain tumors were more likely associated with missense mutations within the DNA-binding surface of the p53 protein that makes contact with the minor groove of DNA (Petitjean et al., 2007). The majority of them were found at codons 175, 245 and 248. These mutant proteins exhibit loss of transactivation function and dominant-negative effects. To date, no other gene has been significantly associated with LFS. Reports that germline mutations in the CHEK2 gene may predispose to LFS have not been substantiated (Vahteristo et al., 2001).

Familial Gliomas: Role of TP53 Gene Brain tumors are the third most common tumors arising from germline TP53 mutations. It is also estimated that not more than a half of families with germline TP53 mutations fulfill the definition of classic LFS. Therefore, the possibility that germline TP53 mutations could play an important role in familial aggregation of gliomas was suggested. Various case reports were first supportive of this hypothesis. Lübbe et al. (1995) described clinical, neuropathological and molecular genetic findings in a Swiss family with four brain tumors in only two generations. The neoplasms observed covered a wide range from a slowly growing lesion already apparent at birth, to anaplastic astrocytoma in a young adult and glioblastomas at the age of less than 10 years. A germline deletion of codon 236 of the TP53 gene was identified as an underlying cause and detected in all affected family members. Vital et al. (1998) reported 2 French families with an identical germline TP53 mutation in codon 248 and a clustering of brain tumors. The youngest patient

5

Familial Gliomas: Role of TP53 Gene

in each family developed a malignant choroid plexus tumor while several young adults of both kindred succumbed to low-grade astrocytoma, anaplastic astrocytoma or glioblastoma. To determine whether TP53 gene plays an important role in familial glioma, the frequency of germline mutations was then investigated. For this purpose, van Meyel et al. (1994) went on to screen 26 members of 16 glioma families for germline TP53 mutations of exons 5 through 9 using a polymerase chain reaction-single-strand conformation polymorphism method. However, no germline mutations were found. Tachibana et al. (2000) investigated 15 familial gliomas by screening exons 5 through 8 and found one GBM patient exhibiting germline TP53 mutation. Careful examination of the corresponding family history showed multiple first degree relatives with GBM and a variety of other cancers (lung, colon, uterine and ovarian), many of them occurring before the age of 45, but not all the criteria for classical LFS were fulfilled; especially no family members developed a sarcoma. Paunu et al. (2001) identified 18 families with two or more gliomas through questionnaires sent to 369 consecutive glioma patients operated at Tampere University Hospital during 1983–1994. Again, sequencing analysis of exons 4 through 10 of the TP53 gene revealed no germline mutations in any of the 18 families. To date, the largest cohort of glioma families was gathered by the Salpetriere group. El Hallani et al. (2009a) examined 79 blood DNA of glioma patients who had at least one family member affected by glioma. Germline TP53 mutation screening was performed on exons 5 through 8. Five patients with variant sequences were identified. Two of them had a known polymorphism at codon 213 and the three other patients had missense TP53 mutations at codon 141, 251, and 273. Only one out of the mutated patient met the criteria for classic LFS. Taken together, only 2.5% (2/78) of familial glioma patients without ascertained LFS carried a mutated TP53 gene. In most studies, it was considered adequate to analyze the highly conserved regions of the TP53 gene, which are exons 5–8. Earlier analyses covering all exons, splice junctions and the promoter region of the TP53 gene suggested that 10% of germline mutations lie outside exons 4–10. Thus, previous reports have only provided a fairly thorough approximation of mutations within the TP53 gene. The fact that the occurrence of germline TP53 mutations among glioma families was a rare event speaks against the claim that

43

such mutations would play a significant role in the genesis of familial gliomas. The relatively rare germline TP53 mutation in glioma families should promptly alert one to an incomplete form of LFS rather than a distinct glioma predisposition. An immunohistochemical case-control analysis of the tumor samples was performed to compare the rate of TP53 mutations and consequent aberrant p53 expression in familial and sporadic gliomas (Paunu et al., 2001). This analysis indicated that overexpressing p53 protein is as common in familial as in sporadic gliomas. In the absence of germline mutations, p53 accumulation probably arises from somatic mutations. Thus, somatic mutations of the p53 gene seem to be similarly involved in the pathogenesis of familial and sporadic gliomas. Together with the germline mutation analysis, this suggests that the role of the TP53 tumor suppressor gene is similar in the tumorigenesis of most familial and sporadic gliomas.

TP53 in Other Forms of Genetic Predisposition to Gliomas In addition to familial aggregation of gliomas, clinicians should be aware of other overlapping situations in which genetic etiology should be suspected: patients with multifocal glioma, glioma as another primary malignancy, and glioma associated with a family history of cancer without ascertained Li-Fraumeni syndrome. Gliomas usually present as a single lesion and rarely metastasize. However, multifocal gliomas, which exhibit more than one independent focus of involvement, were found to occur with a frequency of 9–11%. Moreover, multifocal gliomas are frequently associated with other primary malignancies. Kyritsis et al. (1994) investigated whether germline TP53 gene mutations may account for this phenomenon. Overall, nine out of 39 patients (23%) with multifocal glioma, secondary malignancies, or a family history of cancer exhibited heterozygous germline TP53 mutations. All these mutations would have resulted in amino acid changes in the p53 protein. They were detected in six out of 19 patients with multifocal glioma (including two with family history of cancer, one with another primary malignancy, and two with all three risk factors); one out of 4 patients with unifocal glioma, another primary malignancy and

44

a family history of cancer; and two out of 15 patients with unifocal glioma and a family history of cancer but no second malignancies. No mutations were detected in the patient with unifocal glioma and another malignancy or in the 12 control patients with unifocal glioma and no second malignancies or family history of cancer. Patients having mutations were younger than other patients in the same group. Combinations of these risk factors increased the percentage of germline TP53 gene mutations to approximately 43% if two factors were present, or 67% in the presence of all three factors. Therefore, screening germline TP53 mutations is recommended if theses factors are combined in order to identify high risk relatives for genetic counseling, early cancer detection, and possible enrollment in chemoprevention trials.

TP53 Polymorphism in Sporadic Gliomas Inherited cancer syndromes are associated with rare and highly penetrant monogenic mutations, but genetic factors also play a role in sporadic cancers. Single Nucleotide Polymorphisms (SNP) can affect protein function, promoter activity, messenger RNA stability, and splice variants and therefore can result in a change in the cellular ability to cope with DNA damage, which contributes to cancer susceptibility. Most of the association studies that aimed at identifying low-penetrance variants for susceptibility to gliomas have been based on a candidate gene approach. One of the critical molecular steps of gliomagenesis involves the p53 pathway. A functional SNP at codon 72 of TP53 gene results in the presence of either proline (Pro) or arginine (Arg) in the amino acid sequence of TP53. This change involves DNA-binding motifs of p53 and the Pro72 variant results in weaker induction of apoptosis than the Arg72 variant (Pim and Banks, 2004). Then this polymorphism has been linked to glioma susceptibility but results from different studies are controversial. The Pro72 variant was also suspected of altering the tumor sensitivity to irradiation and chemotherapy. However, conducted glioma studies suggest that the Pro72 variant does not have any prognostic impact on the survival of GBM patients, neither on less aggressive histologic subgroups of glial tumors. Interestingly, El Hallani et al. (2009b) reported

S. El Hallani and I. Ratbi

that the Pro/Pro genotype is associated with earlier onset in GBM population (49.1 years) compared to Arg/Arg (56.6 years) and Arg/Pro (55.9 years). The Pro/Pro genotype was significantly more represented in young patients (less than 45 years old) than in elderly. The accelerating effect of functional p53 insufficiency on tumorigenesis was previously demonstrated. In a model of mice lacking the xeroderma pigmentosum group A gene (XPA–/– mice) and highly predisposed to tongue tumors when exposed to 4-nitroquinoline 1-oxide, the p53 haploinsufficiency (TP53–/+) and complete inactivation (TP53–/–) did not increase the incidence of cancer but accelerated dramatically tumor development (Ide et al., 2003). Genomic alterations and molecular pathways involved in GBM differ between young and older patients. TP53 inactivation is a prominent mechanism in the GBM of child and young adult compared to the GBM of older patient. As a consequence, young patients might be more sensitive to the functional variation of TP53 codon 72, explaining that Pro/Pro genotype constitutes a potent risk factor in this population. In conclusion, clinicians should inquire about any family history of brain tumors and be familiar with the most important hereditary syndromes with neoplastic manifestations in the nervous system, so that these diagnoses are not missed. Genetic alterations predisposing to familial aggregation of gliomas without ascertained hereditary syndrome are yet to be clarified. Although occasional glioma families carrying germline TP53 mutations have been identified, current data do not support routine screening of TP53 gene, except for atypical combination of multifocality and secondary malignancy. As suggested by segregation analysis, a model of familial glioma predisposition based solely on high-risk mutations seems unlikely, and much of the inherited risk is likely to be a consequence of the coinheritance of multiple low-risk variants. However, our knowledge of predisposition to glioma is developing. The advent of genome-wide association study will enable researchers to identify variants that influence an individual’s susceptibility to develop glioma. Such studies are at the vanguard of the new technologies and started to identify risk loci in gliomas. Identifying the sequence changes responsible for causal associations should thus provide insight into the biological mechanisms of glioma, and this may lead to the development of etiological hypotheses regarding non genetic risk factors.

5

Familial Gliomas: Role of TP53 Gene

References El Hallani S, Boisselier B, Marie Y, Paris S, Idbaih A, Carpentier C, Hoang-Xuan K, Delattre JY, Sanson M (2009a) TP53 mutations but no CHEK2 ∗ 1100DelC variant in familial gliomas. Cancer Genet Cytogenet 188:126–128 El Hallani S, Ducray F, Idbaih A, Marie Y, Boisselier B, Colin C, Laigle-Donadey F, Rodéro M, Chinot O, Thillet J, Hoang-Xuan K, Delattre JY, Sanson M (2009b) TP53 codon 72 polymorphism is associated with age at onset of glioblastoma. Neurology 72:332–336 Evans DG, Birch JM, Thorneycroft M, McGown G, Lalloo F, Varley JM (2002) Low rate of TP53 germline mutations in breast cancer/sarcoma families not fulfilling classical criteria for Li-Fraumeni syndrome. J Med Genet 39:941–944 Ide F, Kitada M, Sakashita H, Kusama K, Tanaka K, Ishikawa T (2003) p53haploinsufficiency profoundly accelerates the onset of tongue tumors in mice lacking the xeroderma pigmentosum group A gene. Am J Pathol 163:1729–1733 Kyritsis AP, Bondy ML, Xiao M, Berman EL, Cunningham JE, Lee PS, Levin VA, Saya H (1994) Germline p53 gene mutations in subsets of glioma patients. J Natl Cancer Inst 86:344–349 Li FP, Fraumeni JF, Mulvihill JJ, Blattner WA, Dreyfus MG, Tucker MA, Miller RW (1988) A cancer family syndrome in twenty-four kindreds. Cancer Res 48:5358–5362 Lübbe J, von Ammon K, Watanabe K, Hegi ME, Kleihues P (1995) Familial brain tumour syndrome associated with a p53 germline deletion of codon 236. Brain Pathol 5:15–23 Malkin D, Li FP, Strong LC, Fraumeni JF, Nelson CE, Kim DH, Kassel J, Gryka MA, Bischoff FZ, Tainsky MA (1990) Germ line p53 mutations in a familial syndrome of breast cancer, sarcomas, and other neoplasms. Science 250:1233–1238 Malmer B, Haraldsson S, Einarsdottir E, Lindgren P, Holmberg D (2005) Homozygosity mapping of familial glioma in Northern Sweden. Acta Oncol 44:114–119 Paunu N, Lahermo P, Onkamo P, Ollikainen V, Rantala I, Helén P, Simola KO, Kere J, Haapasalo H (2002) A novel low-

45 penetrance locus for familial glioma at 15q23-q26.3. Cancer Res 62:3798–3802 Paunu N, Sallinen SL, Karhu R, Miettinen H, Sallinen P, Kononen J, Laippala P, Simola KO, Helén P, Haapasalo H (2000) Chromosome imbalances in familial gliomas detected by comparative genomic hybridization. Genes Chromosomes Cancer 29:339–346 Paunu N, Syrjäkoski K, Sankila R, Simola KO, Helén P, Niemelä M, Matikainen M, Isola J, Haapasalo H (2001) Analysis of p53 tumor suppressor gene in families with multiple glioma patients. J Neurooncol 55:159–165 Petitjean A, Mathe E, Kato S, Ishioka C, Tavtigian SV, Hainaut P, Olivier M (2007) Impact of mutant p53 functional properties on TP53 mutation patterns and tumor phenotype: lessons from recent developments in the IARC TP53 database. Hum Mutat 28:622–629 Pim D, Banks L (2004) p53 polymorphic variants at codon 72 exert different effects on cell cycle progression. Int J Cancer 108:196–199 Tachibana I, Smith JS, Sato K, Hosek SM, Kimmel DW, Jenkins RB (2000) Investigation of germline PTEN, p53, p16(INK4A)/p14(ARF), and CDK4 alterations in familial glioma. Am J Med Genet 92:136–141 Vahteristo P, Tamminen A, Karvinen P, Eerola H, Eklund C, Aaltonen LA, Blomqvist C, Aittomäki K, Nevanlinna H (2001) p53, CHK2, and CHK1 genes in Finnish families with Li-Fraumeni syndrome: further evidence of CHK2 in inherited cancer predisposition. Cancer Res 61:5718–5722 van Meyel DJ, Ramsay DA, Chambers AF, Macdonald DR, Cairncross JG (1994) Absence of hereditary mutations in exons 5 through 9 of the p53 gene and exon 24 of the neurofibromin gene in families with glioma. Ann Neuro. 35:120–122 Vital A, Bringuier PP, Huang H, San Galli F, Rivel J, Ansoborlo S, Cazauran JM, Taillandier L, Kleihues P, Ohgaki H (1998) Astrocytomas and choroid plexus tumors in two families with identical p53 germline mutations. J Neuropathol Exp Neurol 57:1061–1069

Chapter 6

The Role of IDH1 and IDH2 Mutations in Malignant Gliomas Yukihiko Sonoda, Ichiyo Shibahara, Ryuta Saito, Toshihiro Kumabe, and Teiji Tominaga

Abstract A recent study identified mutations in the active sites of isocitrate dehydrogenase 1 and 2 (IDH1and IDH2) genes in several types of glioma. All mutations affected a single amino acid located in the binding site of isocitrate (R132 of IDH1 and R172 of IDH2). We analyzed the genomic region spanning wild-type R132 of IDH1 and R172 of IDH2 by direct sequencing in 125 glial tumors. A total of 39 IDH1 mutations and one IDH2 mutation were observed. IDH1 and IDH2 mutations were frequently present in astrocytic and oligodendroglial tumors. However, primary glioblastomas were characterized by a low frequency of mutations (5%) at amino acid position 132 of IDH1. Mutations of IDH1 and IDH2 genes were significantly associated with improved outcome in patients with anaplastic astrocytomas. IDH1 and IDH2 mutations seem to play an important role in early tumor progression of specific types of glioma and might arise from a common glial precursor. The infrequency of IDH1 mutation in primary glioblastomas revealed that these subtypes are entities that are genetically distinct from other glial tumors. Analyses of IDH1 and IDH2 status have significant utility for diagnosis and treatment of patients with gliomas. Keywords IDH1 and IDH2 · Mutations · Gliomas · Glioblastomas · Grades of glioma · Arginine

Introduction Gliomas are the most common type of primary brain tumor and are grouped into four grades according Y. Sonoda () Department of Neurosurgery, Tohoku University Graduate School of Medicine, Aoba-ku, Sendai-shi, Miyagi 980-8577, Japan e-mail: [email protected]

to the World Health Organization (WHO) criteria (Louis et al., 2007). Gliomas include several specific histological subtypes; the most common are astrocytomas, oligodendrogliomas, and ependymomas. Glioblastomas (GBMs) (WHO grade IV), the most malignant type of glioma, may develop very rapidly de novo (primary glioblastoma) in elderly patients, or develop more slowly from low-grade diffuse (DA) (WHO grade II) or anaplastic astrocytoma (AA) (WHO grade III) (secondary glioblastoma) in younger patients (Ohgaki and Kleihues, 2009). Malignant gliomas are believed to develop as the result of stepwise accumulations of genetic lesions. Several genes, including TP53, PTEN, CDKN2A/CDKN2B, and EGFR, are altered in gliomas (Ohgaki and Kleihues, 2009). These alterations tend to occur in a defined order during the progression to a high-grade tumor. TP53 mutation appears to be a relatively early event during the development of an astrocytoma, whereas the loss or mutation of PTEN and amplification of EGFR are characteristic of higher-grade tumors (Furnari et al., 2007). Oligodendrogliomas (O) (WHO grade II) and anaplastic oligodendrogliomas (AO) (WHO grade III) typically show 1p/19q codeletion (Reifenberger et al., 1994). Recent genome-wide mutational analysis has revealed somatic mutations of the cytosolic nicotinamide adenine dinucleotide phosphate (NADP+ )dependent isocitrate dehydrogenase (IDH1) gene at 2q33 in approximately 12% of GBMs (Parsons et al., 2008). Isocitrate dehydrogenase catalyzes the oxidative decarboxylation of isocitrate to alphaketoglutarate thereby enabling NADPH production. Mutations were found that affect the amino acid arginine in position 132 of the amino acid sequence, which is evolutionarily highly conserved, and is located in

M.A. Hayat (ed.), Tumors of the Central Nervous System, Volume 2, DOI 10.1007/978-94-007-0618-7_6, © Springer Science+Business Media B.V. 2011

47

48

the binding site of isocitrate (Xu et al., 2004). In the vast majority of these cases, wild-type arginine in position 132 was replaced by histidine (R132H) (Balss et al., 2008; Parsons et al., 2008; Ichimura et al., 2009; Sonoda et al., 2009; Watanabe et al., 2009; Yan et al., 2009b). The IDH2 gene is homologous to IDH1, which uses NADP+ as an electron receptor. Gliomas without IDH1 mutations were often found to have mutations at the analogous amino acid (R172) of the IDH2 gene (Yan et al., 2009b). Both IDH1and IDH2 mutations reduced enzymatic activity of the encoded protein.

Methodology We obtained 125 samples of tumor tissue from surgical patients diagnosed and treated at Tohoku University Hospital. Resected specimens were quick-frozen in liquid nitrogen and kept at –80◦ C until nucleic acid extraction (Sonoda et al., 2009). The series included 2 DA, 21 AA, 58 primary glioblastomas (prGBM), 3 secondary glioblastomas (secGBM), 8 O, 14 AO, 4 anaplastic oligoastrocytomas of WHO grade III (AOA), 8 gangliogliomas of WHO grade II (GG), and 5 anaplastic gangliogliomas (AG) of WHO grade III. Exon 4 of the IDH1 gene including codon 132 was amplified using 100 ng each of sense primer 5 -CGGTCTTCAGAGAAGCCATT and antisense primer 5 -GCAAAATCACATTATTGCCAAC. A fragment of 219 bp in length spanning the catalytic domain of IDH2 including codon 172 was amplified using 100 ng each of sense primer 5 CAAGCTGAAGAAGATGTGGAA-3 and antisense primer 5 -CAGAGACAA GAGGATGGCTA-3 . PCR was performed using standard buffer conditions, namely, 100 ng of DNA and Ex-Taq HS DNA Polymerase (Takara Bio Inc., Shiga, Japan) employed for 30 cycles with denaturing at 95◦ C for 30 s, annealing at 56◦ C for 30 s, and extension at 72◦ C for 40 s in a total volume of 50 ul. The PCR products were purified using a highly pure PCR product purification kit (Roche, Basel, Switzerland). All sequence reactions were performed using the GenomeLabTM DTCS quick-start kit (Beckman Coulter, Inc., Fullerton, CA). The reactions were carried out in an automated DNA analyzer (CEQ 8000; Beckman Coulter).

Y. Sonoda et al.

Results IDH Mutations in Gliomas We detected 39 mutations in the IDH1 gene in 125 tumors. All mutations were heterozygous with one wild-type allele being present. Only codon 132 of IDH1 was affected by mutations and all mutations were of the R132H type. In addition, the mutation of IDH2 at codon 172 was detected in one AA without IDH1 mutation. The type of mutation was R172S (AGG to AGT). Mutations of IDH1 or IDH2 were frequently observed in AA (62%), secGBM (67%), O (67%), AO (50%), AOA (75%), GG (38%), and AG (60%). Only a few mutations occurred in prGBM (5%). Despite no mutations in DA, we could not draw any reliable conclusions because of the fact that there were only two cases.

Patient Survival and IDH Mutation Patients with GBM carrying an IDH1 mutation (n = 5) had a median survival of 66 months, which was longer than the 17-month survival in patients with wild-type IDH1 (n = 57). However, there was no significant difference because there were too few tumors of GBM with IDH1 mutation (P = 0.1; log-rank test). Mutations of the IDH1 and IDH2 genes were associated with prolonged overall survival in AA patients; the median overall survival was 50 months for patients with mutation (n = 13) and 22 months for those without mutation (n = 8) (P < 0.001; log-rank test).

Discussion Detection of IDH1 and IDH2 Mutations The current routine procedure for assessing IDH gene status is DNA sequencing. All mutations of the IDH1 gene were somatic and missense mutations at codon 132 (arginine). Of these, almost all mutations were of the R132H type; however, 5 other mutations leading to R132C, R132S, R132G, R132L, and R132V were found. Similarly, all IDH2 mutations were found in codon 172; these mutations resulted in amino acid exchanges from arginine to guanine, methionine,

6

The Role of IDH1 and IDH2 Mutations in Malignant Gliomas

serine, and lysine (Balss et al., 2008; Hartmann et al., 2009; Sonoda et al., 2009; Yan et al., 2009b). DNA extraction and sequencing are not available at every institute; however, recently mutation-specific IDH1 antibodies for the most frequent mutation of the R132H type have shown useful properties of high sensitivity and selectivity (Kato et al., 2009; CameloPiragua et al., 2010; Capper et al., 2010).

IDH1 and IDH2 Mutations are Frequently Found in Various Grades of Glioma According to previous reports, no pilocytic astrocytoma with an IDH1 mutation has been found, with the exception of a single case (Balss et al., 2008; Korshunov et al., 2009). High frequencies of IDH1 mutations were found in DA, AO, O, and AO (Watanabe et al., 2009; Yan et al., 2009b). Similarly, IDH1 mutations were more frequent in secGBM than prGBM (Yan et al., 2009b). Among other CNS tumors, IDH1 mutation has been reported in pleomorphic xanthoastrocytoma, primitive neuroectodermal tumor, and GG (Balss et al., 2008; Sonoda et al., 2009; Yan et al., 2009b). On the other hand, no IDH1 mutation was found in ependymal tumors, medulloblastomas, meningeal tumors, and schwannoma (Balss et al., 2008; Ichimura et al., 2009; Yan et al., 2009b). Although the frequency of IDH1 mutation in younger

Fig. 6.1 A model of glioblastoma development. Abbreviations: IDH1/2, isocitrate dehydrogenase 1/2; mut, mutation; 1p19q codel., 1p 19q codeletion; CDKN2A, B, cyclin-dependent kinase inhibitor 2A, B; HD, homozygous deletion; PTEN, phosphatase tensin homolog; EGFR, epidermal growth factor receptor; amp,

49

patients is higher than that in older patients, this mutation has rarely been found in pediatric tumors (Balss et al., 2008). Among non-CNS tumors, IDH1 mutation was found in two prostate cancers and one B-ALL (Bleeker et al., 2009; Kang et al., 2009). In addition, IDH1 mutation has been identified in a subset of acute myeloid leukemia cases (Mardis et al., 2009). IDH2 mutations were only found in WHO grade II or III gliomas, usually without IDH1 mutation. According to previous reports, only one grade III glioma showed both IDH1 and IDH2 mutations (Wick et al., 2009).

IDH1 and IDH2 Mutations Are Early Events in Gliomagenesis IDH1 and IDH2 mutations are commonly found in both astrocytomas and oligodendrogliomas, but the pattern of other genetic alterations is different. Usually, TP53 mutations and 1p/19q codeletion were exclusive, with the exception of only a few cases (Ichimura et al., 2009; Watanabe et al., 2009). The majority of lowgrade diffuse astrocytomas have both TP53 and IDH1 mutations, whereas most oligodendrogliomas show both IDH1 mutations and 1p/19q codeletion (Ichimura et al., 2009; Watanabe et al., 2009). Therefore, IDH1 and IDH2 mutations are very early events in gliomagenesis and may affect common glial precursors (Fig. 6.1).

amplification; DA, diffuse astrocytoma; AA, anaplastic astrocytoma; O, oligodendroglioma; AO, anaplastic astrocytoma; prGBM, primary glioblastoma; secGBM, secondary glioblastoma

50

The pattern of other genetic alterations in gliomas with IDH mutations is entirely different from that in gliomas without IDH mutations. Alterations of PTEN, EGFR, or CDKN2A/CDKN2B are frequently found in GBMs with wild-type IDH1 and IDH2, but in only 5% of cases of AA and GBM with IDH mutation (Yan et al., 2009a; b). Similarly, 1p/19q codeletion was rarely found in oligodendroglial tumors without IDH mutation (Watanabe et al., 2009). TP53 mutations were frequently found in AA with IDH1 mutation; however, these were often found in prGBMs without IDH1 mutations (Ohgaki and Kleihues, 2009; Yan et al., 2009a).

IDH Mutation as a Favorable Prognostic Factor in Glioma Patients Clinically, IDH1 mutation is strongly correlated with good prognosis in patients with several grades of glioma (Sanson et al., 2009; Sonoda et al., 2009; Wick et al., 2009; Yan et al., 2009b). Multivariate analysis has confirmed that IDH1 mutation is an independent favorable prognostic marker in anaplastic gliomas and glioblastomas after adjustment for other factors (age, histology, MGMT promoter methylation status, and genomic profile) and treatment modality (extent of resection) (Sanson et al., 2009; Wick et al., 2009).

Fig. 6.2 The roles of IDH1 and IDH2 in cellular metabolism. Abbreviations: IDH1/2, isocitrate dehydrogenase 1/2; α-KG, α-ketoglutarate; 2-HG, 2-hydroxyglutarate; 2HGD, PHD,

Y. Sonoda et al.

2-Hydroxyglutarate (2-HG) Levels Are Elevated in Gliomas with IDH1 Mutations Although the biological function of IDH mutation is not fully understood, IDH mutations are always monoallelic and do not result in simple loss of function. IDH1 mutation impairs the enzyme’s affinity for its substrate and dominantly inhibits wild-type IDH1 activity through the formation of catalytically inactive heterodimers (Zhao et al., 2009). Consequently, R132H mutation disrupts the function of IDH1 to convert isocitrate to α-ketoglutarate. More recently, IDH1 mutations were found to result in a new ability of the enzyme to catalyze the NADPH-dependent reduction of α-ketoglutarate to R(-)-2-HG (2-HG) (Fig. 6.2) (Dang et al., 2009). 2-HG levels were remarkably elevated in human malignant gliomas with IDH1 mutations (Dang et al., 2009). Structural studies demonstrated that the replacement of arginine 132 by histidine results in a shift of the residues in the active site to produce structural changes consistent with reduced oxidative decarboxylation of isocitrate and acquisition of the ability to convert α-ketoglutarate to 2-HG. The accumulation of 2-HG is found in patients with inherited metabolic disorder 2-hydroxyglutaric aciduria. This disease is caused by a deficiency in the enzyme 2-HG dehydrogenase, which converts 2-HG to α-ketoglutarate (Struys et al., 2005). Patients with 2-HG dehydrogenase deficiency have

2-hydroxyglutarate dehydrogenase prolyl hydroxylase; HIF-1, hypoxia-inducible factor 1; HIF-1-OH, hydroxylated HIF-1

6

The Role of IDH1 and IDH2 Mutations in Malignant Gliomas

an increased risk of developing brain tumors (Aghili et al., 2009). The effect of IDH1 mutation on cellular metabolism requires more investigation, but reduction of α-ketoglutarate by 2-HG or mutant IDH1 results in a lower level of prolyl hydroxylases and promotes the accumulation of hypoxia-inducible factor 1 (HIF-1) (Fig. 6.2). Alterations in HIF-1 may result from mutant IDH1 protein expression. In addition to IDH mutations in gliomas, mutations of other metabolic enzymes such as fumarate hydratase and succinate dehydrogenase occur in paraganglioma and leiomyoma, respectively (King et al., 2006). Energy is produced predominantly by aerobic glycolysis in the cytoplasm of most cancer cells, rather than by mitochondrial oxidative phosphorylation as in normal differentiated cells, a phenomenon termed the “Warburg effect.” Aerobic glycolysis may provide cancer cells with a growth advantage by supplying required metabolites for incorporation into the biomass to produce new cells. Aerobic glycolysis may also be important for glioma progression. Both PI3K and tyrosine kinase signaling are involved in growth control and glycolysis. PI3K activation by PTEN mutation and/or tyrosine kinase activation by EGFR alteration are frequently found in prGBMs. However, PTEN mutation and EGFR alterations are rarely found in AA and GBM patients with IDH mutation, suggesting that the mechanism of cellular metabolism in glioma might depend on IDH status.

Conclusions IDH mutations seem to play an important role in the formation of astrocytic and oligodendroglial tumors. Detection of IDH mutations is easier to perform and interrupt than the determination of 1p/19q codeletion and MGMT promoter methylation. Such information could be useful to improve the diagnostic and therapeutic strategies for gliomas. Furthermore, measurement of 2-HG production will enable identification of patients with IDH1 mutant brain tumors. Of course, further analysis of IDH1 and IDH2 in glioma model systems will be necessary to clarify the genetic mechanisms involved in the initiation and malignant progression of this disease. In addition, extensive genetic profiling of gliomas may allow the molecular classification of gliomas to replace the current histological classification in the near future. Patients with

51

IDH1 mutation may benefit from treatment modalities designed to inhibit the mutant IDH1 expression. Inhibition of 2-HG production might also have therapeutic potential in the treatment of gliomas.

References Aghili M, Zahedi F, Rafiee E (2009) Hydroxyglutaric aciduria and malignant brain tumor: a case report and literature review. J Neurooncol 91:233–236 Balss J, Meyer J, Mueller W, Korshunov A, Hartmann C, Von Deimling A (2008) Analysis of the IDH1 codon 132 mutation in brain tumors. Acta Neuropathol 116:597–602 Bleeker FE, Lamba S, Leenstra S, Troost D, Hulsebos T, Vandertop WP, Frattini M, Molinari F, Knowles M, Cerrato A, Rodolfo M, Scarpa A, Felicioni L, Buttitta F, Malatesta S, Marchetti A, Bardelli A (2009) IDH1 mutations at residue p.R132 (IDH1(R132)) occur frequently in high-grade gliomas but not in other solid tumors. Hum Mutat 30:7–11 Camelo-Piragua S, Jansen M, Ganguly A, Kim JC, Louis DN, Nutt CL (2010) Mutant IDH1-specific immunohistochemistry distinguishes diffuse astrocytoma from astrocytosis. Acta Neuropathol 119(4):509–511 Capper D, Weissert S, Balss J, Habel A, Meyer J, Jager D, Ackermann U, Tessmer C, Korshunov A, Zentgraf H, Hartmann C, Von Deimling A (2010) Characterization of R132H mutation-specific IDH1 antibody binding in brain tumors. Brain Pathol 20:245–254 Dang L, White DW, Gross S, Bennett BD, Bittinger MA, Driggers EM, Fantin VR, Jang HG, Jin S, Keenan MC, Marks KM, Prins RM, Ward PS, Yen KE, Liau LM, Rabinowitz JD, Cantley LC, Thompson CB, Vander Heiden MG, Su SM (2009) Cancer-associated IDH1 mutations produce 2-hydroxyglutarate. Nature 462:739–744 Furnari FB, Fenton T, Bachoo RM, Mukasa A, Stommel JM, Stegh A, Hahn WC, Ligon KL, Louis DN, Brennan C, Chin L, Depinho RA, Cavenee WK (2007) Malignant astrocytic glioma: genetics, biology, and paths to treatment. Genes Dev 21:2683–2710 Hartmann C, Meyer J, Balss J, Capper D, Mueller W, Christians A, Felsberg J, Wolter M, Mawrin C, Wick W, Weller M, Herold-Mende C, Unterberg A, Jeuken JW, Wesseling P, Reifenberger G, Von Deimling A (2009) Type and frequency of IDH1 and IDH2 mutations are related to astrocytic and oligodendroglial differentiation and age: a study of 1,010 diffuse gliomas. Acta Neuropathol 118:469–474 Ichimura K, Pearson DM, Kocialkowski S, Backlund LM, Chan R, Jones DT, Collins VP (2009) IDH1 mutations are present in the majority of common adult gliomas but rare in primary glioblastomas. Neuro-oncology 11:341–347 Kang MR, Kim MS, Oh JE, Kim YR, Song SY, Seo SI, Lee JY, Yoo NJ, Lee SH (2009) Mutational analysis of IDH1 codon 132 in glioblastomas and other common cancers. Int J Cancer 125:353–355 Kato Y, Jin G, Kuan CT, Mclendon RE, Yan H, Bigner DD (2009) A monoclonal antibody IMab-1 specifically recognizes IDH1R132H, the most common glioma-derived mutation. Biochem Biophys Res Commun 390:547–551

52 King A, Selak MA, Gottlieb E (2006) Succinate dehydrogenase and fumarate hydratase: linking mitochondrial dysfunction and cancer. Oncogene 25:4675–4682 Korshunov A, Meyer J, Capper D, Christians A, Remke M, Witt H, Pfister S, Von Deimling A, Hartmann C (2009) Combined molecular analysis of BRAF and IDH1 distinguishes pilocytic astrocytoma from diffuse astrocytoma. Acta Neuropathol 118:401–405 Louis DN, Ohgaki H, Wiestler OD, Cavenee WK, Burger PC, Jouvet A, Scheithauer BW, Kleihues P (2007) The 2007 WHO classification of tumours of the central nervous system. Acta Neuropathol 114:97–109 Mardis ER, Ding L, Dooling DJ, Larson DE, Mclellan MD, Chen K, Koboldt DC, Fulton RS, Delehaunty KD, Mcgrath SD, Fulton LA, Locke DP, Magrini VJ, Abbott RM, Vickery TL, Reed JS, Robinson JS, Wylie T, Smith SM, Carmichael L, Eldred JM, Harris CC, Walker J, Peck JB, Du F, Dukes AF, Sanderson GE, Brummett AM, Clark E, Mcmichael JF, Meyer RJ, Schindler JK, Pohl CS, Wallis JW, Shi X, Lin L, Schmidt H, Tang Y, Haipek C, Wiechert ME, Ivy JV, Kalicki J, Elliott G, Ries RE, Payton JE, Westervelt P, Tomasson MH, Watson MA, Baty J, Heath S, Shannon WD, Nagarajan R, Link DC, Walter MJ, Graubert TA, Dipersio JF, Wilson RK, Ley TJ (2009) Recurring mutations found by sequencing an acute myeloid leukemia genome. N Engl J Med 361:1058–1066 Ohgaki H, Kleihues P (2009) Genetic alterations and signaling pathways in the evolution of gliomas. Cancer Sci 15: 6002–6007 Parsons DW, Jones S, Zhang X, Lin JC, Leary RJ, Angenendt P, Mankoo P, Carter H, Siu IM, Gallia GL, Olivi A, Mclendon R, Rasheed BA, Keir S, Nikolskaya T, Nikolsky Y, Busam DA, Tekleab H, Diaz LA Jr., Hartigan J, Smith DR, Strausberg RL, Marie SK, Shinjo SM, Yan H, Riggins GJ, Bigner DD, Karchin R, Papadopoulos N, Parmigiani G, Vogelstein B, Velculescu VE, Kinzler KW (2008) An integrated genomic analysis of human glioblastoma multiforme. Science 321:1807–1812 Reifenberger J, Reifenberger G, Liu L, James CD, Wechsler W, Collins VP (1994) Molecular genetic analysis of oligodendroglial tumors shows preferential allelic deletions on 19q and 1p. Am J Pathol 145:1175–1190

Y. Sonoda et al. Sanson M, Marie Y, Paris S, Idbaih A, Laffaire J, Ducray F, El Hallani S, Boisselier B, Mokhtari K, Hoang-Xuan K, Delattre JY (2009) Isocitrate dehydrogenase 1 codon 132 mutation is an important prognostic biomarker in gliomas. J Clin Oncol 27:4150–4154 Sonoda Y, Kumabe T, Nakamura T, Saito R, Kanamori M, Yamashita Y, Suzuki H, Tominaga T (2009) Analysis of IDH1 and IDH2 mutations in Japanese glioma patients. Cancer Sci 100:1996–1998 Struys EA, Salomons GS, Achouri Y, Van Schaftingen E, Grosso S, Craigen WJ, Verhoeven NM, Jakobs C (2005) Mutations in the D-2-hydroxyglutarate dehydrogenase gene cause D-2hydroxyglutaric aciduria. Am J Hum Genet 76:358–360 Watanabe T, Nobusawa S, Kleihues P, Ohgaki H (2009) IDH1 mutations are early events in the development of astrocytomas and oligodendrogliomas. Am J Pathol 174: 1149–1153 Wick W, Hartmann C, Engel C, Stoffels M, Felsberg J, Stockhammer F, Sabel MC, Koeppen S, Ketter R, Meyermann R, Rapp M, Meisner C, Kortmann RD, Pietsch T, Wiestler OD, Ernemann U, Bamberg M, Reifenberger G, Von Deimling A, Weller M (2009) NOA-04 randomized phase III trial of sequential radiochemotherapy of anaplastic glioma with procarbazine, lomustine, and vincristine or temozolomide. J Clin Oncol 27:5874–5880 Xu X, Zhao J, Xu Z, Peng B, Huang Q, Arnold E, Ding J (2004) Structures of human cytosolic NADP-dependent isocitrate dehydrogenase reveal a novel self-regulatory mechanism of activity. J Biol Chem 279:33946–33957 Yan H, Bigner DD, Velculescu V, Parsons DW (2009a) Mutant metabolic enzymes are at the origin of gliomas. Cancer Res 69:9157–9159 Yan H, Parsons DW, Jin G, Mclendon R, Rasheed BA, Yuan W, Kos I, Batinic-Haberle I, Jones S, Riggins GJ, Friedman H, Friedman A, Reardon D, Herndon J, Kinzler KW, Velculescu VE, Vogelstein B, Bigner DD (2009b) IDH1 and IDH2 mutations in gliomas. N Engl J Med 360:765–773 Zhao S, Lin Y, Xu W, Jiang W, Zha Z, Wang P, Yu W, Li Z, Gong L, Peng Y, Ding J, Lei Q, Guan KL, Xiong Y (2009) Glioma-derived mutations in IDH1 dominantly inhibit IDH1 catalytic activity and induce HIF-1alpha. Science 324: 261–265

Chapter 7

Malignant Glioma: Isocitrate Dehydrogenases 1 and 2 Mutations Zachary J. Reitman and Hai Yan

Abstract Mutations in the cytoplasmic NADP+ dependent isocitrate dehydrogenase, IDH1, frequently occur in gliomas. The mutations are somatic, almost always heterozygous, and occur at R132, an active site residue of IDH1 that is important for its catalytic function. IDH1 mutations occur in >70% of WHO grades II and III astrocytomas and oligodendrogliomas, as well as WHO grade IV secondary glioblastomas, and more rarely in other glioma subtypes. IDH2 is the mitochondrial homolog of IDH1, and mutations in IDH2 R172, the analogous residue to IDH1 R132, also occur in these subtypes of gliomas, albeit at a much lower frequency. R132H and R172K are the most common IDH1 and IDH2 mutations observed in gliomas, respectively, though alterations to other amino acids at these hotspots have also been observed. IDH1 mutations frequently co-occur with loss of chromosomes 1p and 19q or point mutations of TP53, and they appear to occur before other genetic alterations during glioma pathogenesis. Furthermore, IDH1 mutations are associated with a younger age at diagnosis and a better prognosis for many glioma subtypes. The frequency of IDH1 and IDH2 mutations in specific glioma subtypes suggests that testing for these mutations may help to guide clinical decision-making for glioma patients, and PCR- and antibody-based tests have been developed to determine tumor mutation

H. Yan () The Pediatric Brain Tumor Foundation Institute, The Preston Robert Tisch Brain Tumor Center, Duke University Medical Center, Durham, NC 27710, USA; The Department of Pathology, Duke University Medical Center, Durham, NC 27710, USA e-mail: [email protected]

status. The IDH1 and IDH2 mutations abolish the normal function of the encoded enzymes to oxidize isocitrate to α-ketoglutarate and confer a neomorphic gain of enzymatic activity to reduce α-ketoglutarate to 2-hydroxyglutarate. This gain of function suggests that IDH1 and IDH2 mutations are oncogenic, but the mechanism by which this promotes glioma pathogenesis remains unknown. Keywords Malignant glioma · IDH1 · Mutation · R132 · Enzyme · Dehydrogenase

Introduction The study of genetic alterations that arise in tumors of the central nervous system represents a decades-old field. Recent whole-genome analyses have advanced this field by integrating years of careful observations and by revealing genetic changes that were not obvious in previous studies. Foremost among discoveries of novel genetic changes are frequent mutations of IDH1 and IDH2 in gliomas. These mutations stand apart from other genetic alterations in cancer due in part to the fact that they occur with striking frequency in certain glioma subtypes. Also, while the vast majority of genetic alterations in cancer affect genes involved in cellular signaling and growth control, the IDH1 and IDH2 mutations affect cellular metabolic enzymes without known signaling or growth regulating properties. Furthermore, rather than simply activate or inactivate the normal function of the genes as is typically the result for alterations in cancer, the IDH1 and IDH2 mutations bestow a novel neomorphic activity on the encoded enzymes. Study of these unique mutations may further our understanding of the pathology of

M.A. Hayat (ed.), Tumors of the Central Nervous System, Volume 2, DOI 10.1007/978-94-007-0618-7_7, © Springer Science+Business Media B.V. 2011

53

54

tumors of the central nervous system and possibly provide avenues to better manage these diseases. Mutations in IDH1 were discovered in a wholegenome sequencing analysis of coding exons in glioblastoma samples (Parsons et al., 2008). IDH1 is the cytoplasmic nicotinamide adenine dinucleotide phosphate (NADP+ )-dependent isocitrate dehydrogenase. IDH1 mutations observed in glioma specifically alter codon 132, which encodes an arginine (R132) in the enzyme’s substrate binding site. The mutations are somatic and almost always heterozygous. Additionally, mutations in the analogous R172 codon of IDH2, which encodes the mitochondrial NADP+ -dependent isocitrate dehydrogenase which is the only homolog of IDH1, also occur in many of the same subtypes of glioma at a much lower frequency (Yan et al., 2009). The most frequent IDH1 alteration observed in gliomas is R132H, with other alterations at this codon more rarely observed (Balss et al., 2008). This chapter considers the epidemiology and clinical characteristics of central nervous system tumors with IDH mutations, and briefly discusses the biochemical function of the mutated enzymes.

Tumor Type Distribution IDH mutations occur frequently in specific glioma subtypes. Gliomas are primary tumors of the central nervous system that histologically resemble glia, the supporting cells of the brain. Like other CNS tumors, gliomas are classified by the World Health Organization from grades I to IV, with grades III–IV tumors considered malignant gliomas. The IDH mutations occur in gliomas classified as astrocytomas and oligodendrogliomas, which histologically resemble astrocytes and oligodendrocytes, respectively. Grade I pilocytic astrocytomas predominantly occur in children and seldom progress to higher-grade tumors. Grade II gliomas comprise diffuse astrocytomas, welldifferentiated oligodendrogliomas, and mixed oligoastrocytomas. These can progress to grade III gliomas of the same histology, with grade II diffuse astrocytomas progressing to grade III anaplastic astrocytomas, grade II well-differentiated oligodendrogliomas progressing to grade III anaplastic oligodendrogliomas, and grade II oligoastrocytomas progressing to grade III anaplastic oligoastrocytomas. Glioblastomas are

Z.J. Reitman and H. Yan

grade IV astrocytomas that have a particularly varied and dysplastic histology and carry a grim prognosis. Secondary glioblastomas, by definition, arise from grade II and III astrocytomas. In contrast to this, the majority of glioblastomas are primary glioblastomas which arise de novo, rather than from a previous lowergrade tumor (Louis et al., 2007). Figure 7.1 depicts the glioma subtypes and pathways of tumor progression discussed above. IDH1 mutations are common (>70%) in grades II and III astrocytomas and oligodendrogliomas, and in secondary glioblastomas (Balss et al., 2008). They occur more rarely in primary glioblastomas (3–16%) (Balss et al., 2008; Parsons et al., 2008; Bleeker et al., 2009; Yan et al., 2009), and are seldom found in grade I pilocytic astrocytomas (0–2%) (Balss et al., 2008; Ichimura et al., 2009; Korshunov et al., 2009; Yan et al., 2009). IDH2 mutations share a similar tumor type distribution with IDH1 mutations (Yan et al., 2009). They are generally found in tumors that do not contain IDH1 mutations, and occur much less frequently than IDH1 mutations (Hartmann et al., 2009; Yan et al., 2009). As an example, in a group of 1,010 grades II and III gliomas, 71% contained IDH1 mutations while only 3% contained IDH2 mutations (Hartmann et al., 2009). The percentage of astrocytic and oligodendroglial tumors containing IDH1 or IDH2 mutations based on two large studies is shown in Fig. 7.1. Gangliogliomas seldom (17% or less) contain IDH1 mutations (Sonoda et al., 2009), and extensive analysis of other CNS tumors such as medulloblastomas, meningiomas, and ependymomas has not revealed IDH mutations. IDH1 and IDH2 mutations are found in tumors but not in normal tissues from the same patient, and are almost always heterozygous with a wild-type allele (Yan et al., 2009). IDH1 and IDH2 mutations in gliomas may be specific for humans, as analogous mutations were not found in grades II–IV canine gliomas which are histopathologically indistinguishable from human gliomas (Reitman et al., 2009). However, IDH1 and IDH2 mutations also occur in up to 23% of human acute myelogenous leukemias (Mardis et al., 2009; Ward et al., 2010), and rare cases of IDH1 R132 mutations in colorectal cancer (Sjoblom et al., 2006), prostate cancer (Kang et al., 2009), and acute lymphoblastic leukemia, B-type (Kang et al., 2009) have been reported.

7

Malignant Glioma: Isocitrate Dehydrogenases 1 and 2 Mutations

55

Fig. 7.1 Schematic of pathogenesis and progression of astrocytic and oligodendroglial tumors, frequency of tumors carrying IDH1 R132 and/or IDH2 R172 mutations for each tumor type, and other common genetic alterations that occur in the pathogenesis or progression of each tumor type. Tumor grade is indicated to the left. Arrows indicate de novo pathogenesis from a normal cell, which has been speculated to be a normal stem cell, normal glial progenitor, or other normal cell, or progression from the indicated lower-grade tumor type. Grade I tumors: PA, pilocytic astrocytoma. Grade II tumors: O, welldifferentiated oligodendroglioma; OA, oligoastrocytomas; A,

diffuse astrocytoma. Grade III tumors: AO, anaplastic oligodendroglioma; AOA, anaplastic oligoastrocytoma; AA, anaplastic astrocytomas. Grade IV tumors: pGBM, primary glioblastoma; sGBM, secondary glioblastoma; sGBMO, secondary glioblastomas with oligodendroglial component. HD, homozygous deletion. N.S., not studied. Percentages were reported by Yan et al. (2009), except for OA and AOA which were reported by Hartmann et al. (2009) since relatively few OA and AOA tumors were analyzed in Yan et al. Figure reproduced from Reitman and Yan (2010) with permission and adapted to include percentages

Association with Other Genetic Alterations

group of a variety of glioma subtypes, 92% of tumors with IDH1 R132 mutations also contained either TP53 mutations or 1p/19q loss, while IDH1 wildtype tumors rarely contained these changes (Ichimura et al., 2009). Groups of grade II diffuse astrocytomas, grade III anaplastic astrocytomas, primary glioblastomas, and secondary glioblastomas analyzed for both IDH1 and TP53 in the same tumors often have a higher prevalence of TP53 mutations among the IDH1-mutated group than the IDH1–wild type group (Parsons et al., 2008; Ichimura et al., 2009; Watanabe et al., 2009a; Yan et al., 2009). IDH1 R132 mutation is significantly associated with the presence of any alteration in the TP53 pathway, including alterations in MDM2, MDM4, and p14ARF, in the combined group of grades II and III astrocytomas (Ichimura et al., 2009). Interestingly, 5 astrocytoma patients with Li-Fraumeni syndrome, which is defined by germline TP53 mutation and predisposes to astrocytomas and other cancers, had somatic IDH1 mutations (Watanabe et al., 2009b). Together, these findings illustrate that IDH mutations generally co-occur with TP53 mutations in the astrocytic tumors, and suggest that both

In gliomas, IDH mutations strongly associate with TP53 mutation or codeletion of chromosomes 1p and 19q, but are inversely associated with other genetic changes common in glioma, such as EGFR amplification. TP53 mutations are commonly present in grades II and III astrocytomas and oligoastrocytomas as well as secondary glioblastomas, but only about 30% of primary glioblastomas contain TP53 mutations; they are rare in pure oligodendroglial tumors. Pure oligodendroglial tumors frequently possess total 1p/19q loss and rarely contain TP53 mutations. 1p/19q loss is rare, however, in the astrocytic tumors. In grades II and III oligoastrocytomas, either TP53 mutation or 1p/19q loss are often present, and these changes are inversely associated. Thus, TP53 mutation and 1p/19q loss are generally mutually exclusive and distributed among the grade II and III astrocytic and oligodendroglial tumors, respectively. IDH1 and IDH2 mutations frequently co-occur with either TP53 mutations or 1p/19q loss. In one large

56

are crucial steps in transformation of normal cells into astrocytoma cells. Oligodendroglial tumors frequently contain 1p/19q codeletions. IDH mutations often co-occur with 1p/ 19q loss in oligodendrogliomas, with one study finding 1p/19q loss in 85% of IDH1 or IDH2-mutated oligodendrocytic tumors, but not in oligodendrocytic tumors with wild-type IDH genes (Yan et al., 2009). In another study, all oligodendroglial tumors with 1p/19q loss harbored mutations in either IDH1 or IDH2, while not all tumors with IDH mutations harbored 1p/19q loss in this group (Ichimura et al., 2009). Overall, IDH mutations seem to be a common feature of astrocytomas and oligodendrogliomas, with oligodendrogliomas defined by 1p/19q loss and astrocytomas defined by TP53 mutations. IDH mutations are inversely associated with many of the hallmark changes of primary glioblastomas, including EGFR amplification, CDKN2A or CDKN2B deletion, and PTEN mutations. These changes are uncommon in secondary glioblastomas and lowergrade tumors. IDH mutations and EGFR amplification are inversely correlated among grades II–IV gliomas (Sanson et al., 2009), grades II–IV astrocytic tumors (Ichimura et al., 2009), grade III–IV astrocytic tumors (Yan et al., 2009), or glioblastomas (Parsons et al., 2008; Nobusawa et al., 2009). Similarly, changes in PTEN and CDKN2A/CDKN2B are common among IDH wild-type, but not IDH-mutated, grade III astrocytomas and glioblastomas (Yan et al., 2009). IDH1 mutations have been shown to be negatively associated with PTEN mutations as well (Ichimura et al., 2009). Secondary glioblastomas with wild-type IDH1 have genetic features, such as frequent EGFR amplification, that are characteristic for primary glioblastomas. This suggests that the IDH1 wild-type secondary glioblastomas are actually primary glioblastomas for which the glioblastoma was initially mistaken for a lowergrade lesion (Nobusawa et al., 2009). On the other hand, primary glioblastomas with mutated IDH1 more frequently have alterations, such as TP53 mutation, that are more common in secondary glioblastoma, suggesting that these tumors may have progressed from a lower-grade astrocytoma that was not clinically evident (Nobusawa et al., 2009). IDH1 mutations also associate with characteristic transcriptional and epigenetic profiles. Gliomas with IDH1 R132 mutations frequently display a pattern of hypermethylation throughout the genome compared to

Z.J. Reitman and H. Yan

gliomas without IDH1 mutations (Noushmehr et al., 2010). Additionally, gene expression analysis has revealed that gliomas typically cluster into proneural, neural, classical, or mesenchymal gene expression signatures (Phillips et al., 2006), and that IDH1mutated gliomas associate with the proneural signature (Verhaak et al., 2010). The proneural signature resembles the gene expression signature of normal oligodendrocytes, and has previously been found to associate with grade II and III gliomas, and with younger age at diagnosis and better prognosis compared to gliomas with the other signatures (Phillips et al., 2006). Because IDH1-mutated gliomas have genetic, epigenetic, and gene expression alterations that are distinct from gliomas without IDH1 mutations, it stands to reason that pathogenesis and biology of IDH1-mutated gliomas may be fundamentally different from other gliomas. Since IDH2 mutations are rarer in gliomas than IDH1 mutations, it is not clear whether IDH2 mutated tumors have similar genetic, epigenetic, and gene expression profiles to the IDH1-mutated tumors.

Timing of Mutations IDH1 and IDH2 mutations appear to be early events in glioma formation. The proportion of IDH1 and IDH2 mutated tumors does not increase with grade (Balss et al., 2008), suggesting that the mutations arise during tumor formation, rather than during progression to a higher grade. The mutations are frequent in grade II gliomas, indicating that they are important for early steps in tumor formation. In a group of 51 grade II glioma patients with two or more biopsies, 42 (82%) had an IDH1 mutation at the first biopsy, and had an identical mutation at later biopsies. Two of the remaining nine patients initially did not have an IDH1 mutation, but developed an IDH1 mutation by a later biopsy (Watanabe et al., 2009a). This indicates that IDH1 mutation is an early event in glioma pathogenesis that persists throughout tumor progression. IDH mutations also occur before other common genetic alterations in glioma. Of 51 patients with multiple biopsies, four grade II glioma patients with IDH1 mutations at their first biopsy had TP53 mutations as well at their last biopsy. Another three of these patients

7

Malignant Glioma: Isocitrate Dehydrogenases 1 and 2 Mutations

had IDH1 mutation alone at their first biopsy and 1p/19q loss by the time of their last biopsy (Watanabe et al., 2009a). However, no cases of IDH1 or IDH2 mutation occurring after acquisition of a somatic TP53 mutation or 1p/19q loss have been reported. The fact that IDH mutations occur frequently in astrocytomas containing TP53 mutations and in oligodendrogliomas with 1p/19q loss, and that IDH mutations occur before the other genetic alterations, has led to speculation that the mutations occur in a precursor cell that can give rise to both the astrocytic and oligodendroglial tumor types. Other genetic pathways to tumorigenesis that do not involve frequent IDH mutations likely lead to gliomas such as grade I pilocytic astrocytomas and grade IV primary glioblastomas by separate pathways, as evidenced by a different constellation of genetic changes observed for those tumor types. The common genetic changes found in grades I-IV gliomas, and the suggestion that IDH1 and IDH2 mutations arise in cell that can give rise to both astrocytic and oligodendroglial grade II gliomas, have been incorporated in Fig. 7.1.

Patient Age IDH mutations occur rarely in pediatric glioma patients, and occur in the younger of adult glioma patients. Pediatric patients that do have IDH1 or IDH2 mutations are generally older teenagers with grades II–III tumors rather than young children (De Carli et al., 2009). In studies of adult patients with grades II or III astrocytomas, primary glioblastomas, grade III anaplastic oligoastrocytomas, any grade II glioma, any grade III glioma, or any glioblastoma, patients with IDH1 mutations had a younger median age at diagnosis than those with the same tumor type who did not have IDH1 mutations (Balss et al., 2008; Parsons et al., 2008; Ichimura et al., 2009; Yan et al., 2009). For example, Watanabe et al. found that grade II diffuse astrocytoma patients with IDH1 mutations had a median age of 34 compared to 52 for IDH1 wildtype patients (Watanabe et al., 2009a), and Yan et al. found that grade III anaplastic astrocytoma patients with IDH1 or IDH2 mutations had a median age of 34 compared to 53 for the IDH wild-type patients (Balss et al., 2008). However, a significantly younger age for IDH1-mutated patients has not been observed

57

for patients with pure oligodendroglial tumors. The older median age for grade II and III astrocytoma patients without IDH mutations has been speculated to be caused by misclassification of some primary glioblastomas as lower-grade tumors (Balss et al., 2008). Misclassification of primary glioblastomas as grade III anaplastic astrocytomas could also contribute to the lower frequency of IDH1 mutations in grade III anaplastic astrocytomas compared to grade II diffuse astrocytomas that has been observed in multiple studies (Balss et al., 2008; Ichimura et al., 2009; Yan et al., 2009).

Patient Survival Glioma patients with IDH mutations survive longer than patients with wild-type IDH1 and IDH2. In one study, glioblastoma patients with IDH1 mutations survived 3.7 years compared to 1.1 years for IDH1 wildtype glioblastoma patients (Parsons et al., 2008). Other studies have found that patients with IDH-mutated grade III anaplastic astrocytomas and groups of grade II, III, or IV gliomas survive longer than patients with IDH wild-type tumors of the same type (Dubbink et al., 2009; Ichimura et al., 2009; Sanson et al., 2009; Yan et al., 2009). Since IDH-mutated patients are younger than non-mutated patients, and since younger age is associated with longer survival, these univariate analyses do not show whether IDH mutation predicts better survival independently from age. Thus, some of the large differences in survival reported for these tumors may be reflect the fact that IDH mutations are markers for young age, and therefore better prognosis. Despite the association with survival, no association was found between IDH1 status and response to temozolomide in grade II diffuse astocytomas (Dubbink et al., 2009). Multivariate analyses that take into account age and other factors shed the most light on IDH mutations as independent predictors of survival. Results of such analyses have been mixed. In one study, a Cox regression multivariate analysis including age, tumor type, grade, TP53 mutation status, and 1p/19q codeletion status failed to identify IDH1 status as an independent prognostic factor (Ichimura et al., 2009). Other studies, however, have found IDH1 status to be a predictor of good outcome. A Cox multivariate analysis that did not use age as a covariate was

58

performed on a randomized clinical trial of grade III glioma patients found IDH1 status to be more strongly associated with longer progression-free survival than other prognostic factors, such as 1p/19q status and MGMT promoter methylation (Wick et al., 2009). In another large group of patients, multivariate analysis adjusting for grade, age above or below 48, MGMT status, genomic profile, and treatment, confirmed that IDH1 mutation as an independent favorable prognostic marker (Sanson et al., 2009). A multivariate analysis of glioblastoma patients that adjusted for age found a significant association of IDH1 mutation with longer survival (Nobusawa et al., 2009).

Clinical Testing The specificity of IDH mutations for certain tumor types and their association with outcome makes testing for IDH mutation status potentially useful for the diagnosis and prognosis of gliomas. The high frequency of IDH mutations in grades II–III gliomas and secondary glioblastomas makes them highly sensitive and specific for these tumors compared to other CNS tumors. This has been exploited to show that IDH1 status can help distinguish between grade II diffuse astrocytomas and grade I pilocytic astrocytomas, which may be helpful in cases for which scant material is available for histopathological analysis (Korshunov et al., 2009). In another study, the sensitivity and specificity of IDH1 mutations were 73.3% and 96.3%, respectively, for secondary glioblastomas compared to primary glioblastomas (Nobusawa et al., 2009). Given the potential diagnostic utility for determining IDH mutation status, several methods have been developed to determine whether IDH1 or IDH2 are mutated in tumor samples. Polymerase chain reaction (PCR)-based assays have been developed to accurately detect IDH1 and IDH2 mutation in the DNA of tumor cells (Meyer et al., 2009), and these methods can determine IDH status in formalin-fixed, paraffinembedded tissues even in histologically normal tissue that contains little tumor material (Horbinski et al., 2009). Additionally, monoclonal antibodies have been developed to specifically stain tumors containing the IDH1 R132H protein (Capper et al., 2009; Kato et al.,

Z.J. Reitman and H. Yan

2009), which could be put to use in typical diagnostic pathology laboratories.

Properties of Mutated Enzymes The frequency, specificity, and early timing of IDH mutations provide strong evidence for their importance in glioma tumorigenesis, and functional studies have begun to reveal the properties of the mutated enzymes. The IDH1 and IDH2 mutations observed in cancer diminish the normal isocitrate dehydrogenase activity of the enzymes, shown on the left side of Fig. 7.2, as demonstrated by in vitro assays of either recombinant mutant enzymes (Zhao et al., 2009) or lysates of cells overexpressing the mutant enzymes (Yan et al., 2009). Additionally, recombinant IDH1 R132H can bind to wild-type IDH1, and the resulting heterodimers have diminished isocitrate dehydrogenase activity (Zhao et al., 2009). Since most IDH mutations are heterozygous with a wild-type allele, this has led to speculation that the IDH1 mutants function in cancer to dominant negatively inactivate wild-type IDH1. However, the formation of wild-type:mutant heterodimers has not been demonstrated in vivo, and overexpression of IDH1 R132 and IDH2 R172 mutants has not been shown to lower cellular isocitrate dehydrogenase activity in glioma cell lines, as would be expected for a dominant negative enzyme. Additionally, Park and colleagues have demonstrated that IDH1 and IDH2 protect of the cell from noxious insults (Kim et al., 2007; Park et al., 2008). It seems reasonable that one wild-type allele of either gene could continue to provide protective benefits to the cell when the other allele is mutated in cancer, but it seems unlikely that the major function for IDH mutations cancer would be to inactivate these protective enzymes. The specificity of the IDH mutations for “hotspot” codons is reminiscent of oncogenes such as KRAS and PIK3CA, which are also activated by specific point mutations. This resemblance suggests that IDH mutations also confer a gain of function. Consistent with this, IDH1 R132 and IDH2 R172 mutants gain the neomorphic activity to reduce α-ketoglutarate to (R)-2-hydroxyglutarate as shown on the right side of Fig. 7.2 (Dang et al., 2009;

7

Malignant Glioma: Isocitrate Dehydrogenases 1 and 2 Mutations

59

Fig. 7.2 Wild-type IDH1 and IDH2 catalyze the oxidative decarboxylation of isocitrate to form α-ketoglutarate, but IDH1 R132 and IDH2 R172 mutants catalyze the reduction of α-ketoglutarate to (R)-2-hydroxyglutarate. Wild-type IDH1 and IDH2 normally catalyze a two step reaction, starting with isocitrate on the upper left and resulting in α-ketoglutarate on the bottom center. First, they catalyze the NADP+ -dependent oxidation of the α-hydroxyl of isocitrate (blue) to produce an intermediate, oxalosuccinate. Then, they catalyze the decarboxylation of the β-carboxyl (green) of this intermediate to produce α-ketoglutarate and CO2 . This reaction is reversible: under appropriate conditions, the enzymes can catalyze the addition of CO2 to α-ketoglutarate to form oxalosuccinate. Then, they catalyze the NADPH-dependent reduction of the α-ketone to a hydroxyl to form isocitrate. R132 of IDH1 is analogous to R172

of IDH2. In each enzyme, this arginine residue forms hydrogen bonds with the β-carboxyl of isocitrate. This presumably coordinates removal of the β-carboxyl to release CO2 in the forward reaction. Conversely, the arginine coordinates the addition of CO2 to form the β-carboxyl in the reverse reaction. As expected for enzymes with mutations in a residue that aids in substrate binding and catalysis, IDH1 R132 and IDH2 R172 mutants are impaired in their normal ability to bind isocitrate or convert it to α-ketoglutarate. However, the mutated enzymes still bind α-ketoglutarate. Nevertheless, the enzymes do not coordinate the addition of CO2 to α-ketoglutarate followed by reduction of the α-ketone as normally occurs in the reverse reaction. Instead, only reduction of the α-ketone occurs. This results in conversion of α-ketoglutarate and NADPH to (R)-2-hydroxyglutarate and NADP+ as shown on the right

Gross et al., 2010; Ward et al., 2010). IDH1mutated gliomas contain a 100-fold elevated concentration of (R)-2-hydroxyglutarate (also known as R(-)-2-hydroxyglutarate or D-2-hydroxyglutarate) compared to IDH1-wild-type gliomas (Dang et al., 2009). This neomorphic activity apparently leads to the accumulation of (R)-2-hydroxyglutarate observed in the tumors, and also may alter the cellular NADPH/NADP+ ratio and cause flux away from αketoglutarate. One or more of these metabolic changes could confer a selective advantage to glioma cells and promote glioma formation through a yet-unknown mechanism. If the presence of mutated IDH1 or IDH2 is essential for the malignant properties of glioma cells,

it may be possible to design molecular therapies that target these tumor-specific mutants. Further research will be needed to determine the downstream effects of the gain of function of the IDH mutations.

Mutation Types A spectrum of different mutated codons has been observed at IDH1 R132 and IDH2 R172 in glioma. IDH1 R132H is by far the most common IDH mutation in gliomas, accounting for 88.2–92.8% of IDH1-mutated tumors, followed by R132C (3.4–4.6%), R132S (0.8–2.5%), R132G (0.6–3.9%),

60

R132L (0–4.5%), and R132V (one case) (Balss et al., 2008; Hartmann et al., 2009; Ichimura et al., 2009; Watanabe et al., 2009a; Yan et al., 2009). Only a handful of the rarer IDH2 mutations have been reported, with a total of 24 R172K, 9 R172M, 5 R172W, and 2 R172G cases found in two large studies (Hartmann et al., 2009; Yan et al., 2009). All of the observed mutations result from a single base pair change in the R132 or R172 codon, except for the single observed case of IDH1 R132V, which results from alteration of two base pairs (Balss et al., 2008). Different mutations have a range of impact on enzyme function. For instance, IDH1 R132H has a greater loss of isocitrate dehydrogenase activity and more rapidly reduces α-ketoglutarate than IDH1 R132C, R132L, or R132S (Dang et al., 2009). If these changes in activity provide a selective advantage to a glioma or pre-glioma cell, then perhaps the greater potency for the R132H mutation explains its higher observed frequency in gliomas compared to the other mutations. Interestingly, IDH1 R132C is observed more frequently than IDH1 R132H in non-glioma cancers, including acute myelogenous leukemia and case reports of prostate cancer, colorectal cancer, and acute lymphoblastic leukemia, B-type (Sjoblom et al., 2006; Kang et al., 2009; Mardis et al., 2009). Additionally, all five cases of IDH1 mutations reported in glioma patients with Li-Fraumeni syndrome were R132C mutations (Watanabe et al., 2009b). Also, patients with IDH1 R132C and R132S mutations were significantly younger than those with R132H mutations in one large study (Hartmann et al., 2009). Differences have also been noted between patients with IDH1- and IDH2-mutated tumors. One large study noted a trend for IDH2 mutations to occur more frequently in grade III tumors compared to grade II tumors, and that IDH2 mutations were significantly more frequent in oligodendroglial tumors compared to astrocytic tumors (Hartmann et al., 2009). Though IDH1 and IDH2 mutations are inversely associated, several cases of grade III gliomas contained both alterations (Hartmann et al., 2009). Interestingly, acute myeloid leukemias exhibit a strikingly different distribution of IDH1 and IDH2 mutations than gliomas. Acute myeloid leukemias contain IDH2 mutations slightly more frequently than IDH1 mutations (Ward et al., 2010). Furthermore, almost half of IDH1 and IDH2 the mutations in acute myeloid leukemias occur at IDH2 R140 which also apparently

Z.J. Reitman and H. Yan

confers (R)-2-hydroxyglutarate production to the enzyme (Ward et al., 2010). Of note, this site that has never been found to be mutated in gliomas (Yan et al., 2009). The biological explanation for the distribution of the different IDH mutations in different cancers remains to be elucidated.

Conclusion IDH1 and IDH2 mutations occur frequently in grades II and III gliomas and secondary glioblastomas. These changes associate with other genetic alterations such as TP53 mutation and 1p/19q loss, as well as a younger patient age and longer patient survival among most tumor types. PCR and immunohistochemical techniques can determine mutation status in surgical tumor samples and hold promise as powerful diagnostic and prognostic tests to aid in the clinical management of this difficult set of diseases. The mutations likely occur early in tumorigenesis and their genetic profile suggests that they act as oncogenes. The mutated enzymes have greatly reduced normal enzymatic activity but gains the novel, possibly oncogenic, ability to produce (R)-2-hydroxyglutarate. The frequency and specificity of the IDH1 and IDH2 mutations in glioma suggests that these alterations are central to glioma formation and/or maintenance, but further study is needed to understand the mechanism by which they interact with glioma biology.

References Balss J, Meyer J, Mueller W, Korshunov A, Hartmann C, von Deimling A (2008) Analysis of the IDH1 codon 132 mutation in brain tumors. Acta Neuropathol 116:597–602 Bleeker FE, Lamba S, Leenstra S, Troost D, Hulsebos T, Vandertop WP, Frattini M, Molinari F, Knowles M, Cerrato A, Rodolfo M, Scarpa A, Felicioni L, Buttitta F, Malatesta S, Marchetti A, Bardelli A (2009) IDH1 mutations at residue p.R132 (IDH1(R132)) occur frequently in high-grade gliomas but not in other solid tumors. Hum Mutat 30:7–11 Capper D, Zentgraf H, Balss J, Hartmann C, von Deimling A (2009) Monoclonal antibody specific for IDH1 R132H mutation. Acta Neuropathol 118:599–601 Dang L, White DW, Gross S, Bennett BD, Bittinger MA, Driggers EM, Fantin VR, Jang HG, Jin S, Keenan MC, Marks KM, Prins RM, Ward PS, Yen KE, Liau LM, Rabinowitz JD, Cantley LC, Thompson CB, Vander Heiden

7

Malignant Glioma: Isocitrate Dehydrogenases 1 and 2 Mutations

MG, Su SM (2009) Cancer-associated IDH1 mutations produce 2-hydroxyglutarate. Nature 462:739–744 De Carli E, Wang X, Puget S (2009) IDH1 and IDH2 mutations in gliomas. N Engl J Med 360:2248 author reply 2249 Dubbink HJ, Taal W, van Marion R, Kros JM, van Heuvel I, Bromberg JE, Zonnenberg BA, Zonnenberg CB, Postma TJ, Gijtenbeek JM, Boogerd W, Groenendijk FH, Smitt PA, Dinjens WN, van den Bent MJ (2009) IDH1 mutations in low-grade astrocytomas predict survival but not response to temozolomide. Neurology 73:1792–1795 Gross S, Cairns RA, Minden MD, Driggers EM, Bittinger MA, Jang HG, Sasaki M, Jin S, Schenkein DP, Su SM, Dang L, Fantin VR, Mak TW (2010) Cancer-associated metabolite 2-hydroxyglutarate accumulates in acute myelogenous leukemia with isocitrate dehydrogenase 1 and 2 mutations. J Exp Med 207:339–344 Hartmann C, Meyer J, Balss J, Capper D, Mueller W, Christians A, Felsberg J, Wolter M, Mawrin C, Wick W, Weller M, Herold-Mende C, Unterberg A, Jeuken JW, Wesseling P, Reifenberger G, von Deimling A (2009) Type and frequency of IDH1 and IDH2 mutations are related to astrocytic and oligodendroglial differentiation and age: a study of 1,010 diffuse gliomas. Acta Neuropathol 118:469–474 Horbinski C, Kofler J, Kelly LM, Murdoch GH, Nikiforova MN (2009) Diagnostic use of IDH1/2 mutation analysis in routine clinical testing of formalin-fixed, paraffin-embedded glioma tissues. J Neuropathol Exp Neurol 68:1319–1325 Ichimura K, Pearson DM, Kocialkowski S, Backlund LM, Chan R, Jones DT, Collins VP (2009) IDH1 mutations are present in the majority of common adult gliomas but are rare in primary glioblastomas. Neuro Oncol 11:341–347 Kang MR, Kim MS, Oh JE, Kim YR, Song SY, Seo SI, Lee JY, Yoo NJ, Lee SH (2009) Mutational analysis of IDH1 codon 132 in glioblastomas and other common cancers. Int J Cancer 125:353–355 Kato Y, Jin G, Kuan CT, McLendon RE, Yan H, Bigner DD (2009) A monoclonal antibody IMab-1 specifically recognizes IDH1R132H, the most common glioma-derived mutation. Biochem Biophys Res Commun 390:547–551 Kim SY, Lee SM, Tak JK, Choi KS, Kwon TK, Park JW (2007) Regulation of singlet oxygen-induced apoptosis by cytosolic NADP+-dependent isocitrate dehydrogenase. Mol Cell Biochem 302:27–34 Korshunov A, Meyer J, Capper D, Christians A, Remke M, Witt H, Pfister S, von Deimling A, Hartmann C (2009) Combined molecular analysis of BRAF and IDH1 distinguishes pilocytic astrocytoma from diffuse astrocytoma. Acta Neuropathol 118:401–405 Louis DN, Ohgaki H, Wiestler OD, Cavenee WK, Burger PC, Jouvet A, Scheithauer BW, Kleihues P (2007) The 2007 WHO classification of tumours of the central nervous system. Acta Neuropathol 114:97–109 Mardis ER, Ding L, Dooling DJ, Larson DE, McLellan MD, Chen K, Koboldt DC, Fulton RS, Delehaunty KD, McGrath SD et al. (2009) Recurring mutations found by sequencing an acute myeloid leukemia genome. N Engl J Med 361:1058–1066 Meyer J, Pusch S, Balss J, Capper D, Mueller W, Christians A, Hartmann C, von Deimling A (2010) PCR- and Restriction Endonuclease-Based Detection of IDH1 Mutations. Brain Pathol 20:298–300

61

Nobusawa S, Watanabe T, Kleihues P, Ohgaki H (2009) IDH1 mutations as molecular signature and predictive factor of secondary glioblastomas. Clin Cancer Res 15:6002–6007 Noushmehr H, Weisenberger DJ, Diefes K, Phillips HS, Pujara K, Berman BP, Pan F, Pelloski CE, Sulman EP, Bhat KP, Verhaak RG, Hoadley KA, Hayes DN, Perou CM, Schmidt HK, Ding L, Wilson RK, Van Den Berg D, Shen H, Bengtsson H, Neuvial P, Cope LM, Buckley J, Herman JG, Baylin SB, Laird PW, Aldape K (2010) Identification of a CpG island methylator phenotype that defines a distinct subgroup of glioma. Cancer Cell 17:510–522 Park SY, Lee SM, Shin SW, Park JW (2008) Inactivation of mitochondrial NADP+-dependent isocitrate dehydrogenase by hypochlorous acid. Free Radic Res 42:467–473 Parsons DW, Jones S, Zhang X, Lin JC, Leary RJ, Angenendt P, Mankoo P, Carter H, Siu IM, Gallia GL et al. (2008) An integrated genomic analysis of human glioblastoma multiforme. Science 321:1807–1812 Phillips HS, Kharbanda S, Chen R, Forrest WF, Soriano RH, Wu TD, Misra A, Nigro JM, Colman H, Soroceanu L, Williams PM, Modrusan Z, Feuerstein BG, Aldape K (2006) Molecular subclasses of high-grade glioma predict prognosis, delineate a pattern of disease progression, and resemble stages in neurogenesis. Cancer Cell 9:157–173 Reitman ZJ, Olby NJ, Mariani CL, Thomas R, Breen M, Bigner DD, McLendon RE, Yan H (2010) IDH1 and IDH2 hotspot mutations are not found in canine glioma. Int J Cancer 127(1):245–246 Reitman ZJ, Yan H (2010) Isocitrate dehydrogenase 1 and 2 mutations in cancer: alterations at a crossroads of cellular metabolism. J Natl Cancer Inst 102:932–941 Sanson M, Marie Y, Paris S, Idbaih A, Laffaire J, Ducray F, El Hallani S, Boisselier B, Mokhtari K, Hoang-Xuan K, Delattre JY (2009) Isocitrate dehydrogenase 1 codon 132 mutation is an important prognostic biomarker in gliomas. J Clin Oncol 27:4150–4154 Sjoblom T, Jones S, Wood LD, Parsons DW, Lin J, Barber TD, Mandelker D, Leary RJ, Ptak J, Silliman N, Szabo S, Buckhaults P, Farrell C, Meeh P, Markowitz SD, Willis J, Dawson D, Willson JK, Gazdar AF, Hartigan J, Wu L, Liu C, Parmigiani G, Park BH, Bachman KE, Papadopoulos N, Vogelstein B, Kinzler KW, Velculescu VE (2006) The consensus coding sequences of human breast and colorectal cancers. Science 314:268–274 Sonoda Y, Kumabe T, Nakamura T, Saito R, Kanamori M, Yamashita Y, Suzuki H, Tominaga T (2009) Analysis of IDH1 and IDH2 mutations in Japanese glioma patients. Cancer Sci 100:1996–1998 Verhaak RG, Hoadley KA, Purdom E, Wang V, Qi Y, Wilkerson MD, Miller CR, Ding L, Golub T, Mesirov JP et al (2010) Integrated genomic analysis identifies clinically relevant subtypes of glioblastoma characterized by abnormalities in PDGFRA, IDH1, EGFR, and NF1. Cancer Cell 17: 98–110 Ward PS, Patel J, Wise DR, Abdel-Wahab O, Bennett BD, Coller HA, Cross JR, Fantin VR, Hedvat CV, Perl AE, Rabinowitz JD, Carroll M, Su SM, Sharp KA, Levine RL, Thompson CB (2010) The common feature of leukemia-associated IDH1 and IDH2 mutations is a neomorphic enzyme activity converting alpha-ketoglutarate to 2-hydroxyglutarate. Cancer Cell 17:225–234

62 Watanabe T, Nobusawa S, Kleihues P, Ohgaki H (2009a) IDH1 mutations are early events in the development of astrocytomas and oligodendrogliomas. Am J Pathol 174:1149–1153 Watanabe T, Vital A, Nobusawa S, Kleihues P, Ohgaki H (2009b) Selective acquisition of IDH1 R132C mutations in astrocytomas associated with Li-Fraumeni syndrome. Acta Neuropathol 117:653–656 Wick W, Hartmann C, Engel C, Stoffels M, Felsberg J, Stockhammer F, Sabel MC, Koeppen S, Ketter R, Meyermann R, Rapp M, Meisner C, Kortmann RD, Pietsch T, Wiestler OD, Ernemann U, Bamberg M, Reifenberger G, von Deimling A, Weller M (2009) NOA-04 randomized

Z.J. Reitman and H. Yan Phase III trial of sequential radiochemotherapy of anaplastic glioma with procarbazine, lomustine, and vincristine or temozolomide. J Clin Oncol 27:5874–5880 Yan H, Parsons DW, Jin G, McLendon R, Rasheed BA, Yuan W, Kos I, Batinic-Haberle I, Jones S, Riggins GJ, Friedman H, Friedman A, Reardon D, Herndon J, Kinzler KW, Velculescu VE, Vogelstein B, Bigner DD (2009) IDH1 and IDH2 mutations in gliomas. N Engl J Med 360:765–773 Zhao S, Lin Y, Xu W, Jiang W, Zha Z, Wang P, Yu W, Li Z, Gong L, Peng Y, Ding J, Lei Q, Guan KL, Xiong Y (2009) Gliomaderived mutations in IDH1 dominantly inhibit IDH1 catalytic activity and induce HIF-1alpha. Science 324:261–265

Chapter 8

Metabolic Differences in Different Regions of Glioma Samples Francisca M. Santandreu, Jordi Oliver, and Pilar Roca

Abstract Gliomas are a heterogeneous disease both under a clinical and a pathological point of view. The differential expression of genotypic and metabolic alterations presents a regional distribution within the tumor mass. Even the coexistence of different subpopulations of cancer cells, differing in their sensitivity to apoptosis, autophagy and chemotherapy, has been proved in gliomas. Metabolic alterations are important since they confer adaptive, proliferative, and survival advantages on glioma cells. Thereby, metabolic reprogramming provides substrates for biosynthetic pathway, induces apoptosis resistance. The heterogeneous distribution of oxygen within the tumor determines metabolic differences between regions. Glioma cells located at the periphery of the tumor have higher proliferative capacity, which is accompanied by a greater respiration and mitochondrial oxidative capacity compared to cells located at the center of the tumor; such differences could be related to differential degree of hypoxia or oxidative stress between regions of the tumor. Regional differences within glioma have diagnostic and therapeutic implications, and hinder the prediction of the tumor’s biologic aggressiveness and the patient’s response to standard treatment. Keywords Metabolism · Cancer · Glioma Antioxidant system · Oncogenes · Hypoxia

P. Roca () Department de Biologia Fonamental i Ciències de la Salut. Ed. Guillem Colom, Universitat de les Illes Balears, Carretera Valldemossa Km 7.5, Palma de Mallorca, 07122, Balearic Islands, Spain e-mail: [email protected]

·

Introduction Gliomas are primary tumors of the central nervous system. There are several classifications for gliomas according to cellular type, grade, and location. The World Health Organization subdivides astrocytomas in four grades (I–IV), according to the increasing malignancy determined by pathological evaluation of the tumor. Low-grade gliomas are localized tumors, welldifferentiated and portend a better prognosis for the patient. High-grade gliomas are anaplastic, and have a high growth and invasive capacity; grade IV astrocytomas or glioblastomas constitute the most common and agressive forms. Glioblastoma patients have a median survival expectancy of only 14 months on the current standard treatment of surgical resection to the extent feasible, followed by adjuvant radiotherapy plus temozolomide, given concomitantly with and after radiotherapy. Histological grade, tumor type, age, Karnofsky performance status, tumor location and the magnitude of surgical resection are prognostic factors for gliomas (Hentschel and Sawaya, 2003; Lamborn et al., 2004). Treatment for brain gliomas depends on the location, the cellular type and the grade of malignancy, and often, is a combined approach, using surgery, radiation therapy, and chemotherapy. Highgrade gliomas almost always grow back even after complete surgical excision due to their high tendency to infiltrate (Laerum et al., 1984; Kaba and Kyritsis, 1997). The recurrence of tumor growth, which is the major cause of mortality for patients with gliomas, has been recently associated to the existence of a heterogeneous population of cancer cells within the tumor. From a clinical point of view, gliomas are heterogeneous tumors. The heterogeneity also exists inside

M.A. Hayat (ed.), Tumors of the Central Nervous System, Volume 2, DOI 10.1007/978-94-007-0618-7_8, © Springer Science+Business Media B.V. 2011

63

64

the tumor mass since gliomas express metabolic and histological differences between individual regions. Metabolic alterations in gliomas have a great importance because they promote tumor’s biologic aggressiveness, stimulating cellular proliferation and invasion. In this sense, a better knowledge of the heterogeneous nature of gliomas and their metabolism will facilitate the development of new diagnostic and therapeutic strategies to fight against the disease, overcoming the important limitations of standard therapy. This chapter will describe the heterogeneous nature of glioma, pointing out the presence of regional differences. For their implications on tumor growth and maintenance, we will emphasize the main features of metabolic reprogramming in tumor cells and, in particular, we will describe some metabolic differences between individual regions within glioma tissue.

Human Gliomas: Tumors of Heterogeneous Nature Gliomas are a heterogeneous disease both under a clinical and a pathological point of view. Intratumoral heterogeneity could be defined as a genotypic and/or phenotypic variability within glioma, and it has been recognized in karyotypic and histological analyses, by a variable expression of proliferative markers, growth factors and their receptors, and intrinsic resistance to chemotherapy and radiation therapy (Coons and Johnson, 1993). The differential expression of these cellular features can be diffuse or present a regional distribution. Regional heterogeneity implies that individual cells tend to be similar to their immediate neighbors, whereas more distant cells exhibit different features. Histological analyses indicate that malignant gliomas contain tumor cells with different degrees of dedifferentiation, endothelial cells with variable morphology and blood vessels that contain leukocytes and erythrocytes. Regional differences in proliferative activity, histological features and genotypic alterations have been identified in human gliomas. However, intratumoral heterogeneity in ploidy is not arbitrary, Coons and Johnson described the clustering of regions with similar percentages of aneuploid cells, hypothesizing that glioma progression might be the result of local mutations and subsequent clonal

F.M. Santandreu et al.

expansion. Additionally, the degree of heterogeneity in proliferative markers as well as genotypic heterogeneity increases with the tumor histological grade; most importantly, the higher the intratumoral heterogeneity in these markers of malignancy, the higher the adverse impact in the usefulness of their analysis to predict survival of glioma patients (Coons and Johnson, 1993). Other studies using cell colonies and human paraffin-embedded sections suggest a heterogeneous spatial distribution of cancer cells inside the tumor mass, where the most proliferative cells are located at the periphery of the tumor (Bru et al., 2003). Similarly, magnetic resonance imaging studies evidence a structural heterogeneity within the tumor, where different areas can be distingished from the periphery, which is the most proliferative area of the tumor, to the center. This structural heterogeneity is also correlated with heterogeneity in vascularization (Le Duc et al., 1999). At an additional level of complexity, we must consider that different subpopulations of cancer cells coexist within the tumor mass. The presence of more than one subclone in the primary tumor has been related with tumor recurrence and intratumoral genotypic heterogeneity in gliomas. In relation to this, Ito and colleagues analyzed one case of glioblastoma with oligodendroglial components and determined using a primary cell culture the possible relationship between genetic modifications and chemosensitivity (Ito et al., 2007). While the primary tumor exhibited 1p/19q/10q losses, the recurrent glioblastoma only showed a loss of heterozygosity (LOH) on chromosome 10q. In that study, the chemosensitivity was related to cells with 1p/19q LOH, whereas cells exhibiting 10q LOH were more resistant to chemotherapeutic agents. Shapiro and Shapiro suggested that the karyotypic alterations are associated to cancer cell phenotype. Specifically that hyperdiploid cells, which are unstable in culture, tend to have short doubling times and are usually sensitive to anticancer drugs such as 1,3-bis(2-chloroethyl)1-nitrosourea (BCNU); whereas, cancer cells that are nearly diploid have a normal appearance, are stable in culture, grow more slowly and are more likely to be resistant to nitrosoureas; therefore, it was suggested that the latter population could include stem tumor cells which could be involved in the repopulation of tumor mass (Shapiro and Shapiro, 1985). More recently, the existence of a small fraction of cells, exhibiting features of primitive neural progenitor

8 Metabolic Differences in Different Regions of Glioma Samples

cells, with capacity to induce tumor formation, has been proved in gliomas. These quiescent cells with self-renewal capacity are known as cancer stem cells, which constitute a reservoir of self-sustaining cells with the ability to self-renew and to cause the heterogeneous lineages of cancer cells that comprise the tumor. In 2004, Singh and colleagues succesfully isolated cancer stem cells from different types of brain tumors. Stem cell population was found exclusively in the fraction of tumor cells expressing CD133. The injection of only one hundred CD133+ cells into the NOD-SCID mouse brain led to the growth of a tumor with identical histological features as those displayed by the parenteral tumor. Conversely, the fraction of CD133– tumor cells failed to form tumors, even when 1000-fold more CD133– cells were injected into the brains of the mice. It is important to consider that the existence of tumor cell subpopulations could provide an explanation to the high rate of refractory malignant glioma. In addition, conventional therapeutic approaches targeting the overall population of glioma cells may select the more resistant tumor cells owing to their idiosyncratic properties. In this sense, several studies have found that CD133+ population has an increased resistance to apoptosis (type I programmed cell death), and autophagy (type II programmed cell death) induced by the treatment with chemotherapeutic agents such as temozolomide (Fu et al., 2009). Additionally, gene expression studies in gliomas reveal that CD133+ tumor cells have an elevated expression of multi-drug resistance gene BCRP1, DNA repair genes such as O6-methylguanine-DNA methyltransferase (MGMT) and genes that inhibit apoptosis (Liu et al., 2006; Kang and Kang, 2007).

Interrelationship Between Energy Metabolism and Hallmarks of Cancer Tumors follow the same basic metabolic pathways as normal tissue, but changes in the tumor’s microenvironment and the tumor cells themselves can have important effects on their metabolism. There are two interrelated factors that crucially influence tumor metabolism: oxygen concentration and altered metabolic pathways/enzymes. Hypoxia, a common condition in tumors, determines the way in which cellular energy is produced since the metabolism of

65

most energy substrates such as fatty acids, ketone bodies, amino acids (especially glutamine) and lactate is oxygen dependent (Bouzier et al., 1998). However, glucose can be used for energy production under both normoxic and hypoxic conditions in tumor cells. The presence of oxygen in gliomas, unlike in normal brain tissue, does not inhibit glycolysis (Warburg effect). Additionally, in the presence of oxygen, glioma cells preferentially metabolize the excess of glucose by glycolysis, which provokes a decrease in tumor respiration (Crabtree effect). In tumor cells, the metabolic alterations such as enhanced glycolysis, inhibited tricarboxylic acid (TCA) cycle and enhanced lactate production would result in a net loss of carbon that could have been used for anabolic reactions; however, cancer cells can avoid this by means of a much higher net consumption of glucose than normal cells.

Which Advantages Confer Metabolic Alterations to Tumor Cells? Metabolic alterations promote tumor growth for several reasons (Hsu and Sabatini, 2008; Kroemer and Pouyssegur, 2008). Firstly, conditions of fluctuating oxygen tension, due to inconstant hemodynamics of distant blood vessels, would be lethal for cells that rely on oxidative phosphorylation (OXPHOS) to generate ATP. Secondly, lactate produced as end product of aerobic glycolysis causes acidification of the extracellular milieu, supporting tumor invasion and the suppression of anticancer immune effectors. Thirdly, a major advantage to the cancer cell is that intermediates from the glycolytic pathway can be redirected toward anabolic reactions linked to cellular growth and proliferation. Figure 8.1 summarizes how glycolytic intermadiates are used for the synthesis of macromolecules in the cancer cell. For example, cancer cells enhance their biosynthetic capabilities by expressing a tumor-specific form of pyruvate kinase, M2-PK, which is less active in the conversion of phosphoenolpyruvate (PEP) to pyruvate and thus less efficient at ATP production but facilitates the incorporation of glucose carbons into lipids, amino acids and nucleic acids. In this sense, glucose 6-phosphate becomes available for the synthesis of glycogen and ribose 5-phosphate and dihydroxyacetone phosphate for the synthesis of triacylglycerides and phospholipids. Fourthly, tumor

66

Fig. 8.1 Metabolic reprogramming constitutes a major advantage to cancer cells since intermediates from the glycolytic pathway can be redirected toward anabolic reactions linked to cellular growth and proliferation. NADPH, nicotinamide adenine dinucleotide phosphate; DHAP, dihydroxyacetone phosphate; PEP, phosphoenolpyruvate; OXPHOS, oxidative phosphorylation

cells can metabolize glucose through pentose phosphate pathway to generate nicotinamide adenine dinucleotide phosphate (NADPH), which ensures the antioxidant systems of cancer cells and facilitates their resistance to chemotherapeutic agents. In summary, actively dividing cells not only need great amounts of ATP but also macromolecules such as nucleotides, lipids and proteins, synthesis of which is facilitated by metabolic reprogramming. Additionally, recent studies have shown that several steps in lipid synthesis are required for and may even actively promote tumorigenesis. The molecular mechanisms that underlie metabolic reprogramming of cancer cells are complex. Tumor microenvironment favors a specific metabolic profile. Oxygen levels within a tumor fluctuate both temporally and spatially, and almost always are insufficient to satisfy tumor cell growth, leading to hypoxia and the stabilization of the hypoxia inducible factor 1 (HIF-1), which initiates a transcriptional program that provides multiple solutions to low oxygen availability by decreasing the dependence on aerobic respiration. In parallel, HIF-1 stimulates angiogenesis by upregulating several factors among which vascular endothelial growth factor (VEGF) is included. When hypoxic

F.M. Santandreu et al.

stress is detected by HIF-1 protein, cell metabolism is shifted toward glycolysis by the increased expression of inhibitors of mitochondrial metabolism, glucose transporters and glycolytic enzymes such as lactate dehydrogenase A (LDH-A). Lactate dehydrogenase is the enzyme that catalyzes the last step of glycolysis. This tetrameric protein is formed by subunits A and B, encoded by two different genes, which are differentially regulated. The inhibition of LDH-A prevents the Warburg effect and forces cancer cells to use oxidative phosphorylation to oxidize NADH and produce ATP (Fantin et al., 2006). In this way, LDH-A overexpression facilitates anaerobic metabolism of glucose and conversion of pyruvate to lactate. During tumorigenesis, a shift from A/B isoenzyme pattern to the LDH-A pattern has been suggested. Besides it has been described that tumor growth is attenuated while cells depend on respiration to obtain energy, suggesting that aerobic glycolysis might be essential for cancer progression. Metabolic alterations in the tumor are also promoted by the activation of oncogenes and the loss of tumor suppressors. The major oncogenic events (such as constitutive activation of growth factor pathways, constitutive activation of HIF-1, and inactivation of p53) could constitute the common cause of metabolic reprogramming and hallmarks of cancer such as autonomous growth, resistance against apoptosis, limitless replication, and angiogenesis (Kroemer and Pouyssegur, 2008). A diagram of the intimate relationship between tumor metabolism and the main events of cellular transformation is shown in Fig. 8.2. The activation of oncogenes such as Myc and Akt is involved in the “Warburg effect”, which inhibits oxidative phosphorylation, stimulates glycolysis and increases lactate production. On the other hand, the loss of the tumor suppressor protein p53 prevents SCO2 (synthesis of cytochrome c oxidase 2) gene expression, which is required for assembling cytochrome c oxidase, and thus, interferes in the mitochondrial respiratory chain’s function; it also prevents the activation of one of its cellular effectors, TIGAR (TP53-induced glycolysis and apoptosis regulator), which inhibits glycolysis by decreasing the levels of fructose-2,6-bis-phosphate. Regardless of whether the tumor microenvironment or oncogene activation plays a more important role in driving the development of a distinct cancer metabolism, it is likely that the resulting alterations confer adaptive, proliferative, and survival advantages

8 Metabolic Differences in Different Regions of Glioma Samples

67

Fig. 8.2 Tumor altered metabolism is caused by hypoxic stress but also by oncogene activation and tumor suppressor loss. In turn, metabolic reprogramming in cancer cells stimulates

tumor growth, apoptosis resistance, angiogenesis, avoidance of immunosurveillance, tissue invasion and metastasis. OXPHOS, oxidative phosphorylation; HIF-1, hypoxia inducible factor 1

on the cancer cell. Metabolic alterations provide substrates for biosynthetic pathways, induce apoptosis resistance, and stimulate signaling pathways involved in the pro-oncogenic process. Cancer cells are characterized for their ability to support metabolic alterations, chronic hypoxia as well as oxidative stress (Dang and Semenza, 1999). The mitochondrion of tumor cell constitutes an important point of control of cellular processes such as proliferation/apoptosis, oxidative metabolism and production of reactive oxygen species (ROS). The mitochondria of glioma cells exhibit important functional and ultrastructural alterations (Oudard et al., 1997), which could be related with the high dependence of glucose that glioma cells show. Mitochondrial DNA mutations can appear as a result of tumor progression, in turn, some of them could actively contribute to tumor progression due to concomitant stimulation of ROS production, tumor growth and aerobic glycolysis (Zhou et al., 2007). Primary mitochondrial dysfunctions have direct effects on tumor proliferation and metabolism; mutations of the mitochondrial matrix protein fumarate hydratase and the inner mitochondrial membrane protein succinate dehydrogenase could have causal implications in tumor development (Rustin, 2002). Besides, these mutations induce HIF-1 due to the accumulation of TCA cycle intermediates (fumarate or succinate), which competitively inhibit the alphaketoglutarate-dependent HIF-1α prolyl hydroxylase,

the enzyme that usally targets HIF-1α for oxygendependent destruction. Thus, total o partial defects in OXPHOS can lead to increased glycolysis and inherent resistance to apoptosis. Mitochondria are the main cellular source of ROS in tumor cells. Reactive oxygen species are released as a consequence of incomplete reduction of oxygen during respiration, and it is a price that a cell working in the presence of oxygen has to pay in favor of more efficient bioenergetics. Endogenous production of ROS contributes to intracellular signaling transduction (Benhar et al., 2002). It appears that signals linked to proliferation and survival need ROS for efficient transmission to the nucleus. In this way, ROS act as second messengers stimulating cellular proliferation through the activation of redox-sensitive signal transduction pathways; changes in redox state cause structural modifications and, in consequence, modulate the function of cytosolic enzymes and transcription factors that control, among others, the gene expression of proteins that participate in the protection against oxidative stress. Among the proteins and signaling pathways sensitive to redox changes, some signaling cascades play a crucial role in the regulation of gene expression and prevention of apoptosis such as Ras/Raf/MEK/ERK and PTEN/PI3K/Akt/mTOR/NF-kappaB, components of which are mutated or aberrantly expressed in glioblastomas. Oncogenes such as PI3K (phosphatidylinositol 3-kinase) and Akt have direct effects on cellular

68

metabolic deregulation (Kroemer and Pouyssegur, 2008). Protein kinase B or Akt is a downstream effector of insulin signaling that stimulates the uptake and consumption of glucose through the glycolytic pathway. The activation of PI3K causes the transcriptional downregulation of carnitine palmitoyltransferase 1A (CPT1A), an enzyme located in the outer mitochondrial membrane that esterifies long-chain fatty acids to carnitine, thereby initiating the mitochondrial import of fatty acids and channeling them to betaoxidation. Therefore, PI3K activation inhibits betaoxidation and contributes to the increased dependency of cancer cells on glucose.

Metabolic Differences Between Regions of Human Glioma Malignant cells in a solid tumor are known to be exposed to a heterogeneous environment, with tumor regions characterized by a variable gradient of essential metabolites, nutrients, oxygen and growth factors. Hypoxia in tumors results from an imbalance between oxygen delivery and oxygen demand, and it is heterogeneous within the tumor mass. Oxygen consumption within tumors varies and depends largely on the number of proliferative cells, since these cells consume the most oxygen. A heterogeneous distribution of oxygen within the tumor can be determined by multiple factors including: variations in vascular density/orientation; altered erythrocyte flux and function; longitudinal oxygen gradients along vessels and radial perivascular oxygen gradients. Depending on the oxygenation status, different areas are established within a solid tumor. This oxygen gradient can influence the different metabolic behavior between regions. Essentially, lactate concentrations are higher in hypoxic areas in relation to the more oxygenated ones, thus supporting a switch from aerobic to anaerobic metabolism concomitant with hypoxia. Lactate tumor concentrations are important since this metabolite is a prognostic factor for several human cancers: high lactate levels increase the likelihood of metastases and have a negative impact on disease free survival and overall survival (Walenta et al., 2000). Hypoxia is specially relevant in human gliomas since it influences the tumor’s clinical behavior, increasing its biologic aggressiveness and decreasing its response to therapy (Evans et al., 2004).

F.M. Santandreu et al.

Hypoxic cells tend to be quiescent and chemotherapy agents are more effective in actively dividing cells. Hypoxia also decreases the response of tumors to radiotherapy. Concerning the intratumoral metabolic differences, it has been postulated that depending on the oxygenation status inside the tumor, a symbiotic relationship between two cancer cell populations could exist. In 2008, Sonveaux and colleagues described a “metabolic symbiosis” between hypoxic and aerobic cancer cells: cancer cells under hypoxic conditions convert glucose to lactate and extrude it, whereas aerobic cells take up lactate and utilize it for oxidative phosphorylation. In the case of limited glucose availability for the tumor, this symbiosis would facilitate a more efficient use of glucose. At the molecular level, it has been described that a key component of this symbiotic relationship could be the monocarboxylate transporter 1 (MCT1), this isoform is repressed by hypoxia and transports lactate into, rather than out of, cancer cells. Oxygen-dependent expression of MCT1 allow aerobic cancer cells to efficiently take up and, in concert with the oxygen-dependent expression of lactate dehydrogenase B (LDH-B), utilize lactate as an energy substrate, thereby freeing these cells from the need to take up large quantities of glucose. Interestingly, a similar hypothesis was proposed three years before by Griguer and colleagues (2005), on the basis of their observations of glycolytic metabolism in cultured glioma cells, suggesting that lactate produced by anaerobic glycolisis within the hypoxic region of glioma could also be used as an oxidative substrate in another region where oxygen supply was higher. These authors proved that glioma cells from different established cell lines showed a heterogeneous glucose metabolism. In this sense, there are highly glycolytic glioma cells whereas others preferentially use OXPHOS to obtain energy. The use of OXPHOS to obtain energy enhances cellular survival in the presence of low glucose concentrations, and this metabolic phenotype is related to the expression of lactate dehydrogenase A and B isoenzymes. In glioma cells, the ratio of LDH reductive/oxidative activity is about 7 for the A form and 1 for the B form, suggesting that lactate oxidation is more favorable for cells that express LDH-B. Whereas the expression of LDH-B isoform is independent of extracellular glucose concentration in glioma cells, the expression of LDH-A isoform is highly glucose-dependent. Thereby, LDH-A

8 Metabolic Differences in Different Regions of Glioma Samples

expression is lost in the presence of low glucose concentrations. Taking the aforementioned considerations into account, we characterized the possible differences in energy metabolism and ROS metabolism between the central and peripheral regions of human glioma (Santandreu et al., 2008). In our study, “central region” refers to the tumor core, where biopsy sampling is usually performed, whereas “peripheral region” refers to a region closer to surgical resection margins. Histological analyses showed that each region contained similar cellularity and number of tumor cells and, additionally, that the presence of residual nontumoral tissue was minimal and equal in both regions. Mitochondrial abundance between regions was similar, and the biochemical composition of isolated mitochondria was equal in terms of protein content and mitochondrial DNA copy number. Immunohistochemical analysis using Ki-67 proliferation marker revealed that the highest number of proliferative cells was located at the periphery of the glioma mass. In contrast with the apparent mitochondrial structural homogeneity between glioma regions, mitochondria from the periphery of the tumor showed a significantly higher resting respiration (nonphosphorylating conditions) than mitochondria from the center of the tumor. Both the drop in oxygen consumption and cytochrome c oxidase activity reflected a lower mitochondrial oxidative capacity in the center as opposed to the periphery of the tumor. It is important to mention that the decline in mitochondrial oxidative capacity in the center of the tumor was not caused by variations in mitochondrial abundance, in mitochondrial DNA copy number or in mitochondrial DNA lesions between regions. Therefore, metabolic differences between the center and the periphery of the tumor could be better explained by intratumoral variations in the degree of oxidative stress or hypoxia. In our study an important trend toward a greater lipid oxidation was identified in the center of the tumor, which was accompanied by a significant adaptive increase in mitochondrial enzymatic antioxidant systems (manganese superoxide dismutase or MnSOD) and tissular non-enzymatic antioxidant systems (an increase in the levels of reduced glutathione and increased relationship between reduced vs oxidized glutathione). During resting respiration the maximum hydrogen peroxide (H2 O2 ) production capacity tends to be higher in isolated mitochondria from the periphery of

69

the tumor in relation to those isolated from the center of the tumor, and it is consistent with the notion of an increased mitochondrial functionality in the peripheral region of glioma. The capacity to produce ROS does not reflect the intracellular levels of ROS to which cells are exposed. In contrast, the observations of a trend toward a higher oxidative damage, greater respiratory dysfunction and the adaptive increase in antioxidant systems in the center of the tumor, appear to indicate a greater degree of chronic oxidative stress in this region. Considering that cancer cells located at the periphery of the tumor have higher proliferative activity, there is a higher energy demand in this region. In this situation, it is assumed that oxidative stress is low because cells display a “respiratory control” in which respiration is activated by ADP, and this stimulation stops after the conversion of ADP to ATP. In summary, recent studies support the concept of metabolic heterogeneity within glioma. This metabolic heterogeneity appears not to be random but to be ascribed to defined regions in the spatial structure of the tumor. In this sense, differences in mitochondrial oxidative capacity exist between the central and peripheral region of glioma, and are associated with a differential proliferative capacity; such metabolic differences might be caused by a heterogeneous distribution of oxygen within the tumor. Figure 8.3 represents a hypothetic diagram of metabolic differences between individual regions of glioma and the possible factors involved.

Possible Therapeutic and Diagnostic Implications In this chapter, we have tried to illustrate, in a simplified form, the relevance of cancer cell metabolism, not only for anabolic pathways associated with tumor growth but also for the existence of intimate links between the main oncogenes and metabolic reprogramming in tumors. It should be noted that the knowledge of the high glycolytic rate in tumor cells has led to the development of highly sensitive and specific diagnostic tests for cancer using the glucose analog 2-deoxy-2-[18 F]fluoro-D-glucose by positron emission tomography (FDG-PET). The diagnosis of gliomas using this imaging technique is based on the

70

F.M. Santandreu et al.

Fig. 8.3 Possible metabolic differences between the central and peripheral region of human glioma. Mitochondria from the center of the tumor show a decline in respiration compared to those from the periphery of the tumor which it is probably caused by a higher degree of oxidative and hypoxic stress. Cancer cells located at the periphery of the tumor have higher proliferative

activity, which is coupled to a better mitochondrial oxidative capacity and better maintenance of redox homeostasis, compared with cancer cells located at the center of the tumor. In the graphic, the width of the arrow indicates the intensity of the process. OXPHOS, oxidative phosphorylation; ROS, reactive oxygen species; HIF-1, hypoxia inducible factor 1

preferential incorporation and storage of fludeoxyglucose F18 in tumor tissue compared to healthy tissue. Several studies performed with gliomas point out that differences in energy metabolism between normal cells and cancer cells, could be used as a biochemical basis to develop new therapeutic strategies selectively targeted against cancer cells (Seyfried and Mukherjee, 2005; Nebeling et al., 1995). The inhibition of enzymes and pathways that participate in metabolic reprogramming could have a stong effect on tumor growth, not only limiting bioenergetic flux and anabolic reactions in cancer cells but also reverting the neoplastic phenotype by inducing apoptosis or by blocking invasion and angiogenesis. In other words, interventions such as the inhibition of the PI3K/Akt/mTOR pathway (that would inhibit tumor growth), reestablishment of p53 function (that would restore apoptosis and senescence) or inhibition of transcription factor HIF-1 (that would inhibit angiogenesis), would also normalize metabolic functions in gliomas. Preclinical and clinical evaluation of metabolic inhibitors is still in its early stages, with the exception of mTOR antagonists. The mammalian target of rapamycin (mTOR) is a serine/threonine kinase that enhances cellular growth while inhibiting catabolic reactions mediated by autophagy (Faivre et al., 2006). Although the lack of enzymatic inhibitors with an acceptable degree of specificity is one of the main obstacles at this time, the discovery of selective metabolic inhibitors for anabolic and bioenergetic pathways in gliomas will provide, alone or in combination therapy, a completely new arsenal with which to combat central nervous system tumors.

Intratumoral heterogeneity in gliomas is also important because of its relevance in several clinical aspects, notably, histological grading, patient’s therapeutic response and refractory disease. Since variability in specific histological features (cellular density, necrosis, cytologic and nuclear pleomorfism, mitotic activity and microvascular proliferation) produces different histological grades between individual regions within glioma (Coons and Johnson, 1993), it is likely that errors in histological classification induced by regional heterogeneity could be minimized avoiding limited biopsies (localized sample site(s) and small sample size). Recent studies support the idea that malignant gliomas might be considered as “microecosystems” where tumor cells, microenvironment, vasculature and cancer stem cells are all interrelated. Thus, the most malignant cells could be selected under adverse conditions such as the pressure from the immune system, radiotherapy or chemotherapy. The increase in antioxidant defense systems in glioma cells has been associated with a higher resistance to radiotherapy and chemotherapy (Lee et al., 2004). If one accepts that metabolic differences and differences in antioxidant capacity do exist between individual regions of glioma, then it may follow that standard therapy could retard tumor growth in the short term but might facilitate recurrence in the long term by means of a selective pressure in vivo, which would promote the survival of the more resistant cellular population. In most clinical trials, in which monoclonal antibodies or low molecular-weight kinase inhibitors have been used to control the dysregulation in glioma growth,

8 Metabolic Differences in Different Regions of Glioma Samples

monotherapies have failed to show a survival benefit for patients. If we accept the coexistence of more than one subtype of cancer cells within glioma, it is likely that combination therapy might be the more efficient strategy to eliminate the distinct subpopulations contained in the tumor mass. Additionally, functional imaging techniques for diagnosis such as FDG-PET could help to assess in vivo metabolic heterogeneity in human gliomas (Goldman et al., 1997). In all likelyhood, a better understanding of some aspects of tumor cellular biology such as metabolic heterogeneity could clarify the problem of why glioma patients diagnosed with the same histological grade have different evolutions and respond differentially to standard treatment. Acknowledgements We gratefully acknowledge the support from the Conselleria d’Economia, Hisenda i Innovació del Govern de les Illes Balears and the Fondo de Investigaciones Sanitarias del Ministerio de Sanidad y Consumo del Gobierno Español (PI060266).

References Benhar M, Engelberg D, Levitzki A (2002) ROS, stress-activated kinases and stress signaling in cancer. EMBO Rep 3: 420–425 Bouzier AK, Voisin P, Goodwin R, Canioni P, Merle M (1998) Glucose and lactate metabolism in C6 glioma cells: evidence for the preferential utilization of lactate for cell oxidative metabolism. Dev Neurosci 20:331–338 Bru A, Albertos S, Luis Subiza J, Garcia-Asenjo JL, Bru I (2003) The universal dynamics of tumor growth. Biophys J 85:2948–2961 Coons SW, Johnson PC (1993) Regional heterogeneity in the DNA content of human gliomas. Cancer 72:3052–3060 Dang CV, Semenza GL (1999) Oncogenic alterations of metabolism. Trends Biochem Sci 24:68–72 Evans SM, Judy KD, Dunphy I, Jenkins WT, Hwang WT, Nelson PT, Lustig RA, Jenkins K, Magarelli DP, Hahn SM, Collins RA, Grady MS, Koch CJ (2004) Hypoxia is important in the biology and aggression of human glial brain tumors. Clin Cancer Res 10:8177–8184 Faivre S, Kroemer G, Raymond E (2006) Current development of mTOR inhibitors as anticancer agents. Nat Rev Drug Discov 5:671–688 Fantin VR, St-Pierre J, Leder P (2006) Attenuation of LDH-A expression uncovers a link between glycolysis, mitochondrial physiology, and tumor maintenance. Cancer Cell 9:425–434 Fu J, Liu ZG, Liu XM, Chen FR, Shi HL, Pangjesse CS, Ng HK, Chen ZP (2009) Glioblastoma stem cells resistant to temozolomide-induced autophagy. Chin Med J 122: 1255–1259

71

Goldman S, Levivier M, Pirotte B, Brucher JM, Wikler D, Damhaut P, Dethy S, Brotchi J, Hildebrand J (1997) Regional methionine and glucose uptake in high-grade gliomas: a comparative study on PET-guided stereotactic biopsy. J Nucl Med 38:1459–1462 Griguer CE, Oliva CR, Gillespie GY (2005) Glucose metabolism heterogeneity in human and mouse malignant glioma cell lines. J Neurooncol 74:123–133 Hentschel SJ, Sawaya R (2003) Optimizing outcomes with maximal surgical resection of malignant gliomas. Cancer Control 10(2):109–114 Hsu PP, Sabatini DM (2008) Cancer cell metabolism: warburg and beyond. Cell 134:703–707 Ito M, Wakabayashi T, Natsume A, Hatano H, Fujii M, Yoshida J (2007) Genetically heterogeneous glioblastoma recurring with disappearance of 1p/19q losses: case report. Neurosurgery 61:E168–E169 Kaba SE, Kyritsis AP (1997) Recognition and management of gliomas. Drugs 53:235–244 Kang MK, Kang SK (2007) Tumorigenesis of chemotherapeutic drug-resistant cancer stem-like cells in brain glioma. Stem Cells Dev 16:837–847 Kroemer G, Pouyssegur J (2008) Tumor cell metabolism: cancer’s Achilles’ heel. Cancer Cell 13:472–482 Laerum OD, Bjerkvig R, Steinsvag SK, de Ridder L (1984) Invasiveness of primary brain tumors. Cancer Metastasis Rev 3:223–236 Lamborn KR, Chang SM, Prados MD (2004) Prognostic factors for survival of patients with glioblastoma: recursive partitioning analysis. Neuro Oncol 6:227–235 Le Duc G, Peoc’h M, Remy C, Charpy O, Muller RN, Le Bas JF, Decorps M (1999) Use of T(2)-weighted susceptibility contrast MRI for mapping the blood volume in the gliomabearing rat brain. Magn Reson Med 42:754–761 Lee HC, Kim DW, Jung KY, Park IC, Park MJ, Kim MS, Woo SH, Rhee CH, Yoo H, Lee SH, Hong SI (2004) Increased expression of antioxidant enzymes in radioresistant variant from U251 human glioblastoma cell line. Int J Mol Med 13:883–887 Liu G, Yuan X, Zeng Z, Tunici P, Ng H, Abdulkadir IR, Lu L, Irvin D, Black KL, Yu JS (2006) Analysis of gene expression and chemoresistance of CD133+ cancer stem cells in glioblastoma. Mol Cancer 5:67 Nebeling LC, Miraldi F, Shurin SB, Lerner E (1995) Effects of a ketogenic diet on tumor metabolism and nutritional status in pediatric oncology patients: two case reports. J Am Coll Nutr 14:202–208 Oudard S, Boitier E, Miccoli L, Rousset S, Dutrillaux B, Poupon MF (1997) Gliomas are driven by glycolysis: putative roles of hexokinase, oxidative phosphorylation and mitochondrial ultrastructure. Anticancer Res 17:1903–1911 Rustin P (2002) Mitochondria, from cell death to proliferation. Nat Genet 30:352–353 Santandreu FM, Brell M, Gene AH, Guevara R, Oliver J, Couce ME, Roca P (2008) Differences in mitochondrial function and antioxidant systems between regions of human glioma. Cell Physiol Biochem 22:757–768 Seyfried TN, Mukherjee P (2005) Targeting energy metabolism in brain cancer: review and hypothesis. Nutr Metab 2:30 Shapiro JR, Shapiro WR (1985) The subpopulations and isolated cell types of freshly resected high grade human

72 gliomas: their influence on the tumor’s evolution in vivo and behavior and therapy in vitro. Cancer Metastasis Rev 4:107–124 Singh SK, Hawkins C, Clarke ID, Squire JA, Bayani J, Hide T, Henkelman RM, Cusimano MD, Dirks PB (2004) Identification of human brain tumour initiating cells. Nature 432:396–401 Sonveaux P, Vegran F, Schroeder T, Wergin MC, Verrax J, Rabbani ZN, De Saedeleer CJ, Kennedy KM, Diepart C, Jordan BF, Kelley MJ, Gallez B, Wahl ML, Feron O, Dewhirst MW (2008) Targeting lactate-fueled respiration selectively kills hypoxic tumor cells in mice. J Clin Invest 118:3930–3942

F.M. Santandreu et al. Walenta S, Wetterling M, Lehrke M, Schwickert G, Sundfor K, Rofstad EK, Mueller-Klieser W (2000) High lactate levels predict likelihood of metastases, tumor recurrence, and restricted patient survival in human cervical cancers. Cancer Res 60:916–921 Zhou S, Kachhap S, Sun W, Wu G, Chuang A, Poeta L, Grumbine L, Mithani SK, Chatterjee A, Koch W, Westra WH, Maitra A, Glazer C, Carducci M, Sidransky D, McFate T, Verma A, Califano JA (2007) Frequency and phenotypic implications of mitochondrial DNA mutations in human squamous cell cancers of the head and neck. Proc Natl Acad Sci USA 104:7540–7545

Chapter 9

Glioblastoma Patients: Role of Methylated MGMT Giulio Metro and Alessandra Fabi

Abstract The O6 -methylguanine-DNA methyltransferase (MGMT) protein is a DNA repair enzyme that antagonizes the anti-tumor effects of alkylating agents, particularly temozolomide. Consistent with this mechanism of action, MGMT silencing by gene promoter methylation has been shown to be both predictive and prognostic in clinical trials of newly diagnosed glioblastoma patients treated with temozolomide in combination with radiotherapy and as adjuvant treatment. However, assessment of methylation of the MGMT gene promoter still requires standardization and prospective validation in clinical trials in order to best define the role of this biomarker for individualization of treatment. Nevertheless, even patients with temozolomide-sensitive glioblastoma cannot avoid eventual recurrence. Moreover, the optimal treatment strategy for patients with tumors lacking methylation of the MGMT gene promoter has yet to be determined. Here, we discuss the predictive and prognostic value of MGMT silencing, focusing on the importance of standardizing MGMT assessment as a crucial prerequisite for achieving personalization of treatment according to MGMT status in the next future. Keywords Glioblastoma · MGMT · Methylation · Alkylating agents · Silencing · Pseudoprogression

Introduction The O6 -methylguanine-DNA methyltransferase (MGMT) protein is a DNA repair enzyme which is A. Fabi () Division of Medical Oncology, Regina Elena National Cancer Center Institute, 00144 Rome, Italy e-mail: [email protected]

ubiquitously expressed in normal tissues (Gerson, 2004). Its high conservation throughout evolution suggests that MGMT plays a crucial role in maintaining cell physiology and genome integrity (Gerson, 2004). Although the levels of MGMT expression may vary considerably among different organs, tumors usually exhibit higher levels of expression compared with their tissue of origin (Gerson, 2004). Importantly, MGMT has been implicated in resistance to anticancer therapy with alkylating agents, particularly temozolomide. In fact, MGMT acts by antagonizing temozolomide-induced DNA methylation at the O6 position of guanine by transferring the methyl group to the active site of the enzyme itself. If left unrepaired, O6 -methylguanine undergoes incorrect pairing with thymine instead of cytosine during DNA replication, eventually leading to cell death (Gerson, 2004). Consistent with this mechanism of action, low levels/loss of function of MGMT have been shown to be associated with increased sensitivity to temozolomide treatment in malignant gliomas (Chinot et al., 2007; Friedman et al., 1998; Hegi et al., 2004, 2005). Importantly, loss of function of MGMT in glioblastoma multiforme (GBM) can be attributed almost exclusively to hypermethylation of the MGMT gene promoter, a phenomenon known to determine epigenetic silencing of the MGMT corresponding protein (Weller et al., 2010). Although the enzyme activity of MGMT appears to be the most important mechanism underlying resistance to temozolomide and alkylating agents in general, other mechanisms may exist such as disturbances of the mismatch repair system and increased expression of the chromatinassociated gene poly(ADP-ribose) polymerase-1 (PARP-1), which is involved in nucleotide excision repair system (Curtin et al., 2004; Yip et al., 2009).

M.A. Hayat (ed.), Tumors of the Central Nervous System, Volume 2, DOI 10.1007/978-94-007-0618-7_9, © Springer Science+Business Media B.V. 2011

73

74

G. Metro and A. Fabi

Assessing the Status of MGMT in Tumor Tissue

MGMT Methylation in Newly Diagnosed Glioblastoma: Predictive Role

The potential clinical utility of MGMT as a biomarker in malignant gliomas has led to an ongoing debate on how MGMT status should be assessed and which specific procedure is best suited for routine clinical applications. A test aimed at assessing MGMT status needs to be standardized and reproducible in different laboratories, and must have a clinically relevant cut-off point. This, in order to allow patient selection and individualize therapy. Immunohistochemistry (IHC) detects the levels of the MGMT protein in tumor tissue, while the enzymatic activity of MGMT can be measured by high-performance liquid chromatography (HPLC) (Weller et al., 2010). By contrast, epigenetic silencing of the MGMT gene by promoter methylation can be assessed using methylation-specific PCR (MSP) (Weller et al., 2010). The MSP assay can be performed on DNA extracted from paraffin-embedded tissue samples despite the fact that fixation may cause deterioration of the DNA so that the best results are usually obtained on frozen tissue. Currently, the gelbased MSP assay represents the most widely used technology and is the only test that has been shown to be both predictive and prognostic in clinical trials of GBM patients treated with temozolomide (Hegi et al., 2004, 2005; Stupp et al., 2009). Recently, a real-time PCR MSP assay has been developed for quantitative assessment of the MGMT methylation status (qMSP), proving more amenable to definition of technical cut-off and quality control compared with the qualitative gel-based MSP assay (Vlassenbroeck et al., 2008). Real-time PCR MSP assay is being prospectively validated in the phase III randomized CENTRIC trial (Clinicaltrials.gov NTC00689221) which investigates cilengitide, a new integrin inhibitor, in addition to temozolomide plus radiotherapy for patients with newly diagnosed GBM with methylation of the MGMT gene promoter (Stupp et al., 2010). Given the large variability of MGMT methylation (30–60%) reported in the literature for GBM (Weller et al., 2010), prospective validation of cut-offs for optimal MGMT assessment is crucial for future individualization of treatment according to MGMT methylation status.

The first observation on a potential predictive role of MGMT in malignant glioma patients was made more than 10 years ago (Belanich et al., 1996). Patients with low protein levels of MGMT as assessed by immunofluorescence microscopy showed greater benefit from carmustine (BCNU) compared with patients with high MGMT levels (Belanich et al., 1996). Similarly, low protein levels of MGMT as detected by IHC were found to be predictive of prolonged progression-free survival in GBM patients treated with first-line temozolomide (Friedman et al., 1998) as well as of increased survival in inoperable newly diagnosed GBM treated with neoadjuvant temozolomide (Chinot et al., 2007). Importantly, a correlation with survival was also demonstrated when malignant glioma patients harbouring methylation of the MGMT gene promoter were treated with nitrosoureas (Esteller et al., 2000) or temozolomide plus radiotherapy (Hegi et al., 2004), strongly suggesting that MGMT gene silencing through methylation of the MGMT gene promoter as assessed by gel-based MSP assay could provide both a predictive and prognostic biomarker for increased sensitivity to alkylating agent chemotherapy with or without radiotherapy. This was subsequently confirmed in the randomized European Organization for Research and Treatment of Cancer (EORTC) 26981-22981 – National Cancer Institute of Canada (NCIC) trial, where surgically treated patients for newly diagnosed GBM were randomized to radiotherapy or radiotherapy plus concomitant and adjuvant temozolomide (Stupp et al., 2005). In this trial, consistent with the idea that methylation could predict the benefits of alkylating agent chemotherapy, methylation of the MGMT gene promoter was found to be associated with a strikingly longer progressionfree survival in patients treated with temozolomide plus radiotherapy compared with methylated patients treated with radiotherapy alone (10.3 months versus 5.9 months, respectively, P = 0.001) (Hegi et al., 2005). By contrast, in unmethylated patients only a slight trend was observed towards longer progressionfree survival with temozolomide plus radiotherapy versus radiotherapy alone (5.3 months versus 4.4 months,

9

Glioblastoma Patients: Role of Methylated MGMT

respectively, P = 0.02), suggesting little or no benefit from the addition of temozolomide to radiotherapy in the unmethylated group (Hegi et al., 2005). The modest effect of temozolomide in patients lacking methylation of the MGMT gene promoter has raised an ongoing discussion as to whether assessment of the MGMT status should be made mandatory, and whether temozolomide should be avoided in patients with tumors lacking methylation of the MGMT gene promoter. To this regard, it is important to note that the qualitative gel-based MSP assay used in the EORTC-NCIC trial divided patients into methylated and unmethylated (45% versus 55%, respectively) (Stupp et al., 2005). On the other hand, quantitative assays (e.g. qMSP) suggest the presence of a subgroup of patients with intermediate methylation who may still benefit from temozolomide, thus representing the “gray zone” in the test results. For this reason, at the present time MGMT testing should be considered more “informative” rather than “decision-making” in routine clinical practice, and all patients with newly diagnosed GBM should be treated with standard temozolomide plus radiotherapy followed by adjuvant temozolomide irrespective of the methylation status of the MGMT gene promoter.

MGMT Methylation in Newly Diagnosed Glioblastoma: Prognostic Role Besides being predictive of sensitivity to alkylating agent chemotherapy in newly diagnosed GBM, the methylation status of the MGMT gene promoter appears to be also a strong prognostic factor for outcome (Hegi et al., 2004, 2005; Stupp et al., 2009). In the EORTC-NCIC trial methylation of the MGMT gene promoter was found to be significantly associated with longer survival irrespective of treatment compared with absence of methylation (18.2 months versus 12.2 months, respectively, P = 0.001), which corresponds to a risk reduction of 55% (hazard ratio for death = 0.45; 95% confidence interval, 0.32–0.61). In an updated analysis, the 2 and 5 year survival rates in patients with methylated MGMT were 49 and 14%, respectively, while the corresponding figures for

75

methylated patients treated initially with radiotherapy alone were 24 and 5%. On the other hand, patients with an unmethylated MGMT promoter treated with temozolomide plus radiotherapy had a 2 and 5 year survival of 15 and 8%, respectively, compared with 2 and 0% of the unmethylated patients who received radiotherapy alone (Stupp et al., 2009). The small improvement in survival observed in unmethylated with the addition of temozolomide to radiotherapy could be attributed to the somewhat arbritary separation of patients into methylated and unmethylated (the so-called “gray zone”) as well as to other situations such as differences in post-progression therapy and individual variability due to other unknown prognostic factors. Recently, the prognostic relevance of methylation of the MGMT gene promoter has been confirmed in elderly GBM patients treated with concomitant and adjuvant temozolomide (Brandes et al., 2009). In addition, the observation that 74% of patients with GBM who survive longer than 5 years have MGMT promoter methylation, as opposed to less than 50% in an unselected population of GBM patients, underlines again the prognostic value of MGMT methylation (Krex et al., 2007). Nevertheless, long-term survival can be observed even in the absence of methylation of the MGMT gene promoter, indicating that MGMT is only one aspect of a more complex biological system (Krex et al., 2007).

The Informative Role of MGMT Methylation in Newly Diagnosed Glioblastoma Recent evidence suggests that the importance of knowing the methylation status of the MGMT gene promoter may go beyond its predictive and prognostic role. In fact, MGMT status could help predict the occurrence of a phenomenon known as “pseudoprogression” (Brandes et al., 2008). Pseudoprogression occurs in up to 30% of GBM patients treated with radiotherapy or concomitant temozolomide and radiotherapy, and consists of a radiologically-detected increase in lesion (with or without worsening of

76

neurological symptoms) which subsequently subsides or even decreases in size without further treatment (Brandes et al., 2008; Taal et al., 2008). That is because pseudoprogression likely reflects a treatment-related local tissue reaction with inflammation, oedema and increased abnormal vessel permeability rather than true tumor progression. In a series of 103 glioblastoma patients treated with concomitant temozolomide and radiotherapy, lesion enlargement at MRI occurred in 48.5% of patients, of whom 64% were later on considered as having pseudoprogression (31% of the total) (Brandes et al., 2008). Importantly, in this study MGMT methylated tumors were found to be significantly associated with pseudoprogression, as within the group showing tumor enlargement 91% of the methylated patients had pseudoprogression compared with only 41% of the unmethylated GBMs (P = 0.0002) (Brandes et al., 2008). Considering the diagnosis of pseudoprogression and knowing that it occurs more often in MGMT methylated patients is crucial for planning optimal treatment strategies and designing clinical trials with novel agents.

MGMT Methylation for Predicting Clinical Outcome of Recurrent Glioblastoma In contrast to newly diagnosed GBM, the predictive value of MGMT methylation has been questioned in recurrent glioblastomas. Changes in MGMT methylations at recurrence occur in approximately 40% of patients and appear to be more frequent in patients with methylated tumors at first surgery (Brandes et al., 2010). Importantly, the absence of a strong predictive effect in recurrent GBM with the use of various schedules of administration of temozolomide (Brandes et al., 2006; Wick et al., 2007) might suggest the predominance of MGMT-independent mechanisms of resistance in the setting of recurrent disease. Escape from MGMT methylation-mediated sensitivity to the alkylating drug by selection for mismatch repair deficiency might be involved in this “loss” of prediction (Yip et al., 2009). Few studies investigated the relationship between MGMT methylation and sensitivity to drugs other than temozolomide in recurrent glioblastoma. In 19 patients with recurrent malignant gliomas a study showed that single-agent treatment with a nitrosurea drug,

G. Metro and A. Fabi

namely fotemustine, was associated with a considerable rate of disease control in patients with methylated MGMT (66.5%) compared with those with unmethylated MGMT (0%) (Fabi et al., 2009). Despite the low number of patients analyzed and the heterogeneity of patients histotypes included in this study, these data suggest that the presence of MGMT promoter methylation could predict response to fotemustine. However, assessment of MGMT status was performed at the time of initial surgery and no re-assessment of MGMT was conducted at the time of tumor recurrence. This might be important, since changes in the status of MGMT promoter methylation may occur after primary treatment for newly diagnosed glioblastoma (Brandes et al., 2010).

MGMT Methylation in Grade III Anaplastic Gliomas: Only Prognostic? The possible predictive value of methylation of the MGMT gene promoter in GBM patients has recently been challenged in grade III anaplastic gliomas. In fact, in the NOA-04 trial a significantly longer progressionfree survival and overall survival were shown for anaplastic glioma with MGMT methylation irrespective of initial treatment with radiotherapy or alkylatingbased chemotherapy with temozolomide or PCV (procarbazine, lomustine and vincristine) (Wick et al., 2009). Similarly, in the EORTC trial 26951 testing radiotherapy alone versus radiotherapy followed by adjuvant PCV, progression-free and overall survival were found to be significantly longer in patients whose tumors showed methylation of the MGMT gene promoter irrespective of the administration of chemotherapy (van den Bent et al., 2009). These findings suggest that methylation of the MGMT gene promoter is only prognostic but not predictive for chemotherapy in anaplastic gliomas, thus hypothesizing that other markers might be involved in sensitivity to chemotherapy in anaplastic gliomas. Also, these observation underline the importance of studying grade III and IV gliomas as two distinct entities.

MGMT-Depleting Strategies Alternative, more protracted dosing regimens of temozolomide might have a role in increasing sensitivity to

9

Glioblastoma Patients: Role of Methylated MGMT

treatment through optimal depletion of MGMT activity (Weller et al., 2010). Indirect comparisons of treatment efficacy in recurrent disease suggest that the use doseintense schedules of temozolomide might be superior to conventional temozolomide 150–200 mg/m2 administered 5 days every 4 weeks (Brandes et al., 2006; Perry et al., 2010; Wick et al., 2007; Yung et al., 1999). However, whether this presumed superiority is to be attributed to enhanced MGMT depletion is not known. Unfortunately, in a recent trial randomizing chemonaïve patients with recurrent anaplastic astrocytoma or GBM to conventional temozolomide or continuous temozolomide 100 mg/m2 for 21 days every 4 weeks, an inferior outcome was observed for the continuous dose-intense schedule (Brada et al. 2010). At the present time, studies are underway to test dose-intense temozolomide as upfront adjuvant treatment for newly diagnosed GBM following concomitant temozolomide plus radiotherapy. Several MGMT inhibitors are under active investigation with the goal of modulating resistance to alkylating agent chemotherapy. O6-benzylguanine (O6BG) is a potent MGMT-depleting agent that when used in combination with temozolomide has the potential of restoring temozolomide sensitivity in recurrent malignant gliomas (Quinn et al., 2009). However, O6-BG depletes MGMT nonselectively, thus lowering MGMT levels in normal cells as well, therefore resulting in substantial toxicity, particularly myelosuppression. Similarly to O6-BG, Lomeguatrib, O4-benzylfolates and 2-amino-O4-benzylpteridine are other MGMTdepleting agents that are being investigated in combination with temozolomide or BCNU in recurrent disease progressive to alkylating agent chemotherapy.

77

a prospective manner, and it is presently not recommended to use the MGMT promoter methylation assay to determine who should receive temozolomide and who should not. On the other hand, the knowledge so far gained about MGMT status should be exploited to design thoughtful clinical studies aimed at improving the overall outcome of glioblastoma patients, eventually leading to individualization of treatment. For instance, an interesting approach would be that of trying to overcome resistance to alkylating agent chemotherapy in patients with unmethylated MGMT gene promoter. On this basis, a recent trial has closed accrual after enrolling 1153 patients with the aim of showing whether temozolomide dose-intensification (temozolomide for 21 days every 4 weeks) after chemo-radiotherapy would improve outcome of patients with newly diagnosed glioblastoma (Clinicaltrials.gov NCT00304031). In this study, assessment of MGMT status by qMSP assay has been made mandatory for trial inclusion and used as a stratification factor for assigning patients to the control and experimental arms. Alternative strategies may also be useful in unmethylated patients for enhancing the antitumor effect of radiotherapy during the concomitant phase of treatment (Metro et al., 2010). To this regard, given the synergistic activity shown by enzastaurin, a protein kinase C-beta inhibitor, in conjunction with radiotherapy (Tabatai et al., 2007), a phase II trial is evaluating enzastaurin given with radiotherapy and as adjuvant treatment in newly diagnosed glioblastoma patients without methylation of the MGMT gene promoter (Clinicaltrials.gov NCT00509821).

References Conclusions and Future Directions In patients with malignant gliomas several molecular markers other than MGMT may have clinical relevance, including overexpression of the epidermal growth factor receptor, presence of the epidermal growth factor receptor vIII mutation, loss or mutation of PTEN gene. However, MGMT methylation status seems to be the most valuable marker for prediction of outcome and prognosis of patients treated with alkylating agent chemotherapy. Nevertheless, to date no biomarker has been definitively validated in

Belanich M, Pastor M, Randall T, Guerra D, Kibitel J, Alas L, Li B, Citron M, Wasserman P, White A, Eyre H, Jaeckle K, Schulman S, Rector D, Prados M, Coons S, Shapiro W, Yarosh D (1996) Retrospective study of the correlation between the DNA repair protein alkyltransferase and survival of brain tumor patients treated with carmustine. Cancer Res 56:783–788 Brada M, Stenning S, Gabe R, Thompson LC, Levy D, Rampling R, Erridge S, Saran F, Gattamaneni R, Hopkins K, Beall S, Collins VP, Lee SM (2010) Temozolomide versus procarbazine, lomustine, and vincristine in recurrent high-grade glioma. J Clin Oncol 28:4601–4608 Brandes AA, Franceschi E, Tosoni A, Bartolini S, Bacci A, Agati R, Ghimenton C, Turazzi S, Talacchi A, Skrap M, Marucci G, Volpin L, Morandi L, Pizzolitto S, Gardiman M, Andreoli

78 A, Calbucci F, Ermani M (2010) O(6)-methylguanine DNAmethyltransferase methylation status can change between first surgery for newly diagnosed glioblastoma and second surgery for recurrence: clinical implications. Neuro Oncol 12:283–288 Brandes AA, Franceschi E, Tosoni A, Benevento F, Scopece L, Mazzocchi V, Bacci A, Agati R, Calbucci F, Ermani M (2009) Temozolomide concomitant and adjuvant to radiotherapy in elderly patients with glioblastoma: correlation with MGMT promoter methylation status. Cancer 115:3512– 3518 Brandes AA, Franceschi E, Tosoni A, Blatt V, Pession A, Tallini G, Bertorelle R, Bartolini S, Calbucci F, Andreoli A, Frezza G, Leonardi M, Spagnolli F, Ermani M (2008) MGMT promoter methylation status can predict the incidence and outcome of pseudoprogression after concomitant radiochemotherapy in newly diagnosed glioblastoma patients. J Clin Oncol 26:2192–2197 Brandes AA, Tosoni A, Cavallo G, Bertorelle R, Gioia V, Franceschi E, Biscuola M, Blatt V, Crinò L, Ermani M (2006) Temozolomide 3 weeks on and 1 week off as firstline therapy for recurrent glioblastoma: phase II study from gruppo italiano cooperativo di neuro-oncologia (GICNO). Br J Cancer 95:1155–1160 Chinot OL, Barrié M, Fuentes S, Eudes N, Lancelot S, Metellus P, Muracciole X, Braguer D, Ouafik L, Martin PM, Dufour H, Figarella-Branger D (2007) Correlation between O6 -methylguanine-DNA methyltransferase and survival in inoperable newly diagnosed glioblastoma patients treated with neoadjuvant temozolomide. J Clin Oncol 25:1470– 1475 Curtin NJ, Wang LZ, Yiakouvaki A, Kyle S, Arris CA, CananKoch S, Webber SE, Durkacz BW, Calvert HA, Hostomsky Z, Newell DR (2004) Novel poly(ADP-ribose) polymerase-1 inhibitor, AG14361, restores sensitivity to temozolomide in mismatch repair-deficient cells. Clin Cancer Res 10:881–889 Esteller M, Garcia-Foncillas J, Andion E, Goodman SN, Hidalgo OF, Vanaclocha V, Baylin SB, Herman JG (2000) Inactivation of the DNA-repair gene MGMT and the clinical response of gliomas to alkylating agents. N Engl J Med 343:1350–1354 Fabi A, Metro G, Russillo M, Vidiri A, Carapella CM, Maschio M, Cognetti F, Jandolo B, Mirri MA, Sperduti I, Telera S, Carosi M, Pace A (2009) Treatment of recurrent malignant gliomas with fotemustine monotherapy: impact of dose and correlation with MGMT promoter methylation. BMC Cancer 9:101 Friedman HS, McLendon RE, Kerby T, Dugan M, Bigner SH, Henry AJ, Ashley DM, Krischer J, Lovell S, Rasheed K, Marchev F, Seman AJ, Cokgor I, Rich J, Stewart E, Colvin OM, Provenzale JM, Bigner DD, Haglund MM, Friedman AH, Modrich PL (1998) DNA mismatch repair and O6alkylguanine-DNA alkyltransferase analysis and response to temodal in newly diagnosed malignant glioma. J Clin Oncol 16:3851–3857 Gerson SL (2004) MGMT: its role in cancer aetiology and cancer therapeutics. Nat Rev Cancer 4:296–307 Hegi ME, Diserens AC, Godard S, Dietrich PY, Regli L, Ostermann S, Otten P, Van Melle G, de Tribolet N, Stupp R (2004) Clinical trial substantiates the predictive value of O-6-methylguanine-DNA methyltransferase

G. Metro and A. Fabi promoter methylation in glioblastoma patients treated with temozolomide. Clin Cancer Res 10:1871–1874 Hegi ME, Diserens AC, Gorlia T, Hamou MF, de Tribolet N, Weller M, Kros JM, Hainfellner JA, Mason W, Mariani L, Bromberg JE, Hau P, Mirimanoff RO, Cairncross JG, Janzer RC, Stupp R (2005) MGMT gene silencing and benefit from temozolomide in glioblastoma. N Engl J Med 352:997–1003 Krex D, Klink B, Hartmann C, von Deimling A, Pietsch T, Simon M, Sabel M, Steinbach JP, Heese O, Reifenberger G, Weller M, Schackert G (2007) German glioma network. Long-term survival with glioblastoma multiforme. Brain 130:2596–2606 Lee SM, Brada M, Stenning S, Thompson L, Gabe R (2008). for BR12 Collaborators; NCRI Brain Tumour Clinical Studies Group/GB. A randomised trial of procarbazine, CCNU and vincristine (PCV) vs temozolomide (5-day or 21-day schedule) for recurrent high grade glioma (MRC BR12, ISRCTN83176944). 33rd meeting of the European Society for Medical Oncology, Stockholm, Sweden, 12–16 September Metro G, Fabi A, Mirri MA, Vidiri A, Pace A, Carosi M, Russillo M, Maschio M, Giannarelli D, Pellegrini D, Pompili A, Cognetti F, Carapella CM (2010) Phase II study of fixed dose rate gemcitabine as radiosensitizer for newly diagnosed glioblastoma multiforme. Cancer Chemother Pharmacol 65:391–397 Perry JR, Bélanger K, Mason WP, Fulton D, Kavan P, Easaw J, Shields C, Kirby S, Macdonald DR, Eisenstat DD, Thiessen B, Forsyth P, Pouliot JF (2010) Phase II trial of continuous dose-intense temozolomide in recurrent malignant glioma: RESCUE study. J Clin Oncol 28:2051–2057 Quinn JA, Jiang SX, Reardon DA, Desjardins A, Vredenburgh JJ, Rich JN, Gururangan S, Friedman AH, Bigner DD, Sampson JH, McLendon RE, Herndon JE 2nd, Walker A, Friedman HS (2009) Phase II trial of temozolomide plus O6-benzylguanine in adults with recurrent, temozolomideresistant malignant glioma. J Clin Oncol 27:1262–1267 Stupp R, Hegi ME, Mason WP, van den Bent MJ, Taphoorn MJ, Janzer RC, Ludwin SK, Allgeier A, Fisher B, Belanger K, Hau P, Brandes AA, Gijtenbeek J, Marosi C, Vecht CJ, Mokhtari K, Wesseling P, Villa S, Eisenhauer E, Gorlia T, Weller M, Lacombe D, Cairncross JG, Mirimanoff RO (2009) European Organisation for Research and Treatment of Cancer Brain Tumour and Radiation Oncology Groups; National Cancer Institute of Canada Clinical Trials Group. Lancet Oncol 10:459–466 Stupp R, Hegi ME, Neyns B, Goldbrunner R, Schlegel U, Clement PM, Grabenbauer GG, Ochsenbein AF, Simon M, Dietrich PY, Pietsch T, Hicking C, Tonn JC, Diserens AC, Pica A, Hermisson M, Krueger S, Picard M, Weller M (2010) Phase I/IIa study of cilengitide and temozolomide with concomitant radiotherapy followed by cilengitide and temozolomide maintenance therapy in patients with newly diagnosed glioblastoma. J Clin Oncol 27:5881–5886 Stupp R, Mason WP, van den Bent MJ, Weller M, Fisher B, Taphoorn MJ, Belanger K, Brandes AA, Marosi C, Bogdahn U, Curschmann J, Janzer RC, Ludwin SK, Gorlia T, Allgeier A, Lacombe D, Cairncross JG, Eisenhauer E, Mirimanoff RO (2005) European Organisation for Research and Treatment of Cancer Brain Tumor and Radiotherapy Groups; National Cancer Institute of Canada Clinical Trials

9

Glioblastoma Patients: Role of Methylated MGMT

Group. Radiotherapy plus concomitant and adjuvant temozolomide for glioblastoma. N Engl J Med 352:987–996 Taal W, Brandsma D, de Bruin HG, Bromberg JE, SwaakKragten AT, Smitt PA, van Es CA, van den Bent MJ (2008) Incidence of early pseudo-progression in a cohort of malignant glioma patients treated with chemoirradiation with temozolomide. Cancer 113:405–410 Tabatabai G, Frank B, Wick A, Lemke D, von Kürthy G, Obermüller U, Heckl S, Christ G, Weller M, Wick W (2007) Synergistic antiglioma activity of radiotherapy and enzastaurin. Ann Neurol 61:153–161 van den Bent MJ, Dubbink HJ, Sanson M, van der Lee-Haarloo CR, Hegi M, Jeuken JW, Ibdaih A, Brandes AA, Taphoorn MJ, Frenay M, Lacombe D, Gorlia T, Dinjens WN, Kros JM (2009) MGMT promoter methylation is prognostic but not predictive for outcome to adjuvant PCV chemotherapy in anaplastic oligodendroglial tumors: a report from EORTC Brain Tumor Group Study 26951. J Clin Oncol 27:5881–5886 Vlassenbroeck I, Califice S, Diserens AC, Migliavacca E, Straub J, Di Stefano I, Moreau F, Hamou MF, Renard I, Delorenzi M, Flamion B, DiGuiseppi J, Bierau K, Hegi ME (2008) Validation of real-time methylation-specific PCR to determine O6 -methylguanine-DNA methyltransferase gene promoter methylation in glioma. J Mol Diagn 10:332–337 Weller M, Stupp R, Reifenberger G, Brandes AA, van den Bent MJ, Wick W, Hegi ME (2010) MGMT promoter methylation

79 in malignant gliomas: ready for personalized medicine? Nat Rev Neurol 6:39–51 Wick A, Felsberg J, Steinbach JP, Herrlinger U, Platten M, Blaschke B, Meyermann R, Reifenberger G, Weller M, Wick W (2007) Efficacy and tolerability of temozolomide in an alternating weekly regimen in patients with recurrent glioma. J Clin Oncol 25:3357–3361 Wick W, Hartmann C, Engel C, Stoffels M, Felsberg J, Stockhammer F, Sabel MC, Koeppen S, Ketter R, Meyermann R, Rapp M, Meisner C, Kortmann RD, Pietsch T, Wiestler OD, Ernemann U, Bamberg M, Reifenberger G, von Deimling A, Weller M (2009) NOA-04 randomized phase III trial of sequential radiochemotherapy of anaplastic glioma with procarbazine, lomustine, and vincristine or temozolomide. J Clin Oncol 27:5874–5880 Yip S, Miao J, Cahill DP, Iafrate AJ, Aldape K, Nutt CL, Louis DN (2009) MSH6 mutations arise in glioblastomas during temozolomide therapy and mediate temozolomide resistance. Clin Cancer Res 15:4622–4629 Yung WK, Prados MD, Yaya-Tur R, Rosenfeld SS, Brada M, Friedman HS, Albright R, Olson J, Chang SM, O’Neill AM, Friedman AH, Bruner J, Yue N, Dugan M, Zaknoen S, Levin VA (1999) Multicenter phase II trial of temozolomide in patients with anaplastic astrocytoma or anaplastic oligoastrocytoma at first relapse. Temodal Brain Tumor Group. J Clin Oncol 17:2762–2771

Chapter 10

Brain Tumor Angiogenesis and Glioma Grading: Role of Tumor Blood Volume and Permeability Estimates Using Perfusion CT Rajan Jain

Abstract Perfusion imaging of brain tumors has been done using various tracer and non-tracer modalities and can provide additional physiologic and hemodynamic information, which is not available with routine morphologic imaging. Tumor vascular perfusion parameters obtained using CT or MR perfusion have been used for tumor grading, prognosis and treatment response in addition to differentiating treatment/radiation effects from recurrent neoplasms. This chapter is an overview of utility of perfusion CT for assessment of brain tumors, description of the technique, its advantages and limitations. Keywords Perfusion CT · Glioma grading · Angiogenesis · Tumor blood volume · Permeability

Introduction Gliomas, the most common primary brain neoplasms in adults are very heterogeneous tumors. High-grade gliomas can be highly invasive and extremely vascular tumors. Two of the most important factors in determining malignancy of gliomas are their ability to infiltrate the brain parenchyma and to recruit or synthesize vascular networks for further growth i.e., neoangiogenesis (Folkman, 1992; Jain et al., 2002).

R. Jain () Division of Neuroradiology, Department of Radiology and Department of Neurosurgery, Henry Ford Health System, Detroit, MI 48202, USA e-mail: [email protected]

Malignant brain tumors are characterized by neovascularity and increased angiogenic activity with a higher proportion of immature and highly permeable vessels. Glioma grading is currently based on the histological assessment of the tumor, which is achieved by either brain biopsy or cytoreductive surgery; however, there are inherent limitations with these techniques and their interpretation (Law et al., 2008). In vivo perfusion imaging techniques provide additional information regarding tumor physiology and hemodynamics, which may help in better characterizing glioma and may also overcome some of the limitations of histopathologic grading and conventional morphologic imaging. Perfusion imaging has been used to assess tumor grade, prognosis, and recently to assess treatment response, which has caught more attention due to advent of newer therapeutic options including antiangiogenic agents. Traditionally, perfusion imaging of brain tumors has been done with magnetic resonance imaging (MRI), using various perfusion imaging techniques and estimating tumor blood volume, blood flow, and permeability (Roberts et al., 2000; Law et al., 2004, 2008). However, perfusion CT (PCT), which has also been used recently for glioma grading (Ellika et al., 2007; Jain et al., 2008), provides a linear relationship between tissue signal and tissue concentration of a contrast agent unlike perfusion MR and, hence, probably provides a more robust and less biased estimation of physiologic and hemodynamic parameters. In view of the wider availability, faster scan times, and low cost combined with its ease of quantification of various perfusion parameters as compared to MR perfusion, PCT is potentially well suited to study brain tumors and monitoring tumor response to antiangiogenic agents (Ellika et al., 2007; Jain et al., 2008).

M.A. Hayat (ed.), Tumors of the Central Nervous System, Volume 2, DOI 10.1007/978-94-007-0618-7_10, © Springer Science+Business Media B.V. 2011

81

82

In Vivo Perfusion Imaging Versus Histopathology The current standard for tumor grading is histopathologic assessment of tissue, which has inherent limitations such as sampling error, interobserver variation and wide variety of classification systems that are available, the most commonly of which used is the WHO grading system (Law et al., 2008). Most of the gliomas, especially high grade gliomas, show a high degree of regional heterogeneity within the tumor based on tumor cellularity, edema, and necrosis. Many of these factors are inherently dependent on the blood supply. Histopathological evaluation of tumor angiogenesis using various markers such as microvascular density (MVD), microvascular cellular proliferation (MVCP), and total vascular area (TVA) is also limited by this regional heterogeneity, and its confounding effect is worsened by its small size and a limited number of samples obtained with surgical biopsy. This can frequently result in inaccurate classification and grading of gliomas due to sampling errors. Hence, there is a need for noninvasive in vivo clinical imaging tools which can study perfusion in the entire tumor, can be used to assess much larger volumes than small biopsy samples, and probably guide biopsy and excision sites for better results. Brain tumor angiogenesis is a continuously evolving process which can also be affected by various treatment modalities. Hence, in vivo perfusion imaging that can be repeated, unlike invasive procedures such as surgical excision or biopsy, can help assess continued evolution of these tumors as well as treatment response. Another important limitation of histopathological grading system is that gliomas having similar grades respond differently to similar treatment regimens as has been noted with different molecular and genetic markers (Cairncross et al., 1998), suggesting that there is a role for other biomarkers such as perfusion parameters in predicting progression or survival apart from histopathological grading (Law et al., 2008).

Perfusion Parameters and Their Importance Dynamic contrast-enhanced imaging techniques using MRI or CT have been used to obtain measures of

R. Jain

tumor vascular physiology and hemodynamics. After the rapid administration of a contrast agent and the acquisition of serial images at short intervals (seconds), an analysis that uses a pharmacokinetic model of the time dependence of contrast can produce imaging biomarkers such as tumor blood volume, blood flow, vascular permeability, and size of extravascular extracellular space. Many of these parameters have been correlated with tumor grade, aggressiveness, and prognosis (Law et al., 2004, 2006, 2008; Jain et al., 2008).

Tumor Blood Volume Regional tumor blood volume measurements reflect an assessment of tumor vasculature and perfusion, and have been correlated with glioma grading as well as prognosis. Measurement of tumor blood volume is a good surrogate marker for MVD, a measure of angiogenesis and an important prognostic indicator (Leon et al., 1996; Li et al., 1994; Weidner, 1995) in many human cancers. The association between MVD and tumor aggressiveness can be explained by the following: (1) solid tumors are composed of two interdependent components which are the malignant cells and the stroma that they induce, and MVD could be a measure of the success that a tumor has in forming this stromal component; (2) endothelial cells in this stromal component stimulate the growth of tumor cells; thus, the more intratumoral vessels there are, the more endothelial cells and paracrine growth stimulation; (3) intratumoral MVD is a direct measure of the vascular window through which tumor cells pass to spread to distant sites (Weidner, 1995). Tumoral MVD, however, does not distinguish new blood vessels from the native ones, does not mark actively proliferating endothelial cells, and does not correlate with the degree of endothelial cell proliferation. However, these limitations do not seem to diminish the clinical value of this measure. Cha et al. (2003) showed a strong correlation of CBV (cerebral blood volume) measurements in mouse gliomas with MVD and suggested that rCBV (regional CBV) may be elevated due to increase in vessel size or total number of vessels or both. Aronen et al. (2000) also showed a strong correlation of CBV and tumor energy metabolism with MVD using MR perfusion and FDG-PET imaging, respectively.

10 Brain Tumor Angiogenesis and Glioma Grading

Tumor Vascular Permeability It is a fact that tumor blood vessels have defective and leaky endothelium. Hypoxia or hypoglycemia that occurs in rapidly growing tumors increases the expression of vascular endothelial growth factor (VEGF) which is not only a potent angiogenic factor, but also a potent permeability factor (Plate et al., 1992; Shweiki et al., 1992). VEGF leads to the development of neoangiogenic vessels which are immature, tortuous (Jain et al., 2002) and also have increased permeability to macromolecules due to large endothelial cell gaps, incomplete basement membrane, and absence of smooth muscles. These abnormal tumor vessels can be used as potential markers to assess the tumor grade. Thus, in vivo measurement of tumor vessel permeability is important for various reasons: (1) it can be used for grading of tumors because increased permeability is associated with immature blood vessels, which is seen with neoangiogenesis; (2) it can be used to study the response of tumors to various therapies especially antiangiogenic therapy (Bhujwalla et al., 2003; Raatschen et al., 2008). (3) understanding the concept of permeability can help in understanding the mechanism of entry of therapeutic agents into the central nervous system and; (4) development of methods to selectively alter the blood brain barrier (BBB) to enhance drug delivery (Provenzale et al., 2005).

In Vivo Blood Volume and Permeability Quantification: Imaging Techniques, Limitations and Controversies MR (Roberts et al., 2000; Cha et al., 2003; Law et al., 2004, 2008) and CT perfusion (Cenic et al., 2000; Purdie et al., 2001; Ellika et al., 2007; Jain et al., 2008) have been used for in vivo characterization of tumor angiogenesis in various animal and human studies. One of the major limitations of clinically used imaging techniques is low spatial resolution which limits assessment of the complex and heterogenous vascular microenvironment in detail (McDonald and Choyke, 2003; Vajkoczy and Menger, 2000). Another limitation is the current use of low molecular weight contrast agents for human subjects. Because permeability measurements depend on the leakage of particles from the

83

blood to the interstitial space, it is hypothesized that low molecular weight contrast agents leak relatively easily from the blood into the interstitium. This might also lead to overestimation of the permeability which will be influenced by and will approximate tumor blood flow (de Lussanet et al., 2005). High molecular weight contrast agents leak from the vessels and move through the interstitium with relative difficulty and are flow independent. Thus, macromolecular permeability and vascular volumes may be best measured by high molecular weight contrast agents (de Lussanet et al., 2005). However, currently used CT or MRI contrast agents are of low molecular weight (∼0.5–0.9 kDa) which might limit the accurate differentiation of vascular permeability and blood volume (Choyke, 2005). Even though CT and MR contrast agents do not differ much in their molecular weight, different charges of nonionic CT contrast as compared to ionic MR contrast may also be responsible for the differences in permeability measured with these two modalities. Another controversial aspect of measuring permeability with various perfusion imaging techniques is the scanning time. Delayed permeability due to slow leakage of contrast from a leaky blood vessel may not be accurately measured with first pass of the contrast agent using a 45 or 60 s scanning time (Goh et al., 2005), and can be only measured with longer acquisition times; however, there is no real consensus regarding the optimal acquisition time. Jain et al. (2008) used PCT for tumor vascular permeability quantification and reported slightly higher permeability values than those reported in literature using other techniques, and this could be explained by the mathematical model. Longer acquisition time of 170 s could also explain the higher values that they obtained because the scanning sequence was optimized to include delayed permeability also as described above. Gliomas, particularly high-grade gliomas, can have extremely variable and heterogeneous blood flow due to the complex tumor vasculature which can influence permeability (McDonald and Choyke, 2003; Vajkoczy and Menger, 2000). Other factors which can also influence permeability include luminal surface area, interstitial, hydrostatic, and osmotic pressure across the endothelium. Slow blood flow or low osmotic gradients which can occur in the high-grade tumor with considerable vasogenic edema and also in the central parts of large tumors can lead to a larger component of delayed permeability, requiring longer acquisition times.

84

Quantification of vessel permeability with various perfusion techniques requires two or more compartment pharmacokinetic models with an arterial input function, making these studies more complex than CBV estimation (Tofts et al., 1999), and is dependent on the imaging technique employed and the mathematical model that is used. Post gadolinium T1 weighted MRI gives a rough estimate of the disruption of BBB, and has been used in the past for quantitative estimation of permeability but dynamic imaging acquisition provides a better estimate of vascular permeability (Tofts et al., 1999). Dual echo gradient echo perfusion weighted imaging (Uematsu and Maeda, 2006) is based on a simple two compartment kinetic model (Boxerman et al., 2006), and has more recently been used to measure vascular permeability and to correctly estimate blood volume in brain tumors with deficient or absent BBB. There are several studies which have found a correlation between increased vascular permeability and higher tumor grade (Law et al., 2004; Johnson et al., 2004; Jain et al., 2008; Roberts et al., 2001). However, MR perfusion techniques have certain limitations because of the nonlinear relationship of the signal intensity with the contrast, both for dynamic contrast enhanced imaging with T1-weighting (Yankeelov et al., 2005), and for dynamic susceptibility contrast imaging with T2or T2∗ -weighting. In the latter case, when the contrast agent remains intravascular, the method is widely accepted as a relative estimate of CBF (cerebral blood flow) and CBV, although there is a possibility for artifacts because of difficulties in assessing the shape and timing of the arterial input function (Conturo et al., 2005). In the event that substantial leakage of a contrast agent from intra- to extra-vascular space takes place, a strong and competing T1 contrast effect is often noticed in the areas of pathology because of the necessity of short (∼1 s) repetition times needed to estimate CBF. As a first-order tactic to minimize the competing T1 contrast, preloading with contrast agent has been proposed with some success (Boxerman et al., 2006). However, this approach does not allow an estimate of Ktrans . An alternative has also been proposed (Johnson et al., 2004). To decrease the T1 effect, this approach used a slower repetition time, lengthening the repetition time of the experiment, and undermining the estimation of CBF; thus, yielding estimates of only CBV and Ktrans . A further refinement, allowing the estimate

R. Jain

of blood volume and producing an index of transfer constant, has been suggested (Boxerman et al., 2006), and a dual-echo gradient echo sequence (Uematsu and Maeda, 2006) also shows some potential for an index of blood volume and transfer constant. Despite the partial success of these rapid imaging studies, in contrast to CT perfusion, there does not appear to be an MRI technique that will reliably quantify CBF, CBV, and Ktrans in one experiment. PCT also has the advantage of providing absolute measures of these perfusion parameters whereas MR estimates are mostly relative to the normal brain parenchyma. Another major disadvantage of MR perfusion is susceptibility artifacts due to hemorrhage and various mineral depositions, which can be a major issue in posttreatment tumor patients. Despite these facts, MR perfusion has been more often utilized because MRI is the standard of care for brain tumor patients; whereas, PCT requires a separate examination with iodinated contrast agent and moderate exposure to ionizing radiation.

Perfusion CT Tracer Kinetics Theory The model proposed by Johnson and Wilson (1966) is a convenient model for tumor application and is based on the assumption that the distribution volume of CT contrast medium in tissue consists of the extra-cellular space in capillaries and interstitial space. Central Volume Principle relates tissue (tumor) blood flow (TBF), tissue (tumor) blood volume (TBV) and mean transit time (MTT) in the following equation: TBF = TBV/MTT (24). Perfusion studies are obtained by monitoring the passage of an iodinated contrast bolus through the cerebral vasculature. There is a linear relationship between contrast agent concentration and attenuation, with the contrast agent causing a transient rise in attenuation proportional to the amount of contrast in a given region. Contrast agent timeconcentration curves are generated in an arterial region of interest (ROI), a venous ROI as well as in each pixel. Deconvolution of arterial and tissue enhancement curves, according to the adiabatic approximation of the Johnson and Wilson model (St. Lawrence and Lee, 1998), gives the blood flow scaled impulse residue function (IRF), from which TBF, TBV, and MTT can be determined as prescribed in the reference (Lee et al., 2003). Specifically, blood flow is equal to the height of the blood flow scaled IRF. CBV is calculated as an

10 Brain Tumor Angiogenesis and Glioma Grading

area under the initial plateau (intravascular phase) of the blood flow scaled IRF. Permeability surface-area product characterizes the diffusion of some of the contrast agent from the blood vessels into the interstitial space due to deficient or leaky BBB. Permeability is related to the diffusion coefficient of contrast agent in the assumed water filled pores of the capillary endothelium. The diffusion flux of contrast agent across the capillary endothelium is dependent on both the diffusion coefficient and the total surface area of the pores. PS is computed from the IRF. Contrast agent diffusion appears in the IRF as a residual enhancement that occurs after the initial impulse response and which decreases exponentially with time. The IRF is used to estimate the first-pass fraction of contrast agent that remains in the tissue, the extraction fraction, E (Purdie et al., 2001). The extraction fraction is related to the rate at which contrast leaks out of the vasculature via the following relationship: PS

E =1−eF , where PS is the permeability surface-area product and F is flow. The PS product has the same dimensions as flow, and thus the ratio PS F is dimensionless. In physiological terms, PS is the rate at which contrast agent flows into the extravascular tissues; it is related to another commonly stated parameter of vascular leakage, the transfer constant by the following: K trans = EF, Ktrans

is the transfer constant with, again, the where same dimensions as flow. It is easily demonstrated trans ≈ PS. In normal cerethat, if PS F 1.2). Seven of 18 low grade (II) gliomas (38.9%) that did not show contrast enhancement on MRI with Gd-DTPA were not detected in MET-PET. Thus, of all the 62 gliomas studied, 55 (88.7%) were imaged by MET-PET. On the other hand, FLT-PET detected all 27 high grade gliomas. Five of 9 low grade gliomas (55.6%) were not visually detected in FLT-PET. Thus, of all the 36 gliomas studied, 31 (86.1%) were imaged by FLTPET. Tumors that were false-negative on MET-PET were also false-negative on FLT-PET. No tumors were detected by FLT-PET only. By using both tracers, the sensitivity rate in tumor detection was 88.9% (32 of 36 cases) in our series, because one false negative low grade glioma with FLT had a positive finding in MET-PET.

a

• MET-PET and FLT-PET can detect all high-grade gliomas (100% sensitivity). • MET-PET cannot detect about 40% of grade II gliomas and FLT-PET cannot detect about 56% of grade II gliomas. • MET-PET exhibits a slightly higher sensitivity in tumor detection than FLT-PET.

Tracer Uptake and Tumor Grade Normal brain parenchyma had low MET uptake with the SUVmean of 1.46 ± 0.33 (range, 0.93–2.62). The average MET SUVmax in WHO grade II (n = 18), III (n = 20) and IV (n = 24) gliomas was 2.80 ± 1.37, 3.94 ± 1.36 and 4.54 ± 1.31, respectively, and the average MET T/N ratio in grade II, III and IV gliomas was 1.83 ± 0.84, 2.81 ± 0.98 and 3.31 ± 0.94, respectively (Fig. 12.1a). There was a highly significant correlation between individual MET SUVmax and T/N ratio (r = 0.89, P < 0.001). The difference in the MET SUVmax and T/N ratio between grade II and grade IV gliomas was statistically significant

b

10

FLT SUVmax

MET SUVmax

8 6

* 4

3

2

# 1

2

0

0 Grade II

Grade III

Grade II

Grade IV

8

18

7

16

6

14 FLT TN ratio

MET TN ratio

4

5 4 3

*

10 6 4 2

Grade III (n = 20)

Grade IV (n = 24)

Fig. 12.1 PET tracer uptake in different WHO tumor grade

#

8

1 Grade II (n = 18)

Grade IV

12

2

0

Grade III

0

Grade II (n = 9)

Grade III (n = 9)

Grade IV (n = 18)

12 Newly Diagnosed Glioma

107

(P < 0.001). Again, the difference in the MET SUVmax and T/N ratio between grade II and grade III gliomas was statistically significant (∗ : P < 0.05). The difference in the MET SUVmax and T/N ratio between grade III and grade IV gliomas was not statistically significant. A scatterplot figure demonstrated that there was a significant overlap of MET uptake in the tumors despite the significant difference between grade II and grade III gliomas (Fig. 12.1a). This is due to relatively high MET uptake in grade II oligodendroglioma and oligoastrocytoma (arrows in Fig. 12.1a) compared to that in grade II astrocytoma. Normal brain parenchyma had very weak FLT uptake with the SUVmean of 0.19 ± 0.04 (range, 0.11– 0.28). The average FLT SUVmax in WHO grade II (n = 9), III (n = 9) and IV (n = 18) gliomas was 0.27 ± 0.08, 0.69 ± 0.42 and 1.98 ± 0.80, respectively, and the average FLT T/N ratio in grade II, III and IV gliomas was 1.48 ± 0.61, 3.87 ± 2.11 and 10.22 ± 3.79, respectively (Fig. 12.1b). There was a highly significant correlation between individual FLT SUVmax and T/N ratio (r = 0.93, P < 0.001). The differences in the FLT SUVmax and T/N ratio between grade III and grade IV gliomas were statistically significant (#: P < 0.001). However, the differences in the FLT SUVmax and T/N ratio between grade II and grade III gliomas were not statistically significant probably due to relatively high variation of FLT uptake in grade III gliomas. Four patients with grade III

anaplastic astrocytoma (Ki-67 index, 5–10%) had no contrast enhancement on MR images with Gd-DTPA. FLT uptake in these patients was similar to the low levels observed in grade II gliomas (SUVmax, 0.28–0.37) (arrows in scatterplot, Fig. 12.1b). • MET and FLT uptakes in gliomas increase with tumor grade. • Average MET uptake in grade III gliomas is significantly higher than that in grade II gliomas (P < 0.05). • MET uptake in grade II oligodendroglioma and oligoastrocytoma is relatively high compared to that in grade II astrocytoma. • Average FLT uptake in grade IV gliomas is significantly higher than that in grade III gliomas (P < 0.001). • FLT uptake in non-contrast enhanced grade III gliomas is relatively low compared to that in contrast enhanced grade III gliomas.

MET and FLT Uptake in Individual Tumors In 36 patients who underwent PET examination with both tracers, a significant but relatively weak correlation was observed between the individual SUVmax of MET and FLT (r = 0.52, P < 0.05) (Fig. 12.2a)

a

b

4

20

n = 36

16

3

14 FLT T/N ratio

FLT SUVmax

n = 36

18

2

12 10 8 6

1

4

r = 0.52, P < 0.05 0 1

2

3

4

5

6

7

8

9

MET SUVmax

Fig. 12.2 Correlation between MET uptake and FLT uptake

r = 0.49, P < 0.05

2 10

0 0

1

2

3

4

5

MET T/N ratio

6

7

8

108

and T/N ratio of MET and FLT (r = 0.49, P < 0.05) (Fig. 12.2b). The most distant three points (arrows in the figures) from the 95% confidence interval line, which exhibited high FLT uptake with moderate MET uptake, were glioblastoma cases with the highest Ki-67 indices (50–65%) in this series. • A significant but weak correlation is observed between the individual uptake of MET and FLT, and it seems to be difficult to precisely predict one result from the other.

Tracer Uptake and Tumor Proliferation Activity Histopathology was obtained from all 62 patients by surgery after the PET study. In all tumors examined, linear regression analysis showed a significant correlation between tracer uptake and tumor proliferation activity determined by Ki-67 labeling index. The analysis indicated a more significant correlation of the Ki-67 index with FLT SUVmax (r = 0.81, P < 0.001) and T/N ratio (r = 0.91, P < 0.001) (Fig. 12.3b) than with MET SUVmax (r = 0.40, P < 0.05) and T/N ratio (r = 0.48, P < 0.001) (Fig. 12.3a). • A significant correlation is observed between tracer uptake and tumor proliferation activity determined by Ki-67 labeling index. • Ki-67 labeling index is more significantly correlated with FLT uptake than that with MET uptake.

Discussion Positron emission tomography (PET) with 2-deoxy2-[18 F]fluoro-D-glucose (FDG) is a well established method in the diagnosis and management of patients suffering from brain tumors. FDG uptake is generally associated with histological tumor grade in gliomas. FDG uptake in low-grade gliomas (which are mostly grade II in adults) is usually similar to that of the normal white matter, whereas most grade III anaplastic gliomas have an FDG uptake exceeding that of the normal white matter or similar to that of the normal gray matter. Untreated glioblastomas show high FDG

N. Kawai et al.

uptake, which might be higher than that of the normal gray matter. One problem with the use of FDG for the diagnosis of gliomas is the high background uptake of FDG in glucose-dependent brain tissue. Thus, accurate evaluation of gliomas with FDG-PET is difficult especially in low grade gliomas. MET-PET possesses high specificity in tumor detection (Herholz et al., 1998), tumor delineation (Borbély et al., 2006), and differentiation of benign from malignant lesions (Ogawa et al., 1995). Various studies have found a correlation between tumor grade and MET uptake in gliomas (Ceyssens et al., 2006; Hatakeyama et al., 2008; Kaschten et al., 1998). Amino acids, including MET readily cross the intact BBB through neutral amino acid transporters (large amino acid transporter 1: LAT-1) and are incorporated into the area with active tumor. Our recent study has demonstrated that the grade of LAT1 immunostaining increases with glioma grade and the expression of LAT1 is significantly correlated with MET uptake in newly diagnosed gliomas (Okubo et al., 2010). MET uptake in gliomas significantly correlates with cell proliferation, in vitro Ki-67 labeling index (Hatakeyama et al., 2008; Kim et al., 2005), and proliferating cell nuclear antigen expression (Sato et al., 1999). Increased MET uptake seems to be caused by increased carrier-mediated and passive transport rather than elevated protein synthesis and is highly correlated with microvessel density (angiogenesis) in gliomas (Kracht et al., 2003; Nojiri et al., 2009). Many previous studies enrolled patients with newly diagnosed and previously treated recurrent cases in their materials. Radiation therapy used as an adjuvant therapy of gliomas can cause loosening of the endothelial tight junction, vascular leakage, or endothelial cell death and increase vascular permeability (Cao et al., 2005). Radiation could act to increase vascular permeability not only in the BBB but also in the blood-tumor barrier (BTB) (Cao et al., 2005), and thus, potentially increasing passive, non carriermediated transport of MET through the endothelial cells to tumor cells. In an earlier study on suspected gliomas recurrence, there was no significant relationship between the primary tumor histopathology (grade) and MET uptake ratio (Van Laere et al., 2005), in contrast to the findings of this study performed in a pre-therapeutic setting. Also, no significant differences were found in the SUVmax and T/N ratio of MET between recurrent malignant glioma and radionecrosis following stereotactic radiosurgery (Tsuyuguchi

12 Newly Diagnosed Glioma

b 80

80 n = 62

70

r = 0.40, P < 0.05

60

Ki-67 labeling index

Ki-67 labeling index

a

109

50 40 30 20 10 0

n = 36

70 60 50 40 30 20 10

r = 0.81, P < 0.001

0 –10

–10 1

2

3

4

6

5

7

8

1

0

9

MET SUVmax 80

3

4

80 n = 62

70

r = 0.48, P < 0.001

60 50 40 30 20 10

Ki-67 labeling index

Ki-67 labeling index

2

FLT SUVmax

60 50 40 30 20 10

0

0

–10

–10 0

1

2

3

4

5

6

7

MET T/N ratio

n = 36

70

r = 0.91, P < 0.001

0

2

4

6

8

10

12

14

16

18

FLT T/N ratio

Fig. 12.3 Correlation between PET tracer uptake and Ki-67 index

et al., 2003). As for the mechanism of tracer uptake, disruption of the BBB and BTB might be responsible for increased availability and transport, and such rupture might also be present in low grade glioma and radionecrosis without active tumor cells and mislead the interpretation of PET findings. In this study, therefore, we only enrolled patients with newly diagnosed glioma to reconfirm the usefulness of MET- and FLT-PET in tumor detection, noninvasive grading and assessment of the proliferation rate. Our study revealed a significant difference in the MET SUVmax and T/N ratio between grade II and grade III gliomas. Several studies have demonstrated the impossibility of differentiation between grade II and grade III gliomas with MET-PET only (Ceyssens et al., 2006; Hatakeyama et al., 2008; Kaschten et al., 1998). We found that 4 patients with grade II oligodendroglioma and oligoastrocytoma had a higher MET uptake (SUVmax, 3.45–6.8) (Fig. 12.1a, arrows in scatterplot) than did grade II astrocytoma (average

SUVmax, 2.21 ± 0.76, n = 13). Higher MET uptake in grade II oligodendroglioma compared with grade II astrocytoma has also been reported in a recent study (Nojiri et al., 2009). This difference might reflect specific metabolic properties of oligodendroglial cells (such as myelin synthesis) and different cellular densities and rates of cell turnover of oligodendroglioma. A recent study has also shown that an increase in the microvessel area in oligodendroglioma may contribute to the higher MET uptake among low-proliferative grade II gliomas (Nojiri et al., 2009). Increased MET uptake was observed not only in pure oligodendroglioma but also in mixed oligoastrocytoma (n = 3) in our study. This finding is contrary to the finding of a previous MET-PET study on increased MET uptake in mixed oligoastrocytoma. This could be explained by the relatively high oligodendroglial tumor components in our cases. Our results indicate that there was a significant overlap of MET uptake in the tumors despite a significant difference between grade II and grade III

110

gliomas. Therefore, MET-PET is not an ideal tool to precisely predict tumor malignancy and a surrogate for pathologic grading. Also, caution must be paid to evaluate the tumors with oligodendroglial components in MET-PET. Some studies have concluded that MET-PET has more clinical usefulness in assessing the tumor invasion (Kim et al., 2005; Ogawa et al., 1995) but not the tumor grade. Although we have not compared the MET-PET and MRI findings, a previous study reported that the gadolinium-enhanced area on MR images covered, on average, only 60% of the MET accumulation area in malignant gliomas (Miwa et al., 2004). In this study we did not include the case with pilocytic astrocytoma. Recently, Galldiks et al. (2009) showed that pilocytic astrocytomas demonstrated relatively high MET uptake when comparison was made with WHO grade II astrocytomas. The histopathological features of pilocytic astrocytoma include microvascular proliferation and infiltration, but are not signs of malignancy in this tumor entity. MET transport may be increased by an increased number of microvessels combined with a higher density of transporters (LAT-1) in the endothelial cells in pilocytic astrocytomas (Okubo et al., 2010). Again, caution must be paid to evaluate the tumors with pilocytic astrocytoma in MET-PET. FLT is a newly developed PET tracer, which allows for noninvasive assessment of tumor proliferation (Chen et al., 2005; Choi et al., 2005; Hatakeyama et al., 2008; Saga et al., 2006). In contrast to MET, which provides only an indirect measure of the proliferation status as amino acid uptake in the tumor, FLT allows the direct measurement of cellular TK1 activity, which has been reported to be proportional to the proliferation activity of the tumor. In fact, our linear regression analysis indicated a much more significant correlation of the Ki-67 labeling index with FLT uptake than with MET uptake in gliomas. As uptake of FLT is low in intact brain tissue, FLT-PET provides a lowbackground cerebral image, and is thus considered to be an ideal and attractive PET tracer for the imaging of gliomas. However, the sensitivity for the detection of tumors with FLT-PET has been reported to be lower compared with MET-PET, especially for low-grade gliomas. This is the case in our small series, absolute uptake of FLT in all tumors examined is lower than that of MET (SUVmax, 1.23 ± 0.98 vs. 3.84 ± 1.50), and the FLT uptake ratio to normal brain tissue

N. Kawai et al.

is higher than that of MET (T/N ratio, 6.45 ± 4.84 vs. 2.72 ± 1.10) because of extremely low uptake of FLT in normal brain tissue. Although sensitivity for tumor detection is lower for FLT-PET (86.1%) than for MET-PET (88.7%) in our series, the difference was not statistically significant. By using both tracers, the sensitivity rate in tumor detection was 88.9% (32 of 36 cases) in our series, because one falsenegative low grade glioma with FLT had a positive finding in MET-PET. Some studies have investigated several parameters simultaneously and concluded that different tracers, such as FDG and MET, can provide different and complimentary information in individual cases (Van Laere et al., 2005). However, low grade gliomas that were false negative on MET-PET were also false-negative on FLT-PET and no tumors were detected by FLT-PET only in our series. Therefore, there is no complimentary information in tumor detection with simultaneous measurements of FLT and MET uptake in cases where the tumor metabolism is only slightly increased (low grade gliomas) from that of the normal brain tissue. When comparison was made between individual MET and FLT uptake in the tumor, there was a significant but relatively weak correlation between the uptake of MET and FLT. The most distant three points from the 95% confidence interval area, which exhibited high FLT uptake with moderate MET uptake, were glioblastoma cases with a highly proliferative nature. This could be explained by the relatively high rate of metabolic trapping of FLT by TK1 compared to the influx effect across the BBB in these patients. Kinetic analysis of FLT uptake could clarify whether the transport effect or metabolic trapping largely contributes to the increased accumulation of FLT in the tumor (Jacobs et al., 2005). The FLT-PET study revealed that the differences in the SUVmax and T/N ratio between grade III and grade IV gliomas were statistically significant, but the differences between grade II and grade III gliomas were not statistically significant in our cases. This may have been due to the relatively small number of grade III gliomas and the relatively wide range of FLT uptake in these tumors (T/N ratio, 1.52–7.10). Four patients with grade III anaplastic astrocytoma (Ki-67 index, 5–10%) had no contrast enhancement on MR images with Gd-DTPA. FLT uptake was similar to the low levels observed in normal brain tissue in these patients (T/N ratio, 1.52–2.18, arrows in scatterplot, Fig. 12.1b).

12 Newly Diagnosed Glioma

These patients exhibited moderately increased MET uptake in the tumor (T/N ratio, 2.83–3.27). This discrepancy might be due to the selectivity of TK1 as a target of FLT for the salvage pathway of DNA synthesis. FLT accumulation is a reliable measure of the salvage pathway of DNA synthesis (Schwartz et al., 2003). In gliomas in which de novo DNA pathways are predominantly used in pyrimidine biosynthesis, the proliferation activity might be underestimated by this tracer (Schwartz et al., 2003). Muzi et al. (2006) reported a similar case in an FLT-PET study and concluded that FLT may be less useful in assessing proliferation in noncontrast-enhancing tumors regardless of histopathology grading. Low tracer access to the tumor tissue may limit the metabolic trapping of FLT by TK1 even in proliferative tumors. In conclusion, we have shown here that PET studies using MET and FLT are useful for tumor detection in 62 newly diagnosed histologically verified gliomas. MET has a slightly higher sensitivity in tumor detection than FLT (but not statistically significant) in newly diagnosed gliomas; and both tracers are 100% sensitive for malignant gliomas. However, there is no complimentary information with simultaneous measurements of MET and FLT in tumor detection of low grade gliomas. Our results reveal that MET-PET and FLT-PET appear to be a valuable tool for noninvasive tumor grading in newly diagnosed gliomas, however, caution must be paid to evaluate the tumors with oligodendroglial components for MET-PET and with noncontrast-enhancement for FLT-PET. FLT-PET seems to be superior to MET-PET in assessment of cellular proliferation activity in gliomas of different grade. This study could be adequately powered to make clear-cut distinctions for glioma grading in newly diagnosed gliomas when the sample size of FLTPET is appropriately increased. Further studies might identify the individual role of MET and FLT in gliomas of different types and grades.

References Been LB, Suumeijer AJH, Cobben DCP, Jager PL, Hoekstra HJ, Elsinga PH (2004) [18 F]FLT-PET in oncology: current status and opportunities. Eur J Nucl Med Mol Imaging 31: 1659–1672 Borbély K, Nyáry I, Tóth M, Ericson K, Gulyás B (2006) Optimization of semi-quantification in metabolic PET studies with 18 F-fluorodeoxyglucose and 11 C-methionine in

111 the determination of malignancy of gliomas. J Neurol Sci 246:85–94 Cao Y, Tsien CI, Shen Z, Tatro DS, Ten Haken R, Kessier ML, Chenevert TL, Lawrence TS (2005) Use of magnetic resonance imaging to assess blood-brain/blood-gliomas barrier opening during conformal radiotherapy. J Clin Oncol 23:4127–4136 Ceyssens S, Van Laere K, de Groot T, Goffin J, Bormans G, Mortelmans L (2006) [11 C]methionine PET, histopathology, and survival in primary brain tumors and recurrence. AJNR Am J Neuroradiol 27:1432–1437 Chen W, Cloughesy T, Kamdar N, Satyamurthy N, Bergsneider M, Liau L, Mischel P, Czernin J, Phelps ME, Silverman DHS (2005) Imaging proliferation in brain tumors with 18 F-FLT PET: comparison with 18 F-FDG. J Nucl Med 46:945–952 Chen W, Delaloye S, Silverman DHS, Geist C, Czernin J, Sayre J, Satyamurthy N, Pope W, Lai A, Phelps ME, Cloughesy T (2007) Predicting treatment response of malignant gliomas to bevacizumab and irinotecan by imaging proliferation with [18 F] fluorothymidine positron emission tomography: a pilot study. J Clin Oncol 25:4714–4721 Choi SJ, Kim JS, Kim JH, Oh SJ, Lee JG, Kim CJ, Ra YS, Yeo JS, Ryu JS, Moon DH (2005) [18 F]3 -deoxy-3 fluorothymidine PET for the diagnosis and grading of brain tumors. Eur J Nucl Med Mol Imaging 32:653–659 Galldiks N, Kracht LW, Berthold F, Miletic H, Klein JC, Herholz K, Jacobs AH, Heiss WD (2009) [11 C]-L-methionine positron emission tomography in the management of children and young adults with brain tumors. J Neurooncol 96:231–239 Galldiks N, Kracht LW, Burghaus L, Thomas A, Jacobs AH, Heiss WD, Herholz K (2006) Use of 11 C-methionine PET to monitor the effects of temozolomide chemotherapy in malignant gliomas. Eur J Nucl Med Mol Imaging 33:516–524 Hatakeyama T, Kawai N, Nishiyama Y, Yamamoto Y, Sasakawa Y, Ichikawa T, Tamiya T (2008) 11 C-methionine (MET) and 18 F-fluorothymidine (FLT) PET in patients with newly diagnosed glioma. Eur J Nucl Med Mol Imaging 35:2009–2017 Herholz K, Hölzer T, Bauer B, Schröder R, Voges J, Ernestus RI, Mendoza G, Weber-Luxenburger G, Löttgen J, Thiel A, Wienhard K, Heiss WD (1998) 11 C-methionine PET for differential diagnosis of low-grade gliomas. Neurology 50:1316–1322 Jacobs AH, Thomas A, Kracht LW, Li H, Dittmar C, Garlip G, Galldiks N, Klein JC, Sobesky J, Hilker R, Vollmar S, Herholz K, Wienhard K, Heiss WD (2005) 18 Ffluoro-L-thymidine and 11 C-methylmethionine as markers of increased transport and proliferation in brain tumors. J Nucl Med 46:1948–1958 Kaschten B, Stevenaert A, Sadzot B, Deprez M, Degueldre C, Fiore GD, Luxen A, Reznik M (1998) Preoperative evaluation of 54 gliomas by PET with fluorine-18fluorodeoxyglucose and/or carbon-11-methionine. J Nucl Med 39:778–785 Kim S, Chung JK, Im SH, Jeong JM, Lee DS, Kim DG, Jung HW, Lee MC (2005) 11 C-methionine PET as a prognostic marker in patients with glioma: comparison with 18 F-FDG PET. Eur J Nucl Med Mol Imaging 32:52–59 Kracht LW, Friese M, Herholz K, Schroeder R, Bauer B, Jacobs A, Heiss WD (2003) Methyl-[11 C]-L-methionine uptake as measured by positron emission tomography correlates to

112 microvessel density in patients with glioma. Eur J Nucl Med Mol Imaging 30:868–873 Miwa K, Shinoda J, Yano H, Okumura A, Iwama T, Nakahashi T, Sakai N (2004) Discrepancy between lesion distributions on methionine PET and MR images in patients with glioblastoma multiforme: insight from a PET and MR fusion image study. J Neurol Neurosurg Psychiatry 75:1457–1462 Muzi M, Spence AM, O’Sullivan F, Mankoff DA, Wells JM, Grierson JR, Link JM, Krohn KA (2006) Kinetic analysis of 3 -deoxy-3 -18 F-fluorothymidine in patients with gliomas. J Nucl Med 47:1612–1621 Nariai T, Tanaka Y, Wakimoto H, Aoyagi M, Tamaki M, Ishikawa K, Senda M, Ishii K, Hirakawa K, Ohno K (2005) Usefulness of L-[methyl-11 C] methionine-positron emission tomography as a biological monitoring tool in the treatment of glioma. J Neurosurg 103:498–507 Nojiri T, Nariai T, Aoyagi M, Senda M, Ishii K, Ishiwata K, Ohno K (2009) Contributions of biological tumor parameters to the incorporation rate of L-[methyl-11 C] methionine into astrocytomas and oligodendrogliomas. J Neurooncol 93:233–241 Nuutinen J, Sonninen P, Lehikoinen P, Sutinen E, Valavaara R, Eronen E, Norrgård S, Kulmala J, Teräs M, Minn H (2000) Radiotherapy treatment planning and long-term follow-up with [11 C]methionine PET in patients with low-grade astrocytoma. Int J Radiat Oncol Biol Phys 48:43–52 Ogawa T, Hatazawa J, Inugami A, Murakami M, Fujita H, Shimosegawa E, Noguchi K, Okudera T, Kanno I, Uemura K, Hadeishi H, Sasajima T (1995) Carbon-11-methionone PET evaluation of intracerebral hematoma: distinguishing neoplastic from non-neoplastic hematoma. J Nucl Med 36: 2175–2179 Okubo S, Zhen HN, Kawai N, Nishiyama Y, Haba R, Tamiya T (2010) Correlation of L-methyl-11 C-methionine (MET) uptake with L-type amino acid transporter 1 in human gliomas. J Neurooncol 99:217–225 Ribom D, Engler H, Blomquist E, Smits A (2002) Potential significance of 11 C- methionine PET as a marker for the radiosensitivity of low grade gliomas. Eur J Nucl Med 29:632–640 Ribom D, Eriksson A, Hartman M, Engler H, Nilsson A, Långström B, Bolander H, Bergström M, Smits A (2001) Positron emission tomography 11 C-methionine and

N. Kawai et al. survival in patients with low-grade gliomas. Cancer 92: 2541–2549 Saga T, Kawashima H, Araki N, Takahashi JA, Nakashima Y, Higashi T, Oya N, Murai T, Hojo M, Hashimoto N, Manabe T, Hiraoka M, Togashi K (2006) Evaluation of primary brain tumors with FLT-PET: usefulness and limitations. Clin Nucl Med 31:774–780 Sato N, Suzuki M, Kuwata N, Kuroda K, Wada T, Beppu T, Sera K, Sasaki T, Ogawa A (1999) Evaluation of the malignancy of glioma using 11 C-methionine positron emission tomography and proliferating cell nuclear antigen staining. Neurosurg Rev 22:210–214 Schwartz JL, Tamura Y, Jordan R, Grierson JR, Hrohn KA (2003) Monitoring tumor cell proliferation by targeting DNA synthetic processes with thymidine and thymidine analogs. J Nucl Med 44:2027–2032 Shiels AF, Grierson JR, Dohmen BM, Machulla HJ, Stayanoff JC, Lawhorn-Crews JM, Obradovich JE, Muzik O, Mangner TJ (1998) Imaging proliferation in vivo with [F-18]FLT and positron emission tomography. Nat Med 4:1334–1336 Tsuyuguchi N, Sunada I, Iwai Y, Yamanaka K, Tanaka K, Takami T, Otsuka Y, Sakamoto S, Ohata K, Goto T, Hara M (2003) Methionine positron emission tomography of recurrent metastatic brain tumor and radiation necrosis after stereotactic radiosurgery: is a differential diagnosis possible?. J Neurosurg 98:1056–1064 Tsuyuguchi N, Takami T, Sunada I, Iwai Y, Yamanaka K, Tanaka K, Nishikawa M, Ohata K, Torii K, Morino M, Nishio A, Hara M (2004) Methionine positron emission tomography for differentiation of recurrent brain tumor and radiation necrosis after stereotactic radiosurgery -in malignant glioma. Ann Nucl Med 18:291–296 Ullrich RT, Kracht L, Brunn A, Herholz K, Frommolt P, Miletic H, Deckert M, Heiss WD, Jacobs AH (2009) Methyl-L11 C-methionine PET as a diagnostic marker for malignant progression in patients with glioma. J Nucl Med 50: 1962–1968 Van Laere K, Ceyssens S, Van Calenbergh F, de Groot T, Menten J, Falmen P, Bormans G, Mortelmans L (2005) Direct comparison of 18 F-FDG and 11 C-methionine PET in suspected recurrence of glioma: sensitivity, inter-observer variability and prognostic value. Eur J Nucl Med Mol Imaging 32:39–51

Chapter 13

Role of Diffusion Tensor Imaging in Differentiation of Glioblastomas from Solitary Brain Metastases Sumei Wang, Harish Poptani, Elias R. Melhem, and Sungheon Kim

Abstract Differentiation between glioblastomas and solitary brain metastases is an important clinical problem as the treatment strategy can significantly differ depending on the tumor type. However, due to their similar appearance on conventional magnetic resonance imaging, accurate distinction between glioblastomas and brain metastases remains challenging and often necessitates an invasive surgical biopsy for a definitive diagnosis. Recently, diffusion tensor magnetic resonance imaging (DTI) has been applied in solving this problem and has demonstrated potential as a noninvasive imaging biomarker to differentiate glioblastomas from solitary brain metastases. This chapter provides a brief review on the fundamental concepts of DTI and how this new imaging method can be applied in differentiation between glioblastomas and solitary brain metastases. It also includes discussions on some practical issue to further improve the reliability of the method and potential metrics sensitive to tumor that can be derived from DTI data. Keywords Glioblastomas · DTI · Metastases · Neoplasms · MRI · ADC and FA

Introduction Glioblastomas and brain metastases are the two most common brain neoplasms in adults. The management of these two neoplasms is vastly different and can

S. Kim () Department of Radiology, Center for Biomedical Imaging, New York University School of Medicine, New York, NY 10016, USA e-mail: [email protected]

potentially affect the clinical outcome (Giese and Westphal, 2001; Soffietti et al., 2002). For example, patients with glioblastomas almost always undergo surgical resection (Giese and Westphal, 2001), while patients with suspected brain metastases without clinical history of systemic cancer undergo a complicated systemic staging to determine the site of primary carcinoma and evaluation for distant metastases before any surgical intervention or medical therapy (Soffietti et al., 2002). In some cases, clinical history and presence of multiple peripheral enhancing lesions in the brain makes the diagnosis of brain metastases relatively straightforward. However, appearance of the solitary brain metastases on magnetic resonance imaging (MRI) can be nonspecific. Similarly, glioblastomas can also occasionally present as multiple peripheral enhancing lesions. Although glioblastomas typically present as a solitary mass, a solitary brain metastasis may be the first manifestation of disease in about 30% of patients with systemic cancer. Hence, accurate distinction between glioblastomas and brain metastases remains challenging, which often necessitates an invasive surgical biopsy for a definitive diagnosis. Conventional MRI is limited in making an accurate distinction between glioblastomas and brain metastases due to their similar appearance. Both neoplasms may exhibit ring-enhancement and extensive edema. Over the past few years, diffusion tensor imaging (DTI) has been increasingly used to study pathologic changes in brain tumors (Murakami et al., 2009; Yamasaki et al., 2005). DTI has also been investigated in differentiating glioblastomas from metastases (Lu et al., 2004; Tsuchiya et al., 2005; Wang et al., 2009). In this chapter, we will briefly explain the DTI technique and its application in differentiating glioblastomas from solitary brain metastases.

M.A. Hayat (ed.), Tumors of the Central Nervous System, Volume 2, DOI 10.1007/978-94-007-0618-7_13, © Springer Science+Business Media B.V. 2011

113

114

Pathophysiology of Brain Tumors Glioblastoma is the most common and most malignant primary brain tumor. According to the World Health Organization (WHO) classification, it is also listed as Grade IV astrocytoma. Pathologically, it is composed of highly cellular anaplastic cells with marked nuclear atypia and mitotic activity, accompanied by marked regional heterogeneity and cellular polymorphism as well as microvascular proliferation and necrosis. Glioblastoma is associated with a high proliferation rate (Ki-67 labeling about 15–20%). As glioblastomas are biologically aggressive tumors, they tend to grow in an infiltrative manner, invading the surrounding tissues, especially the white matter tracts (Rees et al., 1996). Brain metastases are tumors originating from other organs of the body which spread into the brain via hematogenous routes. Neoplastic cells from the primary site can break away and enter the body’s circulatory blood stream. These neoplastic cells circulate, travel and take up residence in the brain. The malignant cells multiply in the endothelial cell layer and grow into the brain parenchyma in a non-infiltrating

Fig. 13.1 (a) Geometric interpretation of the tensor using the diffusion ellipsoid. The re-parameterized tensor obtained from a set of diffusion-weighted measurements has principal, intermediate, and minor eigenvalues (λ1 , λ2 , λ3 ) and corresponding eigenvectors (e1 , e2 , e3 ). These values define the relative magnitude and direction of diffusion along three spatial dimensions within each voxel respectively. Depending on the underlying probability of intravoxel diffusion, the tensor may take on one of three cardinal shapes – prolate (CL), spherical (CS), or oblate

S. Wang et al.

advancing pattern. The most common brain metastases are from lung, breast and melanoma. Histological appearance and cellularity of the brain metastases are identical to those of the primary tumor (Zhang and Olsson, 1997), and they may vary markedly among the primary tumor types.

Methodology-Diffusion Tensor Imaging Diffusion imaging probes the self-diffusional movement of water molecules, which can be detected by the attenuation of MRI signal using a diffusion weighted sequence. When unimpeded, water molecules move in a random manner (isotropic diffusion). However, the presence of obstacles, such as axonal membranes and myelin sheaths in white matter fiber tracts, restricts and/or hinders this molecular motion in a particular direction resulting in anisotropic diffusion. Apparent diffusivity of water is generally higher in directions parallel to fiber tracts than in the perpendicular direction (Beaulieu, 2002). Three dimensional probability distribution of diffusivity can be described by a diffusion tensor ellipsoid with three eigenvectors and the corresponding eigenvalues (λ1 and λ2 λ3 , Fig. 13.1a).

(CP). Reprinted with permission from Hess CP et al. (2007). (b) Bivariate normal distribution of the tissue shape measurement in 3P space. The distribution is highly overlapping in three groups of tissues: (1) corpus callosum (CC) and internal capsule (IC), (2) arcuate fasciculus (AF) and subcortical white matter (SCW), and (3) gray matter (GM), lentiform nucleus (LEN) and thalamus (TH). Reprinted with permission from Alexander et al. (2000)

13 Role of Diffusion Tensor Imaging in Differentiation of Glioblastomas

The eigenvector associated with the largest eigenvalue denotes the predominant orientation of fibers in a given imaging voxel. If a particular voxel has a high degree of anisotropy, one of the eigenvalues will be much higher than the other two. Most commonly used indices for diffusion tensor are apparent diffusion coefficient (ADC) and fractional anisotropy (FA) (Basser and Pierpaoli, 1996), which can be calculated according to Eqs. (13.1) and (13.2), respectively ADC = (λ1 + λ2 + λ3 )/3

(13.1)

  3 (λ1 − λ)2 + (λ2 − λ)2 + (λ3 − λ)2 FA = 2 λ21 + λ22 + λ23 (13.2) where λ denotes mean of the three eigenvalues. ADC is a measure of the directionally averaged magnitude of diffusion and is related to cell density, size and parenchyma permeability. FA represents the degree of diffusion anisotropy, and reflects the degree of alignment of cellular structure (Basser and Pierpaoli, 1996). Although FA is a good indicator of diffusion anisotropy, it does not provide information on the shape of the diffusion ellipsoid. For example, it cannot distinguish a flat ellipsoid from an oblong one. Westin et al. (2002) have modeled diffusion anisotropy using a set of three basic metrics that depend on the shape of the diffusion tensor: linear anisotropy coefficient (CL) where diffusion is mainly along the direction corresponding to the largest eigenvalue; planar anisotropy coefficient (CP) where diffusion is mainly restricted to the plane spanned by the two eigenvectors corresponding to the two largest eigenvalues; and spherical anisotropy coefficient (CS), which indicates isotropic diffusion (Hess and Mukherjee, 2007) (Fig. 13.1a). The CL, CP and CS values can be calculated using the following equations: CL = (λ1 − λ2 )/(λ1 + λ2 + λ3 )

(13.3)

CP = 2(λ2 − λ3 )/(λ1 + λ2 + λ3 )

(13.4)

CS = 3λ3 /(λ1 + λ2 + λ3 ).

(13.5)

The CL, CP and CS values lie in the range from 0 to 1 and the sum of these three metrics is equal to 1 for a given voxel: CL + CP + CS = 1.

(13.6)

115

Information on the geometric nature of diffusion tensor can aid in quantitative characterization of the local structure in tissue and provide further tensor shape differentiation in comparison to FA (Alexander et al., 2000; Westin et al., 2002; Zhang et al., 2004). The DTI features of white matter are quite different for each of the anisotropy measures. These differences arise from the relative contribution of the linear, planar and spherical shape components of the diffusion tensor. Figure 13.1b shows the distribution of tensor shape measurements from the gray matter, deep nuclei and white matter regions in three phase (3P) tensor shape diagram. The commissural and deep projection (internal capsule and corticospinal) tracts are dominated by linear shape, whereas arcuate fasciculus and peripheral radiation fibers have a significant planar feature. The gray matter appears isotropic with high CS and mainly clustered near the top of the plot. Deep nuclei (thalamus and lentiform) regions contain some axon connections and are slightly more anisotropic than the gray matter (Westin et al., 2002; Zhang et al., 2004). These studies suggest that tensor shape measurements allow one to explore the tissue microstructural difference.

Diffusion Tensor Imaging in Tumor Classification The diffusion properties of water molecules are directly related to the microstructure of the medium in which they reside, which determines the DWI and DTI contrast for diagnostic purposes. As such, diffusion MR imaging has been used extensively in clinical practice because of its exquisite sensitivity to cellular status, cytotoxic edema, cellular density, and directional organization of the tissue (Chenevert et al., 2006). In brain tumors, DTI provides a promising tool for detecting microscopic difference in tissue properties. Diagnostic brain tumor imaging involves not only the characterization of the tumor, but also detection of reactive and infiltrative changes surrounding the tumor. The neoplastic mass can be generally subdivided into two regions; the contrast-enhancing region representing the solid part of the tumor and the central area with no or slight enhancement representing the necrotic or cystic part of the tumor. Similarly, the peritumoral edematous region can be separated into two regions; proximal region surrounding the enhancing part of

116

the tumor potentially including infiltrative tumor cells, and more distal region mainly comprised of vasogenic edema. These four sub-regions of a tumor lesion can be substantially different from each other in terms of their DTI indices. Hence, when comparing the observations of different studies, it is often important to note where the measurements were made. In addition, systematic analysis of DTI parameters from these different areas may provide a robust way for characterization of brain neoplasms.

Correlation of ADC with Tumor Cellularity ADC reflects the rate of diffusional motion of the water molecules. Of all the histologic features used in tumor classification, cellularity has been the main target of assessment with DTI. The higher the tumor cellularity (and hence the higher volume of intracellular space), the lower the ADC value due to decreased water diffusivity caused by a relative reduction in extracellular space for the water molecules to move about (Beaulieu, 2002; Chenevert et al., 2006). This inverse correlation between ADC and cellularity has been reported in both glial (Yamasaki et al., 2005) and nonglial tumors (Guo et al., 2002) (Fig. 13.2).

Fig. 13.2 (a) Scatterplot of FA versus ADC measured in solid enhancing tumors and contralateral NAWM. Measurements from lymphomas tend to cluster in the area corresponding with FA and ADC decreases, whereas measurements from GBMs tend to cluster in area corresponding with FA and ADC increases. Reprinted with permission from Toh et al. (2008). (b) Scatterplot of FA versus MD measured in solid tumors. FA plotted as a function of MD. FA and MD showed an inverse

S. Wang et al.

ADC values have also been used in differentiating glioblastomas from metastases, however, with mixed results (Calli et al., 2006; Lu et al., 2004; Morita et al., 2005; Oh et al., 2005; Tsuchiya et al., 2005; Yamasaki et al., 2005). Some reports have suggested that ADC (Lu et al., 2004; Morita et al., 2005) is helpful for the differentiation, while others indicated the limited use of ADC in the differentiation of neoplasms (Calli et al., 2006; Oh et al., 2005; Yamasaki et al., 2005). In our previous study (Wang S et al., 2009), we did not observe a significant difference in the ADC values from the enhancing areas between glioblastomas and metastases, indicating limited sensitivity and specificity of this parameter in tumor differentiation (Fig. 13.3b). Besides cellularity, other factors such as extracellular matrix, viscosity and mucins may also affect the measurement of ADC (Zamecnik, 2005).

Role of FA from the Enhancing Region of the Tumor FA reflects the orientation of tissue microstructure, and as such its use may not be limited to the white matter tracts alone (Beaulieu, 2002). Regions of relatively high anisotropy have been reported in brain abscesses (Kumar et al., 2007), glioblastomas (Beppu et al.,

correlation (R = 0.70). Measurements from lymphomas tend to cluster in the area corresponding with FA increases and MD decreases, whereas measurements from high grade tumors tend to cluster in area corresponding with FA decreases and MD increases. G2; grade 2 glioma, G3 and G4; grade 3 and grade 4 glioma, ML; malignant lymphoma. Reprinted with permission from Kinoshita et al. (2008)

13 Role of Diffusion Tensor Imaging in Differentiation of Glioblastomas

Fig. 13.3 (a) A 68 year old female with a glioblastoma in the left frontal lobe. No hemorrhage was noted based on T1 and T2-weighted images (not shown). Transverse contrast-enhanced T1-weighted (1760/3.1) (A) and FLAIR (9420/141) (B) images show ring-enhancement and extensive edema. ADC map (C) shows restricted diffusion of the enhancing part. FA (D), CL (E) and CP (F) from the enhancing part are lower than normal appearing white matter. (b) Box plot of imaging characteristics in glioblastomas (white box) and brain metastases (gray box). Boxes represent the median, the 25th and the 75th percentiles, bars indicate the range of data distribution. Circles

117

represent outliers (values more than 1.5 box length from the 75th/25th percentile). ∗ indicates significant differences (p < 0.05) between glioblastomas and brain metastases. CR: central region. ER: enhancing region. IPR: immediate peritumoral region. DPR: distant peritumoral region. (c) Receiver operative characteristic (ROC) curves for FA, CL, CP, ADC and logistic regression model (LRM) from the enhancing region of the tumor. ADC+FA+CP is the best predictor for differentiation of glioblastomas from brain metastases with area under the curve (AUC) 0.98. Reprinted with permission from Wang S et al. (2009)

118

2003) and areas of hemorrhage (Kumar et al., 2007), indicating that FA is also related to structural orientation of the tissue/cells other than the white matter. In contrast to ADC, the relationship between FA and tumor cellularity is unclear, as both positive (Beppu et al., 2003; Kinoshita et al., 2008) and negative (Toh et al., 2008) correlation has been reported (Fig. 13.2). While Wang W et al. (2009) and Reiche et al. (2010) reported lower FA from the enhancing regions of glioblastomas compared with brain metastases, we observed higher FA in the enhancing regions of glioblastomas than in those of metastases (Wang S et al., 2009) (Fig. 13.3a). One likely reason for these contradictory results is the lack of standardized methods, both for acquisition as well as postprocessing and selection of region of interest (ROI). Our recent work (Wang et al., 2010) demonstrated high FA in glioblastomas compared with both brain metastases and primary cerebral lymphomas. Among these three tumor types, lymphomas have the highest cellularity, followed by glioblastomas and brain metastases (Koeller et al., 1997; Rees et al., 1996; Zhang and Olsson, 1997). These findings indicate that diffusion anisotropy may not directly correlate with tumor cellularity. It has been reported that FA of tumor can be affected by several factors including extracellular to intracellular space ratio, extracellular matrix, tortuosity and vascularity (Zamecnik, 2005).

Shape Based Diffusion Tensor Metrics in Differentiation of Brain Tumors FA is a scalar metric, which indicates the degree of anisotropy, regardless of the shape of the diffusion ellipsoid. Diffusion tensor can represent tubular, planar, or spherical shape of diffusion pattern, measured by CL, CP and CS, respectively (Alexander et al., 2000; Westin et al., 2002; Zhang et al., 2004). Both CL and CP values contribute to FA observed in tissue and their relative values indicate the shape of diffusion ellipsoid (Alexander et al., 2000). Tensor shape measurements provide further differentiation of tumors in comparison to that based on FA and/or ADC (Westin et al., 2002; Zhang et al., 2004). Zhang et al. (2004) reported lower CL in brain metastases than contralateral normal brain. Higher CP values in fibroblastic meningiomas were reported in

S. Wang et al.

comparison with other subtypes, such as endothelial meningiomas (Tropine et al., 2007). Kumar et al. (2007) reported high CP and low CL in the abscess cavity compared with normal white matter thus distinguishing true from pseudo white matter tracts. It has also been reported that epidermoid cysts have high CP (Santhosh et al., 2009) and tuberculomas showed lower CL, CP and higher CS (Gupta et al., 2008) compared with normal white matter. We have earlier demonstrated higher FA, CL and CP from the enhancing part of glioblastomas in comparison to brain metastases (Wang S et al., 2009) (Fig. 13.3b). These results suggest that tensor shape measurements provide additional information about tissue characteristics, which may further aid in tumor classification. A ring with high CP has been reported in glioblastomas, brain metastases and meningiomas. While the potential reason for the observation of this ring remains speculative, its presence may reflect compression of surrounding tissue by the tumor (Tropine et al., 2007; Zhang et al., 2004).

DTI in Assessing Tumor Infiltration Generally speaking, peritumoral region is defined as the area of abnormality surrounding the enhancing part of the tumor. In metastatic brain tumors, peritumoral edema is widely regarded as vasogenic edema. In this region, increased extracellular water is present due to leakage of plasma from altered tumor capillaries. Also this region does not include any tumor cells. In glioblastomas, on the other hand, the peritumoral region includes both vasogenic edema and infiltrating tumor cells. Discrimination of tumor-infiltrated edema from pure vasogenic edema may be beneficial for accurate preoperative diagnosis of glioblastomas and metastases. Most DTI studies on this subject have focused on the peritumoral regions of the tumor (Lu et al., 2004; Morita et al., 2005; Tsuchiya et al., 2005; Zhang and Olsson, 1997). In particular, Lu et al. (2004) reported a significant difference between tumor-infiltrated edema and pure vasogenic edema using a parameter called “tumor infiltration index”, which measures departure from a linear relationship between ADC and FA. These authors also reported higher ADC in metastasis compared to glioblastomas. However, another study

13 Role of Diffusion Tensor Imaging in Differentiation of Glioblastomas

demonstrated lower ADC in the peritumoral region of metastases compared to that of glioblastomas (Morita et al., 2005). In contrast, van Westen et al. (2006) reported no difference in ADC and FA values in the peritumoral region of glioblastomas, metastases and meningiomas. Recently, Kinoshita et al. (2010) claimed that “tumor infiltration index” could not differentiate vasogenic edema from tumor infiltrated edema. The difference in defining the ROIs for the peritumoral region in these studies may in part be responsible for the discrepancy. A number of studies have focused on the area close to the enhancing region (peritumoral region) by either manually placing a number of small ROIs around the tumor (Morita et al., 2005) or by using a band of arbitrarily chosen thickness around the tumor (Law et al., 2002; Oh et al., 2005). In our previous study (Wang S et al., 2009), the peritumoral areas were further subdivided into immediate peritumoral region and distant peritumoral region with the hypothesis that the immediate peritumoral region may have a higher degree of tumor infiltration in glioblastomas. We observed a significant difference in FA, CL, and CP between glioblastomas and metastases in the immediate peritumoral region. In the distant peritumoral region, only FA and CP measurements reached significant difference between the two tumor types (Wang S et al., 2009). While statistical significance was observed, the overall sensitivity, specificity and accuracy for all the DTI metrics in the peritumoral areas were lower than in the enhancing part of the tumor (Fig. 13.3c). Since the edematous region contains areas of increased extracellular water, tumor infiltration and varying fractional composition of normal white/gray matter, it is difficult to determine which factor dominates the DTI metrics. These confounding factors may further explain the conflicting reports of DTI characteristics in the peritumoral regions.

119

limited role in tumor classification. We have previously reported that the single best predictor for differentiation between glioblastomas and brain metastases is FA, with a sensitivity 89%, specificity 80% and AUC of 0.90 (Wang S et al., 2009). Accurate characterization of complicated tissue, such as a tumor, may require two or more imaging parameters. To date, only a limited number of studies have investigated the role of DTI parameters in combination for tumor classification. In a recent publication, it was suggested that a combination of minimum ADCs and difference of ADC facilitates more accurate grading of astrocytomas than either parameter measured individually (Murakami et al., 2009). We have previously reported that a multivariate logistic regression analysis can determine an optimal combination of DTI parameters to differentiate glioblastomas from brain metastases (Wang S et al., 2009). From a study with sixty three patients, we reported that a combination of ADC, FA and CP from the enhancing part of the tumor provides most accurate tumor classification, with a sensitivity of 92%, specificity of 100% and AUC 0.98 (Fig. 13.3c).

Conclusion In this chapter, we have discussed how glioblastomas and metastases can be characterized by DTI metrics, such as ADC, FA, CL, CP and CS. These DTI metrics can be used individually or in combination, to differentiate glioblastomas from metastases. Further investigations on a larger patient population and histological validation will be necessary to determine the robustness of these parameters in differentiating tumor types. Combined with rapidly growing MRI technology for faster imaging and higher resolution, DTI holds a great promise to elucidate morphological and functional characteristics of brain tumors noninvasively.

Combined DTI Metrics for Classification References As mentioned previously in the introduction, in vivo DTI data provides a number of parameters that provide information about the shape, magnitude and degree of diffusion anisotropy, which may in turn be used to differentiate different tumor types. However, these parameters, by themselves individually, have a

Alexander AL, Hasan K, Kindlmann G, Parker DL, Tsuruda JS (2000) A geometric analysis of diffusion tensor measurements of the human brain. Magn Reson Med 44:283–291 Basser PJ, Pierpaoli C (1996) Microstructural and physiological features of tissues elucidated by quantitative-diffusion-tensor MRI. J Magn Reson B 111:209–219

120 Beaulieu C (2002) The basis of anisotropic water diffusion in the nervous system – a technical review. NMR Biomed 15:435–455 Beppu T, Inoue T, Shibata Y, Kurose A, Arai H, Ogasawara K, Ogawa A, Nakamura S, Kabasawa H (2003) Measurement of fractional anisotropy using diffusion tensor MRI in supratentorial astrocytic tumors. J Neurooncol 63:109–116 Calli C, Kitis O, Yunten N, Yurtseven T, Islekel S, Akalin T (2006) Perfusion and diffusion MR imaging in enhancing malignant cerebral tumors. Eur J Radiol 58:394–403 Chenevert TL, Sundgren PC, Ross BD (2006) Diffusion imaging: insight to cell status and cytoarchitecture. Neuroimaging Clin N Am 16:619–632 Giese A, Westphal M (2001) Treatment of malignant glioma: a problem beyond the margins of resection. J Cancer Res Clin Oncol 127:217–225 Guo AC, Cummings TJ, Dash RC, Provenzale JM (2002) Lymphomas and high-grade astrocytomas: comparison of water diffusibility and histologic characteristics. Radiology 224:177–183 Gupta RK, Haris M, Husain N, Saksena S, Husain M, Rathore RK (2008) DTI derived indices correlate with immunohistochemistry obtained matrix metalloproteinase (MMP-9) expression in cellular fraction of brain tuberculoma. J Neurol Sci 275:78–85 Hess CP, Mukherjee P (2007) Visualizing white matter pathways in the living human brain: diffusion tensor imaging and beyond. Neuroimaging Clin N Am 17:407–426 Kinoshita M, Goto T, Okita Y, Kagawa N, Kishima H, Hashimoto N, Yoshimine T (2010) Diffusion tensor-based tumor infiltration index cannot discriminate vasogenic edema from tumor-infiltrated edema. J Neurooncol 96:409–415 Kinoshita M, Hashimoto N, Goto T, Kagawa N, Kishima H, Izumoto S, Tanaka H, Fujita N, Yoshimine T (2008) Fractional anisotropy and tumor cell density of the tumor core show positive correlation in diffusion tensor magnetic resonance imaging of malignant brain tumors. NeuroImage 43:29–35 Koeller KK, Smirniotopoulos JG, Jones RV (1997) Primary central nervous system lymphoma: radiologic-pathologic correlation. Radiographics 17:1497–1526 Kumar M, Gupta RK, Nath K, Rathore RK, Bayu G, Trivedi R, Husain M, Prasad KN, Tripathi RP, Narayana PA (2007) Can we differentiate true white matter fibers from pseudofibers inside a brain abscess cavity using geometrical diffusion tensor imaging metrics?. NMR Biomed 21:581–588 Law M, Cha S, Knopp EA, Johnson G, Arnett J, Litt AW (2002) High-grade gliomas and solitary metastases: differentiation by using perfusion and proton spectroscopic MR imaging. Radiology 222:715–721 Lu S, Ahn D, Johnson G, Law M, Zagzag D, Grossman RI (2004) Diffusion-tensor MR imaging of intracranial neoplasia and associated peritumoral edema: introduction of the tumor infiltration index. Radiology 232:221–228 Morita K, Matsuzawa H, Fujii Y, Tanaka R, Kwee IL, Nakada T (2005) Diffusion tensor analysis of peritumoral edema using lambda chart analysis indicative of the heterogeneity of the microstructure within edema. J Neurosurg 102: 336–341 Murakami R, Hirai T, Sugahara T, Fukuoka H, Toya R, Nishimura S, Kitajima M, Okuda T, Nakamura H, Oya N,

S. Wang et al. Kuratsu J, Yamashita Y (2009) Grading astrocytic tumors by using apparent diffusion coefficient parameters: superiority of a one- versus two-parameter pilot method. Radiology 251:838–845 Oh J, Cha S, Aiken AH, Han ET, Crane JC, Stainsby JA, Wright GA, Dillon WP, Nelson SJ (2005) Quantitative apparent diffusion coefficients and T2 relaxation times in characterizing contrast enhancing brain tumors and regions of peritumoral edema. J Magn Reson Imaging 21:701–708 Rees JH, Smirniotopoulos JG, Jones RV, Wong K (1996) Glioblastoma multiforme: radiologic-pathologic correlation. Radiographics 16:1413–1438 Reiche W, Schuchardt V, Hagen T, Il’yasov KA, Billmann P, Weber J (2010) Differential diagnosis of intracranial ring enhancing cystic mass lesions-role of diffusion-weighted imaging (DWI) and diffusion-tensor imaging (DTI). Clin Neurol Neurosurg 112:218–225 Santhosh K, Thomas B, Radhakrishnan VV, Saini J, Kesavadas C, Gupta AK, Kapilamoorthy TR, Nair SN (2009) Diffusion tensor and tensor metrics imaging in intracranial epidermoid cysts. J Magn Reson Imaging 29:967–970 Soffietti R, Ruda R, Mutani R (2002) Management of brain metastases. J Neurol 249:1357–1369 Toh CH, Castillo M, Wong AM, Wei KC, Wong HF, Ng SH, Wan YL (2008) Primary cerebral lymphoma and glioblastoma multiforme: differences in diffusion characteristics evaluated with diffusion tensor imaging. AJNR Am J Neuroradiol 29:471–475 Tropine A, Dellani PD, Glaser M, Bohl J, Ploner T, Vucurevic G, Perneczky A, Stoeter P (2007) Differentiation of fibroblastic meningiomas from other benign subtypes using diffusion tensor imaging. J Magn Reson Imaging 25:703–708 Tsuchiya K, Fujikawa A, Nakajima M, Honya K (2005) Differentiation between solitary brain metastasis and highgrade glioma by diffusion tensor imaging. Br J Radiol 78:533–537 Wang S, Kim S, Chawla S, Wolf R, Knipp D, Vossough A, Orourke D, Judy K, Melhem ER, Poptani H (2010) Differentiation between glioblastomas, brain metastases and primary cerebral lymphomas using diffusion and perfusion weighted imaging. Proc ISMRM 4287 Wang S, Kim S, Chawla S, Wolf RL, Zhang WG, O’Rourke DM, Judy KD, Melhem ER, Poptani H (2009) Differentiation between glioblastomas and solitary brain metastases using diffusion tensor imaging. NeuroImage 44:653–660 Wang W, Steward CE, Desmond PM (2009) Diffusion tensor imaging in glioblastoma multiforme and brain metastases: the role of p, q, L, and fractional anisotropy. AJNR Am J Neuroradiol 30:203–208 Westin CF, Maier SE, Mamata H, Nabavi A, Jolesz FA, Kikinis R (2002) Processing and visualization for diffusion tensor MRI. Med Image Anal 6:93–108 Yamasaki F, Kurisu K, Satoh K, Arita K, Sugiyama K, Ohtaki M, Takaba J, Tominaga A, Hanaya R, Yoshioka H, Hama S, Ito Y, Kajiwara Y, Yahara K, Saito T, Thohar MA (2005) Apparent diffusion coefficient of human brain tumors at MR imaging. Radiology 235:985–991 Zamecnik J (2005) The extracellular space and matrix of gliomas. Acta Neuropathol 110:435–442 Zhang S, Bastin ME, Laidlaw DH, Sinha S, Armitage PA, Deisboeck TS (2004) Visualization and analysis of white

13 Role of Diffusion Tensor Imaging in Differentiation of Glioblastomas matter structural asymmetry in diffusion tensor MRI data. Magn Reson Med 51:140–147 Zhang M, Olsson Y (1997) Hematogenous metastases of the human brain-characteristics of peritumoral brain changes: a review. J Neurooncol 35:81–89

121

van Westen D, Latt J, Englund E, Brockstedt S, Larsson EM (2006) Tumor extension in high-grade gliomas assessed with diffusion magnetic resonance imaging: values and lesion-to-brain ratios of apparent diffusion coefficient and fractional anisotropy. Acta Radiol 47:311–319

Chapter 14 131 I-TM-601 SPECT imaging of Human Glioma Adam N. Mamelak and David Hockaday

Abstract Primary malignant brain tumors (gliomas) typically infiltrate deeply into surrounding normal brain tissue. Currently available methods to determine the true extent of glioma infiltration are limited. For example, T1 weighted MRI tends to underestimate the extent of infiltration, while T2 weighted imaging tends to overestimate it. Most radio-labeled ligands are not tumor specific. Accurate determination of extent of infiltration is important to guide treatment and to measure response to therapy. TM-601 (chlorotoxin), a 36-amino-acid peptide first identified in the venom of a scorpion, selectively binds to glioma cells but not normal brain parenchyma. A phase I/II clinical trial of intracavitary 131 I-TM601 in adult patients with recurrent high grade glioma was performed to determine the bio-distribution and toxicity of this potential therapy. We evaluated the SPECT imaging and bio-distribution data from this trial to determine if 131 I-TM601 might also be useful in determining tumor extent. Adult patients with recurrent high-grade gliomas underwent tumor resection, implantation of an intra-cavitary reservoir, and a single-dose injection of 370 MBq (10 mCi) 131 I-TM601 (0.25–1.0 mg of 131 I-TM601) 2–4 weeks after surgery. Total-body planar scans and whole-brain SPECT scans were obtained on days 0, 1, 2, 3, and 6–8 after injection. Post-resection MRI images were co-registered to the SPECT scans using image analysis software. Analysis of the rate of radioactive decay and biologic elimination from the body and at the cavity site

A.N. Mamelak () Department of Neurosurgery, Cedars-Sinai Medical Center, Los Angeles, CA, USA e-mail: [email protected]

was performed. T1-weighted with gadolinium contrast (T1-Wc), T2-weighted (T2), and SPECT volumes were estimated by stereological Cavalieri sections and compared for overlap. Nonbound 131 I-TM601 was eliminated by 48 h after injection with the remaining radio-labeled peptide bound to tumor for at least 6– 8 day. Biologic decay rates from 24 to 168 h after injection were only slightly shorter than the physical decay of 131 I(6.3 vs. 8.0 day), indicating steady state binding. A comparison of tumor volume estimates using all three imaging parameters indicated that 131 I-TM601–determined tumor volumes more closely paralleled T2 volumes than T1-Wc volumes. Overlap between co-registered MRI and SPECT scans corroborated the presence of radio-labeled peptide in the vicinity of infiltrating tumor up to 168 h after injection. 131 I-TM601 provides a reliable estimate for primary tumor extent. The poor spatial resolution of 131 Iodine remains a major limitation to this tool. Further modification of this radio-peptide with better imaging isotopes such as 123 I (SPECT) or 124 I (PET), may provide clinicians with an important tool for determining tumor extent and differentiating regions of viable tumor from necrosis. Keywords Gliomas · 131 I-TM601 · Ligand Chlorotoxin · Decay rates · SPECT imaging

·

Introduction Accurate imaging of human malignancies in vivo requires methods that define the full extent of tumor, exclude normal tissues, and are spatially resolved enough to permit accurate delineation of tumor size and location. This ideal objective is very rarely

M.A. Hayat (ed.), Tumors of the Central Nervous System, Volume 2, DOI 10.1007/978-94-007-0618-7_14, © Springer Science+Business Media B.V. 2011

123

124

achieved by most currently available imaging methods, and represents a particular challenge for imaging of gliomas in the human brain. Gliomas are the most common primary brain tumors diagnosed annually, comprising approximately 16,500 cases and accounting for nearly 13,000 deaths (Central Brain Tumor Registry of the United States (CBTRUS), 1992–1997; American Brain Tumor Association, Primer of Brain Tumors (ABTA), 2001). The most lethal gliomas are the high-grade gliomas (HGGs): grade III anaplastic astrocytoma, anaplastic oligoendroglioma, and grade IV gliomablastoma multiforme (GBM). The ability to adequately image tumor volume extent is critical for treatment because more than 85% of HGG patients have recurrences near the original tumor site, and extent of tumor resection is one of the most important predictors of long-term survival (Brady et al., 1992; Natali et al., 1991; Debinski et al., 1995). Magnetic Resonance Imaging (MRI) is the primary method for imaging gliomas in vivo. Magnetic Resonance Imaging carries a very high spatial resolution that can be easily achieved, and is relatively inexpensive. Unfortunately T1 and T2 weighted MRI pulse sequences do not reliably differentiate infiltrating tumor from cytotoxic edema or radiation necrosis. The addition of gadolinium-based contrast agents can aide in defining regions of increased capillary permeability vascularity, but these are indirect measures of tumor and often underestimate the true extent of glioma invasion. Magnetic Resonance Spectroscopy (MRS) can detect biochemical abnormalities that are predictive of tumor, but this approach lacks the spatial resolution needed to reliably determine tumor extent, and is inherently limited by magnetic field inhomogeneities. Radio-isotope based imaging methods such as positron emission tomography (PET) and single photon emission computed tomography (SPECT) rely on cell surface receptor mediated binding or cellular uptake of a ligand, with subsequent imaging of the radio-isotope bound to the ligand. The most prominent example is the widespread use of 18 Flouro deoxyglucose (18 FDG). See other chapters in this volume for a more detailed description of the role of 18 FDG for imaging gliomas. 18 FDG is taken up by metabolically active cells utilizing glucose for energy production. Because cancer cells are generally more metabolically active than normal tissues, they preferentially take up the 18 FDG, providing an indirect measure of tumor.

A.N. Mamelak and D. Hockaday

While 18 FDG has been quite useful for staging of metastatic disease in several systemic cancers, it has been less useful for gliomas because glucose is the sole source of metabolic energy in the brain, and often the metabolic activity of normal brain exceeds that of low grade or intermediate grade glioma. 18 FDG PET is only 80% accurate in differentiating glioma from radiation necrosis (Barker et al., 1997; Thompson et al., 1999). 11 C-Methionine is a similar PET agent, although it may be slightly more accurate than 18 FDG in differentiating tumor from necrosis (Bigner et al., 1998). A comparison of these methods is addressed in another chapter in this volume. Spectroscopy has much less intrinsic spatial resolution than PET, and so in general has been even less useful as an imaging platform. Distinguishing viable tumor from vasogenic edema or necrosis, however, remains difficult, even with the use of metabolic imaging agents due to the significant amount of angiogenesis present in high-grade gliomas and the breaches of the blood-brain barrier (BBB) at the edema site. An imaging agent with the ability to bind to only tumor cells would provide a novel platform for determination of tumor extent. An ideal compound would be rapidly diffusible, tumor-specific, and has high resolution. Most of the currently available radioligands do not meet both of these criteria; either they are too large to negotiate the BBB and the interstices of the extracellular matrix (ECM) or they are not specific in their binding to glioma. Monoclonal antibodies (mAbs) such as 81C6, for instance, are too large to pass the BBB and diffuse from the surgically created resection cavity into the parenchyma, where residual tumor may thrive (Mamelak et al., 2006). Moreover, many mAbs lack glioma specificity and have a high degree of catabolism of the radiolabel. Chlorotoxin (CTX), a scorpion-derived 36 amino acid peptide, has been shown to specifically bind to a phosphatidyl inositide, a phosphorylated lipid on lamellipodia of tumor cells. Several immunohistochemical in vitro and in vivo experiments indicate that CTX does not bind to normal brain parenchyma or endothelial cells but specifically binds to tumescent glial cells. TM-601, a synthetic version of CTX, has been shown to selectively label human gliomas in vivo and in vitro (Soroceanu et al., 1998). Unlike many other radioligands, TM-601 is relatively small at 4 kDa: it passes the BBB and diffuses readily

14

131 I-TM-601 SPECT imaging of Human Glioma

through the brain parenchyma and may spread across the volume of the whole tumor. These properties make TM-601 potentially useful for imaging, as well as targeting local therapy. Several immunohistochemistry and cell-binding assays demonstrated a positive correlation between peptide-glioma binding and increasing tumor malignancy (Soroceanu et al., 1998). Intracranial injections in animals reliably showed long-term tumor-specific uptake. These findings and supporting in vivo efficacy and toxicology data in animals justified a Phase I clinical trial of TM-601 conjugated to 131 I in patients with recurrent high-grade glioma (Rajapakse et al., 2000). The primary goals of this study were to determine the safety profile and biodistribution of 131 I-TM-601. To preliminary evaluate the imaging potential of TM-601, the radiographic studies from subset of patients in this trial were analyzed in detail. Of note, 131 I-TM-601 was initially developed as a potential therapeutic agent, and that is why 131 I was chosen as the radioligand. As such the extremely poor spatial resolution of 131 I was a known shortcoming for imaging purposes, but more spatially resolved imaging agents could not provide any potential therapy, rendering their use unfeasible in this initial trial.

Methods Nine adult patients with recurrent high-grade gliomas underwent tumor resection implantation of an intracavitary reservoir and a single dose injection of 10 (±1) mCi 131 I-TM-601 (0.25–1.0 mg) 2–4 weeks after surgery. Protocol, eligibility criteria, and details of the clinical trial are presented elsewhere (Mamelak et al., 2006).

125

Prior to delivery of 131 I-TM-601 to the patient, a 1 cc volume of 0.1–1 mCi 111 In-DTPA was injected into the reservoir to assure catheter potency. At the time of 111 In-DTPA administration, a 256 × 256 matrix image was taken with a gamma camera equipped with a high resolution, low-energy (247 keV, 15% window) collimator to monitor delivery of the 111 In-DTPA via catheter and observe any leakage. Upon completion, dynamic planar images of the head were continuously acquired for 15 min. A 57 Co transmission scan was performed with and without the patient on the scanner bench to connect subsequent imaging data acquired for total body planar images. Prior to injection, the 57 Co source was placed 3 cm below the scanner table with the anterior camera 3 cm above the patient. A high-energy collimator centered at 122 keV with a 20% window was used and images were acquired with a gantry speed of 10 cm/min and 256 × 1024 matrix size.

Injection of Radiolabeled Peptide One to two hours after labeling, patients received 131 ITM-601 intracranially via a 25 gauge butterfly needle and a shielded 5 cc syringe with a volume of 2–5 cc. Initially, 25% of the total radiolabeled peptide dose was given. After 5 min. the remaining dose was administered, followed by a 1–2 cc saline flush. Blood and urine samples were collected for assessment of serum radioactivity. Urine samples were also collected over the period of days 1–2, 2–3, 3–4, and 4–6 or 8, while blood samples were collected 1, 2, and 4 h after the completion of the infusion, and days 2, 3, 4, and the time of imaging on day 6 or 8.

Planar Images Preparation of 131 I-TM-601 and Preliminary Scans Presealed, sterilized vials containing lyophilized TM601 were radiolabeled with 10 mCi I131 via the Iodogen bead method. Extent of labeling was determined via immediate thin layer chromatography (ITLC) with >92% efficiency required to justify human use.

After intracranial injection of 131 I-TM-601, anterior and posterior whole-body planar images of matrix size 1024 × 256 were acquired on a dual-detector gamma camera equipped with high-energy collimator set at 364 keV (15% window). Image acquisition included a 20 mL, 200 μCi calibrated 131 I source, which had been placed 10 cm from the feet of the patient. Subsequent images were acquired days 2, 3, 4, and either day 6 or 8 post injection.

126

Spectroscopy A 180◦ whole brain spectroscopy scan was performed on a dual headed gamma camera (Toshiba 7200 GCA), with fiducial markers containing10 μCi 131 I-TM-601 placed on each zygoma and on the nasion to orient SPECT scans with regard to subsequent MR images. A 1 mCi 131 I-TM-601 standard was also placed 10 cm above the patient head. Data was acquired at 3 degrees per step at 30 s per step. For these images, a 128 × 128 matrix was employed, the high-energy collimator was kept centered at 364 keV with a 15% window, and a ramp filter was applied for 3D reconstruction. Data were acquired in a proprietary format and converted to Interfile 3.0 for data analysis. Subsequent SPECT scans were performed with the total body planar images on days 2, 3, 4, and 6 or 8 after injection.

Magnetic Resonance Imaging Postoperative MRI scans were also acquired for each patient prior to 131 I-TM-601 injections. For each patient, 20 transverse slices were acquired on a 16 bits per pixel scale with matrix size of 256 × 256. Sequences included coronal, axial, and sagittal T1weighted images (with and without gadolinium contrast), T2 axial, FLAIR, and 3D-SPGR images. All MRI scans were stored as DICOM files for future image processing and analysis.

Spectroscopy Image Processing All image analysis was performed with Analyze 5.0 (AnalyzeDirect, Lenexa, KS). For each patient, the SPECT scan, comprised of 128 × 128 × 68 voxels of 79.5-mm3 , for each day (1–8) was loaded on Analyze 5.0, with the slices containing the standard excluded. For each SPECT scan, the intensity range on a 32 bit per pixel scale was calculated and recorded with the expectation that the magnitude (the range of intensity values) of the brightest voxel should correspond roughly to the decay and/or the dispersion of radioisotope. Each value was normalized for the duration of the scans, and the normalized intensities for each patient were then averaged for all patients on each day of scanning. To check the accuracy of this method,

A.N. Mamelak and D. Hockaday

the magnitude of the brightest voxel of the standard was recorded for five patients and measured against the physical decay of I131 , accomplished by defining a volume of interest (VOI) which only included the standard (the slices removed from the cranial VOI) and measuring the calculated intensities within this VOI. Once normalized and averaged, the intensities were fitted exponentially.

Formation of Magnetic Resonance Imaging Volumes For each patient, T1 weighted with gadolinium contrast (T1-Wc) and T2 volumes were formed from 2-D images, on the Analyze 5.0 platform. Resolution of the coronal and sagittal images was considerably lower than the axial resolution: voxel height was 7 mm; length and width, 0.83 mm. Therefore, the Force Cubic option (under Load As) was used to load the volumes as linearly interpolated, isotropic voxels of 0.83 mm × 0.83 mm × 0.83 mm or 0.57 mm3 . The volumes were primarily analyzed in the axial (transverse) plane.

Image Analysis: Stereological Estimation of Volume Tumor volumes for each patient were determined from the T1-Wc, T2, and SPECT scans. Given the arbitrary nature of gray-scale threshold methods and the substantial differences that result in volumes of extracted objects (Roberts et al., 2000) stereological Cavalieri sections were used in Analyze 5.0 (Shen et al., 2005). One of every three slices was sampled for each scan, and the highlighted or tumor associated region demarcated. A reasonable minimum volume was determined via the exclusion of ambiguous sample points; a reasonable maximum volume was determined via those sample points inclusion. The mean was then calculated for each volume with the accompanying error equal to half the spread between the recorded minimum and maximum. In addition, for SPECT scans (where limited resolution of 131 I rounds-off tumor extent) a full width at half-maximum (FWHM) was intermittently calculated across the cross-sectional diameter (transverse plane) using the Line Profile option, with the

14

131 I-TM-601 SPECT imaging of Human Glioma

127

assumption that intensity falloff would more or less follow a normal (Gaussian) curve. The diameter of the stereological volume was then checked by the sampled FWHM for accuracy. For MR images, two volumes were determined for each patient: T1-Wc and T2. For SPECT, the tumor volumes for the second day post injection and the final day were catalogued for each patient.

Co-registration After separate image processing, the isotropic MR volumes were co-registered to the SPECT scan-volumes using Analyze 5.0. Under the 3-D Registration program, SPECT scans were matched manually and fused to MR images for each patient, using the alignment of fiducial markers on each SPECT volume to anatomical structures of the nasion and zygoma on the MR image. Gray-scale intensities were converted to a 24bit red, green, blue (RGB) “Hot-Metal” colormap for visual affect.

Results Bioelimination Total body planar-scans had been previously used by Shen et al. (2005) to estimate effective half-lives in various organs; in the present nine patient subset, these half-lives were recalculated as 47±5 h for bodily elimination, 45±5 h for elimination from the brain, and 39±6 h for elimination from the cavity. Qualitatively, these results are seen in the total-body planar scans on day 2 and day 6–8 (Fig. 14.1). For all nine patients, retention of radiolabeled peptide was observed in the tumor cavity region for the duration of scans. All SPECT scans also showed clear evidence of 131 I-TM-601 remaining at the tumor site on Day 8 (168 h post injection) at an activity comparable to that of the 1 mCi standard. For the standard, the decay rate of the value of the brightest voxel roughly followed the theoretical (physical) rate of decay. The calculated exponential of the standard yielded a half-life of 6.3±1 days, comparable to the known half-life for 131 I of 8.04 days.

Fig. 14.1 Total body planar scans of a single patient on days 1 and 8. Note the intense uptake and long-term retention of radioisotope in the right frontal lobe injection site, and minimal uptake anywhere else in the body. The 100 μCi standard is evident at the patient’s foot

Tumor Volume Stereological estimation of tumor volumes in the nine patients showed 131 I-TM-601 estimated volumes were intermediate between the T1-Wc estimated volumes on the low end and T2 determined volumes on the high end. Over the course of scanning, the estimated volume 24 h post-injection was significantly larger and above or equal to T2 estimates for most patients; by 168 h post-injection most peptide determined volumes are smaller, better centered between T1-Wc and T2 volumes, and considerably less uncertain. Overall, SPECT determined volumes on Day 2 and Day 8 were always larger than the T1-Wc determined volumes, and most were smaller than the T2 determined volumes by Day 8. Moreover, FWHM measures across the span of the SPECT volume corresponded to the smaller volumes of Day 8 stereological estimates.

Fusion of Images The fusion of SPECT volumes to MR images via the use of fiducial markers yielded high-overlap and wellcentered regions of tumor extent 168 h post-injection (Fig. 14.2). As the SPECT-estimated volumes were rounded, they tended to approximate the area around the edema (highlighted region in T1-Wc) and necrotic

128

Fig. 14.2 Co-registration of MRI and SPECT images demonstrate excellent overlap and permit better definition of region of tracer uptake. Day 8 imaging has better resolution and closer approximates the true tumor extent than day 2 estimates

region (dark center region in T1-Wc) and resection area (also dark) and extend to regions in the T2determined volume. Generally, the SPECT-determined volume was centered in this region and agreed with the highlighted T1-Wc and T2 areas. Spectroscopy images performed on Day 1 exhibited overwhelming scatter and thus were not useful for image analysis. By Day 2, fused images showed considerable isolation of the highlighted SPECT region to that of the tumor region on the MR image. Further differentiation was evident on Day 3, with SPECTdetermined volume better centered and shifted from the site of injection and the cavity towards the necrotic core. By Day 8, residual glow from the cavity and injection site had vanished and remaining radiolabeled isotope had an intensity level comparable to the standard.

Discussions Tumor Volumes For the present study, 131 I was chosen for its potential therapeutic benefit compared with less energetic but better imaging radioisotopes such as 123 I or 124 I. With the limited resolution of the current isotope, considerable difficulty arises in defining exact tumor extent. Stereological methods offer better statistical estimates and more reproducible results than traditional threshold methods when the exact extent of the region of interest is unknown (Roberts et al., 2000; Shen et al.,

A.N. Mamelak and D. Hockaday

2005; Gadeberg et al., 1999). As a practical measure, stereology also allows the sampling of a smaller number of slices with the same or better accuracy than ROI drawing. Therefore, as far as 131 I is concerned, volume of distribution of TM-601 is best estimated from stereology. Of course, higher resolution radioisotopes may allow more precise methods to be used in the foreseeable future. As it stands, spectroscopy-estimated volumes of distribution must be approached with several factors in mind. Spectroscopy-estimated volumes appear rounded-off due to the limited resolution of the radioisotope. Smaller, more peripheral features are lost among the scatter. Their shape is primarily spherical or ellipsoidal, despite their more irregular appearance in T1-Wc or T2 MRI, and the uncertainty of the estimate is generally large. Similar qualifications exist for MRI-determined volumes, where the tumor per se is not measured but rather the associated enhanced regions of increased capillary permeability associated with tumor but also possibly postoperative or radiation-induced effects, or inflammatory clogs. Generally for T1-Wc, gliomatumor volume is underestimated as invasive fringe tumor cells lack the increased angiogenesis necessary to register in this modality. In contrast, T2 signal hypersensitivity may represent tumor or radiationinduced, vasogenic edema, infiltrating tumor, and/or cytotoxic effects, and as such, T2 estimates tend to exaggerate the true extent of tumor. In the present study, SPECT determined volumes using 131 I-TM-601 appear to provide better estimates than T1-Wc or T2 determined estimates. The larger volume and uncertainties in Day 2 stereological estimates corresponded with the presence of more diffusely highlighted voxels than in Day 8, where steep falloff allowed the highlighted SPECT volume to be readily discernible. Scattering effects diminished resolution beyond practicality for the initial day, and no distinguishable trends could be registered within the interval of a day, i.e., the volumes of consecutive days were effectively the same. By Day 8, this uncertainty had diminished substantially and the volumes were more easily estimated, because a sample marker could be readily identified as belonging to the set of volume points. Therefore, lower energy imaging isotopes of higher spatial resolution may provide substantially better tumor definition.

14

131 I-TM-601 SPECT imaging of Human Glioma

On the whole, estimated SPECT volumes were smaller on Day 8 suggesting that the non-bound radiolabeled peptide is rapidly eliminated from the brain, while bound peptide persists. The biodistribution data also support this contention. The volumes are hedged between the stereologically estimated T1-Wc and the T2 volumes. Most SPECT volumes were in fact closer to the T2 estimate than the T1-Wc by Day 8, suggesting that 131 I-TM-601 diffuses and binds to the full extent of primary tumor, while the limited resolution may explain overestimation. Total body planar images and SPECT scans frequently showed some “hot spots” of increased tracer uptake removed from the primary tumor volume. These hot spots may indicate diffusion of the peptide to sites distal from the main tumor mass. This finding, while speculative and preliminary, would be very valuable for determining true tumor extent and predicting sites of recurrence or treatment failure.

Overlap of Fusion Volumes The detailed anatomical clarity of MRI coregistered to SPECT scans allows more accurate differentiation of actual tumor and imaging artifacts then seen in the SPECT scans alone. The increased utility of image fusion is well documented and likely to be important in determining the utility of this peptide for longterm use. Other imaging technologies such as “Digital Subtraction” may also prove useful in increasing the imaging sensitivity with TM-601.

Indications of Decay Rate Qualitative evidence in the SPECT scans supports the elimination calculated by biodistribution in the planar scans. Although calibration was performed for planarscans, quantitative results were unavailable for the SPECT scans. No calibration between count-rate and intensity was available for the particular SPECT scanner used, and the attenuation and scatter corrections remain unknown. Even if these factors were evident, defining the VOI (the exact extent of enhanced region) with the limited resolution of 131 I would be practically ineffective in the clinical setting.

129

Nevertheless, the maximum and minimum intensities for each scan had been calculated, the difference of which yielded the intensity value of the brightest voxel. An ad-hoc approach was taken to compare the intensity of the injected radioisotope to that of the standard, though with admittedly limited statistics. These values show substantial slowing of decay 24– 48 h post-injection to rates towards that expected of purely physical decay, in agreement with the calculated elimination by biodistribution in the planar images, 39±6 h for elimination from the cavity. Rates continue mild slowing through 168 h post-injection as they approach physical decay rates. If the above methods are indeed reliable, these rates may represent a shift from radioisotope dispersion to physical decay 24–48 h post-injection.

Future Directions TM-601 appears to represent a very attractive ligand for radioisotope-based imaging of gliomas, and potentially many other malignancies. It meets the ideal criteria of being tumor-specific, non-toxic and rapidly diffusible in solid organs. It appears to exhibit longterm binding to tumor at low doses. Several steps are required to determine the ultimate utility of this approach. First and foremost, less energetic and more spatially selective imaging isotopes must be tested in both animal models and humans. 123 I is an obvious first choice due to the relative ease of labeling. However, it is likely that PET sensitive isotopes such as 124 I or 64 Cu may be even more appealing. Methods to maximize peptide-isotope binding conditions still need to be worked out. In conclusion, the specificity and stability of 131 ITM-601 allow adequate imaging of high-localization and limited resolution in the clinical setting, which is promising for the development of higher resolution radioligands of TM-601 such as 124 I or 64 Cu for PET imaging. Although SPECT determined volumes using 131 I-TM-601 appear to provide better estimates than T1-Wc or T2 determined estimates the overlap of fusion images suggests 131 I-TM-601 fully estimates the extent of primary brain tumor. Imaging studies following intravenous injection will also be important in determining the ultimate utility of this novel peptide in the clinical setting.

130

References American Brain Tumor Association, Primer of Brain Tumors (2001) Chapter 3: facts and statistics, 7th edn. ABTA, DesPlaines, IL Barker FGII, Chang SM, Valk PE, Pounds TR, Prados MD (1997) 18-Flourodeoxyglucose uptake and survival of patients with suspected recurrent malignant glioma. Cancer 79(1):115–126 Bigner DD, Brown MT, Friedman AH, Coleman RE, Akabani G, Friedman HS, Thorstad WL, McLendon RE, Bigner SH, Zhao X-G, Pegram CN, Wikstrand CJ, Herndon JEII, Vick NA, Paleologos N, Cokgor I, Provenzale JM, Zalutsky MR (1998) Iodine-131-labeled antitenascin monoclonal antibody 81C6 treatment of patients with recurrent malignant gliomas: phase I trial results. J Clin Oncol 16:2202–2212 Brady LW, Miyamoto C, Woo DV, Rackover M, Emrich J, Bender H, Dadparvar S, Steplewski Z, Koprowski H, Black P, Lazzaro B, Nair S, McCormack T, Nieves J, Morabito M, Eshleman J (1992) Malignant astrocytomas treated with iodine-125 labeled monoclonal antibody 425 against epidermal growth factor receptor: a phase II trial. Int J Radiat Oncol Biol Phys 22(1):225–230 Central Brain Tumor Registry of the United States (CBTRUS), 1992–1997 data. From www.cbtrus.org. 3 Debinski W, Obiri NI, Powers SK, Pastan I, Puri RK (1995) Human glioma cells overexpress receptors for interleukin 13 and are extremely sensitive to a novel chimeric protein composed of interleukin 13 and pseudomonas exotoxin. Clin Cancer Res 1(11):1253–1258

A.N. Mamelak and D. Hockaday Gadeberg P, Gundersen HJG, Tågehøj F (1999) How Accurate are Measurements on MRI? A Study on Multiple Sclerosis Using Reliable 3D Stereological Methods. J MRI 10:72–79 Mamelak AN, Rosenfeld S, Bucholz R, Raubitschek A, Nabors LB, Fiveash JB, Shen S, Khazaeli MB, Colcher D, Liu A, Osman MH, Guthrie B, Schade-Bijur S, Hablitz DM, Alvarez VL, Gonda MA (2006) Phase I single-dose study of intracavitary administered 131 I-TM-601 in adults with recurrent high-grade glioma. J Clin Oncol 24(22): 3644–3649 Natali PG, Nicotra MR, Bigotti A, Botti C, Castellani P, Risso AM, Zardi L (1991) Comparative analysis of the expression of the extracellular matrix protein tenascin in normal human fetal, adult and tumor tissues. Int J Cancer 47(6): 811–816 Rajapakse JC (2000) Random-grid stereologic volumetry of MR head scans. J MRI 12:833–841 Roberts N, Puddephat MJ, McNulty V (2000) The benefit of stereology for quantitative radiology. Br J Radiol 73: 679–697 Shen S, Khaaeli MB, Gillespie G, Alvarez VL (2005) Radiation dosimetry of 131 I-chlorotoxin for targeted radiotherapy in glioma-bearing mice. J Neurooncol 71:113–119 Soroceanu L, Gillespie Y, Khazaeli MB, Sontheimer H (1998) Use of chlorotoxin for targeting of primary brain tumors. Cancer Res 58(12):4871–4879 Thompson TP, Lunsford LD, Kondziolka D (1999) Distinguishing recurrent tumor and radiation necrosis with positron emission tomography versus stereotactic biopsy. Stereot Funct Neurosurg 73(1–4):9–14

Chapter 15

Assessment of Biological Target Volume Using Positron Emission Tomography in High-Grade Glioma Patients Habib Zaidi and Srinivasan Senthamizhchelvan

Abstract High-grade gliomas (HGG) are the most challenging brain tumors to treat. Even though various sophisticated options exist to treat patients with gliomas, the disease invariably leads to death over months or years. The major obstacles encountered in treating gliomas are in determining the exact location, extent, and metabolic activity of the tumor. Molecular imaging of energy metabolism, amino acid transport, cell proliferation and cell death have been found helpful in identifying the biologically active tumor tissues for therapy. It allows a better understanding of pathology at the molecular level. This ability is especially useful in brain tumors where tissue sampling in vivo is associated with significant risks. Positron emission tomography (PET) is one of the most prominent molecular imaging modalities utilized for imaging pathophysiology of tumors at an early stage. In this chapter, the applicability of PET in assessing the biologically active tumor volumes in high-grade glioma patients for radiation therapy treatment planning and therapy monitoring will be reviewed. We will focus on the concept of biological target volume (BTV) and associated methods of image segmentation available for delineating tumor volumes in connection with their applicability in high-grade gliomas. Keywords High-grade gliomas · PET · Molecular imaging · BTV · Necrosis · Dose painting

H. Zaidi () Division of Nuclear Medicine, Geneva University Hospital, CH-1211 Geneva 4, Switzerland e-mail: [email protected]

Introduction The success of cancer treatment depends on multiple factors. One of the most important factors is the accuracy of the information about tumor location, extent and magnitude of disease. Traditionally this information is obtained, through anatomical imaging methods such as x-ray computed tomography (CT), magnetic resonance imaging (MRI), and ultrasound (US). However, it has become clear now, that the acquisition of molecular and physiological information by noninvasive molecular imaging modalities such as positron emission tomography (PET) could vastly enhance our ability to fight cancer at an early stage (Weissleder, 2006). Molecular imaging has the potential to detect physiological alterations that signal the existence of cancer when it is still at a curable stage. Advances in genomics and proteomics technologies have shown the potential to transform the way in which cancer is clinically managed today. Molecular imaging is poised to play a key role in this transformation, since it will allow the integration of molecular and physiological information specific to each individual case with anatomical information obtained through conventional imaging methods. As a noninvasive molecular imaging method PET exploits the unique decay characteristics of positron-emitting isotopes. The isotopes of fluorine, oxygen, carbon, and others have been routinely used in the development of diagnostically useful biological tracers that are available for PET imaging of functional and/ or metabolic assessment of normal tissues or disease state. Conventional stand-alone PET has now been replaced by PET/CT for improved patient throughput and most importantly for the availability of

M.A. Hayat (ed.), Tumors of the Central Nervous System, Volume 2, DOI 10.1007/978-94-007-0618-7_15, © Springer Science+Business Media B.V. 2011

131

132

complementary information of molecular PET images and anatomical CT images in one imaging session. There has been a tremendous expansion of clinical applications of PET in oncology for the diagnosis, staging and restaging of cancer patients. More than half of all patients with cancer receive radiation therapy (RT) at some stage during the course of their disease management. Applications of PET in RT have been reported in lung, head and neck, breast, lymphoma, prostate and many other cancers (Zaidi et al., 2009). Studies have also found that PET has advantages over CT in the standardization of tumor volume delineation in the reduction of the risk for geometrical misses, in the minimization of radiation dose to the normal organs, and in the assessment of tumor burden, blood flow, tissue inflammation, and hypoxia. Integration of functional PET data with anatomical CT data has become a standard in RT particularly in treatment planning of various cancers. Imaging plays a key role in the state-of-the-art high-precision RT techniques like three-dimensional conformal radiation therapy (3DCRT), intensity modulated radiation therapy (IMRT), image guided radiation therapy (IGRT), tomotherapy, and stereotactic radiation therapy/surgery (SRT/SRS). These high-precision radiation delivery methods allow better dose distributions within the targeted tumor volume while sparing a larger portion of adjacent normal tissues. Success of these RT techniques requires accurate tumor volume delineation, tumor characterization, and response assessment during and after treatment. Conventionally these tasks are achieved through anatomical imaging (CT, MRI, and US). Of late PET has been increasingly used in high-precision RT for tumor volume delineation and characterization, because PET brings in the crucial functional and molecular information which enable the direct evaluation of tumor metabolism, cell proliferation, apoptosis, hypoxia and angiogenesis. This is a significant advance in cancer imaging with great potential for optimizing RT treatment planning and to the management of cancer patients. The availability of PET and CT in a single imaging system (PET/CT) to obtain a fused anatomical and functional dataset has made the applicability of PET in radiation oncology clinics much easier (Yap et al., 2004). The most recent introduction of PET-MR instrumentation dedicated for concurrent high resolution brain imaging is now revolutionizing the use of multimodality imaging in tumor brain imaging (Boss et al., 2010). Numerous reports are available in the

H. Zaidi and S. Senthamizhchelvan

literature in support of the routine use of PET for RT target volume delineation in non-small cell lung cancer (NSCLC), head and neck cancers, lymphoma and in esophageal cancers, with promising preliminary data in many other cancers (Gregoire et al., 2007; Nestle et al., 2009; Zaidi et al., 2009). The focus of this chapter is to update the readers on the potential use of PET imaging in the management of high-grade gliomas (HGG), with particular emphasis being given to the role of PET in the assessment of biological target volumes (BTVs) for RT.

PET-Guided Biological Target Volume (BTV) Delineation Over the past two decades radiation oncology community has seen a paradigm shift from 2D treatment planning to 3D conformal treatment planning for RT. One of the major advantages of 3D treatment planning and associated treatment delivery techniques is the leverage offered for the dose escalation to tumor volumes while preserving tolerance doses for normal structures. The state-of-the-art 3D-CRT techniques (IMRT, IGRT, tomotherapy, volumetric arc therapy and SRS/SRT etc.) developed to deliver highly conformal radiation beams directed towards targets warrant equally precise imaging modalities for accurate delineation of the tumor extent. Customizing dose delivery to various parts of the treatment volumes (“dose painting” and/or “dose sculpting”) based on their dose requirements are possible by 3D-CRT techniques today (Ling et al., 2000). However it is important to know what needs to be “painted”, and how much “paint” is required to take complete advantage of these highly conformal radiation delivery techniques (Rickhey et al., 2010). The principal objective of all form of radiation delivery techniques is to achieve highest possible tumor dose without exceeding the dose level of surrounding normal tissue toxicity. This is achieved in RT today by selective dose escalation also known as “dose painting” with sharp dose gradient along the tumor boundary. The rationale for dose painting is that, treating a selective tumor region to high dose should result in higher tumor control probability and lower normal tissue complication probability. The expected outcome of 3D-CRT is minimal complications and side

15 Assessment of Biological Target Volume Using PET in HGG Patients

effects, while achieving maximum possible dose to tumor tissues. Accurate delineation of the tumor volumes and assessment of the tumor response to the ongoing treatment regimens are necessary for successful implementation of modern RT techniques. In order to achieve the above said goals it is highly desirable to precisely locate and visualize metabolic tumor extensions and delineate boundaries. Volumetric patient imaging forms the basis for 3D-CRT treatment planning and also helps in the design of radiation fields and dose distributions. Since 3D imaging is used to delineate target volumes and normal structures, the quality of the imaging and information obtained from the images have a direct impact on the patient treatment and potentially on the outcomes and complications. Traditionally, anatomical imaging like CT and MRI are used for radiotherapy treatment planning, monitoring and follow-up evaluation. The major advantage of anatomical imaging is their high resolution, which enables clear visualization of morphological changes. However, the precise delineation of tumor regions with anatomical imaging has some significant limitations. The CT and MRI measure the differential density and magnetic properties of the tissues respectively, both of which may not necessarily be tumor specific characteristics. The morphological changes similar to malignancy can occur due to other confounding factors like infection, treatment (radiotherapy, chemotherapy and surgery etc.) induced inflammation, which makes the distinction between tumor biology from other pathological conditions difficult. Moreover biological changes manifest first and the time frame for the development of detectable morphological changes is too long, during which the disease might have progressed to an advanced stage. The applicability of anatomical imaging for precise delineation of tumor extension is hampered, when tumor harboring sites have unchanged morphology, density and magnetic properties similar to normal tissue. Successful treatment planning requires information about tumor biological characteristics such as proliferation and hypoxia and the ability to distinguish treatment related scar, edema and necrosis from malignant cells (Sun et al., 2011). Anatomical imaging modalities lack the aforementioned qualities which make them insufficient for the delineation of target volume in most of the clinical presentations and as well treatment response evaluation. Even though PET has

133

lower spatial resolution in comparison to its anatomical imaging counterparts the sensitivity of PET for detecting tumor biology is high with required activity concentration of nmol to pmol range to detect biological signals (Nestle et al., 2009). A typical example where PET plays an important role in defining biologically active tumor volume over anatomical imaging is shown in Fig. 15.1. The term “biological imaging” was coined by (Ling et al., 2000) and has become popular in radiation oncology practice. Biological imaging for cancer detection, staging, therapy monitoring and follow up has gained vast interest as the evidence of treatment success with this modality is keep accumulating. The diagnostic utility of PET as the most prominent biological imaging system is evident from the increased interest and utility of PET imaging for cancer imaging and therapy planning for the last few years. The term “biological target volume” (BTV) was also first proposed by (Ling et al., 2000). It is a new method of defining tumor volumes based on metabolism, physiology and molecular biology of tissues. The definition of BTV, apart from the standard definitions of tumor volumes as gross tumor volume (GTV), clinical target volume (CTV) and planning target volume (PTV), is aimed at providing information on the location and extent of tumor margins based on tumor biology. The BTV should also be helpful in assessing the biological response of tumor to therapy. In radiotherapy planning, biological imaging will also guide the radiation oncology community in defining biologically interesting sub-volumes (field in field) of the tumor like a target within the GTV, which could be irradiated with a higher dose through conformal RT (Nestle et al., 2009). Figure 15.2 shows clinical examples comparing biologically defined tumor volumes with anatomically defined tumor volumes. Till date, many PET tracers have been evaluated diagnostically in different cancers. New PET tracers are being developed for staging, RT planning, treatment monitoring and response evaluation during and after completion of RT has gaining more attention in biological imaging.

Techniques for PET-Guided Biological Target Volume (BTV) Delineation Traditionally CT and MRI have been used as primary imaging modalities for radiation therapy treatment

134

H. Zaidi and S. Senthamizhchelvan

A

B

D

E

C

F

Fig. 15.1 Example of a patient with a glioblastoma (WHO IV) in the left temporal and frontal areas. The images shown on the top row (temporal area) correspond to gadolinium enhanced T2-weighted MRI (A), coregistered 18 F-FET (B) and fused PET/MR (C) of the first study. The same is shown in the bottom

Patient#1

row for the same study in the frontal area (D, E and F). The 18 FFET PET study revealed an additional lesion missed on MRI. In addition, the T2-weighted MRI and the 18 F-FET PET show substantially different gross tumor volume extension for radiation therapy treatment planning

Patient#2

Patient#3

BTVPET GTVMR Fig. 15.2 Biological (BTV, blue) and morphological gross tumour (GTV, red) volume defining the clinical target volume in patients with high-grade glioma. Note the common volume between the tumour volumes (yellow chicken wire). Good

BTV-GTV matching is shown (left) in 1 patient, while substantial BTV-GTV mismatch is also detailed (center and right) in 2 other patients. Adapted from Weber et al. (2008)

15 Assessment of Biological Target Volume Using PET in HGG Patients

planning. X-ray computed tomography and MRI offer excellent spatial resolution, and soft tissue contrast, but both imaging modalities fail to provide functional properties of the imaged tissues. In cases where the true extent of the disease may extend beyond anatomically defined volumes, despite its inferior resolution compared with CT and MRI, PET has been shown valuable for defining the extent of target volumes. Some of the key contributions of PET in addition to or in combination with other imaging modalities are delineation of tumor volumes, biological characterization of tumor, and assessment of treatment response. With the widespread adoption of hybrid PET/CT scanners, in radiotherapy clinics, PET-based delineation of target volumes appears to be an attractive option in RT treatment planning. One of the most difficult issues facing PET-based RT treatment planning is the accurate delineation of target regions from typical noisy functional images. The major problems encountered in functional volume quantification are image segmentation and imperfect system response function. Image segmentation is defined as the process of classifying the voxels of an image into a set of distinct classes. The difficulty in image segmentation is compounded by the low spatial resolution and high-noise characteristics of PET images. Medical image segmentation has been identified as the key problem of medical image analysis and remains a popular and challenging area of research. Despite the difficulties and known limitations, several image segmentation approaches have been proposed and used in clinical setting including thresholding, region growing, classifiers, clustering, edge detection, Markov random field models, artificial neural networks, deformable models, atlas-guided, and many other approaches (Zaidi and El Naqa, 2010). Medical image segmentation remains an unsolved problem that has captured the imagination of image analysis scientists over the past three decades. Manual segmentation methods available on most commercial software packages to identify lesion boundaries and to quantify GTVs in terms of standardized uptake value are very laborious and tedious. They discourage physicians from taking advantage of the inherently quantitative data and compel them to use qualitative means in their diagnosis, therapy planning, and assessment of patient response to therapy. Semi- or fully automated segmentation methods enable physicians to easily extract maximum and mean standardized uptake value estimates from a lesion volume. This also allows

135

the physician to track changes in lesion size and uptake after radio/chemotherapy. At present, various methods are used in practice to delineate PET-based target volumes. Manual delineation of target volumes using different window level settings and look up tables is the most common and widely used technique in the clinic. However, the method is highly operatordependent and is subject to high variability between operators. Rather large intra-observer variability was reported for many localizations including HGG as shown in Fig. 15.3 (Weber et al., 2008). In this respect, semi- or fully-automated delineation techniques might offer several advantages over manual techniques by reducing operator error/subjectivity, thereby improving reproducibility.

Assessment of PET-Guided Biological Target Volumes in High-Grade Gliomas (HGG) Accurate determination of the tumor boundary at the microscopic level, assessment of the tumor sensitivity, and/or prognosis to the therapy is essential for successful treatment planning. Glioma cells are one of the most treatment resistant cells and their inherent heterogeneous cell populations, diffuse infiltration into normal brain tissues are the biggest challenge in the tumor localization and precise delineation of tumor extent. Prognosis of cerebral gliomas has continued to remain poor for several decades, albeit significant advances in multimodality diagnostic and therapeutic procedures. Therefore, development of a highly specific and sensitive non-invasive imaging modality is required. The ability to closely correlate diagnosis with pathology, distinguish inflammation and necrosis from tumors, differentiate tumor grades, accurately delineate tumor volume, monitor treatment responses, and identify residual tumor/recurrence are the desirable goals for imaging brain tumors to improve their clinical management. The specificity of anatomical imaging modalities in distinguishing neoplastic disease from vascular or inflammatory processes can be problematic in highgrade gliomas. Treatment effects including surgical trauma, corticosteroid-induced reduction of edema and contrast enhancement, and radionecrosis cannot

136

H. Zaidi and S. Senthamizhchelvan Interobserver Varibility in BTV delineation

80.0

70.0

60.0

Volume cm3

50.0

40.0

Observer #1 Observer #2 Observer #3

30.0

20.0

10.0

0.0 1

2

3

4

5

6

7

8

9

10 11 Case No.

12

13

14

15

16

17

18

19

Fig. 15.3 Biological tumor volume measurements by three observers for each high-grade glioma case (1 through 19)

always reliably be distinguished from tumor recurrence or response to therapy. Molecular imaging allows a better understanding of pathology at molecular level. This ability is especially useful in the brain tumors where tissue sampling in vivo is associated with significant risks. Availability of a suitable imaging modality or a multimodality combination to obtain the information of interest noninvasively is also vital for the basic research and development of novel and effective experimental therapeutics required to improve prognosis. Accurate quantitative information on the metabolic state of glioma tumor cells can be achieved through biological imaging using PET. Depending on the radiotracer used, various molecular processes can be visualized through PET imaging, most of them relating to an increased cell proliferation, metabolic rates, and DNA synthesis as well as abnormal microvessel density and thereby define tumor extent better than morphologic imaging in malignant gliomas (Tsien et al., 2009). PET radiotracers are especially helpful (1) in the localization, grading and finding the extent of glioma cells; (2) in the identification of metabolically active residual tumor after therapy; (3) monitoring of tumor progression; and (4) most importantly in the differentiation between recurrent tumor and radiation

necrosis. Various metabolism and biochemical pathways are exploited by PET tracers for glioma imaging. Energy metabolism of cells is imaged by [F-18]-2fluoro-2-deoxyglucose (FDG). Amino acid transport and incorporation of tumor cells are tracked by L -methyl-[C-11]methionine (MET), L -[C-11]tyrosine, L -[F-18]fluorotyrosine. DNA synthesis is imaged by 2-[C-11]thymidine, methyl-[C-11]thymidine, [F-18]-3 deoxy-3 -fluorothymidine (FLT). Cell membrane/lipid biosynthesis is tracked by 1-[C-11]acetate, [C-11]choline, [F-18]fluorocholine. Hypoxia is an important aspect to consider for assessing the aggressiveness of tumor and predicting the outcome of therapy (Sun et al., 2011). Tumor hypoxia is imaged by [F-18]fluoromisonidazole (FMISO) and many other tracers. The preferential uptake of malignant glioma cells in comparison to normal cells are exploited by tracers like 18 F-FDG, 11 C-MET, 18 F-FET and 18 F-FLT, depending on the tumor grade as a reflection of increased activity of membrane transporters for amino acids (11 CMET and 18 F-FET) and nucleosides (18 F-FLT) as well as increased expression of cellular hexokinase (18 FFDG) and thymidine kinase (18 F-FLT) genes, which phosphorylate 18 F-FDG and 18 F-FLT, respectively.

15 Assessment of Biological Target Volume Using PET in HGG Patients

Many hypoxia tracers (18 F-FMISO, 18 F-FAZA, 64 CuATSM and 18 F-EF5) have already shown their importance in target volume delineation or patient management in RT (Grosu et al., 2005a). The most commonly available PET tracer 18 F-FDG has the potential to detect abnormal metabolic rate, through increased cellular glucose metabolism in brain tumors. However, its use in target definition is complicated by the high level of intrinsic glucose uptake in the brain. FDG imaging is useful in distinguishing low-grade gliomas (LGG) from HGG based on tumorto-cortex (T/C) uptake ratio and tumor-to-white matter (T/WM) uptake ratio. Selecting the optimum site for tumor biopsy can be done based on the maximum uptake of FDG for sampling of the most malignant areas of tumors. For assessing response by pre- to posttreatment comparisons, FDG appears to be limited in clinical usefulness. The ability of 18 F-FDG PET to differentiate recurrent tumor from radiation necrosis is also limited. The false-positive and false-negative FDG-PET could result in unacceptably low sensitivity, specificity, and negative predictive values. The goal of PET imaging with radio-labeled amino acids is to assess the protein synthetic process of tumor growth. Amino acid uptake in normal brain tissues is low relative to FDG uptake so that the tumor to normal tissue contrast is better with amino acid imaging than with FDG. Radiolabeled amino acids can also penetrate the blood–brain barrier independently of its disturbance. A variety of 11 C- and 18 F-labeled amino acids such as 11 C-methionine (11 C-MET) and 18 F-fluoro-ethyl-tyrosine (18 F-FET) have been studied for potential use in oncologic PET. Most brain tumors show an increased uptake of amino acids compared with normal brain tissue. In particular, the uptake of 18 F-FET by brain tumors especially by high-grade glioma cells is intense relative to the low uptake in normal cerebral tissue and has shown the potential in the detection of primary and recurrent brain tumors with high sensitivity and specificity. Compared with 11 C-MET, 18 F-FET PET findings in brain tumors are similar. One of the advantages of 18 F-FET over 11 C-MET is that the half life, which makes it possible to be used in clinics not having on-site cyclotron. 11 C- methionine PET and 18 F-fluorothymidine (FLT), provide better differentiation of the tumor from brain background signals than 18 F-FDG PET. The 11 C-methionine PET scan reflects metabolic activity through increased transport of amino acid carriers

137

at the level of the blood-brain barrier that is highly expressed in malignant tumors compared with low uptake in the normal brain. On the other hand FLT, is a thymidine analog that is incorporated exclusively into DNA. It measures the activity of cellular thymidine kinase, which increases several-fold as cells enter the S-phase and begin DNA synthesis. Increased FLT uptake and therefore thymidine kinase levels provide a direct measure of the cellular proliferation rate. Tumor hypoxia remains the most challenging condition for treatment. Though oxygen metabolism in gliomas differs from that of normal brain tissue, the lack of oxygen appears to be an important factor in determining glioma aggressiveness and response to therapy. It has been documented in several types of cancers that low levels of oxygen tension are associated with persistent tumor following RT and with the subsequent development of local recurrences. In gliomas, spontaneous necrosis suggests the presence of hypoxic regions that are radioresistant. 18 F-FMISO imaging of hypoxic glioma cells shows significant promise, however larger patient population studies are required to ascertain its clinical impact. Identifying the regional distribution of hypoxia may improve planning of resections and allow targeting higher doses of radiotherapy more precisely to the hypoxic areas. Glioma cell membrane biosynthesis is imaged using 11 C-acetate and 11 C-choline. The rationale for imaging membrane and lipid biosynthesis is that tumor growth requires both of these processes in parallel with DNA and protein synthesis. These will likely show retention in tumor tissue but not by gray matter, an important advantage over FDG. A very few studies have been done so far on this front to show the potential advantages of this approach.

PET Imaging for Differentiating Recurrent Brain Tumor from Radiation Necrosis Radiation damage to vascular endothelial cells and oligodendrocytes causes necrosis. Differentiating tumor growth from post-treatment radiation effect (PTRE) remains a common challenge in high-grade glioma tumors. On MRI, appearances of radiation necrosis and of recurrent tumor are quite similar, as both causes’ areas of increased signal intensity. Conventional MRI/MRSI are currently used for the

138

detection of early treatment-bed changes, though accurate diagnosis is challenging because tumor growth, PTRE, and admixed lesions can all have identical MR imaging appearances. Microscopic tissue analysis distinguishes these entities and can document intra-lesion heterogeneity by resolving distinct subregions of tumor from pure PTRE within different locations of the same lesion. Early work on utilizing PET in differentiating radiation induced necrosis from recurrent brain tumor was conducted by Patronas and coworkers using FDG (Patronas et al., 1982). The rationale for using FDG is that radiation necrosis is expected to show decreased uptake in comparison with recurrent tumors. However, in many cases, distinguishing recurrent tumor from radiation necrosis is found to be difficult based on FDG-PET alone (Hustinx et al., 2005). Radiolabeled amino acid analogues like 11 C-MET, 18 F-FET and proliferation marker 18 F-FLT are suggested to perform better in PTRE evaluation, than FDG in detecting residual and recurrent tumors after fractionated irradiation (Hustinx et al., 2005; Reinhardt et al., 1997). Though the exact incidence of true radiation necrosis is largely unknown, differentiating it from recurrent tumor has a larger clinical implication in the clinical management of patients and PET tracers seem to play a major role in it.

Current Practice of Target Volume Delineation in HGG Conventional target volume definitions in high grade gliomas (HGG) have not incorporated PET. The gross tumor volume (GTV) is mostly defined on postoperative MRI and includes the contrast-enhancing lesion as well as the surgical cavity (Fueger et al., 2010). Clinical target volumes (CTVs) include an MRI defined volume with an addition of uniform margin (about 2 cm) within the brain, which includes areas of microscopic extension and peritumoral edema. When therapy is planned for a second CTV (boost volume) a margin of typically 1.5 cm around the GTV is added to account for areas of microscopic disease. These CTVs are then further expanded by a uniform margin of 0.3–0.5 cm to create a PTV to account for treatment setup uncertainties. Initial imaging for HGG is usually done with CT or MRI, which provides accurate information about lesion anatomy

H. Zaidi and S. Senthamizhchelvan

and location. However, follow-up assessment of primary HGG tumors after radiation therapy, chemotherapy and surgery, is often difficult, since the anatomical imaging modalities are usually not able to differentiate recurrent tumor from radiation necrosis, surgical scar or inflammation. Identifying radiation necrosis and differentiating them from tumor recurrence pose a potential diagnostic challenge because the accurate diagnosis has important implications for the patient management.

PET-Guided Biological Target Volumes in HGG Grosu and co-workers have investigated the use of amino-acid PET and single-photon emission computed tomography (SPECT) in gross tumor volume definition for radiotherapy treatment planning of gliomas (Grosu et al., 2005a, c). They have shown that 11 CMET-PET offers significant additional information about tumor extent in HGG, compared to CT and MRI alone. This study suggested that integration of amino-acid PET in target volume definition might contribute to an improved outcome in HGG patients treatment. It was also showed that abnormal MET PET activity was detected beyond the area of the contrast-enhancing lesion on MRI. In a retrospective study on newly diagnosed glioblastoma multiforme (GBM) who underwent MET PET before radiation, the area of MET uptake was found to be larger than the contrast-enhancing gadolinium volume in 29 (74%) of 39 patients (Grosu et al., 2005c). In the same study it was showed that patients who underwent treatment planning based on 11 C-MET PET/SPECT imaging had improved survival compared with treatment planning based on CT/MRI in recurrent gliomas, with a median survival of 9 months versus 5 months, respectively. Other investigators have also suggested that 11 C-MET PET has the potential to improve target volume definition in the radiation treatment planning of high-grade gliomas by identifying residual tumor after resection and also in recurrent gliomas. Yamane et al. (2010) have recently showed that 11 C-MET PET can provide useful information in initial diagnosis and differentiating tumor recurrence from radiation necrosis. They have also claimed that use of 11 C-MET PET has changed the intended clinical management in half of the patients.

15 Assessment of Biological Target Volume Using PET in HGG Patients

Our group (Vees et al., 2009) has studied the contribution of 18 F-FET PET in the delineation of GTV in HGG patients compared with MRI alone. In this study PET based tumor volumes were delineated in 18 patients using seven image segmentation techniques. The PET image segmentation techniques included manual delineation of contours (GTV(man) ), a 2.5 standardized uptake value (SUV) cutoff (GTV(2.5) ), a fixed threshold of 40 and 50% of the maximum signal intensity (GTV(40%) and GTV(50%) ), signal-to-background ratio (SBR)-based adaptive thresholding (GTV(SBR) ), gradient find (GTV(GF) ), and region growing (GTV(RG) ). Figure 15.4 shows an example of image segmentation using these techniques. Overlap analysis was also conducted to assess geographic mismatch between the GTVs delineated using the different techniques. Contours defined using GTV(2.5) failed to provide successful delineation technically in three patients (18% of cases) as SUV(max) 36 months 12 months

13 months

17 months

13 months

NM

8 months

5 months

Survival

16 Skin Metastases of Glioblastoma 145

L, fronto-parietal L, frontal

13, M

58, F 41, M

48, F

53, F

Mentrikoski et al. (2008)

Senetta et al. (2009)

L, frontotemporal

L, frontal NM

Therapy

Craniotomy

Craniotomy, RT, CT

Craniotomy, RT, CT (TMZ) Craniotomy, RT and intratumoral seed implants Craniotomy, RT, CT

Craniotomy, RT, Local CT with DTI-015 Stereotactic biopsy, RT, CT (2 cycles of BCNU) Craniotomy, RT, CT (TMZ)

Craniotomy, RT Craniotomy, RT (55.8 Gy)

7 months

14 months

14 months

16 months 2 months

NM

11 months

12 months

12 months 48 months

Timing to cutaneous progression

Yes

No

No

Yes NM

Yes

Yes

Yes

Yes Yes

Intracranial progression

Parotid gland, LN, bone (L4) No

No

Leptomeninges, lung, liver No No

Axial skeleton

No LN, leptomeniges No

Additional metastatic organs

FNAc

Total removal and focal RT Biopsy and focal RT

Biopsy NM

FNAc/partial excision CT

Excision

No treatment CT

10 months

25 months

26 months

NM NM

10 months

12 months

13 months

14 months 52 months

Diagnosis/therapy of cutaneous lesions Survival

NM not mentioned, F female, M male, L left, R right, RT radiotherapy, CT chemotherapy, TMZ temozolomide, FTM fotemustine, PC procarbazine and CCNU, LN limph nodes, FNAc fine-needle aspiration cytology

63, M

L, frontal

60, F

Bouillot-Eimer et al. (2005) Saad et al. (2007)

Miliaras et al. (2009)

L, parietal

74, F

Schultz et al. (2005)

Supra-tentorial L, Temporooccipital L, temporal

Allan (2004) Moon et al. (2004)

Tumor location

Age Sex

60, M 35, F

Author, year

Table 16.1 (continued)

146 R. Senetta and P. Cassoni

16 Skin Metastases of Glioblastoma

147

Fig. 16.1 Histopathology and immunophenotype shift in one case skin metastatic GBM. The intracranial GBM showed (at haematoxylin and eosin) a marked hypercellularity with an area of classical pseudo-palisade necrosis and prominent glomeruloid vascular proliferation (a). At immunohistochemistry, a diffuse

and intense GFAP reactivity was present (b). The skin metastatic tumour was composed of small neoplastic cells with a scant cytoplasm and hyper-chromatic nuclei (c) with only few neoplastic elements positive to GFAP (d) and an intense vimentin immunoreactivity (e)

mass appeared after 6 cycles of chemotherapy, in presence of a partial response of the intracranial tumour (reduction of about 90% of the enhancing area). This observation induced us to verify in these two cases if primary and metastatic lesions still shared a main phenotypical consistency or if the skin localization acquired a prevalent divergent phenotype (Fig. 16.1). In fact, the immunophenotypical profile of cutaneous tumours revealed a strongly reduced GFAP and EGFR expression, paralleled by an increased vimentin and YKL-40 (which is a marker of radio-resistance in GBM) (Pelloski et al., 2005) staining at IHC (Senetta et al., 2009). Other authors similarly observed an intense vimentin-immunoreactivity in the cutaneous GBM sites associated either to a focal (as in our cases) or a robust GFAP expression (Jain et al., 2005; Mentrikoski et al., 2008; Schultz et al., 2005; Senetta et al., 2009). As previously discussed in our paper (Senetta et al., 2009), these changes altogether may suggest a “glial-to-mesenchymal transition” of the cutaneous metastases: the newly acquired immunophenotype could account for the therapy resistance of the skin lesions in contrast to the synchronous responsiveness of the intracranial GBM. This hypothesis is in agreement with previous evidence of the overexpression of mesenchymal and angiogenesis related genes in GBMs with more aggressive behaviour (Phillips et al., 2006) and could then be extended to their metastatic spread: still, it should be understood if such shift towards a mesenchymal de-differentiation in skin lesions can be a spontaneous event in the pathway of tumour progression or if

radiotherapy could favour the sarcomatous transformation and/or the selection of resistant “mesenchymal” cell clones (Burger et al., 1979; Schiffer et al., 1990).

Diagnostic and Therapeutic Approaches to Skin Metastases of GBM As summarized in Table 16.1, the diagnosis of a cutaneous dissemination of GBM is mainly performed on a biopsy, a partial resection or a total excision of the skin lesion. In two cases a fine-needle aspiration cytology (FNAc) approach was used, and the diagnosis was then re-confirmed on the tumour excision histology (Schultz et al., 2005; Miliaras et al., 2009). Some authors highlighted the importance of patient neoplastic history (whether intracranial or extra-neural) for an accurate diagnosis (Schultz et al., 2005; Mentrikoski et al., 2008): without a knowledge of intracranial tumour story, the differential diagnosis must include, above all, a melanoma, followed by a sarcomatoid squamous cell carcinoma and atypical fibroxanthoma (Mentrikoski et al., 2008). In fact, a cutaneous lesions with atypical morphological appearance at histology (spindle-shape cells, nuclear pleomorphism, irregular nuclear membranes, prominent nucleoli and mitoses) associated to immunohistochemical positivity for S-100 protein (as frequently described in GBMs) (Bouillot-Eimer et al., 2005; Figueroa et al., 2002; Mentrikoski et al., 2008) should first be considered as a possible melanoma localization, even in presence of

148

a known history of GBM. Schultz and colleagues listed the cytological features of malignant neoplasm that can arise in the scalp and that must be distinguished from a metastatic GBM, when an anamnesis of glioma is present (Schultz et al., 2005). The treatment of GBM scalp metastases can consist in surgery alone (with a partial or total lesion excision with surrounding subcutaneous and bone tissue), or in surgery in combination to chemotherapy or radiotherapy, or in chemotherapy and radiotherapy alone. In the two cases of GBM skin metastases that we reported, total surgery+local radiotherapy (total dose: 40 Gy) proved to be more effective in preventing further cutaneous relapses compared to biopsy+local radiotherapy (total dose: 55 Gy)+6 cycles of Temozolomide. Therefore, in our experience, a radical surgical treatment seems to allow the best local tumour control (Senetta et al., 2009). In conclusion, GBM metastases are rare events, and therefore do not represent an ordinary clinicaloncologic concern: however, comprehension of the mechanisms involved in promoting their development could bear relevance in elucidating GBM biology overall, and considered as a unique model to focus on the crucial steps necessary to GBM progression.

References Allan RS (2004) Scalp metastasis from glioblastoma. J Neurol Neurosurg Psychiatry 75:559 Astner ST, Pihusch R, Nieder C, Rachinger W, Lohnerm H, Tonn JC, Molls M, Grosu AL (2006) Extensive local and systemic therapy in extraneural metastasized glioblastoma multiforme. Anticancer Res 26:4917–4920 Bauchet L, Mathieu-Daudé H, Fabbro-Peray P, Rigau V, Fabbro M, Chinot O, Pallusseau L, Carnin C, Lainé K, Schlama A, Thiebaut A, Patru MC, Bauchet F, Lionnet M, Wager M, Faillot T, Taillandier L, Figarella-Branger D, Capelle L, Loiseau H, Frappaz D, Campello C, Kerr C, Duffau H, Reme-Saumon M, Trétarre B, Daures JP, Henin D, Labrousse F, Menei P, Honnorat J; with the participation of Société Française de Neurochirurgie (SFNC) and the Club de NeuroOncologie of the Société Française de Neurochirurgie (CNOSFNC) and Société Française de Neuropathologie (2010) Oncological patterns of care and outcome for 952 patients with newly diagnosed glioblastoma in 2004. Neurooncol. 12:725–735 Bernstein JJ, Woodard CA (1995) Glioblastoma cells do not intravasate into blood vessels. Neurosurgery 36:124–132 Bouillot-Eimer S, Loiseau H, Vital A (2005) Subcutaneous tumoral seeding from a glioblastoma following stereotactic biopsy: case report and review of the literature. Clin Neuropathol 24:247–251

R. Senetta and P. Cassoni Burger PC, Mahley MS Jr, Dudka L, Vogel FS (1979) The morphologic effects of radiation administered therapeutically for intracranial gliomas. A post-mortem study of 25 cases. Cancer 44:1256–1272 Campora RG, Salaverri CO, Ramirez FV, Villadiego MS, Davidson HG (1993) Metastatic glioblastoma multiforme in cervical lymph nodes: report of a case with diagnosis by fine needle aspiration. Acta Cytol 37:938–942 Cervio A, Piedimonte F, Salaberry J, Alcorta SC, Salvat J, Diez B, Sevlever G (2001) Bone metastases from secondary glioblastoma multiforme: a case report. J Neurooncol 52:141–148 Datta CK, Weinstein JD, Bland JE, Brager PM, Stewart MA (1998) A case of cervical lymph node metastasis resulting from glioblastoma multiforme. W V Med J 94:276–278 Figueroa P, Lupton JR, Remington T, Olding M, Jones RV, Sekhar LN, Sulica VI (2002) Cutaneous metastasis from an intracranial glioblastoma multiforme. J Am Acad Dermatol 46:297–300 Hata N, Katsuta T, Inoue T, Arikawa K, Yano T, Takeshita M, Iwaki T (2001) Extracranial metastasis of glioblastoma to the lung and the heart with a histological resemblance to small cell carcinoma of the lung: an autopsy case. No Shinkei Geka 29:433–438 Houston SC, Crocker IR, Brat DJ, Olson JJ (2000) Extraneural metastatic glioblastoma after interstitial brachytherapy. Int J Radiat Oncol Biol Phys 48:831–836 Hsu E, Keene D, Ventureyra E, Matzinger MA, Jimenez C, Wang HS, Grimard L (1998) Bone marrow metastasis in astrocytic gliomata. J Neurooncol 37:285–293 Jain N, Mirakhur M, Flynn P, Choudhari A (2005) Cutaneous metastasis from glioblastoma. Br J Neurosurg 19:65–68 Laraqui L, Amarti A, Zouiadia F, Maher M, Kettani F, Saidi A (2001) Pulmonary metastasis from glioblastoma: a case report. Rev Pneumol Clin 5713:225–228 Louis DN, Ogaki H, Wiestler OD, Cavanee WK (2007) WHO classification of tumours of the central nervous system, 4th edn. IARC Lyon Press, Lyon Matsuyama J, Mori T, Hori S, Nakano T, Yamada A (1989) Gliosarcoma with multiple extracranial metastases. Case report. Neurol Med Chir 29:938–943 Mentrikoski M, Johnson MD, Korones DN, Scott GA (2008) Glioblastoma multiforme in skin: a report of 2 cases and review of the literature. Am J Dermatopathol 30: 381–384 Miliaras G, Tsitsopoulos PP, Markoula S, Kyritsis A, Polyzoidis KS, Malamou-Mitsi V (2009) Multifocal glioblastoma with remote cutaneous metastasis: a case report and review of the literature. Cen Eur Neurosurg 70:39–42 Moon KS, Jung S, Lee MC, Kim IY, Kim HW, LEE JK, Kim TS (2004) Metastatic glioblastoma in cervical lymph node after repeated craniotomies: report of a case with diagnosis by fine needle aspiration. J Korean Med Sci 19:911–914 Mourad PD, Farrell L, Stamps LD, Chicoine MR, Silbergeld DL (2005) Why are systemic glioblastoma metastases rare? Systemic and cerebral growth of mouse glioblastoma. Surg Neurol 63:511–519 Mujic A, Hunn A, Taylor AB, Lowenthal RM (2006) Extracranial metastases of a glioblastoma multiforme to the pleura, small bowel and pancreas. J Clinic Neurosci 13:677–681

16 Skin Metastases of Glioblastoma Newton HB, Rosenblum MK, Walker RW (1992) Extraneural metastases of infratentorial glioblastoma multiforme to the peritoneal cavity. Cancer 69:2149–2153 Pelloski CE, Mahajan A, Maor M, Chang EL, Woo S, Gilbert M, Colman H, Yang H, Ledoux A, Blair H, Passe S, Jenkins RB, Aldape KD (2005) YKL-40 expression is associated with poorer response to radiation and shorter overall survival in glioblastoma. Clin Cancer Res 11: 3326–3334 Phillips HS, Kharbanda S, Chen R, Forrest WF, Soriano RH, Wu TD, Misra A, Nigro JM, Colman H, Soroceanu L, Williams PM, Modrusan Z, Feuerstein BG, Aldape K (2006) Molecular subclasses of high-grade glioma predict prognosis, delineate a pattern of disease progression, and resemble stages in neurogenesis. Cancer Cell 9:157–173 Rajagopalan V, El Kamar FG, Thayaparan R, Grossbard ML (2005) Bone marrow metastases from glioblastoma multiforme: a case report and review of the literature. J Neurooncol 72:157–161 Saad AG, Sachs J, Turner CD, Proctor M, Marcus KJ, Wang L, Lidov H, Ullrich NJ (2007) Extracranial metastases of glioblastoma in a child: case report and review of the literature. J Pediatr Hematol Oncol 29:190–194

149 Sadik AR, Port R, Garfinkel B, Bravo J (1984) Extracranial metastasis of cerebral glioblastoma multiforme: case report. Neurosurgery 15:549–551 Santos AV, Saraiva PF, Santiago B (2003) Extracranial metastasis of glioblastoma multiforme. Acta Med Port 16:209–211 Schiffer D, Chiò A, Giordana MT, Mauro A, Migheli A, Soffietti R, Vigliani MC (1990) Vascular response to irradiation in malignant gliomas. J Neurooncol 8:73–84 Schultz S, Pinsky GS, Wu NC, Chamberlain MC, Rodrigo AS, Martin SE (2005) Fine needle aspiration diagnosis of extracranial glioblastoma multiforme: case report and review of the literature. CytoJournal 2:19 Senetta R, Trevisan E, Rudà R, Benech F, Soffietti R, Cassoni P (2009) Skin metastases of glioblastoma in the absence of intracranial progression are associated with a shift towards a mesenchymal immunophenotype: report of two cases. Acta Neuropathol 118:313–316 Wallace CJ, Forsyth PA, Edwards DR (1996) Lymph node metastases from glioblastoma multiforme. Am J Neuroradiol 17:1929–1931 Yasuhara T, Tamiya T, Meguro T, Ichikawa T, Sato Y, Date I, Nakashima H, Ohmoto T (2003) Glioblastoma with metastasis to the spleen–case report. Neurol Med Chir (Tokyo) 43:452–456

Part II

Therapy

Chapter 17

Diffuse Low-Grade Gliomas: What Does “Complete Resection” Mean? Johan Pallud

Abstract Supratentorial hemispheric diffuse low-grade gliomas (LGG), i.e., World Health Organization grade II gliomas, are generally revealed by seizures in young adults with no or only mild neurological deficits. These progressive tumors are characterized by a continuous growth, by tumor recurrences and by a progression into a higher grade of malignancy. Maximal safe resection preserving eloquent brain areas, when possible, is currently considered as the optimal primary treatment modality of LGG. Imaging determinations of the spatial extent of LGG are of paramount importance in evaluating the risk-to-benefit ratio of surgical resection. However, it is not yet clear how accurately MRI can delineate LGG. Indeed, LGG may recur postoperatively even after a MRI-based complete resection and recurrences generally occurred in the resection margins. The value of conventional MRI in determining the spatial extent of LGG is thus questionable. As demonstrated by multi-scale correlative approaches with histological and imaging data, conventional MRI underestimates the actual spatial extent of LGG, even when they are well delineated on T2-weigthed and FLAIR sequences. Cycling isolated tumor cells are present beyond MRI-defined abnormalities and permeate surrounding “normal” parenchyma at sites up to at least 15 mm outside MRI-defined tumor limits. Clear tumor boundaries do not actually exist as LGG are diffusely infiltrative tumors with a decrease of tumor cell density as a function of distance from MRI-defined

J. Pallud () Service de Neurochirurgie, Hôpital Sainte-Anne, Paris Cedex 14, France e-mail: [email protected]

abnormalities. This implies that a MRI-based complete resection of a LGG leaves isolated tumor cells beyond the surgical field. These results should be considered when planning a surgical resection of a LGG: (1) a maximal safe resection preserving eloquent brain area is recommended because of the infiltrative nature of LGG and the frequent juxtaposition close to and/or within critically eloquent brain areas; (2) surgical resection should be tailored according to cortico-subcortical functional boundaries rather than MRI-based limits; (3) an extended resection of a margin beyond MRI-defined abnormalities, whenever feasible in non-eloquent brain areas, might increase the survival of patients harboring a LGG; (4) an early surgical treatment while the LGG is smaller might delay tumor progression by decreasing the number of residual isolated tumor cells. Keywords LGG · Complete resection · WHO · Grade II gliomas · Risk-to-benefit ratio · MRI

Introduction Supratentorial hemispheric diffuse low-grade gliomas (LGG), i.e., World Health Organization (WHO) grade II gliomas are generally revealed by seizures in young adults with a normal life and no or only mild neurological deficits (Duffau, 2005). However, these progressive tumors are characterized by a continuous growth (Mandonnet et al., 2003; Pallud et al., 2006), by tumor recurrences and by a progression into a higher grade of malignancy that is the major cause of mortality (Duffau, 2005; Pallud et al., 2006).

M.A. Hayat (ed.), Tumors of the Central Nervous System, Volume 2, DOI 10.1007/978-94-007-0618-7_17, © Springer Science+Business Media B.V. 2011

153

154

As a consequence, the better knowledge of the natural history of LGG resulted in an active therapeutic management and maximal safe resection preserving eloquent brain areas, when possible, is currently considered as the optimal primary treatment modality (Berger et al., 1994; Duffau, 2005, 2009; Sanai and Berger, 2008; Soffietti et al., 2010). Indeed, despite the lack of Class I evidence, there is growing evidence that the extent of resection is a significant prognostic factor for progression free and overall survivals (Sanai and Berger, 2008; Soffietti et al., 2010). The goal of surgery is thus to optimize the extent of resection while preserving the quality of life. Therefore, individual determinations of the spatial extent of LGG and of the cortical and subcortical functional organization are of paramount importance in evaluating the risk-to-benefit ratio of surgery (Duffau, 2009). The spatial extent of LGG is defined preoperatively on conventional MRI although it is not clear how accurately MRI can delineate LGG. Indeed, they may recur after a MRI-based complete resection and recurrences generally occurred at the site of the initial tumor, in the resection margins (Kelly, 1992, 2004). The value of MRI in determining the actual spatial extent of LGG is questionable. There is thus a need in assessing how MRI reflects their actual tumor limits. Such data are best provided by pathological analysis of spatially oriented surgical samples obtained from sites outside MRI-defined abnormalities. Studies are lacking for LGG as such samples performed are rarely available in a current neurosurgical practice. Therefore, the aim of the present chapter is to clarify the value of conventional MRI in delineating LGG and determining their actual spatial extent. The absence of a clear tumor boundary will be stressed at the light of histologic-imaging correlative studies focused on LGG.

Multiscale Approach of the Spatial Configuration of Cerebral Gliomas Microscopic Scale The understanding of the spatial configuration and growth pattern of a particular glioma is important for planning the best therapeutic strategy. The first

J. Pallud

data were acquired on postmortem analyses before the modern imaging era (Daumas-Duport and Szikla, 1981). They showed that cerebral diffuse gliomas: • Lack of a definite “tumor capsule” and spread diffusely throughout the brain in many cases. • May present without formation of a definite mass and without pronounced destruction of the preexisting tissue. • Present a microscopic tumor extension greater than that observed grossly with a tumor cell infiltration beyond the gross demarcation of the tumor. The brain invasion of diffuse gliomas was found to follow particular anatomical structures, such as myelinated fiber tracts, basement membranes of blood vessels or other basement membrane–like structures, subependyma and the perivascular spaces (Giese et al., 2003). The study of spatially oriented pathological samples obtained from image-based serial stereotactic biopsies of all components of the imaging abnormalities allowed the assessment of in vivo growth pattern of diffuse gliomas (Daumas-Duport et al., 1983, 1987; Daumas-Duport and Szikla, 1981; Kelly et al., 1987). Daumas-Duport et al. studied the histopathological tridimentionnal architectural pattern of diffuse gliomas and identified two distinct tumor components: • The “solid tumor tissue” component where the tumor cells are in contact with each other, containing microvascular proliferation and only residual brain parenchyma or none. • The “isolated tumor cells” component associated with edema but no microvascular proliferation, permeating functional brain parenchyma. This allowed proposing a spatial classification scheme with three tridimentionnal architectural subtypes for diffuse gliomas (Fig. 17.1) (Daumas-Duport et al., 1983, 1987; Daumas-Duport and Szikla, 1981; Kelly et al., 1987): • Type I: solid tumor tissue only without any peripheral isolated tumor cells. • Type II: solid tumor tissue and peripheral parenchyma infiltrated by isolated tumor cells. • Type III: brain parenchyma infiltrated by isolated tumor cells only without solid tumor tissue.

17 Diffuse Low-Grade Gliomas: What Does “Complete Resection” Mean?

Fig. 17.1 (a, b) Histopathological tridimentionnal architectural pattern of diffuse gliomas. The three different architectural subtypes comprise a “solid tumor tissue” component where the tumor cells are in contact with each other, containing microvascular proliferation and only residual brain parenchyma or none and an “isolated tumor cells” component associated with edema but no microvascular proliferation, permeating functional brain parenchyma. (Architectural subtype I: solid tumor tissue only without any peripheral isolated tumor cells; Subtype II: solid tumor tissue and peripheral parenchyma infiltrated by isolated tumor cells; Subtype III: brain parenchyma infiltrated by isolated tumor cells only without solid tumor tissue). (c, d, e, f) The WHO grade IV gliomas are architectural type II tumors.

155

A solid tumor tissue component is observed peroperatively (c). They comprised generally a contrast-enhanced mass on axial T1-weighted sequences after gadopentetate dimeglumine (d) surrounded by peripheral hypointensity on T1-weighted sequences (f) and by hyperintensity on FLAIR sequences (e). (g, h, i, j) The WHO grade II gliomas are architectural type III tumors. Isolated tumors cells infiltrate the brain parenchyma and expand the involved gyri, as observed peroperatively (h). They comprise generally an hypointensity on axial T1-weighted sequences (i) and an hyperintensity on axial FLAIR sequences (j) without definite contrast-enhancement on axial T1-weighted sequences after gadopentetate dimeglumine (g)

156

The identification of different architectural subtypes for diffuse gliomas has provided a guide for surgical decision-making. Indeed, the spatial architectural subtype III of LGG, implying a possible tumor invasion of functional brain areas, argues for the use of functional preoperative and peroperative mapping methods to optimize the benefit-to-risk ratio of surgical resection. In addition, this spatial classification scheme is generally correlated with WHO tumor grades: • Pilocytic astrocytomas are primarily type I tumors. • Grade II gliomas (LGG) are frequently type III tumors comprised only of a volume of brain parenchyma infiltrated by isolated tumor cells. However, there are some LGG that are type II because patches of tumor tissue exist within areas of infiltrated parenchyma. • The majority of higher-grade diffuse gliomas (WHO grade III and IV) are type II tumors; they have a central area of solid tumor tissue surrounded by a zone of peripheral infiltration.

Macroscopic Scale The MRI findings are correlated with the architectural subtypes of the spatial classification scheme described previously (Fig. 17.1): • Type I corresponds to a contrast-enhancing mass with no peripheral hypointensity on T1-weighted and no hyperintensity on T2-weighted or FLAIR sequences. • Type II corresponds to a contrast-enhanced mass surrounded by peripheral hypointensity on T1weighted and by hyperintensity on T2-weighted or FLAIR sequences. • Type III corresponds to an hypointense on T1weighted and hyperintense on T2-weighted or FLAIR sequences area without definite contrastenhancement. The architectural subtype of LGG explains the typical absence of contrast enhancement and why its occurrence is associated with a worsened prognosis (Pallud et al., 2009). Indeed, contrast enhancement should be considered as a key event in the malignant progression of LGG as it reflects macroscopically the microscopic neoangiogenesis.

J. Pallud

The tumor extension along myelinated white matter fiber tracts is easily demonstrated on MRI. Several reports have illustrated that glioma cells migrate along intra-hemispheric tracts, inter-hemispheric tracts and descending pathways and use them as preferential ways of invasion (Mandonnet et al., 2006; Pallud et al., 2005).

Dynamic Scale Mandonnet et al. first showed quantitatively the spontaneous radiological growth of LGG on successive MRIs in a subset of 27 patients that were followed before oncological treatment (Mandonnet et al., 2003). The study suggested a linear evolution of the mean tumor diameter over time (velocity of diametric expansion) quantified at a 4 mm/year average rate. These results were confirmed on a larger series of 143 LGG that ranged individual velocities of diametric expansion from 1 to 36 mm/year and demonstrated the strong prognostic significance of individual tumor growth rates on overall survival (Pallud et al., 2006). As other teams reported similar observations (Hlaihel et al., 2010; Peyre et al., 2010; Ricard et al., 2007), it is now accepted that LGG present a spontaneous and continuous radiological growth before any transformation into a higher grade of malignancy.

MRI-Based Estimation of the Spatial Extent of LGG Static Approach Conventional MRI is widely used in practice to study the anatomical extent of diffuse gliomas (Laster et al., 1984; Price, 2009). LGG appear as a hypointense area on T1-weighted sequence and hyperintense area on T2-weighted or FLAIR sequences. The margins of LGG is usually indistinct on T1-weighted sequence and is better demonstrated on T2-weighted and FLAIR sequences that show maximal visible imaging abnormalities (Bynevelt et al., 2001; Connor et al., 2004; Price, 2009). The hyperintense area is either with well-circumscribed margins or with indistinct borders,

17 Diffuse Low-Grade Gliomas: What Does “Complete Resection” Mean?

particularly on oligodendrogliomas with a loss of heterozygothy at chromosome 1p19q (Daumas-Duport et al., 1997; Jenkinson et al., 2006). It is suggested that Proton Magnetic Resonance Spectroscopic Imaging (H1 -MRSI) may be more valuable than conventional MRI to assess the spatial extent of diffuse gliomas, as abnormal metabolic areas may exceed areas of MRI-defined abnormalities (Croteau et al., 2001; Ganslandt et al., 2005). One previous report studied histological data of peripheral areas of gliomas that appeared normal on MRI and abnormal on H1 -MRSI and identified tumor infiltration, using proliferation markers, in four studied LGG (Ganslandt et al., 2005). However, this study could not provide reliable data regarding the true spatial extent of LGG from H1 -MRSI since it included gliomas of different histopathological subtypes and grades without imaging data, the biopsy samples were performed with a frameless stereotactic system within one centimeter from MRI-defined abnormalities. In addition, the MRSI techniques are currently limited by the poor resolution (Price et al., 2006). Studies have shown that diffusion tensor imaging reveals peritumoral abnormalities in diffuse gliomas that are not apparent on conventional MRI and that diffusion tensor imaging may delineate tumor margins better than conventional MRI (Price et al., 2006).

Dynamic Approach As LGG present a continuous growth, studying their radiological changes over time by the mean of repeated MRI is a simple way to follow the changes in the spatial extent of LGG (Mandonnet et al., 2008). Imaging changes are correlated with changes in the natural history of LGG and have been shown to be associated with a worsened prognosis, as they reflect the progression into a higher grade of malignancy:

• An elevated tumor growth rate with a velocity of diametric expansion at 8 mm/year or more (Pallud et al., 2006). • The change over time of a contrast enhancement pattern or the occurrence of a new contrast enhancement (Pallud et al., 2009).

157

Does MRI Reflect the Actual Spatial Extent of LGG? The actual spatial extent of LGG and their histological margins of tumor infiltration are difficult to determine. Previous studies have demonstrated that conventional MRI underestimate the spatial extent of malignant gliomas and that isolated tumor cells are observed well beyond MRI-defined abnormalities on T2-weighted sequence (Daumas-Duport et al., 1983, 1987; DaumasDuport and Szikla, 1981; Earnest et al., 1988; Kelly, 1992; Kelly et al., 1987). The accurate correlation of MRI with histopathological findings remains the only means of characterizing imaging abnormalities and of establishing tumor margins. However, such data are scarse for LGG as samples performed beyond MRI-defined abnormalities are rarely available. Iwama et al. reported a correlation study between MRI and postmortem histopathological examination in cerebral gliomas (Iwama et al., 1991). In one studied LGG, scattered isolated tumor cells were identified beyond MRI-defined abnormalities on T2weighted sequence. These findings might possibly be explained by tumor growth, since a 19-week interval elapsed between MRI and post-mortem studies. An ideal correlative study would imply correlating MRI and histopathological specimens taken simultaneously. Watanabe et al. identified isolated tumor cells in spatially oriented surgical samples obtained from sites outside MRI-defined abnormalities on T2weighted sequence, from five of the eight studied LGG (Watanabe et al., 1992). These results also question the accuracy of histological and imaging correlation procedures as the location of specimens was performed intraoperatively, based on the depth below the anatomically identified cortex. Accurate real-time correlative studies between MRI and spatially oriented specimens could be allowed by image-based serial stereotactic biopsies. Stereotactic spatially oriented histological specimens from LGG peripheral areas that appeared normal on MRI have already been presented. Using T1weighted MRI-based stereotactic biopsies, Jenkinson et al. (2006) showed an insidious change in cellularity at the radiological margin on T1-weighted sequence in most of 43 LGG. The spatially limited sampling 2–3 mm beyond the MRI-defined tumor border did not allow a precise definition of the spatial extent

158

of LGG. Earnest et al. showed the tumor to be confined to the lesion margins as demonstrated by MRI on T1-weighted and T2-weighted in one case of LGG (Earnest et al., 1988). In contrast, Kelly et al. (1987) reported a study of serial stereotactic biopsies obtained from 10 untreated LGG and identified isolated tumor cells in three of the five biopsy specimens obtained beyond MRI-defined abnormalities on T2-weighted sequence. However, these studies included a small number of LGG, were based on histological morphological criteria alone, and did not specify the MRI tumor delineation of the studied LGG.

MRI Underestimates the Spatial Extent of LGG In a recent study, Pallud et al. (2010) retrospectively analyzed all MRI-based serial stereotactic biopsies that were performed according to the Talairach stereotactic method in their institution for the diagnosis of diffuse cerebral gliomas. Such stereotactic biopsies were performed until 2002 to sample all components of the imaging abnormalities for an accurate pathological diagnosis, for classification into histological subtypes and grading and to establish the real tumor extent in a surgical context (Daumas-Duport et al., 1983; Daumas-Duport and Szikla, 1981). They selected, from a series of 109 procedures performed for the diagnosis of a WHO grade II glioma from January 1992 to December 2001, 16 untreated adult patients harboring a well-delineated and non contrast-enhanced supratentorial LGG for whom MRI-based serial stereotactic biopsies were performed, in part, beyond MRI-defined abnormalities before any oncological treatment. In addition to classical morphological histological criteria, the identification of isolated tumor cells was ascertained using multiple immunostainings and cell cycle marker (MIB-1 immunostaining). Moreover, the inaccuracy in the histological and imaging correlation procedure has been limited by several features of this study: • The short interval between preoperative MRI and histological diagnosis should restrict tumor growth between imaging and biopsy. • Biopsy samples were performed in a single operation before any treatment, so that artificially induced

J. Pallud

ultrastructural changes of the tumor should be limited and the use of a Sedan-Vallicioni biopsy cannula should act to reduce brain movements during biopsy procedures. • The superimposition of reformatted preoperative MRI and intraoperative teleradiographic X-rays taken after each biopsy samples allowed each biopsy site to be directly displayed on MRI on the same reference plane. The superimposition accuracy was controlled with postoperative MRI or CT-scan demonstrating the biopsy tract. • Only isolated tumor cells detected in biopsy samples that were at least 10 mm distant from MRIdefined abnormalities were considered. A total of 101 biopsy samples (median 6, range 3–9 per patient) were performed in the 16 patients. A total of 37 biopsy samples (median 2.2, range 1–5 per patient) were performed beyond MRI-defined abnormalities on T2-weighted and FLAIR sequences. The maximal distance of the biopsy samples sites from MRI-defined abnormalities on T2-weighted sequence ranged from 10 to 26 mm. • In all biopsy samples performed beyond MRIdefined abnormalities, the cortex and the white matter had a normal appearance on routine staining. There was no increased cell density, no edema and no gliosis. • MIB-1 immunostaining revealed MIB-1 positive cells (i.e., cycling cells), in all but two of the 37 samples where MIB-1 positive cells were indifferently distributed within the cortex and the white matter and their number was often variable from one BS to one another in a same patient. • None of the MIB-1 positive cells coexpressed glial fibrillary acidic protein and that all MIB-1 positive cells coexpressed OLIG2, thus excluding the possibility that MIB-1 positive cells correspond to reactive astrocytes or activated microglia. MIB-1 positive cells identified beyond MRI-defined abnormalities on T2-weighted and FLAIR sequences were cycling isolated tumor cells, since (Fig. 17.2): • Their morphological characteristics reflected those of tumor cells. • Their number was significantly higher than that of non-tumor controls.

17 Diffuse Low-Grade Gliomas: What Does “Complete Resection” Mean?

159

Of note, in this study MIB-1 immunostaining was used to detect cycling tumor cells in brain parenchyma surrounding MRI-defined LGG. MIB-1 immunostaining only labels the fraction of the tumor cells, which at a given time are cycling cells. However, the proliferation rates are low for LGG. Therefore, only a

low percentage of tumor cells (i.e. the cycling tumor cells) could be identified using MIB-1 immunostaining. Moreover, glioma cells under migration present a decreased proliferation rate (Giese et al., 2003). As isolated tumor cells may represent invasive gliomas cells, it is then probable that the brain parenchyma around MRI-defined abnormalities on T2-weighted and FLAIR sequences contains a higher proportion of tumor cells than those demonstrated with MIB-1 immunostaining. This study demonstrated the inability of conventional MRI on T2-weighted and FLAIR sequences in determining the actual spatial extent of LGG.

Fig. 17.2 (a) Preoperative MRI of a patient on axial T2weighted sequences demonstrating the location of serial stereotactic biopsy samples performed in a right fronto-insular WHO grade II glioma. Each biopsy samples are defined as inside (black) or outside (white) areas of hypersignal. (b) A significant decreasing gradient of the number of MIB-1 positive cells with the distance from the MRI-defined abnormalities is observed. The number of cycling cells is expressed as MIB-1 positive cells/cm2 and the distance from MRI-defined abnormalities is expressed in mm. (c, d, e, f) Comparative histological features of biopsy samples performed within (c) and outside (d, e, f) MRIdefined abnormalities. Conventional Hemalun-Phloxin stainings

showed that samples performed within the MRI-defined tumor limits (c) are constituted, in the white matter, of infiltrative tumor cells associated by interstitial edema and gliosis (×400). In samples performed outside MRI-defined abnormalities (d), the white matter cell density appeared normal without any edema or gliosis (×400). Double immunostainings revealing that cycling cells do not shared astrocytic marker but correspond to Olig2 positive cells. Double immunufluorescent labelling (e) showed that all MIB-1 positive cells (green) co-expressed the oligodendrocyte cell marker olig2 (red). Double chromogenic immunostaining (f) revealed that MIB-1 positive cells (red) do not share Glial Fibrillary Acidic Protein astrocytic marker (brown) (×600)

• Their number per square centimeter significantly decreased with distance from the MRI-defined tumor borders. • Their number was significantly correlated with the tumor growth fraction.

160

Practical Conclusions As demonstrated by Pallud et al. (2010) using a multi-scale correlative approach with histological and imaging data, conventional MRI underestimates the actual spatial extent of LGG, even when they are well delineated on T2-weigthed and FLAIR sequences: • Cycling isolated tumor cells are present beyond MRI-defined abnormalities in all LGG studied and permeate surrounding “normal” parenchyma. • Isolated tumor cells can be detected at sites up to at least 15 mm beyond MRI-defined abnormalities. It is clearly difficult for present imaging techniques to distinguish between intact brain parenchyma and that infiltrated by scattered isolated tumor cells. As already explicited by Kelly (2004), tumor cells could be found far from any MRI-defined abnormality, even in LGG. Thus, clear tumor boundaries do not actually exists and LGG are diffusely infiltrative tumors with a decrease of tumor cell density as a function of distance from the tumor component associated with abnormalities on conventional MRI. This implies that a gross total removal of a LGG either macroscopically or on postoperative imaging leaves isolated tumor cells beyond the surgical field. This observation may explain clinical events hampering the natural history of LGG: • The observation of tumor recurrences in the margin of the resection (Kelly, 1992, 2004) were isolated tumor cells are the more numerous. • The prognostic value of the extent of resection for progression free survival (Sanai and Berger, 2008) as extensive resection decreases the number of remaining isolated tumor cells. • The prognostic value of the extent of resection for overall survival (Berger et al., 1994; Sanai and Berger, 2008) as extensive resection theoretically decreases the cumulative odds of malignant progression of residual tumor cells. • The prognostic value of the tumor volume (Berger et al., 1994) as larger LGG probably have more surrounding isolated tumor cells and more residual tumor cells after surgical resection. As a practical consequence, the knowledge of the inability of MRI in determining the spatial extent of

J. Pallud

LGG is of paramount importance in evaluating the risk-to-benefit ratio of surgery, which goal is optimizing the extent of resection while preserving the quality of life. These results should be considered when planning a surgical resection of a LGG and several points have to be highlighted: • A maximal safe resection preserving eloquent brain area is recommended because of the infiltrative nature of LGG and the frequent juxtaposition close to and/or within critically eloquent brain areas. • Surgical resection should be tailored according to cortico-subcortical functional boundaries rather than MRI-based limits. • An extended resection of a margin beyond MRIdefined abnormalities, whenever feasible in noneloquent brain areas, might increase the survival of patients harboring a LGG. • An early surgical treatment while the LGG is smaller might delay tumor progression by decreasing the number of residual isolated tumor cells that escape the gross total removal. Acknowledgments Johan Pallud wants to thank FrançoisXavier Roux, Edouard Dezamis and Bertrand Devaux of the department of Neurosurgery, Catherine Daumas-Duport and Pascale Varlet of the department of Neuropathology, JeanFrançois Meder and Catherine Oppenheim of the department of Neuroradiology of the Sainte-Anne Hospital Center in Paris, France. Johan Pallud wants to thanks all the members of the French Glioma Network (REG, Réseau d’Etude des Gliomes) and particularly Emmanuel Mandonnet, Laurent Capelle, Luc Taillandier and Hugues Duffau.

References Berger MS, Deliganis AV, Dobbins J, Keles GE (1994) The effect of extent of resection on recurrence in patients with low grade cerebral hemisphere gliomas. Cancer 74: 1784–1791 Bynevelt M, Britton J, Seymour H, MacSweeney E, Thomas N, Sandhu K (2001) FLAIR imaging in the follow-up of low-grade gliomas: time to dispense with the dual-echo? Neuroradiology 43:129–133 Connor SE, Gunny R, Hampton T, O’Gorman R (2004) Magnetic resonance image registration and subtraction in the assessment of minor changes in low grade glioma volume. Eur Radiol 14:2061–2066 Croteau D, Scarpace L, Hearshen D, Gutierrez J, Fisher JL, Rock JP, Mikkelsen T (2001) Correlation between magnetic resonance spectroscopy imaging and image-guided biopsies: semiquantitative and qualitative histopathological analyses of patients with untreated glioma. Neurosurgery 49:823–829

17 Diffuse Low-Grade Gliomas: What Does “Complete Resection” Mean? Daumas-Duport C, Meder JF, Monsaingeon V, Missir O, Aubin ML, Szikla G (1983) Cerebral gliomas: malignancy, limits and spatial configuration. Comparative data from serial stereotaxic biopsies and computed tomography (a preliminary study based on 50 cases). J Neuroradiol 10:51–80 Daumas-Duport C, Monsaigneon V, Blond S, Munari C, Musolino A, Chodkiewicz JP, Missir O (1987) Serial stereotactic biopsies and CT scan in gliomas: correlative study in 100 astrocytomas, oligo-astrocytomas and oligodendrocytomas. J Neurooncol 4:317–328 Daumas-Duport C, Szikla G (1981) Definition of limits and 3D configuration of cerebral gliomas. Histological data, therapeutic incidences (author’s transl). Neurochirurgie 27: 273–284 Daumas-Duport C, Varlet P, Tucker ML, Beuvon F, Cervera P, Chodkiewicz JP (1997) Oligodendrogliomas. Part I: Patterns of growth, histological diagnosis, clinical and imaging correlations: a study of 153 cases. J Neurooncol 34:37–59 Duffau H (2005) Lessons from brain mapping in surgery for lowgrade glioma: insights into associations between tumor and brain plasticity. Lancet Neurol 4:476–486 Duffau H (2009) Surgery of low-grade gliomas: towards a ‘functional neurooncology’. Curr Opin Oncol 21:543–549 Earnest Ft, Kelly PJ, Scheithauer BW, Kall BA, Cascino TL, Ehman RL, Forbes GS, Axley PL (1988) Cerebral astrocytomas: histopathologic correlation of MR and CT contrast enhancement with stereotactic biopsy. Radiology 166: 823–827 Ganslandt O, Stadlbauer A, Fahlbusch R, Kamada K, Buslei R, Blumcke I, Moser E, Nimsky C (2005) Proton magnetic resonance spectroscopic imaging integrated into image-guided surgery: correlation to standard magnetic resonance imaging and tumor cell density. Neurosurgery 56:291–298 Giese A, Bjerkvig R, Berens ME, Westphal M (2003) Cost of migration: invasion of malignant gliomas and implications for treatment. J Clin Oncol 21:1624–1636 Hlaihel C„ Guilloton L, Guyotat J, Streichenberger N, Honnorat J, Cotton F (2010) Predictive value of multimodality MRI using conventional, perfusion, and spectroscopy MR in anaplastic transformation of low-grade oligodendrogliomas. J Neurooncol 97:73–80 Iwama T, Yamada H, Sakai N, Andoh T, Nakashima T, Hirata T, Funakoshi T (1991) Correlation between magnetic resonance imaging and histopathology of intracranial glioma. Neurol Res 13:48–54 Jenkinson MD, du Plessis DG, Smith TS, Joyce KA, Warnke PC, Walker C (2006) Histological growth patterns and genotype in oligodendroglial tumors: correlation with MRI features. Brain 129:1884–1891 Kelly PJ (1992) Stereotactic resection and its limitations in glial neoplasms. Stereotact Funct Neurosurg 59:84–91 Kelly PJ (2004) Technology in the resection of gliomas and the definition of madness. J Neurosurg 101:284–286 Kelly PJ, Daumas-Duport C, Kispert DB, Kall BA, Scheithauer BW, Illig JJ (1987) Imaging-based stereotaxic serial biopsies in untreated intracranial glial neoplasms. J Neurosurg 66:865–874 Laster DW, Ball MR, Moody DM, Witcofski RL, Kelly DL Jr (1984) Results of nuclear magnetic resonance with cerebral glioma. Comparison with computed tomography. Surg Neurol 22:113–122

161

Mandonnet E, Capelle L, Duffau H (2006) Extension of paralimbic low grade gliomas: toward an anatomical classification based on white matter invasion patterns. J Neurooncol 78:179–185 Mandonnet E, Delattre JY, Tanguy ML, Swanson KR, Carpentier AF, Duffau H, Cornu P, Van Effenterre R, Alvord EC Jr, Capelle L (2003) Continuous growth of mean tumor diameter in a subset of grade II gliomas. Ann Neurol 53:524–528 Mandonnet E, Pallud J, Clatz O, Taillandier L, Konukoglu E, Duffau H, Capelle L (2008) Computational modeling of the WHO grade II glioma dynamics: principles and applications to management paradigm. Neurosurg Rev 31:263–269 Pallud J, Capelle L, Taillandier L, Fontaine D, Mandonnet E, Guillevin R, Bauchet L, Peruzzi P, Laigle-Donadey F, Kujas M, Guyotat J, Baron MH, Mokhtari K, Duffau H (2009) Prognostic significance of imaging contrast enhancement for WHO grade II gliomas. Neurooncology 11:176–182 Pallud J, Devaux B, Daumas-Duport C, Oppenheim C, Roux FX (2005) Glioma dissemination along the corticospinal tract. J Neurooncol 73:239–240 Pallud J, Mandonnet E, Duffau H, Kujas M, Guillevin R, Galanaud D, Taillandier L, Capelle L (2006) Prognostic value of initial magnetic resonance imaging growth rates for World Health Organization grade II gliomas. Ann Neurol 60:380–383 Pallud J, Varlet P, Devaux B, Geha S, Badoual M, Deroulers C, Page P, Dezamis E, Daumas-Duport C, Roux FX (2010) Diffuse low-grade oligodendrogliomas extend beyond MRIdefined abnormalities. Neurology 74:1724–1731 Peyre M, Cartalat-Carel S, Meyronet D, Ricard D, Jouvet A, Pallud J, Mokhtari K, Guyotat J, Jouanneau E, Sunyach MP, Frappaz D, Honnorat J, Ducray F (2010) Prolonged response without prolonged chemotherapy: a lesson from PCV chemotherapy in low-grade gliomas. Neurooncology 20:1078–1082 Price SJ (2009) Advances in imaging low grade gliomas. Adv Tech Stand Neurosurg 35:1–34 Price SJ, Jena R, Burnet NG, Hutchinson PJ, Dean AF, Peña A, Pickard JD, Carpenter TA, Gillard JH (2006) Improved delineation of glioma margins and regions of infiltration with the use of diffusion tensor imaging: an image-guided biopsy study. AJNR Am J Neuroradiol 27:1969–1974 Ricard D, Kaloshi G, Amiel-Benouaich A, Lejeune J, Marie Y, Mandonnet E, Kujas M, Mokhtari K, Taillibert S, LaigleDonadey F, Carpentier AF, Omuro A, Capelle L, Duffau H, Cornu P, Guillevin R, Sanson M, Hoang-Xuan K, Delattre JY (2007) Dynamic history of low-grade gliomas before and after temozolomide treatment. Ann Neurol 61: 484–490 Sanai N, Berger MS (2008) Glioma extent of resection and its impact on patient outcome. Neurosurgery 62:753–764 Soffietti R, Baumert BG, Bello L, von Deimling A, Duffau H, Frenay M, Grisold W, Grant R, Graus F, Hoang-Xuan K, Klein M, Melin B, Rees J, Siegal T, Smits A, Stupp R, Wick W (2010) Guidelines on management of low-grade gliomas: report of an EFNS-EANO Task Force. Eur J Neurol 17:1124–1133 Watanabe M, Tanaka R, Takeda N (1992) Magnetic resonance imaging and histopathology of cerebral gliomas. Neuroradiology 34:463–469

Chapter 18

Quantitative Approach of the Natural Course of Diffuse Low-Grade Gliomas Johan Pallud and Emmanuel Mandonnet

Abstract Supratentorial hemispheric diffuse lowgrade gliomas (LGG), i.e., World Health Organization grade II gliomas, are a heterogeneous group of tumors with distinct clinical, histological and molecular characteristics. The prognosis of LGG varies between series and reflects their heterogeneity with different subgroups harboring specific intrinsic properties. The natural course of LGG, as observed in clinical practice can be summarized as a three-step process although there is an actual continuum from “low-grade” to “high-grade” of malignancy. The two first steps correspond to the histological “low-grade” of malignancy with an initial silent period before clinical revelation followed by a symptomatic period. The third step corresponds to the progression to a higher grade of malignancy leading to neurological disability and ultimately to death. It has been well demonstrated that LGG are progressive tumors that present a systematic, spontaneous and continuous radiological growth all along their natural course, even during the initial silent period and during the symptomatic period before any transformation into a higher grade of malignancy. The Velocity of Diametric Expansion (VDE), estimated from the evolution of the mean tumor diameter over time, can easily quantify the radiological tumor growth. The median VDE is at about 4 mm/year for LGG albeit with a great heterogeneity. Several intrinsic factors may influence spontaneous VDE (1p19q codeletion and p53 overexpression) or not (histological subtype) and extrinsic factors may modify VDE

J. Pallud () Service de Neurochirurgie, Hôpital Sainte-Anne, 14 Paris Cedex, France e-mail: [email protected]

(hormonal changes during pregnancy). The spontaneous VDE has a strong prognostic significance on overall and progression free survivals. As a consequence, the analysis of the spontaneous VDE, a dynamic macroscopic parameter easily available in clinical practice, may be a useful tool to overcome biological diversity of LGG and could be integrated along with the other “static” parameters (histological and molecular analyses) in a multi-scale approach to understand better the individual natural course of LGG. The VDE remains unchanged after surgical resection, whereas it decreases markedly during and after chemotherapy with temozolomide or PCV. Thus, a precise assessment of the VDE obtained before and after treatment would help guiding and analyzing the effects of different oncological treatment modalities on an individual basis. Keywords LGG · VDE · Heterogeneity · MTD · codeletion · overexpression

Introduction Supratentorial hemispheric diffuse low-grade gliomas (LGG), i.e., World Health Organization grade II gliomas, are a heterogeneous group of tumors with distinct clinical, histological and molecular characteristics (Soffietti et al., 2010). The prognosis of LGG varies between series and reflects their heterogeneity with different subgroups harboring specific intrinsic properties. Indeed, the 5-year overall and progression-free survival rates in randomized studies range from 58 to 72% and 37 to 55%, respectively (Soffietti et al., 2010). In addition, the natural history of LGG is complex and poorly understood. During their natural course,

M.A. Hayat (ed.), Tumors of the Central Nervous System, Volume 2, DOI 10.1007/978-94-007-0618-7_18, © Springer Science+Business Media B.V. 2011

163

164

LGG grow continuously and tend to progress to a higher grade of malignancy, leading to neurological disability and ultimately to death. Hence, reference to these lesions as benign gliomas has been abandoned (Lang and Gilbert, 2006). However, a “wait and see” policy has been proposed in the past years in the initial management of LGG before their progression towards a higher grade of malignancy. Thus, long clinical and radiological follow-ups with repeated MRI of untreated LGG are available (Wessels et al., 2003). They allow correlating the quantitative radiological growth rates over time with clinical and histological features. Therefore, the aim of the present chapter is to review quantitative data regarding the natural history of LGG. A 3-step natural history scheme will be proposed. We will show that LGG present systematically a spontaneous and continuous radiological tumor growth before oncological treatment and before progression into a higher grade of malignancy. The practical implications will be detailed, mainly the prognostic significance of spontaneous growth rates of LGG regarding overall survival and progression-free survival, and the effects of oncological treatments on radiological growth rates.

The Three-Step Natural History of LGG A Step-By-Step Approach of a Continuum The natural course of LGG, as observed in clinical practice can be summarized as a three-step process. The two first steps correspond to the histological “lowgrade” of malignancy with an initial silent period before clinical revelation followed by a symptomatic period. The third step corresponds to the progression to a higher grade of malignancy. Beyond this artificial clustering of the LGG natural course, inspired by WHO histological classifications, one should keep in mind the actual continuum from “low-grade” to “high-grade” of malignancy as the tumor acquires progressively genotypic and phenotypic characteristics leading to malignancy.

The Initial Silent Period The natural course of LGG during the initial silent period and their “date of birth” are unknown and

J. Pallud and E. Mandonnet

only analyses of incidental LGG could help the understanding. However, the natural course of an incidental LGG has been very scarcely reported. In a recent report, Pallud et al. (2010a) reviewed 1296 cases of histologically confirmed supratentorial hemispheric LGG in adults and identified 47 incidental tumors, defined as a previously undetected and incidental LGG at the time of radiological diagnosis that was unexpectedly discovered and unrelated to the purpose of the MRI examination in asymptomatic patients. They showed that incidental LGG are similarly proliferative tumours continuously growing on sequential MRI at the same rates than the symptomatic LGG: • The radiological tumour growth and the proliferation rates of incidental LGG were within reported ranges of symptomatic LGG. • When not treated, incidental LGG became symptomatic with time, at a median interval of 48 months since radiological discovery. • Incidental LGG present tumor recurrence and progression to higher grade of malignancy and led to death in several cases despite oncological treatment. In addition, the data suggested that incidental LGG represent an earlier step than symptomatic LGG in the tumor natural course as: • Age at radiological discovery was lower. • Incidental LGG presented with significantly smaller tumour volumes. Accordingly, tumours were limited to one lobe in most of cases, corpus callosum involvement was rare and no midline crossing was observed. • Only rare cases of contrast enhancement were observed, whereas contrast enhancement has been reported in ~16% of cases for all LGG (Pallud et al., 2009a). As the “date of birth” of a LGG is not known, the tumor genesis is still a mater of debate. They probably occur during the young adult period as suggested by rare reports of a normal MRI examination several years before the LGG discovery. This finding is in agreement with literature, which showed that LGG are distinct entities in pediatric and adult populations with different natural courses and prognoses (Bristol, 2009).

18 Quantitative Approach of the Natural Course of Diffuse Low-Grade Gliomas

The Symptomatic Period of Low-Grade of Malignancy Histologically, LGG comprise generally only a volume of isolated tumor without solid tumor tissue associated with edema but no microvascular proliferation permeating brain parenchyma preferentially along particular anatomical structures (Daumas-Duport et al., 1997; Giese et al., 2003). They correspond to a spatial architectural subtype III, as described by Daumas-Duport et al. (1983, 1987) and Daumas-Duport and Szikla (1981). However, there are some LGG that are type II because patches of solid tumor tissue exist within areas of infiltrated parenchyma by isolated tumor cells. In addition, the proliferation rates of LGG are very low (Soffietti et al., 2010). These spatial architectural findings are correlated with the MRI findings: • Repeated MRI examinations demonstrate a spontaneous and slow growth pattern of LGG (Chen et al., 2010; Mandonnet et al., 2003). • Isolated tumor cells present with an area of hypointensity on T1-weighted sequence and hyperintensity on T2-weighted and FLAIR sequences. MRI abnormalities are either with well-circumscribed margins or with indistinct borders, involve the cortex from the surface to its deeper part, extend more or less in the adjacent white matter and occasionally expand the involved gyrus (Daumas-Duport et al., 1997). • Definite contrast enhancement is typically absent in accordance with the architectural subtype III of LGG. Indeed, contrast enhancement reflects macroscopically the microvascular proliferation thus explaining why its occurrence is associated with a worsened prognosis (Pallud et al., 2009a). • The preferential brain invasion along myelinated white matter fiber tracts is easily demonstrated on MRI. Several reports have illustrated that glioma cells migrate along intra-hemispheric tracts, inter-hemispheric tracts and descending pathways (Mandonnet et al., 2006; Pallud et al., 2005). Of note, it has to be underlined that conventional MRI underestimates the actual spatial extent of LGG, even when they are well delineated on T2weigthed and FLAIR sequences since cycling isolated tumor cells can be detected at sites up to at least

165

15 mm beyond MRI-defined abnormalities (Pallud et al., 2010c). Taken together, histological and imaging abnormalities explain the silent initial period and the long symptomatic period with no or mild neurological disorders: • The invasive behavior of LGG explains why functional areas may persist within brain parenchyma permeated by isolated tumor cells (Daumas-Duport et al., 1997). • The slow tumor growth allows functional compensation by brain plasticity mechanisms to be efficiently implemented, explaining the lack of neurological deficits despite frequent involvement of eloquent areas (Duffau, 2005, 2009). • The superficial tumor location and the frequent cortical involvment explains why seizures are the commonest presenting symptom and occur in ~90% of LGG (Soffietti et al., 2010). • The lack of solid tumor tissue component explains the rare occurrence of focal neurological deficit and of increased intracranial pressure, despite clear radiological mass effects (Soffietti et al., 2010). Of note, LGG may remain asymptomatic until progression to a higher grade of malignancy.

The Malignant Progression The time to progression to a higher grade of malignancy is variable from one patient to one another since: • LGG are a heterogeneous group of tumors with distinct clinical, histological and molecular characteristics. • LGG are diagnosed at various steps along the continuum of their natural course. Histologically, a solid tumor tissue component where the tumor cells are in contact with each other, containing microvascular proliferation and only residual brain parenchyma or none, occurs within the isolated tumor cells component. The tumor now corresponds to a spatial architectural subtype II, as described by Daumas-Duport et al. (1983, 1987) and Daumas-Duport and Szikla (1981). The malignant transformation is spatially heterogeneous and areas of

166

high-grade and low-grade of malignancy may coexist. It is sometimes difficult to perform an accurate histological grading due to the limitations of histological sampling. In addition, proliferation rates increase. These spatial architectural findings are correlated with the MRI findings: • A change of the radiological growth pattern with growth rates at ranges over those of LGG (Pallud et al., 2006). • The occurrence of nodular-like and ring-like contrast enhancement patterns reflecting macroscopically the microvascular proliferation within the architectural subtype II. They are known to be associated with a worsened prognosis (Pallud et al., 2009a). • The occurrence of necrosis within the contrast enhanced areas. Taken together, histological and imaging abnormalities explain functional deficits that are associated with malignant transformation. Neurological disability occurs as: • Brain plasticity mechanisms are overtaken by the fast growing tumor. • The increased mass effect and intracranial pressure injure the peripheral brain areas that were initially recruited by brain plasticity mechanisms. • The solid tumor tissue component destroys the remaining functional infiltrated brain parenchyma. This malignant transformation leads ultimately to death (Soffietti et al., 2010).

Image-Based Monitoring of the Natural Course of LGG Evolution Over Time of the Radiological Growth Pattern of LGG Only rare papers have analyzed qualitatively the evolution of the radiological growth pattern over time of

J. Pallud and E. Mandonnet

LGG. Chen et al. reported that tumors originating from the grey matter will remain bulky while continously growing, whereas tumors originating at the junction of gray and white matters will grow predominantly along the adjacent white matter fiber tracts (Chen et al., 2010). Mandonnet et al., demonstrated the dynamic invasion of major white matter fiber tracts surrounding the insular lobe for paralimbic LGG (Mandonnet et al., 2006).

Quantitative Radiological Growth Rates of LGG All quantitative longitudinal studies of radiological growth patterns demonstrated a spontaneous and continuous growth of LGG (Fig. 18.1a). Mandonnet et al. first quantified the spontaneous radiological growth of LGG on successive MRIs in 27 histologically proven LGG followed before oncological treatment (Fig. 18.1b) (Mandonnet et al., 2003). These initial results were confirmed on a larger series of 143 histologically proven LGG that ranged individual velocities of diametric expansion from 1 to 36 mm/year (Fig. 18.1c) (Pallud et al., 2006). Several independent groups have since reported the same results (Hlaihel et al., 2010; Pallud et al., 2010c; Peyre et al., 2010; Ricard et al., 2007). The quantitative analyses demonstrated a linear evolution of the spontaneous Mean Tumor Diameter (MTD) over time and quantified the Velocity of Diametric Expansion (VDE) at about 4 mm/year average rate (Fig. 18.1b) (Mandonnet et al., 2003). Further studies, performed by distinct groups, have confirmed this range value (Hlaihel et al., 2010; Pallud et al., 2010c; 2006; Peyre et al., 2010; Ricard et al., 2007). Of note, these quantitative studies never reported indolent untreated LGG or LGG alternating indolent and growing periods. In addition, Pallud et al., recently demonstrated that incidental LGG present a similar radiological growth during the silent period preceding their clinical revelation with VDE in the known ranges for symptomatic LGG (Pallud et al., 2010a). As different independent groups have reported similar observations, it has to be accepted that LGG are progressive tumors that present a spontaneous and

18 Quantitative Approach of the Natural Course of Diffuse Low-Grade Gliomas

Fig. 18.1 (a) Example of the natural course of a diffuse lowgrade glioma through the evolution of its Velocity of Diametric Expansion (VDE) over time. A right fronto-temporo-insular glioma was discovered incidentally in a 29-year-old righthanded woman. The tumor was initially followed and consecutive MRI showed a spontaneous growth with a VDE at 5.3 mm/year (e). About 3 years after radiological discovery, partial seizures occurred. A subtotal surgical resection (general anaesthesia) was performed as the first oncological treatment and confirmed a WHO grade II oligodendroglioma on pathological analysis. After surgery, the residual tumor progressed with a VDE at 2.3 mm/year. Three years after histological diagnosis, the patient refused further clinical and radiological follow-up. One year later, epilepsy reccured and the patient presented in emergency with a left hemiparesis. Images

167

demonstrated a radiological progression toward a higher grade of malignancy with a VDE increase at 28.4 mm/year and the occurrence of a nodular-like area of contrast enhancement (e). A second subtotal surgical resection (general anaesthesia) was performed and an external conformational radiotherapy plus concomitant and adjuvant chemotherapy with temozolomide was started. b. The evolution of the Mean Tumor Diameter (MTD) is comparable among patients with a median VDE at about 4 mm/year (Adapted from Mandonnet et al., 2003). c. Distribution of patients by individual radiological growth rates. The VDE vary considerably among patients but their distribution results in two groups. About 85% presented a VDE less than 8 mm/year (blue bars) and about 15% presented a VDE at 8 mm/year or more (purple bars) (adapted from Pallud et al., 2006)

168

J. Pallud and E. Mandonnet

continuous radiological growth all along their natural course (i.e., during the initial silent period and during the symptomatic period before any transformation into a higher grade of malignancy).

Radiological Growth Rates Reflect Different Natural Courses for LGG

Effects of Intrinsic and Extrinsic Factors on Spontaneous Tumor Growth Rates of LGG

Pallud et al. first studied the prognostic value of spontaneous MRI growth rates on overall survival in a retrospective series of 143 histologically proven LGG with measurements of the evolution of the MTD over time (Pallud et al., 2006). The low growth rates subgroup (VDE lower than 8 mm/year) exhibited a significantly longer overall survival than the high growth rates subgroup (VDE at 8 mm/year or more) (Fig. 18.2c). In multivariate analysis, tumor growth rates and initial tumor volume were two independent prognostic factors significantly associated with overall survival. The prognostic significance of spontaneous MRI growth rates on predicting progression into a higher grade of malignancy was addressed by Hlailel et al. (2010). They showed that an elevated VDE higher than 3 mm/year was correlated with a greater risk of progression into a higher grade of malignancy with an average VDE at 7.87 mm/year in transformers group versus an average VDE at 2.14 mm/year in non transformers group. Brasil Caseiras et al. (2009) proved that tumor growth within 6 months was better than baseline volumes, rCBV, or ADC in predicting time to malignant transformation in untreated LGG using the evolution of the tumor volume over time.

Although the different LGG histological subtypes (oligodendroglioma, astrocytoma, mixed glioma) are associated with different prognoses (Pignatti et al., 2002; Soffietti et al., 2010), they do not exhibit significant differences regarding the radiological tumor growth rates (Mandonnet et al., 2009; Pallud et al., 2006; Ricard et al., 2007). As an example, Mandonnet et al. quantified the median VDE at 3.35 mm/year in the astrocytoma subgroup, at 4.2 mm/year in the oligodendroglioma subgroup, and at 3.53 mm/year in the mixed glioma subgroup, without any statistical differences between these subgroups (Fig. 18.2a) (Mandonnet et al., 2009). Ricard et al. (2007) investigated the link between LGG radiological growth rates and genetic alterations. They showed that LGG with 1p-19q codeletion grew significantly slower than LGG without 1p-19q codeletion and that LGG with overexpression of p53 grew significantly faster than LGG without overexpression of p53 (Fig. 18.2a). There is little documentation of the effects of extrinsic factors on glioma growth properties. It has been suggested that pregnancy can have a negative impact on the natural course of glioma, by the mean of hormonal changes (Pallud et al., 2009b). Only one study investigated quantitatively the impact of pregnancy on the radiological growth rates of LGG in 12 pregnancies in 11 adult women harbouring a LGG prior to pregnancy (Pallud et al., 2010b). Pallud et al. (2010b) showed that LGG accelerated their radiological growth rates during pregnancy (Fig. 18.2a). Indeed, VDE significantly increased during pregnancy above levels detected either before pregnancy or after delivery in 75% of cases, changes in tumour growth were associated with an increase in seizure frequency in 40% of cases and radiological and clinical changes during pregancy motivated further oncological treatment after delivery in 25% of cases.

Prognostic Significance of the Spontaneous Growth Rates

Radiological Growth Rates as a New Prognostic Factor for LGG? The prognosis of LGG varies between series and reflects the heterogeneity of the observed natural histories. The 5-year overall (OS) and progression-free survival (PFS) rates in randomized studies range from 58 to 72% and 37 to 55%, respectively (Soffietti et al., 2010). Regarding clinical findings, age over 40 years and presence of pre-operative neurological deficits are adverse prognostic factors (Pignatti et al., 2002; Soffietti et al., 2010). Regarding radiological findings,

18 Quantitative Approach of the Natural Course of Diffuse Low-Grade Gliomas

Fig. 18.2 (a) Factors influencing the spontaneous Velocity of Diametric Expansion (VDE) of WHO grade II gliomas. The different histological subtypes do not exhibit significant differences of VDE (black bars, adapted from Mandonnet et al., 2009). The 1p-19q codeletion molecular status is associated with a significantly slower VDE (dark grey bars, adapted from Ricard et al., 2007). The p53 overexpression molecular status is associated with a significantly faster VDE (light grey bars, adapted from Ricard et al., 2007). Pregnancy increases significantly VDE as compared to pre-pregnancy and post-delivery rates (white bars, adapted from Pallud et al., 2010b). (b) Effects of oncological treatments on the Velocity of Diametric Expansion (VDE). VDE

169

are expressed before and after surgery (black bars, adapted from Mandonnet et al.), before and during Temozolomide (TMZ) chemotherapy (dark grey bars, adapted from Ricard et al., 2007), before and during PCV chemotherapy (light grey bars, adapted from Peyre et al., 2010). (c) Kaplan-Meier estimates of overall survival by individual Velocity of Diametric Expansion (VDE). The subgroup with a VDE less than 8 mm/year (blue line) presents a significant longer overall survival (median survival more than 15 years) than the subgroup with a VDE at 8 mm/year or more (purple line) where the overall survival is closer to that of more malignant gliomas (median survival at 5.16 years) (adapted from Pallud et al., 2006)

170

J. Pallud and E. Mandonnet

larger tumors and tumors crossing the midline correlate with a short OS and PFS (Pignatti et al., 2002; Soffietti et al., 2010) and growth rates are inversely correlated with OS (Pallud et al., 2006). There are conflicting reports as to whether contrast enhancement is associated with a worsened prognosis (Soffietti et al., 2010). However, Pallud et al. recently showed that the presence of contrast enhancement alone, regardless of its pattern, had no prognostic value and that only the presence of a nodular-like pattern or of progressive contrast enhancement over time were associated with a worsened prognosis (Pallud et al., 2009a). A low Cerebral Blood Volume and a low uptake of 11C-methionine correlate with longer PFS and OS (Soffietti et al., 2010). Regarding pathological and molecular findings, the astrocytoma histological subtype is associated with a worsened prognosis (Pignatti et al., 2002; Soffietti et al., 2010). 1p loss (with or without 19q loss) is a favorable prognostic factor (Soffietti et al., 2010). IDH1 codon 132 mutation has been recently suggested as an independent favorable prognostic factor (Soffietti et al., 2010). When risk factors are cumulated, the survival of LGG is dramatically decreased with survival times closer to those of higher grade gliomas. The limitation of the histological diagnosis explains, in part, this prognostic heterogeneity. Analyses on microscopic and macroscopic, static and dynamic, scales are required to refine the prognostic evaluation. A multi-scale approach is a way to circumvent the diagnostic limitations of the histological examination. Thus, analysis of VDE, a dynamic macroscopic parameter easily available in clinical practice by the mean of repeated measurements of MTD over time, may be a useful tool to overcome biological diversity of LGG (Mandonnet et al., 2008; Pallud et al., 2006). As a practical consequence, VDE could be integrated along with the other “static” parameters (histological and molecular analyses) in a multi-scale approach to understand better the individual natural course of LGG.

as promising tool in the follow-up of LGG and in the monitoring of the different oncological treatments. Mandonnet et al. have recently shown that VDE remain unchanged after surgical resection in a retrospective study of 54 resected LGG (Fig. 18.2b) (Mandonnet et al., 2009). This suggests that the survival benefit of surgery is mediated by a cytoreductive effect. Of note, 2 patients exhibited a decrease of their tumor growth rates greater than 3 mm/year whereas in 2 other patients, surgery failed to stop an ongoing malignant transformation, resulting in an increase of 3 mm/year on the post-operative tumor growth rates. Thus, the precise quantitative assessment of the VDE obtained pre and post-operatively would help analyzing the effects of surgical resection on an individual basis and guiding the postoperative oncological treatment. Similarly, the tumor response to chemotherapy can be quantified by the VDE evolution over time. Ricard et al. first quantified the tumor response after temozolomide and almost all LGG exhibited an initial decrease of the VDE after temozolomide onset (Fig. 18.2b) (Ricard et al., 2007). They evidenced different patterns of response, depending on the 1p-19q codeletion status as tumor relapse occured more frequently and earlier in tumors without 1p-19q codeletion. Peyre at al. first quantified the tumor response after PCV chemotherapy and all LGG presented an initial VDE decrease after PCV onset (Fig. 18.2b) (Peyre et al., 2010). They demonstrated that the median VDE decrease after PCV onset was very close to the tumor growth decrease after temozolomide onset found by Ricard et al. (2007). In addition, they showed an ongoing VDE decrease after PCV continuation that was prolonged more than 2 years in 60% of the LGG under study. These results raise the issue of a chemotherapy monitoring based on the quantitative changes of the tumor VDE. Of note, at that time, no study has quantified the radiological response, using VDE evolution over time, after radiation therapy.

Quantitative Assessment of Treatment Efficacy

Conclusions

Along with clinical response, the quantitative assessment of the individual VDE changes by the mean of the MTD evolution over time on consecutive MRI, appears

Quantitative analyses on MRI have demonstrated that LGG are progressive tumors that present a spontaneous and continuous radiological growth all along their natural course during the initial silent period and

18 Quantitative Approach of the Natural Course of Diffuse Low-Grade Gliomas

during the symptomatic period before any transformation into a higher grade of malignancy. LGG are a heterogeneous group of tumors with distinct prognoses and a multi-scale approach may help overpassing the diagnostic limitations of the sole histological examination. Thus, analysis of VDE, a dynamic macroscopic parameter easily available in clinical practice by the mean of repeated measurements of MTD over time (for technical details, see (Mandonnet et al., 2008)), may be a useful tool to understand the biological diversity of LGG. At the light of its strong prognostic significance, VDE could be integrated along with the other “static” parameters (histological and molecular analyses) in a multi-scale approach to understand better the individual natural course of LGG. In addition, the precise quantitative assessment of the VDE obtained before and after treatment would help guiding and analyzing the effects of different oncological treatment modalities on an individual basis. Acknowledgments Johan Pallud and Emmanuel Mandonnet want to thanks all the members of the French Glioma Network (REG, Réseau d’Etude des Gliomes) and particularly Laurent Capelle, Luc Taillandier and Hugues Duffau. Johan Pallud wants to thank François-Xavier Roux, Edouard Dezamis and Bertrand Devaux of the department of Neurosurgery, Catherine DaumasDuport and Pascale Varlet of the department of Neuropathology, Catherine Oppenheim and Jean-François Meder of the department of Neuroradiology of the Sainte-Anne Hospital Center in Paris, France.

References Brasil Caseiras G, Ciccarelli O, Altmann DR, Benton CE, Tozer DJ, Tofts PS, Yousry TA, Rees J, Waldman AD, Jager HR (2009) Low-grade gliomas: six-month tumor growth predicts patient outcome better than admission tumor volume, relative cerebral blood volume, and apparent diffusion coefficient. Radiology 253:505–512 Bristol RE (2009) Low-grade glial tumors: are they all the same? Semin Pediatr Neurol 16:23–26 Chen X, Dai J, Jiang T (2010) Supratentorial WHO grade II glioma invasion: a morphologic study using sequential conventional MRI. Br J Neurosurg 24:196–201 Daumas-Duport C, Meder JF, Monsaingeon V, Missir O, Aubin ML, Szikla G (1983) Cerebral gliomas: malignancy, limits and spatial configuration. Comparative data from serial stereotaxic biopsies and computed tomography (a preliminary study based on 50 cases). J Neuroradiol 10:51–80 Daumas-Duport C, Monsaigneon V, Blond S, Munari C, Musolino A, Chodkiewicz JP, Missir O (1987) Serial stereotactic biopsies and CT scan in gliomas: correlative study

171

in 100 astrocytomas, oligo-astrocytomas and oligodendrocytomas. J Neurooncol 4:317–328 Daumas-Duport C, Szikla G (1981) Definition of limits and 3D configuration of cerebral gliomas. Histological data, therapeutic incidences (author’s transl). Neurochirurgie 27:273– 284 Daumas-Duport C, Varlet P, Tucker ML, Beuvon F, Cervera P, Chodkiewicz JP (1997) Oligodendrogliomas. Part I: patterns of growth, histological diagnosis, clinical and imaging correlations: a study of 153 cases. J Neurooncol 34:37–59 Duffau H (2005) Lessons from brain mapping in surgery for lowgrade glioma: insights into associations between tumor and brain plasticity. Lancet Neurol 4:476–486 Duffau H (2009) Surgery of low-grade gliomas: towards a ‘functional neurooncology’. Curr Opin Oncol 21:543–549 Giese A, Bjerkvig R, Berens ME, Westphal M (2003) Cost of migration: invasion of malignant gliomas and implications for treatment. J Clin Oncol 21:1624–1636 Hlaihel C, Guilloton L, Guyotat J, Streichenberger N, Honnorat J, Cotton F (2010) Predictive value of multimodality MRI using conventional, perfusion, and spectroscopy MR in anaplastic transformation of low-grade oligodendrogliomas. J Neurooncol 97:73–80 Lang FF, Gilbert MR (2006) Diffusely infiltrative low-grade gliomas in adults. J Clin Oncol 24:1236–1245 Mandonnet E, Capelle L, Duffau H (2006) Extension of paralimbic low grade gliomas: toward an anatomical classification based on white matter invasion patterns. J Neurooncol 78:179–185 Mandonnet E, Delattre JY, Tanguy ML, Swanson KR, Carpentier AF, Duffau H, Cornu P, Van Effenterre R, Alvord EC Jr, Capelle L (2003) Continuous growth of mean tumor diameter in a subset of grade II gliomas. Ann Neurol 53: 524–528 Mandonnet E, Pallud J, Clatz O, Taillandier L, Konukoglu E, Duffau H, Capelle L (2008) Computational modeling of the WHO grade II glioma dynamics: principles and applications to management paradigm. Neurosurg Rev 31:263–269 Mandonnet E, Pallud J, Fontaine D, Taillandier L, Bauchet L, Peruzzi P, Guyotat J, Bernier V, Baron MH, Duffau H, Capelle L (2009) Inter- and intrapatients comparison of WHO grade II glioma kinetics before and after surgical resection. Neurosurg Rev 33:91–96 Pallud J, Capelle L, Taillandier L, Fontaine D, Mandonnet E, Guillevin R, Bauchet L, Peruzzi P, Laigle-Donadey F, Kujas M, Guyotat J, Baron MH, Mokhtari K, Duffau H (2009a) Prognostic significance of imaging contrast enhancement for WHO grade II gliomas. Neuro-Oncology 11:176–182 Pallud J, Devaux B, Daumas-Duport C, Oppenheim C, Roux FX (2005) Glioma dissemination along the corticospinal tract. J Neurooncol 73:239–240 Pallud J, Duffau H, Razak RA, Barbarino-Monnier P, Capelle L, Fontaine D, Frenay M, Guillet-May F, Mandonnet E, Taillandier L (2009b) Influence of pregnancy in the behavior of diffuse gliomas: clinical cases of a French glioma study group. J Neurol 256:2014–2020 Pallud J, Fontaine D, Duffau H, Mandonnet E, Sanai N, Taillandier L, Peruzzi P, Guillevin R, Bauchet L, Bernier V, Baron MH, Guyotat J, Capelle L (2010a) Natural history of incidental world health organization grade II gliomas. Ann Neurol 68:727–733

172 Pallud J, Mandonnet E, Deroulers C, Fontaine D, Badoual M, Capelle L, Guillet-May F, Page P, Peruzzi P, Jouanneau E, Frenay M, Cartalat-Carel S, Duffau H, Taillandier L Club de Neuro-Oncologie de la Société Française de Neurochirurgie (SFNC), Association des Neuro-Oncologues d’Expression Française (ANOCEF) (2010b) Pregnancy increases the growth rates of World Health Organization grade II gliomas. Ann Neurol 67:398–404 Pallud J, Mandonnet E, Duffau H, Kujas M, Guillevin R, Galanaud D, Taillandier L, Capelle L (2006) Prognostic value of initial magnetic resonance imaging growth rates for World Health Organization grade II gliomas. Ann Neurol 60:380–383 Pallud J, Varlet P, Devaux B, Geha S, Badoual M, Deroulers C, Page P, Dezamis E, Daumas-Duport C, Roux FX (2010c) Diffuse low-grade oligodendrogliomas extend beyond MRIdefined abnormalities. Neurology 74:1724–1731 Peyre M, Cartalat-Carel S, Meyronet D, Ricard D, Jouvet A, Pallud J, Mokhtari K, Guyotat J, Jouanneau E, Sunyach MP, Frappaz D, Honnorat J, Ducray F (2010) Prolonged response without prolonged chemotherapy: a lesson from PCV chemotherapy in low-grade gliomas. Neuro-Oncology 10:1078–1082 Pignatti F, van den Bent M, Curran D, Debruyne C, Sylvester R, Therasse P, Afra D, Cornu P, Bolla M, Vecht C, Karim

J. Pallud and E. Mandonnet AB European organization for research and treatment of cancer brain tumor cooperative group, European organization for research and treatment of cancer radiotherapy cooperative group (2002) Prognostic factors for survival in adult patients with cerebral low-grade glioma. J Clin Oncol 20: 2076–2084 Ricard D, Kaloshi G, Amiel-Benouaich A, Lejeune J, Marie Y, Mandonnet E, Kujas M, Mokhtari K, Taillibert S, LaigleDonadey F, Carpentier AF, Omuro A, Capelle L, Duffau H, Cornu P, Guillevin R, Sanson M, Hoang-Xuan K, Delattre JY (2007) Dynamic history of low-grade gliomas before and after temozolomide treatment. Ann Neurol 61: 484–490 Soffietti R, Baumert BG, Bello L, Von Deimling A, Duffau H, Frénay M, Grisold W, Grant R, Graus F, Hoang-Xuan K, Klein M, Melin B, Rees J, Siegal T, Smits A, Stupp R, Wick W (2010) Guidelines on management of low-grade gliomas: report of an EFNS-EANO∗ Task Force. Eur J Neurol 17:1124–1133 Wessels PH, Weber WE, Raven G, Ramaekers FC, Hopman AH, Twijnstra A (2003) Supratentorial grade II astrocytoma: biological features and clinical course. Lancet Neurol 2:395–403

Chapter 19

Impact of Extent of Resection on Outcomes in Patients with High-Grade Gliomas Debraj Mukherjee and Alfredo Quinones-Hinojosa

Abstract Throughout the course of human history, surgical therapy has remained a viable option for the treatment of certain, well-defined lesions. However, particularly given the technological advancements over the last half-century, the fields of neurosurgery and neurooncology have been able to embrace multimodal forms of therapy, including maximal safe surgical resection, radiotherapy, and both local and systemic chemotherapy, in an effort to improve survival and decrease the odds of developing new post-operative deficits in patients with highly malignant gliomas. Some studies do appear to indicate that extent of resection plays a favorable role in survival of patients with high grade gliomas, although the exact relationship between extent of resection and survival or poor outcomes is not completely defined. Future prospective, randomized trials may need to be developed to assess the effect of multimodal forms of imaging, motor mapping, and combination forms of therapy upon shortand long-term patient outcomes. Keywords High-grade astrocytomas · Mapping · Central nervous system · Temozolomide · Chemotherapy · Radiotherapy

Introduction High-grade astrocytomas are the most common malignant primary central nervous system tumors in adults,

D. Mukherjee () Maxine Dunitz Neurosurgical Institute, Cedars-Sinai Medical Center, Los Angeles, California 90048 e-mail: [email protected]

encompassing both anaplastastic astrocytomas (also known as WHO Grade III tumors) and glioblastoma multiforme (also known as WHO Grade IV tumors). Despite advances in medical and surgical therapy, the median survival remains less than 2 years. While mean survival for patients with high-grade gliomas have remained relatively short, individual patient survival is heterogeneous. Thus, there has been an emphasis on studying intraoperative factors that hold prognostic significance in predicting survival in patients with malignant astrocytomas. For many solid organ malignant tumors, gross total resection with clear margins has been associated with significantly extended survival in well-designed studies. However, the effects of such extensive resection on prolonging survival in patients with malignant astrocytomas is less clear as no prospectively, randomized trials have been conducted in this areas of study. Extensive resection of highgrade gliomas is made increasingly difficult because these tumors are frequently invasive, widely infiltrative, and often involve eloquent areas. Improvements in surgical adjuncts including functional MR imaging, cortical mapping, and intraoperative MR imaging have made it possible to achieve extensive resection of many gliomas. It has remained incompletely understood whether more extensive resection of malignant astrocytomas is associated with significantly prolonged survival in well-designed studies.

Historical Considerations Dating back to the early attempts by Neolithic man, surgery of the brain, meninges, or skull have slowly evolved over 12,000 years. Through most of this history, interventions were technically crude by

M.A. Hayat (ed.), Tumors of the Central Nervous System, Volume 2, DOI 10.1007/978-94-007-0618-7_19, © Springer Science+Business Media B.V. 2011

173

174

modern measures and aimed almost exclusively toward the removal of extracranial, overtly visualized lesions of the skull. Within the modern era, this work progressed toward surgical intervention of the coverings of the brain by Zanobi Pecchioli, who resected a fungus of the dura mater in a patient presenting with a large extracranial mass in 1835. Yet, there continued to be no widely reported surgical interventions to remove primary intraparenchymal lesions of the brain through the late nineteenth century, as only a limited amount was known of localized areas of cortical functioning until the work of Paul Broca and Hughlings Jackson in the 1860s. With modest improvements in neurosurgical techniques and in the collective understanding of cortical functioning, the first widely recognized resection of a primary brain tumor was performed by Rickman J. Godlee in 1884 on a 25-year old who presented with a nearly 3 year history of intractable seizures. Although removal of this seemingly low-grade glioma was deemed an operative success, the patient died of complications from meningitis 3-weeks after surgical intervention. Yet, at autopsy, the patient was noted to have no residual tumor, likely alluding, in part, to the implicit importance Godlee may have placed in achieving a gross total resection. Attempts at intracranial surgery continued to some degree through the early part of the twentieth century and reached a milestone with Walter Dandy’s 1928 case series of patients undergoing hemispherectomy for removal of invasive high-grade gliomas. This series included one patient surviving three and a half years following surgical resection. Nevertheless, most patients died within 3 months of surgery, often of infectious complications, including meningitis and pneumonia. Over the following decades, advancements in surgical science, including the appropriate use of antibiotics and sterile technique, decreased immediate postoperative morbidity and mortality. Furthermore, the advent of cerebral angiograms in the 1930s, computerized topography (CT) scans in the 1970s, and magnetic resonance imaging (MRI) technology in the 1980s further facilitated in operative planning for these complex operations. With the last two decades, advancements in microsurgery have further increased the possibility of achieving operative gross total resection in patients with both low and high grade gliomas. However, the lack of well-designed, prospective studies have made the role of extensive resection of

D. Mukherjee and A. Quinones-Hinojosa

high-grade astrocytomas incompletely defined, despite the great advances made in developing adjuvant radiotherapy and chemotherapy protocols that significant improve survival among these patients (Stupp et al., 2005). For instance, many past studies used varying cut-offs for defining extent of resection thresholds for defining gross total, near total, or subtotal resection, including varying methods of assessing degree of resection (particularly among the non-volumetric studies). Additionally, a majority of these studies did not use standard WHO tumor grading criteria, thus making it difficult to fully generalize the findings of these studies to the general neuro-oncological literature. Furthermore, although most studies has comparable levels of adjuvant radiotherapy or chemotherapy between patients of varying degrees of extent of resection, all studies did not adjust likely predictors of overall survival, including patient comorbidities, location of tumor (for instance near eloquent cortical areas), or known predictors of survival, including 1p19q status among oligodendrogliomas and MGMT status among patient with glioblastoma multiforme. Furthermore, the use of particular pre-operative imaging modalities possibly helpful for operative planning, including DTI or fMRI for patients with lesions involving motor tracts or eloquent cortical areas, respectively, were not accounted for in most studies.

The Impact of Extent of Resection: Lessons Learned from Low-Grade Gliomas Over approximately 20 years, at least eleven studies have independently and retrospectively reviewed the impact of extent of resection on outcomes in patients with low-grade astrocytomas, as reviewed by Sanai and Berger (2008a). Of these, most have used non-volumetric means, including for instance intraoperative surgeon report or gross radiographic evidence of residual disease, to categorize extent of resection. A majority also presented evidence either in univariate or multivariate analysis of supporting greater 5-year progression-free survival or overall survival with greater extent of resection. Only the Johannesen et al. (2003) report from Norway studied 5-year overall survival and found no statistically significant difference in survival between biopsy, subtotal

19 Impact of Extent of Resection on Outcomes in Patients with High-Grade Gliomas

resection, or gross total resection, although there was a trend toward greater survival in those with greater extent of resection. Additionally, a few studies have retrospectively reviewed this association using more precise radiographic volumetrics, as opposed to selfreport or two-dimensional impressions of extent or resection from post-operative radiographic imaging, to assess degree of resection. All retrospective studies demonstrated significantly improved 5-year in those with greater extent of resection, although the threshold for definitely greatest degree of resection ranged from 75 to 100% in these studies. Taken in total, all major retrospective studies analyzing the association between extent of resection and 5-year survival have demonstrated at least a trend toward better outcomes in those with greater resection. These findings have been supported in two reviews and one meta-analyses on this area of study (Pouratian et al., 2007; Sanai and Berger, 2008a; Smith et al., 2008). However, findings in those with low-grade gliomas may not necessarily translate directly to patient with more infiltrative highgrade gliomas with often inherent much worse overall prognoses.

The Impact of Extent of Resection on Survival in Patients with High-Grade Gliomas: A Review of the Literature In the modern era, at least 29 studies have assessed the impact of extent of resection on progression-free or overall survival in patients with high-grade astrocytomas, as reviewed by Sawaya (1999) as well as Sanai and Berger (2008a). Of these, a majority used non-volumetric means of assessing extent or resection while just four studies used the more precise measure of volumetrics (Keles et al., 1999, 2006; Lacroix et al., 2001; Pope et al., 2005). Among the non-volumetric studies, an equal number demonstrated a statistically significant improvement in overall survival in multivariate analysis as those that showed no significant association. Among all studies of high-grade glioma patients, only three have assessed progression-free survival as an outcome (Jeremic et al., 1994; SandbergWollheim et al., 1991; Ushio et al., 2005). Within these studies, two reports demonstrated a statistically significant association between greater extent of resection and longer progression-free survival in multivariate

175

analysis (Jeremic et al., 1994; Ushio et al., 2005), while one reports showed no significant association in multivariate analysis (Sandberg-Wollheim et al., 1991). Four volumetric studies have thus far been widely published on the topic (Keles et al., 1999, 2006; Lacroix et al., 2001; Pope et al., 2005). Among these studies, two reports demonstrated significantly improved overall survival in multivariate analysis in those with greater extent of resection (Keles et al., 1999; Lacroix et al., 2001), while the remaining two studies showed no statistically significant relationship in multivariate analysis (Keles et al., 2006; Pope et al., 2005). Only two volumetric studies have thus far assessed the impact of degree of resection on progression-free survival, with one study showing significant association (Keles et al., 1999) and the other demonstrating no significance (Keles et al., 2006), both in multivariate analyses.

The Importance of Multimodal Therapy in Improving Survival Although some progress has been made in understanding the impact of extensive resection on survival, the treatment of high-grade astrocytomas remains multimodal in approach, combining surgical resection with both adjuvant chemotherapy and radiation. The efficacy of combination local chemotherapy following extensive tumor resection was found to improve survival in patient with recurrent malignant gliomas from 23 to 31 weeks after revision resection (Brem et al., 1995). Amongst all high-grade glioma patients, a major advance in the treatment of highgrade gliomas occurred when Valtonen et al. (1997) found that median time from surgery to death in primary, high-grade glioma patients was 58.1 weeks for patients who received carmustine-loaded biodegradable polymers delivering localized chemotherapy in surgically resection tumor beds versus 39.9 weeks for those who only underwent surgical resection. More specifically patients with glioblastoma multiforme, the most deadly form of high-grade gliomas, survived 53.3 weeks compared to 39.9 weeks in those given carmustine-loaded wafer with resection, as compared to those undergoing resection alone. Truly multimodal therapy, incorporating extensive surgical resection along with both adjuvant

176

D. Mukherjee and A. Quinones-Hinojosa

temozolamide chemotherapy and adjuvant radiotherapy demonstrated a further significant improvement in glioblastoma survival through the Stupp et al. (2005) report, which has now become standard of care for all patients with glioblastoma multiforme. Additionally, recent retrospective work has demonstrated that combination surgical resection and both local and systemic chemotherapy may significantly improvement outcomes among these patients as well (McGirt et al., 2009b). In a recent analysis, glioblastoma multiforme patients receiving extensive surgical resection along with both local carmustineimpregnated wafer chemotherapy and systemic temozolamide chemotherapy had 9 months of greater survival relative to counterparts only receiving surgical resection and localized wafer chemotherapy; the two groups did not have significantly different rates of complication (Fig. 19.1). Thus, surgical resection coupled with carmustine-impregnanted, biodegradable wafers, systemic temozolamide, and radiation therapy are currently considered the most effective treatment modalities available for treating high-grade gliomas. Additionally approaches to modifying multimodal therapy, such as using temozolomide in a dose-dense relationship in patient with MGMT silencing, or using temozolomide in combination with MGMT inhibitors (such as O6 -benzylguanine), are currently under further investigation (Quinn et al., 2009).

The Impact of New Post-operative Deficits on Survival

Fig. 19.1 Kaplan-Meier plot of survival after primary resection of glioblastoma multiforme and radiotherapy in patients ≤ 70 years old. Patient receiving concomitant temozolamide according to the Stupp protocol in addition to Gliadel wafer

implantation demonstrated improved survival relative to those receiving Gliadel wafters and radiotherapy alone (n = 78). Mean survival was 21.3 versus 12.4 months, respectively (p = 0.005). Source: McGirt (2009b)

While studies have suggested an association between greater degree of resection among malignant astrocytomas and improved survival, the negative effects that new post-operative deficits, potentially induced by greater extensive resection, may have upon survival is less clear. A single retrospective analysis of 306 patients undergoing primary resection for glioblastoma multiforme was conducted at a single, high-volume institution to assess the relationship between new post-operative motor or language deficits on survival (McGirt et al., 2009a). Even after adjusting for patient age, preoperative KPS score, adjuvant therapy, and extent of resection, those patients with newly acquired motor and language deficits had reductions in median survival of 3.9- and 3.3-months, respectively, relative to their deficit-free counterparts (Fig. 19.2). Those patients who only demonstrated transient deficits in the immediately post-operative period had a smaller decrease in median survival relatively to those who developed permanent deficits. Whether this association was one of causation or simple statistical association currently remains unknown given the retrospective nature of the study (McGirt et al., 2009a). Although a definitive mechanism to explain the observed association is not yet known, past work

19 Impact of Extent of Resection on Outcomes in Patients with High-Grade Gliomas

177

Fig. 19.2 Kaplan-Meier plot demonstrating survival after resection of glioblastoma multiforme in patients without a new post-operative neurological deficit (mean survival 12.8 months), with a surgically acquired language deficit (mean survival 9.6 months), or with surgically acquired motor deficits (mean

survival 9.0 months). Those with new deficits experienced significantly worse survival (p60 Gy the risk of radiation related necrosis increases (Schultheiss et al., 1995). Yaacov et al. (2010) found a 5% risk of necrosis for at BED of 120 Gy (100–140 Gy) and a 10% risk at BED of 150 Gy (140–170 Gy) for fraction sizes < 2.5 Gy (equivalent to 72 and 89 Gy in 2 Gy per day fractions respectively, α/β= 3). This risk rises significantly for twice daily fractionation and becomes unpredictable for fractionation >2.5 Gy. Based on their analysis, they conclude dose, fraction size, and treatment volume are major players in the risk of necrosis, while location, use of chemotherapy, and, diabetes mellitus are less well established risk factors. They estimate a median time of 1–2 years between initial radiation and development of radiation necrosis.

218

The development of diffuse white matter injury is also a potential part of the late effect syndrome and has been documented after both whole brain radiation and focal radiotherapy. For the latter, the changes can extend beyond the peripheral of the high dose volume. They found resulting neurological sequelae ranging from mild lethargy to progressive memory loss leading to dementia. Histopathologically the damaged white matter shows reactive astrocytosis and white matter pallor. White matter injury resulting in cognitive changes seems to be related to volume, dose and fraction size. Leukoencephalopathy describes diffuse white matter injury in the setting of chemotherapy with or without radiation. It is noted in adults beyond 12–18 months and has a spectrum of symptoms with lethargy, dysarthia to progressive severe CNS damage resulting in ataxia, memory loss, and dementia. Imaging changes are similar to those described above and differ with late development of calcifications mostly limited to the basal ganglia and grey-white matter interface. Histopathology analysis show diffuse gliosis and multiple non-inflammatory foci of necrosis (Schultheiss et al., 1995). Late cerebrovascular effect causing stroke like deficits as a result of obliteration of vessels is of less concern in patients with malignant gliomas as these side effects are experienced by those who experience long survivorship, more commonly documented in children and for those receiving treatments for base of skull tumors.

Recovery of Occult Injury and Re-irradiation The recovery of occult radiation injury is important in understanding the safety of retreatments and determination of re-irradiation dose. Rodent studies have shown that the spinal cord is able to recover significant and that the extent of repair is dependant on the initial dose received and time interval to retreatment (Bauman et al., 1996). Studies on monkeys suggest that most of the occult injury induced by an initial treatment is recovered within 2 years; however, there is variability in the rate of recovery (Schultheiss et al., 1995). In this study, the time to late toxicity was similar in both the retreatment and single treatment group.

A. Hallock and L. VanderSpek

However, a retrospective clinical experience of 300 cases of radiation myelopathy from 76 papers showed that latency period for development of toxicity was reduced for the lumbar spine and with increasing doses of radiation. For the brain, given the lack of similar studies, an estimate of 40–50% of the previous radiation dose after 1–2 years maybe accounted for in retreatments (Bauman et al., 1996). It is important to recognize the inadequate understanding of impact of high dose per fractions in setting of retreatments on the brain. Both a recent literature review by Mayer et al. (2009) and Yaacov et al. (2010) highlight the limitations of the present evidence evaluating the sequelae of reirradiation. Most clinical data is derived from single institution studies that often lack details in target volumes, spatial factors or time between radiation treatments. Reporting toxicity as only a crude rate (ratios to number of patients treated) instead of an actuarial risk (ratios to number of survivors at risk) can result in underestimation of true risks (Yaacov et al., 2010). These limitations acknowledged, Mayer et al. (2009) evaluated 21 studies using techniques of radiosurgery and FSRT and conventionally fractionated radiotherapy. They concluded that radiation can be given safely using different techniques with necrosis becoming evident at (normalized total doses) NTDcumulative >100 Gy. The NTDcumulative is the total dose delivered in 2 Gy per fraction using the linear quadratic formulation (accounting for differences in fraction size and total dose) to sum radiation courses. In their review, the major factor contributing to risk of necrosis was total dose and the time interval to retreatment (minimum interval of 3 months) did not significantly correlate to risk of necrosis. Previous published reports on salvage therapies including those for radiosurgery and brachytherapy estimate a necrosis of 10–20% (Bauman et al., 1996) and ranging from 4 to 40% for SRT or 3DCRT (Butowki et al., 2006). The timelines of progression to these symptoms in the setting of reirradiation is not well established particularly for those patients retreated for malignant glioma where competing risk of tumor progression may underestimate the potential for long term toxicity. As systemic therapies combined with focal therapies become a more common component of salvage regimes; it is important to gain an understanding of the impact of both cytotoxic chemotherapy and targeted agents on both acute and late toxicities in the setting of retreatments.

23 Recurrent Malignant Glioma Patients

Re-irradiation of Malignant Glioma There is no consensus on the general management of recurrent malignant gliomas or re-irradiation for recurrence. No comparative studies of the salvage options of surgery, radiation, or systemic therapies are available in the literature. Feasibility of salvage option (i.e. resectability and operability when considering reoperation), expected therapeutic response, and potential for cumulative toxicity should be guiding principles in the management of patients with recurrent malignant gliomas. Figure 23.1 provides a suggested approach. Conventional, fractionated stereotactic radiotherapy, brachytherapy, radiosurgery (Linac based, GammaKnife, and Cyberknife) have been utilized in the salvage setting; the highly focal techniques allow the possibility of maximizing therapeutic efficacy, tolerance, and convenience. Radiosurgery for progressive malignant glioma is the subject of another chapter.

219

For this section of the chapter, a literature search for re-irradiation studies in recurrent malignant gliomas between 2000 and 2010 was conducted. Studies with less than ten patients, involving brachytherapy, radiosurgery, or inadequate information about treatment regimes were excluded. Tables 23.2, 23.3 and 23.4 lists these retrospective, prospective and phase I trials and are provided as a guide to the doses, time intervals and regimes used and for an appreciation of tolerances of normal brain. It is important to reemphasize the challenges in making comparison and caution against making overarching general conclusions based on these studies. These studies span eras that have seen significant advancements in technology such as improved neuro-imaging and neurosurgical techniques, have inherent issues of selection biases, and differing institution treatment practices. Also, there is no gold standard for imaging changes to diagnose true normal tissue necrosis; many of the studies did not grade their necrosis or report on re-operation rates or confirm suspected necrosis with biopsy.

Fig. 23.1 A suggested approach to the management of recurrent malignant glioma

220

A. Hallock and L. VanderSpek

Figure 23.2 shows these studies in a temporal relationship, with median overall survival from time of retreatment.

calculation for specific patient scenarios, the following formulas have been provided (Table 23.1): NTDcumulative = BEDcumulative /2;

3D Conformal/Stereotactic Radiotherapy As stated earlier in this chapter, the customization of salvage radiation treatment regime including radiation technique, total dose, and dose per fraction for individual patients should account for initial dose with first treatment, time interval to retreatment, new treatment volume, and when appropriate consider placement in a clinical trial. As suggested by Mayer and Sminia (2009), attention should be paid to the NTDcumulative . They found that both NTDcumulative and total retreatment dose can be increased when using conformal techniques or radiosurgery, without added risk of necrosis. Based on their review, the range of NTDcumulative for the conventional technique re-irradiation was 81.6– 101.9 Gy; for FSRT was 90–133.9 Gy. To aid in

and BEDcumulative = BEDinitial + BEDreirradiation BED = nd(1 + d/[α/β]) D = dose/fraction (Gy); n = number of fractions; α/β = repair capacity of tissue. Table 23.2 includes studies that have evaluated treatments of recurrence with different radiation techniques and suggest that re-irradiation can be delivered safely by various techniques. There is significant heterogeneity in treatments that patients received within institutions in terms of prior chemotherapy regimes, resections, etc. The increased exploration of FSRT started after some early radiosurgery studies suggested high rates of reoperation for necrosis. FSRT allows for

22

Fig. 23.2 Scatter plot of re-irradiation studies in recurrent malignant gliomas. Median survival is from time of retreatment. XRT (conventional radiation), FSRT (fractionated stereotactic radiotherapy), CTX (chemotherapy), TG (targeted therapy)

Median Overall Survival

20 18 XRT XRT+CTX FSRT FSRT+CTX FSRT+TG

16 14 12 10 8 6 2000

Table 23.1 BED for examples of common radiation regimes Dose/fraction BEDα/β=10 BEDα/β=2 Dose (Gy) Fractions (Gy) (Gy) (Gy)

2005 Year of Study

NTD (Gy)

2010

Comments

60 30 2 72 120 60 Standard initial treatment 35 7 5 53 122.5 61 Hypofractionated Re-XRT 20 1 20 60 220 110 Radiosurgerya a The linear quadratic model was derived for fractionated radiotherapy. Its reliability for small number of fractions or single fractions is uncertain. There are models suggested for radiosurgery but these do not account for all possible variables, and are not well validated.

45–61 Gy/1.8–2 Gy daily OR 45 Gy/3 Gy daily OR 54.8 Gy/1.8 Gy bid+CT (94%) 60 Gy/2 Gy per day/5 days per week +CT (16 patients) 59.4 Gy/2 Gy per day/5 days per week +CT (16 patients) 45–60 Gy/Dose/fx: NA

50–60 Gy/2 Gy per day/5 days per week

30 Gy/5 Gy per day/5 days per week

46 Gy/2 Gy per day/5 days per week

19

33

11

15

NA

12

7.9 15.4 (AA)

6.9

8

Acute edema (1)/necrosis (2): clinical necr + cognitive decline (1) at 62 months/all late toxicity at BED> 204 Gy. Dose >40 Gy major risk factor. Necrosis(2) Mixed (tumor +radiation) necrosis (1)

GBM(51) 36 Gy/2 Gy per day/5 10 49 8 Necrosis(0) GS(1) days per week GCG(1) Combs et al. (2005a) AA(22) 36 Gy/2 Gy per day/5 34 56.2 16 Necrosis (0) AO(10) days per week AOA(8) Ernst-Stecken et al. GBM(11) 35 Gy/5 Gy per day/5 10 22.4 12 Necrosis (0) (2007) AA(4) days per week AA (anaplastic astrocytoma), AO (anaplastic oligodendroglioma), AOA (Anaplastic oligoastrocytoma), AE (Anaplastic Ependymoma), GBM (glioblastoma multiforme), GS (gliosarcoma), GCG (giant cell glioblastoma), LG (low grade), DE (Dose escalation), NA (not available), CT (chemotherapy), fx (fraction). a Median interval is from the time from the first radiation treatment. b Median survival is from the time of retreatment.

Vordermark et al. (2005)

Combs et al. (2005b)

GBM(14) AA(5)

Veninga et al. (2001)

20-36 Gy/4–6 Gy per day/5 days per week

Necrosis(0)

60 Gy/2 Gy per day/5 days per week

Selch et al. (2000)

GBM(15) AOA(3) LG(3) Unk(1) AA(29)

Graded toxicity(n)/necrosis (n)/comments

Table 23.2 Re-irradiation with conventional radiotherapy, fractionated stereotactic radiotherapy, or LINAC based stereotactic radiosurgery First treatment radiation Median Median Median (median or range)/dose/ Median re-radiation/ volume intervala survivalb Author Histology (n) (months) (months) schedule dose/schedule treated (cc)

23 Recurrent Malignant Glioma Patients 221

222

delivery for high dose per fraction treatment while taking radiobiological advantage for normal tissue repair and maintaining short overall treatment times. Cho et al. (1999) compared FSRT to SRS and found lower complication rates with FSRT without impact on survival rates. Other studies of FSRT given alone or in conjunction with systemic therapies (Table 23.3) support the results of Cho et al. (1999). Given high progression post retreatments, newer trials are exploring the role of dose escalation and enhancing radiosensitization.

Stereotactic Radiotherapy with Systemic Therapies Brain tumors provide unique challenges to treatment with systemic therapies alone with issues of diffuse infiltrative disease, blood brain barrier, and rapid evolution of resistance. There is only modest benefit to conventional chemotherapy in treatment for recurrent malignant gliomas, with greater benefit observed in patients with grade three gliomas. Chemotherapeutic agents used for recurrent gliomas include temozolomide, nitrosourea, carboplatin, procarbazine, irinotecan, etoposide and carmustine wafers (Wen and Kesari, 2008). Concurrent chemoradiation is becoming increasing common as evidenced by its use in radical treatments of gynecologic, gastrointestinal, head & neck, and thoracic malignancies. The addition of chemotherapy to repeat radiation treatment of recurrent malignant glioma is another treatment strategy and has been described in the literature for several chemotherapeutic agents (Table 23.3). Chemotherapy may interact with radiation through several mechanisms: spatial cooperation, additive or synergistic cell killing or non-overlapping toxicity. In the retreatment of glioma, chemotherapy may potentially address microscopic tumor outside the radiated volume. Improved tumor cell kill within the treated volume may result from the additive effects of chemotherapy and radiation and allow for treatment intensification without the increased risk of necrosis seen when higher dose re-irradiation is used as a single modality due to non-overlapping toxicity. Synergistic effects can include alteration of cell cycle

A. Hallock and L. VanderSpek

by chemotherapy resulting in greater number of cells within the radiosensitive G2/M phase, inhibition of radiation repair and conversion of sub-lethal single strand radiation injury to lethal double strand breaks. The role of Cisplatin (CDDP), Paclitaxel, Lomustine, Topotecan, Temozolomide and hyperbaric oxygen with radiation in recurrent gliomas have been reported (Table 23.3). Cisplatin (CDDP) is used as a radiosensitizer in several tumor types (anal, cervical, head and neck, bladder), but has shown only limited benefit in the setting of re-irradiation for malignant glioma (VanderSpek et al., 2008; Glass et al., 1997). Topotecan has been shown to be efficacious in combination with radiation, both because of its synergistic effect and substantial CNS penetration. Temozolomide is an oral alkylating agent that readily crosses the blood brain barrier that has had the greatest impact on treatment of GBM (Stupp et al., 2005), and its efficacy was first found in recurrent gliomas. Continuous administration of this temozolomide appears to "deplete" the DNA repair enzyme MGMT (O6 methylguanine-DNA methyltransferase), which is increased during radiotherapy (Stupp et al., 2005). Despite these biologic correlations, clinical studies of combined modality treatment for recurrent glioma suggest limited benefit. While the morbidity experienced with combined use of systemic therapy and re-irradiation seems acceptable there do not appear to be any apparent gains in disease control or outcomes compared to re-irradiation alone (Table 23.3, Fig. 23.2). Given the limited benefit of conventional systemic therapies, there is great interest in the role of targeted molecular agents and antiangiogenic agents as our understanding of the molecular pathogenesis of gliomas increases. For malignant gliomas, agents that target Endothelial Growth Factor Receptor (EGFR), Platelet Growth Factor Receptor (PDGFR), Vascular Growth Factor Receptor (VEGFR), mammalian target of rapamycin (mTOR), farnesyltransferase, and P13k are of interest (Wen and Kesari, 2008). Malignant gliomas are known to be highly vascular tumors, expressing VEGF as an important angiogenesis factor and have some response to older, likely less potent antiangiogenic agents like Thalidomide. It is hypothesized that the normalization of abnormal blood vessels with anti-angiogenic therapy targeting

GBM (5) AA (7) AO(2)

HG (25)

60 Gy/1.8–3 Gy 30 Gy/5 Gy per per day/5 days day/5 days per per week + week chemo (20 pts)

NA

20-30 Gy (2 × 1.2 Gy per day) 60 Gy/2 Gy per 30 Gy/2 Gy per day/5 days per day/5 days per week week

Median radiation/ dose/schedule

TMZ (66%)

TMZ

TMZ

Chemotherapy

16

14

NA

Median intervala (months)

11 (received functional imaging)

163 (PTV)

NA

Median volume treated (cc)

20

7.5

9.3

Median survivalb (months)

No G 3 or 4 toxicities/necrosis(0) Grade 3 hematologic (1) Cephalgia(1) Mental degradation (1,query tumor progression) Late toxicity: NA No acute >grade 3 neuro toxicity; hematological toxicity:NA Mixed necrosis (3) Improved survival with PET (SPECT /CT/ MRI compared to CT/ MRI alone; 9 vs. 5 months) 84% previous chemotherapy for recurrence

Graded toxicity/ necrosis/comments

60 Gy/2 Gy per 25–30 Gy/5–6 Gy Topotecan NA 14.5 16.5 day/5 days per per day week OR 54.4 Gy/1.6 Gy twice daily/3.5 weeks Combs et al. GBM(8) 60 Gy/2 Gy per 36 Gy/2 Gy per TMZ 36 50 8 No severe toxicities (2008) HG(10) day/5 days per day/5 days per (PTV) Necrosis(0) week week GBM (glioblastoma multiforme), AA (anaplastic astrocytoma), AO (anaplastic oligodendroglioma), AA (anaplastic astrocytoma), AO (anaplastic oligodendroglioma), AOA (Anaplastic oligoastrocytoma), GS(gliosarcoma), GCG (giant cell glioblastoma), HG (high grade), LG (low grade), TMZ (Temozolomide), NA (not available) a Median Interval is from time from the first radiation treatment. b Median Survival is from the time of retreatment.

Wurm et al. (2006) GBM(18) GCG(2) AA(3) AOA(2)

Grosu et al. (2005) GBM(33) GS(2) AA(8) AO(1)

Schonekaes et al. (2002) Schafer et al. (2004)

Table 23.3 Reirradiation and chemotherapy First treatment r radiation (median or range)/Dose/ Author Histology (n) schedule

23 Recurrent Malignant Glioma Patients 223

224

either VEGF or VEGFR will increase the therapeutic benefit of radiation and chemotherapies. While it is thought that glioma stem cells produce VEGF, use of these single agent VEGF inhibiters have shown have limited benefit (Wen and Kesari, 2008). This has prompted evaluation of these agents in combination with chemotherapy. Work from Friedman et al. (2009) has resulted in the approval of bevacizumab in combination with irinotecan for use in recurrent GBM. Given these findings, potential role of targeted therapies with radiation is exciting especially for brain tumors, but we are still in the preliminary stages of our understanding. There is a suggestion that promotion of angiogenesis may occur during radiation, implying the potential benefit of combining anti-angiogenic agents during radiotherapy. In vivo and vitro studies show that exposure to radiation induces VEGF expression in several tumor cell lines and that an anti-VEGF treatment potentiates radiation induced tumor kill (Gorski et al., 1999). Blocking of VEGF-receptor 2 potentiates radiation induced tumor control in vivo human tumor xenograft of glioblastoma multiforme (U87) (Gutin et al., 2009). Gutin et al. (2009) studied the benefits of combination of 7 cycles of bevacizumab with FSRT in recurrent gliomas and reported a 6 months PFS of 65%, with 50% overall response rate. However, three out of 25 patients discontinued treatments due to grade 3 toxicities of intratumoral hemorrhage, wound dehiscence and bowel perforation (but there was no noted necrosis). Evaluation of responses of anti-angiogenic therapies may be complicated by the decreased enhancement on imaging seen with these drugs, independent of their anti-tumor effect. Thus trials evaluating these agents should include survival endpoints (such as 6 month survival) as a complement to radiographic endpoints. Another interesting advantage of combining anti-angiogenic agents is the potential reduction in necrosis suggesting that a potential protective role for this agent in combination with re-irradiation, perhaps sequentially administered given the toxicity seen with concurrent administration in the series by Gutin et al. (2009). Endothelial Growth Factor Receptors (EGFR) are involved in pathways known to cell differentiation, angiogenesis, metastatic spread, resistance to apoptotic death and thought to be expressed highly in cells resistant to radiation and chemotherapy. The majority

A. Hallock and L. VanderSpek

of glioblastomas show a deregulation as a result of mutations, overexpressions and gene amplifications in the molecular pathway involving EGFR. Small molecule inhibitors targeted against the EGFR tyrosine kinase pathways or the EGFR receptor has been evaluated in glioblastoma. For example, the epidermal growth factor receptor tyrosine kinase inhibitor (EGFR TKI), Gefitinib (ZD1839 or Iressa), has demonstrated a median survival for patients with recurrent GBM or 39.4 weeks when used as a single agent after first relapse (Schwer et al., 2008). Combination of the EGFR TKI with radiation can have potentiating effects through greater number of cells in the radiosensitive phase of the cell cycle; greater radiation induced apoptosis and decreased DNA repair mechanisms within the cancer cells. The addition of molecular targeted therapies with re-irradiation provides an exciting new therapeutic option for patients with recurrent malignant gliomas. There are too few studies to make any meaningful conclusions and even these studies report outcomes comparable other studies for salvage therapies (Table 23.4). At present, there is no clear evidence for the standard use of targeted therapies with re-irradiation. This suggests the need for further exploration into their use and identifying patients who may benefit such therapies. For example, EGFR TKI agents are small molecules and should have a theoretical benefit in terms of intracranial delivery, however, only 10-20% of GBM respond to these targeted agents. Mellinghoff et al. (2005) showed that the sensitivity of GBMs to these EFGR inhibitors was based on co-expression of EGFRvIII and PTEN, implying that therapeutic benefit may not be for all patients. There are currently phase 2 trials evaluating multi tyrosine kinase inhibitors underway to determine if therapeutic outcomes can be improved by targeting multiple molecular abnormalities. Consideration of enrolment of patients in trials evaluating molecular targeted therapies with radiation is recommended to help establish clear benefits.

Fractionated Stereotactic Reirradiation Techniques The radiation planning volumes for FSRT used in reirradiation of malignant glioma need to take into account the proximity of the lesion to eloquent areas

GBM(11) AA(4)

60 Gy/1.8–2 Gy per day/5 days per week + TMZ 18–36 Gy/6–12 Gy per day/ consecutive days

Median radiation/dose/ schedule Gefitinib (Iressa)

Target therapy 12

Median intervala (months)

41 (PTV)

Median volume treated (cc)

GBM(20) 59.4 Gy 30 Gy/6 Gy per Bevacizumab 15 34 AA(4) Fractionation: NA day/5 days per (Avastin) AO(1) week GBM (glioblastoma multiforme), AA (anaplastic astrocytoma), AO (anaplastic oligodendroglioma), AO (anaplastic oligodendroglioma) a Median Interval is from the time of first radiation treatment. b Median Survival is from the time of retreatment.

Gutin et al. (2009)

Schwer et al. (2008)

Table 23.4 Re-irradiation with molecular targeted agents First treatment radiation (median or range)/dose/ Author Histology (n) schedule

12.5

10

Median survivalb (months)

Necrosis (2): Grade 1; BCNU wafers given postoperative at recurrence for 5 patients Grade 3 CNS toxicity Necrosis(0)

Graded toxicity/ radiation(n)/comments

23 Recurrent Malignant Glioma Patients 225

226

of the brain and location of the recurrence in relation to the previously radiated volume, as well as the method of immobilization, treatment planning system, and delivery techniques available at each institution which influence the accuracy of treatment delivery (Bauman et al., 2006). The typical planning process involves fabrication of any immobilization devices followed by acquisition of a volumetric computed tomotherapy (CT) scan with the patient in the immobilization device for radiation treatment planning purposes (called CT simulation). CT simulation scans are typically acquired using 3 mm slices thickness and with intravenous contrast administration for optimal delineation of the target. Typically a stereotactic image fusion is performed between the planning CT and diagnostic imaging (MRI or PET/SPECT). Most commonly the diagnostic highresolution T1-weighted MRI with Gadolinium contrast is used to assist in target volume delineation. The tumor visible on imaging is defined as the Gross Tumor Volume (GTV) and serves as the target for radiation treatment planning. For reirradiation, delineation of a region of risk (clinical target volume, CTV) surrounding the GTV (as usually performed for initial treatment to cover regions of microscopic infiltration) is omitted in order to minimize the retreated volume. The planning target volume (PTV) is intended to account for geometric errors in setup and is dependent on the immobilization system used. For example for traditional radiosurgery systems, invasive stereotactic frame systems with submillimeter localization accuracy are used, justifying the omission of a PTV margin. For relocatable frame systems, selection of a PTV margin is dependent on the system used and institutional practice. For instance, Hudes et al. (1999) defined the target volume as the contrast-enhanced tumor edge (GTV) without a PTV margin for patients immobilized in the GTC (Gill-Thomas-Cosman) frame. Others utilize a margin around the GTV, such as Schwer et al. (2008) where the PTV was defined as the GTV + 2 mm as they used a technique with patients immobilized using a custom-molded removable head frame and had documented reproducibility within 2 mm. Overall the choice of a PTV margin depends on the measured accuracy of the overall radiation delivery at each institution as well as the proximity to organs at risk and clinical judgment (Bauman et al., 2006). Acknowledgements The authors would like to thank Dr. Glenn Bauman, London Regional Cancer Program, University of

A. Hallock and L. VanderSpek Western Ontario, for his constructive feedback during the development of the chapter.

References Bauman GS, Sneed PK, Wara WM, Staplers LJA, Chang SM, McDermott MW, Gutin PH, Larson DA (1996) Reirradiation of primary CNS tumors. Int J Radiat Oncol Biol Phys 36:433–441 Bauman G, Wong E, McDermott MW (2006) Fractionated radiotherapy techniques. Neurosurg Clin N Am 17:99–110, v, 20 Butowski NA, Sneed PK, Chang SM (2006) Diagnosis and treatment of recurrent high-grade astrocytoma. J Clin Oncol 24: 1273–1280 Chan JL, Lee SW, Fraass BA, Normolle DP, Greenberg HS, Junck LR, Gebarski SS, Sandler HM (2002) Survival and failure patterns of high-grade gliomas after threedimensional conformal radiotherapy. J Clin Oncol 20: 1635–1642 Chang EL, Akyurek S, Avalos T, Rebueno N, Spicer C, Garcia J, Famiglietti R, Allen PK, Chao C, Mahajan A, Woo SY, Maor MH (2007) Evaluation of peritumoral edema in the delineation of radiotherapy clinical target volumes for glioblastoma. Int J Radiat Oncol Biol Phys 68:144–250 Chang SM, Butowski NA, Sneed PK, Garner IV (2006) Standard treatment and experimental targeted drug therapy for recurrent glioblastoma multiforme. Neurosurg Focus 20:E4 Cho KH, Hall WA, Gerbi BJ, Higgins PD, McGuire WA, Clark HB (1999) Single dose versus fractionated stereotactic radiotherapy for recurrent high-grade gliomas. Int J Radiat Oncol Biol Phys 45:1133–1141 Combs SE, Bischof M, Welzel T, Oertel S, Debus J, SchulzErtner D (2008) Radiochemotherapy with temozolomide as re-irradiation using high precision fractionated stereotactic radiotherapy (FSRT) in patients with recurrent gliomas. J Neurooncol 89:205–210 Combs SE, Debus J, Schulz-Ertner D (2007) Radiotherapeutic alternatives for previously irradiated recurrent gliomas. BMC Cancer 7:167–177 Combs SE, Gutwein S, Thilmann C, Debus J, Schulz-Ertner D (2005a) Reirradiation of recurrent WHO grade III astrocytomas using fractionated stereotactic radiotherapy (FSRT). Strahlenther Onkol 181:768–773 Combs SE, Gutwein S, Thilmann C, Huber P, Debus J, Schulz-Ertner D (2005b) Stereotactically guided fractionated re-irradiation in recurrent glioblastoma multiforme. J Neurooncol 74:167–171 Dhermain F, Ducreux D, Bidault F, Bruna A, Parker F, Roujeau T, Beaudre A, Armand JP, Haie-Meder C (2005) Use of functional imaging modalities in radiation therapy treatment planning in patients with glioblastoma. Bull Cancer 92:333–342 Ernst-Stecken A, Ganslandt O, Lambrecht U, Sauer R, Grabenbauer G (2007) Survival and quality of life after hypofractionated stereotactic radiotherapy for recurrent malignant glioma. J Neurooncol 81:287–294 Friedman HS, Prados MD, Wen PY, Mikkelsen T, Schiff D, Abrey LE, Yung WK, Paleologos N, Nicholas MK,

23 Recurrent Malignant Glioma Patients Jensen R, Vredenburgh J, Huang J, Zheng M, Cloughesy T (2009) Bevacizumab alone and in combination with irinotecan in recurrent glioblastoma. J Clin Oncol 27: 4733–4740 Glass J, Silverman C, Axelrod R, Corn B, Andrews D (1997) Fractionated stereotactic radiotherapy with cis-platinum radiosensitization in the treatment of recurrent, progressive, or persistent malignant astrocytoma. Am J Clin Oncol 20:226–229 Gòmez-Rio M, Rodriguez-Fernàez A, Ramos-Font C, LòpezRamirez E, Llamas-Elvira JM (2008) Diagnostic accuracy of 201 Thallium-SPECT and 18 F-FDG-PET in the clinical assessment of glioma recurrence. Eur J Nucl Med Mol Imaging 35:966–975 Gorski DH, Beckett MA, Jaskowiak. NT, Calvin DP, Mauceri HJ, Salloum RM, Seetharam S, Koons A, Hari DM, Kufe DW et al (1999) Blockage of the vascular endothelial growth factor stress response increases the antitumor effects of ionizing radiation. Cancer Res 59:3374–3378 Grosu AL, Weber WA, Franz M, Stark S, Piert M, Thamm R, Gumprecht H, Schwaiger M, Molls M, Nieder C (2005) Reirradiation of recurrent high-grade gliomas using amino acid PET (SPECT)/CT/MRI image fusion to determine gross tumor volume for stereotactic fractionated radiotherapy. Int J Radiat Oncol Biol Phys 63:511–519 Gutin PH, Iwamoto FM, Beal K, Mohile NA, Karimi S, Hou BL, Lymberis S, Yamada Y, Chang J, Abrey LE (2009) Safety and efficacy of bevacizumab with hypofractionated stereotactic irradiation for recurrent malignant gliomas. Int J Radiat Oncol Biol Phys 75:156–163 Halperin EC, Bentel G, Heinz ER, Burger PC (1989) Radiation therapy treatment planning in supratentorial glioblastoma multiforme: an analysis based on post mortem topographic anatomy with CT correlations. Int J Radiat Oncol Biol Phys 17:1347–1350 Hochberg FH, Pruitt A (1980) Assumptions in the radiotherapy of glioblastoma. Neurology 30:907–911 Hudes RS, Corn BW, Werner-Wasik M, Andrews D, Rosenstock J, Thoron L, Downes B, Curran WJ Jr (1999) A phase I dose escalation study of hypofractionated stereotactic radiotherapy as salvage therapy for persistent or recurrent malignant glioma. Int J Radiat Oncol Biol Phys 43:293–298 Laperriere NJ, Leung PMK, McKenzie S (1998) Randomized study of brachytherapy in the initial management of patients with malignant astrocytoma. Int J Radiat Oncol Biol Phys 41:1005–1011 Mayer R, Sminia P (2009) Reirradiation tolerance of the human brain. Int J Radiat Oncol Biol Phys 70:1350-1360 Mellinghoff IK, Wang MY, Vivanco I, Haas-Kogan DA, Zhu S, Dia EQ, Lu KV, Yoshimoto K, Huang JH, Chute DJ, Riggs BL, Horvath S, Liau LM, Cavenee WK, Rao PN, Beroukhim R, Peck TC, Lee JC, Sellers WR, Stokoe D, Prados M, Cloughesy TF, Sawyers CL, Mischel PS (2005) Molecular determinants of the response of glioblastoma to EGFR kinase inhibitors. N Engl J Med 353: 2012–2024

227 Oppitz U, Maessen D, Zunterer H, Richter S, Flentje M (1999) 3D-recurrence-patterns of glioblastomas after CT-planned postoperative irradiation. Radiother Oncol 53:53–57 Peca C, Pacelli R, Elefante A, Del Basso De Caro ML, Vergara P, Mariniello G, Giamundo A, Maiuri F (2009) Early clinical and neuroradiological worsening after radiotherapy and concomitant temozolomide in patients with glioblastoma; tumor progression or radionecrosis? Clin Neurol Neurosurg 111:331–334 Schafer U, Micke O, Schuller P, Schuck A, Willich N (2004) The effect of sequential radiochemotherapy in preirradiated malignant gliomas in a phase II study. J Neurooncol 67:233– 239 Schonekaes K, Mucke R, Panke J, Rama B, Wagner W (2002) Combined radiotherapy and temozolomide in patients with recurrent high grade glioma. Tumori 88:28–31 Schultheiss TE, Kun LE, Ang KK, Stephens LC (1995) Radiation response of the central nervous system. Int J Radiat Oncol Biol Phys 10:1109–1115 Schwer AL, Damek DM, Kavanagh BD, Gaspar LE, Lillehei K, Stuhr K, Chen C (2008) A phase I dose-escalation study of fractionated stereotactic radiosurgery in combination with gefitinib in patients with recurrent malignant gliomas. Int J Radiat Oncol Biol Phys 70:993–1001 Selch MD, DeSalles AA, Solberg TD, Wallace RE, Do TM, Ford J, Cabatan-Wong C, Withers HR (2000) Hypofractionated stereotactic radiotherapy for recurrent malignant gliomas. J Radiosurgery 3:3–12 Stupp R, Mason WP, Van den Bent MJ, Weller M, Fisher B, Taphoorn MJ, Belanger K, Brandes AA, Marosi C, Bogdahn U, Curschmann J, Janzer RC, Ludwin SK, Gorlia T, Allgeier A, Lacombe D, Cairncross JG, Eisenhauer E, Mirimanoff RO (2005) Radiotherapy plus concomitant and adjuvant temozolomide for glioblastoma. N Engl J Med 352:987–996 VanderSpek L, Fisher B, Bauman G, Macdonald D (2008) 3D conformal radiotherapy and cisplatin for recurrent malignant glioma. Can J Neurol Sci 35:57–64 Veninga T, Langendijk HA, Slotman BJ, Rutten EH, van der Kogel AJ, Prick MJ, Keyser A, van der Maazen RW (2001) Reirradiation of primary brain tumours: survival, clinical response and prognostic factors. Radiother Oncol 59:127– 137 Vordermark D, Kolbl O, Ruprecht K, Vince GH, Bratengeier K, Flentje M (2005) Hypofractionated stereotactic reirradiation: treatment option in recurrent malignant glioma. BMC Cancer 5:55–59 Wen PY, Kesari S (2008) Malignant gliomas in adults. N Engl J Med 359:492–507 Wurm RE, Kuczer DA, Schlenger L, Matnjani G, Scheffle D, Cosgrove VP, Ahswede J, Woiciechowsky C, Budach V (2006) Hypofractionated stereotactic radiotherapy combined with topotecan in recurrent malignant glioma. Int J Radiat Oncol Biol Phys 66:S26–S32 Yaacov RL, Li AX, Naqa IE, Hahn C, Marks LB, Merchant TE, Dicker AP (2010) Radiation dose-volume effects in the brain. Int J Radiat Oncol Biol Phys 76:S20–S27

Chapter 24

Glioblastoma: Boron Neutron Capture Therapy Tetsuya Yamamoto, Kei Nakai, and Hiroaki Kumada

Abstract Boron neutron capture therapy (BNCT) is a unique method that can deliver tumor-cell-selective high-linear energy transfer (LET) particle radiotherapy to an extended target area encompassing a microscopic invasion while avoiding radiation damage to the surrounding normal brain tissue. The process of BNCT is based on the nuclear interaction of 10 B with thermal neutrons with the release of high LET α and 7 Li particles through the boron neutron capture reaction, 10 B(n, α) 7 Li. The very short path length ( 1h BPA:700 mg/kg, 6 h Prescribed peak BSH:100 mg/kg,1 h dose 15 Gy >

BSH:100 mg/kg, 1 h

BPA: 290–500 mg/kg, 2h BPA: 900 mg/kg, 6 h

BPA: 250–330 mg/kg, 2h BPA: 250–350 mg/kg, 1–2 h BSHc : 100 mg/kg/min

14.1

23.5

26.9–65.4

ND

21.9 for 450 mg/kg cohort 14.2e

13.2 for 10.4 Gy cohortd

13 (1 field: 14.8, 2 fields: 12.1, 3 fields: 11.9) 12

16.3–63.0

ND

15.5–54.3

ND

ND

18–55 (data from 38 of 53 subjects) 7.3–24.8

Studsvik AB Sweden LVR-15, NRI Rez, Czech Republic KUR, KURRI and JRR-4, JAEA, Japan

Single fraction (No) Single fraction (No) Single fraction (No) Single fraction (20–30 Gy/10– 20 fraction) Single fraction (30 Gy/15 fr or 30.6 Gy/17 fraction)

Fir1, Helsinki, Finland

HFR, Petten, the Netherlands

MITR-II, M-67, MIT, USA

BMRR, BNL, USA

Reactor, institute, country

Single fraction (No)

4 fractions (No)

1 or 2 fractions (No)

Single fraction (No)

Neutron irradiation, (photon radiation)

University of 8 (1998–2007), 65 BPA:250 mg/kg,1 h 8.4–14.1/2.5–3.4 15.5–42.5 27.1/11.9 JRR-4, JAEA, Tsukuba and years BSH:5 g/body,1 h Japan JAEA, Phase I/II a Weighted dose D w b Includes 2 melanomas. Patients underwent BNCT at MITR-II c Four fractions on consecutive days, 100 mg/kg was administered the first day, and this dose was sufficient to keep the average blood concentration at 30 μg/g during treatment days 2–4 d Mean survival time for the 3rd dose group e Survival time calculated from the BNCT treatment

NRI Rez, Phase I

Harvard/MIT, Phase I

53 (1994–1999), (56.5 years for 1 field) 20 (1997–1999)b , 56 years

BNL, Phase I/II

Table 24.1 Summary of currently underway or recently completed external beam BNCT clinical trials for newly diagnosed GBM Number of evaluated Minimum tumor patients (year), Normal brain Median survival dose in GTVa median age of Boron drug: dose, dosea peak/ave. Trial (reference) (Gy) (months) (Gy) patients infusion time

24 Glioblastoma: Boron Neutron Capture Therapy 233

234

at the Studsvik BNCT facility for 29 patients suffering from GBM, who received 900 mg/kg BPA-F in a 6-h infusion. The minimum dose to the tumor volume and to the target volume (defined as tumor plus edema plus a 2 cm margin) ranged from 15.4 to 54.3 Gy and 8.8 to 30.5 Gy, respectively. Four patients developed grade 3–4 toxic events including epileptic seizures, hematuria, thrombosis and erythema. The median time from BNCT treatment to tumor progression was 5.8 months, and the MST after BNCT was 14.2 months, suggesting the survival benefit of long-time infusion compared to conventional 2-h infusion in BPA administration (Henriksson et al., 2008; Sköld et al., 2009). The combined use of BPA and BSH was based on experimental data that showed different uptake characteristics of the cell-cycle dependency of tumor cells (Yoshida et al., 2002). Other experimental data have also suggested that the combination of BNCT and photon radiation leads to significant gains in survival (Barth et al., 2005). In the trial conducted by Osaka Medical College, the first 10 patients suffering from GBM were administered 100 mg/kg of BSH and 250 mg/kg of BPA in a 1-h infusion (protocol 1), and the latter 11 patients were administered 100 mg/kg of BSH and 700 mg/kg of BPA in a 6-h infusion (protocol 2). X-ray irradiation was added in protocol 2 for a total dose of 20–30 Gy. The MSTs for all patients and for protocol 2 patients were 15.6 and 23.5 months, respectively (Kawabata et al., 2008). In the trial at the University of Tsukuba and Tokushima University at Japan Research Reactor No.4 (JRR-4) of the Japan Atomic Energy Agency (JAEA), a low dose (250 mg/kg) of BPA was administered over 1 h, and 5 g BSH/kg bodyweight was infused over 1 h in 8 patients with a single irradiation field. These patients received additional photon radiation to the volume which was defined as a peritumoral high intensity in the T2-weighted MRI after completion of BNCT. The median OS and the time to progression were 27.1 and 11.9 months, respectively. The 1-year and 2-year survival rates were 87.5 and 62.5%, respectively. The small group of patients showed good tolerance to the treatment and there were no severe acute or subacute adverse events. Although it is not certain whether additional photon irradiation could play a role in the clinical response to BNCT, the survival of this small cohort seems to be favorable evidence (Yamamoto et al., 2009).

T. Yamamoto et al.

Boron Compounds for Clinical BNCT Successful BNCT is dependent on the selective accumulation and absolute level of 10 B atoms in tumor cells. There are two boron delivery agents available for clinical BNCT trials for high-grade glioma, BPA and BSH (Fig. 24.1). Boron is an essential plant nutrient, and necessary for the optimal health of mammals as a rare element, that may be beneficial for bone growth and maintenance, central nervous system function, and the inflammatory response, though its physiological role in mammals is not yet fully understood. Natural boron consists of 80% 11 B and 20% 10 B, and the latter 10 B isotope has high efficiency in capturing thermal neutrons to generate the boron neutron capture reaction,10 B(n, α) 7 Li (Nielsen, 2009). BPA, the amino acid analogue, is remarkably nontoxic. Large doses have been administered to experimental animals without perturbing hematopoietic, hepatic or renal function (FDA-IND #43,317). BPA has structural characteristics similar to those of a melanin precursor, and promising clinical results were shown in the pilot study of BNCT for skin melanoma (Coderre et al., 2003). BPA is usually administrated intravenously as a soluble fructose complex, BPA-F, at doses ranging from 250 mg BPA/kg to 900 mg BPA/kg. BPA can penetrate across the blood-brain barrier into the normal brain, and is actively transported through the tumor cell membrane due to the elevated rate of amino acid transport in proliferating cells. The average macroscopic concentration of boron in tumor is 2–4 times higher than that in blood. However, non-proliferating cells escape the radiation damage following BNCT. The primary mode of BSH distribution is passive diffusion from blood to tumor tissue via the disrupted blood brain barrier. The boron concentration in the normal brain with an intact blood brain barrier remains minimal, while the tumor 10 B concentration is related to both the tumor vessel density and blood 10 B level. A tumor-to-blood boron concentration ratio ranging from 0.5 to 1.0 has been reported in human patients treated with BSH-mediated BNCT. The recent clinical trials have used a combination of BPA and BSH with the intention that these different compounds would accumulate in different subpopulations of tumor cells (Yamamoto et al., 2008).

24 Glioblastoma: Boron Neutron Capture Therapy

New Boron Compound and Delivery System Previous studies showed that a boron concentration of approximately 20–40 μg10 B/g tissue, or 109 10 B atoms/cell, could control the tumor progression. The coefficient rate of BSH is low, thus only 2% or less of the total amount administered participated in the neutron capture reaction in the irradiation field. Roughly translated, these results mean that these agents must be as safe as H2 O or glucose, because the required intracellular boron level could be achieved by drug administration on the order of grams. This is one of the reasons why the development of new boron compounds is difficult. The required characteristics for BNCT agents are as follows: (a) non-toxic at the clinically effective dose; (b) the achievement of at least 10–30 μg10 B/g of tumor; (c) high tumor/brain and tumor/blood concentration ratios; (d) rapid clearance from blood circulation and normal tissues but persistence in tumor; (e) water solubility; and (f) chemical stability (Coderre et al., 2004; Barth et al., 2005). To date, a variety of boron delivery agents have been investigated, including porphyrins, amino acids, polyhedral borane, carbohydrates, polyamines, biochemical precursors, DNA-binding agents, antisense agents, peptides, liposomes and monoclonal antibodies. Low molecular weight boron agents are peptides, porphyrins, amino acids such as BPA or carborane derivatives such as BSH. The high molecular weight group contains boronated antibodies and proteins. Liposomes are categorized independently of boron delivery systems. BOPP, a boronated porphyrin, and GB-10, polyhedral borane Na2 B10 H10 , have been evaluated for their pharmacokinetics in humans. A phase I trial for potential application of BOPP to photodynamic therapy was interrupted because of its general toxicity. Recently, improvement of the synthesis pathway and novel agents such as porphyrin-cobaltacarborane conjugates or 10 B-enriched carboranyl-containing phthalocyanine have been under examination. GB-10 has been proposed as a boron neutron capture agent for an additional dose in fast neutron therapy. In a trial in which GB-10 was used in 15 lung cancer patients, no major toxicity was recorded. Using the principle of antibody-antigen reaction for active tumor targeting, numerous molecular target

235

drugs were developed and are clinically used at present. Nevertheless, with respect to boron delivery, this method does not achieve a high cellular concentration of boron, even with the high selectivity and specificity. Most drug delivery systems (DDS) applied to BNCT are liposomal delivery systems. Such liposoms may include a solution containing a high concentration of boron, or boron-containing lipids. Active targeting is achieved by modifying the surface molecules. The current marketing of several liposomal anticancer agents and photodynamic therapy agents for macular degeneration shows the steady growth of the technical improvements in this field. Direct intratumoral injection and convectionenhanced delivery require the use of a pump to achieve a pressure gradient and establish the bulk flow of boron-containing agents during the interstitial infusion. Blood brain barrier disruption with hyperosmotic mannitol or receptor-mediated permeabilizer-7 (CMP7), have helped to achieve high tumor boron concentrations in experimental animals (Yang et al., 2009). Recent reports have revealed that preloading with LDOPA increases the uptake of BPA, that buthionine sulfoximine may similarly enhance the uptake of BSH, and that tumor growth is delayed following administration of either agents and neutron irradiation (Capuani et al., 2008; Yoshida et al., 2008).

Neutron Beam Facilities for BNCT The first clinical trials of BNCT for malignant brain tumors were performed at the Brookhaven Medical Research Reactor (BMRR) of BNL from 1959 to 1961. Another group of patients with malignant gliomas was treated using the reactor at MIT over the same period. Several research reactors, such as the Hitachi Training Reactor, Musashi Institute of Technology Research Reactor (MuITR), Kyoto University Research Reactor (KUR) and Japan Research Reactor No. 2 (JRR-2) in JAEA, have been applied to clinical studies in Japan. In these early BNCT trials, a low energy thermal neutron beam was used for the irradiation. To deliver thermal neutrons to deeper regions of the brain, the BNCT procedure included craniotomy, such as skin flap reopening and bone removal. Since 1994, external beam BNCT using epithermal neutron beams was initiated using the BMRR and the High Flux

236

T. Yamamoto et al.

Fig. 24.4 Cross-section of the neutron irradiation facility of Japan Research Reactor No. 4 (JRR-4) (upper) and process flow of the BNCT dose planning system JCDS and the setting simulation (lower)

Reactor (HFR) in Petten, the Netherlands. By using epithermal neutron beams the therapeutic range was expanded more deeply into the brain, compared to that of the thermal neutron beam irradiation, and with the application of the epithermal neutron beam it became possible to perform non-operative irradiation (Yamamoto et al., 2008). At present, neutron beams exclusively for use in BNCT are available at JRR-4 (JAEA, Japan), KUR (Kyoto University, Japan), LVR-15 (NRI Rez, Czech Republic), Studsvik AB (Sweden), Fir1 (Finland), HFR (Petten, the Netherlands), MITR-II and M-67 (MIT, USA), and the TRIGA reactor (Palva, Italy).

JRR-4 in JAEA can generate a thermal neutron beam as well as an epithermal neutron beam (Fig. 24.4). Neutron beam facilities consist of several neutron spectrum control devices, such as a heavy water tank, cadmium shutter, bismuth filter or collimator. The heavy water tank has four separated heavy water layers. The heavy water thickness in the tank can be changed by filling the corresponding layers with heavy water. And the cadmium shutter can be put in or out at the beam line. When the cadmium shutter is put at the beam line, the beam facility generates an epithermal neutron beam. The bismuth filter is installed at the beam line to reduce gamma-rays from the reactor

24 Glioblastoma: Boron Neutron Capture Therapy

core. The utilization of the heavy water tank and the cadmium shutter enables the generation of neutron beams appropriate for the patient’s disease condition. The BNCT clinical trial for malignant brain tumors using the facility of JRR-4 was initiated in 1999 using a thermal neutron beam, including craniotomy, and beginning in 2003 this treatment was switched to external beam BNCT with an epithermal neutron beam (Yamamoto et al., 2009). The epithermal neutron beam allowed the performance of clinical trials for other cancers, including head-and-neck and pulmonary tumors (Kankaanranta et al., 2007; Haginomori et al., 2008). Now the thermal neutron beam is used only for shallow-seated tumors such as skin melanomas.

Treatment Planning System The treatment planning for BNCT differs markedly from the planning for photon or electron irradiation in conventional radiotherapy and, in some ways, is significantly more complex. The boron neutron capture reaction provides a tumor-selective boron dose, 10 B(n, α) 7 Li, while other non-selective dose components are present and consist of proton recoils due to fast neutrons, 1 H(n, n )p, 0.54 MeV protons from the nitrogen capture reaction, 14 N(n, p)14 C, γ-rays arising from the contamination in the primary beam and 2.2 MeV prompt γ-rays from the hydrogen capture. Treatment planning for BNCT involves computation and analysis of these different components of radiation doses and their distribution in a patient to determine the neutron beam orientation and radiation fluence enabling the delivery of an optimized radiation dose and distribution. The optimized combination will comply with the dose prescription, optimize the dose to the target volume, and respect the dose limits to normal tissues and organs at risk. The therapeutic advantage in BNCT is obtained principally through the tumor-selectivity of a neutron capture agent (e.g., 10 B) rather than through precise geometric targeting of multiple well-collimated radiation fields onto the target volume. In conventional photon radiotherapy, only a single low-LET dose component must be computed, the dose from primary photons or electrons, both of which are ultimately delivered by electrons. In contrast, in BNCT the radiation field is a complex mixture of high- and low-LET dose components with varying biological

237

effectiveness that depends on both the tissue and the chemical form of the neutron capture agent. The spatial distribution of each dose component is different and depends on the tissue composition, as well as the neutron and photon fluence spectra. Three BNCT treatment planning systems (TPSs) have been reported and are being applied to clinical trials NCTPlan, SERA and JCDS (Nigg, 2003; Kumada et al., 2007). All three TPSs are noncommercial and have been developed by small research teams, usually with expertise in nuclear engineering. The NCTPlan was developed at MIT and is being applied for the treatment planning of BNCT at MITR and the RA-6 reactor in Argentina. SERA was developed by the Idaho National Engineering and Environmental Laboratory (INEEL) and Montana State University (MSU) in the USA. The system has been used for the planning of BNCT studies that are being performed at the FiR-1 reactor in Finland and KUR in Japan. JCDS was developed by JAEA to perform BNCT at JRR-4, and the system is still being developed in order to improve the dose calculation accuracy (Kumada et al., 2007; 2009). The treatment planning procedure using BNCT TPSs is almost the same as that of the conventional TPSs for photon radiotherapy, and the user interface of TPSs is also similar to that of photon radiotherapy. The algorithms generally used for dose calculations in clinical photon radiotherapy are computationally efficient, simple, and empirical. On the other hand, BNCT TPSs exclusively rely on a Monte Carlo simulation for dose calculations because of the complex, scatterdominated nature of the radiation transport processes involved. Figure 24.4 shows the outline of the procedure for BNCT treatment planning with JCDS. In the process of computational dosimetry, a patient’s CT and MRI images are first used to create a 3-dimensional model. The CT images are used to automatically differentiate compositions of the human body into bone, soft tissues and air according to their CT values. On the other hand, the MRI data are used to define several regions of interest (ROIs), such as the tumor region and organs at risk. By superimposing the MRI images onto the CT images, a detailed 3D model incorporating the information on body composition and ROIs is created. The latest version of JCDS can deal with PET images in addition to CT and MRI images to pick out tumor regions in accordance with the T/N ratio on PET images. To compute the distributions of several dose

238

components and neutron fluxes in the body effectively, the detailed 3D model is converted into a voxel calculation model. The initial version of the JCDS made the voxel calculation model by dividing the space into 10 × 10 × 10 mm3 voxel cells that contain the proper material data in each. JCDS has been improved continuously to perform more accurate dosimetry, and the current version of JCDS can produce a voxel calculation model of 2 × 2 × 2 mm3 voxel cells. The distributions of the dose components and fluxes in the voxel model are determined by Mote Carlo transport calculation. JCDS evaluates the detailed distributions of the dose components based on the calculation results, and outputs a two-dimensional distribution superimposed on CT or MRI. Finally, physicians decide the irradiation conditions for the patient based on the JCDS results. In conclusion, BNCT is the most complicated radiotherapy; however, it may one day be a valuable alternative treatment in cases of GBM that invades the surrounding healthy normal brain. BNCT seems to be preferable to currently existing high-dose radiotherapies, since in theory it allows tumor-cell-selective high-dose particle radiation while sparing normal cells. It remains uncertain whether BNCT can control GBM sufficiently and lead to distinct survival benefits because of the limitations of the present clinical trials, which include the lack of a contemporary control arm. Randomized trials of comparably selected patients are required to demonstrate the efficiency of this emerging modality more conclusively. To improve the effectiveness of BNCT, several critical issues must be addressed. Single or combined use of currently available boron agents must be optimized, and more tumor-selective boron agents and/or delivery or targeting systems are needed. The combined use of photon irradiation, chemotherapy, immunotherapy, and/or molecular targeting therapy with BNCT should also be evaluated. Finally, the development of an accelerator-based neutron source could lead to an in-hospital BNCT, which would greatly enhance both basic and clinical research on BNCT. Acknowledgments This study was supported in part by a Grant-in-Aid for Society Collaboration from the Ministry of Education, Science and Culture of Japan (20390379). We thank Professor Akira Matsumura, Dr. Tadashi Nariai and Dr. Alexander Zaboronok for their kind support in the manuscript preparation.

T. Yamamoto et al.

References Barth RF, Coderre JA, Vicente GH, Blue TE (2005) Boron neutron capture therapy of cancer: current status and future prospects. Clin Cancer Res 11:3987–4002 Buatti J, Ryken TC, Smith MC, Sneed P, Suh JH, Mehta M, Olson JJ (2008) Radiation therapy of pathologically confirmed newly diagnosed glioblastoma in adults. J Neurooncol 89:313–337 Burian J, Marek M, Rataj J, Flibor S, Rejchrt J, Viererbl L, Sus F, Honova H, Petruzelka L, Prokes K, Tovarys F, Dbaly V, Benes V, Kozler P, Honzatko J, Tomandl I, Mares V, Marek J, Syrucek M (2002) Report on the first patient group of the phase I BNCT trial at the LVR-15 reactor. In: Surewein W, Moss R, Wittig A (eds) Research and development in neutron capture therapy. Monduzzi Editore, Bologna, pp 1107–1112 Busse PM, Harling OK, Palmer MR, Kiger WS III, Kaplan J, Kaplan I, Chuang CF, Goorley JT, Riley KJ, Newton TH, Santa Cruz GA, Lu XQ, Zamenhof RG (2003) A critical examination of the results from the Harvard-MIT NCT program phase I clinical trial of neutron capture therapy for intracranial disease. J Neurooncol 62:111–121 Capuani S, Gili T, Bozzali M, Russo S, Porcari P, Cametti C, D’Amore E, Colasanti M, Venturini G, Maraviglia B, Lazzarino G, Pastore FS (2008) L-DOPA preloading increases the uptake of borophenylalanine in C6 glioma rat model: a new strategy to improve BNCT efficacy. Int J Radiat Oncol Biol Phys 72:562–567 Chanana AD, Capala J, Chadha M, Coderre JA, Diaz AZ, Elowitz EH, Iwai J, Joel DD, Liu HB, Ma R, Pendzick N, Peress NS, Shady MS, Slatkin DN, Tyson GW, Wielopolski L (1999) Boron neutron capture therapy for glioblastoma multiforme: interim results from the phase I/II doseescalation studies. Neurosurgery 44:1182–1193 Coderre JA, Hopewell JW, Turcottea JC, Rileyc KJ, Binnsc PJ, Kiger WS III, Harling OK (2004) Tolerance of normal human brain to boron neutron capture therapy. Appl Radiat Isot 61:1083–1087 Coderre JA, Turcotte JC, Riley KJ, Binns PJ, Harling OK, Kiger WS III (2003) Boron neutron capture therapy: Cellular targeting of high linear energy transfer radiation. Technol Cancer Res Treat 5:355–375 Diaz AZ (2003) Assessment of the results from the phase I/II boron neutron capture therapy trials at the Brookhaven National Laboratory from a clinician’s point of view. J Neurooncol 62:101–109 Fitzek MM, Thornton AF, Rabinov JD, Lev MH, Pardo FS, Munzenrider JE, Okunieff P, Bussiere M, Braun I, Hochberg FH, Hedley-Whyte ET, Liebsch NJ, Harsh GR 4th (1990) Accelerated fractionated proton/photon irradiation to 90 cobalt gray equivalent for glioblastoma multiforme: results of a phase II prospective trial. J Neurosurg 91:251–260 Haginomori S, Miyatake S, Inui T, Kawabata S, Takamai A, Lee K, Takenaka H, Kuroiwa T, Uesugi Y, Kumada H, Ono K (2008) Planned fractionated boron neutron capture therapy using epithermal neutrons for a patient with recurrent squamous cell carcinoma in the temporal bone: a case report. Head Neck 31:412–418 Henriksson R, Capala J, Michanek A, Lindahl SA, Salford LG, Franzén L, Blomquist E, Westlin JE, Bergenheim T

24 Glioblastoma: Boron Neutron Capture Therapy (2008) Boron neutron capture therapy (BNCT) for glioblastoma multiforme: A phase II study evaluating a prolonged high-dose of boronophenylalanine (BPA). Radiother Oncol 88:183–191 Joensuu H, Kankaanranta L, Seppälä T, Auterinen I, Kallio M, Kulvik M, Laakso J, Vähätalo J, Kortesniemi M, Kotiluoto P, Seren T, Karila J, Brander A, Järviluoma E, Ryynänen P, Paetau A, Ruokonen I, Minn H, Tenhunen M, Jääskeläinen J, Färkkilä M, Savolainen S (2003) Boron neutron capture therapy of brain tumors: clinical trials at the Finnish facility using boronophenylalanine. J Neurooncol 62:123–134 Kankaanranta L, Koivunoro H, Kortesniemi M, Välimäki P, Seppälä T, Kotiluoto P, Auterinen I, Kouri M, Savolainen S, Joensuu H (2008) BPA-based BNCT in the treatment of glioblastoma multiforme In: Zonta A, Altieri S, Roveda L, Barth R (eds) A dose escaration study. Proceedings of the 13th international congress of neutron capture therapy. ENEA, p 30 Kankaanranta L, Seppälä T, Koivunoro H, Saarilahti K, Atula T, Collan J, Salli E, Kortesniemi M, Uusi-Simola J, Mäkitie A, Seppänen M, Minn H, Kotiluoto P, Auterinen I, Savolainen S, Kouri M, Joensuu H (2007) Boron neutron capture therapy in the treatment of locally recurred head and neck cancer. Int J Radiat Oncol Biol Phys 69:475–482 Kawabata S, Miyatake S, Kuroiwa T, Yokoyama K, Doi A, Iida K, Miyata S, Nonoguchi N, Michiue H, Takahashi M, Inomata T, Imahori Y, Kirihata M, Sakurai Y, Maruhashi A, Kumada H, Ono K (2008) Boron neutron capture therapy for newly diagnosed glioblastoma. J Radiat Res (Tokyo) 50:51–60 Kumada H, Nakamura T, Komeda M, Matsumura A (2009) Development of a new multimodal Monte-Carlo radiotherapy planning system. Appl Radiat Isot 67:118–121 Kumada H, Yamamoto K, Matsumura A, Yamamoto T, Nakagawa Y (2007) Development of JCDS, a computational dosimetry system at JAEA for boron neutron capture therapy. J Phy Conf Ser 74:1–7 Nariai T, Ishiwata K, Kimura Y, Inaji M, Momose T, Yamamoto T, Matsumura A, Ishii K, Ohno K (2009) PET pharmacokinetic analysis to estimate boron concentration in tumor and brain as a guide to plan BNCT for malignant cerebral glioma. Appl Radiat Isot 67:S348–S350 Nielsen FH (2009) Micronutrients in parenteral nutrition: boron, silicon, and fluoride. Gastroenterology 137:S55–S60 Nigg DW (2003) Computational dosimetry and treatment planning considerations for neutron capture therapy. J Neurooncol 62:75–86

239 Sköld K-H, Stenstam B, Diaz AZ, Giusti V, Pellettieri L, Hopewell JW (2009) Boron neutron capture therapy for glioblastoma multiforme: advantage of prolonged infusion of BPA-f. Acta Neurol Scand. [Epub ahead of print] Stupp R, Mason WP, van den Bent MJ, Weller M, Fisher B, Taphoorn MJB, Belanger K, Brandes AA, Marosi C, Bogdahn U, Curschmann J, Janzer RC, Ludwin SK, Gorlia T, Allgeier A, Lacombe D, Cairncross JG, Eisenhauer E, Mirimanoff RO (2005) Radiotherapy plus concomitant and adjuvant temozolomide for glioblastoma. N Engl J Med 352:987–996 Tanaka M, Ino Y, Nakagawa K, Tago M, Todo T (2005) High-dose conformal radiotherapy for supratentorial malignant glioma: a historical comparison. Lancet Oncol 6: 953–960 Wittig A, Hideghety K, Paquis P, Heimans J, Vos M, Goetz C, Haselsberger K, Grochulla F, Moss R, Morrissey J, BourhisMartin E, Rassow J, Stecher-Rasmussen F, Turowski B, Westler M, deVris MJ, Fankhauser H, Gabel D, Sauerwein W (2002) Current clinical results of the EORTC-study 11961. In: Surewein W, Moss R, Wittig A (eds) Research and development in neutron capture therapy. Monduzzi Editore, Bologna, pp 1117–1122 Yamamoto T, Nakai K, Kageji T, Kumada H, Endo K, Matsuda M, Shibata Y, Matsumura A (2009) Boron neutron capture therapy for newly diagnosed glioblastoma. Radiother Oncol 91:80–84 Yamamoto T, Nakai K, Matsumura A (2008) Boron neutron capture therapy for glioblastoma. Cancer Lett 262: 143–152 Yang W, Barth RF, Wu G, Huo T, Tjarks W, Ciesielski M, Fenstermaker RA, Ross BD, Wikstrand CJ, Riley KJ, Binns PJ (2009) Convection enhanced delivery of boronated EGF as a molecular targeting agent for neutron capture therapy of brain tumors. J Neurooncol 95:355–365 Yoshida F, Matsumura A, Shibata Y, Yamamoto T, Nakauchi H, Okumura M, Nose T (2002) Cell cycle dependence of boron uptake from two boron compounds used for clinical neutron capture therapy. Cancer Lett 187:135–141 Yoshida F, Yamamoto T, Nakai K, Kumada H, Shibata Y, Tsuruta W, Endo K, Tsurubuchi T, Matsumura A (2008) Combined use of sodium borocaptate and buthionine sulfoximine in boron neutron capture therapy enhanced tissue boron uptake and delayed tumor growth in a rat subcutaneous tumor model. Cancer Lett 263: 253–258

Chapter 25

Glioblastoma: Anti-tumor Action of Cyclosporin A and Functionally Related Drugs Bozena Kaminska, Magdalena Tyburczy, Konrad Gabrusiewicz, and Malgorzata Sielska

Abstract Human malignant gliomas are highly resistant to current therapeutic approaches. Major signaling pathways that have been identified as playing important roles in glioblastomas are: the PTEN/PI3K/Akt/mTOR and the Ras/Raf/MEK/ERK signaling cascades, which support cell invasion, survival and prevent apoptosis. In the face of tumor resistance to apoptosis, novel agents which can overcome resistance or/and affect cell survival by nonapoptotic mechanisms such as necrosis, senescence, autophagy and mitotic catastrophe, are highly desirable. The present chapter focuses on anti-tumor action of cyclosporin A (CsA) and rapamycin that besides their well known immunosuppressive abilities appear to be multitarget kinase inhibitors and moderately effective anti-tumor agents in glioblastomas in vitro, in vivo and in clinical trials. A compelling evidence shows that cyclosporin A induces growth arrest and programmed cell death in cultured rat and human glioblastoma cells. The molecular mechanism involves accumulation of a cell cycle inhibitor – p21Cip1/Waf1, even in the absence of functional p53 tumor suppressor. In C6 glioma cells with functional TP53 and PTEN tumor suppressors CsA treatment up-regulates fasL expression, activates p53 and intrinsic mitochondrial death pathway, while in human glioblastoma cells with defects in either TP53 or PTEN, none

of those effects were observed. Molecular analysis revealed that CsA, trough yet unknown mechanisms, down-regulates PI3K/Akt and mTOR signaling pathways in glioblastoma cells, and interferes with proinvasive activity of tumor-infiltrating microglia. A systemically applied CsA significantly reduced growth of intracranial gliomas, tumor invasion and angiogenesis. Pharmacological inhibitors of the mTOR pathway: rapamycin, temsirolimus, everolimus and AP23573 were tested as potential targeted drugs in human glioblastoma cultures and in animal models. However, rapamycin and derivatives show moderate efficacy in patients with recurrent glioblastoma multiforme, they deserve further clinical studies, particularly in combination with PI-3K pathway inhibitors. Defects of innate and adaptive immunity are common in glioblastoma patients contributing to a lack of effective anti-tumor responses. Thus, immunosuppressants such as CsA, rapamycin and its derivatives may be an effective novel strategy to treat drug-resistant gliomas or complement apoptosis based-therapies. Keywords Glioblastomas · Anti-tumor drugs · Immunosuppressants · Apoptosis · Rapamycin · Inhibitors

Introduction B. Kaminska () Laboratory of Transcription Regulation, Department of Cell Biology, Nencki Institute of Experimental Biology, 02-093 Warsaw, Poland e-mail: [email protected]

Glioblastomas are the most frequent and devastating primary malignant brain tumor in adults. Glioblastomas are highly resistant to current therapeutic approaches, in which surgery is followed

M.A. Hayat (ed.), Tumors of the Central Nervous System, Volume 2, DOI 10.1007/978-94-007-0618-7_25, © Springer Science+Business Media B.V. 2011

241

242

by radiotherapy with concomitant and adjuvant chemotherapy. The prognosis remains poor with a median survival in the range of 12–15 months (Clarke et al., 2010). Common genetic abnormalities in glioblastoma are associated with multiple molecular mechanisms involved in drug resistance, including drug detoxification, aberrant activation or suppression of cellular signal transduction pathways, deficiencies in tumor suppressors, apoptosis mediators and death ligand/receptor signaling. Significantly high frequency of alterations in cell cycle regulators such as the TP53 tumor suppressor and the p16INK4A cyclin dependent kinase inhibitor results in reduced sensitivity to a majority of anti-cancer drugs (Ohgaki and Kleihues, 2009). Major signaling pathways that have been identified as playing important roles in glioblastomas are: the PTEN/PI3K/Akt/mTOR and the Ras/Raf/MEK/ERK signaling cascades, which support cell proliferation, survival, invasion, and prevent apoptosis (McCubrey et al., 2006). Components of these pathways are frequently mutated, aberrantly expressed or constitutively activated in glioblastomas. Monoclonal antibodies or small molecular-weight kinase inhibitors targeting specific pathways are the most common classes of agents in cancer treatment. However, highly selective or specific blocking of only one of the kinases has been associated with limited or sporadic responses. Therefore, multitargeted kinase inhibitors and combinations of single-target kinase inhibitors should be more effective to overcome therapeutic resistance. Some agents could be used together with radiation, chemotherapy, or immunotherapy to enhance treatment efficacy. Furthermore, in the face of tumor resistance to apoptosis, novel agents which can overcome resistance or/and induce cell death by non-apoptotic mechanisms such as necrosis, senescence, autophagy (type II programmed cell death) and mitotic catastrophe, are highly desirable. The present chapter focuses on anti-tumor action of drugs such as cyclosporin A and rapamycin that besides their well known immunosuppressive abilities appear to be kinase inhibitors and potential anti-tumor agents in glioblastomas. The present chapter summarizes a rationale and results of preclinical/clinical studies of these agents in therapy of glioblastomas.

B. Kaminska et al.

Mechanisms of Cyclosporin A Induced Cell Death in Rat C6 Glioma Cells Mode of Immunosuppressant Action Cyclosporin A (CsA), FK506 (tacrolimus, Prograf) and rapamycin (sirolimus) are short polypeptides which have revolutionized transplantology due to ability to block the activation of lymphocytes and other immune system cells. CsA, FK506 and rapamycin bind to specific intracellular proteins called immunophilins: CsA binds to cyclophilin, FK506 and rapamycin bind to FKBP (FK506-binding protein). Drugimmunophilin complexes bind to a regulatory subunit of calcineurin and inhibit its activity. Calcineurin is a calcium- and calmodulin-dependent threonine/serine phosphatase. CsA and FK506 exert immunosuppressive effects by inhibiting of calcineurin-mediated dephosphorylation of NFAT (nuclear factor of activated T cells), thus preventing transcriptional induction of several cytokines and their receptors genes (Fig. 25.1). NFAT family proteins are transcription factors that regulate the expression of a variety of target genes and are implicated in many functions, including cell growth, survival, invasion and angiogenesis (Mancini and Toker, 2009). Several members of NFAT family were detected in C6 glioma cells pointing to a new mechanism of transcription regulation in glioma cells. A transient receptor potential 6 (TRPC6) which is required for the development of the aggressive glioblastoma phenotype and causes a sustained elevation of intracellular calcium, is coupled to activation of the calcineurin-NFAT pathway. Pharmacologic inhibition of this pathway reduced the development of the aggressive glioblastoma phenotype under hypoxia.

Molecular Mechanisms of Pro-apoptotic Action of Cyclosporin A in C6 Glioma Cells It was reported by Mosieniak et al. (1997) that cyclosporin A at concentrations at the range of 30– 60 μM inhibits proliferation of rat C6 glioma cells and induces cell death. CsA-induced cell death was an

25 Glioblastoma: Anti-tumor Action of Cyclosporin A and Functionally Related Drugs

243

Fig. 25.1 Mechanism of immunosuppressant action (a) Immunosuppressants bind to immunophilines: CsA to cyclophilin (cyc) and FK506 to FKBP (FK506-binding protein); subsequently complexes bind to the calcineurin and inhibit its activity. Calcineurin is a calcium- and calmodulin-dependent phosphatase which dephosphorylates NFAT transcription factors allowing them to translocate to the nucleus, where NFAT proteins in cooperation with other transcription factors (f.e. AP-1) regulate the expression of genes coding for chemokines, cytokines and their receptors. The immunosuppressantimmunophilin complex may interfere with MAPK signaling

pathways. (b) Summary of pro-apoptotic mechanisms induced by CsA in C6 glioma cells. CsA induces activation of JNK (c-Jun amino-terminal kinase) and MKK3 (MAP kinaseactivated protein kinase)-p38 MAPK signaling pathways. It leads to activation of AP-1 and transcriptional up-regulation of FasL (ligand Fas) which binds receptor Fas expressed on glioma cells and induces cell death. Activation of p38 MAPK signaling leads to accumulation of p53 and induction of p53-dependent expression of pro-apoptotic genes involved in mitochondrial death pathway

active process requiring expression of new genes and proteins, with typical features of apoptosis: oligonucleosomal DNA fragmentation and caspase 3 activation (Mosieniak et al., 1997; Pyrzynska et al., 2000). Two major mechanisms responsible for induction of

apoptotic death by CsA were identified. Apoptotic cell death induced by CsA was associated with a persistent activation of mitogen activated protein kinases (MAPK), in particular c–Jun N-terminal kinase (JNK) and p38 MAPK. Prolonged activation of JNK led

244

to accumulation of phosphorylated c-Jun and ATF2 (main substrates of JNK) and formation of the AP-1 transcription factor followed by transcriptional activation of the Fas Ligand expression (Pyrzynska et al., 2000). Further studies with promoter constructs depleted of DNA binding sites for particular transcription factors revealed that activation of the FasL gene promoter was only partially AP-1-dependent and collaborative action of other transcription factors was required for promoter activation. It was demonstrated that CsA down-regulates Akt signaling to facilitate activation of Forkhead family members resulting in transcriptional activation of the FasL expression (Ciechomska et al., 2003). Down-regulation of Akt signaling was necessary to permit Forkhead transcription factor translocation to the nucleus and pre-requisite to transcriptional activation of the FasL expression. Treatment of glioma cells with lower doses of CsA (1–10 μM) was sufficient to reduce Akt phosphorylation and sensitized cells to doxorubicin, and UVC treatments (unpublished). Another event resulting from prolonged activation of MAP kinases, in particular p38 MAPK, was accumulation of the tumor suppressor p53 in glioma cells. The p53-family of transcription factors consists of three genes – p53, p63, and p73 sharing significant structural and functional similarities. p53 is a potent inducer of apoptosis and tumor suppression. Many anti-cancer agents, from traditional chemo- and radiation therapies to more recently developed small molecules, exert their effects by enhancing the antiproliferative effects of p53 and transactivating p63/p73 proteins. In normal cells the p53 is expressed at low, constitutive level and localized predominantly in cytoplasm. The latent form of p53 is stabilized and activated by posttranslational modifications. The activation of p53 occurs in response to DNA damage or stress such as hypoxia, nutrients or nucleotide deprivation. p53-mediated cell cycle arrest is largely brought about by induction of p21Waf1, an inhibitor of cyclindependent kinases. Activation of p53 may also result in apoptosis via transcriptional activation of a number of pro-apoptotic proteins including Bax, Fas, p85, IGF-BP3, and PIG3 and apoptotic protease activating factor-1 (apaf-1). It was reported by Pyrzynska et al. (2002) that CsA treatment results in up-regulation of p53 protein level and its accumulation in cell nuclei. Concomitantly, the levels of p21Waf1 and Bax proteins increased,

B. Kaminska et al.

and Bcl-xL decreased in CsA-treated glioma cells. Bax protein translocated to mitochondria, as revealed by immunofluorescence and double staining with a mitochondrial marker, and likely induced mitochondrial apoptotic pathway. Contribution of p53 to CsAinduced cell death was further confirmed in experiments, in which glioma cells stably transfected with a mutant p53 (p53Val135) failed to increase p21 and Bax protein levels and were less sensitive to CsA-induced apoptosis. Also primary fibroblasts from p53-/knockout mice were significantly more resistant to CsA-induced apoptosis compared to their corresponding counterparts containing functional p53 (Pyrzynska et al., 2002). Accumulation and activity of p53 can be regulated by phosphorylation that occurs at several Ser and Thr residues, and a number of cellular kinases have been proposed to directly phosphorylate p53, including casein kinase I, casein kinase II, double-strandedRNA-dependent protein kinase, ATM, CDK7, DNAactivated protein kinase, Jun-NH2 kinase and p38 MAP kinase. The induction of cell death by CsA was associated with a persistent activation of MKK3-p38 MAP kinase signaling pathway. Overexpression of a dominant negative form of MKK3, an upstream activator of p38, abrogated phosphorylation of p38 MAPK and p53 accumulation (Pyrzynska et al., unpublished). Together, the compelling evidence demonstrate that the apoptotic program activated by CsA can be mediated by multiple pathways: via activation of p53 transcription factor and p53-mediated apoptosis, and through up-regulation of FasL expression by JNK-AP-1 and Forkhead (Fig. 25.1).

Cyclosporin A Induces Cell Death or Growth Arrest and Senescence of Human Glioblastoma Cells In rat C6 glioma cells apoptotic cell death induced by CsA was associated with induction of intrinsic death pathway: p53- and mitochondria-dependent, as well as extrinsic, death ligand-dependent pathway. However, the experiments performed on glioma cells lacking a functional p53, namely the cell line stably expressing the temperature-sensitive p53 mutant, indicated that up to 20% of cells died even in the absence of functional

25 Glioblastoma: Anti-tumor Action of Cyclosporin A and Functionally Related Drugs

p53, suggesting also a p53-independent mode of CsA action. Therefore, we tested efficacy of CsA towards three human glioblastoma cell lines that are radioand chemotherapy resistant, in part due to mutations in the tumor suppressor PTEN or/and TP53 genes: T98G (mutated PTEN and TP53), U373-MG (mutated TP53 and PTEN) and U87-MG cells (wild type TP53/mutated PTEN). Cultured cells were exposed to increasing concentrations of CsA (Fig. 25.2) and 30 μg/ml CsA affected growth of all studied glioma cells, produced dramatic changes in cell number and morphology in 24 h. The inhibitory effect of CsA on cell growth and survival was dose-dependent, and progressed with the time of exposure. CsA at the concentration range of 5–20 μg/ml had no visible cytotoxic effects (Zupanska et al., 2005). Morphological alterations induced by 30 μg/ml CsA were characterized by shrinkage of the cells, rounding up of the cell body and detachment from the bottom of the plate at 24– 48 h after the treatment (Fig. 25.2). Cell death induced by CsA was blocked by cycloheximide (a protein synthesis inhibitor) and showed no signs of necrosis, such as swelling or disruption of cells. The appearance of numerous large vacuoles was observed in CsAtreated T98G cells and to less extent in U373-MG cells (Zupanska et al., 2005). Features of CsA-triggered cell death in human glioblastoma cells do not fulfill all criteria of apoptosis. For example, biochemical hallmarks of apoptosis such as phosphatidylserine exposure in the external leaflet of the plasma membrane bilayer, the “ladder-like” oligonucleosomal DNA fragmentation or appearance of the subG1 population were not detected, although condensation of chromatin, deformation of nuclei and sparse TUNEL labeling were observed. Cells in CsA-treated cultures showed “bean” shaped nuclei with condensed chromatin or nuclei with irregular clumps of dense chromatin. Such nuclear alterations as “bean” shaped nuclei with condensed chromatin usually preceded oligonucleosomal DNA fragmentation in CsA-treated rat glioma cells (Mosieniak et al., 1997). In T98G glioblastoma cells nuclear alterations were completely abolished by cycloheximide (CHX) treatment, indicating dependency of cell death on de novo protein synthesis. Measurement of changes in the mitochondrial membrane potential in response to CsA treatment with the fluorescent probe JC-1 did not show alterations in mitochondrial potential, which excludes

245

a possibility of initiation of mitochondrial pathway along with apoptosome formation (Zupanska et al., 2005). Western blot analysis using specific antibodies recognizing intact and cleaved caspases 7, 3 and PARP (poly (ADP-ribose) polymerase-1) was employed to further characterize cell death induced by CsA. The treatment of human glioblastoma cells with 30 μM CsA resulted in activation of the caspase cascade (as evidenced by the appearance of cleaved caspase 7, 3 and cleaved PARP) in T98G cells, and to lesser extent in U373-MG cells. No caspase activation was detected in U87 glioma cells (Fig. 25.2). Growth arrest may lead to replicative senescence or differentiation. Senescent cells in contrast to presenescent, proliferating or quiescent cells express a beta-galactosidase activity detectable at pH 6.0. Since flatten morphology of CsA-treated U87 cells and reduction of the number of proliferating cells suggested the induction of differentiation process, a senescence-associated β-galactosidase (SA-beta-Gal) staining was performed (Fig. 25.2). Untreated U87MG and T98G cells demonstrated very rare cells positive for the SA-beta-Gal staining. After 24 h of CsA treatment the percentage of cells positive for SA-beta-Gal was doubled in U87-MG cells compared to control, while only few T98G cells were positively stained. SA-beta-Gal positive U87-MG cells increased in size and flattened out, thereby attaining morphology of senescent-like cells (Zupanska et al., 2005). Cell proliferation is controlled by cell cycle regulatory factors which include cyclins and cyclindependent kinases. p21WAF1/Cip1 protein is an universal inhibitor of cyclin kinases and plays an important role in inhibiting cell proliferation. The levels of p21WAF1/Cip1 protein were increased after exposure to 30 μM CsA in all examined glioblastoma cell lines. The highest up-regulation of p21WAF1/Cip1 protein was observed in T98G cells and the moderate accumulation was detected in U373-MG and U87MG cells. Further studies revealed transcriptional activation of p21WAF1/Cip1 expression preceded in time by a long-term activation of ERK1/2 signaling, subsequent c-Jun phosphorylation and accumulation of AP-1 complex in CsA-treated glioblastoma cells. Pre-treatment with ERK pathway inhibitors or overexpression of dominant negative mutants MKK1, ERK2 and c-Jun reduced the p21WAF1/Cip1

246

Fig. 25.2 CsA induces programmed cell death or growth arrest/senescence of malignant glioblastoma cells. (a) Morphological alterations induced by CsA treatment in human glioma cells. T98G cells cultured in DMEM with 10% fetal bovine serum, were exposed to 30 μM CsA alone or with 1 μg/ml cycloheximide (CHX). Upper panel shows that a majority of cells lost processes and became round in CsA-treated cultures. Lower panel shows Hoechst 33258 staining revealing deformations of cell nuclei and condensation of chromatin. CHX inhibits morphological alterations in CsA-treated T98G cells. (b) Caspase cascade activation in CsA-triggered cell death.

B. Kaminska et al.

A representative immunoblot shows caspase activation in total protein extracts of human glioma cells treated with 30 μM CsA. Specific antibodies recognizing intact and cleaved caspases 7, 3 and cleaved PARP (Cell Signaling, USA) were employed. (c) Detection of the Senescence-Associated beta-galactosidase (SA-beta-Gal) staining in CsA-treated U87-MG glioma cells. The substantial increase in cellular volume and the blue staining reflecting SA-beta-Gal activity was observed in U87-MG cells treated with CsA. Original magnification for larger photos is ×10; for insets ×20

25 Glioblastoma: Anti-tumor Action of Cyclosporin A and Functionally Related Drugs

expression, confirming involvement of this pathway. Transcriptional activation of p21WAF1/Cip1 expression by CsA was independent of p53 and preceded CsA-induced growth arrest in glioblastoma cells (Zupanska et al., 2007). In conclusion, the findings above presented demonstrate an ability of CsA to induce growth arrest or programmed, but non-apoptotic cell death in human glioblastoma cells at the concentration of 30 μM or higher. Many forms of programmed cell death different from “classical” apoptosis have been described by the criteria of morphology, biochemistry, and response to apoptosis inhibitors, particularly in transformed cells which contain endogenous inhibitors preventing a particular pathway.

Anti-tumor Effects of CsA in Glioma Models Good cytostatic and cytotoxic efficacy of CsA in rat and human glioblastoma cultures encouraged studies in more complex models: in organotypic brain slice cultures and in murine glioma model. Organotypic brain slice cultures injected with glioblastoma cells recapitulate many features of glioblastoma and are very useful for investigating the cellular and molecular mechanisms of glioma invasion under conditions most analogous to those of normal brains in vivo. It was reported by Markovic et al. (2005) that microglial cells contribute significantly to invasion of glioma cells in cultured brain slices. Microglia are the intrinsic immune cells of the brain, serving principally to control the innate and the adaptive immune responses in the central nervous system, to initiate host-defense and tissue repair mechanisms. Microglial cells are attracted towards glioma (glioma tissue consists of up to 30% of microglial cells) and microglia density in gliomas positively correlates with malignancy, invasiveness and grading of the tumors (Watters et al., 2005). When glioma cells were injected into brain slices depleted of endogenous microglia (by incubation with clodronate-filled liposomes for 96 h), the invasiveness of the tumors was significantly decreased. Inoculation of exogenous microglia together with glioma cells into cultured brain slices increased the infiltrative behavior of glioma cells. Experiments with

247

co-culture of microglia with glioma cells revealed that soluble factors released from glioma cells strongly stimulate metalloprotease-2 (MMP-2) activity increasing breakdown of extracellular matrix and thereby promoting tumor invasiveness (Markovic et al., 2005). Further studies showed that membrane type 1 metalloproteinase (MT1-MMP) is up-regulated in gliomaassociated microglia. Microglial MT1-MMP in turn activates glioma-derived pro-MMP-2 and promotes glioma expansion. Tumor growth and invasion in ex vivo model using MT1-MMP deficient brain tissue and in microglia-depleted animals were strongly reduced (Markovic et al., 2009). It was reported by Sliwa et al. (2007) that migration/invasion of fluorescently labeled GL261 glioma cells in murine brain slices significantly decreased upon treatment with CsA, even at low concentrations: 1, 10 and 30 μM. Anti-invasive effects of lower, non-cytotoxic doses of CsA suggested an existence of the additional mechanism of CsA action on tumor invasion. When glioma cells were injected into slices devoid of endogenous microglia, the inhibitory effect of 1 μM CsA on glioma invasiveness mostly vanished. It indicated that CsA abolishes microglia-promoting effects on glioma malignancy. The inhibitory action of CsA on microglia function has been directly demonstrated in microglia-glioma co-cultures. Glioma-derived factors in co-cultures or glioma conditioned medium induce morphological transformation of microglial cells into amoeboid phagocytes, activate MAPK signaling pathways and production of some cytokines. CsA at low doses of 0.1. and 1 μM blocked transformation of microglial cells stimulated by glioma-conditioned medium and abolished pro-invasive effects of microglia (Sliwa et al., 2007). A recent study points to a crucial role of microglia-derived transforming growth factor beta 1 (TGF-β1) in regulation of glioma invasion (Wesolowska et al., 2008). Blockade of TGF-β1 signaling by silencing of TGF-β1 type II receptor with shRNA or neutralizing anti-TGF-β1 antibody abolished the promoting effect of microglia on glioma invasion. In vivo tumor models developed by intracranial or subcutaneous implantation of glioma cell lines in rodents are used to test novel therapies. The advantages of these glioma models are their highly efficient gliomagenesis, reproducible growth rates, and knowledge of tumor location. However, these

248

models have been criticized for not recapitulating all pathological features of human glioblastoma multiforme (GBM), they allow to target different features of GBM, including location in the brain, invasion of brain parenchyma, angiogenesis, and secretion of immune suppressive molecules. In studies by Sliwa et al. (2007) fluorescently labeled GL261 glioma cells were implanted into the striatum of DBA/2 J mice and developed gliomas of considerable size in 14 days. CsA (Sandimmun, Novartis) was administrated intraperitoneally (i.p.) every second days at doses of 2 or 10 mg/kg of body weight starting from the second day after cell inoculation. Systemically applied 2 or 10 mg/kg CsA significantly decreased tumor volumes (by 70%) with similar efficacy. Further studies in syngenic murine models, such as GL261 mouse glioma cells implanted to C57BL6 mice, have confirmed anti-tumor action of CsA administrated intraperitoneally every second day at doses of 2 and 10 mg/kg. Immunofluorescence studies with Iba1 antibody demonstrated that CsA blocks accumulation and activation of glioma-associated microglia confirming the in vitro data (unpublished). It was reported that macrophage–colony stimulating factor (M-CSF) and other cytokines secreted by high-grade glioma cells stimulate differentiation of tumor-infiltrating microglia/macrophages into cells acquiring the anti-inflammatory (M2) phenotype. M-CSF was significantly correlated with histological malignancy and with the proportion of M2 microglia/macrophages in vivo. Such M2 cells instead of initiating immune responses support tumor growth and invasion. CsA with its strong effect on pro-tumorigenic activity of microglia could be effective drug in modulation of anti-inflammatory M2 phenotype of microglia/macrophages in glioma therapy. CsA has been used to block the immune reaction towards human glioblastoma cells transplanted to rodent brains. It was reported by Mathiesen et al. (1989) that human glioblastoma cells implanted to cerebrum of rat hosts survive longer in immunosuppressed animals but aggressive growth of glioblastoma cells and tumor proliferation was not observed. Human glioblastoma cells implanted into the brain of Wistar rats immunosuppressed with CsA applied at 12 mg/kg/daily survived under such conditions but formed dense and poorly diffused tumors (Strojnik

B. Kaminska et al.

et al., 2006). Up to now CsA has not been tested as anti-invasive/cytotoxic agent in clinical trials.

Anti-tumor Action of Rapamycin and Derivatives Mechanism of Action of Rapamycin Rapamycin (Sirolimus) is a macrolide discovered ∼1970 as a product of the bacterium Streptomyces hygroscopicus in a soil sample from the Easter Island (Rapa Nui). Rapamycin was originally developed as an antifungal agent, but turned out to be a potent immunosuppressive drug and since 1997 has been used to prevent host-rejection in kidney transplantation. An immunosuppressive effect of rapamycin is due to the inhibition of interleukin 2 (IL-2)-mediated T-cell proliferation and activation. The mode of action of rapamycin is to bind the cytosolic immunophilin FK-binding protein 12 (FKBP12). This complex inhibits the mammalian target of rapamycin (mTOR) pathway by directly binding the mTOR Complex 1 (mTORC1). mTOR is a kinase playing a key role in the regulation of cell growth and proliferation by regulating ribosomal biogenesis and protein translation. Among others, mTOR can be activated by growth factors and hormones, which is mediated by the induction of PI-3-kinase (PI-3K). Active PI-3K generates phosphatidylinositol (3,4,5)trisphosphate [PtdIns(3,4,5)P3 ] which activates kinase Akt phosphorylating and inactivating the tuberous sclerosis complex 2 (TSC2). TSC2 acts as a GTPaseactivating protein for the small GTPase RHEB (Ras homolog enriched in brain). Inhibition of TSC2 activity leads to elevated RHEB-GTP levels and activation of mTORC1 which by its downstream effectors, such as ribosomal protein S6 kinases (S6K) and 4E-binding protein (4E-BP1), further up-regulates ribosome biogenesis and protein translation (Fig. 25.3). This results in an increase in cell size and mass, and enhanced proliferation. On the other hand, inhibition of mTOR promotes autophagy in some cell types. Autophagy is a process of degradation of organelles and macromolecules to supply cells with nutrients and can function as a cytoprotective mechanism or alternative cell death process.

25 Glioblastoma: Anti-tumor Action of Cyclosporin A and Functionally Related Drugs

249

Fig. 25.3 Rapamycin inhibits proliferation or induces cell death of malignant glioblastoma cells. (a) mTOR signaling network. mTORC1 stimulates cell growth upon activation by growth factors or insulin (for details see the text). (b) Rapamycin inhibits the kinase activity of mTOR measured by reduction of S6K phosphorylation in T98G cells. Western blot analysis of the phospho-S6K (T389) level in T98G glioma cells after treatment with 10 nM rapamycin for 24 h. S6K served as a protein loading control. (c) Rapamycin affects proliferation of T98G cells. Cell proliferation was determined 24 h after exposure to 1, 10, 30

or 60 μM rapamycin using BrdU test. Bars represent the ratio of proliferating cells treated with the inhibitor related to control cells (means ± SEM of three independent experiments, each in triplicate). Statistical analysis was done by one-way ANOVA followed by Newman-Keuls test, ∗∗∗ p < 0.001. (d) Rapamycin induces morphological alterations in T98G cells. Phase-contrast images of T98G glioma cells performed 24 h after treatment with 10, 30 or 60 μM rapamycin. Cell death was observed after using the highest drug concentration. Original magnification is ×10

Rapamycin and Its Analogs in Pre-clinical and Clinical Trials in Glioblastomas

mTOR is a crucial downstream component of the PTEN/Akt signaling, pre-clinical studies with pharmacological inhibitors of the mTOR pathway were initiated in glioblastomas. Apart from rapamycin, its analogs with improved pharmacokinetic properties were synthesized. CCI779 (temsirolimus; Wyeth), RAD001 (everolimus; Novartis) and AP23573 (Ariad) are currently being tested as potential targeted drugs in glioblastomas. Pharmacological inactivation of mTOR decreased tumor cells proliferation and tumor size in PTEN-deficient mice (Rajasekhar et al., 2003). Moreover, xenografts generated from PTEN null U87 glioma cells in mice had inhibited growth and reduced proliferation rate after rapamycin treatment.

Enhanced or constitutive activation of the PI3K/Akt/mTOR pathway is an important factor in gliomagenesis (Guertin and Sabatini, 2007). In human glioblastomas Akt is activated in approximately 70% of tumors due to loss of PTEN (phosphatase and tensin homolog deleted on Chromosome 10), the tumor suppressor and negative regulator of PI-3 K/Akt signaling, and/or up-regulation of epidermal growth factor receptor (EGFR) and platelet-derived growth factor receptor (PDGFR) tyrosine kinases. PTEN mutations are present in 20–40% of GBM. Since

250

Conversely, rapamycin did not decrease size of PTEN wild-type LN229 tumors (Wei et al., 2008). These findings provided grounds to initiate clinical trials of mTOR inhibitors in glioblastomas. The results of six studies using sirolimus or CCI-779 (a dihydroxymethyl propionic acid ester of sirolimus) as a single agent, or sirolimus and RAD001 in combination with inhibitors of EGFR in patients with recurrent glioblastoma multiforme (GBM) are shown in Table 25.1. Treatment with CCI-779 or sirolimus alone had only a limited activity in recurrent GBM, particularly in unselected patients. In the phase II study by Chang et al. (2005) CCI-779 was administered weekly to forty-three patients with recurrent GBM (14 patients were on enzyme-inducing antiepileptic drug). Initially CCI-779 was administered at a dose of 250 mg intravenously but the dose was reduced to 170 mg because of side effects. CCI-779 was well tolerated; however, failed to demonstrate any efficacy as a single agent in patients with recurrent GBM. Despite initial disease stabilization in approximately 50% of patients, the durability of response was short. In another phase II study CCI-779 was administered in a 250-mg intravenous dose weekly to sixty-five recurrent GBM patients with ≥1 chemotherapy regimen (Galanis et al., 2005). The study reported that 20 of 65 patients with recurrent GBM (36%) had radiographic improvement. Progression-free survival at 6 months was 7.8% and median overall survival was 4.4 months. Median time to progression for all patients was 2.3 months and was significantly longer for responders

B. Kaminska et al.

(5.4 months) versus non-responders (1.9 months). The authors assessed activation of the PI3K pathway by examining tumor specimens for total/phosphorylated Akt and p70s6 kinase, and found p70s6 kinase staining indices being significantly more frequent in responders versus non-responders (Galanis et al., 2005). It was reported by Cloughesy et al. (2008) that rapamycin treatment leads to substantial inhibition of tumor cell proliferation in seven of 14 patients, as demonstrated by reduction of Ki-67 staining in a subset of patients with PTEN loss. Inhibition of mTOR signaling was observed in tumor tissue. However, rapamycin led to the activation of Akt in some patients, which correlated with faster tumor progression the authors concluded that rapamycin has anti-cancer activity in PTEN-deficient glioblastoma and deserves further clinical study alone or in combination with PI3K pathway inhibitors. More recently, based on evidence of the synergism between inhibitors of mTOR and EGFR (Rao et al., 2005), several studies investigated such a combination. In two trials a partial radiological response (≥50% decrease in the product of perpendicular diameters of contrast enhancing mass without new lesions) was seen. Doherty et al. (2006) reported a trial on 22 GBMs and six anaplastic gliomas patients treated with either gefitinib 500 mg or erlotinib 150 mg orally/daily in combination with sirolimus administered at a dose of 6 mg orally the first day followed by 4 mg orally once daily thereafter. Both medications were given daily in 28-day cycles. Out of this cohort, 19% of patients

Table 25.1 Results of targeted therapy trials of mTOR inhibitors used alone or in combination with inhibitors of EGFR in recurrent glioblastomas Radiological Median PFS 6-month Median OS Study Treatment Patients (n) response (%) (months) PFS (%) (months) Chang et al. (2005)

CCI-779, 170–250 mg 43 5 weekly Galanis et al. (2005) CCI-779, 250 mg weekly 65 0 Cloughesy et al. Sirolimus, 2, 5, or 10 mg 15 14 (2008) daily Doherty et al. Sirolimus, 4 mg daily; 22 18 (2006) gefitinib, 500 mg, or erlotinib, 150 mg daily Kreisl et al. (2009) RAD001, 70 mg weekly; 22 14 gefitinib, 250 mg daily Reardon et al. Sirolimus, 5–10 mg daily; 32 0 (2010) erlotinib, 150–450 mg daily PFS, progression-free survival; OS, overall survival; NA, not available

2.3

3

NA

2.3 NA

8 NA

4.4 NA

3

25

NA

2.6

5

5.8

1.8

3

8.5

25 Glioblastoma: Anti-tumor Action of Cyclosporin A and Functionally Related Drugs

experienced a partial response and 50% had stable disease; 6-month progression-free survival (6 M-PFS) was 25%. In a second study on 22 patients (all had received prior radiation and chemotherapy) receiving gefitinib (250 mg daily) and everolimus (70 mg weekly), 14% of patients had a partial response and 36% of patients stable disease; median overall survival was 5.8 months (Kreisl et al., 2009). Overall response rate was 19% and a 6 M-PFS of 25% in GBM patients. No molecular markers of response were described. A retrospective review of eight consecutive negative phase II trials in recurrent malignant gliomas from the M.D. Anderson Cancer Center found a 6 MPFS for GBM of 15%. For comparison, among GBM patients treated with temozolomide at 200 mg/m2 /day orally for the first 5 days of a 28-day cycle, PFS at 6 months was 18%; median progression-free survival and median overall survival were 2.1 and 5.4 months, respectively. The 6-month survival rate was 46% (Brada et al., 2001). Although the trials demonstrate moderate clinical benefits from the combination of two drugs, ongoing phase I/II studies are investigating the efficacy of using mTOR inhibitors with other targeted agents and chemoradiation, namely: sirolimus plus vandetanib, CCI-779 or RAD001 plus temozolomide, CCI-779 plus radiation or sorafenib or bevacizumab or perifosine, RAD001 plus AEE788, and RAD001 plus Gleevec and hydroxyurea (http://clinicaltrials.gov).

General Considerations A use of CsA may raise reservations due to employment of drug blocking the immune system in tumor patients, poor drug accessibility due to blood-brain barrier or general toxicity. However, a growing evidence shows both innate and adaptive immunity impaired in glioblastomas (Yang et al., 2010). Glioblastomas are poorly immunogenic and do not express specific tumor antigens (Watters et al., 2005). Microglia infiltrating glioblastomas are converted into tumor supportive cells and contribute to tumor growth (Markovic et al., 2005, 2009; Sliwa et al., 2007). Microglial cells operate as the first line innate and adaptive immunity of the central nervous system (CNS). In the normal CNS, microglia express low levels of major histocompatibility complex (MHC)

251

class I and class II molecules and co-stimulatory molecules such as CD86 and CD40. Upon activation, microglia convert to an active phenotype, upregulate MHC class I and class II and co-stimulatory molecules and take part in CD4-and CD8-specific T cell responses. Furthermore, malignant glioblastoma cells secrete chemoattractants and growth factors, including monocyte chemoattractant protein 1 (MCP-1), colony stimulating factor 1 (CSF-1), granulocyte-macrophage colony stimulating factor (GM-CSF), hepatocyte growth factor (HGF) and recruit microglia. Even though glioblastoma accumulate many microglia, macrophages and a small population of lymphocytes, the defense mechanisms are down-regulated. It was reported by Hussain et al. (2006) that glioblastoma-infiltrating microglia isolated from human tumor tissue samples do not produce inflammatory cytokines: interleukin 1β (IL-1β), interleukin 6 and tumor necrosis factor α (TNF-α), cytokines critical for developing effective innate immune responses. On the contrary, glioma-associated microglia/ macrophages might promote tumor growth by inducing immunosuppression in the tumor microenvironment. Glioblastoma cells and accumulating microglia release several cytokines, such as interleukin 10 (IL10) and TGF-β which may inhibit T-cell activation and contribute to the local glioma immunosuppressive milieu (Watters et al., 2005). An effective antitumor T-cell response is decreased in glioblastomas, because expression of MHC class II and co-stimulatory B7 molecules is reduced and microglia appear deficient in proper antigen presentation. Thus, potential immunossuppressive effects of CsA can be neglected because the immune system in glioblastoma patients is already paralyzed and anti-tumor responses are nonfunctioning. Although, tumors incidence in transplant recipients is a recognized consequence, a recent examination of reported cases of tumor in the transplant population of the Israel Penn International Transplant Tumor Registry shows that primary brain tumors, including gliomas, do not appear to be overrepresented in the Registry, indicating that they may not arise with increased frequency in transplant recipients (Schiff, 2004). As it has been shown in animal glioma model by Sliwa et al. (2007), CsA has a systemic effect, can affect cell interactions in tumor microenvironment and block pro-tumorigenic activity of tumor infiltrating

252

microglia/macrophages. A recent study demonstrates that mTOR signaling controls microglial activation in response to cytokines and appears to play a crucial role in regulating of microglial viability (Dello Russo et al., 2009). Thus anti-tumor action of rapamycin in vivo may be partly due to interference with cell interactions in tumor microenvironment. The blood-brain barrier (BBB) possesses a significant impediment for the delivery of therapeutic drugs into the brain. In the normal adult brain, the BBB mainly consists of vascular endothelial cells, astrocytes and pericytes. In malignant glioblastomas, the blood vessel network is broken down. Increased permeability of tumor blood vessels is induced by angiogenic factors released by malignant glioma cells, such as vascular endothelial growth factor (VEGF) that is associated with intravascular migration of glioma cells into blood vessels and their transfer to distant areas in the brain via blood flow. The mechanism of vascular permeability is used by glioma cells to facilitate migration and invasiveness, but can be also employed to facilitate delivery of anti-tumor medicines into the brain (Tate and Aghi, 2009). Previous studies demonstrated that a systemic injection of 100 mg/kg/i.p. CsA results in a blood concentration of 10 μM maintained for at least 18 h (Ciechomska et al., 2005). Thus, concentrations of CsA (2 and 10 mg/kg) tested in animal glioma model correspond to blood concentrations of 0.2 or 1 μM, respectively. A permeable blood-brain barrier in brain tumors may result in similar CsA concentrations in the brain as detected in the blood. In contrast to widely accepted notion that CsA does not cross or poorly cross the blood-brain barrier, studies on immunosuppressed patients demonstrated the presence of CsA in cerebrospinal fluids suggesting that the drug is able to cross the blood-brain barrier. Furthermore, CsA may impair the brain endothelial barrier function by accelerating NO production in the brain endothelial and astroglial cells. Notably, CsA at lower micromolar concentrations does not affect survival of non-transformed cells. Neurons from mixed neuronal-glial cultures developed from hippocampal dentate gyrus are affected by CsA at the concentrations higher than 8–10 μM (Kaminska et al., 2001). Astrocytes are more resistant and CsA at concentration of 40 μM or higher affects the survival of astrocytes from neonatal brain cultures. Altogether, these results suggest that CsA, particularly

B. Kaminska et al.

at lower concentrations affecting pro-invasive activity of microglia, could be effective anti-tumor drug without inducing neurotoxicity. CsA induces cell death via multiple mechanisms and some of them are able to prevail alterations of growth regulatory and apoptotic pathways, caused by common mutations in human glioblastomas. Rapamycin and its derivatives, besides the well described inhibition of mTOR pathway in glioblastoma cells, may additionally affect proinvasive activity of glioblastoma infiltrating microglia. The unique mechanism of action of CsA and pharmacological inhibitors of the mTOR pathway justifies further research on their anti-tumoral properties either alone or in combined therapies. Acknowledgements This work was supported by grant P-N/024/2006.

References Brada M, Hoang-Xuan K, Rampling R, Dietrich PY, Dirix LY, Macdonald D, Heimans JJ, Zonnenberg BA, Bravo-Marques JM, Henriksson R, Stupp R, Yue N, Bruner J, Dugan M, Rao S, Zaknoen S (2001) Multicenter phase II trial of temozolomide in patients with glioblastoma multiforme at first relapse. Ann Oncol 12:259–266 Chang SM, Wen P, Cloughesy T, Greenberg H, Schiff D, Conrad C, Fink K, Robins HI, De Angelis L, Raizer J, Hess K, Aldape K, Lamborn KR, Kuhn J, Dancey J, Prados MD (2005) Phase II study of CCI-779 in patients with recurrent glioblastoma multiforme. Invest New Drugs 23: 357–361 Ciechomska I, Legat M, Golab J, Wesolowska A, Kurzaj Z, Mackiewicz A, Kaminska B (2005) Cyclosporine A and its non-immunosuppressive derivative NIM811 induce apoptosis of malignant melanoma cells- in vitro and in vivo studies. Int J Cancer 117:59–67 Ciechomska I, Pyrzynska B, Kazmierczak P, Kaminska B (2003) Inhibition of Akt kinase signalling and activation of Forkhead are indispensable for up-regulation of FasL expression in apoptosis of glioma cells. Oncogene 22: 7617–7627 Clarke J, Butowski N, Chang S (2010) Recent advances in therapy for glioblastoma. Arch Neurol 67:279–283 Cloughesy TF, Yoshimoto K, Nghiemphu P, Brown K, Dang J, Zhu S, Hsueh T, Chen Y, Wang W, Youngkin D, Liau L, Martin N, Becker D, Bergsneider M, Lai A, Green R, Oglesby T, Koleto M, Trent J, Horvath S, Mischel PS, Mellinghoff IK, Sawyers CL (2008) Antitumor activity of rapamycin in a phase I trial for patients with recurrent PTEN-deficient glioblastoma. PLoS Med 5:e8 Dello Russo C, Lisi L, Tringali G, Navarra P (2009) Involvement of mTOR kinase in cytokine-dependent microglial activation and cell proliferation. Biochem Pharmacol 78:1242–1251

25 Glioblastoma: Anti-tumor Action of Cyclosporin A and Functionally Related Drugs Doherty L, Gigas DC, Kesari S, Drappatz J, Kim R, Zimmerman J, Ostrowsky L, Wen PY (2006) Pilot study of the combination of EGFR and mTOR inhibitors in recurrent malignant gliomas. Neurology 67:156–158 Galanis E, Buckner JC, Maurer MJ, Kreisberg JI, Ballman K, Boni J, Peralba JM, Jenkins RB, Dakhil SR, Morton RF, Jaeckle KA, Scheithauer BW, Dancey J, Hidalgo M, Walsh DJ (2005) Phase II trial of temsirolimus (CCI-779) in recurrent glioblastoma multiforme: a North Central Cancer Treatment Group study. J Clin Oncol 23:5294–5304 Guertin DA, Sabatini DM (2007) Defining the role of mTOR in cancer. Cancer Cell 12:9–22 Hussain SF, Yang D, Suki D, Grimm E, Heimberger AB (2006) Innate immune functions of microglia isolated from human glioma patients. J Transl Med 30:15–24 Kaminska B, Figiel I, Pyrzynska B, Czajkowski R, Mosieniak G (2001) Treatment of hippocampal neurons with cyclosporin a results in calcium overload and apoptosis which are independent on NMDA receptor activation. Br J Pharmacol 133:997–1004 Kreisl TN, Lassman AB, Mischel PS, Rosen N, Scher HI, Teruya-Feldstein J, Shaffer D, Lis E, Abrey LE (2009) A pilot study of everolimus and gefitinib in the treatment of recurrent glioblastoma (GBM). J Neurooncol 92:99–105 Mancini M, Toker A (2009) NFAT proteins: emerging roles in cancer progression. Nat Rev Cancer 9:810–820 Markovic DS, Glass R, Synowitz M, Rooijen N, Kettenmann H (2005) Microglia stimulate the invasiveness of glioma cells by increasing the activity of metalloprotease-2. J Neuropathol Exp Neurol 64:754–762 Markovic DS, Vinnakota K, Chirasani S, Synowitz M, Raguet H, Stock K, Sliwa M, Lehmann S, Kälin R, van Rooijen N, Holmbeck K, Heppner FL, Kiwit J, Matyash V, Lehnardt S, Kaminska B, Glass R, Kettenmann H (2009) Gliomas induce and exploit microglial MT1-MMP expression for tumor expansion. Proc Natl Acad Sci USA 106:12530–12535 Mathiesen T, Collins VP, Olson L, Granholm L (1989) Prolonged survival and vascularization of xenografted human glioblastoma cells in the central nervous system of cyclosporine a treated rats. Cancer Lett 44:151–156 McCubrey JA, Steelman LS, Abrams SL, Lee JT, Chang F, Bertrand FE, Navolanic PM, Terrian DM, Franklin RA, D’Assoro AB, Salisbury JL, Mazzarino MC, Stivala F, Libra M (2006) Roles of the RAF/MEK/ERK and PI3K/PTEN/AKT pathways in malignant transformation and drug resistance. Adv Enzyme Regul 46:249–279 Mosieniak G, Figiel I, Kaminska B (1997) Cyclosporin A, an immunosuppressive drug, induces programmed cell death in rat C6 glioma cells by a mechanism that involves the AP-1 transcription factor. J Neurochem 68:1142–1149 Ohgaki H, Kleihues P (2009) Genetic alterations and signaling pathways in the evolution of gliomas. Cancer Sci 100: 2235–2241 Pyrzynska B, Mosieniak G, Kaminska B (2000) Changes of the trans-activating potential of AP-1 transcription factor during cyclosporin a-induced apoptosis of glioma cells are mediated by phosphorylation and alterations of AP-1 composition. J Neurochem 74:42–51

253

Pyrzynska B, Serrano M, Martinez-A C, Kaminska B (2002) Tumor suppressor p53 mediates apoptotic cell death triggered by cyclosporine A. J Biol Chem 277: 14102–14108 Rajasekhar VK, Viale A, Socci ND, Wiedmann M, Hu X, Holland EC (2003) Oncogenic Ras and Akt signaling contribute to glioblastoma formation by differential recruitment of existing mRNAs to polysomes. Mol Cell 12: 889–901 Rao RD, Mladek AC, Lamont JD, Goble JM, Erlichman C, James CD, Sarkaria JN (2005) Disruption of parallel and converging signaling pathways contributes to the synergistic antitumor effects of simultaneous mTOR and EGFR inhibition in GBM cells. Neoplasia 7:921–929 Reardon DA, Desjardins A, Vredenburgh JJ, Gururangan S, Friedman AH, Herndon JE 2nd, Marcello J, Norfleet JA, McLendon RE, Sampson JH, Friedman HS (2010) Phase 2 trial of erlotinib plus sirolimus in adults with recurrent glioblastoma. J Neurooncol 96:219–230 Schiff D (2004) Gliomas following organ transplantation: analysis of the contents of a tumor registry. J Neurosurg 101: 932–934 Sliwa M, Markovic D, Gabrusiewicz K, Synowitz M, Glass R, Zawadzka M, Wesolowska A, Kettenmann H, Kaminska B (2007) The invasion promoting effect of microglia on glioblastoma cells is inhibited by cyclosporin A. Brain 130:476–489 Strojnik T, Kavalar R, Lah TT (2006) Experimental model and immunohistochemical analyses of U87 human glioblastoma cell xenografts in immunosuppressed rat brains. Anticancer Res 26:2887–2900 Tate MC, Aghi MK (2009) Biology of angiogenesis and invasion in glioma. Neurotherapeutics 6:447–457 Watters JJ, Schartner JM, Badie B (2005) Microglia function in brain tumors. J Neurosci Res 81:447–455 Wei LH, Su H, Hildebrandt IJ, Phelps ME, Czernin J, Weber WA (2008) Changes in Tumor Metabolism as Readout for Mammalian Target of Rapamycin Kinase Inhibition by Rapamycin in Glioblastoma. Clin Cancer Res 14: 3416–3426 Wesolowska A, Kwiatkowska A, Slomnicki L, Dembinski M, Master A, Sliwa M, Franciszkiewicz K, Chouaib S, Kaminska B (2008) Microglia-derived TGF-beta as an important regulator of glioblastoma invasion-an inhibition of TGF-beta-dependent effects by shRNA against human TGF-beta type II receptor. Oncogene 27:918–930 Yang I, Han SJ, Kaur G, Crane C, Parsa AT (2010) The role of microglia in central nervous system immunity and glioma immunology. J Clin Neurosc 17:6–10 Zupanska A, Adach A, Dziembowska M, Kaminska B (2007) Alternative pathway of transcriptional induction of p21WAF1/Cip1 by cyclosporine A in p53deficient human glioblastoma cells. Cell Signal 19: 1268–1278 Zupanska A, Dziembowska M, Ellert-Miklaszewska A, Gaweda-Walerych K, Kaminska B (2005) Cyclosporine A induces growth arrest or programmed cell death of human glioma cells. Neurochem Int 47:430–441

Chapter 26

Glioblastoma Patients: Chemotherapy with Cisplatin, Temozolomide and Thalidomide Fable Zustovich and Giuseppe Lombardi

Abstract When relapse occurs the treatment of glioblastoma patients remains a challenge for oncologists, the prognosis is poor and no therapy has proven benefit on survival. There are preclinical evidences of a synergism between cisplatin and temozolomide and between cisplatin and thalidomide. In the contest of a phase I and a subsequent phase II study we treated 17 patients with relapsed high-grade malignant glioma with cisplatin plus temozolomide and then with the same combination plus thalidomide. Toxicity was generally manageable but an incidence of 36.4% of vascular thrombo-embolic events was observed in patients treated with thalidomide. Sixteen patients were evaluable for response: 1 patient had a partial response, 8 patients had stable disease and 5 patients had progression. The median time to progression was 4 months (range 0.9–14.9), the progression-free survival at 6 month was 28% and the median overall survival was 7.8+ months (range 2.6–24.2+). Due to the high incidence of thalidomide-associated thrombo-embolic events we continued the phase II study treating patients only with cisplatin and temozolomide. The availability of new antiangiogenetic agents as bevacizumab could in the future improve the efficacy and tolerability of such combinations. Keywords Cisplatin · Temozolomide · Thalidomide · Toxicity · AGAT · Glioblastoma

F. Zustovich () Oncologia Medica 1, I.O.V. – IRCCS, Ospedale Busonera, 35128 Padova, Italy e-mail: [email protected]

Introduction Malignant gliomas are relatively rare diseases. Although prognosis is poor, recent improvement in the multi-disciplinary approach, such as earlier diagnosis, better and repeated surgery plus local chemotherapy, improved radiotherapy and more effective systemic chemotherapy, has prolonged the survival. Long-term survivors are more frequent in the clinical practice, the median overall survival of patients with glioblastoma has reached the 15 months in phase III studies and ~20 months for selected patients in phase II studies. In a phase III study, temozolomide at a daily dose of 75 mg/m2 when associated with standard radiotherapy, and then administered at a dose of 200 mg/m2 day for 5 days every 28 days for at least 6 cycles, has proven to provide a statistically significant advantage in overall 2-year survival (26.5 vs. 10.4%) in comparison to standard radiotherapy alone in patients with glioblastoma after surgical resection (Stupp et al., 2005). The further analysis of the O6 -methylguanine-DNA methyltransferase (MGMT) methylation status in the tumour cells gave more details on survival results and in particular how patients with methylated-MGMT (45% of the total) had a median survival of 18.2 months compared to 12.2 months of the non-methylated counterpart (Hegi et al., 2005). There are no phase III trials investigating second line therapy for patients progressing during or after temozolomide treatment. Surgery or further chemotherapy is always considered whenever possible. Some data are available using more temozolomide treatment with modified schedules, the combination of procarbazine, carmustine and vincristine (PCV regimen), irinotecan and carmustine, fotemustine and

M.A. Hayat (ed.), Tumors of the Central Nervous System, Volume 2, DOI 10.1007/978-94-007-0618-7_26, © Springer Science+Business Media B.V. 2011

255

256

carmustine, among the other drugs. Unfortunately, all these attempts of treatment are characterized by low response rates and the lack of durability when responses occur. In contrast, the use of inhibitors of the angiogenetic pathways appeared very promising. In a phase II study bevacizumab, a monoclonal antibody, targeted against the vascular endothelial growth factor (VEGF), in combination with irinotecan, gave an overall response rate (ORR) of 57% in 35 patients with progressive refractory glioblastoma, the 6-month progression-free survival (PFS-6) was 46% and the 6-month overall survival was 77% (Goli et al., 2007).

Rationale Temozolomide is an alkylating agent which belongs to the family of imidazotetrazine. It can be orally used, and permeates through the brain-blood barrier. Temozolomide has an excellent tolerability profile at the usual dose of 200 mg/m2 a day for five consecutive days every 4 weeks. The dose limiting toxicity is thrombocytopenia. Initial phase II studies in patients with relapsed glioblastomas showed that temozolomide given in monotherapy was able to obtain an ORR of 19% and a PFS-6 of 24% (Brandes et al., 2000). Temozolamide acts through methylation of the O6 position of guanylic acid in DNA with intrachain links. Consequent mismatch-repair system activation leads to apoptotic cell death. Resistance to temozolomide is due to high intracellular levels of O6 -alkyl-guaninealkyltransferase (AGAT) which is involved in DNArepair mechanisms, transferring an alkyl group from DNA to another cysteine residue. Therefore, cells with increased concentrations of AGAT or with a deficiency in mismatch repair before drug administration, may be resistant to temozolomide (D’Incalci et al., 1988). Cisplatin is able to reduce in vitro intracellular AGAT levels in proportion to the drug concentration and the duration of cell exposure to the drug with an AGAT nadir after 24–48 h (D’Atri et al., 2000). The promoting region of AGAT gene is rich in guanine and cytosine (GC) sequences that have a high affinity for cisplatin. The presence of cisplatin reduces the activity of the gene encoding for AGAT and then impair one of the potential mechanisms of resistance to temozolomide.

F. Zustovich and G. Lombardi

This is the basis for a synergistic effect between temozolomide and cisplatin. Furthermore, platinum compounds have their own cytotoxic activity against malignant gliomas (Yung et al., 1991; Bertolone et al., 1989). A combination of cisplatin and temozolomide in chemo-naïve patients with glioblastoma was investigated in a first-line phase II study. Cisplatin was administered at 75 mg/m2 every 28 days followed by oral temozolomide at a dose of 130 mg/m2 bolus followed by nine doses of 70 mg/m2 every 12 h escalated to 1000 mg/m2 in 5 days if no toxicity occurred. The ORR was 20.4% with a PFS-6 of 34% (Brandes et al., 2004). Thalidomide has a well known anti-angiogenic activity in vivo (D’Amato et al., 1994). Malignant gliomas are rich in blood vessels and vascularization is proportional to the grade of malignancy (Takahashi et al., 1992). Many authors evaluated the feasibility and effectiveness of thalidomide in phase I–II studies. Fine et al. (2000) administered daily doses of thalidomide from 800 to 1200 mg in 39 patients with relapsed high grade malignant gliomas, obtaining two partial responses (PR), two minor responses (MR), and 12 instances of stable disease (SD). Among 18 patients with relapsed high grade malignant gliomas treated with daily 100 mg of thalidomide, Short et al. (2001) obtained 1 PR and 2 SD. Marx et al. (2001) treated 38 patients with relapsed glioblastoma with a daily escalating dose of thalidomide from 100 to 500 mg, obtaining 2 PR and 16 SD. Morabito et al. (2004) administered daily 400 mg of thalidomide to 18 patients with relapsed glioblastoma obtaining 1 MR response and 8 SD. A synergism between thalidomide and cisplatin was documented in vitro. Thalidomide seems to enhance the vascular permeability and to improve the tumour vascularization enhancing intratumoral cisplatin concentration (Murphy et al., 2007). Unfortunately, so far, there are no published clinical data evaluating the combination of thalidomide and cisplatin other than a phase I study published by Zustovich et al. (2007). In conclusion, the combination of temozolomide, cisplatin and thalidomide is based on two different in vitro demonstrated synergies, one between temozolomide and cisplatin and the other between cisplatin and thalidomide.

26 Glioblastoma Patients: Chemotherapy with Cisplatin, Temozolomide and Thalidomide

Clinical Experience The clinical experience using temozolomide, cisplatin, and thalidomide in patients with high grade malignant gliomas is based on a published dose-finding phase I study (Zustovich et al., 2007) and the following attempt to perform a phase II trial. Unfortunately, this study concluded prematurely due to the high incidence of vascular thrombo-embolic events (VTE). Nevertheless, the data analysis led to interesting conclusions that overcome the mere dose-finding investigation and revealed some interesting concerns regarding myelotoxicity and vascular venous complications, as described in the following paragraphs. Patients with relapsed malignant primitive brain tumors (not eligible for further local treatment) were enrolled. Patients that had received prior chemotherapy were admitted, including those treated with temozolomide if not combined with other alkylating agents. Treatment was started with cisplatin at 75 mg/m2 day 1 in combination with daily temozolomide 100 mg/m2 for 5 days every 3 weeks. Progressively increasing dose levels were established to verify, firstly, the tolerance to the 3-week schedule of cisplatin and temozolomide with the usual cisplatin dosage of 75 mg/m2 and temozolomide at the same dose-intensity of the standard 4-week schedule; thus, 150 mg/m2 , for 5 days, instead of 200 mg/m2 . Thalidomide was then added in 50 mg step escalating doses starting from 100 mg total dose per day. From April 2002 to March 2005, we enrolled 17 consecutive eligible patients. Of

Fig. 26.1 Left inferior limb arteriography of the patients with artery occlusion probably induced by thalidomide administration

257

the 14 patients with glioblastoma, 11 had previously undergone cytoreductive surgery and 13 had received radiotherapy. All 17 patients were evaluable for toxicity, which was generally mild. The most frequent grade 1–2 toxicities were anemia, nausea, and vomiting. Four patients had grade 1–2 cutaneous/allergic toxicity, one of them also complained of dyspnoea. Grade 3–4 toxicity was febrile neutropenia in one patient, asymptomatic neutropenia in 6 patients, grade 4 thrombocytopenia in 1 patient, deep venous thrombosis (DVT) in 3 patients and pulmonary embolism (PE) in one patient, with an overall VTE incidence of ~23.5% and of 36.4% in the sub-group receiving thalidomide. In these patients, thalidomide was discontinued and not readministered, even after the onset of an anti-coagulant therapy. Sixteen patients were evaluable for response and all for time to progression (TTP) and overall survival (OS). Among the 14 patients with glioblastoma, 1 patient had a PR, 8 patients had SD, and 5 patients had progressive disease. Median TTP was 4 months (range 0.9–14.9 months), PFS-6 was 28% and median OS was 7.8+ months (range 2.6–24.2+ months). As stated before, we planned to continue with a phase II study but an early case of artery thrombotic occlusion occurred (see Fig. 26.1). As a consequence, and in consideration of the high rate VTE in the dose-finding trial, we decided to continue with the sole administration of temozolomide and cisplatin, without thalidomide (see the paragraph “Future perspectives”).

258

Mielotoxicity The dose-finding escalation of thalidomide in combination with cisplatin and temozolomide (75 mg/m2 day 1 and 150 mg/m2 days 1–5, respectively, every 3 weeks) was concluded at the level of 200 mg daily total dose. Dose limiting toxicity (DLT) was hemathological with grade 4 thrombocytopenia in one patient and grade 4 febrile neutropenia in another patient. Moreover, 6 out of the 7 patients who developed grade 3–4 neutropenia belonged to the group treated with thalidomide. Thus, we can infer that thalidomide impairs the bone-marrow function in proportion to the dose administered. In a series of 44 patients with multiple myeoloma treated with thalidomide there were 10 cases of grade 3–4 neutropenia and 5 patients had a bone-marrow hypoplasia without any increase of myeloma cell count (Hattori et al., 2004). The dynamic contrast enhanced magnetic resonance (dMRI) in patients with multiple myeloma evidenced as thalidomide is able to reduce the microcirculation in the bone marrow (Scherer et al., 2002). This reduction is greater when thalidomide is administered in combination with chemotherapy (Wasser et al., 2004). A reduction of bone marrow micro vessels density in patients with multiple myeloma tretaed with thalidomide is also been demonstrated (Kumar et al., 2004).

Vascular Complications At first, we did not consider vascular venous grade 3–4 toxicity as a dose limiting toxicity in consideration of the high incidence of VTE in patients with high-grade glioma. In a series of 68 patients with highgrade gliomas, the rate of VTE was 19% and authors reported as risk factors chemotherapy administration (p = 0.027) and the presence of paresis (p = 0.031) (Dhami et al., 1993). The overall rate of VTE in our series of patients treated with the combination of temozolomide, cisplatin and thalidomide was 36.4% and we supposed that thalidomide might have been a VTE risk factor in high grade glioma patients. There are previous reports regarding thalidomideinduced vascular venous grade 3–4 toxicity in cancer patients. Data from FDA’s MedWatch showed an incidence of VTE of 4.6% in patients treated with

F. Zustovich and G. Lombardi

thalidomide alone, of 15.0% in patients receiving concomitant dexamethasone and of 30.9% in patients receiving concomitant chemotherapy (Bennet et al., 2001). In multiple myeloma patients there was a higher incidence of VTE when thalidomide was associated to chemotherapy regimen containing dexamethasone, vincristine, doxorubicin, cyclophosphamide, etoposide and cisplatin (28% vs. 4%; p = 0.002) (Zangari et al., 2001). A low incidence of VTE occurred when thalidomide was administered in combination with a non-doxorubicin containing chemotherapy regimen (dexamethasone, cyclo-phosphamide, and etoposide) (Moehler et al., 2002). A comparison between 2 chemotherapy regimens differing only for the presence or not of doxorubicin revealed as doxorubicin was a VTE risk factor (16.0% vs 2.5%; p = 0.02) (Zangari et al., 2002). A phase II study of temozolomide and thalidomide in patients with brain metastases of melanoma was prematurely interrupted after the enrolment of 16 patients for a high incidence of VTE. Authors reported 3 cases of PE and 1 case of DVT with an overall VTE rate of 25% (Krown et al., 2006). That was not far from 19% reported in high-grade glioma patients series (Dhami et al., 1993). Unfortunately, besides ours, no published data regarding the combination of cisplatin and thalidomide are available. We suppose that the concomitant administration of cisplatin and thalidomide, like doxorubicin in multiple myeloma patients, might be a VTE risk factor in patients with malignant high-grade gliomas.

Future Perspectives The new generation of drugs classified as “Target therapy” have been tested in malignant gliomas. The inhibitors of the EGFR pathways gave poor result but antiangiogenetic drugs, such as bevacizumab, seemed to be very promising. Bevacizumab is a monoclonal antibody targeted against VEGF. In a phase II study, 35 patients with progressive refractory glioblastoma received bevacizumab in combination with irinotecan. Overall response rate was 57%, PFS-6 was 46% and the 6-month OS was 77% (Goli et al., 2007). In a recent phase II study bevacizumab alone gave an ORR of 35% in heavily pre-treated patients with a PFS-6 of

26 Glioblastoma Patients: Chemotherapy with Cisplatin, Temozolomide and Thalidomide

29% while the addition of irinotecan did not seem to improve the outcome (Kreisl et al., 2008). I believe that the combination of temozolomide, cisplatin, and bevacizumab could be safe and even more effective. Bevacizumab has been combined to temozolomide in high-grade glioma patients and to cisplatin in non-small cell lung cancer patients without the evidence of unexpected or increased vascular and haematological toxicity. Our still unpublished data treating temozolomide refractory glioblastoma patients with the combination of temozolomide and cisplatin are promising with a PR rate of 36% with 2 long lasting PFS of 16.9 and 13.2+ months after 14 enrolled patients. A phase I–II study evaluating the combination of temozolomide, cisplatin, and bevacizumab in glioblastoma patients has been designed, but until sufficient clinical data are gathered, it will remain an appropriate suggestion.

References Bennet CL, Schumok GT, Kwaan HC, Raisch DW (2001) High incidence of Thalidomide-associated deep vein thrombosis and pulmonary emboli when chemotherapy is also administered. Blood 98:863a (Abstract) Bertolone SJ, Baum ES, Krivit W, Hammond GD (1989) A phase II study of cisplatin therapy in recurrent childhood brain tumors. A report from the childrens cancer study group. J Neurooncol 1:5–11 Brandes AA, Basso U, Reni M, Vastola F, Tosoni A, Cavallo G, Scopece L, Ferreri AJ, Panucci MG, Monfardini S, Ermani M (2004) First-line chemotherapy with cisplatin plus fractionated temozolomide in recurrent glioblastoma multiforme: a phase II study of the Gruppo Italiano Cooperativo di Neuro-Oncologia. J Clin Oncol 9:1598–1604 Brandes AA, Pasetto LM, Vastola F, Monfardini S (2000) Temozolomide in patients with high grade gliomas. Oncology 3:181–186 D’Amato RJ, Laughnan MS, Flynn E, Folkman J (1994) Thalidomide is an inhibitor of angiogenesis. Proc Natl Acad Sci USA 9:4082–4085 D’Atri S, Graziani G, Lacal PM, Nistico V, Gilberti S, Faraoni I, Watson AJ, Bonmassar E, Margison GP (2000) Attenuation of O(6)-methylguanine-DNA methyltransferase activity and mRNA levels by cisplatin and temozolomide in jurkat cells. J Pharmacol Exp Ther 2:664–671 Dhami MS, Bona RD, Calogero JA, Helman RM (1993) Venous thromboembolism and high grade gliomas. Thromb Haemost 3:393–396 D’Incalci M, Citti L, Taverna P, Catapano CV (1988) Importance of the DNA repair enzyme O6 -alkyl guanine alkyltransferase (AT) in cancer chemotherapy. Cancer Treat Rev 4: 279–292

259

Fine HA, Figg WD, Jaeckle K, Wen PY, Kyritsis AP, Loeffler JS, Levin VA, Black PM, Kaplan R, Pluda JM, Yung WK (2000) Phase II trial of the antiangiogenic agent thalidomide in patients with recurrent high-grade gliomas. J Clin Oncol 4:708–715 Goli KJ, Desjardins JE, Herndon JE, Rich JN, Reardon DA, Quinn JA, Sathornsumetee S, Bota DA, Friedman HS, Vredenburgh JJ (2007) Phase II trial of bevacizumab and irinotecan in the treatment of malignant gliomas. Proc ASCO 2007:2003 J Clin Oncol 25(18s) (abstract) Hattori Y, Kakimoto T, Okamoto S, Sato N, Ikeda Y (2004) Thalidomide-induced severe neutropenia during treatment of multiple myeloma. Int J Hematol 3:283–288 Hegi ME, Diserens AC, Gorlia T, Hamou MF, de Tribolet N, Weller M, Kros JM, Hainfellner JA, Mason W, Mariani L, Bromberg JE, Cairncross JG, Janzer RC, Stupp R (2005) MGMT gene silencing and benefit from temozolomide in glioblastoma. N Engl J Med 10:997–1003 Kreisl TN, Kim L, Moore K, Duic P, Royce C, Stroud I, Garren N, Mackey M, Butman JA, Camphausen K, Park J, Albert SA, Fine HA (2008) Phase II trial of single-agent bevacizumab followed by bevacizumab plus irinotecan at tumor progression in recurrent glioblastoma. J Clin Oncol 15(10):3617–3623 (Published on line ahead of print Dec 29 2008) Krown SE, Niedzwiecki D, Hwu WJ, Hodgson L, Houghton AN, Haluska FG et al (2006) Phase II study of temozolomide and thalidomide in patients with metastatic melanoma in the brain. Cancer 107(8):1883–1890 Kumar S, Witzig TE, Dispenzieri A, Lacy MQ, Wellik LE, Fonseca R, Lust JA, Gertz MA, Kyle RA, Greipp PR, Rejkumar SV (2004) Effect of thalidomide therapy on bone marrow angiogenesis in multiple myeloma. Leukemia 3:624–627 Marx GM, Pavlakis N, McCowatt S, Boyle FM, Levi JA, Bell DR, Cook R, Biggs M, Little N, Wheeler HR (2001) Phase II study of thalidomide in the treatment of recurrent glioblastoma multiforme. J Neurooncol 1:31–38 Moehler TM, Dchlenzka J, Kaspar B, Neben K, Egerer G, Ho AD, Goldschmidt H (2002) Low incidence of deep venous thrombosis in poor prognosis multiple myeloma patients treated with thalidomide and CED chemotherapy. Blood 100:5136a (abstract) Morabito A, Fanelli M, Carillio G, Gattuso D, Sarmiento R, Gasparini G (2004) Thalidomide prolongs disease stabilization after conventional therapy in patients with recurrent glioblastoma. Oncol Rep 1:93–95 Murphy S, Davey RA, Gu XQ, Haywood MC, McCann LA, Mather LE, Boyle FM (2007) Enhancement of cisplatin efficacy by thalidomide in a 9L rat gliosarcoma model. J Neurooncol 85:181–189 Scherer A, Strupp C, Wittsack HJ, Engelbrecht V, Willers R, Germing U, Gattermann N, Haas R, Modder U (2002) Dynamic contrast-enhanced MRI for evaluating bone marrow microcirculation in malignant haematological diseases before and after thalidomide therapy. Radiologie 3:222–230 Short SC, Traish D, Dowe A, Hines F, Gore M, Brada M (2001) Thalidomide as an anti-angiogenic agent in relapsed gliomas. J Neurooncol 1:41–45 Stupp R, Mason WP, van den Bent MJ, Weller M, Fisher B, Taphoorn MJ, Belanger K, Brandes AA, Marosi C, Bogdahn

260 U, Curschmann J, Janzer RC, Ludwin SK, Gorlia T, Allgeir A, Lacombe D, Cairncross JG, Eisenhauer E, Mirimanoff RO (2005) European organisation for research and treatment of cancer brain tumor and radiotherapy groups; National Cancer Institute of Canada clinical trials group. 2005. Radiotherapy plus concomitant and adjuvant temozolomide for glioblastoma. N Engl J Med 10:987–996 Takahashi JA, Fukumoto M, Igarashi K, Oda Y, Kikuchi H, Hatanaka M (1992) Correlation of basic fibroblast growth factor expression levels with the degree of malignancy and vascularity in human gliomas. J Neurosurg 5: 792–798 Wasser K, Moehler T, Neben K, Nosas S, Heiss J, Goldschmidt H, Hillengass J, Duber C, Kauczor HU, Delorme S (2004) Dynamic MRI of the bone marrow for monitoring multiple myeloma during treatment with thalidomide as monotherapy or in combination with CED chemotherapy. Rofo 9: 1285–1295

F. Zustovich and G. Lombardi Yung WK, Mechtler L, Gleason MJ (1991) Intravenous carboplatin for recurrent malignant glioma: a phase II study. J Clin Oncol 5:860–864 Zangari M, Anaissie E, Barlogie B, Badros A, Desikan R, Gopal AV, Morris C, Toor A, Siegel E, Fink L, Tricot G (2001) Increased risk of deep-vein thrombosis in patients with multiple myeloma receiving thalidomide and chemotherapy. Blood 5:1614–1615 Zangari M, Siegel E, Barlogie B, Anaissie E, Saghafifar F, Fassas A, Morris C, Fink L, Tricot G (2002) Thrombogenic activity of doxorubicin in myeloma patients receiving thalidomide: implications for therapy. Blood 4: 1168–1171 Zustovich F, Cartei G, Ceravolo R, Zovato S, Della Puppa A, Pastorelli D, Mattiazzi M, Bertorelle R, Gardiman MP (2007) A phase I study of cisplatin, temozolomide and thalidomide in patients with malignant brain tumors. Anticancer Res 27(2):1019–1024

Chapter 27

Glioblastoma: Role of Galectin-1 in Chemoresistance Florence Lefranc and Robert Kiss

Abstract Malignant gliomas, especially glioblastomas, are associated with a poor prognosis. This prognosis can be at least partly explained by the fact that glioma cells diffusely infiltrate the brain parenchyma and exhibit decreased levels of apoptosis and are thus resistant to cytotoxic drugs. Galectin-1, the expression of which is stimulated by hypoxia, is a potent modulator of glioblastoma cell migration and a pro-angiogenic molecule. Galectin-1 participates in the resistance of cancer cells, including glioma cells, to chemotherapy and to radiotherapy and is involved in the activation of the Ras oncogenic pathway. Our recent data reveal that temozolomide, the standard treatment for glioma patients, increases Galectin-1 expression in various glioblastoma models both in vitro and in vivo. Consequently, reducing Galectin-1 expression in these models increases the anti-tumor effects of various chemotherapeutic agents, in particular temozolomide. Reducing Galectin-1 expression in glioblastoma cells does not induce apoptotic or autophagic features, but rather modulates p53 transcriptional activity and decreases p53-targeted gene expression. The decrease in Galectin-1 expression also impairs the expression levels of several genes implicated in chemoresistance: ORP150, HERP, GRP78/Bip, TRA1, BNIP3L, GADD45B and CYR61. The involvement of Galectin1 in different steps of glioma malignant progression, such as migration, angiogenesis or chemoresistance, makes it a potential target for the development of new drugs to combat these malignant tumors.

F. Lefranc () Laboratoire de Toxicologie, Faculté de Pharmacie, Université Libre de Bruxelles (ULB), 1050 Brussels, Belgium e-mail: [email protected]

Keywords GBM · Apoptosis · Galectin-1 · Cell death · Chemoresistance · Ras

Introduction Malignant gliomas, especially glioblastomas (GBM), are characterized by the diffuse invasion of distant brain tissue by a myriad of single migrating cells with reduced levels of apoptosis (type I programmed cell death [PCD]) and consequent resistance to the cytotoxic insults of pro-apoptotic drugs (Lefranc et al., 2005). In contrast, GBM cells are less resistant to autophagy-related cell death (type II PCD) than to apoptosis (Lefranc et al., 2005, 2006). Current recommendations are that patients with GBM should undergo maximum surgical resection followed by concurrent radiation and chemotherapy with temozolomide (Lefranc et al., 2006; Stupp et al., 2009). Galectin-1 (Gal1), a lectin with specificity for β-galactosides (Liu and Rabinovich, 2005; Camby et al., 2006; (LeMercier et al., 2009), markedly influences glioma cell migration both in vitro and in vivo (Camby et al., 2001, 2002, 2005). High-grade glioma patients whose gliomas markedly express Gal1 have a significantly shorter survival period than individuals whose gliomas express less Gal1 (Camby et al., 2002). Decreasing Gal1-expression in human orthotopic GBM xenografts significantly increases the survival of GBM tumorbearing mice (Camby et al., 2002; LeMercier et al., 2008b). Gal1-expression is increased under hypoxic conditions; hypoxia also confers cellular resistance to conventional chemotherapy and accelerates malignant progression (Le et al., 2005). Gal1 is negatively regulated by p53 (Puchades et al., 2007),

M.A. Hayat (ed.), Tumors of the Central Nervous System, Volume 2, DOI 10.1007/978-94-007-0618-7_27, © Springer Science+Business Media B.V. 2011

261

262

and Gal1 reciprocally has been shown to modify p53 biological functions (Camby et al., 2006). The reciprocal control exerted by Gal1 or p53 on the other could be implicated in GBM chemoresistance, as could also be the case in the partnership between Ras and Gal1. Ras is implicated in gliomagenesis and modulates glioma aggressiveness (Goldberg and Kloog, 2006). Ras signaling and oncogenesis depend on the dynamic interplay of Ras with distinctive plasma membrane micro-domains and various intracellular compartments. Such interactions are dictated by individual elements in the carboxy-terminal domain of the Ras proteins, one of which is recognized by Gal1, galectin3 and cGMP phosphodiesterase delta (Ashery et al., 2006). Gal1 thereby promotes H-Ras signaling to Raf at the expense of phosphoinositide 3-kinase (PI3K) and Ral guanine nucleotide exchange factor (RalGEF) (Ashery et al., 2006). Gal1 could therefore be involved in these Ras-related pathways implicated in GBM resistance to apoptosis (Lefranc et al., 2005; LeMercier et al., 2009).

F. Lefranc and R. Kiss

when on the current standard treatment (surgical resection to the extent feasible, followed by adjuvant radiotherapy plus temozolomide chemotherapy, given concomitantly with and after radiotherapy) (Lefranc et al., 2005, 2006; Stupp et al., 2009). Malignant gliomas are associated with such dismal prognoses because glioma cells can actively migrate through the narrow extracellular spaces in the brain, often traveling relatively long distances, making them elusive targets for effective surgical management (Lefranc et al., 2005, 2006). Additionally, after surgical resection and adjuvant treatment of malignant gliomas, the residual cancer cells peripheral to the excised lesion give rise to a recurrent tumor that in more than 90% of cases develops immediately adjacent to the resection margin (Lefranc et al., 2005). Clinical and experimental data have also demonstrated that invasive malignant glioma cells show a decrease in proliferation rate and a relative resistance to apoptosis as compared with the highly dense cellular center of the tumor. This may contribute to their resistance to conventional pro-apoptotic chemotherapy and radiotherapy (Lefranc et al., 2005, 2006).

Gliomas: An Overview Galectins: An Overview Gliomas account for more than 50% of all primary brain tumors and are by far the most common primary brain tumor in adults (Lefranc et al., 2005, 2006). Gliomas include tumors that are composed predominantly of astrocytes (astrocytomas), oligodendrocytes (oligodendrogliomas), ependymal cells (ependymomas) or a mixture of various glial cells (e.g., oligoastrocytomas) (Louis et al., 2007). The World Health Organization grading system classifies gliomas as grade I to IV based on the degree of malignancy, as determined by histopathological criteria. Grade I gliomas are generally well-circumscribed and behave in a benign fashion, whereas grade II through IV gliomas are malignant and diffusely infiltrate the brain (Louis et al., 2007). Among gliomas, astrocytomas are the most common and are comprised of pilocytic astrocytomas (grade I), diffuse astrocytomas (grade II), anaplastic astrocytomas (grade III) and GBM (grade IV) (Louis et al., 2007). Glioblastomas are characterized by a very dismal prognosis (Lefranc et al., 2005, 2006; Stupp et al., 2009). GBM patients have a median survival expectancy of only 14 months

Galectins are a structurally-related family of animal lectins defined by two properties: (i) an affinity for β-galactoside sugars; and (ii) sequence homology (Barondes et al., 1994; Liu and Rabinovich, 2005; Camby et al., 2006; Le Mercier et al., 2009, 2010). This consensus sequence corresponds to the carbohydrate-recognition domain (CRD), which is a beta sandwich of about 135 amino acids long and is responsible for β-galactoside binding (Barondes et al., 1994; Liu and Rabinovich, 2005; Camby et al., 2006; Le Mercier et al., 2009). To date, 15 galectins have been characterized; they are numbered according to the chronology of their discovery (galectin-1 to galectin-15) (Barondes et al., 1994; Liu and Rabinovich, 2005; Camby et al., 2006; Le Mercier et al., 2009, 2010). The galectins known so far have either one or two CRDs within a single polypeptide chain, and neither CRD is associated with other types of well-defined protein domains. The mono-CRD galectins can be biologically active as monomers (galectin-5, -7, -10)

27 Glioblastoma: Role of Galectin-1 in Chemoresistance

or as homodimers (galectin-1, -2, -11, -13, -14, -15); the bi-CRD galectins (galectin-4, -6, -8, -9, 12) are active as monomers and might also associate into oligomers (Leffler et al., 2004). Galectin-3, a monoCRD galectin, is unique in that it contains a short proline, glycine and tyrosine rich N-terminal domain fused onto the CRD that therefore allows the formation of oligomers (Leffler et al., 2004). Galectins can segregate into multiple cell compartments. Although these proteins lack the signal sequence that would be required for secretion through the classical secretory pathway, some galectins show extracellular localization, suggesting that they are secreted through a non-classical pathway (Camby et al., 2006). Galectins are present both inside and outside cells. They function extracellularly by interacting with cell surface and extracellular matrix (ECM) glycoproteins and intracellularly by interacting, in a carbohydrate-independent manner, with cytoplasmic and nuclear proteins (Elola et al., 2007). They play a role in a wide range of processes, including cell adhesion, regulation of cell growth, apoptosis, embryonic development and immune processes-like inflammation (Perillo et al., 1998; Moiseeva et al., 1999; Liu and Rabinovich, 2005; Camby et al., 2006). A large amount of experimental evidence has been reported to support the important roles of galectins in cancer biology (Liu and Rabinovich, 2005; Rabinovich, 2005; Lahm et al., 2001; Stillman et al., 2005), including tumor angiogenesis (Moiseeva et al., 1999; Thijssen et al., 2006; Le Mercier et al., 2008b), tumor immune escape (Liu and Rabinovich, 2005) and cancer cell migration (Camby et al., 2001, 2002, 2005; Elola et al., 2007; Jung et al., 2008). In the following section, we focus our attention on the biological roles exerted by Gal1 in gliomas.

Galectin-1 in Glioma Biology The role of Gal1 in glioma biology was first suggested by Yamaoka et al. (2000) and Gunnersen et al. (2000). These groups analyzed the mRNA expression of Gal1 by northern blot in glioma specimens and glioma cell lines. Increased expression of Gal1 mRNA was shown to correlate with increased malignancy in human astrocytic tumors ranging from low-grade astrocytomas to malignant gliomas (Yamaoka et al., 2000). However,

263

no statistical analysis was made (Yamaoka et al., 2000). Two studies from our own group using clinical samples have shown that Gal1 is expressed in all glioma types and that the level of Gal1 expression makes a distinction between non-invasive (WHO grade I) and invasive (WHO grade II-IV) gliomas of the astrocytic tumor (Rorive et al., 2001; Camby et al., 2001). Specifically, we quantitatively determined (by computer-assisted microscopy) the immunohistochemical expression of Gal1 in 220 gliomas, including 151 astrocytic, 38 oligodendroglial and 31 ependymal tumors (Rorive et al., 2001). Our data revealed the expression of Gal1 in all human glioma types with no striking variation in levels among astrocytic, oligodendroglial and ependymal tumors; the level of galectin-1 expression within astrocytic tumors, however, significantly correlated with tumor grade (Rorive et al., 2001). Furthermore, expression levels of Gal1 in highgrade astrocytic tumors from patients with short-term survival periods were significantly higher than those in tumors from patients with long-term survivals (Rorive et al., 2001). The expression of Gal1 was shown to be higher in the invasive parts of xenografted GBM than in the less invasive parts, suggesting their involvement in tumor astrocyte invasion of the brain parenchyma (Camby et al., 2001). Our group has focused on the role of Gal1 in glioma cell migration. We xenografted three human glioblastoma cell lines (H4, U87 and U373) into the brains of nude mice in order to characterize the in vivo galectin-1 expression pattern in relation to the tumor invasion of the normal brain parenchyma. Immunohistochemical analysis of galectin-1 expression in human U87 and U373 glioblastoma xenografts revealed a higher level of galectin-1 expression in invasive areas as compared to the non-invasive areas of the xenografts (Camby et al., 2001; Rorive et al., 2001). Moreover, nude mice intracranially grafted with U87 or U373 cells that were constitutively expressing low levels of galectin-1 (by stable transfection with an expression vector containing the antisense galectin-1 mRNA) had longer survival periods than those grafted with U87 or U373 cells unchanged in expression levels of galectin-1 (Camby et al., 2002). Complementary in vitro studies have shown that adding purified galectin-1 to U87 human GBM cells enhanced tumor cell motility in a lactose-inhibited manner (Camby et al., 2005). This effect appeared to be related to an increase

264

F. Lefranc and R. Kiss

in polymerized filamentous actin and the expression of the small guanosine triphosphatase RhoA (Camby et al., 2002). Finally, using cDNA microarray analysis and confirmation at protein levels, we observed that the U87 GBM cells that were galectin-1 deficient by means of an antisense galectin-1-stable transfection displayed increased protein levels for p21waf/cip1, cullin-2, p53, ADAM-15 and MAP-2 (Camby et al., 2005). Major differences in ADAM-15 expression and the actin stress fiber organization were also observed (Camby et al., 2005). The ADAM family of membraneanchorage glycoproteins encompasses a catalytically active MMP domain and a disintegrin domain and may thus be involved both in the proteolytic cleavage of cell-surface proteins and in integrin-mediated cell adhesion (including alpha9beta1 integrin/ADAM-15 interactions) via the RGD-dependent and -independent binding. Collectively, these data indicate that galectin-1 enhances the migratory capabilities of tumor astrocytes and, therefore, their biological aggressiveness. These features have been recently confirmed by Strik et al. (2007) and by Jung et al. (2008). Progression-associated genetic alterations are common to different glioma types and target growthpromoting and cell-cycle control pathways resulting in focal hypoxia, necrosis and angiogenesis (Louis et al., 2007). GBM is distinguished pathologically from

lower grade tumors by necrosis and microvascular hyperplasia (Louis et al., 2007). Necrotic foci are typically surrounded by “pseudopalisading” cells – a configuration that is relatively unique to malignant gliomas and has long been recognized as an ominous prognostic feature (Louis et al., 2007). Recent investigations have demonstrated that pseudopalisades are severely hypoxic, overexpress hypoxia-inducible factor 1 (HIF-1), and secrete proangiogenic factors such as vascular endothelial growth factor (VEGF) and IL-8 (Louis et al., 2007). HIF-1 is a potent activator of angiogenesis and invasion by upregulating target genes critical for these functions; activation of the HIF-1 pathway is a common feature of gliomas and may explain the intense vascular hyperplasia often seen in GBM. Galectin-1 is also a hypoxia-regulated protein (Le et al., 2005) that has been shown recently to display major roles in angiogenesis (Thijssen et al., 2006) in both gliomas (Le Mercier et al., 2008b) and melanomas (Mathieu et al., 2007) (Fig. 27.1). Galectin-1 involvement in tumor angiogenesis was first suggested after the discovery that both vascular smooth muscle and endothelial cells express the protein (Moiseeva et al., 1999). Clausse et al. (1999) have previously shown that galectin-1 was upregulated in capillaries associated with carcinoma cells. In addition, they found that galectin-1 could mediate interactions between tumors and endothelial cells in

Fig. 27.1 In stress conditions (black squares) such as hypoxia, radiotherapy and/or chemotherapy, increase of galectin-1 expression in GBM cells may activate migration, radio/chemoresistance, immune escape and angiogenesis.

Intracellular galectin-1 could be targeted by siRNA strategies. Alternatively, extracellular galectin-1 could be the target of small inhibitors

27 Glioblastoma: Role of Galectin-1 in Chemoresistance

vitro, suggesting a potential role for galectin-1 in modulating angiogenesis. Finally, Thijssen et al. (2006) have shown that treatment with either galectin1-specific antisense oligodeoxynucleotides or with polyclonal anti-galectin-1 antibodies resulted in inhibition of endothelial cell proliferation and migration, demonstrating an essential role for galectin-1 during angiogenesis. The role of galectin-1 in tumor angiogenesis is further highlighted in galectin-1-null mice, in which tumor growth is markedly impaired because of insufficient tumor angiogenesis (Thijssen et al., 2006). We ourselves have also put forward evidence for the role of galectin-1 in the process of angiogenesis using human glioma cells. To determine how galectin-1 exerts its pro-angiogenic effects, we investigated galectin-1 signaling in the human Hs683 glioma cell line. We observed that galectin-1 signals through the endoplasmic reticulum transmembrane kinase/ribonuclease inositol-requiring 1alpha (IRE1alpha) that regulates the expression of oxygenregulated protein 150 (ORP150), which in turn controls VEGF maturation. Galectin-1 also modulates the expression of six other hypoxia-related genes (CTGF, ATF3, PPP1R15A, HSPA5, TRA1 and CYR61) that are implicated in angiogenesis. Moreover, we have recently reported that downregulating galectin-1 expression in Hs683 human glioma cells through targeted small interfering RNAs provokes a marked decrease in the expression of the brain expressed Xlinked gene (BEX2), a feature which confers increased survival in Hs683 orthotopic xenograft-bearing mice. This decrease in BEX2 expression impairs vasculogenic mimicry channel formation in vitro and angiogenesis in vivo, and modulates glioma cell adhesion, motility and invasive features (Le Mercier et al., 2009).

Galectin-1 in Glioblastoma Chemo/Radioresistance Resistance of human tumors to anticancer drugs is most often ascribed to gene mutations, gene amplification or epigenetic changes that influence the uptake, metabolism or export of drugs from single cells (Trédan et al., 2007). Another important, yet littleappreciated, cause of anticancer drug resistance is the limited ability of drugs to penetrate tumor tissue and

265

to reach all of the tumor cells at a potentially lethal concentration (Trédan et al., 2007). Although now recognized as a major contributor to malignant cancer progression and to treatment failure, the precise role of hypoxia signaling in cancer and in prognosis still needs to be further defined (Koumenis, 2006). As emphasized earlier, intra-tumoral hypoxia causes genetic changes in malignant gliomas that produce a microenvironment that selects for cells of a more aggressive phenotype. Hypoxia can initiate cell demise by apoptosis/necrosis but also prevent cell death by provoking adaptive responses that, in turn, facilitate cell proliferation or angiogenesis, thus contributing to tumor malignant progression (Zhou et al., 2006). Zhou et al. (2006) emphasize that considering that activation of HIF-1 provokes pro-survival as well as pro-death decisions under hypoxia, it will be crucial to understand decision making processes in regulating cell death, adaptation and chemoresistance. Hypoxia is also known to modulate the unfolded protein response, a coordinated program that promotes cell survival under conditions of endoplasmic reticulum (ER) stress (Koumenis, 2006), and which is known to contribute to tumor malignant progression and drug resistance of solid tumors. As mentioned earlier, hypoxia is known to activate galectin-1 expression (Le et al., 2005), and galectin-1 was found to be negatively regulated by transfection with TP53 in glioma cells (Puchades et al., 2007). We recently reported that temozolomide, the standard treatment for glioma patients (Lefranc et al., 2006; Stupp et al., 2009), increases galectin-1 expression in Hs683 glioma cells both in vitro and in vivo. In addition, galectin-1 expression was shown to be upregulated by ionizing irradiation of glioma cell lines in vitro (Strick et al., 2007). Reducing galectin-1 expression in these cells by siRNA increases the antitumor effects of various chemotherapeutic agents, in particular temozolomide both in vitro and in vivo in an orthotopic xenograft mouse model (Le Mercier et al., 2008a). We also observed that decreasing galectin-1 expression by means of an anti-galectin-1 siRNA in the mouse B16F10 metastatic melanoma model (a syngenic model whereby B16F10 melanoma cells are injected in the tail vein) also increases the therapeutic benefits contributed by temozolomide in vivo (Mathieu et al., 2007). This decrease of galectin1 expression in the B16F10 mouse melanoma cells does not modify their sensitivity to apoptosis nor

266

autophagy. However, it does induce heat shock protein 70-mediated lysosomal membrane permeabilization (LMP), a process associated with cathepsin B release into the cytosol, which in turn is believed to sensitize the cells to the proautophagic effects of temozolomide when grafted in vivo (Mathieu et al., 2007). In Hs683 glioma cells a decrease in galectin-1 expression does not induce apoptotic or autophagic features and does not induce LMP, but is found to modulate p53 transcriptional activity and decrease p53-targeted gene expression including DDIT3/GADD153/CHOP, DUSP5, ATF3 and GADD45A (Le Mercier et al., 2008a). In addition, the decrease in galectin-1 expression impairs the ER stress response, which is believed to play a role in drug resistance, and also impairs the expression levels of seven other genes implicated in chemoresistance: ORP150; HERP; GRP78/Bip; TRA1; BNIP3; GADD45B; and CYR61, some of which are also known to be modified by hypoxia (Le Mercier et al., 2008a).

Conclusions Galectins are known to play an important role in malignant cancer progression and especially malignant gliomas (Camby et al., 2006; Le Mercier et al., 2009; Le Mercier et al., 2010). Galectin-1 is involved in many different steps of glioma biology, such as migration, angiogenesis and resistance to chemotherapy and radiotherapy. We have already shown that decreasing galectin-1 expression in human GBM orthotopic xenografts in mouse brains by siRNA administration enhances the therapeutic benefits of temozolomide (Le Mercier et al., 2008b). Thus, galectins, especially galectin-1, could be important targets for the development of new anticancer drugs not only for gliomas but for other types of cancer as well (Ingrassia et al., 2006). The novel aspects of Gal1-related function in the ERS response highlighted in the present study and pertinent to the chemoresistance of glioma cells may be amenable to therapeutic manipulation. This manipulation might be achieved either through in vivo delivery of anti-Gal1 siRNA as demonstrated in one of our preliminary studies (Le Mercier et al., 2008b) or through compounds that suppress Gal1 biological activity (Ingrassia et al., 2006; Camby et al., 2008). The in vivo delivery of anti-Gal1 siRNA directly into

F. Lefranc and R. Kiss

the brain could be achieved in GBM patients by means of Ommaya reservoirs, thus minimizing/avoiding systemic exposure and potential hepatotoxicity. Acknowledgments F. Lefranc is a clinical research fellow and R. Kiss a director of research with the Belgian National FNRS (Belgium). The present chapter was supported by grants awarded by the Fonds de la Recherche Scientifique Médicale (FRSM, Belgium) and by the Fonds Yvonne Boël (Brussels, Belgium).

References Ashery U, Yizhar O, Rotblat B, Elad-Sfadia G, Barkan B, Haklai R, Kloog Y (2006) Spatiotemporal organization of ras signaling: rasosomes and the galectin switch. Cell Mol Neurobiol 26:471–495 Barondes SH, Castronovo V, Cooper DN, Cummings RD, Drickamer K, Feizi T, Gitt MA, Hirabayashi J, Hughes C, Kasai K, Leffler H, Liu FT, Lotan R, Mercurio AM, Monsigny M, Pillai S, Poirer F, Raz A, Rigby PWJ, Rini JM, Wang JL (1994) Galectins: a family of animal betagalactoside-binding lectins. Cell 76:597–598 Camby I, Belot N, Lefranc F, Sadeghi N, de Launoit Y, Kaltner H, Musette S, Darro F, Danguy A, Salmon I, Gabius HJ, Kiss R (2002) Galectin-1 modulates human glioblastoma cell migration into the brain through modifications to the actin cytoskeleton and levels of expression of small GTPases. J Neuropathol Exp Neurol 61:585–596 Camby I, Belot N, Rorive S, Lefranc F, Maurage CA, Lahm H, Kaltner H, Hadari Y, Ruchoux MM, Brotchi J, Zick Y, Salmon I, Gabius HJ, Kiss R (2001) Galectins are differentially expressed in supratentorial pilocytic astrocytomas, astrocytomas, anaplastic astrocytomas and glioblastomas, and significantly modulate tumor astrocyte migration. Brain Pathol 11:12–26 Camby I, Decaestecker C, Lefranc F, Kaltner H, Gabius HJ, Kiss R (2005) Galectin-1 knocking down in human U87 glioblastoma cells alters their gene expression pattern. Biochem Biophys Res Commun 335:27–35 Camby I, Le Mercier M, Lefranc F, Kiss R (2006) Galectin1: a small protein with major functions. Glycobiology 16: 137R–157R Camby I, Le Mercier M, Mathieu V, Ingrassia L, Lefranc F, Kiss R (2008) Galectin-1 as a potential therapeutic target for cancer progression. Drug Fut 33:12 Clausse N, van den Brûle F, Waltregny D, Garnier F, Castronovo V (1999) Galectin-1 expression in prostate tumor-associated capillary endothelial cells is increased by prostate carcinoma cells and modulates heterotypic cell-cell adhesion. Angiogenesis 3:317–325 Elola MT, Wolfenstein-Todel C, Troncoso MF, Vasta GR, Rabinovich GA (2007) Galectins: matricellular glycanbinding proteins linking cell adhesion, migration, and survival. Cell Mol Life Sci 64:1679–1700 Goldberg L, Kloog Y (2006) A Ras inhibitor tilts the balance between Rac and Rho and blocks phosphatidylinositol 3-kinase-dependent glioblastoma cell migration. Cancer Res 66:11709–11717

27 Glioblastoma: Role of Galectin-1 in Chemoresistance Gunnersen JM, Spirkoska V, Smith PE, Danks RA, Tan SS (2000) Growth and migration markers of rat C6 glioma cells identified by serial analysis of gene expression. Glia 32:146–154 Ingrassia L, Camby I, Lefranc F, Mathieu V, Nshimyumukiza P, Darro F, Kiss R (2006) Anti-galectin compounds as potential anti-cancer drugs. Curr Med Chem 13:3513–3527 Jung TY, Jung S, Ryu HH, Jeong YI, Jin YH, Jin SG, Kim IY, Kang SS, Kim HS (2008) Role of galectin-1 in migration and invasion of human glioblastoma multiforme cell lines. J Neurosurg 109:273–284 Koumenis C (2006) ER stress, hypoxia tolerance and tumor progression. Curr Mol Med 6:55–69 Lahm H, Andre S, Hoeflich A, Fischer JR, Sordat B, Kaltner H, Wolf E, Gabius HJ (2001) Comprehensive galectin fingerprinting in a panel of 61 human tumor cell lines by RT-PCR and its implications for diagnostic and therapeutic procedures. J Cancer Res Clin Oncol 127:375–386 Le QT, Shi G, Cao H, Nelson DW, Wang Y, Chen EY, Zhao S, Kong C, Richardson D, O’Byrne KJ, Giaccia AJ, Koong AC (2005) Galectin-1: a link between tumor hypoxia and tumor immune privilege. J Clin Oncol 23:8932–8941 Leffler H, Carlsson S, Hedlund M, Qian Y, Poirier F (2004) Introduction to galectins. Glycoconj J 19:433–440 Lefranc F, Brotchi J, Kiss R (2005) Possible future issues in the treatment of glioblastomas: special emphasis on cell migration and the resistance of migrating glioblastoma cells to apoptosis. J Clin Oncol 23:2411–2422 Lefranc F, Sadeghi N, Camby I, Metens T, Dewitte O, Kiss R (2006) Present and potential future issues in glioblastoma treatment. Expert Rev Anticancer Ther 6:719–732 Le Mercier M, Fortin S, Mathieu V, Kiss R, Lefranc F (2010) Galectins and gliomas. Brain Pathol 20:17–27 Le Mercier M, Fortin S, Mathieu V, Roland I, Spiegl-Kreinecker S, Haibe-Kains B (2009) Galectin-1 pro-angiogenic and promigratory effects in the Hs683 oligodendroglioma model are partly mediated through the control of BEX2 expression. Neoplasia 11:485–496 Le Mercier M, Lefranc F, Mijatovic T, Debeir O, Haibe-Kains B, Bontempi G, Decaestecker C, Kiss R, Mathieu V (2008a) Evidence of galectin-1 involvement in glioma chemoresistance. Toxicol Appl Pharmacol 229:172–183 Le Mercier M, Mathieu V, Haibe-Kains B, Bontempi G, Mijatovic T, Decaestecker C, Kiss R, Lefranc F (2008b) Knocking down galectin 1 in human hs683 glioblastoma cells impairs both angiogenesis and endoplasmic reticulum stress responses. J Neuropathol Exp Neurol 67:456–469 Liu FT, Rabinovich GA (2005) Galectins as modulators of tumour progression. Nat Rev Cancer 5:29–41 Louis DN, Ohgaki H, Wiestler OD, Cavenee WK (2007) WHO classification of tumours of the central nervous system. International Agency for Research on Cancer (IARC), Lyon Mathieu V, Le Mercier M, De Neve N, Sauvage S, Gras T, Roland I, Lefranc F, Kiss R (2007) Galectin-1 knock-

267 down increases sensitivity to temozolomide in a B16F10 mouse metastatic melanoma model. J Invest Dermatol 127: 2399–2410 Moiseeva EP, Spring EL, Baron JH, de Bono DP (1999) Galectin 1 modulates attachment, spreading and migration of cultured vascular smooth muscle cells via interactions with cellular receptors and components of extracellular matrix. J Vasc Res 36:47–58 Perillo NL, Marcus ME, Baum LG (1998) Galectins: versatile modulators of cell adhesion, cell proliferation, and cell death. J Mol Med 76:402–412 Puchades M, Nilsson CL, Emmett MR, Aldape KD, Ji Y, Lang FF, Liu TJ, Conrad CA (2007) Proteomic investigation of glioblastoma cell lines treated with wild-type p53 and cytotoxic chemotherapy demonstrates an association between galectin-1 and p53 expression. J Proteome Res 6:869–875 Rabinovich GA (2005) Galectin-1 as a potential cancer target. Br J Cancer 92:1188–1192 Rorive S, Belot N, Decaestecker C, Lefranc F, Gordower L, Micik S, Maurage CA, Kaltner H, Ruchoux MM, Danguy A, Gabius HJ, Salmon I, Kiss R, Camby I (2001) Galectin-1 is highly expressed in human gliomas with relevance for modulation of invasion of tumor astrocytes into the brain parenchyma. Glia 33:241–255 Stillman BN, Mischel PS, Baum LG (2005) New roles for galectins in brain tumors–from prognostic markers to therapeutic targets. Brain Pathol 15:124–132 Strik HM, Schmidt K, Lingor P, Tonges L, Kugler W, Nitsche M, Rabinovich GA, Bähr M (2007) Galectin-1 expression in human glioma cells: modulation by ionizing radiation and effects on tumor cell proliferation and migration. Oncol Rep 18:483–488 Stupp R, Mayer M, Kann R, Weder W, Zouhair A, Betticher DC, Roth AD, Stahel RA, Majno SB, Peters S, Jost L, Furrer M, Thierstein S, Schmid RA, Hsu-Schmitz SF, Mirimanoff RO, Ris HB, Pless M (2009) Neoadjuvant chemotherapy and radiotherapy followed by surgery in selected patients with stage IIIB non-small-cell lung cancer: a multicentre phase II trial. Lancet Oncol 10:785–793 Thijssen VL, Postel R, Brandwijk RJ, Dings RP, Nesmelova I, Satijn S, Verhofstad N, Nakabeppu Y, Baum LG, Bakkers J, Mayo KH, Poirier F, Griffioen AW (2006) Galectin-1 is essential in tumor angiogenesis and is a target for antiangiogenesis therapy. Proc Natl Acad Sci USA 103:15975–15980 Trédan O, Galmarini CM, Patel K, Tannock IF (2007) Drug resistance and the solid tumor microenvironment. J Natl Cancer Inst 99:1441–1454 Yamaoka K, Mishima K, Nagashima Y, Asai A, Sanai Y, Kirino T (2000) Expression of galectin-1 mRNA correlates with the malignant potential of human gliomas and expression of antisense galectin-1 inhibits the growth of 9 glioma cells. J Neurosci Res 59:722–730 Zhou J, Schmid T, Schnitzer S, Brune B (2006) Tumor hypoxia and cancer progression. Cancer Lett 237:10–21

Chapter 28

Glioma-Initiating Cells: Interferon Treatment Atsushi Natsume, Masasuke Ohno, Kanako Yuki, Kazuya Motomura, and Toshihiko Wakabayashi

Abstract Type I interferons (IFNs) are cytokines that exhibit immunomodulatory, cell differentiative, antiangiogenic, and antiproliferative effects against various neoplasms, particularly glial tumors, by classically activating the Janus Kinase (JAK)/Signal Transducers and Activators of Transcription (STAT) pathways. In contrast to other chemotherapeutic agents, responses to IFN are slow and gradual, often requiring years to develop. However, such responses can be durable. Emerging evidence suggests that the activity of IFNs is directed principally at small populations of cancer stem cells. Glioma stem cells are reported to be resistant to a wide variety of chemotherapeutic agents and possess the remarkable ability to recover from cytotoxic therapy. Interestingly, IFN-β elicits significant antiproliferative effects on glioma stem cells although such effects are not elicited by temozolomide. This cytokine induces terminal differentiation of glioma stem cells to the oligodendroglial cell lineage and downregulates an anti-apoptotic microRNA, miR-21, which is overexpressed in glioma stem cells. IFN-mediated STAT3 phosphorylation may play a crucial role in these events. Previous studies have indicated that chemokines such as bone morphogenetic protein (BMP) and leukemia inhibitory factor/ciliary neurotrophic factor (LIF/CTNF) induce STAT3-mediated differentiation of glioma stem cells. The results of

A. Natsume () Department of Neurosurgery, Nagoya University School of Medicine, 65 Tsurumai-cho, Showa-ku, Nagoya 466-8550, Japan e-mail: [email protected]

these studies provide a new strategy for therapeutic approaches that can induce glioma stem cells to undergo terminal differentiation. Keywords INFs · Cytokine · STAT pathways · Glioma stem cells · microRNA · Phosphorylation

Introduction Type I interferons (IFNs), including IFN-α and IFN-β, are cytokines that exhibit immunomodulatory, cell differentiative, antiangiogenic, and antiproliferative effects against various neoplasms, particularly glial tumors, by classically activating the Janus Kinase (JAK)/Signal Transducers and Activators of Transcription (STAT) pathways (Borden et al., 2007). One of the type I IFNs, IFN-β has multifaceted functions related to antitumor activity, such as cytostatic effects, participation in the differentiation of cytotoxic T lymphocytes and potentiation of their antitumor immunological responses, and ability to act as a drug sensitizer to enhance toxicity against various malignant neoplasms. Combination therapy with IFN-β and nitrosourea has been particularly useful in the treatment of malignant gliomas. We previously reported that IFN-β markedly enhances chemosensitivity to temozolomide, an alkylating agent (Natsume et al., 2005). This revealed that a major mechanism by which IFN-β enhances chemosensitivity is via the downregulation of O6-methylguanine-DNA methyl-transferase (MGMT) transcription through augmentation with TP53. This effect was also confirmed in an experimental animal model (Natsume et al., 2008). In this

M.A. Hayat (ed.), Tumors of the Central Nervous System, Volume 2, DOI 10.1007/978-94-007-0618-7_28, © Springer Science+Business Media B.V. 2011

269

270

chapter, we show that IFN-β elicits a remarkable antiproliferative effect on glioma stem cells (GSCs) and induces terminal differentiation of GSCs to the oligodendroglial cell lineage by the activation of STAT3. Further, we focus on microRNA regulation by type I IFNs in GSCs.

Cancer Stem Cells and Type I IFNs Cancers harbor small cell populations that have the potential to sustain growth and initiate tumorigenesis. These cells, which are known as cancer stem cells (CSCs) or cancer initiating cells, have been identified in leukemia, multiple myeloma, breast cancer, and glioma. In solid tumors, CSCs share many properties with normal stem cells, including self-renewal and multi-potency, and furthermore are known to initiate the growth of tumors. A long-recognized paradox of response and survival in cancer therapies is that although most cancer patients respond to therapy, few are completely cured. The objective clinical response to treatment may not predict substantial improvements in overall survival. Huff et al. (2006) coined the term “dandelion phenomenon” to describe this paradox in cancer. The pattern of the clinical responses displayed by patients with chronic myeloid leukemia (CML) toward imatinib and IFN are quite different. The rapid responses induced by imatinib are probably a consequence of its impressive activity against differentiated CML cells that make up the bulk of the leukemia. However these early effective responses often do not last. The evidence could be explained by the development of CML stem cell resistance to imatinib. Conversely, responses to IFN are slow and gradual, often requiring years to develop, but can be durable; this is consistent with the data suggesting that IFN’s activity is directed principally at small populations of CML stem cells. The pattern of imatinib activity may be analogous to cutting a dandelion off at ground level and leaving the unseen root behind, while IFN mimics specific attack of the root of the dandelion. This response–survival paradox also applies to other malignancies. In ovarian cancer, Moserle et al. (2008) demonstrated that IFN exerts marked antiproliferative effects on side population cells with higher proliferation rates. However, the mechanism of action underlying this effect remains undefined.

A. Natsume et al.

Establishment of Glioma Stem Cells Glioblastoma multiforme (GBM) is the most lethal primary glioma. Patients diagnosed with this type of glioma typically have a median survival of less than 12 months, despite various therapies, including surgical resection, radiotherapy, and chemotherapy. For many years, gliomas such as GBM have been considered to represent the bulk of heterogeneous tumors that consist of differentiated cells and undifferentiated cells with the capability of self-renewal and partial differentiation (Lee et al., 2006). Therefore, failure to eradicate glioma cells may be responsible for the small fraction of undifferentiated tumor cells that re-grow. Nestin is an intermediate filament cytoskeletal protein found in neuroepithelial stem cells and progenitor cells (Reynolds and Weiss, 1992). It has been shown that glial-derived neoplasms also express nestin, and the level of expression is higher in high-grade gliomas compared to low-grade gliomas (Dahlstrand et al., 1992). Using the neurosphere assay, Ignatova et al. (2002) were able to isolate nestin-expressing cells capable of forming clones that exhibit intraclonal heterogeneity in the expression of neural lineage-specific proteins. Subsequently, when Singh et al. (2003) successfully cultured glioma cells in serum-free medium containing stem cell growth factors bFGF and EGF, it was observed that a small percentage of all glioma cells ranging from 0.3 to 25.1%, depending on the aggressiveness of the tumor, were capable of self-replication and formed non-adherent neurospheres and maintained tumor culture over time via multiple passages. These self-renewing and tumor culture-maintaining cells not only stained positive for the undifferentiated neural stem cell marker nestin, they also stained positive for CD133, a hematopoietic stem cell marker also present on normal human neural stem cells. However, these cells were found to lack the expression of β tubulin III and glial fibrillary acidic protein (GFAP), which can be used as markers for differentiated neuronal lineage. In stark contrast to this small group of CD133+ cells, the majority of glioma cells were found to be CD133- and incapable of forming self-sustaining neurospheres. As a general observation, the fraction of CD133+ GSCs from aggressive tumors exhibited an increased rate of self renewal relative to less aggressive tumors. Under culture conditions that promote differentiation, GSCs lose the expression of primitive markers like

28 Glioma-Initiating Cells: Interferon Treatment

CD133 and nestin and instead express differentiated markers for the cell of origin. When serially transplanted into the brain of a NOD-SCID (non-obese diabetic, severe combined immunodeficient) mouse, these GSCs were able to produce exact phenocopies of the original tumor with all the histopathological features and cell surface markers of the original tumor. Immunohistological staining of these GBM xenografts that showed differential staining for CD133 and GFAP underlines the fact that GSCs can differentiate into mature progeny. Even though the presence of GSC can clearly account for the inherent heterogeneous nature of gliomas, cellular and genetic analysis of GSCs showed that these cells were genetically transformed with enhanced self-renewal properties and possessed an abnormal karyotype, which is not only limited to CD133+ cells, but also present in both CD133+ and CD133- cells. This suggests that all cancer cells were clonally derived. Even though consensus has not yet been reached on the exact mechanism of gliomagenesis, our new found understanding of the presence of functional hierarchy in glioma indicates that it will be important to investigate the slow growing mutated GSCs in order to gain an understanding of the mechanisms of treatment failure. GSCs are reported to be resistant to a wide variety of chemotherapeutic agents and possess the remarkable ability to recover from cytotoxic therapy (Eramo et al., 2006). Kang et al. (2007) reported that a small population of multipotent CD 133+ cancer cells can survive and proliferate upon exposure of GBM cells to a lethal dose of carmustine (BCNU). When transplanted into a SCID mouse brain, the original tumor was reproduced. A significantly higher level of CD133+ cells was reported to be present in previously treated GBM when compared with newly diagnosed GBM (Liu et al., 2006). The gene profile of CD133+ cells exhibits a high level of expression of anti-apoptotic genes and chemotherapy resistance genes such as BCRP1, and MGMT (Liu et al., 2006), rendering these cells resistant to many commonly used chemotherapeutic agents, including temozolomide, carboplatin, paclitaxel (Taxol), and etoposide (VP16). In the same tumor, examples of genes such as multi drug resistance-associated proteins 1 and 3 were found to be markedly elevated in GSCs when compared to non-GSCs (Salmaggi et al., 2006). This emphasizes their role in chemoresistance.

271

GSCs not only play a crucial role in chemoresistance, but are vital with respect to the failure of radiation therapy since tumors surviving radiotherapy are found to be enriched in CSCs. In a study conducted by Bao et al. (2006), irradiation of an in vivo glioma xenograft led to a 3–5-fold increase in the CD133+ cell population relative to untreated xenografts. This suggests that irradiation leads to the enrichment of CD133+ cells in the tumor and subsequent formation of more aggressive tumors with decreased latency after serial transplantation. Given the pattern of treatment failure observed with current standard therapy, an alternative strategy involving selectively targeting this functionally distinct chemo- and radiation-resistant small group of GSCs instead of the bulk of the tumor might provide better success in treating this deadly disease.

Terminal Differentiation of GSC by IFN STAT3 activation is crucial for stem cell function, differentiation, and survival. Recent studies demonstrated that bone morphogenetic protein (BMP) and leukemia inhibitory factor/ciliary neurotrophic factor (LIF/CTNF) induced STAT3-mediated differentiation of neural and GSCs (Lee et al., 2008). These studies indicate the potential for therapeutic approaches that can induce GSCs to undergo terminal differentiation. Our previous study showed that STAT3 expression is strongly upregulated in GSCs, and that IFN-β phosphorylates a tyrosine residue of STAT3. IFN-β treatment elicits a remarkable antiproliferative effect and enhances the expression of oligodendroglial markers, including oligodendrocyte-specific protein (OSP), galactosylceraminidase (GalC), and myelin basic protein (MBP), on GSCs. However, the expression of these markers was inhibited by a peptide that specifically inhibits STAT3 (Fig. 28.1). Furthermore, IFNβ treatment did not induce the expression of OSP, GalC, and MBP in human neural stem cells. In contrast, while the expression of other neural lineage markers (i.e., GFAP and neuron-specific class III βtubulin) was unaffected, the expression of nestin and CD133 was reduced in the GSCs after IFN-β treatment. IFN-β-treated GSCs were unable to form tumors in NOD/SCID mice (Yuki et al., 2009). Therefore, IFN may represent a potential therapeutic agent for inducing terminal differentiation of GSCs.

272

A. Natsume et al.

Fig. 28.1 IFN-β phosphorylates a tyrosine residue of STAT3. IFN-β treatment elicits a remarkable antiproliferative effect. IFN-β induces glioma stem cells to undergo terminal differentiation to mature oligodendrocytes that

express oligodendrocyte-specific protein (OSP), galactosylceraminidase (GalC), and myelin basic protein (MBP), and inhibits gliomagenesis

Regulation of MicroRNAs by IFN in GSCs

its translation (Esquela-Kerscher and Slack, 2006). Emerging evidence suggests that miRNAs are involved in crucial biological processes, including development, differentiation, apoptosis, and proliferation of mammalian cells. In humans, miRNAs have been proposed to contribute to oncogenesis because they possess multifaceted functions either as tumor suppressors or as oncogenes (Schickel et al., 2008). RNA interference induces a multitude of responses in addition to the knockdown of a gene. This is best understood in the context of an antiviral immune response. In particular, double-stranded RNA, a nucleic acid associated with viral replication, is involved in numerous interactions contributing to the induction, activation, and regulation of antiviral mechanisms. One particularly intriguing function of double-stranded RNA is to stimulate important protective responses such as the activation of Dicer-related antiviral pathways, induction of type I IFN (IFN-α/β), and stimulation of double-stranded RNA-activated protein kinase and oligoadenylate synthase (Karpala et al., 2005).

MicroRNAs (miRNAs) are small noncoding RNAs consisting of 20–22 nucleotides that participate in the posttranslational regulation of gene expression by means of RNA interference (RNAi). The miRNA genes are transcribed by RNA polymerase II in the nucleus to form large pri-miRNA transcripts. These pri-miRNA transcripts are processed by the Drosha enzyme to release the pre-miRNA precursor product, which is less than 70 nucleotides in length. After the pre-miRNA is transported into the cytoplasm, another enzyme known as Dicer processes the intermediate to generate a mature 22nucleotide miRNA. This mature miRNA is integrated into the RNA-induced silencing complex and forms double-stranded RNA with complementary mRNAs. Depending on the degree of homology between the miRNA and the mRNA, the RNA-induced silencing complex could inhibit mRNA function by either promoting its cleavage or by inhibiting

28 Glioma-Initiating Cells: Interferon Treatment

IFN-α/β regulates the levels of crucial mediators of the antiviral response, such as protein kinase R, the 2 –5 oligoadenylate synthase/RNase L system, adenosine deaminase ADR1, and Mx GTPase (Katze et al., 2002). Thus, RNA interference might be involved in the IFN-mediated antiviral response. It was recently reported that the levels of liver-specific miRNA, miR-122, and several other miRNAs are regulated by IFN-β in human hepatoma cells, and that IFN-β rapidly modulates the expression of miRNAs that target the hepatitis C virus genomic RNA, thereby inhibiting viral replication (Pedersen et al., 2007). In addition to its ability to interfere with viral replication, IFN-β is also known for its antiproliferative effects in a variety of neoplasms such as hepatocellular carcinoma, chronic myeloblastic leukemia, melanoma, renal cancer, and glioma (Borden et al., 2007). Therefore, the possibility is recognized that IFN-β might induce or downregulate cellular miRNAs in human neoplasms and thereby use the RNA interference system in its action against tumor progression. We tested whether IFN-β can alter the expression of cellular miRNAs in human glioma cells by using the data obtained from a genome-wide microarray (Ohno et al., 2009). On the basis of the initial screening, we found that the expression levels of several miRNAs were increased (miR-187 and miR-194) or attenuated (miR-100, miR-21, and let-7 family miRNAs) in response to IFN-β treatment. Of the miRNAs regulated by IFN-β, we focused on miR-21 because it is one of the best known miRNAs associated with tumorigenesis and progression in gliomas. miR-21 also modulates tumorigenesis through the regulation of genes such as bcl-2, PTEN, tropomyosin-1, and PDCD4 (Chen et al., 2008; Meng et al., 2007; Zhu et al., 2007). Previously, miR-21 was suggested to be aberrantly expressed and was recognized as one of the major anti-apoptotic factors in malignant gliomas (Chan et al., 2005). We demonstrated that the overexpression of miR-21 occurs in a surgical specimen of glioblastoma by performing an in situ hybridization (Fig. 28.2a–c). We also compared the miR-21 expression levels in glioma cell lines, GSCs, and normal brain tissue. The miR-21 was found to be overexpressed in glioma cells relative to normal brain cells. Notably, the expression level of miR-21 was found to be greater in GSCs than in the glioma cell lines (Fig. 28.2d). This finding may indicate that miR-21 plays a crucial role in the initiation and progression of glioma.

273

Investigations using quantitative real timepolymerase chain reaction (qRT-PCR) revealed that IFN-β downregulates miR-21 in cultured glioma cell lines and GSCs, and that the systemic delivery of IFN-β reduces the level of miR-21 in intracranial GSC xenografts in mice. In the time course experiments, the IFN-β treatment showed a relatively fast response in reducing miR-21 levels, suggesting that the negative regulation of miR-21 might be mediated directly by IFN-β, for example, through the phosphorylation of JAK/STAT. Our finding that IFN-β also suppresses the expression of pri-miR-21 and pre-miR-21 suggests that it regulates miR-21 transcription. The putative regulatory region of the miR-21 gene is located within an intron of the overlapping transmembrane protein 49 gene (TMEM49). This regulatory region contains 2 consensus STAT3-binding sites located approximately 800 bp upstream from the transcription start site. We used a luciferase reporter gene system with a fused full-length pri-miR-21 promoter, including the STAT3 binding site, to demonstrate that IFN-β inactivates STAT3-mediated miR-21 promoter activity and that the transcription activity could be recovered by using a STAT3 inhibitor. These data suggest that miR-21 expression is negatively regulated by STAT3 activation induced by IFN-β. In conclusion, the downregulation of miR-21 in response to IFN-β treatment contributes to the antitumor effects of this cytokine in GSCs. Furthermore, miR-21 expression is negatively regulated by STAT3 activation (Fig. 28.3). These results highlight the importance of understanding the transcriptional regulation of miRNAs, which would be involved in oncogenesis of gliomas.

Discussion Here, we show that STAT3 activation is crucial for oligodendrogenesis and miR-21 downregulation in GSCs. However, the role of STAT3 activation is debatable because its overactivation has been reported to be oncogenic in other neoplasms (Frank, 2007). Loffler et al. (2007) demonstrated that IL-6dependent STAT3 activates the transcription of miR-21 in multiple myeloma cells. While IL-6 induces proliferation of myeloma cells, IFN-β reduces the growth of glioma cells or induces apoptosis in these cells.

274

Fig. 28.2 (a) miR-21 overexpression in glioma-initiating cells (GICs). miR-21–specific probe was hybridized in situ with glioblastoma tissue obtained from a surgical specimen. The miR-21–specific probe clearly stains the glioblastoma tissue but doesnot stain the normal cortextissue. (b) Tumor cells express significant amounts of miR-21, as observed under high

A. Natsume et al.

magnification, (c) whereas nontumorous tissue does not express miR-21. (d) qRT-PCR shows that miR-21 is overexpressed to a great extent in glioma cells than in normal brain cells. Notably, the amount of miR-21 is greater in GICs than in the glioma cell lines. Columns, mean; bars, SD (normal brain expressed as 1)

28 Glioma-Initiating Cells: Interferon Treatment

275

Fig. 28.3 IFN-β inactivates STAT3-mediated miR-21 promoter activity. The miRNA genes are transcribed by RNA polymerase II in the nucleus to form large pri-miRNA transcripts. These pri-miRNA transcripts are processed by the Drosha enzyme to release the pre-miRNA precursor product, which is less than 70 nucleotides in length. After the pre-miRNA is transported into

the cytoplasm, another enzyme known as Dicer processes the intermediate to generate a mature 22-nucleotide miRNA. This mature miRNA is integrated into the RNA-induced silencing complex and forms double-stranded RNA with complementary mRNAs of genes such as bcl-2, PTEN, tropomyosin-1, and PDCD4

A possible explanation for this seemingly paradoxical role of STAT3 activation is that the STAT pathway is context-dependent and that various intracellular and/or environmental cues play a pivotal role in determining the outcome of the activation of the pathway. This discrepancy may arise from differences in cytokine stimulus and cell type (Loffler et al., 2007). An unresolved question pertains to the identity of the regulators of oligodendrogenesis of GSCs. Oligodendroglial lineage arises from undifferentiated progenitor cells, which are predominately found in the subventricular zone. These progenitor cells mature into myelinating oligodendrocytes as a result of interactions between BMP, sonic hedgehog, and proteins of the Notch signaling pathways (Nicolay et al., 2007). IFN may activate the BMP/Smad pathway and Notch signaling. There is a strong evidence suggesting that thyroid hormones (THs) directly regulate the differentiation and maturation of oligodendroglial cells. Triiodothyronine (T3) is an active hormone that

regulates gene expression by binding to specific intracellular TH receptors. Our group previously showed that 95% of glioma cells express a member of the thyroid/steroid hormone receptor superfamily – peroxisome proliferative-activated receptor gamma (PPARγ), which activates the transcription of target genes after forming heterodimers with retinoid X receptors (RXR). PPARs are known to be responsible for deciding the fate of specific neural stem cells. It has been reported that PPARs regulate the proliferation, migration, and differentiation of neural stem cells via STAT3 activation. Although further confirmative studies are warranted in this regard, we were able to show that treatment of GSCs with IFN-β causes the release of T3 and that this is inhibited by the STAT3 inhibitor. Treatment with exogenous T3 results in the formation of MBP-positive mature oligodendrocytes displaying a multibranched morphology. This suggests that GSCs may possess neuroendocrinal properties and differentiate into mature oligodendrocytes.

276

References Bao S, Wu Q, McLendon RE, Hao Y, Shi Q, Hjelmeland AB, Dewhirst MW, Bigner DD, Rich JN (2006) Glioma stem cells promote radioresistance by preferential activation of the DNA damage response. Nature 444:756–760 Borden EC, Sen GC, Uze G, Silverman RH, Ransohoff RM, Foster GR, Stark GR (2007) Interferons at age 50: past, current and future impact on biomedicine. Nat Rev Drug Discov 6:975–990 Chan JA, Krichevsky AM, Kosik KS (2005) MicroRNA-21 is an antiapoptotic factor in human glioblastoma cells. Cancer Res 65:6029–6033 Chen Y, Liu W, Chao T, Zhang Y, Yan X, Gong Y, Qiang B, Yuan J, Sun M, Peng X (2008) MicroRNA-21 downregulates the expression of tumor suppressor PDCD4 in human glioblastoma cell T98G. Cancer Lett 272:197–205 Dahlstrand J, Collins VP, Lendahl U (1992) Expression of the class VI intermediate filament nestin in human central nervous system tumors. Cancer Res 52:5334–5341 Eramo A, Ricci-Vitiani L, Zeuner A, Pallini R, Lotti F, Sette G, Pilozzi E, Larocca LM, Peschle C, De Maria R (2006) Chemotherapy resistance of glioblastoma stem cells. Cell Death Diff 13:1238–1241 Esquela-Kerscher A, Slack FJ (2006) Oncomirs – microRNAs with a role in cancer. Nat Rev Cancer 6:259–269 Frank DA (2007) STAT3 as a central mediator of neoplastic cellular transformation. Cancer Lett 251:199–210 Huff CA, Matsui W, Smith BD, Jones RJ (2006) The paradox of response and survival in cancer therapeutics. Blood 107: 431–434 Ignatova TN, Kukekov VG, Laywell ED, Suslov ON, Vrionis FD, Steindler DA (2002) Human cortical glial tumors contain neural stem-like cells expressing astroglial and neuronal markers in vitro. Glia 39:193–206 Kang MK, Kang SK (2007) Tumorigenesis of chemotherapeutic drug-resistant cancer stem-like cells in brain glioma. Stem Cells Dev 16:837–847 Karpala AJ, Doran TJ, Bean AG (2005) Immune responses to dsRNA: implications for gene silencing technologies. Immunol Cell Biol 83:211–216 Katze MG, He Y, Gale M Jr (2002) Viruses and interferon: a fight for supremacy. Nat Rev Immunol 2:675–687 Lee J, Kotliarova S, Kotliarov Y, Li A, Su Q, Donin NM, Pastorino S, Purow BW, Christopher N, Zhang W, Park JK, Fine HA (2006) Tumor stem cells derived from glioblastomas cultured in bFGF and EGF more closely mirror the phenotype and genotype of primary tumors than do serumcultured cell lines. Cancer Cell 9:391–403 Lee J, Son MJ, Woolard K, Donin NM, Li A, Cheng CH, Kotliarova S, Kotliarov Y, Walling J, Ahn S, Kim M, Totonchy M, Cusack T, Ene C, Ma H, Su Q, Zenklusen JC, Zhang W, Maric D, Fine HA (2008) Epigeneticmediated dysfunction of the bone morphogenetic protein pathway inhibits differentiation of glioblastoma-initiating cells. Cancer Cell 13:69–80 Liu G, Yuan X, Zeng Z, Tunici P, Ng H, Abdulkadir IR, Lu L, Irvin D, Black KL, Yu JS (2006) Analysis of gene expression and chemoresistance of CD133+ cancer stem cells in glioblastoma. Mol Cancer 5:67

A. Natsume et al. Loffler D, Brocke-Heidrich K, Pfeifer G, Stocsits C, Hackermuller J, Kretzschmar AK, Burger R, Gramatzki M, Blumert C, Bauer K, Cvijic H, Ullmann AK, Stadler PF, Horn F (2007) Interleukin-6 dependent survival of multiple myeloma cells involves the Stat3-mediated induction of microRNA-21 through a highly conserved enhancer. Blood 110:1330–1333 Meng F, Henson R, Wehbe-Janek H, Ghoshal K, Jacob ST, Patel T (2007) MicroRNA-21 regulates expression of the PTEN tumor suppressor gene in human hepatocellular cancer. Gastroenterology 133:647–658 Moserle L, Indraccolo S, Ghisi M, Frasson C, Fortunato E, Canevari S, Miotti S, Tosello V, Zamarchi R, Corradin A, Minuzzo S, Rossi E, Basso G, Amadori A (2008) The side population of ovarian cancer cells is a primary target of IFN-alpha antitumor effects. Cancer Res 68:5658–5668 Natsume A, Ishii D, Wakabayashi T, Tsuno T, Hatano H, Mizuno M, Yoshida J (2005) IFN-beta down-regulates the expression of DNA repair gene MGMT and sensitizes resistant glioma cells to temozolomide. Cancer Res 65:7573–7579 Natsume A, Wakabayashi T, Ishii D, Maruta H, Fujii M, Shimato S, Ito M, Yoshida J (2008) A combination of IFN-beta and temozolomide in human glioma xenograft models: implication of p53-mediated MGMT downregulation. Cancer Chemother Pharmacol 61:653–659 Nicolay DJ, Doucette JR, Nazarali AJ (2007) Transcriptional control of oligodendrogenesis. Glia 55:1287–1299 Ohno M, Natsume A, Kondo Y, Iwamizu H, Motomura K, Toda H, Ito M, Kato T, Wakabayashi T (2009) The modulation of microRNAs by type I IFN through the activation of signal transducers and activators of transcription 3 in human glioma. Mol Cancer Res 7:2022–2030 Pedersen IM, Cheng G, Wieland S, Volinia S, Croce CM, Chisari FV, David M (2007) Interferon modulation of cellular microRNAs as an antiviral mechanism. Nature 449: 919–922 Reynolds BA, Weiss S (1992) Generation of neurons and astrocytes from isolated cells of the adult mammalian central nervous system. Science 255:1707–1710 Salmaggi A, Boiardi A, Gelati M, Russo A, Calatozzolo C, Ciusani E, Sciacca FL, Ottolina A, Parati EA, La Porta C, Alessandri G, Marras C, Croci D, De Rossi M (2006) Glioblastoma-derived tumorospheres identify a population of tumor stem-like cells with angiogenic potential and enhanced multidrug resistance phenotype. Glia 54:850–860 Schickel R, Boyerinas B, Park SM, Peter ME (2008) MicroRNAs: key players in the immune system, differentiation, tumorigenesis and cell death. Oncogene 27: 5959–5974 Singh SK, Clarke ID, Terasaki M, Bonn VE, Hawkins C, Squire J, Dirks PB (2003) Identification of a cancer stem cell in human brain tumors. Cancer Res 63:5821–5828 Yuki K, Natsume A, Yokoyama H, Kondo Y, Ohno M, Kato T, Chansakul P, Ito M, Kim SU, Wakabayashi T (2009) Induction of oligodendrogenesis in glioblastoma-initiating cells by IFN-mediated activation of STAT3 signaling. Cancer Lett 284:71–79 Zhu S, Si ML, Wu H, Mo YY (2007) MicroRNA-21 targets the tumor suppressor gene tropomyosin 1 (TPM1). J Biol Chem 282:14328–14336

Chapter 29

Glioblastoma: Anti-tumor Action of Natural and Synthetic Cannabinoids Aleksandra Ellert-Miklaszewska, Iwona Ciechomska, and Bozena Kaminska

Abstract The past few decades have seen renewed interest in medicinal cannabis or rather cannabinoids – active compounds derived from the Cannabis plant, as well as their endogenous counterparts and a still growing set of synthetic derivatives. One of the most extensively studied and promising applications of cannabinoids is their potential use as anti-cancer agents in malignant tumors, such as glioblastomas. The anti-tumor action of cannabinoids is mediated via the CB1 and CB2 cannabinoid receptors. Growing array of data suggest that significant alterations of a balance in the cannabinoid system between the levels of endogenous ligands and their receptors occur along with the malignant transformation in various types of cancer. Increased CB2 receptor expression has been observed in glioblastoma cells, invading microglia/macrophages and endothelial cells of the tumor blood vessels as compared to non-tumor brain samples. Thus, among various approaches to avoid CB1-receptor-mediated psychodysleptic side effects of some cannabinoids, special attention is paid to substances, which selectively stimulate the CB2 receptors, putatively overexpressed in target tumor cells. Induction of cell death by cannabinoid treatment relies on the generation of a pro-apoptotic sphingolipid ceramide and disruption of signaling pathways crucial for regulation of cellular proliferation, differentiation or apoptosis. Increased ceramide levels lead also to ER-stress and autophagy in drug treated glioblastoma

B. Kaminska () Laboratory of Transcription Regulation, Department of Cell Biology, Nencki Institute of Experimental Biology, 02-093 Warsaw, Poland e-mail: [email protected]

cells. Cannabinoids have displayed a great potency in reducing glioma tumor growth in experimental animal models without producing the generalized toxic effects unavoidable with most conventional chemotherapeutic drugs. Apparently their effectiveness in vivo has been attributed to several mechanisms of action. Cannabinoids have recently emerged as compounds that beyond inhibition of tumor cell proliferation and survival impair tumor angiogenesis, invasiveness and even gliomagenesis. The good safety profile observed in a pilot clinical trial, together with remarkable antitumor effects reported in preclinical studies may set the basis for further research aimed at better evaluation of the potential anti-cancer activity of cannabinoids. A unique mechanism of cannabinoid action among standard oncology remedies justifies further research on their anti-tumoral properties either alone or in combined therapies. Keywords Glioblastoma · Cannabinoids · Anti-cancer agents · Receptors · Ligands · Autophagy

Introduction Preparations from Cannabis sativa, the hemp plant, have been used for centuries for both medicinal and recreational purposes (Howlett et al., 2002; Mackie, 2006). Isolation of the active components of the plant, called cannabinoids, in 1960s, as much as subsequent cloning of cannabinoid receptors, discovery of their endogenous ligands and development of synthetic cannabinoids contributed to an intensive burst of cannabinoid research. Along with our expanding comprehension of mechanisms of cannabinoids action,

M.A. Hayat (ed.), Tumors of the Central Nervous System, Volume 2, DOI 10.1007/978-94-007-0618-7_29, © Springer Science+Business Media B.V. 2011

277

278

targeting cannabinoid signaling for therapeutic purposes has inevitably emerged as an interesting area of scientific and clinical investigations. One of the most extensively studied applications of cannabinoids is their potential use as anti-cancer agents. Anti-proliferative effects of cannabinoids have been reported in various cultured cancer cells, including neural, breast, prostate, skin, thyroid cancer cells and leukemia cells. Several studies demonstrated anti-tumor activity of cannabinoids in animal models (Guzman, 2003). Since the first publication by Sanchez and co-workers in 1998 providing evidence that cannabinoids can be an effective tool against glioma cells, growing interest of several groups including ours has been focused on therapeutic potential of cannabinoids in glial tumors. Current scientific data relevant to the use of cannabinoids in treatments of glioblastomas are reviewed here including both in vitro and in vivo results, proposed mechanisms of action, reports from limited human studies and prospects for patients therapy in clinics.

Cannabinoid System in Health and Brain Tumors Cannabinoids are a group of structurally heterogenous but pharmacologically related compounds classified into three subtypes. Plant-derived cannabinoids (phytocannabinoids) are uniquely found in the cannabis plant. (–)-trans-9 -tetrahydrocannabinol (9 -THC) is the most potent and abundant out of approximately 70 identified phytocannabinoids. Endogenous cannabinoids (also known as endocannabinoids) such anandamide and 2-arachidonoylglycerol are produced by mammals as rapidly inactivated lipid mediators, which levels are strictly controlled by a transporter system and hydrolyzing enzymes. Various modifications of the chemical structure of natural cannabinoids based on structure-activity relationship studies have led to generation of a still growing set of synthetic cannabinoids (Fig. 29.1). Cannabinoids elicit a wide range of central and peripheral effects, which are mediated mostly through cannabinoid receptors (Howlett et al., 2002). There are two types of specific Gi/o -protein-coupled receptors cloned so far, called CB1 and CB2, although an existence of additional cannabinoid-binding receptors

A. Ellert-Miklaszewska et al.

has been suggested (Howlett et al., 2002; Stella, 2004). CB1 and CB2 differ in their predicted amino acid sequence, tissue distribution, physiological role and signaling mechanisms. CB1 is abundantly expressed in the central nervous system and peripheral nerve terminals, where it mediates inhibition of neurotransmitter release, but it is also present in some extra-neural sites, such as testis, uterus, vascular endothelium, eye, spleen, and tonsils (Howlett et al., 2002). By contrast, the CB2 receptor is predominantly expressed in cells and organs of the immune system (Howlett et al., 2002). The role of peripheral CB2 receptor activation under baseline physiologic conditions is still not well defined. It is suggested to be involved in B-cell differentiation and modulation of immune response. Increased levels of CB2 are reported in tissues during development, inflammation, injury and cancer, revealing a critical role for the CB2 receptor in regulating these processes (Howlett et al., 2002). The CB2 receptor was believed to be absent from healthy brain, however, its expression has been detected recently in microglia – brain macrophages (Stella, 2004; Gong et al., 2006), as well as in a small subpopulation of neurons (Van Sickle et al., 2005; Gong et al., 2006). The function of CB2 receptor in the brain is completely unknown. Nevertheless, experiments performed in animals show that CB2-selective agonists appear to be free from widespread psychoactive effects attributed to activation of the CB1 receptor (Guzman, 2003; Valenzano et al., 2005). Growing array of data suggest that significant alterations of a balance in the cannabinoid system between the levels of endogenous ligands and their receptors occur along with the malignant transformation in various types of cancer. While non-transformed astrocytes were shown to express only the CB1 cannabinoid receptor, both types of functional cannabinoid receptors have been found in several established human glioblastoma cell lines, as well as in primary cultures derived from the most malignant brain tumor, glioblastoma multiforme (GBM) (Galve-Roperh et al., 2000; Sanchez et al., 2001; Howlett et al., 2002) (and our unpublished data). Immunohistochemical analysis of human low and high grade glioma surgical specimens revealed increased CB2 receptor expression in tumor cells, invading microglia/macrophages and endothelial cells of the tumor blood vessels as compared to non-tumor brain samples (Sanchez et al., 2001; EllertMiklaszewska et al., 2007; Schley et al., 2009). We

29 Glioblastoma: Anti-tumor Action of Natural Synthetic Cannabinoids

279

Fig. 29.1 Chemical structures of plant-derived 9 -THC ((-)-trans-9 -tetrahydrocannabinol), endocannabinoid anandamide (N-arachidonoylethanolamine) and synthetic cannabinoids JWH133, selective for non-psychoactive CB2 receptor ((6aR,10aR)-3-(1,1-Dimethylbutyl)-6a,7,10,10a-tetrahydro-6,6,

9-trimethyl-6H-dibenzo[b,d]pyran), and WIN55,212-2, a CB1/ CB2 receptor agonist ((R)-(+)-[2,3-Dihydro-5-methyl-3-(4morpholinylmethyl)pyrrolo[1,2,3-de]-1,4-benzoxazin-6-yl]-1naphthalenylmethanone)

detected the presence of CB2 receptors in all analyzed biopsies of astrocytomas and glioblastomas. The proportion of malignant tumors expressing high levels of CB2 (10 out of 16, 62.5%) was over 2-fold higher than that seen in the tumors of lower grade (7 out of 29, 24%) (Ellert-Miklaszewska et al., 2007). Thus, the extent of CB2 expression correlated with tumor malignancy. Moreover, as observed by Sanchez et al. (2001) CB2 receptor immunoreactivity markedly prevailed over detected CB1 receptor levels in grade IV astrocytomas (Sanchez et al., 2001). The levels of CB1 receptor expression in tumor and tumor-associated endothelial cells were not significantly different from the control tissue and showed no dependence on tumor grade (Sanchez et al., 2001; Schley et al., 2009). As brain tumors constitute the second most common malignancy in children and the prevalence of histological types of brain tumors vary significantly between the adult and pediatric populations, we compared the expression of the CB2 receptor in paraffin-embedded sections from primary brain tumors of adult and pediatric patients (Ellert-Miklaszewska et al., 2007). Most glioblastomas obtained from children expressed very high levels of CB2 receptors. Interestingly,

some benign pediatric astrocytic tumors, such as subependymal giant cell astrocytoma (SEGA), which may occasionally cause mortality owing to progressive growth in some patients, also displayed high CB2imunoreactivity. The high levels of CB2 expression would predestine astrocytic tumors to be vulnerable to cannabinoid treatment. Significant CB2 immunostaining was also found in limited cases of other histological types of brain tumors, primarily in high grade ependymomas and meningiomas (Ellert-Miklaszewska et al., 2007). However, we observed that the CB2 staining intensity was much higher in astrocytomas than in oligodendrogliomas, ependymomas or meningiomas of the same grade. It is worth to mention that all examined cases of embryonal tumors (medulloblastoma and S-PNET), the most frequently diagnosed malignant brain tumors in childhood, showed no or trace CB2 immunoreactivity. Our results suggest that the level of CB2 receptor expression correlates primarily with the histopathological origin of the brain tumor cells and their differentiation state, reflected in the tumor grade. A link between CB2 expression and malignancy grade of the tumor has been reported also in prostate,

280

breast and pancreatic cancer, and the level of CB2 expression in transformed cells was higher than in the respective normal tissue (Sarfaraz et al., 2005; Caffarel et al., 2006; Carracedo et al., 2006a). In examined tumors, CB2 receptors were usually located in the areas of intense tissue proliferation and invading cells. The enhancement of cannabinoid receptor expression in malignant versus healthy tissues, might suggest a possible role of the endocannabinoid system in the tonic suppression of cell divisions and cancer growth. This hypothesis is partly supported by the finding of increased levels of anandamide and decreased levels of one of endocannabinoid metabolizing enzymes in human glioblastoma compared to human non-tumor brain tissue (Petersen et al., 2005).

Anti-tumoral Action of Cannabinoids in Glioma Cells and In Vivo Tumor Models The pioneer study to show that cannabinoids had antitumor effect and prolonged the life of mice bearing Lewis lung adenocarcinoma was reported by Munson et al. in 1975 (Munson et al., 1975). Further research in this area was not performed until late 1990s, when an interest in the role of cannabinoids in cancer therapy was renewed. Since then, plant-derived, synthetic and endogenous cannabinoids have been successfully used to block proliferation and invasive potential of various cancer types (glioma, lymphoma, lung, thyroid, skin, pancreas, breast and prostate carcinoma) in both in vitro and in vivo models (Guzman, 2003). The research on therapeutic potential of cannabinoids in cancer treatment is the most advanced in gliomas. Programmed cell death of glioma cells after cannabinoid treatment was first described by Manuel Guzman and his co-workers (Sanchez et al., 1998). They showed that 9 -THC is able to inhibit growth of rat C6 glioma cells in vitro and induce cell death with features typical for apoptosis (Sanchez et al., 1998). We reported an apoptotic death triggered by a mixed CB1/CB2 synthetic agonist WIN 55,212-2 in rat glioma cells (Ellert-Miklaszewska et al., 2005). Recent studies (Gomez del Pulgar et al., 2002; Salazar et al., 2009) and our own observations show effectiveness of 9 -THC, WIN55,212-2 and a CB2-selective synthetic cannabinoid to induce apoptosis of cultured human glioblastoma cells and tumor-derived primary cultures.

A. Ellert-Miklaszewska et al.

The anti-tumor action of cannabinoids, mediated via the CB1 or CB2 cannabinoid receptors, has been also proved in vivo revealing a significant regression of malignant gliomas in cannabinoid-treated animals (Galve-Roperh et al., 2000; Sanchez et al., 2001; Massi et al., 2004; Duntsch et al., 2006). Local administration of 9 -THC or WIN55,212-2 reduced the size of tumors generated by intracranial inoculation of C6 glioma cells in rats, leading to complete eradication of gliomas and increased survival in one third of the treated rats (Galve-Roperh et al., 2000). There has been no recurrence in any of the surviving rats as monitored periodically by MRI scan. Studies performed in mouse xenograft models with intratumoral and intraperitoneal drug administration demonstrated that non-psychoactive phytocannabinoid cannabidiol (Massi et al., 2004), a CB2selective agonist JWH133 (Sanchez et al., 2001) or a novel cannabinoid KM-233 (Duntsch et al., 2006) blocked the proliferation of human astrocytoma cells implanted subcutaneously in the flank of immune-deficient mice. Our preliminary data suggest that systemic cannabinoid administration can effectively hamper intracranial tumor growth in rats (unpublished).

Mechanism of Cannabinoids Pro-apoptotic Action – Inhibition of Pro-survival Pathways Several events and signal transduction pathways triggered mostly by stimulation of the CB1 and CB2 receptors have already been described to participate in the cannabinoid-induced apoptosis, a programmed cell death process (Guzman et al., 2001; Guzman, 2003). They include inhibition of PKA, superoxide generation, and strong increase in intracellular calcium concentration (Howlett et al., 2002). However, the best characterized mechanism of cannabinoidinduced cell death involves sustained accumulation of pro-apoptotic sphingolipid ceramide, which modulates signaling pathways crucial in the control of tumour cell growth and survival (Galve-Roperh et al., 2000; Sanchez et al., 2001). Cannabinoid receptor activation triggers two peaks of ceramide generation in glioma cells (Galve-Roperh et al., 2000; Sanchez et al., 2001; Gomez del Pulgar

29 Glioblastoma: Anti-tumor Action of Natural Synthetic Cannabinoids

et al., 2002). Treatment with 9 -THC or another CB1/CB2 receptor agonist produces a rapid release of ceramide via enzymatic hydrolysis of sphingomyelin from the cell membrane. The second ceramide peak is generated within hours or days after receptor activation and depends on increase of ceramide synthesis de novo via induction of serine palimitoyltransferase, a regulatory enzyme of sphingolipid biosynthesis (Gomez del Pulgar et al., 2002). Selective CB2 receptor agonists, such as JWH133, are supposed to stimulate only the ceramide synthesis process, which is sufficient for turning on the cell death program (Sanchez et al., 2001; Gomez del Pulgar et al., 2002). Thus, enhanced production of ceramide de novo is considered as an important event in cannabinoids-induced apoptosis. Galve-Roperh and co-workers (2000) postulated that the increased ceramide levels reported upon cannabinoid challenge led to prolonged activation of Raf-1/MEK/Erk signaling cascade and thus mediated glioma cell cycle arrest and cell death. The same authors showed also that pharmacological inhibition of ceramide synthesis de novo prevented the inhibition of protein kinase B (also known as Akt) triggered by cannabinoids (Gomez del Pulgar et al., 2002). Our studies revealed that rather down-regulation of Erk activity, together with inhibition of PI3K/Akt pathway, contributed to C6 glioma cell death induced by WIN55,212-2 (Ellert-Miklaszewska et al., 2005). The serine/threonine protein kinase Akt, activated downstream of phosphoinositide 3-kinase (PI3K), as well as the Ras-activated Raf1/MEK/Erk pathway are widely recognized as key mediators of growth factor-promoted cell survival in gliomas (Kapoor and O’Rourke, 2003). Both survival pathways converge on a small pro-apoptotic member of a Bcl-2 family of proteins, Bad. Phosphorylation of Bad by Akt and Erk retains the protein in the cytosol, where it is recognized by certain regulatory proteins and sequestered (Zha et al., 1996). Otherwise, Bad translocates to mitochondria, and formation of heterodimers between nonphosphorylated Bad and anti-apoptotic proteins, such as Bcl-XL or Bcl-2, may result in a loss of integrity of the outer mitochondrial membrane (Zha et al., 1996). The release of apoptogenic proteins, including cytochrome c, triggers the executive phase of programmed cell death. We proposed a mechanism, in which the decrease of mitogenic/prosurvival signaling evoked by the synthetic cannabinoid

281

WIN55,212-2 promoted the pro-apoptotic function of Bad (Fig. 29.2). Accordingly, we demonstrated changes in Bad phosphorylation level followed by collapse of the mitochondrial membrane potential in C6 glioma cells treated with WIN55,212-2. These events preceded activation of caspase 9 by factors released from disrupted mitochondria, subsequent processing of effector caspases, cysteine proteases, which play essential roles in apoptosis, and finally oligonucleosomal DNA fragmentation. Our further studies, as well as some published data suggest that human glioma cells treated with cannabinoids enter the suicide cell death using the same mitochondria-dependent (intrinsic) pathway (Carracedo et al., 2006b). This mechanism contributes to the induction of apoptosis by cannabinoids also in other types of tumor cells (Velasco et al., 2007). However, an involvement of an alternative, death receptor–dependent (extrinsic) pathway in the apoptotic process triggered by these compounds cannot be ruled out.

The Role of ER Stress and Autophagy in Cannabinoid-Induced Cell Death Different experimental approaches shown that the pro-apoptotic and tumor growth-inhibiting activity of cannabinoids relies on the accumulation of de novosynthesized ceramide, an event that occurs in the ER (endoplasmic reticulum) and eventually leads to execution of cell death via mitochondrial intrinsic pathway. A recent study by Carracedo et al. 2006b) suggested a new link between the two events, alternative to inhibition of pro-survival pathways. They showed that 9 THC treatment of glioma cells leads to up-regulation of the stress-regulated protein p8 and ER stress-related downstream targets: ATF-4 (activating transcription factor 4), CHOP (the C/EBP-homologous protein) and TRB3 (tribbles homologue 3). Selective knockdown of ATF-4 and TRB3 blocked cannabinoid-induced apoptosis in glioma cells. Inhibition of ceramide synthesis de novo prevented 9 -THC-induced p8, ATF-4, CHOP and TRB3 up-regulation as well as ER dilation, indicating that ceramide accumulation is an early event in the cannabinoid-triggered ER stress and apoptosis in glioma cells (Carracedo et al., 2006b). Furthermore, ER-stress–evoked stimulation of

282

A. Ellert-Miklaszewska et al.

Fig. 29.2 Mechanisms of multimode anti-tumoral action of cannabinoids. Natural and synthetic cannabinoids induce tumor cell death, block angiogenesis and invasiveness crucial for tumor progression. Increased ceramide synthesis de novo via induction of serine palimitoyltransferase (SPT) cell death. Inhibition of pro-survival pathways (Akt and Erk signaling) stimulates translocation of Bad to the outer mitochondrial membrane and its pro-apoptotic function. Interaction between Bad and Bcl-2 triggers a decrease of the mitochondrial membrane potential () and release of pro-apoptotic factors (such as cytochrome c)

to the cytosol, where apoptosis is executed by caspase cascade. Alternatively, induction of apoptosis by cannabinoids can be mediated by ER (endoplasmic reticulum)-stress and autophagy. Up-regulation of the stress-regulated protein p8 and ER-stress-related downstream targets: ATF-4 (activating transcription factor 4), CHOP (the C/EBP-homologous protein) and TRB3 (tribbles homologue 3) leads to inhibition of Akt, an upstream activator of mTOR. Decreased activity of Akt/mTOR pathway contributes to initiation of autophagy, that precedes apoptosis of glioma cell

the p8/TRB3 pathway preceded the inhibition of the Akt/mTOR (mammalial target of rapamycin) axis, considered a key step in the early triggering of autophagy (Salazar et al., 2009). Irradiation and chemotherapeutic drugs kill cancer cells typically through induction of apoptosis. However most cancers (including gliomas) are resistant to therapies that induce apoptosis (type I programmed cell death). Macroautophagy, hereafter called autophagy, is an evolutionarily conserved catabolic process, where a cell self-digests its cytoplasmic contents. Autophagy is accompanied by the progressive development of vesicle structures from autophagosomes to autolysosomes. Autophagy is activated in response to multiple stresses during cancer progression, such as nutrient starvation, the unfolded protein response (ER stress) and hypoxia.

Indeed, a detailed analysis of human astrocytoma cell lines and a primary culture of human glioma cells indicated that 9 -THC treatment led to formation of autophagosomes in tumor cells (Salazar et al., 2009). The authors also observed that ER stress occurred earlier than autophagy and it preceded apoptosis in cannabinoid-induced human and mouse cancer cell death. Administration of 9 -THC to mice bearing tumors derived from human astrocytoma cells produced increased TRB3 expression, inhibition of mTOR signaling pathway, occurrence of autophagy markers and caspase-3 activation. These findings indicate that cannabinoid promotes the autophagy-mediated cell death through stimulation of ER stress in human glioma cells. Classically autophagy has been proposed to promote cell survival but paradoxically autophagy has

29 Glioblastoma: Anti-tumor Action of Natural Synthetic Cannabinoids

been reported to contribute to cell death (type II programmed cell death). Recently it was also demonstrated that autophagy, not apoptosis, is induced in glioblastoma cell by radiation and several chemotherapeutic agents (Aoki et al., 2008). The most striking evidence for pro-autophagic chemotherapy to overcome apoptosis resistance in cancer cells comes from the use of temozolomide (Aoki et al., 2008), a cytotoxic drug, which has demonstrated therapeutic benefits in glioblastoma patients. However, the role of autophagy in causing cell death, rather than occurring along with cell death, is still not clear. As reported by Salazar et al. (2009) pharmacological or genetic inhibition of autophagy at different levels prevented 9 -THC-induced cell death, suggesting that autophagy contributes to cannabinoid anti-tumoral action.

Other Targets of Cannabinoid Action Cannabinoids have displayed a great potency in reducing glioma tumor growth in experimental animal models. Apparently their effectiveness in vivo has been attributed to several mechanisms of action. Cannabinoids have recently emerged as compounds that beyond inhibition of tumor cell proliferation and survival impair tumor angiogenesis, invasiveness and even gliomagenesis (Blazquez et al., 2003; Aguado et al., 2007). Increased demand for oxygen and nutrients supply to proliferating cancer cells makes angiogenesis a critical factor for the progression of solid tumors and a popular target for oncologic therapies. Local administration of the CB2 selective cannabinoid JWH133 in a mouse flank inoculation model of glioma turned the vascular hyperplasia characteristic of actively growing tumors to a pattern of blood vessels characterized by small, differentiated and impermeable capillaries, thus proving antiangiogenic potential of the cannabinoid (Blazquez et al., 2003). This was associated with a reduced expression of pro-angiogenic factors, such as vascular endothelial growth factor (VEGF) and angiopoietin-2 in cannabinoid treated tumors and could be partly related to a direct influence of the cannabinoid on endothelial cells migration and survival as well as on VEGF signaling in tumor cells. This and subsequent studies of the same authors (Blazquez

283

et al., 2003, 2008) showed that 9 -THC or JWH133 administration to glioma-bearing mice decreased the activity and expression of matrix metalloproteinase-2 (MMP-2). MMP-2 is a proteolytic enzyme that allows tissue breakdown and remodeling during angiogenesis and metastasis and its up-regulation is associated with high progression and poor prognosis of gliomas. 9 THC inhibited MMP-2 expression and cell invasion in cultured glioma cells (Blazquez et al., 2008). Blocking the release of pro-angiogenic factors or the extracellular matrix degrading enzyme by in vivo administered cannabinoids correlated with decreased tumor volumes (Blazquez et al., 2003, 2008). Although the identification of the cell of origin of gliomas is still a matter of debate, recent findings support the existence of brain cancer stem cells generated by transformation of the normal neural stem cell. Aguado and coworkers (2007) showed that activation of CB1 and CB2 receptors with HU-210 and JWH133 promoted differentiation of glioma stem-like cells derived from GBM biopsies and from glioma cell lines. Moreover, cannabinoid challenge decreased neurosphere formation and cell proliferation in secondary xenografts, which correlated with the decreased efficiency of cannabinoid-treated glioma stem-like cells to initiate glioma formation in vivo (Aguado et al., 2007). Multimode action of cannabinoids, including inhibition of gliomagenesis may have important implications for development of cannabinoid-based therapeutic strategies.

Perspectives for the Clinical Use of Cannabinoids in Glioma Patients The past few decades have seen renewed interest in medicinal cannabis. Capsules of THC and its synthetic analogue nabilone are approved in several countries as a prescription-only medicine for nausea and vomiting caused by cytotoxic chemotherapy, unresponsive to other anti-emetics. Other potential palliative effects of cannabinoids in oncology – supported by phase III clinical trials – include apetite stimulation and pain inhibition (Guzman, 2003). However, discussions on controlled medical use of cannabis and cannabinoids are now beyond its application for symptomatic relief. There is quite a large number of emerging evidence that cannabinoids can in fact modify the progression

284

of certain diseases. The efficacy of intra-tumorally or systemically administered cannabinoids against high-grade glioma in rat and mouse models (GalveRoperh et al., 2000; Sanchez et al., 2001; Massi et al., 2004; Duntsch et al., 2006), raised promises of using cannabinoid-based therapies against malignant gliomas.

Safety Profile of Therapeutic Cannabinoids Standard chemotherapeutics are a double edged sword, they eliminate cancer cells but affect severely healthy cells in the body. Based on evidence from in vitro and in vivo studies cannabinoids appear to have a favorable safety profile and do not produce the generalized toxic effects as most conventional chemotherapeutic drugs. Several mechanisms could be responsible for cannabinoid targeting only the cancer cells. In contrast to pro-apoptotic action of CB1/CB2-activating cannabinoids, such as 9 -THC and WIN55,212-2, on transformed cells, treatment of primary cultured astrocytes with these compounds did not trigger ceramide accumulation or induction of ER stress-related genes. Furthermore, cannabinoids promote survival of glial cells and neurons in different models of injury, suggesting that the anti-proliferative effect of cannabinoids is selective for brain tumor cells, viability of normal brain cells remains unaffected or even favored by cannabinoid challenge (Guzman, 2003). In our studies astrocytes and glioma cells were vulnerable to WIN55,212-2 which was in line with data on human primary mixed glial cultures and glioblastoma cells treated with the synthetic cannabinoid (McAllister et al., 2005). Administration of JWH133 although effective toward tumor cells, did not affect survival or morphology of normal astrocytes (unpublished). Apparently negligible expression of CB2 receptor in the brain and its abundance in high grade gliomas seems to confer a relative safety of CB2-selective agonists for targeted glioma therapy. Concomitantly, a novel cannabinoid chemotherapeutic exhibiting over 10-times higher affinity for the CB2 vs. CB1 receptor produced a minimal toxicity to healthy cultured brain slices at micromolar concentrations required to eradicate U87 glioma cells (Duntsch et al., 2006).

A. Ellert-Miklaszewska et al.

As reported by the same authors, clinical monitoring of mice with U87-derived subcutaneous tumors proved that treatment with the cannabinoid was effective in reducing tumor volume and produced no general toxic effects under intratumoral or systemic administration. More importantly, some animal studies show also a good safety profile of CB1-activating agents. In the studies by Galve-Roperh et al. (2000) 9 THC and WIN55,212-2 were delivered intratumorally via an osmotic pump for 7 days leading to partial remission or complete eradication of the tumors generated intracranially by inoculation of C6 glioma cells The same effective doses of cannabinoids were evaluated in tumor-free animals for potential adverse effects in the brain. MRI analysis showed no signs of damage related to necrosis, edema, infection or trauma. Cannabinoid administration to tumor-bearing and control rats induced no substantial modification in behavioral parameters, in food and water intake or in body weight. No abnormalities in biochemical and hematological parameters nor markers of tissue damage have been revealed during 7-day delivery period and for at least 2 months after cannabinoid treatment (Galve-Roperh et al., 2000). The strategy to use CB2 selective compounds in glioma treatment has at least one more advantage over the non-selective cannabinoids. The medical use of cannabinoids is limited mainly by their undesirable side-effects attributed to marijuana abuse. Due to the well known psychotropic activity of 9 -THC and related compounds mimicking its action on the CB1 receptors, potential application of these agents to patients raises a number of clinical and ethical considerations. Among various approaches to avoid CB1receptor-mediated psychodysleptic side effects, special attention is paid to substances, which selectively stimulate the CB2 receptors, putatively overexpressed in target tumor cells (Guzman et al., 2001; Duntsch et al., 2006).

Results from a Pilot Clinical Study On the basis of successful preclinical findings Guzman and his collaborators conducted a pilot phase I clinical trial, in which nine patients with actively growing recurrent glioblastoma multiforme were administered THC intratumorally via an infusion catheter placed

29 Glioblastoma: Anti-tumor Action of Natural Synthetic Cannabinoids

into the resection cavity (Guzman et al., 2006). The enrolled patients had previously failed standard therapy (surgery and radiotherapy) and had clear evidence of tumor progression. The primary endpoint of the study was to determine the safety of intracranial 9 THC administration in a dose escalation regime. The initial dose of 9 -THC delivered to the patients was 20–40 μg at day 1, increasing progressively for 2–5 days up to 80–180 μg/day. Over the median treatment duration of 15 days patients underwent continuous physical, neurological, biochemical and hematological examinations as well as frequent magnetic resonance and computed tomography scans of the brain. The cannabinoid treatment did not result in any marked alterations of clinical and laboratory tests. There was no evidence of hemorrhage, oedema or injury. The good safety profile observed in this pilot clinical trial, together with remarkable anti-tumor effects reported in preclinical studies may set the basis for further research aimed at better evaluation of the potential anti-cancer activity of cannabinoids.

Risks, Constrains and Benefits The controlled clinical use of natural or synthetic cannabinoids should not be perceived equally as the use of smoked cannabis, which is a declared addictive drug. Noteworthy, central and peripheral effects of cannabinoids pronounced in marihuana smokers are not apparent in a controlled clinical setting (Guzman, 2003). Moreover, tolerance to these unwanted effects of cannabinoids develops rapidly in humans and laboratory animals. The possibility that tolerance also develops after therapeutic application as well as psychotic and addictive potential of cannabinoids have not been thoroughly examined. Although the general consensus indicates that cannabinoids have anti-tumor effects there are a few studies that have shown contradictory results or duality of cannabinoids effects on tumor cell fate depending of the dose range used (Guzman, 2003). Some authors questioned whether cannabinoid receptors do constitute relevant molecular targets to treat glioblastoma in humans, as the high concentrations of the currently available cannabinoids required to trigger apoptosis in glioma cells in culture are not pharmacologically relevant (Held-Feindt et al., 2006). This

285

emphasizes the need for comprehensive dose-response studies in the future correlated with analysis of the cannabinoid receptor expression. In general, existing data encourage to survey of new synthetic compounds in order to replace 9 -THC or related cannabinoids with more potent and selective mimetics. Another possibility would be to prolong the effectiveness of endocannabinoids by development of the inhibitors of their breakdown. Due to genetic and epigenetic alterations malignant glioblastomas are highly resistant to radiation and chemotherapy. Mainstream therapeutic strategies for the management of all primary brain tumors are still mostly palliative, known to leave survivors with devastating neurological deficits and frequently with a high risk of the disease recurrence. The potency of synthetic cannabinoids to induce apoptosis in glioblastoma cells has been tested by us and others on several cell lines and primary cell cultures derived from biopsies of human tumors, which to some extent may reflect the heterogeneity of glioma molecular characteristics (Galve-Roperh et al., 2000; Sanchez et al., 2001; Guzman, 2003; Massi et al., 2004; Duntsch et al., 2006; Ellert-Miklaszewska et al., 2005; McAllister et al., 2005). Thus, cannabinoids were able to override alterations of growth regulatory and apoptotic pathways, caused by common mutations reported in primary and secondary glioblastomas. Cannabinoid apoptotic action relies on the generation of ceramide and disruption of signaling pathways crucial for regulation of cellular proliferation, differentiation or apoptosis (Galve-Roperh et al., 2000; Gomez del Pulgar et al., 2002; Ellert-Miklaszewska et al., 2005; Salazar et al., 2009). Their unique mechanism of action among standard oncology remedies justifies further research on their anti-tumoral properties either alone or in combined therapies.

References Aguado T, Carracedo A, Julien B Velasco G, Milman G, Mechoulam R, Alvarez L, Guzman M, Galve-Roperh I (2007) Cannabinoids induce glioma stem-like cell differentiation and inhibit gliomagenesis. J Biol Chem 282:6854–6862 Aoki H, Kondo Y, Aldape K, Yamamoto A, Iwado E, Yokoyama T, Hollingsworth EF, Kobayashi R, Hess K, Shinojima N, Shingu T, Tamada Y, Zhang L, Conrad C, Bogler O, Mills G, Sawaya R, Kondo S (2008) Monitoring autophagy in

286 glioblastoma with antibody against isoform B of human microtubule-associated protein 1 light chain 3. Autophagy 4:467–475 Blazquez C, Casanova ML, Planas A, Del Pulgar TG, Villanueva C, Fernandez-Acenero MJ, Aragones J, Huffman JW, Jorcano J L, Guzman M (2003) Inhibition of tumor angiogenesis by cannabinoids. FASEB J 2:2 Blazquez C, Salazar M, Carracedo A, Lorente M, Egia A, Gonzalez-Feria L, Haro A, Velasco G, Guzman M (2008) Cannabinoids inhibit glioma cell invasion by downregulating matrix metalloproteinase-2 expression. Cancer Res 68:1945–1952 Caffarel MM, Sarrio D, Palacios J, Guzman M, Sanchez C (2006) Delta9-tetrahydrocannabinol inhibits cell cycle progression in human breast cancer cells through Cdc2 regulation. Cancer Res 66:6615–6621 Carracedo A, Gironella M, Lorente M, Garcia S, Guzman M, Velasco G, Iovanna JL (2006a) Cannabinoids induce apoptosis of pancreatic tumor cells via endoplasmic reticulum stress-related genes. Cancer Res 66:6748–6755 Carracedo A, Lorente M, Egia A, Blazquez C, Garcia S, Giroux V, Malicet C, Villuendas R, Gironella M, Gonzalez-Feria L, Piris MA, Iovanna J L, Guzman M, Velasco G (2006b) The stress-regulated protein p8 mediates cannabinoid-induced apoptosis of tumor cells. Cancer Cell 9:301–312 Duntsch C, Divi MK, Jones T, Zhou Q, Krishnamurthy M, Boehm P, Wood G, Sills A, Ii BM (2006) Safety and efficacy of a novel cannabinoid chemotherapeutic, KM-233, for the treatment of high-grade glioma. J Neurooncol 77(2):143–152 Ellert-Miklaszewska A, Grajkowska W, Gabrusiewicz K, Kaminska B, Konarska L (2007) Distinctive pattern of cannabinoid receptor type II (CB2) expression in adult and pediatric brain tumors. Brain Res 1137:161–169 Ellert-Miklaszewska A, Kaminska B, Konarska L (2005) Cannabinoids down-regulate PI3K/Akt and Erk signalling pathways and activate proapoptotic function of bad protein. Cell Signal 17:25–37 Galve-Roperh I, Sanchez C, Cortes ML, del Pulgar TG, Izquierdo M, Guzman M (2000) Anti-tumoral action of cannabinoids: involvement of sustained ceramide accumulation and extracellular signal-regulated kinase activation. Nat Med 6:313–319 Gomez del Pulgar T, Velasco G, Sanchez C, Haro A, Guzman M (2002) De novo-synthesized ceramide is involved in cannabinoid-induced apoptosis. Biochem J 363:183–188 Gong JP, Onaivi ES, Ishiguro H, Liu QR, Tagliaferro PA, Brusco A, Uhl GR (2006) Cannabinoid CB2 receptors: immunohistochemical localization in rat brain. Brain Res 1071:10–23 Guzman M (2003) Cannabinoids: potential anti-cancer agents. Nat Rev Cancer 3:745–755 Guzman M, Duarte MJ, Blazquez C, Ravina J, Rosa MC, GalveRoperh I, Sanchez C, Velasco G, Gonzalez-Feria L (2006) A pilot clinical study of Delta9-tetrahydrocannabinol in patients with recurrent glioblastoma multiforme. Br J Cancer 95:197–203 Guzman M, Sanchez C, Galve-Roperh I (2001) Control of the cell survival/death decision by cannabinoids. J Mol Med 78:613–625 Held-Feindt J, Dorner L, Sahan G, Mehdorn HM, Mentlein R (2006) Cannabinoid receptors in human astroglial tumors. J Neurochem 98:886–893

A. Ellert-Miklaszewska et al. Howlett AC, Barth F, Bonner TI, Cabral G, Casellas P, Devane WA, Felder CC, Herkenham M, Mackie K, Martin BR, Mechoulam R, Pertwee RG (2002) International union of pharmacology. XXVII. Classification of cannabinoid receptors. Pharmacol Rev 54:161–202 Kapoor GS, O’Rourke DM (2003) Receptor tyrosine kinase signaling in gliomagenesis: pathobiology and therapeutic approaches. Cancer Biol Ther 2:330–342 Mackie K (2006) Cannabinoid receptors as therapeutic targets. Annu Rev Pharmacol Toxicol 46:101–122 Massi P, Vaccani A, Ceruti S, Colombo A, Abbracchio MP, Parolaro D (2004) Anti-tumor effects of cannabidiol, a nonpsychoactive cannabinoid, on human glioma cell lines. J Pharmacol Exp Ther 308:838–845 McAllister SD, Chan C, Taft RJ, Luu T, Abood ME, Moore DH, Aldape K, Yount G (2005) Cannabinoids selectively inhibit proliferation and induce death of cultured human glioblastoma multiforme cells. J Neurooncol 74:31–40 Munson AE, Harris LS, Friedman MA, Dewey WL, Carchman RA (1975) Antineoplastic activity of cannabinoids. J Natl Cancer Inst 55:597–602 Petersen G, Moesgaard B, Schmid PC, Schmid HH, Broholm H, Kosteljanetz M, Hansen HS (2005) Endocannabinoid metabolism in human glioblastomas and meningiomas compared to human non-tumour brain tissue. J Neurochem 93:299–309 Salazar M, Carracedo A, Salanueva IJ, Hernandez-Tiedra S, Lorente M, Egia A, Vazquez P, Blazquez C, Torres S, Garcia S, Nowak J, Fimia GM, Piacentini M, Cecconi F, Pandolfi PP, Gonzalez-Feria L, Iovanna JL, Guzman M, Boya P Velasco G (2009) Cannabinoid action induces autophagy-mediated cell death through stimulation of ER stress in human glioma cells. J Clin Invest 119:1359–1372 Sanchez C, de Ceballos ML, del Pulgar TG, Rueda D, Corbacho C, Velasco G, Galve-Roperh I, Huffman JW, Ramon y Cajal S, Guzman M (2001) Inhibition of glioma growth in vivo by selective activation of the CB(2) cannabinoid receptor. Cancer Res 61:5784–5789 Sanchez C, Galve-Roperh I, Canova C, Brachet P, Guzman M (1998) Delta9-tetrahydrocannabinol induces apoptosis in C6 glioma cells. FEBS Lett 436:6–10 Sarfaraz S, Afaq F, Adhami VM, Mukhtar H (2005) Cannabinoid receptor as a novel target for the treatment of prostate cancer. Cancer Res 65:1635–1641 Schley M, Stander S, Kerner J, Vajkoczy P, Schupfer G, Dusch M, Schmelz M, Konrad C (2009) Predominant CB2 receptor expression in endothelial cells of glioblastoma in humans. Brain Res Bull 79:333–337 Stella N (2004) Cannabinoid signaling in glial cells. Glia 48:267–277 Valenzano KJ, Tafesse L, Lee G, Harrison JE, Boulet JM, Gottshall SL, Mark L, Pearson MS, Miller W, Shan S, Rabadi L, Rotshteyn Y, Chaffer SM, Turchin PI, Elsemore DA, Toth M, Koetzner L, Whiteside GT (2005). Pharmacological and pharmacokinetic characterization of the cannabinoid receptor 2 agonist, GW405833, utilizing rodent models of acute and chronic pain, anxiety, ataxia and catalepsy. Neuropharmacology 48:658–672 Van Sickle MD, Duncan M, Kingsley PJ, Mouihate A, Urbani P, Mackie K, Stella N, Makriyannis A, Piomelli D, Davison JS, Marnett LJ, Di Marzo V, Pittman QJ, Patel KD, Sharkey

29 Glioblastoma: Anti-tumor Action of Natural Synthetic Cannabinoids KA (2005) Identification and functional characterization of brainstem cannabinoid CB2 receptors. Science 310:329–332 Velasco G, Carracedo A, Blazquez C, Lorente M, Aguado T, Haro A, Sanchez C, Galve-Roperh I, Guzman M (2007) Cannabinoids and gliomas. Mol Neurobiol 36:60–67

287

Zha J, Harada H, Yang E, Jockel J, Korsmeyer SJ (1996) Serine phosphorylation of death agonist BAD in response to survival factor results in binding to 14-3-3 not BCL-X(L). Cell 87:619–628

Chapter 30

Patients with Recurrent High-Grade Glioma: Therapy with Combination of Bevacizumab and Irinotecan Benedikte Hasselbalch, Ulrik Lassen, and Hans Skovgaard Poulsen

Abstract Despite surgery, radiotherapy, and chemotherapy, most patients with recurrent high-grade glioma have a poor prognosis. High-grade gliomas are known to be vascular tumors that produce the vascular endothelial growth factor (VEGF), one of the mediators necessary for angiogenesis which facilitates for tumor growth. Treatment of recurrent high-grade gliomas with the VEGF neutralizing antibody, bevacizumab in combination with the topoisomerase inhibitor irinotecan has demonstrated promising response rates and enhanced progression free survival when compared to historical controls. However, it still needs to be confirmed, that bevacizumab and irinotecan improves overall survival. Moreover, as not all patients benefit from this treatment, there is an ongoing search for one or more predictive biomarkers. Keywords HGG · Bevacizumab · Irinotecan · VEGF · Topoisomerase I · Hypoxia

Introduction Malignant primary brain tumors represents 1–2% of all newly diagnosed tumors and account for about 2% of all cancer-related deaths. Overall, malignant gliomas account for 78% of malignant brain tumors. Primary brain tumors (PBT) are of neuroepithelial origin and according to WHO classification there are three main

H.S. Poulsen () Department of Radiation Biology, The Finsen Center, Sec 6321, Copenhagen University Hospital, DK-2100 Copenhagen, Denmark e-mail: [email protected]

types which usually can be distinguished by their histological features; oligodendrogliomas, mixed oligoastrocytomas and astrocytomas (or gliomas) (Louis et al., 2007). Through analyzing the most malignant region of the tumors, PBT are graded as low-grade tumors (WHO grades I and II), or as high-grade tumors (WHO grades III and IV) dependent on four main features: nuclear atypia, mitoses, microvascular proliferation, and necrosis. By the degree of increasing anaplasia, the types of astrocytomas usually include pilocytic astrocytoma (grade I), diffuse astrocytoma (grade II), anaplastic astrocytoma (grade III) and the most malignant form, glioblastoma multiforme (grade IV/GBM). In addition, grade III malignant tumors include anaplastic oligodendroglioma, anaplastic oligoastrocytoma and mixed glioma. Areas of vascular proliferation and/or necrosis are mandatory for GBM whereas occasional proliferation of tumor vessels can occur in grade III astrocytomas (Louis et al., 2007). The pronounced vascularization arises because of increased angiogenesis. The dense vascularity does not prevent the HGG tumor from being hypoxic, partly because of the dysfunctional nature of the tumor vessels. Standard treatment for HGG is debulking surgery if possible. For GBM, this is followed by concomitant R temozolomide (Temodal ); an oral alkylating agent, plus radiotherapy and adjuvant temozolomide (Stupp et al., 2005). The introduction of temozolomide has improved the survival of GBM significantly, increasing the 2-year survival from 10 to 27% and 5-year survival from 1.9 to 9.8%, compared to previous treatment regimens (Stupp et al., 2009). However, nearly all patients with GBM will eventually relapse. The prognosis for recurrent GBM is even worse with a median survival of

M.A. Hayat (ed.), Tumors of the Central Nervous System, Volume 2, DOI 10.1007/978-94-007-0618-7_30, © Springer Science+Business Media B.V. 2011

289

290

3–9 months when using traditional chemotherapeutic agents. The survival benefit of temozolomide as part of a primary setting for grade III astrocytomas still needs to be determined. Accordingly, the primary treatment of grade III astrocytoma varies between the continents, e.g. most patients in north America receives the same treatment regimen as GBM, whereas most places in Europe use surgery and radiotherapy as primary treatment, giving temozolomide at recurrence, -if possible, after tumor-reductive surgery. The median survival is longer for this group of patients (2–4 years), although most typically recur and may progress to GBM. The primary treatment of anaplastic oligodendroglioma includes debulking surgery followed by radiotherapy, whereas the timing and survival benefit from the addition of chemotherapy still remains to be established. Two large randomized control trials failed to show that the addition of chemotherapy with procarbazine, lomustin and vincristin (PCV) either adjuvant or neoadjuvant to radiotherapy significantly prolonged disease free survival (Cairncross et al., 2006; van den Bent et al., 2006). However, as most patients randomly assigned to RT alone received chemotherapy at progression, interpretation of the survival data from these studies are difficult. HGG are still considered incurable and accordingly there is a pivotal need for improved treatment strategies for this malignancy. The aim of therapy for patients with relapsed HGG is to offer a better quality of life in terms of improvement in neurological function, neuro-cognitive function and steroid requirements, ideally coupled with a prolongation of survival. Despite numerous studies using temozolomide, PCV, carboplatin, lomustin (CCNU), carmustine (BCNU), or imatinib with or without the combination of hydroxyurea, no regimen has emerged as a standard for relapsed HGG after radiation and temozolomide. Recently, the use of bevacizumab and irinotecan has shown promising results in recurrent HGG, improving progression free survival (PFS) and quality of life compared with historical results (Friedman et al., 2009; Wong et al., 1999). However, it still needs to be established if a regimen including bevacizumab for recurrent HGG, improve overall survival (OS) compared to other treatment regimens used at recurrence. Furthermore, there is an ongoing search for one or more biomarkers that can predict which patients will benefit from treatment with bevacizumab.

B. Hasselbalch et al.

Hypoxia Hypoxia plays a prominent role in tumor growth, invasion, angiogenesis, resistance to chemo- and radiotherapy and decreased patient survival in various cancer types, including HGG. The characteristic necrotic regions of GBM are assumed to be regions of hypoxia, although this involvement is not conclusively determined. When available blood flow cannot fulfill the requirements for maintaining oxygen homeostasis, the partial oxygen pressure of these tumor areas become low, i.e. hypoxic, or close to zero, anoxic. The diffusion limit for oxygen is approximately 100 μm and oxygen transport over further distances requires red blood cells. Tumor hypoxia evolves as a consequence of insufficient oxygen delivery and is a feature of most solid tumors. Tumor growth results in increased distance to existing blood vessels, which in combination with insufficient neo-vascularization, contributes to a tumor microenvironment with low oxygen tension. Moreover, tumor vessels are leaky, leading to tumor edema and increased intratumoral pressure, which further increases hypoxia. Cancer cells undergo numerous changes that enable them to adapt to and survive hypoxia, contributing to a more aggressive behavior of the tumor. The hypoxia inducible factors (HIF), HIF-1α and HIF-2α, are critical for this adaptive response. The characteristic necrotic regions of GBM are surrounded by a cluster of cells known as pseudopalisading that are suspected to be regions of hypoxia, although this has not been conclusively proven. These necrotic areas do not seem to be related to tumor size, as they have been found in both small and large tumors. Furthermore, it has been demonstrated in animal glioma models that tumors 58 mo; median OS 35.8 mo

Long-lasting tumor-free survival (> 5 year after vaccination) Partial response (n=1); tumor-free survival (n=2; 5 year after vaccination) Partial response (n=4); mixed response (n=1)

Significant increase in median survival

32 High-Grade Gliomas: Dendritic Cell Therapy 323

Publication

Clin Cancer Res 11:4160–4167 (2005)

J Transl Med 5:67 (2007)

Cancer Res 68:5955–5964 (2008)

J Clin Neurosci 15:114–121 (2008)

Author

Yamanaka et al.

Okada et al.

Wheeler et al.

Walker et al.

Table 32.2 (continued)

13 (phase I)

34 (phase II)

2 (A)/5 (B)

24 (phase I-II)

No. of patients (type of trial)

Single-cell suspension of autologous tumor cells

IL-4 transfected fibroblasts + apoptotic glioma cells (A)/autologous tumor lysate (B) Autologous tumor lysate

Autologous tumor lysate

Cell product

Intradermal

Subcutaneous

Intradermal

Intradermal or intradermal + intratumoral (Ommaya)

Administration

Priming phase consisting of 6 vaccinations at 2-week intervals; further vaccinations at 6-week intervals

3 vaccinations at 2-week intervals with fourth vaccination 6 week after the third

2 vaccinations with 2-week interval

1–10 vaccinations at 3-week intervals

Treatment schedule

Post-vaccine antigen-directed IFNγ response in PBMC (qPCR-based assay) (n=17); DTH-test resulted in cutaneous GBM in 1 pat (DTH was subsequently discontinued) Increased T cell infiltration in post-vaccination tumor specimens compared to pre-vaccination specimens (n=3)

A: increased T cell reactivity; B: no response

Positive DTH reaction to tumor lysate (n=8); positive IFNγ Elispot (n=7)

Immune response

9-month survival (9/13); 12-month survival (6/13); 18 months or longer survival (3/13)

Significant positive correlation between post-vaccine response magnitude on one hand and TTS and TTP spanning chemotherapy on the other hand

Partial response (n=1); minor response (n=3); significant increase in median survival A: clinical and radiological response in both patients; B: no response

Clinical response

324 H. Ardon et al.

Clin Cancer Res 14:3098–3104 (2008)

Pediatr Blood Cancer 54:519–525 (2010a)

J Neurooncol 99:261–272 (2010b)

De Vleeschouwer et al.

Ardon et al.

Ardon et al.

8 (pilot study)

45 children (phase I-II) (33 HGG)

56 (phase I-II)

No. of patients (type of trial)

Autologous tumor lysate

Autologous tumor lysate

Autologous tumor lysate

Cell product

Intradermal

Intradermal

Intradermal

Administration

Cohort comparison

Cohort comparison

Treatment schedule

Not reported

Positive DTH reaction to tumor lysate (9/21 at time of diagnosis and 9/17 after two vaccinations)

Immune response

PFS 3 months; OS 9.6 months; 2-year OS 14.8%; total resection is predictor for better PFS; younger age and total resection are predictors for better OS in univariable analysis; tendency towards improved PFS when faster DC vaccination schedule with tumor lysate boosting was applied PFS HGG 4.4 months – GBM 4.3 months; OS HGG 13.5 months – GBM 12.2 months; 2-year OS 18% 6-month PFS 75%; PFS 18 months; OS 24 months

Clinical response

Incorporated in Positive DTH RChT – 4 reaction to tumor induction lysate (3/7); vaccinations positive IFNγ at 1-week Elispot (5/8); interval; then increase CTL 3 boost (6/7) vaccinations with 1-month interval, followed by 3-monthly boostvaccinations CTL, cytotoxic T lymphocyte; DC, dendritic cell; DTH, delayed-type hypersensitivity; GBM, glioblastoma multiforme; IFNγ, interferon-γ; IL-4, interleukin-4; JAM, just another method; MHC, major histocompatibility complex; mo, month; NK, natural killer; no., number; OS, overall survival; pat, patient; PBMC, peripheral blood mononuclear cell; PFS, progression-free survival; qPCR, quantitative polymerase chain reaction; rhIL-12, recombinant human interleukin-12; RTE, naive recent thymic emigrant T cell; TTP, time to progression; TTS, time to survival; w, week

Publication

Author

Table 32.2 (continued)

32 High-Grade Gliomas: Dendritic Cell Therapy 325

326

T cells are involved. Protective immunity against intracranial glioma growth obtained through immunization with either lysate- or RNA-loaded DC was reported by Insug et al. (2002), while adding recombinant IL-12 to the vaccine regimen further improved its efficacy. Vaccination with lysate-pulsed DC combined with IFN-β gene therapy resulted in a survival benefit, as was demonstrated by Saito et al. (2004). Similarly, the group of Okada et al. (2007) revealed that the sequential intratumoral delivery of an IFN-α encoding adenoviral vector and DC induced longterm survival and specific CTL activity. Furthermore, the same group showed that intratumoral administration of DC, genetically engineered to secrete IFNα, enhanced the efficacy of peripheral vaccines with cytokine gene-transduced tumor cells. Using vaccines created through electrofusion of DC and irradiated tumor cells, Kjaergaard et al. (2005) observed complete tumor regression of established intracranial tumors with infiltration of both CD4+ and CD8+ T cells. Ciesielski et al. (2006) have focused on survivin, a member of the apoptosis inhibition family of proteins, as GL261 TAA. In particular, the authors exploited the xenogeneic differences between human and murine survivin sequences to develop a more immunogenic tumor vaccine. The efficacy of systemic immunotherapy with DC loaded with GL261 antigens was confirmed by Pellegatta et al. (2006a). Additionally, these authors introduced the concept of cancer stem cells in this model and reported that DC targeting of such stem cells within the GL261 tumor cell pool provides a higher level of protection against GL261 glioma (Pellegatta et al., 2006b). Recently, Grauer et al. (2007 and 2008) illustrated the pronounced impact of FoxP3+ Treg in the GL261 model and suggested that Treg elimination is a prerequisite for successful eradication of established glioma using tumor lysate pulsed DC. Maes et al. (2009) showed that DC loaded with tumor antigens in the form of RNA molecules are capable of inducing a T cellmediated antitumoral immune response. Furthermore, the impact of Treg in this model was evident and elimination of Treg resulted in long-term surviving mice. Strikingly, upon rechallenge of long-term surviving mice after Treg depletion and/or DC vaccination, only those mice that were treated with DC vaccination depicted a prolonged antitumoral protective immune response.

H. Ardon et al.

Dendritic Cell Therapy for Human Gliomas Seventeen clinical phase I/II studies and case reports have been published in the literature since 2000 (Table 32.2). The median patient number in these reports was only 12, ranging from 1 to 56. The inclusion criteria, immunotherapeutic designs and interpretations varied significantly among and even within the reports, being based entirely on single-center approaches and hypotheses. Hence, no firm conclusions can be drawn regarding the optimal immunotherapeutic strategy, nor would it make sense to perform a meta-analysis on these data. The importance of a minimal residual disease setting for adjuvant postoperative DC vaccination in HGG patients was already stressed in the first reports. In the series published by De Vleeschouwer et al. (2008), the importance of a gross total resection prior to DC vaccination for GBM patients was confirmed in a multivariate analysis. General awareness and agreement on this important issue is mounting and even culminating in the assumption that therapeutic tumor vaccinations should be used for early disease states. Most of the reports describe patients treated with immunotherapy at a stage of minimal residual disease, after gross total or near-total resection. Interestingly, patients have been under maintenance corticosteroid treatment just prior and/or during immunotherapy in some trials, which one would expect to have a negative effect on the generation of an effective immune response. In order to avoid this confounding variable, other trials have been restricted to patients who could obtain a total or near-total resection and be rapidly weaned from corticosteroids. There are multiple arguments supporting the need for these restrictions. First, because there are major immune-suppressive mechanisms at play within the tumor microenvironment and even systemically, only a gross total resection results in a clinically effective recovery of normal immune system function. Second, it might be difficult to produce good-quality DC out of monocytes isolated at the time of corticosteroid treatment. Third, an overwhelming peritumoral inflammatory reaction was seen in one patient with bulky residual tumor (Rutkowski et al., 2004). In this patient, the vaccine-induced inflammatory immune reaction occurred with a progressively

32 High-Grade Gliomas: Dendritic Cell Therapy

327

increasing delay after the vaccine injection. This is in line with the hypothesis that the growing tumor and, maybe as such, the increasingly induced immune suppression counteract the vaccine-induced immune response. This observation points to a tumor-specific reaction induced by the vaccine, but maybe also to a gradually shifting balance in favor of tumor-induced immune suppression.

monitoring tools directed toward the affected organ, that is, the brain, will have to take into account that the presence of tumor-infiltrating lymphocytes and other immune effector cells does not unequivocally correlate with a good or a bad outcome. Proof of efficacy of immunotherapy asks for impact on OS in patients with HGG, of whom the different prognostic risk factors are clearly categorized.

Selection of Outcome Measures

Nature of Dendritic Cells

Immunological responses, assessed by a wide variety of immune monitoring tools, have been demonstrated in 50% of reported cases, ranging from 20 to 100%. Only recently, Wheeler et al. (2008) found stronger evidence between immunological responses and survival in vaccinated glioma patients. However, the use of peripheral immune monitoring tools might prove inadequate in case the peripheral immune system does not mirror in any way local immune events in the brain tumor microenvironment. Although immunological responses are mandatory to map the way and to provide proof of the principle for this therapeutic strategy, “proof of efficacy” can only be established in large well-designed comparative clinical trials, ideally in controlled randomized trials. Especially in HGG, an old paradigm states that pre-treatment prognostic variables might have more impact on outcome than any (new) potentially active therapy or treatment strategy. To that end, recursive partitioning analysis (RPA) as originally described and validated for newly diagnosed malignant gliomas by the Radiation Therapy Oncology Group (RTOG) in 1993, provides us with an excellent model of prognostic classes of pre-treatment and treatment-related patients’ variables. Using this RPA classification, the impact of new treatments on OS in HGG can be examined in clinically similar patient groups and compared with large databases of conventionally treated patients. To date, there is consensus that clinical responses, as defined by the Macdonald criteria (Macdonald et al., 1990), do not apply for biological treatment modalities such as tumor vaccination. The paradigm of a therapeutic vaccine leading to (detectable) immunological responses and hence to (detectable) clinical responses with increased OS is at least incomplete. Even immune

All clinical reports have used monocyte-derived DC for their trials. The methodology for DC preparation is now fairly well established, and gives a sufficient yield of mature DC (DCm) for injections into patients. Most of the older trials have used immature DC, while others have used maturation stimuli like TNFα, penicillin-killed Streptococcus pyogenes (OK-432), monocyte-conditioned medium, IFN-γ and TNF-α in combination with IL-4-secreting fibroblasts (Van Gool et al., 2009). Only one phase I study focused on the dose of DC and could not find any dose-limiting toxicity (Liau et al., 2005). At the time of the design of most published clinical trials, it was not known if and which lymph nodes would be the optimal destination of injected DC. Later on, however, data became available that priming of T cells by DC within the cervical lymph nodes induced an integrin homing pattern toward intracerebral locations. Other preclinical in vivo brain tumor models clearly showed an enrichment of tumor-specific Treg in the blood and cervical lymph nodes. As pointed out by Tuyaerts et al. (2007), the question whether DC should be injected intradermally, intravenously or in the lymph nodes is not yet resolved. This becomes particularly important as only a small amount of the intradermally injected DC ultimately reaches the T cell area of lymph nodes. However, taking into account the documented failure rate of even experienced radiologists injecting DC in lymph nodes under ultrasound guidance, and the desire to spread DC vaccination technology to more centers in order to perform largescale clinical trials, most researchers in the field favor the intradermal route for DC injection. A very interesting issue is the potential role that local injection of DC directly into the tumor region

328

might play in immunotherapeutical approaches. In their study, Yamanaka et al. (2005) showed some benefit from intratumoral plus intradermal injections of DC as compared with only intradermal injections, although one should note that this observation was not derived from a prospective randomized approach. Nevertheless, a local change from an immune suppressive into a more immune stimulatory microenvironment might be induced by intratumoral administration of DC.

Monitoring Immune responses following vaccination have been monitored in most trials. These analyses have included positive DTH skin reaction, T cell reactivity and NK cell enrichment in peripheral blood, as well as measuring T cell infiltration in tumoral tissue taken after vaccination. The reported immune monitoring remains very global in these clinical trials, mainly due to the lack of specific antigens to be targeted. Although data are still preliminary, advanced magnetic resonance imaging (MRI) eventually combined with positron emission tomography (PET) may soon provide better tools to monitor the effects of immunotherapy of HGG.

Results of Clinical Studies Because all clinical trials reported thus far comprise case reports, phase I, phase I/II or phase II trials, the questions answered in these trials relate to feasibility, toxicity and early efficacy, both immunologically and/or oncologically (Table 32.2). In all trials, it seemed to be feasible to administer DCbased vaccines. Moreover, all reports conclude that the toxicity is minimal except for the case reported by Rutkowski et al. (2004) where an overwhelming inflammatory reaction around a residual tumor was observed. Of major importance, none of the reports describes autoimmune phenomena induced by DC vaccination. Besides feasibility, toxicity and clinical outcome, all published studies on DC-based immunotherapy have aimed to assess immune responses. Several immune assays have been developed for this purpose,

H. Ardon et al.

but so far, results have been difficult to interpret. Moreover, immunological responses are not detected in all patients. A correlation between immunologically measurable effects and clinical outcome was reported in several papers. In the study by Wheeler et al. (2008), a significant progressive (logarithmic) association was found between time to survival after inclusion and post-vaccine IFN-γ enhancement. This allowed them to propose that single-vaccine response metrics can quantitatively reflect inhibition of tumor progression by T cells in GBM patients. Only a limited number of research groups are actively performing clinical trials world-wide. All these trials are small, and include heterogeneous groups of patients with regard to histology and/or time of vaccination. Some studies included patients with all subtypes of HGG, while others focused solely on patients with GBM. Some studied patients with relapsed HGG, while others included patients with a first diagnosis of HGG, either exclusively or combined with patients treated at recurrence. Only one group reported on children with relapsed HGG. Moreover, as pointed out before, the vaccination protocol varied enormously among research groups. Due to this heterogeneity in study design and execution, sound comparative analyses on clinical outcome are difficult. Nevertheless, in most trials, some clinical effect is observed and some important clues for future clinical studies are provided, especially from trials where there are comparisons with historical controls or between internal subgroups. For example, Yu et al. (2001) reported that seven patients with newly diagnosed GBM who received three biweekly intradermal vaccinations with peptide-pulsed DC had median survival times of 455 days as compared to 257 for 42 nonvaccinated but otherwise comparably treated patients in the same institute. Although compared with historical control patients and although no statistics are provided, the prolongation of the median survival is an interesting observation in this phase I clinical trial. The same group reported on vaccinations at the time of relapse in eight GBM patients, and made a similar comparison to 26 historical control patients (Yu et al., 2004). Both patient groups underwent second craniotomy for relapsed GBM. Whereas the median survival was 30 weeks in the control group, subsequent vaccination after new resection resulted in a significant change of median survival to 133 weeks, with some patients surviving for more than 250 weeks.

32 High-Grade Gliomas: Dendritic Cell Therapy

In a recent paper, the same group demonstrated that vaccine responders exhibited more favorable clinical outcomes relative to non-responders. Moreover, the vaccine-induced responses elicited therapeutic benefits primarily by sensitizing tumors to subsequent chemotherapy (Wheeler et al., 2008). In the phase I study reported by Liau et al. (2005), clinical outcome in the patients with stable tumors or no residual disease at time of DC vaccination compared favorably, even when compared with historical/concurrent data for the best prognostic subgroup of GBM patients treated during the same time period at their institute. In contrast, Okada et al. (2007) did not show any benefit in PFS in patients treated with DC loaded with autologous tumor lysate in combination with IL-4 producing fibroblasts. In the study on recurrent HGG patients reported by Yamanaka et al. (2003), radiological responses were observed in the group of five patients who received intratumoral plus intradermal vaccinations, while no response was detected in the five patients who were only injected intradermally. De Vleeschouwer et al. (2008) report on a large group of patients with relapsed GBM treated with DC loaded with tumor lysate in a stepwise process using a cohort-comparison study concept (HGG-IMMUNO2003), comprising of four cohorts (A-D). Following this strategy, the authors have been able to compare each cohort within the trial with the other cohorts. Thus far, three major issues in the complexity of DC immunotherapy have been addressed: the vaccination schedule, the boosting and the maturation cocktail. In this large group of patients, a complete resection improved the median PFS and OS as compared to an incomplete resection. Interestingly, a stepwise improvement in PFS curves could be seen by shortening the interval between the induction DC vaccines, with further improvements noted by adapting the boosting intervals and technology, and finally by changing the maturation cocktail. Reported 2-year survival was 16.7% in cohort C and 27.8% in cohort D. In the report by Ardon et al. (2010b), DC-based immunotherapy was integrated into the primary treatment regimen for eight pilot patients with newly diagnosed GBM. After radiochemotherapy, 4 weekly DC-based vaccines were injected, and further boosting was done with lysate vaccines during the maintenance temozolomide phase. Median PFS in these eight patients was 17.8 months and median OS was 24.3 months. PFS at 6 months was seen in six of the eight

329

patients (75%). The report on the results of a large phase I/II trial HGG-2006 with 77 patients included, is in preparation.

Optimization of Immunotherapy Although several studies have reported promising results, and activation of the immune system against gliomas seems possible, immunotherapeutical approaches are counteracted by rapid tumor progression and glioma-induced immune suppression. There is a clear need for improvement of the immunotherapeutical strategies and several options exist to enhance the immune response. It has been shown in murine models that co-stimulation of T cells can be enhanced by using agonistic antibodies to costimulatory molecules, resulting in a better activation of T cells. The use of antagonistic antibodies to coinhibitory molecules (such as “cytotoxic lymphocyteassociated antigen-4” (CTLA-4) and “programmeddeath-1 receptor”) can also lead to a stronger T cell activation (Grauer et al., 2009; McDermott, 2009). It is also important to target Treg that suppress the antitumoral immune response. As already mentioned, it has been shown in mouse models that elimination of Treg, either in combination with anti-CTLA-4 antibodies or not, results in increased immunity against glioma and a stronger response to DC-based immunotherapy (Grauer et al., 2008; Maes et al., 2009). In humans, studies have been carried out using daclizumab (anti-CD25 antibody), ONTAK (“denileukin diftitox”) and CD25-specific immunotoxins (LMB-2) (Grauer et al., 2009; Rech and Vonderheide, 2009). Clinical responses, however, have been limited. Low-dose metronomic temozolomide diminished Treg in a rodent glioma model and in melanoma patients, and therefore seems a promising tool to use in glioma patients. Unfortunately, this tool cannot be combined with DCbased immunotherapy since metronomic temozolomide administration would counteract the induction of an immune response (Banissi et al., 2009; Su et al., 2004). On the other hand, low-dose metronomic oral cyclophosphamide has been propagated as an effective Treg-depleting regimen that does not counteract immunotherapy (Ghiringhelli et al., 2004, 2007). The local immune suppression by the glioma itself can be targeted via suppressive cytokines, such as

330

TGF-β2, that are produced by the tumor. TGF-β is capable of inhibiting T cell and B cell activation and proliferation, it suppresses the activity of NK cells, reduces the production of cytokines like IL-2, IL-6, IL-10, IFN-γ, and suppresses the expression of MHCII molecules on glioma cells. In preclinical studies it has been shown that the expression of TGF-β2 by GBM cells can be diminished by exposing them to antisense oligonucleotides (AON) against mRNA of TGF-β2 The most advanced AON for the therapy of high-grade gliomas is a phosphorothioate-modified AON, AP 12009 (trabedersen), which targets mRNA encoding TGF-β2. AP 12009 is administered intratumorally using convection-enhanced delivery. A series of Phase I and II clinical trials have evaluated the toxicity profile and optimal dose of the substance, and a randomized, controlled Phase III study is ongoing in patients with recurrent or refractory anaplastic astrocytoma after standard radio- and chemotherapy (Hau et al., 2007, 2009; Schneider et al., 2008). Finally, production of DC can be optimized; for example, maturation of DC can be improved by using “Toll-like receptor” agonists (De Vleeschouwer et al., 2008; Grauer et al., 2009; Van Gool et al., 2009).

Conclusion In recent years insights into tumor immunology have been refined and it is now clear that both the adaptive and innate immune system play an important role. Moreover, it has been shown that immunological processes take place within the CNS and the brain should be looked upon as an immune-distinct environment. Many glioma-specific antigens have been described and interactions between gliomas and the immune system are being elucidated: gliomas will grow and evade the antitumoral immune responses if the balance between tumor cell proliferation and elimination tips in favor of immune-resistant cells (“immune escape”). Immunotherapy for patients with HGG is a novel therapeutic approach that opens new opportunities for enhanced survival without major toxicity. This is of particular importance, taking into account that HGG cause a relatively high community burden, not only with many years of life lost due to cancer,

H. Ardon et al.

but also because of the morbidity from the tumor and subsequent treatments. In the assessment of immunotherapeutical results for patients with HGG, the improvement of PFS and, particularly, the significant increase of OS with satisfactory quality of life are, of course, by far the most important variables, much more than immunologic surrogate markers for response. The studies in preclinical animal models for human gliomas, especially the murine models, are interesting in order to obtain more insight into the basic biology underlying disease evolution and immunotherapeutic mechanisms. Still, timely translation of new concepts into clinical practice should be the primary objective. With several groups entering into clinical practice in this field, new data are now being generated. However, it is of paramount importance to study this promising approach based on data from well designed trials with appropriate end points. Thus far, the observations with regard to both immunological responses and clinical responses are promising and beneficial effects are reproducible. Of particular interest is the fact that no induction of autoimmunity or other major toxicities have been observed to date. However, larger and more homogeneous patient groups with more refined inclusion criteria should be studied for more rapid progress in this field. Moreover, confirmatory studies with an appropriate randomization versus a control patient group, preferably even stratified for prognostic markers, including molecular tumor signature, should be implemented in the near future. For this, collaborative efforts between wellorganized and experienced vaccination centers should be established.

Acknowledgements This translational research program has been supported by the Olivia Hendrickx Research Fund (http://www.olivia.be). Support was also obtained from Electrabel Netmanagement Vlaanderen, CAF Belgium, Baxter, the Herman Memorial Research Fund (http://www.hmrf.be), the James E. Kearney Memorial Fund and gifts from private families and service clubs. Additionally, grants were obtained from “Stichting tegen Kanker,” IWT (TBM project), the Stem Cell Institute Leuven, the Emmanuel van der Schueren Fund, the International Union against Cancer, the Klinisch Onderzoeksfonds UZ Leuven, and the Fund for Scientific Research – Flanders (FWO-V). We are very grateful for the technical assistance from KatjaVandenbrande, Goedele Stegen, Vallentina Schaiko, Elke Nackers and Anaïs Van Hoylandt. We thank the neurooncology team in the hospital for fruitful patient discussion, and the staff of the Laboratory of Experimental Immunology for basic scientific discussions.

32 High-Grade Gliomas: Dendritic Cell Therapy

References Abou-Ghazal M, Yang DS, Qiao W, Reina-Ortiz C, Wei J, Kong LY, Fuller GN, Hiraoka N, Priebe W, Sawaya R, Heimberger AB (2008) The incidence, correlation with tumor-infiltrating inflammation, and prognosis of phosphorylated STAT3 expression in human gliomas. Clin Cancer Res 14:8228–8235 Aoki H, Mizuno M, Natsume A, Tsugawa T, Tsujimura K, Takahashi T, Yoshida J (2001) Dendritic cells pulsed with tumor extract-cationic liposome complex increase the induction of cytotoxic T lymphocytes in mouse brain tumor. Cancer Immunol Immunother 50:463–468 Ardon H, De Vleeschouwer S, Van Calenbergh F, Claes L, Kramm CM, Rutkowski S, Wolff JE, Van Gool SW (2010a) Adjuvant dendritic cell-based tumour vaccination for children with malignant brain tumours. Pediatr. Blood Cancer 54:519–525 Ardon H, Van Gool S, Lopes IS, Maes W, Sciot R, Wilms G, Demaerel P, Bijttebier P, Claes L, Goffin J, Van Calenbergh F, De Vleeschouwer S (2010b) Integration of autologous dendritic cell-based immunotherapy in the primary treatment for patients with newly diagnosed glioblastoma multiforme: a pilot study. J Neurooncol. [Epub ahead of print] Balkwill F (2004) The significance of cancer cell expression of the chemokine receptor CXCR4. Semin Cancer Biol 14: 171–179 Banissi C, Ghiringhelli F, Chen L, Carpentier AF (2009) Treg depletion with a low-dose metronomic temozolomide regimen in a rat glioma model. Cancer Immunol Immunother 58:1627–1634 Bao S, Wu Q, McLendon RE, Hao Y, Shi Q, Hjelmeland AB, Dewhirst MW, Bigner DD, Rich JN (2006) Glioma stem cells promote radioresistance by preferential activation of the DNA damage response. Nature 444:756–760 Borghaei H, Smith MR, Campbell KS (2009) Immunotherapy of cancer. Eur J Pharmacol 625:41–54 Bradbury MW, Cserr HF, Westrop RJ (1981) Drainage of cerebral interstitial fluid into deep cervical lymph of the rabbit. Am J Physiol 240:F329–F336 Bradbury PA, Shepherd FA (2008) Immunotherapy for lung cancer. J Thorac Oncol 3:S164–S170 Choi BD, Archer GE, Mitchell DA, Heimberger AB, McLendon RE, Bigner DD, Sampson JH (2009) EGFRvIII-targeted vaccination therapy of malignant glioma. Brain Pathol 19: 713–723 Ciesielski MJ, Apfel L, Barone TA, Castro CA, Weiss TC, Fenstermaker RA (2006) Antitumor effects of a xenogeneic survivin bone marrow derived dendritic cell vaccine against murine GL261 gliomas. Cancer Immunol Immunother 55:1491–1503 Copier J, Dalgleish AG, Britten CM, Finke LH, Gaudernack G, Gnjatic S, Kallen K, Kiessling R, Schuessler-Lenz M, Singh H, Talmadge J, Zwierzina H, Hakansson L (2009) Improving the efficacy of cancer immunotherapy. Eur J Cancer 45:1424–1431 Cserr HF, Knopf PM (1992) Cervical lymphatics, the blood-brain barrier and the immunoreactivity of the brain: a new view. Immunol Today 13: 507–512

331 De Vleeschouwer S, Fieuws S, Rutkowski S, Van Calenbergh F, Van Loon J, Goffin J, Sciot R, Wilms G, Demaerel P, Warmuth-Metz M, Soerensen N, Wolff JE, Wagner S, Kaempgen E, Van Gool SW (2008) Postoperative adjuvant dendritic cell-based immunotherapy in patients with relapsed glioblastoma multiforme. Clin Cancer Res 14:3098–3104 De Vleeschouwer S, Rapp M, Sorg RV, Steiger HJ, Stummer W, Van Gool S, Sabel M (2006) Dendritic cell vaccination in patients with malignant gliomas: current status and future directions. Neurosurgery 59:988–999 De Vleeschouwer S, Van Calenbergh F, Demaerel P, Flamen P, Rutkowski S, Kaempgen E, Wolff JE, Plets C, Sciot R, Van Gool SW (2004) Transient local response and persistent tumor control of recurrent malignant glioma treated with combination therapy including dendritic cell therapy. J Neurosurg Pediatrics 100:492–497 De Vleeschouwer S, Van Gool SW, Van Calenbergh F (2005) Immunotherapy for malignant gliomas: emphasis on strategies of active specific immunotherapy using autologous dendritic cells. Childs Nerv Syst 21:7–18 Dhodapkar KM, Cirignano B, Chamian F, Zagzag D, Miller DC, Finlay JL, Steinman RM (2004) Invariant natural killer T cells are preserved in patients with glioma and exhibit antitumor lytic activity following dendritic cell-mediated expansion. Int J Cancer 109:893–899 Dix AR, Brooks WH, Roszman TL, Morford LA (1999) Immune defects observed in patients with primary malignant brain tumors. J Neuroimmunol 100:216–232 Dunn GP, Bruce AT, Ikeda H, Old LJ, Schreiber RD (2002) Cancer immunoediting: from immunosurveillance to tumor escape. Nat Immunol 3:991–998 Dunn GP, Dunn IF, Curry WT (2007) Focus on TILs: prognostic significance of tumor infiltrating lymphocytes in human glioma. Cancer Immun 7:12 Dunn GP, Old LJ, Schreiber RD (2004a) The immunobiology of cancer immunosurveillance and immunoediting. Immunity 21:137–148 Dunn GP, Old LJ, Schreiber RD (2004b) The three Es of cancer immunoediting. Annu Rev Immunol 22:329–360 Editorial note Future Oncol (2009) Phase III trial indicates potential for melanoma vaccine. Future Oncol 5:759–761 Elliott LH, Brooks WH, Roszman TL (1987) Activation of immunoregulatory lymphocytes obtained from patients with malignant gliomas. J Neurosurg 67:231–236 Engell-Noerregaard L, Hansen TH, Andersen MH, Thor Straten P, Svane IM (2009) Review of clinical studies on dendritic cell-based vaccination of patients with malignant melanoma: assessment of correlation between clinical response and vaccine parameters. Cancer Immunol Immunother 58:1–14 Eramo A, Ricci-Vitiani L, Zeuner A, Pallini R, Lotti F, Sette G, Pilozzi E, Larocca LM, Peschle C, De Maria R (2006) Chemotherapy resistance of glioblastoma stem cells. Cell Death Differ. 13:1238–1241 Fecci PE, Mitchell DA, Whitesides JF, Xie W, Friedman AH, Archer GE, Herndon JE, Bigner DD, Dranoff G, Sampson JH (2006) Increased regulatory T-cell fraction amidst a diminished CD4 compartment explains cellular immune defects in patients with malignant glioma. Cancer Res 66:3294–3302 Ghiringhelli F, Larmonier N, Schmitt E, Parcellier A, Cathelin D, Garrido C, Chauffert B, Solary E, Bonnotte B, Martin F (2004) CD4+CD25+ regulatory T cells suppress tumor

332 immunity but are sensitive to cyclophosphamide which allows immunotherapy of established tumors to be curative. Eur J Immunol 34:336–344 Ghiringhelli F, Menard C, Puig PE, Ladoire S, Roux S, Martin F, Solary E, Le Cesne A, Zitvogel L, Chauffert B (2007) Metronomic cyclophosphamide regimen selectively depletes CD4+CD25+ regulatory T cells and restores T and NK effector functions in end stage cancer patients. Cancer Immunol Immunother 56:641–648 Gilboa E, Nair SK, Lyerly HK (1998) Immunotherapy of cancer with dendritic-cell-based vaccines. Cancer Immunol Immunother 46:82–87 Grange JM, Krone B, Stanford JL (2009) Immunotherapy for malignant melanoma–tracing Ariadne’s thread through the labyrinth. Eur J Cancer 45:2266–2273 Grauer OM, Nierkens S, Bennink E, Toonen LW, Boon L, Wesseling P, Sutmuller RP, Adema GJ (2007) CD4+FoxP3+ regulatory T cells gradually accumulate in gliomas during tumor growth and efficiently suppress antiglioma immune responses in vivo. Int J Cancer 121:95–105 Grauer OM, Sutmuller RP, van Maren W, Jacobs JF, Bennink E, Toonen LW, Nierkens S, Adema GJ (2008) Elimination of regulatory T cells is essential for an effective vaccination with tumor lysate-pulsed dendritic cells in a murine glioma model. Int J Cancer 122:1794–1802 Grauer OM, Wesseling P, Adema GJ (2009) Immunotherapy of diffuse gliomas: biological background, current status and future developments. Brain Pathol 19:674–693 Hau P, Jachimczak P, Bogdahn U (2009) Treatment of malignant gliomas with TGF-beta2 antisense oligonucleotides. Expert Rev Anticancer Ther 9:1663–1674 Hau P, Jachimczak P, Schlingensiepen R, Schulmeyer F, Jauch T, Steinbrecher A, Brawanski A, Proescholdt M, Schlaier J, Buchroithner J, Pichler J, Wurm G, Mehdorn M, Strege R, Schuierer G, Villarrubia V, Fellner F, Jansen O, Straube T, Nohria V, Goldbrunner M, Kunst M, Schmaus S, Stauder G, Bogdahn U, Schlingensiepen KH (2007) Inhibition of TGF-beta2 with AP 12009 in recurrent malignant gliomas: from preclinical to phase I/II studies. Oligonucleotides 17: 201–212 Heath WR, Carbone FR (2001) Cross-presentation, dendritic cells, tolerance and immunity. Annu Rev Immunol 19:47–64 Heimberger AB, Abou-Ghazal M, Reina-Ortiz C, Yang DS, Sun W, Qiao W, Hiraoka N, Fuller GN (2008) Incidence and prognostic impact of FoxP3+ regulatory T cells in human gliomas. Clin Cancer Res 14:5166–5172 Hemmati HD, Nakano I, Lazareff JA, Masterman-Smith M, Geschwind DH, Bronner-Fraser M, Kornblum HI (2003) Cancerous stem cells can arise from pediatric brain tumors. Proc Natl Acad Sci USA 100:15178–15183 Insug O, Ku G, Ertl HC, Blaszczyk-Thurin M (2002) A dendritic cell vaccine induces protective immunity to intracranial growth of glioma. Anticancer Res 22:613–621 Kakimi K, Nakajima J, Wada H (2009) Active specific immunotherapy and cell-transfer therapy for the treatment of non-small cell lung cancer. Lung Cancer 65:1–8 Kikuchi T, Akasaki Y, Abe T, Fukuda T, Saotome H, Ryan JL, Kufe DW, Ohno T (2004) Vaccination of glioma patients with fusions of dendritic and glioma cells and recombinant human interleukin 12. J Immunother 27: 452–459

H. Ardon et al. Kikuchi T, Akasaki Y, Irie M, Homma S, Abe T, Ohno T (2001) Results of a phase I clinical trial of vaccination of glioma patients with fusions of dendritic and glioma cells. Cancer Immunol Immunother 50:337–344 Kjaergaard J, Wang LX, Kuriyama H, Shu S, Plautz GE (2005) Active immunotherapy for advanced intracranial murine tumors by using dendritic cell-tumor cell fusion vaccines. J Neurosurg 103:156–164 Liau LM, Black KL, Martin NA, Sykes SN, Bronstein JM, Jouben-Steele L, Mischel PS, Belldegrun A, Cloughesy TF (2000) Treatment of a Glioblastoma patient by vaccination with autologous dendritic cells pulsed with allogeneic Major Histocompatibility Complex Class I-matched tumor peptides: case report. Neurosurgical Focus 9:e8 Liau LM, Prins RM, Kiertscher SM, Odesa SK, Kremen TJ, Giovannone AJ, Lin JW, Chute DJ, Mischel PS, Cloughesy TF, Roth MD (2005) Dendritic cell vaccination in glioblastoma patients induces systemic and intracranial T-cell responses modulated by the local central nervous system tumor microenvironment. Clin Cancer Res 11:5515–5525 Liu G, Yuan X, Zeng Z, Tunici P, Ng H, Abdulkadir IR, Lu L, Irvin D, Black KL, Yu JS (2006) Analysis of gene expression and chemoresistance of CD133+ cancer stem cells in glioblastoma. Mol Cancer 5:67 Macdonald DR, Cascino TL, Schold SC Jr, Cairncross JG (1990) Response criteria for phase II studies of supratentorial malignant glioma. J Clin Oncol 8:1277–1280 Maes W, Rosas GG, Verbinnen B, Boon L, De Vleeschouwer S, Ceuppens JL, Van Gool SW (2009) DC vaccination with anti-CD25 treatment leads to long-term immunity against experimental glioma. Neuro Oncol 11:529–542 Maldonado-Lopez R, Moser M (2001) Dendritic cell subsets and the regulation of Th1/Th2 responses. Semin Immunol 13:275–282 Malmberg KJ (2004) Effective immunotherapy against cancer: a question of overcoming immune suppression and immune escape? Cancer Immunol Immunother 53:879–892 McDermott DF (2009) Immunotherapy of metastatic renal cell carcinoma. Cancer 115:2298–2305 McMahon EJ, Bailey SL, Miller SD (2006) CNS dendritic cells: critical participants in CNS inflammation? . Neurochem Int 49:195–203 Mu LJ, Gaudernack G, Saeboe-Larssen S, Hammerstad H, Tierens A, Kvalheim G (2003) A protocol for generation of clinical grade mRNA-transfected monocyte-derived dendritic cells for cancer vaccines. Scand J Immunol 58:578–586 Ni HT, Spellman SR, Jean WC, Hall WA, Low WC (2001) Immunization with dendritic cells pulsed with tumor extract increases survival of mice bearing intracranial gliomas. J Neurooncol 51:1–9 Norden AD, Drappatz J, Wen PY (2009) Antiangiogenic therapies for high-grade glioma. Nat Rev Neurol 5:610–620 Okada H, Lieberman FS, Walter KA, Lunsford LD, Kondziolka DS, Bejjani GK, Hamilton RL, Torres-Trejo A, Kalinski P, Cai Q, Mabold JL, Edington HD, Butterfield LH, Whiteside TL, Potter DM, Schold SC Jr, Pollack IF (2007) Autologous glioma cell vaccine admixed with interleukin-4 gene transfected fibroblasts in the treatment of patients with malignant gliomas. J Transl Med 5:67 Pellegatta S, Poliani PL, Corno D, Grisoli M, Cusimano M, Ubiali F, Baggi F, Bruzzone MG, Finocchiaro G

32 High-Grade Gliomas: Dendritic Cell Therapy (2006a) Dendritic cells pulsed with glioma lysates induce immunity against syngeneic intracranial gliomas and increase survival of tumor-bearing mice. Neurol Res 28: 527–531 Pellegatta S, Poliani PL, Corno D, Menghi F, Ghielmetti F, Suarez-Merino B, Caldera V, Nava S, Ravanini M, Facchetti F, Bruzzone MG, Finocchiaro G (2006b) Neurospheres enriched in cancer stem-like cells are highly effective in eliciting a dendritic cell-mediated immune response against malignant gliomas. Cancer Res 66:10247–10252 Ransohoff RM, Kivisakk P, Kidd G (2003) Three or more routes for leukocyte migration into the central nervous system. Nat Rev Immunol 3:569–581 Rech AJ, Vonderheide RH (2009) Clinical use of anti-CD25 antibody daclizumab to enhance immune responses to tumor antigen vaccination by targeting regulatory T cells. Ann NY Acad Sci 1174:99–106 Roszman T, Elliott L, Brooks W (1991) Modulation of T-cell function by gliomas. Immunol Today 12:370–374 Rutkowski S, De Vleeschouwer S, Kaempgen E, Wolff JEA, Kuhl J, Demaerel P, Warmuth-Metz M, Flamen P, Van Calenbergh F, Plets C, Sörensen N, Opitz A, Van Gool SW (2004) Surgery and adjuvant dendritic cell-based tumour vaccination for patients with relapsed malignant glioma, a feasibility study. Br J Cancer 91:1656–1662 Saito R, Mizuno M, Nakahara N, Tsuno T, Kumabe T, Yoshimoto T, Yoshida J (2004) Vaccination with tumor cell lysate-pulsed dendritic cells augments the effect of IFN-beta gene therapy for malignant glioma in an experimental mouse intracranial glioma. Int J Cancer 111:777–782 Schadendorf D, Ugurel S, Schuler-Thurner B, Nestle FO, Enk A, Brocker EB, Grabbe S, Rittgen W, Edler L, Sucker A, Zimpfer-Rechner C, Berger T, Kamarashev J, Burg G, Jonuleit H, Tuttenberg A, Becker JC, Keikavoussi P, Kämpgen E, Schuler G (2006) Dacarbazine (DTIC) versus vaccination with autologous peptide-pulsed dendritic cells (DC) in first-line treatment of patients with metastatic melanoma: a randomized phase III trial of the DC study group of the DeCOG. Ann Oncol 17:563–570 Schneider T, Becker A, Ringe K, Reinhold A, Firsching R, Sabel BA (2008) Brain tumor therapy by combined vaccination and antisense oligonucleotide delivery with nanoparticles. J Neuroimmunol 195:21–27 Singh SK, Clarke ID, Terasaki M, Bonn VE, Hawkins C, Squire J, Dirks PB (2003) Identification of a cancer stem cell in human brain tumors. Cancer Res 63:5821–5828 Singh SK, Hawkins C, Clarke ID, Squire JA, Bayani J, Hide T, Henkelman RM, Cusimano MD, Dirks PB (2004) Identification of human brain tumour initiating cells. Nature 432:396–401 Sorg RV, Ozcan Z, Brefort T, Fischer J, Ackermann R, Muller M, Wernet P (2003) Clinical-scale generation of dendritic cells in a closed system. J Immunother 26:374–383 Su YB, Sohn S, Krown SE, Livingston PO, Wolchok JD, Quinn C, Williams L, Foster T, Sepkowitz KA, Chapman PB (2004) Selective CD4+ lymphopenia in melanoma patients treated with temozolomide: a toxicity with therapeutic implications. J Clin Oncol 22:610–616 Terando AM, Faries MB, Morton DL (2007) Vaccine therapy for melanoma: current status and future directions. Vaccine 25(Suppl 2):B4–B16

333 Thurner B, Roder C, Dieckmann D, Heuer H, Kruse M, Glaser A, Keikavoussi P, Kämpgen E, Bender A, Schuler G (1999) Generation of large numbers of fully mature and stable dendritic cells from leukapheresis products for clinical application. J Immunol Methods 223:1–15 Tuyaerts S, Aerts JL, Corthals J, Neyns B, Heirman C, Breckpot K, Thielemans K, Bonehill A (2007) Current approaches in dendritic cell generation and future implications for cancer immunotherapy. Cancer Immunol Immunother 56: 1513–1537 Tuyaerts S, Noppe SM, Corthals J, Breckpot K, Heirman C, De Greef C, Van Riet I, Thielemans K (2002) Generation of large numbers of dendritic cells in a closed system using Cell Factories. J Immunol Methods 264:135–151 Van Gool S, Maes W, Ardon H, Verschuere T, Van Cauter S, De Vleeschouwer S (2009) Dendritic cell therapy of high-grade gliomas. Brain Pathol 19:694–712 Van Poppel H, Joniau S, Van Gool SW (2009) Vaccine therapy in patients with renal cell carcinoma. Eur Urol 55: 1333–1342 Verhoeff JJ, van Tellingen O, Claes A, Stalpers LJ, van Linde ME, Richel DJ, Leenders WP, van Furth WR (2009) Concerns about anti-angiogenic treatment in patients with glioblastoma multiforme. BMC Cancer 9:444 Walker DG, Laherty R, Tomlinson FH, Chuah T, Schmidt C (2008) Results of a phase I dendritic cell vaccine trial for malignant astrocytoma: potential interaction with adjuvant chemotherapy. J Clin Neurosci 15:114–121 Wheeler CJ, Black KL, Liu G, Mazer M, Zhang XX, Pepkowitz S, Goldfinger D, Ng H, Irvin D, Yu JS (2008) Vaccination elicits correlated immune and clinical responses in glioblastoma multiforme patients. Cancer Res 68:5955–5964 Wheeler CJ, Black KL, Liu G, Ying H, Yu JS, Zhang W, Lee PK (2003) Thymic CD8(+) T cell production strongly influences tumor antigen recognition and age-dependent glioma mortality. J Immunol 171:4927–4933 Yamanaka R (2009) Dendritic-cell- and peptide-based vaccination strategies for glioma. Neurosurg Rev 32:265–273 Yamanaka R, Abe T, Yajima N, Tsuchiya N, Homma J, Kobayashi T, Narita M, Takahashi M, Tanaka R (2003) Vaccination of recurrent glioma patients with tumour lysatepulsed dendritic cells elicits immune responses: results of a clinical phase I/II trial. Br J Cancer 89:1172–1179 Yamanaka R, Homma J, Yajima N, Tsuchiya N, Sano M, Kobayashi T, Yoshida S, Abe T, Narita M, Takahashi M, Tanaka R (2005) Clinical evaluation of dendritic cell vaccination for patients with recurrent glioma: results of a clinical phase I/II trial. Clin Cancer Res 11:4160–4167 Yu JS, Liu G, Ying H, Yong WH, Black KL, Wheeler CJ (2004) Vaccination with tumor lysate-pulsed dendritic cells elicits antigen-specific, cytotoxic T-cells in patients with malignant glioma. Cancer Res 64:4973–4979 Yu JS, Wheeler CJ, Zeltzer PM, Ying H, Finger DN, Lee PK, Yong WH, Incardona F, Thompson RC, Riedinger MS, Zhang W, Prins RM, Black KL (2001) Vaccination of malignant glioma patients with peptide-pulsed dendritic cells elicits systemic cytotoxicity and intracranial T-cell infiltration. Cancer Res 61:842–847 Zitvogel L, Tesniere A, Kroemer G (2006) Cancer despite immunosurveillance: immunoselection and immunosubversion. Nat Rev Immunol 6:715–727

Chapter 33

Glioblastoma Multiforme: Use of Adenoviral Vectors Thomas Wirth, Haritha Samaranayake, and Seppo Ylä-Herttuala

Abstract Glioblastoma multiforme is a tumour of the brain with extremely poor prognosis. It constitutes approximately 60–70% of all primary brain tumours with a mean survival of 14.6 months after diagnosis. Current standard therapy includes surgical resection followed by adjuvant radiotherapy and/or chemotherapy. With the introduction of gene therapy a new therapeutic area has been opened and numerous approaches are currently been studied in pre-clinical as well as in clinical settings for the treatment of brain tumours. These include pro-drug activation/suicide gene therapy, anti-angiogenic gene therapy, oncolytic virotherapy, immune modulation and many other strategies, which will be discussed in this review. The use of suicide gene therapy has been probably one of the most studied therapeutic approaches and will be also discussed in this review in more detail. Keywords Glioblastoma multiforme · Adenoviral vector · Pro-drug activation · Proliferation · Invasion · Suicide gene therapy

Introduction One of the most lethal malignancies are gliomas, which constitute approximately 60–70% of all primary brain tumours with an overall global incidence

T. Wirth () Department of Biotechnology and Molecular Medicine, A. I. Virtanen Institute, University of Kuopio, FI-70211 Kuopio, Finland e-mail: [email protected]

of 4–6 new cases per 100,000 population per year. Current standard therapy includes surgical resection followed by adjuvant radiotherapy and/or chemotherapy. The overall prognosis of patients diagnosed with glioblastoma multiforme (GBM) and treated with current therapies has remained poor for the last few decades, with a mean survival of 14.6 months after diagnosis and a 5-year survival of 95% efficiency and also observed cross-inhibiton at around 50%. Combined knock down resulted in tumors with hardly detectable amounts of VEGF and IL6. On a morphological level, combined inhibition of IL6 and VEGF is significantly more efficient than single inhibition as evidenced by reduced size, proliferation and neovascularization of the tumors. This comparative analysis can be easily led on significant numbers of tumors (n>20 per group) providing robust quantitative data. On the left panel of Fig. 35.3, we show an example of vascular changes to different treatments. In anti-IL6 or anti-VEGF conditions alone a similar reduction in tumor angiogenesis can be observed, compared to control tumors. Combinatory treatment completely inhibits vascularization of the tumors. Interestingly, the experimental glioma allows investigating invasive properties that may be enhanced under anti-angiogenic treatment, a feature that neurooncologists are particularly sensitized to when treating glioma patients with Avastin (de Groot et al., 2010). In control glioma, tumor cells are rarely detected at distance from the primary implantation site in the supporting CAM. Under metabolic/hypoxic stress induced by IL6 or VEGF knockdown, tumor cells invade the CAM tissue (Fig. 35.3 right panels), with a 3–4 fold increase in the number of tumor cells detected in the CAM. Different modes of invasion can be observed. In the anti-VEGF condition, cells invade in a collective manner, whereas IL6 knock down mostly induces single-cell spreading. In combined treated tumors, the invasive phenotype was significantly reduced, to a similar degree as seen in controls. These results have been confirmed in an orthotopic mouse glioma model. The experimental glioma is a fast, cost-effective alternative to screen different combination of inhibitors, including siRNAs for their effect on standard tumor parameters such as angiogenesis, but also recapitulates biological phenomena observed in mammalian brains such as induction of infiltration after treatment with single angiogenesis inhibitors.

35 Cellular and Molecular Characterization of Anti-VEGF and IL-6 Therapy

367

Fig. 35.3 Left panel: representative illustrations of angiogenic and invasive phenotypes revealed by triple immunofluorescence staining (blue: DAPI; red: anti-vimentin; green: SNA-lectin; scale bar left panels = 50 μm; right panels = 100 μm). Note that control tumors are characterized by the presence of a dense immature vascular network. Knockdown of IL6 and/or VEGF strongly decreases tumor vascularization and combined knockdown is even more efficient at reducing vascular density. Right

panel: invasive behavior of tumor cells in the CAM. Sections of tumor and surrounding CAM tissue 7 days after implantation were stained with DAPI (blue) and anti-vimentin (red). A representative field of the margin between tumor (asterisk) and the CAM (dashed line) is shown. Both, anti-IL6 and antiVEGF cause increased invasion in the CAM stroma, whereas IL6/VEGF combined knockdown cells barely invade the CAM

Exploring Molecular Pathways Associated with Inhibition

tissue than murine models and cross hybridization between chicken and human sequences is negligible (see Fig. 35.1c). Tumors with or without treatment (single or double knock down) harbored a specific gene signature. As the phenotypes of anti-VEGF and antiIL6 treated tumors are very similar, with an equivalent reduction in tumor volume and vascularization, and

As detailed previously, the CAM experimental glioma is well suited to perform comparative gene profiling as it allows more precise isolation of tumor

368

the two induced pathways are partially overlapping, highlighting differences at the level of the molecular responses may be predictive for a better therapeutic outcome. For instance there were twice as many genes down regulated in anti-IL6 tumors compared to antiVEGF tumors. In addition and as anticipated from the small volumes of the tumors, we found out that the cell cycle machinery is severely affected when both VEGF and IL6 pathways are inhibited. Around 25 positive regulators of mitosis and 10 major components of chromatin assembly (mainly histone subunits) are significantly down regulated in the combined treatment. Interestingly, many genes involved in invasion were over expressed in tumors after single inhibition, but were strongly inhibited when VEGF and IL6 were inhibited. Amongst those factors, CXCL12/SDF1 (Tu et al., 2009) and PDPLN (Wicki and Christofori, 2007) merit further investigation as potential markers of poor prognosis and inducers of invasive behavior glioma cells. We showed that the experimental glioma on the CAM allows to predict the efficiency (reduction of tumor growth) and adverse effect (induction of invasiveness) of anti-VEGF and anti-IL6 pathways inhibitors. Humanized antibodies specifically raised against Interleukin-6 receptor are available and used in the clinic for rheumatoid arthritis (Tocilizumab, Actemra, Roche) but no trials have yet been engaged to estimate anti-glioma effects of this new drug. We are currently investigating effects on tumor progression and infiltration of a combination of Avastin and Tocilizumab in mouse glioma models.

Future Directions The experimental glioma model is a useful tool to predict treatment efficacy of new anti-angiogenic drugs and to validate new targets. An advantage of the model is fast development of tumors and the possibility to visually evaluate treatment efficacy. There are a growing number of humanized antibodies, aptamers and small molecules inhibitors of pro-angiogenic key factors. It will therefore be of interest to test different combinations of these drugs to find a treatment paradigm that efficiently blocks tumor vascularization and invasion at the lowest dose possible. This combination is likely to give similar results in other

S. Javerzat and M. Hagedorn

pre-clinical models and ultimately may be of benefit for glioblastoma patients. The experimental glioma model on the CAM might considerably speed up the search for efficient anti-glioma treatments. Acknowledgements This work was supported La Ligue Contre le Cancer, Comitée de la Dordogne (to SJ) and l’Agence Nationale de la Recherche, ANR (“Glioma Model”, JC05_0060, to MH).

References Casanovas O, Hicklin DJ, Bergers G, Hanahan D (2005) Drug resistance by evasion of antiangiogenic targeting of VEGF signaling in late-stage pancreatic islet tumors. Cancer Cell 8:299–309 Chang CY, Li MC, Liao SL, Huang YL, Shen CC, Pan HC (2005) Prognostic and clinical implication of IL-6 expression in glioblastoma multiforme. J Clin Neurosci 12:930–933 Cheng SY, Huang HJ, Nagane M, Ji XD, Wang D, Shih CC, Arap W, Huang CM, Cavenee WK (1996) Suppression of glioblastoma angiogenicity and tumorigenicity by inhibition of endogenous expression of vascular endothelial growth factor. Proc Natl Acad Sci USA 93:8502–8507 Clauss A, Ng V, Liu J, Piao H, Russo M, Vena N, Sheng Q, Hirsch MS, Bonome T, Matulonis U, Ligon AH, Birrer MJ, Drapkin R (2010) Overexpression of elafin in ovarian carcinoma is driven by genomic gains and activation of the nuclear factor kappaB pathway and is associated with poor overall survival. Neoplasia 12:161–172 Dai C, Holland EC (2001) Glioma models. Biochim Biophys Acta 1551:M19–M27 de Groot JF, Fuller G, Kumar AJ, Piao Y, Eterovic K, Ji Y, Conrad CA (2010) Tumor invasion after treatment of glioblastoma with bevacizumab: radiographic and pathologic correlation in humans and mice. Neuro Oncol 12:233–242 Dumartin L, Quemener C, Laklai H, Herbert J, Bicknell R, Bousquet C, Pyronnet S, Castronovo V, Schilling MK, Bikfalvi A, Hagedorn M (2010) Netrin-1 mediates early events in pancreatic adenocarcinoma progression, acting on tumor and endothelial cells. Gastroenterology 138: 1595–1606 Ferrara N (2010) Pathways mediating VEGF-independent tumor angiogenesis. Cytokine Growth Factor Rev 21:21–26 Hagedorn M, Javerzat S, Gilges D, Meyre A, de Lafarge B, Eichmann A, Bikfalvi A (2005) Accessing key steps of human tumor progression in vivo by using an avian embryo model. Proc Natl Acad Sci USA 102:1643–1648 Hanahan D, Weinberg RA (2000) The hallmarks of cancer. Cell 100:57–70 Holland EC, Celestino J, Dai C, Schaefer L, Sawaya RE, Fuller GN (2000) Combined activation of Ras and Akt in neural progenitors induces glioblastoma formation in mice. Nat Genet 25:55–57 Huang J, Frischer JS, Serur A, Kadenhe A, Yokoi A, McCrudden KW, New T, O’Toole K, Zabski S, Rudge JS, Holash

35 Cellular and Molecular Characterization of Anti-VEGF and IL-6 Therapy J, Yancopoulos GD, Yamashiro DJ, Kandel JJ (2003) Regression of established tumors and metastases by potent vascular endothelial growth factor blockade. Proc Natl Acad Sci USA 100:7785–7790 Inoue M, Hager JH, Ferrara N, Gerber HP, Hanahan D (2002) VEGF-A has a critical, nonredundant role in angiogenic switching and pancreatic beta cell carcinogenesis. Cancer Cell 1:193–202 Jendreyko N, Popkov M, Rader C, Barbas CF 3rd (2005) Phenotypic knockout of VEGF-R2 and Tie-2 with an intradiabody reduces tumor growth and angiogenesis in vivo. Proc Natl Acad Sci USA 102:8293–8298 Jensen RL (2009) Brain tumor hypoxia: tumorigenesis, angiogenesis, imaging, pseudoprogression, and as a therapeutic target. J Neurooncol 92:317–335 Knisely JP, Rockwell S (2002) Importance of hypoxia in the biology and treatment of brain tumors. Neuroimaging Clin N Am 12:525–536 Kunkel P, Ulbricht U, Bohlen P, Brockmann MA, Fillbrandt R, Stavrou D, Westphal M, Lamszus K (2001) Inhibition of glioma angiogenesis and growth in vivo by systemic treatment with a monoclonal antibody against vascular endothelial growth factor receptor-2. Cancer Res 61: 6624–6628 Lamour V, Le Mercier M, Lefranc F, Hagedorn M, Javerzat S, Bikfalvi A, Kiss R, Castronovo V, Bellahcene A (2010) Selective osteopontin knockdown exerts anti-tumoral activity in a human glioblastoma model. Int J Cancer 126: 1797–1805 Loeffler S, Fayard B, Weis J, Weissenberger J (2005) Interleukin-6 induces transcriptional activation of vascular endothelial growth factor (VEGF) in astrocytes in vivo and regulates VEGF promoter activity in glioblastoma cells via direct interaction between STAT3 and Sp1. Int J Cancer 115:202–213 Mottet D, Pirotte S, Lamour V, Hagedorn M, Javerzat S, Bikfalvi A, Bellahcene A, Verdin E, Castronovo V (2009) HDAC4 represses p21(WAF1/Cip1) expression in human cancer cells through a Sp1-dependent, p53-independent mechanism. Oncogene 28:243–256 Niola F, Evangelisti C, Campagnolo L, Massalini S, Bue MC, Mangiola A, Masotti A, Maira G, Farace MG, Ciafre SA (2006) A plasmid-encoded VEGF siRNA reduces glioblastoma angiogenesis and its combination with interleukin-4 blocks tumor growth in a xenograft mouse model. Cancer Biol Ther 5:174–179

369

Pelloski CE, Mahajan A, Maor M, Chang EL, Woo S, Gilbert M, Colman H, Yang H, Ledoux A, Blair H, Passe S, Jenkins RB, Aldape KD (2005) YKL-40 expression is associated with poorer response to radiation and shorter overall survival in glioblastoma. Clin Cancer Res 11:3326–3334 Reilly KM, Loisel DA, Bronson RT, McLaughlin ME, Jacks T (2000) Nf1; Trp53 mutant mice develop glioblastoma with evidence of strain-specific effects. Nat Genet 26:109–113 Saidi A, Hagedorn M, Allain N, Verpelli C, Sala C, Bello L, Bikfalvi A, Javerzat S (2009) Combined targeting of interleukin-6 and vascular endothelial growth factor potently inhibits glioma growth and invasiveness. Int J Cancer 125:1054–1064 Saidi A, Javerzat S, Bellahcene A, De Vos J, Bello L, Castronovo V, Deprez M, Loiseau H, Bikfalvi A, Hagedorn M (2008) Experimental anti-angiogenesis causes upregulation of genes associated with poor survival in glioblastoma. Int J Cancer 122:2187–2198 Strieth S, Eichhorn ME, Sutter A, Jonczyk A, Berghaus A, Dellian M (2006) Antiangiogenic combination tumor therapy blocking alpha(v)-integrins and VEGF-receptor-2 increases therapeutic effects in vivo. Int J Cancer 119:423–431 Tchirkov A, Rolhion C, Bertrand S, Dore JF, Dubost JJ, Verrelle P (2001) IL-6 gene amplification and expression in human glioblastomas. Br J Cancer 85:518–522 Tu H, Zhou Z, Liang Q, Li Z, Li D, Qing J, Wang H, Zhang L (2009) CXCR4 and SDF-1 production are stimulated by hepatocyte growth factor and promote glioma cell invasion. Onkologie 32:331–336 Van Meir E, Sawamura Y, Diserens AC, Hamou MF, de Tribolet N (1990) Human glioblastoma cells release interleukin 6 in vivo and in vitro. Cancer Res 50:6683–6688 Weissenberger J, Loeffler S, Kappeler A, Kopf M, Lukes A, Afanasieva TA, Aguzzi A, Weis J (2004) IL-6 is required for glioma development in a mouse model. Oncogene 23: 3308–3316 Wicki A, Christofori G (2007) The potential role of podoplanin in tumour invasion. Br J Cancer 96:1–5 Willett CG, Boucher Y, di Tomaso E, Duda DG, Munn LL, Tong RT, Chung DC, Sahani DV, Kalva SP, Kozin SV, Mino M, Cohen KS, Scadden DT, Hartford AC, Fischman AJ, Clark JW, Ryan DP, Zhu AX, Blaszkowsky LS, Chen HX, Shellito PC, Lauwers GY, Jain RK (2004) Direct evidence that the VEGF-specific antibody bevacizumab has antivascular effects in human rectal cancer. Nat Med 10:145–147

Chapter 36

Adult Brainstem Gliomas: Diagnosis and Treatment Florence Laigle-Donadey and Jean-Yves Delattre

Abstract Adult brainstem gliomas constitute a heterogeneous group of tumors. A better radiological analysis of these tumors will improve their classification and help to better distinguish prognosis subgroups. New MRI techniques may also contribute to differential diagnosis and help neurosurgeons in removing resectable brainstem tumors. However, biopsy remains indicated in many contrast enhancing brainstem masses in adults because of the great variety of differential diagnosis. Conventional radiotherapy is the standard of care of brainstem glioma and chemotherapy has been disappointing to date. Given the lack of efficacy of conventional drugs, a better understanding of the biology of this tumor is the key to develop targeted therapy. In the future, advances in diagnostic and treatment modalities will probably result in improvement in the pattern of care of brainstem gliomas which remain now associated with a poor prognosis. Keywords Brainstem gliomas · MRI techniques · Differential diagnosis · Exophytic gliomas · Drugs · Radiotherapy

F. Laigle-Donadey () Service de Neurologie Mazarin, Hôpital de la Pitié-Salpêtrière – APHP, 75651 Paris Cedex 13, France e-mail: [email protected]

Introduction Brainstem tumors occur in a region located between the mesencephalon and the medulla. This definition excludes patients with tumors originating in the thalamus and hypothalamus, or lesions originating from the cerebellum, cerebellar peduncles, upper cervical spinal cord, as well as tumors arising from ventricles. “Brainstem glioma” is the most frequent tumor of the region but it constitutes a heterogeneous group of tumors with variable prognoses. A bimodal age distribution has been noted, allowing a clear distinction between BS gliomas in children and adults that differ in their clinical and radiological presentation, histology, biological behaviour, and clinical course. This chapter will focus on adult BS gliomas which is a rare disease. Indeed, in contrast with in children, where brainstem tumors are frequent, accounting for approximately 10–20% of all pediatric CNS tumors, brainstem gliomas are much less frequent in adults, accounting for 2–2.5% of all intracranial tumors in adults (White, 1963). As for children, tissue confirmation is frequently not feasible unless an exophytic component exists, because of the risks associated with direct histologic examination (Packer et al., 1985) and their classification relies mainly on clinical presentation and MRI characteristics (Guillamo et al., 2001a). A striking difference between adults and children is the better prognosis of the “diffuse intrinsic form” in adults (Landolfi et al., 1998; Selvapandian et al., 1999; Guillamo et al., 2001a).

M.A. Hayat (ed.), Tumors of the Central Nervous System, Volume 2, DOI 10.1007/978-94-007-0618-7_36, © Springer Science+Business Media B.V. 2011

371

372

“Diffuse Intrinsic” or “Diffusely Infiltrative Low Grade” Brainstem Gliomas Epidemiology Diffuse intrinsic brainstem gliomas represent the most frequent type of brainstem glioma in adults, accounting for 45–50% of cases (Guillamo et al., 2001a; Salmaggi et al., 2008). These tumors usually occur in young patients, between the age of 20 and 30 years.

Pathology Biopsy is rarely performed in typical intrinsic diffuse low-grade glioma in adults. When a biopsy is performed, a benign histology (grade II glioma) is found in up to 82% of cases in adults, in contrast with children (Guillamo et al., 2001a, b) which is a likely explanation for the better survival of the diffuse intrinsic form in adults (Guillamo et al., 2001a; Salmaggi et al., 2008). For this reason, Guillamo et al. suggest that this subgroup of tumor should be named “diffuse intrinsic low-grade brainstem gliomas” in adults. However, adult patients can rarely present a rapidly growing tumor similar to diffuse intrinsic BSG of children (Guillamo et al., 2001a).

F. Laigle-Donadey and J.-Y. Delattre

pseudomyasthenic presentation (Dirr et al., 1989), as well as orthostatic hypotension (Hsu et al., 1984).

Radiological Presentation MRI reveals a pattern similar to the pattern observed in children with a diffuse enlargement of the brainstem, with high signal on T2-weighted images and low signal on T1-weighted images, located in the pons (70%) or in the medulla (30%). The lesion does not enhance after contrast infusion (Freeman and Farmer, 1998; Guillamo et al., 2001b). MRS showed an elevation of the Cho/NAA ratio at diagnosis in all nine patients from a cohort that was investigated with MRS (Fig. 36.1) (Salmaggi et al., 2008).

Treatment Neurosurgery Biopsy is rarely performed in typical intrinsic diffuse low-grade glioma in adults (Guillamo et al., 2001b). However, this remains a matter of debate. Some authors consider that guided stereotactic biopsy

Clinical Presentation Despite this histological difference, there are no major differences in clinical presentation between adults and children (Selvapandian et al., 1999), except for a longer duration of symptoms and signs in adults, (Landolfi et al., 1998; Selvapandian et al., 1999). Patients typically present with a long-lasting history of facial palsy. Facial myokymia (Selvapandian et al., 1999), as well as hemifacial spasm (Elgamal and Coakham, 2005) may occur. Other symptoms rarely described include snoring associated with Ondine’s curse in the case of medulla oblongata glioma (Marin-Sanabria et al., 2006), central neurogenic hyperventilation (Gaviani et al., 2005), tongue tremor (Saka et al., 2006), visual auras and migraine (Lim et al., 2005),

Fig. 36.1 Axial flair MR in a 35 year-old patient presenting with diplopia. Note diffuse swelling and increased signal intensity due to intrinsic pontine glioma

36 Adult Brainstem Gliomas: Diagnosis and Treatment

is relatively “safe” and believe that an accurate diagnosis should be obtained in all cases and not only when atypical clinical or imaging features suggesting a non-neoplastic lesion (Thomas et al., 1988). Resection of intrinsic diffuse brainstem tumors remains impossible despite technical advances (Tanaka et al., 2005; Mursch et al., 2005). A stereotaxic approach can be a rapid and safe method for evacuation of the contents of cysts (Thomas et al., 1988; Rajshekhar and Chandy, 1995), providing neurological benefit in most cases. A stereotactic placement of an Ommaya reservoir after reaccumulation of the cyst can also be performed. Patients with hydrocephalus may require ventriculostomy or ventriculoperitoneal shunting for symptomatic relief. Radiotherapy Focal radiotherapy is the standard treatment for adult brainstem gliomas and can improve or stabilize patients for years. The conventional dose of radiotherapy ranges from 54 to 60 Gy. Diffuse intrinsic brainstem glioma seems to be more responsive to radiotherapy in adults than in children. However there is an important discrepancy between the clinical and the radiological response. In one series, a majority of patients (62%) clinically improved for a “long” period (over 6 months) after radiotherapy while radiological response was noted in only 19% of cases (Guillamo et al., 2001a). The optimal date of treatment remains unknown. Indeed, some patients may have mild symptoms during extended periods without any treatment, suggesting that radiotherapy may sometimes be deferred until clear evidence of symptomatic tumor progression (Guillamo et al., 2001a).

373

radiological result in one case report of progressive diffuse BS glioma treated with bevacizumab and irinotecan (Torcuator et al., 2009). Symptomatic Treatment Patients with difficulties in swallowing may need gastrostomy, such as percutaneous esophagogastrostomy. Steroids are frequently administered to brainstem tumor patients. However, they should be used sparingly, at the lowest necessary dosage, because of their numerous side effects.

Follow-up The frequently indolent growth pattern of the intrinsic form in adults contrasts with the natural history of this disease in children and is likely related to a lower grade of the tumors. Median survival of diffuse intrinsic lowgrade gliomas in adults is 7.3 years, as compared to less than 1 year in children. Two recent retrospective studies have evaluated the characteristics from adult BSGs. In an American cohort including 101 adult patients, the overall survival at 5 and 10 years was 58 and 41%, respectively, with a median survival of 85 months (range 1–228) (Kesari et al., 2008). In an Italian cohort of 34 adult patients, the median overall survival was 59 months (Salmaggi et al., 2008). However, it is worth noting that rarely some adult patients can present with a rapidly growing tumor similar to children with diffuse intrinsic brainstem glioma (Guillamo et al., 2001a).

Malignant Brainstem Gliomas Epidemiology

Chemotherapy Efficacy of chemotherapy in adult brainstem gliomas remains unproven and adjuvant chemotherapy cannot be currently recommended. Moreover, the effectiveness of chemotherapy on relapse is uncertain, although it may benefit some patients, sometimes with a frank clinical improvement contrasting with a lack of radiological response. The role of new therapies such as anti-angiogenic drugs need to be defined in this case. Torcuator et al. reported an interesting clinical and

Malignant brainstem glioma accounts for 31–39% of adults with BSG (Guillamo et al., 2001a; Salmaggi et al., 2008) and appear in older patients, usually after the age of 40.

Pathology Histological examination shows high-grade (III, IV) gliomas in all cases.

374

Clinical Presentation The clinical history is subacute, consisting of cranial palsies and long tract signs, rapidly leading to an altered performance status.

F. Laigle-Donadey and J.-Y. Delattre

Careful systemic and neurologic work up usually leads to the correct diagnosis. However, a biopsy should be considered in many contrast enhancing brainstem masses when this work up is not conclusive (Abernathey et al., 1989; Rajshekhar and Chandy, 1995).

Radiological Presentation Treatment Contrast enhancement was found in 100% of cases upon initial MRI in Guillamo’s series, and often exhibited a ring-like pattern (Guillamo et al., 2001a), suggesting central necrosis. However, the radiological picture of malignant BS glioma is non specific and justifies histological confirmation in most cases. In previous studies, preoperative radiological diagnoses were found to be incorrect in 10–25% of cases in patients over 20 years of age who presented with a contrast enhancing lesion in the brainstem (Fig. 36.2) (Rajshekhar and Chandy, 1995; Boviatsis et al., 2003).

Differential Diagnoses The differential diagnosis of malignant brainstem gliomas is often difficult and includes many various diseases (Table 36.1).

Fig. 36.2 Axial T1-weighted MR after gadolinium infusion in an adult patient with an exophytic brainstem glioma showing post gadolinium enhancement

Surgery In order to avoid the complications of an empiric and inappropriate therapy, biopsy is indicated in many contrast enhancing brainstem masses because the histological variety is much more important in adults than in childhood (see paragraph above) (Abernathey et al., 1989; Kratimenos and Thomas, 1993; GoncalvesFerreira et al., 2003; Shad et al., 2005). Moreover, a benign, non-neoplastic pathology was diagnosed in 13–17% of adults with a suspected malignant brainstem mass (Rajshekhar and Chandy, 1995). Although it is difficult to estimate, the risk of severe complications of a stereotactic biopsy for a suspected malignant brainstem glioma is considered to be low (probably between 1 and 5.6%) (Rajshekhar and Chandy, 1995; Boviatsis et al., 2003). Usually, the lower the lesion is located in the brainstem, the greater the risks involved. Operative mortality is exceedingly rare (Thomas et al., 1988; Kratimenos and Thomas, 1993; Boviatsis et al., 2003). Morbidity is usually minimal and temporary, consisting of transient neurological deterioration (Kratimenos and Thomas, 1993; Goncalves-Ferreira et al., 2003), or obstructive hydrocephalus. Permanent neurological deficit was estimated to occur in about 1.4% in Rajshekhar’s series (Rajshekhar and Chandy, 1995) and some patients required early re-aspiration of a haematoma (Kratimenos and Thomas, 1993). The optimal method and routes of biopsy are controversial (Abernathey et al., 1989; Kratimenos and Thomas, 1993; Goncalves-Ferreira et al., 2003). CT-or MRI-guided stereotactic surgery of the brainstem appears safe and reliable (Rajshekhar and Chandy, 1995; Massager et al., 2000; Boviatsis et al., 2003). PET-scan may be helpful (Massager et al., 2000). Biopsy provides a high yield of positive histological diagnosis, estimated in the literature around

36 Adult Brainstem Gliomas: Diagnosis and Treatment

375

Table 36.1 Main differential diagnoses of malignant brainstem gliomas Other tumors

Infections

Vascular processes

• Metastasis (Friedman et al., 1989) • Lymphoma (Friedman et al., 1989; Goncalves-Ferreira et al., 2003) • Germinoma and other intracranial germ cell tumors (Friedman et al., 1989) • Ganglioglioma (Goncalves-Ferreira et al., 2003) • Ependymoma (Abernathey et al., 1989) • Dysembryoplastic neuroepithelial tumor (Fujimoto et al., 2000) • Chordoma (Friedman et al., 1989) • Neurocytoma (Swinson et al., 2006) • Dermoid and epidermoid cysts (Rajshekhar and Chandy, 1995) • Cystic trochlear nerve neurinoma (Shenoy and Raja, 2004) • Acoustic neuroma (Yousry et al., 2004)

• Tuberculomas (Rajshekhar and Chandy, 1995) • Toxoplasmosis (Goncalves-Ferreira et al., 2003) • Pyogenic abscess (Rajshekhar and Chandy, 1995) • Aspergillus abscess (Goncalves-Ferreira et al., 2003) • Cryptococcal abscess (Abernathey et al., 1989)

• Haematomas, vasculitis (Goncalves-Ferreira et al., 2003) • Arteriovenous malformations (Abernathey et al., 1989), especially intracranial dural arteriovenous fistula (Crum and Link, 2004) and cavernous angiomas • Hypertensive encephalopathy (Jurcic et al., 2004) • Infarctions (Abernathey et al., 1989) • Radionecrosis (Abernathey et al., 1989)

95% (Thomas et al., 1988; Abernathey et al., 1989; Kratimenos and Thomas, 1993; Massager et al., 2000).

Inflammations and systemic diseases • Neuro behçet (Ben Taarit et al., 2002), • Sarcoidosis (Goncalves-Ferreira et al., 2003), • Progressive multifocal leukencephalopathy (Friedman et al., 1989; Goncalves-Ferreira et al., 2003), • Demyelinating disease (Abernathey et al., 1989; Goncalves-Ferreira et al., 2003)

overlap between adults and children. They include focal tectal gliomas (8%), cystic pilocytic astrocytoma and exophytic posterior glioma.

Radiotherapy These malignant tumors are highly resistant to treatment by radiotherapy. Indeed, after radiotherapy, only 13% of patients showed clinical and radiological improvement (Guillamo et al., 2001a). Gamma Knife radiosurgery has been recently used in some cases of brainstem gliomas, sometimes with satisfying functional outcome (Fuchs et al., 2002), although it seems to be ineffective in other cases. Chemotherapy Chemotherapy has been occasionally used in this subgroup of tumors but data are too scant to draw any recommendation on this issue. Follow-up The overall prognosis remains very poor despite radiotherapy and chemotherapy, with a median survival about 11.2 months (Guillamo et al., 2001a) and 25 months (Salmaggi et al., 2008). The other types of BS tumors identified in the adult do not seem to differ from pediatric types, suggesting

Focal Tectal Brainstem Gliomas Epidemiology Focal tectal gliomas constitute a small subgroup in adults (8% of cases).

Pathology Histopathology, in the few cases where it could be obtained, most often reveals a grade II oligoastrocytoma (Guillamo et al., 2001a). However, some cases of pilocytic astrocytomas have been described. A few cases of glioblastoma have also been reported in this location with an aggressive course.

Clinical and Radiological Presentation Focal tectal glioma usually presents with obstructive hydrocephalus, as in children. They may rarely present with intracranial hemorrhage, which should be removed.

376

Treatment and Follow-up These lesions are associated with long survival (sometimes exceeding 10 years) following ventriculoperitoneal shunting and usually focal radiotherapy (Chamberlain, 1999). Since simple observation after shunting may be an appropriate option in children, the decision to treat adult patients with radiotherapy is also questionable to some authors (Guillamo et al., 2001a), who propose a similar approach in adults.

F. Laigle-Donadey and J.-Y. Delattre

progress has been made. In this group, advances in treatment will require a better knowledge of the biology of these tumors, and innovative approaches, possibly based on the analysis of the molecular profile of the tumor. Experimental animal models of BSG and preclinical human studies to investigate intratumoral drug treatment as well as promising new agents, such as signal transduction inhibitors, are actively being studied in children (Fouladi et al., 2007; Pollack et al., 2007) and could secondly benefit to adult patients. Finally, the criteria to evaluate tumor response during therapy of BSG remain unclear and require novel approaches.

Posterior Exophytic Gliomas Exophytic contrast-enhancing glioma, which is well known in children (up to 10% of cases) and is associated with a good prognosis, is extremely rare in adults, maybe because most exophytic gliomas are pilocytic astrocytomas, a rare tumor type in adults (Guillamo et al., 2001a). A case report of exophytic pontine glioma simulating acoustic neurinoma has been reported in an adult patient (Swaroop and Whittle, 1997). Surgical resection is reserved for such patients with specific tumoral locations as dorsal exophytic tumors protruding into the fourth ventricle. Improvement in neurosurgical techniques (particularly the use of intraoperative ultrasound, intraoperative neurophysiological mapping and computer reconstruction techniques) has facilitated partial resection of tumors previously considered as inoperable (Tanaka et al., 2005), or even gross total removal in some cases, without affecting the functional prognosis.

Brainstem Gliomas Associated with Neurofibromatosis Type 1 Data regarding brainstem glioma associated with NF1 in adult patients are too scarce to draw conclusions, but these tumors may be more aggressive in adults than in children (Guillamo et al., 2001a).

Conclusion In conclusion, despite improvements in the understanding and classification of BSGs, the primary challenge remains diffuse BSG of adults where little

References Abernathey CD, Camacho A, Kelly PJ (1989) Stereotaxic suboccipital transcerebellar biopsy of pontine mass lesions. J Neurosurg 70:195–200 Ben Taarit C, Turki S, Ben Maiz H (2002) Pseudotumoral neurobehcet: a case report. J Mal Vasc 27:93–95 Boviatsis EJ, Kouyialis AT, Stranjalis G, Korfias S, Sakas DE (2003) CT-guided stereotactic biopsies of brain stem lesions: personal experience and literature review. Neurol Sci 24:97–102 Chamberlain MC (1999) Adult brainstem gliomas. Neurology 53:437–438 Crum BA, Link M (2004) Intracranial dural arteriovenous fistula mimicking brainstem neoplasm. Neurology 62: 2330–2331 Dirr LY, Donofrio PD, Patton JF, Troost BT (1989) A falsepositive edrophonium test in a patient with a brainstem glioma. Neurology 39:865–867 Elgamal EA, Coakham HB (2005) Hemifacial spasm caused by pontine glioma: case report and review of the literature. Neurosurg Rev 28:330–332 Fouladi M, Nicholson HS, Zhou T, Laningham F, Helton KJ, Holmes E, Cohen K, Speights RA, Wright J, Pollack IF, Children’s Oncology Group (2007) A phase II study of the farnesyl transferase inhibitor, tipifarnib, in children with recurrent or progressive high-grade glioma, medulloblastoma/primitive neuroectodermal tumor, or brainstem glioma: a Children’s Oncology Group study. Cancer 110:2535–2541 Freeman CR, Farmer JP (1998) Pediatric brain stem gliomas: a review. Int J Radiat Phys 40:265–271 Friedman WA, Sceats J, Nestok B, Ballinger WE (1989) The incidence of unexpected pathological findings in an imageguided biopsy series: a review of 100 consecutive cases. Neurosurgery 25:180–184 Fuchs I, Kreil W, Sutter B, Papaethymiou G, Pendl G (2002) Gamma Knife radiosurgery of brainstem gliomas. Acta Neurochir Suppl 84:85–90 Fujimoto K, Ohnishi H, Tsujimoto M, Hoshida T, Nakazato Y (2000) Dysembryoplastic neuroepithelial tumor of the cerebellum and brainstem. Case report. J Neurosurg 93: 487–489

36 Adult Brainstem Gliomas: Diagnosis and Treatment Gaviani P, Gonzalez RG, Zhu JJ, Batchelor TT, Henson JW (2005) Central neurogenic hyperventilation and lactate production in brainstem glioma. Neurology 64:166–167 Goncalves-Ferreira AJ, Herculano-Carvalho M, Pimentel J (2003) Stereotactic biopsies of focal brainstem lesions. Surg Neurol 60:311–320 Guillamo JS, Doz F, Delattre JY (2001b) Brain stem gliomas. Curr Opin Neurol 14:711–715 Guillamo JS, Monjour A, Taillandier L, Devaux B, Varlet P, Haie-Meder C, Defer GL, Maison P, Mazeron JJ, Cornu P, Delattre JY, Association des Neuro-Oncologues d’Expression Francaise (ANOCEF) (2001a) Brainstem gliomas in adults: prognostic factors and classification. Brain 124:2528–2539 Hsu CY, Hogan EL, Wingfield W Jr, Webb JG, Perot PL Jr, Privitera PJ, Talbert OR, Balentine JD (1984) Orthostatic hypotension with brainstem tumors. Neurology 34:1137–1143 Jurcic V, Ferluga D, Jeruc J, Pogacnik T, Popovic M (2004) Hypertensive encephalopathy mimicking brainstem tumour in psychiatric patient. Folia Neuropathol 42:37–41 Kesari S, Kim RS, Markos V, Drappatz J, Wen PY, Pruitt AA (2008) Prognostic factors in adult brainstem gliomas: a multicenter, retrospective analysis of 101 cases. J Neurooncol 88:175–183 Kratimenos GP, Thomas DG (1993) The role of image-directed biopsy in the diagnosis and management of brainstem lesions. Br J Neurosurg 7:155–164 Landolfi JC, Thaler HT, DeAngelis LM (1998) Adult brainstem gliomas. Neurology 51:1136–1139 Lim EC, Wilder-Smith EP, Chong JL, Wong MC (2005) Seeing the light: brainstem glioma causing visual auras and migraine. Cephalalgia 25:154–156 Marin-Sanabria EA, Kobayashi N, Miyake S, Kohmura E (2006) Snoring associated with Ondine’s curse in a patient with brainstem glioma. J Clin Neurosci 13:370–373 Massager N, David P, Goldman S, Pirotte B, Wikler D, Salmon I, Nagy N, Brotchi J, Levivier M (2000) Combined magnetic resonance imaging- and positron emission tomographyguided stereotactic biopsy in brainstem mass lesions: diagnostic yield in a series of 30 patients. J Neurosurg 93:951–957 Mursch K, Halatsch ME, Markakis E, Behnke-Mursch J (2005) Intrinsic brainstem tumours in adults: results of microneurosurgical treatment of 16 consecutive patients. Br J Neurosurg 19:128–136 Packer RJ, Zimmerman RA, Luerssen TG, Sutton LN, Bilaniuk LT, Bruce DA, Schut L (1985) Brainstem gliomas of childhood: magnetic resonance imaging. Neurology 35:397–401 Pollack IF, Jakacki RI, Blaney SM, Hancock ML, Kieran MW, Phillips P, Kun LE, Friedman H, Packer R, Banerjee A,

377 Geyer JR, Goldman S, Poussaint TY, Krasin MJ, Wang Y, Hayes M, Murgo A, Weiner S, Boyett JM (2007) Phase I trial of imatinib in children with newly diagnosed brainstem and recurrent malignant gliomas: a Pediatric Brain Tumor Consortium report. Neuro Oncol 9:145–160 Rajshekhar V, Chandy MJ (1995) Computerized tomographyguided stereotactic surgery for brainstem masses: a riskbenefit analysis in 71 patients. J Neurosurg 82:976–981 Saka E, Ozkaynak S, Tuncer R (2006) Tongue tremor in brainstem pilocytic astrocytoma. J Clin Neurosci 13:503–506 Salmaggi A, Fariselli L, Milanesi I, Lamperti E, Silvani A, Bizzi A, Maccagnano E, Trevisan E, Laguzzi E, Rudà R, Boiardi A, Soffietti R., Associazione Italiana di Neurooncologia (2008) Natural history and management of brainstem gliomas in adults. A retrospective Italian study. J Neurol 255:171–177 Selvapandian S, Rajshekhar V, Chandy MJ (1999) Brainstem glioma: comparative study of clinico-radiological presentation, pathology and outcome in children and adults. Acta Neurochir (Wien) 141:721–727 Shad A, Green A, Bojanic S, Aziz T (2005) Awake stereotactic biopsy of brain stem lesions: technique and results. Acta Neurochir (Wien) 147:47–49 Shenoy SN, Raja A (2004) Cystic trochlear nerve neurinoma mimicking intrinsic brainstem tumour. Br J Neurosurg 18:183–186 Swaroop GR, Whittle IR (1997) Exophytic pontine glioblastoma mimicking acoustic neuroma. J Neurosurg Sci 41: 409–411 Swinson BM, Friedman WA, Yachnis AT (2006) Pontine atypical neurocytoma: case report. Neurosurgery 58:E990 Tanaka S, Kobayashi I, Utsuki S, Iwamoto K, Takanashi J (2005) Biopsy of brain stem glioma using motor-evoked potential mapping by direct peduncular stimulation and individual adjuvant therapy. Case report. Neurol Med Chir (Tokyo) 45:49–55 Thomas DG, Bradford R, Gill S, Davis CH (1988) Computerdirected stereotactic biopsy of intrinsic brain stem lesions. Br J Neurosurg 2:235–240 Torcuator R, Zuniga R, Loutfi R, Mikkelsen T (2009) Bevacizumab and irinotecan treatment for progressive diffuse brainstem glioma: case report. J Neurooncol 93: 409–412 White HH (1963) Brain stem tumors occurring in adults. Neurology 13:292–300 Yousry I, Muacevic A, Olteanu-Nerbe V, Naidich TP, Yousry TA (2004) Exophytic pilocytic astrocytoma of the brain stem in an adult with encasement of the caudal cranial nerve complex (IX–XII): presurgical anatomical neuroimaging using MRI. Eur Radiol 14:1169–1173

Chapter 37

The Use of Low Molecular Weight Heparin in the Treatment and Prevention of Thromboembolic Disease in Glioma Patients Bo H. Chao and H. Ian Robins

Abstract The neuro-oncologist routinely deals with an extraordinarily high risk patient population relative to thromboembolic disease. With the advent of easily accessible diagnostic studies, e.g., ultrasound and/or spiral CT scans, timely diagnosis of venous thromboembolic events (VTE) is readily accomplished. The introduction of low molecular weight heparin (LMWH) about 20 years ago (in contrast to unfractionated heparin and vitamin K antagonists) has provided a class of agents with a favorable therapeutic index. In the review to follow, the literature regarding the use of LMWH in neuro-oncology patient populations is summarized. Topics addressed include prophylaxis, treatment, as well as consideration of the potential anti-neoplastic properties of these drugs. Keywords Low molecular Thromboembolic disease · Prophylaxis · Antagonist

weight Glioma

heparin · VTE

· ·

Introduction It is well recognized that cancer patients in general (Shlebak and Smith, 1997), and in particular glioma patients are highly predisposed to thromboembolic phenomena (Sawaya and Highsmith, 1988; Marras et al., 2000). Brisman and Mendell (1973) demonstrated an 8.4% incidence of pulmonary emboli in glioma patients; this represents a three fold increase H.I. Robins () Department of Medicine, Human Oncology, and Neurology, University of Wisconsin School of Medicine and Public Health, Madison, WI 53792, USA e-mail: [email protected]

in the incidence seen in non-malignant neurosurgical patients. Similarly, the incidence of deep vein thromboembolism (DVT) in such patients is 27.5% compared to 17% in a control neurosurgical group (Kayser-Gatchalian and Kayser, 1975). Ruff and Posner (1983) noted a 31% incidence of confirmed DVT in 264 glioma patients. Quevedo et al. (1994) noted an incidence of 28% of thromboembolism in a series of 64 high grade glioma patients. Sawaya et al. (1992), using I-fibrinogen scanning, demonstrated DVTs in 60% of glioblastoma multiforme (GBM) patients; the presence of venous thromboembolic events (VTE) did not correlate with ambulatory status, time of surgery, length of operation, or occurrence in a paretic limb. Sawaya and Highsmith (1992) have suggested that malignant brain tumors release a factor responsible for this predisposition to coagulopathy. Previous work suggesting increased platelet adhesiveness in malignant brain tumors is consistent with this supposition (Nathanson and Savitsky, 1952; Millac, 1967). More recent work by Iberti et al. (1994) and Hamilton et al. (1994) supports the concept of an increased coagulable state of brain tumor patients. Coupled to this significant morbidity is the potential risk for the central nervous system hemorrhage when anti-coagulants are used as a therapeutic intervention and/or prophylaxis. In the text to follow we will discuss the existing data for the treatment and prophylaxis of thromboembolic disease in glioma patients. The review will focus on the use of low molecular weight heparin (LMWH), as the published literature and experience center on this class of anti-coagulants relative to both safety and efficacy. Encompassed in this discussion will be a review of the data suggesting the potential for anti-neoplastic effects with the application of LMWH.

M.A. Hayat (ed.), Tumors of the Central Nervous System, Volume 2, DOI 10.1007/978-94-007-0618-7_37, © Springer Science+Business Media B.V. 2011

379

380

Low Molecular Weight Heparin as Treatment for Thromboembolic Disease Based on randomized clinical studies, LMWH is the agent of choice for the initial and long term treatment of VTE in patients with neoplastic disease (Lee, 2009; Lyman et al., 2007). The cumulative risk of symptomatic and recurrent VTE during 6 months of anticoagulation was 17% for patients treated with a vitamin K antagonist, such as warfarin or acenocoumarol, compared with 9% for patients treated with dalteparin, a LMWH (Lee et al., 2003). The relative risk reduction of 52% was statistically significant (hazard ratio = 0.48; P = 0.002). One episode of recurrent VTE is prevented for every 13 patients treated with dalteparin. Regarding side effects, no difference in major or minor bleeding was detected between the groups, with 6% of dalteparin-treated patients and 4% of control patients having major bleeding episodes (Lee et al., 2003). There were no confirmed cases of heparin-induced thrombocytopenia, and patient compliance with selfinjection in these studies was good (Deitcher et al., 2006; Meyer et al., 2002; Lee et al., 2003; Hull et al., 2006). Comparing to unfractionated heparin, LMWH reduces the risk of recurrent thrombosis by 32%, major bleeding by 43%, and death by 24%, according to a Cochrane systemic review summarizing 22 clinical trials (van Dongen et al., 2004). LMWH is the preferred agent for treatment of cancer-associated thrombosis because of its efficacy, safety, and convenience. In the aforementioned studies in a general population of cancer patients with VTE, dalteparin was given at full therapeutic dose of 200 U/kg once daily for a month, followed by dose reduction down to 75% of the full dose for the remainder of the 6-month period (Lee et al., 2003). It should be noted that the concept of a 25% dose reduction after initial therapy was an arbitrary choice, and not evaluated in comparison to other possibilities. Other studies used full therapeutic dose of enoxaparin for 3 months (Deitcher et al., 2006; Meyer et al., 2002). However, in patients with GBM, because of a higher risk of bleeding, we typically treat VTE in GBM patients with full dose LMWH until resolutions of symptoms from their thrombosis, usually within 2 weeks of onset of VTE, followed by 50% dose reduction in the following month, and followed by another 25% dose reduction after 2–3 months. We recommend dividing the total daily dose during the acute phase

B.H. Chao and H.I. Robins

of treatment into twice-a-day dosing to reduce bleeding risk. If the GBM patient develops recurrent VTE during dose reduction of LMWH, we would check the antifactor Xa levels 4 h after an injection and adjust the dose accordingly. There has not been a trial to evaluate this treatment algorithm; it has been used at our center since the introduction of LMWH with an estimated incidence of recurrent VTE less than 5%. This use of a lower dose of LMWH post acute treatment is consistent with the efficacy of LMWH in the prophylaxis setting and had evolved as a conceptual extrapolation. It has been our practice to continue LMWH indefinitely in high grade glioma patients based on the experience in a Eastern Cooperative Oncology Group (ECOG)/Radiation Therapy Oncology Group (RTOG) trial, described in detail in section “Anti-neoplastic Effects of Low Molecular Weight Heparin” below (Robins et al., 2008). Patients who are on LMWH for protracted periods of time are at risk for osteoporosis, (and often have another risk factor for bone demineralization, i.e., steroids). We suggest monitoring such patients with bone mineral density studies, and considering interventions such as bisphosphonates as appropriate. It should be noted that fondaparinux, a synthetic selective inhibitor of activated factor X (with no activity against thrombin), is approved for the initial therapy of VTE (Buller et al., 2004). In terms of administration, convenience, and cost, it is comparable to LMWH. Similar to LMWH, it is contraindicated in patients with significant renal insufficiency. However, it has not been studied in cancer patients. Hence its use in cancer patients is generally limited in those who have allergies to heparin or a history of heparin-induced thrombocytopenia (Lee, 2009). The role of vena cava filters in the treatment of VTE in cancer patients remains undefined as previously reviewed (Lee, 2009). Filters do not treat the underlying hypercoagulable state in cancer patients. The use of filters in patients receiving anticoagulation led to a reduction in symptomatic pulmonary embolism, but more recurrent DVTs and no difference in overall survival (Investigators, 2005). It is our view that the use of filters should be reserved for situations in which anticoagulant therapy is contraindicated, e.g., active bleeding, severe thrombocytopenia, or in preparation for a significant neurosurgical intervention.

37 LMWH in the Treatment and Prevention of Thromboembolic Disease in Glioma Patients

Low Molecular Weight Heparin as Prophylaxis for Thromboembolic Disease As discussed in section “Introduction”, glioma patients are highly predisposed to thromboembolic phenomena with an incidence approaching ~30%. In a phase II study in patients with newly diagnosed glioblastoma multiforme (GBM), patients were given dalteparin 5000 units during and post radiation, as described in detail in section “Anti-neoplastic Effects of LMWH” below (Robins et al., 2008). The primary endpoint was survival, and the secondary end point was the incidence of VTE. This study closed prematurely with only 42 patients accrued due to a change in standard of care for GBM. As the study was conceived it was estimated that over the course of the study, 30% of the enrolled patients would develop a VTE over the entire time course of their disease. With a planned cohort of 72 patients, there was 81% power to detect a decrease from 30 to 15% in VTE using a two-sided 0.05 exact test for a single proportion. In the end the study observed no VTE in patients actively receiving dalteparin. Time on dalteparin ranged from 1.2 to 25.4 months, with a median time of 6.3 months. There were no reports of grade 3/4 bleeding or thrombocytopenia related to dalteparin pre or post progression. There were local site reactions to dalteparin in up to 71% of patients; two patients experienced grade 3 toxicities ascribed to radiation. Two patients developed DVT after stopping dalteparin when they developed progressive GBM. A seemingly contradictory result in the context of a phase III Canadian study was later reported at American Society of Clinical Oncology annual meeting in 2007 (Perry et al., 2007b), and was updated later in the year at another meeting (Perry et al., 2007a). This trial was planned to enroll 512 patients; it closed prematurely, secondary to lack of drug availability, with 186 malignant glioma patients randomized to dalteparin 5000 units daily versus placebo. The primary endpoint was the development of VTE. Although there was a trend for LMWH to decrease VTE, it was not statistically significant (HR = 0.65, 95% CI: 0.27–1.54, p = 0.33). However, the hazard ratio was impressive at 0.65 in favor of the LMWH arm. It remains speculative whether this hazard ratio would have achieved statistical significance if the study had met its accrual

381

objective. There was also a trend for increased bleeding in the LMWH arm 5.1% (n = 5) versus 1.2% (n = 1). These authors in commenting on the “trend toward increased intracranial bleeding” noted “the clinical significance of this is unclear since intratumoral hemorrhage is sometimes part of the natural history of this disease” (Perry et al., 2007a). Potential noteworthy differences in these studies may relate to patient populations. The RTOG study was restricted to a GBM cohort (as opposed to all malignant glioma), as well as a requirement for an unusually good performance status, i.e., 95% ECOG performance status of 0 or 1. To date a final published report regarding the Canadian study has not become available. At present, we agree with the authors of both of these individual studies: The results of these studies taken collectively leave the role of anticoagulant in thromboembolic prophylaxis as unclear for this group of patients. Definitive conclusions regarding the application of VTE prophylaxis strategies will require further investigation. There have, however, been studies of VTE prophylaxis after elective neurosurgery in patients with brain tumors, including but not exclusive to glioma. Two multi-center, randomized, double-blind trials studied the use of LMWH (nadroparin 7500 units subcutaneously once daily, or enoxaparin 40 mg subcutaneously once daily, respectively) versus placebo in conjunction with the use of compression stockings in the prevention of VTE after elective neurosurgery (Nurmohamed et al., 1996; Agnelli et al., 1998). The majority of the accrued patients had brain tumors, up to 30% of whom were diagnosed with glioma (Agnelli et al., 1998). LMWH or placebo was given within 24 h after surgery and continued for up to 10 days or until hospital discharge. Both studies showed that LMWH combined with compression stockings is more effective than compression stockings alone for the prevention of VTE after elective neurosurgery, without inducing any significant increase of major bleeding. The absolute risk reduction of VTE in patients receiving LMWH was 8 and 14%, whereas the absolute risk reduction of proximal VTE or pulmonary embolism in patients receiving LMWH was 4 and 8%, respectively (Nurmohamed et al., 1996; Agnelli et al., 1998). However, it should be noted that LMWH in these studies was given within 24 h after surgery. Another randomized study conducted at the University of Michigan, where LMWH (enoxaparin

382

30 mg subcutaneously every 12 h) was initiated before the induction of anesthesia and was continued throughout the hospital stay, showed that enoxaparin therapy initiated at the time of anesthesia induction significantly increased postoperative intracranial hemorrhage (Dickinson et al., 1998). This study was terminated early because of the increased incidence of adverse events in the enoxaparin treatment group. In conclusion, LMWH given within 24 h after neurosurgery, but not prior to neurosurgery, in combination with compression stockings, is effective for the prevention of VTE in glioma patients. However, the role of VTE prophylaxis beyond the postoperative period after hospital discharge is yet to be determined and still requires further studies.

Anti-neoplastic Effects of Low Molecular Weight Heparin The genesis of a clinical study to test the antineoplastic proprieties of LMWH in GBM patients occurred when investigators anecdotally noted usually good outcome in high grade glioma patients who were receiving LMWH for DVT. In reviewing the literature, a meta-analysis was identified that suggested a reduced mortality in cancer patients receiving LMWH compared with patients receiving standard heparin; further, this benefit was not consequent to a reduced thromboembolism rate (Siragusa, 1996). Additionally, a potential mechanistic basis for this was identified: (i) anti-angiogenesis (Folkman et al., 1983; Norrby 1993), (ii) effects on cellular matrix (Bitan et al., 1995; Collen et al., 2000), and (iii) reduced cell proliferation (Castellot et al., 1986). In this regard, it is of interest to note that based on preclinical studies, anti-angiogenesis agents can putatively serve as radiosensitizers (Teicher et al., 1995; Wachsberger et al., 2003). Also, a preclinical study linked up-regulated tissue factor expression by epidermal growth factor receptor (EGFR) and EGFRvIII and the prothrombotic events that occur in the progression of GBM (Rong et al., 2006). As GBM is among the most vascular of all neoplasms, abundantly expressing vascular endothelial growth factor and platelet-derived growth factor (Dunn et al., 2004), the use of LMWH as a therapeutic adjunct to radiotherapy was viewed as worthy of clinical testing.

B.H. Chao and H.I. Robins

Taking aforementioned considerations into account, the ECOG in cooperation with the RTOG initiated a phase II study of the effect of dalteparin during and post radiation in newly diagnosed patients with GBM (Robins et al., 2008): Dalteparin of 5000 units daily was given subcutaneously with and after conventional radiotherapy to newly diagnosed GBM patients. Forty-five patients were accrued. The study was designed such that at time of progression, patients could continue dalteparin in addition to standard regimens. Pretreatment characteristics included: Median age 61 (range 26–78); ECOG Performance status: 0 = 38%, 1 = 57%, 2 = 5%; gross total resection 45%. There were no grade 3/4 bleeding or thrombocytopenic events, and no VTE occurred while patients were on dalteparin. Median time on dalteparin was 6.3 months, median time to progression was 3.9 months; median survival was 11.9 months. There was no significant improvement in survival when compared to the RTOG GBM database (with various radiation/drug doublets including BCNU) using recursive partitioning analysis (RPA) (Curran et al., 1993). The results were, however, inferior to the use of temozolomide, which became the new standard of care just at the time this study was concluded (Stupp et al., 2004). As noted above, investigators were provided the option of continuing dalteparin after disease progression. The rationale for this included its potential use as an anti-angiogenesis agent, i.e., to slow disease progression, as well as use for DVT prophylaxis. In this study, 22 of 38 patients continued dalteparin post first progression. Median survival for those who stopped dalteparin at first progression versus continued use was 3.2 and 7.8 months, respectively. This analysis is confounded by several factors and should be interpreted with caution. Patients could receive other treatments at progression, as outlined in the protocol, with or without the continuation of dalteparin. The patients who stopped dalteparin at first progression may have been too sick to receive other treatments. For those patients who continued dalteparin at first progression, their survival could be confounded by the varying lengths that they continued dalteparin and/or by any of the other treatments they received. As discussed in section “Low Molecular Weight Heparin as Prophylaxis for Thromboembolic Disease” of this chapter above, historically the incidence of VTE in GBM patients is ~30%, and in this study, no VTE was observed in GBM patients actively receiving

37 LMWH in the Treatment and Prevention of Thromboembolic Disease in Glioma Patients

dalteparin. Because dalteparin might reduce the incidence of VTE, having no significant overlapping toxicities with most other drugs, its testing in a combined modality approach with other medications active in the treatment of GBM might be warranted in future trials.

Summary Comments As outlined above, the neuro-oncologist routinely deals with an extraordinarily high risk patient population relative to VTE. Indeed, it is recommended that new patients be trained in the signs and symptoms of DVT and pulmonary embolism. With the advent of easily accessible diagnostic studies, e.g., ultrasound and/or spiral CT scans, timely diagnosis of VTE is less problematic than in past years. Outpatient therapy is becoming standard practice at most centers for clinically well compensated patients. For more than 5 decades, unfractionated heparin and vitamin K antagonists have been used for the treatment and prevention of VTE. The introduction of LMWH about 20 years ago, however, has provided a class of agents with a unique therapeutic index that is favorable in glioma patients. It is anticipated that more convenient drugs (e.g., oral, or perhaps not requiring monitoring) currently in development, targeting activated factor X (i.e., factor Xa) or thrombin (factor IIa), will become available. As the optimal treatment for VTE is prevention, it is hoped and anticipated that such drug development of oral IIa and Xa inhibitors will provide a new set of efficacious and safe agents for this high risk patient population.

References Agnelli G, Piovella F, Buoncristiani P, Severi P, Pini M, D’Angelo A, Beltrametti C, Damiani M, Andrioli GC, Pugliese R, Iorio A, Brambilla G (1998) Enoxaparin plus compression stockings compared with compression stockings alone in the prevention of venous thromboembolism after elective neurosurgery. N Engl J Med 339:80–85 Bitan M, Mohsen M, Levi E, Wygoda MR, Miao H, Lider O, Svahn CM, Ekre HP, Ishai-Michaeli R, Bar-Shavit R, Vlodavsky I, Peretz T (1995) Structural requirements for inhibition of melanoma lung colonization by heparanase inhibiting species of heparin. Isr J Med Sc 31:106–118

383

Brisman R, Mendell J (1973) Thromboembolism and brain tumor. J Neurosurg 38:337–338 Buller HR, Davidson BL, Decousus H, Gallus A, Gent M, Piovella F, Prins MH, Raskob G, Segers AE, Cariou R, Leeuwenkamp O, Lensing AW (2004) Fondaparinux or enoxaparin for the initial treatment of symptomatic deep venous thrombosis: a randomized trial. Ann Intern Med 140:867–873 Castellot JJ, Jr, Choay J, Lormeau J, Petitou M, Sache E, Kamovsky MJ (1986) Structural determinants of the capacity of heparin to inhibit the proliferation of vascular smooth muscle cells. II. Evidence for a pentasaccharide sequence that contains a 3-0-sulfate group. J Cell Biol 102:1979–1984 Collen A, Smorenburg SM, Peters E, Lupu F, Koolwijk P, van Noordan C, van Hinsbergh VWM (2000) Unfractionated and low molecular weight heparin affect fibrin structure and angiogenisis in vitro. Cancer Res 60:6196–6200 Curran WJ Jr, Scott CB, Horton J, Nelson JS, Weinstein AS, Fischbach AJ, Chang CH, Rotman M, Asbell SO, Krisch RE (1993) Recursive partitioning analysis of prognostic factors in three Radiation Therapy Oncology Group malignant glioma trials. J Natl Cancer Inst 85:704–710 Deitcher SR, Kessler CM, Merli G, Rigas JR, Lyons RM, Fareed J (2006) Secondary prevention of venous thromboembolic events in patients with active cancer: enoxaparin alone versus initial enoxaparin followed by warfarin for a 180-day period. Clin Appl Thromb Hemost 12:389–396 Dickinson LD, Miller LD, Patel CP, Gupta SK (1998) Enoxaparin increases the incidence of postoperative intracranial hemorrhage when initiated preoperatively for deep venous thrombosis prophylaxis in patients with brain tumors. Neurosurgery 43:1074–1081 Dunn IF, Hesse O, Black PM (2004) Growth factors in glioma angiogenesis: FGFs, PDGF, EGF, and TGFs. J Neurooncol 50:121–137 Folkman J, Langer R, Linhardt RJ, Haudenschild C, Taylor S (1983) Angiogenesis inhibition and tumor regression caused by heparin or a heparin fragment in the presence of cortisone. Science 221:719–725 Hamilton MG, Hull RD, Pineo GF (1994) Venous thromboembolism in neurosurgery and neurology patients: A review. Neurosurgery 34:280–296 Hull RD, Pineo GF, Brant RF, Mah AF, Burke N, Dear R, Wong T, Cook R, Solymoss S, Poon MC, Raskob G (2006) Long-term low-molecular-weight heparin versus usual care in proximal-vein thrombosis patients with cancer. Am J Med 119:1062–1072 Iberti TJ, Miller M, Abalos A, Fischer EP, Post KD, Benjamin E, Oropello JM, Wiltshire-Clement M, Rand JH (1994) Abnormal coagulation profile in brain tumor patients during surgery. Neurosurgery 34:389–395 Investigators, P.d.R.d.E.P.p.I.C (2005) Eight-year follow-up of patients with permanent vena cava filters in the prevention of pulmonary embolism: the PREPIC (Prevention du Risque d’Embolie Pulmonaire par Interruption Cave) randomized study. Circulation 112:416–422 Kayser-Gatchalian MC, Kayser K (1975) Thrombosis and intracranial tumors. J Neurol 209:217–224 Lee AY (2009) Anticoagulation in the treatment of established venous thromboembolism in patients with cancer. J Clin Oncol 27:4895–4901

384 Lee AY, Levine MN, Baker RI, Bowden C, Kakkar AK, Prins M, Rickles FR, Julian JA, Haley S, Kovacs MJ, Gent M (2003) Low-molecular-weight heparin versus a coumarin for the prevention of recurrent venous thromboembolism in patients with cancer. N Engl J Med 349:146–153 Lyman GH, Khorana AA, Falanga A, Clarke-Pearson D, Flowers C, Jahanzeb M, Kakkar A, Kuderer NM, Levine MN, Liebman H, Mendelson D, Raskob G, Somerfield MR, Thodiyil P, Trent D, Francis CW (2007) American Society of Clinical Oncology guideline: recommendations for venous thromboembolism prophylaxis and treatment in patients with cancer. J Clin Oncol 25:5490–5505 Marras LC, Geerts WH, Perry JR (2000) The risk of venous thromboembolism is increased throughout the course of malignant glioma: an evidence-based review. Cancer 89:640–646 Meyer G, Marjanovic Z, Valcke J, Lorcerie B, Gruel Y, SolalCeligny P, Le Maignan C, Extra JM, Cottu P, Farge D (2002) Comparison of low-molecular-weight heparin and warfarin for the secondary prevention of venous thromboembolism in patients with cancer: a randomized controlled study. Arch Intern Med 162:1729–1735 Millac P (1967) Platelet stickiness in patients with intracranial tumors. BMJ 4:25–26 Nathanson M, Savitsky JP (1952) Platelet adhesive index studies in multiple sclerosis and other neurological disorders. Bull NY Acad Med 28:462–468 Norrby K (1993) Heparin and angiogenesis: a low-molecularweight fraction inhibits and a highmolecular-weight fraction stimulates angiogenesis systemically. Haemostasis 23(Suppl 1):141–149 Nurmohamed MT, van Riel AM, Henkens CM, Koopman MM, Que GT, d’Azemar P, Buller HR, ten Cate JW, Hoek JA, van der Meer J, van der Heul C, Turpie AG, Haley S, Sicurella A, Gent M (1996) Low molecular weight heparin and compression stockings in the prevention of venous thromboembolism in neurosurgery. Thromb Haemost 75: 233–238 Perry JR, Rogers L, Laperrier N, Julian J, Geerts W, Agnelli G, Malkin M, Sawaya R, Baker R, Levine M (2007a) A phase III randomized placebo-controlled trial of thromboprophylaxis using dalteparin low molecular weight heparin (LMWH) in patients with newly diagonsed malignant glioma updated results. Thromb Res 120:ABS PO–40 Perry JR, Rogers L, Laperriere N, Julian J, Geerts W, Agnelli G, Malkin M, Sawaya R, Baker R, Levine M (2007b) PRODIGE: a Phase III randomized placebo-controlled trial of thromboprophylaxis using dalteparin low molecular weight heparin (LMWH) in patients with newly diagnosed malignant glioma. J Clin Oncol 25:77s (Abstr 2011)

B.H. Chao and H.I. Robins Quevedo JF, Buckner JC, Schmidt JL, Dinapoli RP, O’Fallon JR (1994) Thromboembolism in patients with high-grade glioma. Mayo Clin Proc 69:329–332 Robins HI, O’Neill A, Gilbert M, Olsen M, Sapiente R, Berkey B, Mehta M (2008) Effect of dalteparin and radiation on survival and thromboembolic events in glioblastoma multiforme: a phase II ECOG trial. Cancer Chemother Pharmacol 62:227–233 Rong Y, Durden DL, Van Meir EG, Brat DG (2006) Differential regulation of tissue factor expression by EGFR and EGFRvIII in GBM. Neuro-Oncology 8:401 (ABS CB–27) Ruff RL, Posner JB (1983) Incidence and treatment of peripheral venous thrombosis in patients with glioma. Ann Neurol 13:334–336 Sawaya R, Highsmith RF (1988) Brain tumors and the fibrinolytic enzyme system. In: Kornblith PL, Walker MD (eds) Advances in neuro-oncology. Futura, Mount Kisco, NY, pp 103–157 Sawaya R, Highsmith RF (1992) Postoperative venous thromboembolism and brain tumors: Part III. Biochemical profile. J Neuro-Oncol 14:113–118 Sawaya R, Zuccarrello M, Elkalliny M, Nishiyama H (1992) Postoperative venous thromboembolism and brain tumors: Part I. Clinical Profile. J Neurooncol 14:119–125 Shlebak AA, Smith DB (1997) Incidence of objectively diagnosed thromboembolic disease in cancer patients undergoing cytotoxic chemotherapy and/or hormonal therapy. Cancer Chemother Pharmacol 39:462–466 Siragusa SCB, Piovella F, Hirsh J, Ginsberg S (1996) Lowmolecular-weight heparins and unfractionated heparin in the treatment of patients with acute venous thromboembolism: Results of a meta-analysis. J Am Med 100:269–277 Stupp R, Mason WP, Van Den Bent MJ, Weller M, Fisher B, Taphoorn M, Brandes AA, Cairncross G, Lacombe D, Mirimanoff RO (2004) Conclusive results of a randomized phase III trial by the EROTC Brain and RT groups and NICIC Clinical Trails Group. Concomitant and adjuvant temozolomide and radiotherapy for newly diagnosed glioblastoma multiforme. Proc Am Soc Clin Oncol 23:1 (Abstract 2) Teicher BA, Holden SA, Ara G, Dupuis NP, Liu F, Yuan J, Ikebe M, Kakeji Y (1995) Influence of an antiangiogenic treatment on 9L gliosarcoma: oxygenation and response to cytotoxic therapy. Int J Cancer 61:732–737 van Dongen CJ, van den Belt AG, Prins MH, Lensing AW (2004) Fixed dose subcutaneous low molecular weight heparins versus adjusted dose unfractionated heparin for venous thromboembolism. Cochrane Database Syst Rev CD001100:1–92 Wachsberger P, Burd R, Dicker AP (2003) Tumor response to ionizing radiation combined with antiangiogenesis or vascular targeting agents. Clin Cancer Res 9:1957–1971

Part III

Prognosis

Chapter 38

Brainstem Gliomas: An Overview Marco Antonio Lima

Abstract Brainstem gliomas are a heterogeneous group of tumors that affect all ages, although they are predominant in children. There are several grading systems but they are often divided in four subgroups: tectal, diffuse pontine, cervicomedullary and dorsal exophytic. While diffuse pontine gliomas usually are high-grade tumors, the other three subtypes are low- grade lesions (pilocytic astrocytoma or grade II gliomas). The clinical presentation is diverse, but involvement of cranial nerves and long tracts are frequently present. Imaging is crucial for diagnosis and surgical planning and, in the case of diffuse pontine gliomas, MRI has replaced biopsy as gold standard. Treatment includes general measures (control of pain, seizures, physical and speech therapy) that are best addressed by a multidisciplinary team and surgical resection or palliation depending on the extension, grade and location. Since hydrocephalus is common, ventricular derivation followed by observation is the most frequent approach for tectal gliomas. Surgery and/or radiation therapy are reserved for clinical or radiological progression. Diffuse pontine gliomas are inoperable and radiation therapy is the mainstay of treatment leading to transient control of the symptoms. Most cervicomedullary and dorsally exophytic tumors are treated with partial or complete resection when feasible, followed by radiation therapy at time of recurrence or progression. Prognosis is dismal for pontine gliomas, and most patients do not survive 2 years from diagnosis. Outcome is better for tectal,

M.A. Lima () Department of Neurosurgery, Instituto Nacional de Câncer-INCa, Centro, Rio de Janeiro RJ-22230-130, Brazil e-mail: [email protected]

cervicomedullary and dorsally exophytic tumors and many patients achieve a long-term survival, although neurological sequelae are common. Keywords Brainstem gliomas · Tectal gliomas · Diffuse pontine gliomas · Cervicomedullary gliomas · Dorsal exophytic gliomas · CNS

Introduction Brainstem gliomas are a group of heterogeneous tumors that may affect either children or adults, but are more prevalent in the first years of life. In the pre-imaging era, they were considered a single entity and inoperable, but newer imaging techniques allowed the classification of subgroups with distinct behaviors. Nowadays, the knowledge of course of the disease as well the outcome of these different subtypes permits a better determination of which patients would benefit from surgery. In addition, despite the involvement of a crucial area of the central nervous system, modern neurosurgical procedures currently permit safer resections with minimal additional morbidity.

Epidemiology Brainstem gliomas are more frequent in children and represent 10% of pediatric central nervous system tumors (Walker et al., 1999). In a recent review of 6212 pediatric patients with gliomas, 19.8% were located in

M.A. Hayat (ed.), Tumors of the Central Nervous System, Volume 2, DOI 10.1007/978-94-007-0618-7_38, © Springer Science+Business Media B.V. 2011

387

388

the brainstem (Qaddoumi et al., 2009). There is no gender predilection and the mean age at onset in children is 7–9 years (Jallo et al., 2004). According to Smith et al. in the 1990–1994 period, the age-adjusted incidence of brainstem high-grade gliomas was 0.14 per 105 person-years, while the low-grade gliomas was 0.18 per 105 person-years (Smith et al., 1998). There are few studies of brainstem gliomas in adults since they are a rare entity, representing less than 2% of gliomas in this population (Guillamo et al., 2001). Although they can occur at any age, the peak incidence in adult patients is between the third and fourth decades (Selvapandian et al., 1999). In a recent study of brainstem gliomas in adults, the median age at diagnosis was 36 years (Kesari et al., 2008). Neurofibromatosis 1 (NF1) predisposes to central nervous system tumors, especially pilocytic astrocytomas. Although the optic pathways are the preferred site, other common locations are the brainstem and cerebellum (Ferner, 2007). In a retrospective study of 104 NF1 patients, 17% had brainstem gliomas (Guillamo et al., 2003), but lesions were more indolent in this group.

Classification Several grading systems for classification and staging of brainstem gliomas have been described over the years based on their location, focality, pattern of tumor growth and the presence of necrosis and hydrocephalus (Barkovich et al., 1990; Epstein, 1985). A simple classification would divide brainstem gliomas in four major categories: diffuse intrinsic gliomas, focal tectal, cervicomedullary and dorsal exophytic tumors (LaigleDonadey et al., 2008). • Focal tectal gliomas: these are rare tumors, representing less than 5% of brainstem gliomas in children (Laigle-Donadey et al., 2008). However, in a recent study with 101 adult brainstem glioma patients revealed that 20% of the lesions were tectal (Kesari et al., 2008). Indeed, Guillamo et al. have shown focal tectal gliomas in 8/48 adult patients (16.7%) suggesting that this subtype might be more frequent in adults than previously thought (Guillamo et al., 2001). The lesions are well

M.A. Lima

circumscribed and most are low-grade gliomas (WHO I and II). • Diffuse intrinsic gliomas: these are the most prevalent and aggressive brainstem gliomas, being responsible for 80% of the cases. Although they can occur anywhere in the brainstem, they are usually located at the pons, extending into the midbrain or medulla over the course of the disease. Most diffuse gliomas are malignant fibrillary astrocytomas (grade III or IV), but biopsy is not performed in all cases, since diagnosis is frequently based solely on clinical and radiological presentation. • Cervicomedullary gliomas: lesions in the cervicomedulary region are present in 5–10% with brainstem gliomas and occur predominantely in the pediatric population. Most tumors are low-grade astrocytomas and are similar to intramedullary spinal cord gliomas. According to Epstein et al. their growth is limited rostrally by the decussating fibers and pial elements (Epstein and Farmer, 1993). The lesion is well circumscribed and may be associated with syringomyelia caudally. • Dorsal exophytic gliomas: this subgroup represents between 14 and 24% of the brainstem gliomas. The epicenter is usually within the fourth ventricle and the age at presentation is younger than in other subtypes. The biopsy or surgical resection often reveals a low-grade glioma. Although, total resection is not feasible in many cases, partial resection can improve outcome.

Clinical Manifestations Tectal lesions grow insidiously and usually produce symptoms due to compression of cerebral aqueduct leading to a non-communicating hydrocephalus. The time from onset of symptoms until diagnosis is variable ranging from days to several years (Stark et al., 2005). Headache, vomiting and ataxia are frequently present at time of diagnosis (Daglioglu et al., 2003). Papilledema is observed in close to 50% of the patients. Parinaud syndrome, strabismus and long tract signs are found in a minority of patients (Ternier et al., 2006). Unusual presentations are tremor, irritability, parkinsonism, seizures, vertigo and decline in school performance. After ventricular drainage and treatment

38 Brainstem Gliomas: An Overview

of hydrocephalus, signs and symptoms can remain stable for years. Patients with diffuse pontine gliomas present with rapidly progressive neurological deficits, usually in weeks or few months, but it may be sudden, suggesting a stroke initially (Guillamo et al., 2001). A shorter duration of symptoms is associated with a worse outcome (Ueoka et al., 2009). The clinical triad includes ataxia, long tract signs and multiple cranial nerve palsies, particularly the sixth and seventh nerves. Clinical course is more protracted in adults when compared to children (Guillamo et al., 2001). Headache is common and can be indistinguishable from migraine including the presence of visual aura (Lim et al., 2005). The single involvement of the facial nerve may initially suggest Bell’s palsy (Walker et al., 1999). Very rarely, spontaneous remission can occur (Lenard et al., 1998). Duration of symptoms in patients with cervicomedullary gliomas range from months to years and it is shorter in fibrilary astrocytomas than in pilocytic astrocytomas and gangliogliomas (Di Maio et al., 2009). Due to preponderance of low-grade tumors, the clinical course is protracted and delay in diagnosis is frequent. Clinical manifestations are related to the primary site of the lesion, medulla or spinal cord, but in most patients, symptoms are overlapping. Involvement of medulla leads to lower cranial nerve dysfunction. Nasal speech, swallowing problems, head tilt, palate deviation and facial nerve palsies dominate the clinical picture. Nausea and vomiting are particularly common and related to bulbar compression and not to increased intracranial pressure. Children may present with failure to thrive. Predominant spinal cord lesions produce weakness, muscle wasting, sphincteric disturbances and may lead to respiratory failure. Patients with dorsally exophytic tumors present with headache, vomiting, failure to thrive, torticollis and ataxia. Signs of long tract dysfunction are not common at presentation, but may appear over the course of the disease.

Differential Diagnosis Infectious, autoimmune, vascular, metabolic and tumoral disorders may have a similar clinical or radiological presentation of a brainstem glioma

389 Table 38.1 Differential diagnosis of brainstem gliomas Infectious Tuberculosis Listeria monocytogenes infection Toxoplasmosis Pyogenic abscess Viral rhombencephalitis Progressive multifocal leukoencephalopathy Autoimmune Sarcoidosis NeuroBehçet Multiple sclerosis Neuromyelitis optica Systemic eritematous lupus Vascular Haematomas Vascular malformations Vasculitis Infarction Posterior reversible leukoencephalopathy Metabolic/inherited Alexander disease Neurofibromatosis type 1 Tumoral Lymphoma Metastasis Germinoma Acoustic neuroma

(Table 38.1) (Laigle-Donadey et al., 2008; Park et al., 2007). Many infectious agents can penetrate de CNS and provoke lesions in the brainstem. Usually, the process is acute or subacute and clinical picture evolves over days. Conversely, brainstem tumors grow more insidiously and years may elapse until diagnosis. Fever and systemic signs are clues for an infectious process as well as the presence of immune system dysfunction. Tuberculosis, toxoplasmosis and Listeria monocytogenes infection predominantly occur in immunossupressed individuals. Cerebrospinal fluid analysis can be helpful in order to determine the specific agent. Inflammatory lesions secondary to autoimmune process may present with mass effect and/or contrast enhancement, mimicking brainstem tumors. When the neurological lesion is the first manifestation of the disease, diagnosis is particularly challenging and may be obtained only after brain biopsy. A relapsing remitting course, the predilection for young women, and involvement of other sites of the neuraxis suggests multiple

390

sclerosis. Neuromyelits optica affects mainly the optic nerves and spinal cord, but is not restricted to these sites. Anti-acquaporine 4 antibody is present in more than half of the cases. Sole involvement of CNS in sarcoidosis is rare and other organs such as lung and lymph nodes are affected first. Patients with Behçet disease can present with a pseudotumoral brainstem lesion, which can be hard to differentiate from a tumor. History of eye lesions, arthritis and genital and oral ulcers should be inquired and are crucial for diagnosis. Hemorrhagic or ischemic strokes produce sudden symptoms and signs in most patients and only rarely can be mistaken for a brainstem glioma. However, a progressive or relapsing course can be seen in brainstem vasculitis and vascular malformations, casting doubt in diagnosis. In these cases, imaging (MR angiography, CT angiography and digital subtraction angiography) is particularly helpful. NF1 patients may develop T2-weighted hyperintense lesions on MRI named unidentified bright objects in the brainstem. These lesions spontaneously disappear during childhood but can be confused with a tumoral lesion. Alexander disease is a rare degenerative leukodystrophy secondary to a mutation in the glial fibrillary acidic protein (GFAP) gene. In juvenile and adult types, a predominant bulbar involvement can occur. Diagnosis is made through MRI and genetic testing. Brainstem can be affected by other tumors including primary CNS neoplasms and metastasis. The clinical and radiological aspects of these lesions are reviewed elsewhere in this book.

Diagnosis Imaging is fundamental to determine the location and extension of brainstem gliomas. CT imaging can show cysts, calcification or the presence of hydrocephalus, but its sensitivity for posterior fossa lesions is low and definition is poor. MRI is the preferred method since it can differentiate tumors from inflammatory and vascular lesions and special techniques such as spectroscopy and perfusion/diffusion sequences can help to determine tumor grade before surgery. Tectal gliomas are iso or hypointense in T1-weighted and hyperintense in T2-weighted and FLAIR sequences (Fig. 38.1). Lesions are located

M.A. Lima

Fig. 38.1 A fluid attenuation inversion recovery (FLAIR) MR image showing a hyperintense lesion in tectum. Biopsy revealed a pilocytic astrocytoma

in the tectal plate, well circumscribed and show contrast enhancement in less than 20% of the cases. Extension into the tegmentum or thalamus may occur and ventricular dilatation is frequently present (Squires et al., 1994). MRI is particularly important for diagnosis of diffuse pontine gliomas in children, since it has replaced biopsy as gold standard for treatment decisions. Schumacher et al. conducted a multicentric study with 142 patients with brainstem disease, including 78 patients with brainstem gliomas (Schumacher et al., 2007). When clinical, immunological, MRI and CSF data were considered, it was possible to differentiate tumoral form non-tumoral lesions in more than 95% of the cases, obviating the need of histological confirmation. On the other hand, it is believed that pontine gliomas are uniformly high-grade, but in a recent study with 24 patients that underwent biopsy for pontine gliomas, two had low-grade tumors (grade II glioma and pilocytic astrocytoma) (Roujeau et al., 2007). In the same study, the procedure was considered safe with minimal morbidity. Nowadays,

38 Brainstem Gliomas: An Overview

391

Fig. 38.2 (a) A hypointense sagital T1-weighted and (b) a hyperintense FLAIR lesion with ill defined margins typical of a pontine glioma

Fig. 38.4 A dorsally exophytic tumor with heterogeneous contrast enhancement is seen in T1-weighted MR sequence

Fig. 38.3 T2-weighted sagital MR sequence showing an extensive cervicomedullary tumor from the medulla to C6

most specialized centers indicate biopsy only in atypical cases. Characteristically, there is a diffuse enlargement of pons and tumor limits are ill defined due to

its infiltrative nature (Fig. 38.2). The lesion is hyperintense in T2-weighted and mostly hypointense in T1-weighted sequences. Contrast enhancement is not usually observed and its presence is of no prognostic significance. Engulfment of the basilar artery can occur (Jallo et al., 2004). MR spectroscopy may be a useful adjuvant in the diagnosis and follow-up of the pontine gliomas. Elevated choline/creatine (Cho/Cr)

392

and Cho/N-acetylaspartate (NAA) ratios are usually observed (Thakur et al., 2006). Upon MRI, cervicomedullary gliomas are homogeneous T1-weighted iso or hypointense and T2-weighted hyperintense lesions (Fig. 38.3). Di Maio et al. reported that tumors in 8/9 patients showed contrast enhancement (Di Maio et al., 2009). Their growth is limited rostrally by the decussating fibers of the corticospinal tract and tumor expands caudally into the spinal cord. Dorsally exophytic tumors are well-demarcated lesions located in the floor of the fouth ventricle and expand dorsolaterally (Fig. 38.4). Similar to other brainstem tumor, they show T1-weighted hypointensity and T2-weighted hyperintensitiy. Most often, gadolinium enhancement is homogeneous.

Treatment General measures in the treatment of brainstem gliomas include analgesia, management of bulbar symptoms and motor impairment, relief of hydrocephalus and emotional support. Palliation is the main component of care in patients with diffuse pontine gliomas. These issues are best addressed by a multidisciplinary team that should include oncologists, neurologists, neurosurgeons, speech and physical therapists in a brain tumor center. Tectal gliomas are indolent lesions that usually present with signs of raised intracranial pressure. The initial approach is aimed to relieve hydrocephalus either with ventricular shunt placement or third ventriculostomy. While shunt placement has been used traditionally, many patients need revision due to infection or malfunction. More recently, third ventriculostomy has become the preferred method by many neurosurgeons. It is safe and efficacious in the short and long term. Li et al. attempted third ventriculostomy in 18 tectal glioma patients and was successful in all. At follow-up, 89% of the individuals remained shunt-free (Li et al., 2005). The procedure can be repeated if the ventriculostomy has closed. Surgical resection soon after diagnosis is controversial. While some groups favor early resection if complete removal of the mass seems feasible, others sustain a conservative approach and reserve surgery for clinical or radiological (contrast enhancement, enlargement

M.A. Lima

or cystic changes) progression. Another approach is stereotactic biopsy followed by radiation therapy as described below. In all cases, intraoperative neurophysiological monitoring should be performed in order to avoid excessive damage. Radiation therapy can be used as an initial approach soon after stereotactic biopsy in patients with high grade lesions or, late in the course of the disease, after clinical progression of low grade tumors. Either conformal radiation therapy or stereotactic radiosurgery (SR) have been used depending on the size, presence of distinct tumor margins and experience at the center. Doses are in the 9–12 Gy range to the tumor margins when SR is employed. In order to avoid radiation induced damage, peripheral doses higher than 14 Gy should not be used (Kihlstrom et al., 1994). Chemotherapy has a limited role in the management of tectal gliomas, being used mostly after clinical progression. Surgery has no role for diffuse pontine gliomas, except in cases of CSF diversion for hydrocephalus. The diffuse and infiltrative nature of these tumors does not produce any well-defined margins. As described above, it is consensus that biopsy is not necessary for most patients, and it should be performed only in atypical cases. The mainstay of treatment for diffuse pontine gliomas is focal conventional radiation therapy. Doses from 50 to 60 Gy divided in daily fraction of 1.5–2 Gy over 6 weeks are used (Laigle-Donadey et al., 2008). This approach leads to clinical improvement in over 75% of the patients initially but does not improve survival (Hargrave et al., 2006). The median onset of disease progression after radiation therapy is < 6 months. The radiological response occurs only in half of the cases. In the last two decades, several studies have investigated alternative radiation protocols such as hyperfractionation and dose escalation. In a large phase III trial, 130 patients were randomly assigned to receive local field radiation at doses of 1.8 Gy daily to a total dose of 54 Gy or a twice a day regimen of 1.17 Gy per fraction to a total dose of 70.2 Gy over 6 weeks (Mandell et al., 1999). The results were disappointing: in the conventional radiation group the median time to death was 8.5 months and in the hyperfractionation group was 8 months. In another study, 66 children underwent treatment with 1 Gy twice daily to a total dose of 78 Gy (Packer et al., 1994). This regimen resulted in a prolonged dependency of steroids

38 Brainstem Gliomas: An Overview

and 9/66 patients showed radiological evidence of radiation induced necrosis. There was no improvement in survival when compared to the standard regimen. Various single or multiple drug regimens have been tested in several trials for management of diffuse pontine gliomas either alone or in combination with radiation therapy. Regimens that used chemotherapy prior to radiation therapy usually failed due to early disease progression. Temozolomide, a drug that has extensively used in supratentorial high grade gliomas improving survival, had no impact in diffuse pontine gliomas either with concurrent radiation therapy or at clinical progression (De Sio et al., 2006). Bevacizumab, an antiangiogenic agent that has been used in recurrent high grade supratentorial gliomas also had no impact in the outcome (Torcuator et al., 2009), but may be useful for radiation induced necrosis in pontine glioma patients (Liu et al., 2009). Radiosensitizing drugs also had disappointing results. Overall, chemotherapy has not brought any substantial impact in the prognosis of these patients. Newer approaches such as local delivery of chemotherapeutic agents in order to bypass the blood brain barrier, immunotherapy and gene therapy are currently being evaluated. Due to the poor response to the conventional therapies, trials with experimental treatments should be offered to these patients. Surgical resection is recommended initially for cervicomedullary gliomas, but morbidity of a radical excision may be important. In most series, biopsy or subtotal excision of the tumor can be performed safely. Di Maio et al. achieved subtotal resection in 6/7 patients (Di Maio et al., 2009), while 9/11 patients underwent subtotal resection in Poussaint series (Young Poussaint et al., 1999). Neurological deficits can be transient or permanent, especially bulbar cranial nerve dysfunction and respiratory function. In cases of extensive spinal cord manipulation, intubation for 48–72 h after surgery is advised (Jallo et al., 2004). Early tracheostomy and feeding gastrostomy may be necessary in the presence of swallowing impairment. Cervical kyphosis is a late complication in patients with cervical laminectomies at the time of surgery due to spinal cord extension of the tumor. Since most tumors in the cervicomedullary region are low grade and grow slowly, radiation therapy is deferred until recurrence or clinical progression with the aim of prevents its deleterious effects in the CNS

393

(Laigle-Donadey et al., 2008). In patients with high grade gliomas, radiation therapy can be performed soon after biopsy or resection of the lesion. Chemotherapy role is yet to be defined in cervicomedullary gliomas. It is usually reserved for recurrence, but a recent study showed that chemotherapy soon after subtotal resection may improve tumor/brainstem interface, opening the possibility for a more extensive resection (Di Maio et al., 2009). Surgical resection is indicated for dorsally exophytic tumors. Advances in neurosurgical and imaging techniques allow extensive resection, although total removal of the lesion is not possible in most cases (Jallo et al., 2004). The goal is the removal of the extrinsic component following by debulking of the intrinsic component in order to prevent neurological morbidity (ataxia, long tract dysfunction and nystagmus). Due to the benign nature of these tumors, radiation therapy and chemotherapy are used in cases of recurrence or progression.

Outcome The outcome of brainstem gliomas is related to the location and histological grade of the tumors. Tectal lesions portray a good prognosis and tumors may remain stable for years. Stark et al. followed 12 children for a median time of 9.5 years (Stark et al., 2005). Radiological progression was observed in three patients, but no clinical deterioration occurred. All children had a good neurological performance at the end of follow-up. In Poussaint series, 20/32 children were stable after a mean follow up of 5 years (Poussaint et al., 1998). For the 12 remaining patients who showed clinical or radiological progression, eight showed decreased tumor size and three showed stable residual lesion after surgery and/or radiation therapy. The prognosis of diffuse pontine gliomas remains dismal despite many experimental protocols of radiation and chemotherapy. The mean survival is less than 1 year in most series and the 2-year survival is only 10% (Broniscer and Gajjar, 2004). Age at onset of disease is prognostic since survival is longer in adults then children (Guillamo et al., 2001; Kesari et al., 2008). Tumor control can be achieved in most patients with cervicomedullary gliomas. In a study with 39 patients, the 5-year progression-free and total survivals

394

were 60 and 89% respectively. In another study, 6/11 patients were stable after surgery at a mean follow-up of 5.2 years (Young Poussaint et al., 1999). Dorsally exophytic tumors have a good outcome after surgical resection. Among 18 patients who underwent surgery, only four showed evidence of radiological progression after a median follow-up of 113 months (Pollack et al., 1993).

References Barkovich AJ, Krischer J, Kun LE, Packer R, Zimmerman RA, Freeman CR, Wara WM, Albright L, Allen JC, Hoffman HJ (1990) Brain stem gliomas: a classification system based on magnetic resonance imaging. Pediatr Neurosurg 16(2):73–83 Broniscer A, Gajjar A (2004) Supratentorial high-grade astrocytoma and diffuse brainstem glioma: two challenges for the pediatric oncologist. Oncologist 9(2):197–206 Daglioglu E, Cataltepe O, Akalan N (2003) Tectal gliomas in children: the implications for natural history and management strategy. Pediatr Neurosurg 38(5):223–231 De Sio L, Milano GM, Castellano A, Jenkner A, Fidani P, Dominici C, Donfrancesco A (2006) Temozolomide in resistant or relapsed pediatric solid tumors. Pediatr Blood Cancer 47(1):30–36 Di Maio S, Gul SM, Cochrane DD, Hendson G, Sargent MA, Steinbok P (2009) Clinical, radiologic and pathologic features and outcome following surgery for cervicomedullary gliomas in children. Childs Nerv Syst 25(11):1401–1410 Epstein F (1985) A staging system for brain stem gliomas. Cancer 56(7 Suppl):1804–1806 Epstein FJ, Farmer JP (1993) Brain-stem glioma growth patterns. J Neurosurg 78(3):408–412 Ferner RE (2007) Neurofibromatosis 1 and neurofibromatosis 2: a twenty first century perspective. Lancet Neurol 6(4): 340–351 Guillamo JS, Creange A, Kalifa C, Grill J, Rodriguez D, Doz F, Barbarot S, Zerah M, Sanson M, Bastuji-Garin S, Wolkenstein P (2003) Prognostic factors of CNS tumours in Neurofibromatosis 1 (NF1): a retrospective study of 104 patients. Brain 126(Pt 1):152–160 Guillamo JS, Monjour A, Taillandier L, Devaux B, Varlet P, Haie-Meder C, Defer GL, Maison P, Mazeron JJ, Cornu P, Delattre JY (2001) Brainstem gliomas in adults: prognostic factors and classification. Brain 124(Pt 12):2528–2539 Hargrave D, Bartels U, Bouffet E (2006) Diffuse brainstem glioma in children: critical review of clinical trials. Lancet Oncol 7(3):241–248 Jallo GI, Biser-Rohrbaugh A, Freed D (2004) Brainstem gliomas. Childs Nerv Syst 20(3):143–153 Kesari S, Kim RS, Markos V, Drappatz J, Wen PY, Pruitt AA (2008) Prognostic factors in adult brainstem gliomas: a multicenter, retrospective analysis of 101 cases. J Neurooncol 88(2):175–183 Kihlstrom L, Lindquist C, Lindquist M, Karlsson B (1994) Stereotactic radiosurgery for tectal low-grade gliomas. Acta Neurochir Suppl 62:55–57

M.A. Lima Laigle-Donadey F, Doz F, Delattre JY (2008) Brainstem gliomas in children and adults. Curr Opin Oncol 20(6):662–667 Lenard HG, Engelbrecht V, Janssen G, Wechsler W, Tautz C (1998) Complete remission of a diffuse pontine glioma. Neuropediatrics 29(6):328–330 Li KW, Roonprapunt C, Lawson HC, Abbott IR, Wisoff J, Epstein F, Jallo GI (2005) Endoscopic third ventriculostomy for hydrocephalus associated with tectal gliomas. Neurosurg Focus 18(6A):E2 Lim EC, Wilder-Smith EP, Chong JL, Wong MC (2005) Seeing the light: brainstem glioma causing visual auras and migraine. Cephalalgia 25(2):154–156 Liu AK, Macy ME, Foreman NK (2009) Bevacizumab as therapy for radiation necrosis in four children with pontine gliomas. Int J Radiat Oncol Biol Phys 75(4):1148–1154 Mandell LR, Kadota R, Freeman C, Douglass EC, Fontanesi J, Cohen ME, Kovnar E, Burger P, Sanford RA, Kepner J, Friedman H, Kun LE (1999) There is no role for hyperfractionated radiotherapy in the management of children with newly diagnosed diffuse intrinsic brainstem tumors: results of a Pediatric Oncology Group phase III trial comparing conventional vs. hyperfractionated radiotherapy. Int J Radiat Oncol Biol Phys 43(5):959–964 Packer RJ, Boyett JM, Zimmerman RA, Albright AL, Kaplan AM, Rorke LB, Selch MT, Cherlow JM, Finlay JL, Wara WM (1994) Outcome of children with brain stem gliomas after treatment with 7800 cGy of hyperfractionated radiotherapy. A Childrens Cancer Group Phase I/II Trial. Cancer 74(6):1827–1834 Park KY, Ahn JY, Cho JH, Choi YC, Lee KS (2007) Neuromyelitis optica with brainstem lesion mistaken for brainstem glioma. Case report. J Neurosurg 107(3 Suppl):251–254 Pollack IF, Hoffman HJ, Humphreys RP, Becker L (1993) The long-term outcome after surgical treatment of dorsally exophytic brain-stem gliomas. J Neurosurg 78(6):859–863 Poussaint TY, Kowal JR, Barnes PD, Zurakowski D, Anthony DC, Goumnerova L, Tarbell NJ (1998) Tectal tumors of childhood: clinical and imaging follow-up. AJNR Am J Neuroradiol 19(5):977–983 Qaddoumi I, Sultan I, Gajjar A (2009) Outcome and prognostic features in pediatric gliomas: a review of 6212 cases from the surveillance, epidemiology, and end results database. Cancer 115(24):5761–5770 Roujeau T, Machado G, Garnett MR, Miquel C, Puget S, Geoerger B, Grill J, Boddaert N, Di Rocco F, Zerah M, Sainte-Rose C (2007) Stereotactic biopsy of diffuse pontine lesions in children. J Neurosurg 107(1 Suppl):1–4 Schumacher M, Schulte-Monting J, Stoeter P, Warmuth-Metz M, Solymosi L (2007) Magnetic resonance imaging compared with biopsy in the diagnosis of brainstem diseases of childhood: a multicenter review. J Neurosurg 106(2 Suppl): 111–119 Selvapandian S, Rajshekhar V, Chandy MJ (1999) Brainstem glioma: comparative study of clinico-radiological presentation, pathology and outcome in children and adults. Acta Neurochir (Wien) 141(7):721–726; discussion 726–727 Smith MA, Freidlin B, Ries LA, Simon R (1998) Trends in reported incidence of primary malignant brain tumors in children in the United States. J Natl Cancer Inst 90(17): 1269–1277

38 Brainstem Gliomas: An Overview Squires LA, Allen JC, Abbott R, Epstein FJ (1994) Focal tectal tumors: management and prognosis. Neurology 44(5): 953–956 Stark AM, Fritsch MJ, Claviez A, Dorner L, Mehdorn HM (2005) Management of tectal glioma in childhood. Pediatr Neurol 33(1):33–38 Ternier J, Wray A, Puget S, Bodaert N, Zerah M, SainteRose C (2006) Tectal plate lesions in children. J Neurosurg 104(6 Suppl):369–376 Thakur SB, Karimi S, Dunkel IJ, Koutcher JA, Huang W (2006) Longitudinal MR spectroscopic imaging of pediatric diffuse pontine tumors to assess tumor aggression and progression. AJNR Am J Neuroradiol 27(4):806–809

395 Torcuator R, Zuniga R, Loutfi R, Mikkelsen T (2009) Bevacizumab and irinotecan treatment for progressive diffuse brainstem glioma: case report. J Neurooncol 93(3): 409–412 Ueoka DI, Nogueira J, Campos JC, Maranhao Filho P, Ferman S, Lima MA (2009) Brainstem gliomas–retrospective analysis of 86 patients. J Neurol Sci 281(1–2):20–23 Walker DA, Punt JA, Sokal M (1999) Clinical management of brain stem glioma. Arch Dis Child 80(6):558–564 Young Poussaint T, Yousuf N, Barnes PD, Anthony DC, Zurakowski D, Scott RM, Tarbell NJ (1999) Cervicomedullary astrocytomas of childhood: clinical and imaging follow-up. Pediatr Radiol 29(9):662–668

Chapter 39

Tumor-Associated Epilepsy in Patients with Glioma Anja Smits and Anette Storstein

Abstract Epileptic seizures are common symptoms in patients with glioma, occurring at disease presentation as well as during later disease. The frequency of tumor-related seizures is strongly related to the growth rate of the tumor. In low-grade gliomas, seizures as initial symptom leading to the brain tumor diagnosis occur in 70–90% of all patients. About 50% of these patients continue to have seizures before operation in spite of optimal antiepileptic drug treatment. In high-grade gliomas, seizures at disease presentation are less frequent and occur in only 30–50% of all patients but may be more difficult to control. Apart from the tumor growth rate, the specific location of the tumor in the brain and its proximity to the cortex affect seizure risk and seizure control in glioma. For many patients with tumor-related seizures, optimal seizure control is not achieved by antiepileptic drugs only but requires additional therapy such as surgical resection and/or radio- and chemotherapy. In this review, we discuss different aspects of the clinical presentation of the disease in patients with tumor-related seizures. We also discuss the possible therapeutic strategies for patients with medically refractory seizures, illustrating the importance of a qualified specialized team of neuro-oncologists, neurologists and neurosurgeons for the optimal management of this patient group.

A. Smits () Department of Neuroscience, Neurology, University Hospital Uppsala, S-751 85 Uppsala, Sweden e-mail: [email protected]

Keywords Epilepsy · Glioma · Antiepileptic drugs · Epileptogenesis · Differential diagnoses · Ischemic infarction

Introduction Epileptic seizures are common in patients with gliomas, affecting the majority of patients with lowgrade tumors and a substantial proportion of patients with high-grade glioma. Seizures occur as initial symptoms at disease onset and are frequent symptoms of chronic tumor disease as well. With tumor treatments for patients with glioma becoming more effective resulting in longer patient survival, the burden of chronic epilepsy will add substantial morbidity to this patient group. Several studies have shown that persistent seizures in patients with brain tumors have sustained negative effects on daily functioning and quality of life (Klein et al., 2003). Optimal seizure control is therefore an important part of the clinical management in patients with brain tumors. Cognitive dysfunction, another main symptom in patients with glioma, is often closely related to both refractory seizures and antiepileptic therapy. The clinical management of cognitive dysfunction in patients with brain tumors is discussed elsewhere. The incidence of seizures in patients with glioma is highly variable and is affected by several factors. The location of the tumor in the brain, the proximity to adjacent cortex and the growth rate of the tumor are important parameters determining seizure risk. The majority of patients with slowly growing low-grade gliomas presents with epileptic seizures at disease onset or within the first year from tumor diagnosis,

M.A. Hayat (ed.), Tumors of the Central Nervous System, Volume 2, DOI 10.1007/978-94-007-0618-7_39, © Springer Science+Business Media B.V. 2011

397

398

often as the only symptom. In patients with high-grade gliomas, epileptic seizures occur less frequently both at disease onset as well as during later course of disease. In addition to symptomatic seizures, these tumors cause neurological deficits as a result of the destructive growth in the brain. Apart from the evident variability in seizure risk at disease onset, the seizure control that can be obtained in patients with gliomas is highly variable. Tumor-related seizures in high-grade gliomas are less frequent, but the epileptic symptoms of these patients may be difficult to control. Many patients with slowly growing gliomas suffer from chronic epilepsy during the entire course of disease, whereas others become seizure-free after their first initial seizure when antiepileptic treatment is started. Add-on treatment with new antiepileptic drugs may increase the proportion of seizure-free patients in the glioma population, but further studies need to assess the efficiency of these drugs in glioma-related seizures as well as the side effects induced by polypharmacy. Other treatment options should be considered when antiepileptic drug treatment fails. Surgical resection including resection of the epileptic foci is known to have beneficial effects on seizure control, especially in low-grade gliomas. Recent reports have shown positive effects with seizure reduction obtained by radiotherapy and chemotherapy as well in this patient group. In general, however, current therapy of tumor-related seizures is far from perfect (van Breemen et al., 2007). A deeper understanding of the underlying mechanisms of brain-tumor epilepsy is a prerequisite for improving seizure control and optimizing the clinical management of gliomarelated seizures.

Epidemiology In general, slowly growing tumors are significantly more epileptogenic than fast growing tumors. Epileptic seizures as initial symptoms leading to the brain tumor diagnosis occur in 70–90% of low-grade gliomas and in 30–50% of high-grade gliomas. Some specific histological tumor subtypes characterized by a typically indolent course of disease, such as dysembryoblastic neuroepithelial tumors (DNET) and gangliogliomas, are associated with a seizure frequency of almost 100%. In contrast, the incidence of seizures as first

A. Smits and A. Storstein

symptoms in metastatic brain disease is as low as 20–30% (van Breemen et al., 2007). Around 4–5% of all adult-onset seizures are caused by brain tumors (Olafsson et al., 2005). However, seizures may go unnoticed for long time periods. The difficulty in defining initial seizures in adults as first symptoms at the time point of disease onset is illustrated in a study by King et al. (1998) who investigated 300 consecutive adults and children presenting with an unexplained seizure. A clinical diagnosis of generalized or focal epilepsy could be made in 141 (47%) of these patients, and imaging of the brain by CT or MRI identified 38 epileptogenic lesions (17 of these were tumors). It was demonstrated that 17% of the patients who presented with their first tonic-clonic seizure reported similar events in the past and 28% had earlier experienced focal epileptic symptoms, such as temporal lobe auras due to non-convulsive seizures. This study confirms the common experience by many doctors that patients may not be aware of or not appreciate the importance of minor epileptic symptoms, and fail to mention these in an initial interview. As a consequence, the brain tumor diagnosis may be delayed for months or, in case of slowly growing tumors, up to several years.

Epileptogenesis in Gliomas The mechanisms behind the development of tumorrelated seizures are likely to involve tumor-related as well as individual patient-related factors but are incompletely understood. The seizure risk and the individual response to antiepileptic drug treatment are variable within subgroups of patients with similar tumor localization and histology. Thus, apart from tumor location and tumor growth rate, other factors that are related to the genetic susceptibility of the individual patient and the intrinsic properties of the host brain may contribute to the development of glioma-related seizures. A putative mechanism of acquiring intrinsic epileptic properties is by altered expression of signaling molecules involved in neuronal circuitries. According to the modern theory of neuronal networks, synchronization of neurons in complex networks is the prerequisite for normal neurological functioning. Due to modifications in such existing circuitries by for example tumor growth, active processes of axonal sprouting,

39 Tumor-Associated Epilepsy in Patients with Glioma

synaptogenesis and neurogenesis will take place in the peritumoral region (Brogna et al., 2008). Such excessive and random synchronization of neurons may persist during critical periods of time, typically in slowly growing tumors, and cause epilepsy. This theory provides new insights also in the underlying mechanisms of cognitive disturbances, the other main symptom of disease in low-grade gliomas. Thus, disarrangement of the same neuronal cerebral networks may lead to cognitive disturbances, as a result of destruction, and epileptic seizures, as a result of pathological excess of functional network structures (Brogna et al., 2008).

Pathophysiology The pathophysiology of seizure development in patients with glioma is thought to occur through different mechanisms in high-grade and low-grade tumors. In fast growing high-grade gliomas, focal ischemia and deafferentiation of cortical areas due to mass effect are likely to be important causative factors (Fig. 39.1), whereas gliosis and chronic inflammatory changes in peritumoral regions of slowly growing gliomas are considered to predispose for epileptic seizures (Fig. 39.2). Increased levels of Fe3+ ions deposits in intra- or peritumoral areas, due to small bleedings from pathological blood vessels, typically occur in high-grade gliomas and may further contribute to the development of tumor-associated seizures (Fig. 39.1) (Beaumont and Whittle, 2000; Schaller and Rüegg, 2003). The almost 100% association of epileptic seizures in patients with specific histological types of neuroglial tumors, such as DNET and gangliogliomas, has raised the question whether these tumors contain intrinsic epileptogenic properties. It has been speculated that the specific neurochemical profiles of these lesions predispose for generating epileptic activity (Beaumont and Whittle, 2000).

Tumor Characteristics The inverse relationship between tumor growth rate and seizure risk is true also for patients with tumors of similar histological malignancy grade. Thus, within the group of grade II gliomas, the slowly growing oligodendrogliomas are more epileptogenic than the

399

diffuse astrocytomas. Other important factors underlying the development of epilepsy are the localization of the tumor in the brain and the proximity with the cortical gray matter (Schaller and Ruegg, 2003). Tumors with location in the vicinity of the primary motor cortex and with limbic and perilimbic cortical localization are highly epileptogenic, whereas occipital tumors are less likely to manifest with seizures. Low-grade gliomas are more frequently than de novo glioblastomas situated in eloquent cortico-subcortical regions, such as the supplementary motor area (SMA) and the insular and peri-insular area. These locations are highly epileptogenic areas and strongly associated with medically refractory seizures (Duffau and Capelle, 2004).

Peritumoral Brain In tumor-associated epilepsy, the tumor somehow acts as a generator to produce an epileptic focus in the brain. In many cases, however, the epileptic focus, i.e. the location corresponding with the start of the seizure, is not contiguous with the tumor. The interplay between the tumor histology and the surrounding tissue is of importance for seizure development, especially in chronic epilepsy of patients with slowly growing gliomas. Alterations in the ultrastructure of the peritumoral brain regions have been demonstrated in patients with tumor-associated seizures, as well as changes in metabolic activity and in pH (Beaumont and Whittle, 2000). The presence of necrotic tissue and the hypoxia of fast growing high-grade gliomas due to pathological vascularization may induce changes in the adjacent regions with the potential to increase neuronal excitability (Fig. 39.1). Some patients with chronic tumor-related seizures may develop secondary epileptic foci that do not correspond to the tumor or peritumoral area, although this phenomenon is still far more controversial in human epilepsy compared to animal models (Cibula and Gilmore, 1997). It is clear though that epileptic activity in regions distant to the tumor is more frequently found in patients with a long seizure history and with temporal lobe tumors. Consistent with the existence of separate epileptic foci, these patients tend to have two different types of focal seizures (Cibula and Gilmore, 1997). Unless complete resection of

400

A. Smits and A. Storstein

a c

b d

e

Fig. 39.1 Micrographs of histopathological sections of a glioblastoma (kindly provided by Dr Aronica, Department of Neuropathology, University of Amsterdam, the Netherlands). (a) Hematoxylin-eosin staining shows characteristic tumor morphology with necrotic areas; (b) Immunostaining with

a

d

anti-GFAP-antibody shows invasion of the tumor in the cortex; (c) Hematoxylin-eosin staining at higher magnification shows polymorphic tumor cells and (d) proliferating blood vessels; (e) Immunostaining with anti-Ki6 antibody shows a large amount of proliferating tumor cells

b

e

Fig. 39.2 Micrographs of histopathological sections of an astrocytoma grade II (kindly provided by Dr Aronica, Department of Neuropathology, University of Amsterdam, the Netherlands). (a) Hematoxylin-eosin staining shows a well-differentiated tumor at low, respectively; (b) high magnification; (c) and the infiltration of the tumor in the cortex;

c

f (d) Immunostaining with anti-GFAP antibody shows immunoreactive astrocytic tumor cells; (e) whereas the tumor cells lack immunoreactivity for neuronal nuclear protein; (f) Immunostaining with anti-Ki6 antibody shows few (