Tumors of the Central Nervous System, Volume 5: Astrocytomas, Hemangioblastomas, and Gangliogliomas

Tumors of the Central Nervous System

Tumors of the Central Nervous System Volume 5

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

Tumors of the Central Nervous System Volume 5

Tumors of the Central Nervous System Astrocytomas, Hemangioblastomas, and Gangliogliomas 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 978-94-007-2018-3 e-ISBN 978-94-007-2019-0 DOI 10.1007/978-94-007-2019-0 Springer Dordrecht Heidelberg London New York Library of Congress Control Number: 2011936737 © Springer Science+Business Media B.V. 2012 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

It is recognized that scientific journals and books not only provide current information but also facilitate exchange of information, resulting in rapid progress in the medical field. In this endeavor, the main role of scientific books is to present current information in more detail after careful additional evaluation of the investigational results, especially those of new or relatively new therapeutic methods and their potential toxic side-effects. Although subjects of diagnosis, drug development, therapy and its assessment, and prognosis of tumors of the central nervous system, cancer recurrence, and resistance to chemotherapy are scattered in a vast number of journals and books, there is need of combining these subjects in single volumes. An attempt will be made to accomplish this goal in the projected ten-volume series of handbooks. In the era of cost-effectiveness, my opinion may be minority perspective, but it needs to be recognized that the potential for false-positive or false-negative interpretation on the basis of a single laboratory test in clinical pathology does exist. Interobservor or intraobservor variability in the interpretation of results in pathology is not uncommon. Interpretative differences often are related to the relative importance of the criteria being used. Generally, no test always performs perfectly. Although there is no perfect remedy to this problem, standardized classifications with written definitions and guidelines will help. Standardization of methods to achieve objectivity is imperative in this effort. The validity of a test should be based on the careful, objective interpretation of the tomographic images, photo-micrographs, and other tests. The interpretation of the results should be explicit rather than implicit. To achieve accurate diagnosis and correct prognosis, the use of molecular criteria and targeted medicine is important. Equally important are the translation of molecular genetics into clinical practice and evidence-based therapy. Translation of medicine from the laboratory to clinical application needs to be carefully expedited. Indeed, molecular medicine has arrived. This is the fifth volume in the series, Tumors of the Central Nervous System. As in the case of the four previously published volumes, this volume mainly contains information on the diagnosis, therapy, and prognosis of brain tumors.Various aspects of three types of brain tumors (Astrocytomas, Hemangioblastoma and Ganglioglioma) are discussed. Insights into the understanding of molecular pathways involved in tumor biology are explained, which lead to the development of effective drugs. Information on pathways facilitates targeted therapies in cancer. Tumor models are also presented, which utilize expression data, pathway sensitivity, and genetic abnormalities, representing targets in cancer. vii

viii

Preface

Advantages and limitations of chemotherapy (e.g., Cisplatin/carboplatin combination) for patients with pilomyxoid astrocytoma are discussed. Identification and characterization of biomarkers, including those for metastatic brain tumors, are presented. Genomic analyses for identifying clinically relevant subtypes are included. A number of imaging modalities, including time-resolved laser fluorescence spectroscopy and magnetic resonance- guided laser interstitial thermal therapy are detailed to diagnose and treat brain tumors. Introduction to new technologies and their applications to tumor diagnosis, treatment, and therapy assessment are explained. For example, nanotechnology-based therapy for malignant tumors of the CNS is explained. Molecular profiling of brain tumors to select therapy in clinical trials of brain tumors is included. Several surgical treatments, including resection, and radiosurgery, are discussed. The remaining two volumes in this series will provide additional recent information on this and other aspects of other types of CNS malignancies. By bringing together a large number of experts (oncologists, neurosurgeons, physicians, research scientists, and pathologists) in various aspects of this medical field, it is my hope that substantial progress will be made against this terrible disease. It would be difficult for a single author to discuss effectively the complexity of diagnosis, therapy, and prognosis of any type of tumor in one volume. Another advantage of involving more than one author is to present different points of view on a specific controversial aspect of the CNS cancer. I hope these goals will be fulfilled in this and other volumes of this series. This volume was written by 85 contributors representing 14 countries. I am grateful to them for their promptness in accepting my suggestions. Their practical experience highlights their writings, which should build and further the endeavors of the reader in this important area of disease. I respect and appreciate the hard work and exceptional insight into the nature of cancer provided by these contributors. The contents of the volume are divided into seven subheadings: Introduction, Diagnosis and Biomarkers, Therapy, Tumor to tumor cancer, Imaging methods, Prognosis, and Quality of life for the convenience of the reader. It is my hope that the current volume will join the preceding volumes of the series for assisting in the more complete understanding of globally relevant cancer syndromes. There exist a tremendous, urgent demand by the public and the scientific community to address to cancer, diagnosis, treatment, cure, and hopefully prevention. In the light of existing cancer calamity, government funding must give priority to eradicating this deadly malignancy over military superiority. I am thankful to Dr. Dawood Farahi and Dr. Kristie Reilly for recognizing the importance of medical research and publishing through an institution of higher education. Union, New Jersey April 2011

M.A. Hayat

Contents

Part I

Astrocytomas: Diagnosis and Biomarkers . . . . . . . . . . . .

1

1 Methylation in Malignant Astrocytomas . . . . . . . . . . . . . . . . María del Mar Inda, Juan A. Rey, Xing Fan, and Javier S. Castresana

3

2 Deciphering the Function of Doppel Protein in Astrocytomas . . . . Alberto Azzalin and Sergio Comincini

13

3 Astrocytic Tumors: Role of Antiapoptotic Proteins . . . . . . . . . . Alfredo Conti, Carlo Gulì, Giuseppe J. Sciarrone, and Chiara Tomasello

23

4 Deregulation of the Wnt/β-Catenin/Tcf Signaling Pathway in Astrocytomas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Gangadhara Reddy Sareddy and Phanithi Prakash Babu

35

5 Subependymal Giant Cell Astrocytoma: Role of mTOR Pathway and Its Inhibitors . . . . . . . . . . . . . . . . . . . . . . . . Katarzyna Kotulska and Sergiusz Jó´zwiak

45

6 Role of Progesterone Receptor Isoforms in Human Astrocytomas Growth . . . . . . . . . . . . . . . . . . . . . . . . . . Ignacio Camacho-Arroyo, Valeria Hansberg-Pastor, Edith Cabrera-Muñoz, Olivia Tania Hernández-Hernández, and Aliesha González-Arenas 7 Astrocytic Tumors: Role of Carbonic Anhydrase IX . . . . . . . . . Joonas Haapasalo, Hannu Haapasalo, and Seppo Parkkila 8 Development of Cysts in Pilocytic Astrocytomas: Role of Eosinophilic Granular Bodies (Method) . . . . . . . . . . . . . . . Jai-Nien Tung, Tang-Yi Tsao, Kun-Tu Yeh, Ching-Fong Liao, and Ming-Chung Jiang 9 Role of Synemin in Astrocytoma Cell Migration . . . . . . . . . . . . Quincy Quick, Yihang Pan, and Omar Skalli 10 Diffuse Astrocytomas: Immunohistochemistry of MGMT Expression . . . . . . . . . . . . . . . . . . . . . . . . . . David Capper

57

65

73

81

89

ix

x

11

12

13

14

15

16

Contents

Central Nervous System Germ Cell Tumors: An Epidemiology Review . . . . . . . . . . . . . . . . . . . . . . . . Daniel L. Keene and Donna Johnston

95

RAF Fusion Genes and MAPK Activation in Pilocytic Astrocytomas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Sally R. Lambert and David T.W. Jones

99

Biomarker Discovery in Central Nervous System Neoplasms: Past, Present and Future . . . . . . . . . . . . . . . . . . Anne F. Buckley, Roger E. McLendon, Carol J. Wikstrand, and Darell D. Bigner Astrocytomas: Role of Taurine in Apoptosis Using Magnetic Resonance Spectroscopy . . . . . . . . . . . . . . . . . . . . . . . . . Kirstie S. Opstad

121

Imaging of Hypoxia-Inducible Factor-1-Active Regions in Tumors Using a POS and 123 I-IBB Method . . . . . . . . . . . . . Masashi Ueda and Hideo Saji

129

Diffuse Low-Grade Astrocytomas: P53-Mediated Inhibition of Angiogenesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Timo Gaiser and Markus D. Siegelin

135

17

Spontaneous Regression of Cerebellar Astrocytomas . . . . . . . . . Mansoor Foroughi, Shibu Pillai, and Paul Steinbok

18

Subependymal Giant Cell Astrocytoma: Gene Expression Profiling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Magdalena Ewa Tyburczy and Bozena Kaminska

Part II 19

20

21

22

23

107

Astrocytomas: Therapy . . . . . . . . . . . . . . . . . . . . . .

Time-Resolved Laser Induced Fluorescence Spectroscopy (TRLIFS): A Tool for Intra-operative Diagnosis of Brain Tumors and Maximizing Extent of Surgical Resection . . . . . . . . Pramod Butte and Adam N. Mamelak

143

149

159

161

Magnetic Resonance-Guided Laser Interstitial Thermal Therapy for Brain Tumors . . . . . . . . . . . . . . . . . . . . . . . . Kevin Beccaria, Michael S. Canney, and Alexandre C. Carpentier

173

Nanotechnology-Based Therapy for Malignant Tumors of the Central Nervous System . . . . . . . . . . . . . . . . . . . . . Abraham Boskovitz, Abdullah Kandil, and Al Charest

187

Pilocytic Astrocytoma: Pathological and Immunohistochemical Factors Affecting Surgical Treatment and Surveillance . . . . . . . . . . . . . . . . . . Devon Haydon and Jeffrey Leonard Pilomyxoid Astrocytomas: Chemotherapy . . . . . . . . . . . . . . . Hitoshi Tsugu, Shinya Oshiro, Fuminari Komatsu, Hiroshi Abe, Takeo Fukushima, Tooru Inoue, Fumio Yanai, and Yuko Nomura

195 203

Contents

xi

Part III

Astrocytomas: Prognosis . . . . . . . . . . . . . . . . . . . . .

211

24 Astrocytomas: Predicting Survival and Recurrence Using Cerebral Blood Volume Measurements . . . . . . . . . . . . . . . . . Sotirios Bisdas

213

25 Electronic Patient-Reported Outcome Monitoring (ePROM) in Brain Tumour Patients . . . . . . . . . . . . . . . . . . . . . . . . Lisa M. Wintner, Johannes M. Giesinger, Gabriele Schauer-Maurer, and Bernhard Holzner Part IV

223

Hemangioblastoma . . . . . . . . . . . . . . . . . . . . . . . .

231

26 Intra-operative ICG Use in the Management of Hemangioblastomas . . . . . . . . . . . . . . . . . . . . . . . . . . Loyola V. Gressot and Steven W. Hwang

233

27 Hemangioblastoma Cysts: Diagnosis Using Fluorescence with 5-Aminolevulinic Acid . . . . . . . . . . . . . . . . . . . . . . . Satoshi Utsuki, Hidehiro Oka, and Kiyotaka Fujii

239

28 Hemangioblastoma: Stereotactic Radiosurgery . . . . . . . . . . . . Anand Veeravagu, Bowen Jiang, and Steven D. Chang

245

Part V

Ganglioglioma . . . . . . . . . . . . . . . . . . . . . . . . . . .

251

29 Gangliogliomas: Molecular Pathogenesis and Epileptogenesis . . . . Eleonora Aronica and Pitt Niehusmann

253

30 Epilepsy-Associated Gangliogliomas: Identification of Genes with Altered Expression . . . . . . . . . . . . . . . . . . . . . . . . . Albert J. Becker

267

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

275

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

Contents of Volume 2

1 Introduction 2 Gliomagenesis: Advantages and Limitations of Biomarkers 3 Molecular Subtypes of Gliomas 4 Glioblastoma: Germline Mutation of TP53 5 Familial Gliomas: Role of TP53 Gene 6 The Role of IDH1 and IDH2 Mutations in Malignant Gliomas 7 Malignant Glioma: Isocitrate Dehydrogenases 1 and 2 Mutations 8 Metabolic Differences in Different Regions of Glioma Samples 9 Glioblastoma Patients: Role of Methylated MGMT 10 Brain Tumor Angiogenesis and Glioma Grading: Role of Tumor Blood Volume and Permeability Estimates Using Perfusion CT 11 Vasculogenic Mimicry in Glioma 12 Newly Diagnosed Glioma: Diagnosis Using Positron Emission Tomography with Methionine and Fluorothymidine 13 Role of Diffusion Tensor Imaging in Differentiation of Glioblastomas from Solitary Brain Metastases 14

131 I-TM-601

SPECT imaging of Human Glioma

15 Assessment of Biological Target Volume Using Positron Emission Tomography in High-Grade Glioma Patients 16 Skin Metastases of Glioblastoma 17 Diffuse Low-Grade Gliomas: What Does “Complete Resection” Mean? 18 Quantitative Approach of the Natural Course of Diffuse Low-Grade Gliomas 19 Impact of Extent of Resection on Outcomes in Patients with High-Grade Gliomas

xv

xvi

Contents of Volume 2

20

Glioma Surgery: Intraoperative Low Field Magnetic Resonance Imaging

21

Low-Grade Gliomas: Intraoperative Electrical Stimulations

22

Malignant Gliomas: Present and Future Therapeutic Drugs

23

Recurrent Malignant Glioma Patients: Treatment with Conformal Radiotherapy and Systemic Therapy

24

Glioblastoma: Boron Neutron Capture Therapy

25

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

26

Glioblastoma Patients: Chemotherapy with Cisplatin, Temozolomide and Thalidomide

27

Glioblastoma: Role of Galectin-1 in Chemoresistance

28

Glioma-Initiating Cells: Interferon Treatment

29

Glioblastoma: Anti-tumor Action of Natural and Synthetic Cannabinoids

30

Patients with Recurrent High-Grade Glioma: Therapy with Combination of Bevacizumab and Irinotecan

31

Monitoring Gliomas In Vivo Using Diffusion-Weighted MRI During Gene Therapy-Induced Apoptosis

32

High-Grade Gliomas: Dendritic Cell Therapy

33

Glioblastoma Multiforme: Use of Adenoviral Vectors

34

Fischer/F98 Glioma Model: Methodology

35

Cellular and Molecular Characterization of Anti-VEGF and IL-6 Therapy in Experimental Glioma

36

Adult Brainstem Gliomas: Diagnosis and Treatment

37

The Use of Low Molecular Weight Heparin in the Treatment and Prevention of Thromboembolic Disease in Glioma Patients

38

Brainstem Gliomas: An Overview

39

Tumor-Associated Epilepsy in Patients with Glioma

40

Brain Tumors Arising in the Setting of Chronic Epilepsy

41

Low-Grade Gliomas: Role of Relative Cerebral Blood Volume in Malignant Transformation

42

Angiocentric Glioma-Induced Seizures: Lesionectomy

Contents of Volume 3

1 Introduction 2 Brain Tumor Classification Using Magnetic Resonance Spectroscopy 3 Cellular Immortality in Brain Tumors: An Overview 4 Tumor-to-Tumor Metastasis: Extracranial Tumor Metastatic to Intracranial Tumors 5 Brain Metastases from Breast Cancer: Treatment and Prognosis 6 Brain Metastasis in Renal Cell Carcinoma Patients 7 Coexsistence of Inflammatory Myofibroblastic Tumor in the Lung and Brain 8 Breast Cancer and Renal Cell Cancer Metastases to the Brain 9 Breast Cancer Brain Metastases: Genetic Profiling and Neurosurgical Therapy 10 Central Nervous System Tumours in Women Who Received Capecitabine and Lapatinib Therapy for Metastatic Breast Cancer 11 Functional Role of the Novel NRP/B Tumor Suppressor Gene 12 Brain Tumors: Diagnostic Impact of PET Using Radiolabelled Amino Acids 13 Malignant Peripheral Nerve Sheath Tumors: Use of 18FDG-PET/CT 14 Brain Tumors: Evaluation of Perfusion Using 3D-FSEPseudo-Continuous Arterial Spin Labeling 15 Cerebral Cavernous Malformations: Advanced Magnetic Resonance Imaging 16 Nosologic Imaging of Brain Tumors Using MRI and MRSI 17 Brain Tumor Diagnosis Using PET with Angiogenic Vessel-Targeting Liposomes 18 Frozen Section Evaluation of Central Nervous System Lesions 19 Clinical Role of MicroRNAs in Different Brain Tumors

xvii

xviii

Contents of Volume 3

20

Electrochemotherapy for Primary and Secondary Brain Tumors

21

Brain Tumors: Convection-Enhanced Delivery of Drugs (Method)

22

Brain Metastases: Clinical Outcomes for Stereotactic Radiosurgery (Method)

23

Noninvasive Treatment for Brain Tumors: Magnetic Resonance-Guided Focused Ultrasound Surgery

24

Radioguided Surgery of Brain Tumors

25

Implications of Mutant Epidermal Growth Factor Variant III in Brain Tumor Development and Novel Targeted Therapies

26

Endoscopic Port Surgery for Intraparenchymal Brain Tumors

27

Intracranial Tumor Surgery in Elderly Patients

28

Intracranial Hemangiopericytoma: Gamma Knife Surgery

29

Stereotactic Radiosurgery for Cerebral Metastases of Digestive Tract Tumors

30

Malignant Brain Tumors: Role of Radioresponsive Gene Therapy

31

Brain Tumors: Quality of Life

32

Health-Related Quality of Life in Patients with High Grade Gliomas

33

Epilepsy and Brain Tumours and Antiepileptic Drugs

34

Familial Caregivers of Patients with Brain Cancer

35

Pain Management Following Craniotomy

36

Air Transportation of Patients with Brain Tumours

Contents of Volume 4

1 Epidemiology of Primary Brain Tumors 2 Supratentorial Primitive Neuroectodermal Tumors 3 Epileptic Seizures and Supratentorial Brain Tumors in Children 4 Breast Cancer Metastasis to the Central Nervous System 5 Melanoma to Brain Metastasis: Photoacoustic Microscopy 6 Extraaxial Brain Tumors: The Role of Genetic Polymorphisms 7 Central Nervous System Germ Cell Tumor 8 Microvascular Gene Changes in Malignant Brain Tumors 9 Role of MicroRNA in Glioma 10 Glioblastoma Multiforme: Cryopreservation of Brain Tumor-Initiating Cells (Method) 11 Relationship Between Molecular Oncology and Radiotherapy in Malignant Gliomas (An Overview) 12 High-Grade Brain Tumours: Evaluation of New Brain Lesions by Amino Acid PET 13 Cyclic AMP Phosphodiesterase-4 in Brain Tumor Biology: Immunochemical Analysis 14 Molecular Imaging of Brain Tumours Using Single Domain Antibodies 15 Quantitative Analysis of Pyramidal Tracts in Brain Tumor Patients Using Diffusion Tensor Imaging 16 Differentiation Between Gliomatosis Cerebri and Low-Grade Glioma: Proton Magnetic Resonance Spectroscopy 17 Peripheral Nerve Sheath Tumors: Diagnosis Using Quantitative FDG-PET 18 Tumor Resection Control Using Intraoperative Magnetic Resonance Imaging

xix

xx

Contents of Volume 4

19

Brain Tumors: Clinical Applications of Functional Magnetic Resonance Imaging and Diffusion Tensor Imaging

20

Trigeminal Neuralgia: Diagnosis Using 3-D Magnetic Resonance Multi-Fusion Imaging

21

Epilepsy-Associated Brain Tumors: Diagnosis Using Magnetic Resonance Imaging

22

Growth of Malignant Gliomas In Vivo: High-Resolution Diffusion Tensor Magnetic Resonance Imaging

23

Resection of Brain Lesions: Use of Preoperative Functional Magnetic Resonance Imaging and Diffusion Tensor Tractography

24

Paradigms in Tumor Bed Radiosurgery Following Resection of Brain Metastases

25

Rat Model of Malignant Brain Tumors: Implantation of Doxorubicin Using Drug Eluting Beads for Delivery

26

Electromagnetic Neuronavigation for CNS Tumors

27

Stereotactic Radiosurgery for Intracranial Ependymomas

28

Is Whole Brain Radiotherapy Beneficial for Patients with Brain Metastases?

29

Triggering Microglia Oncotoxicity: A Bench Utopia or a Therapeutic Approach?

30

Preoperative Motor Mapping

31

Intraoperative Monitoring for Cranial Base Tumors

32

Brain Tumours: Pre-clinical Assessment of Targeted, Site Specific Therapy Exploiting Ultrasound and Cancer Chemotherapeutic Drugs

33

Headaches in Patients with Brain Tumors

34

Headache Associated with Intracranial Tumors

35

Patients with Brain Cancer: Health Related Quality of Life

36

Emerging Role of Brain Metastases in the Prognosis of Breast Cancer Patients

Contributors

Hiroshi Abe Department of Neurosurgery, Fukuoka University Faculty of Medicine, Jonan-ku, Fukuoka 814-0180, Japan Eleonora Aronica Department of Neuropathology, Academic Medical Center, 1105 AZ, Amsterdam, The Netherlands, [email protected] Alberto Azzalin Institute of Molecular Genetics, IGM-CNR Pavia via Abbiategrasso 207 (OR via Ferrata 1), 27100 Pavia, Italy, [email protected] Phanithi Prakash Babu Department of Biotechnology, School of Life Sciences, University of Hyderabad, Hyderabad 500 046, India, [email protected] Kevin Beccaria Department of Neurosurgery and Advanced Surgical Technologies Research Team, Hopital de la Pitie-Salpetriere, Assistance Publique Hopitaux de Paris, Université Paris VI – Pierre & Marie Curie, 75013 Paris, France Albert J. Becker Department of Neuropathology, University of Bonn Medical Center, D-53105 Bonn, Germany, [email protected] Darell D. Bigner Section(s) of Surgical Pathology, Duke University Medical Center, Durham, NC 27710, USA Sotirios Bisdas Department of Diagnostic and Interventional Neuroradiology, Karls Eberhard University, Tübingen, Germany, [email protected] Abraham Boskovitz Department of Neurosurgery, Tufts Medical Center, Tufts University, Boston, MA 02111, USA Anne F. Buckley Section(s) of Surgical Pathology, Duke University Medical Center, Durham, NC 27710, USA Pramod Butte Department of Neurosurgery, Cedars-Sinai Medical Center, Los Angeles, CA 90048, USA, [email protected] Edith Cabrera-Muñoz Facultad de Quimica, Universidad Nacional Autonoma de Mexico, Coyoacan, Mexico 04510, D.F. Mexico Ignacio Camacho-Arroyo Facultad de Quimica, Universidad Nacional Autonoma de Mexico, Coyoacan, Mexico 04510, D.F. Mexico, [email protected]

xxi

xxii

Michael S. Canney Department of Neurosurgery and Advanced Surgical Technologies Research Team, Hopital de la Pitie-Salpetriere, Assistance Publique Hopitaux de Paris, Université Paris VI – Pierre & Marie Curie, 75013 Paris, France David Capper Department of Neuropathology, Institute of Pathology, Ruprecht-Karls-University, 69120 Heidelberg, Germany, [email protected] Alexandre C. Carpentier Department of Neurosurgery and Advanced Surgical Technologies Research Team, Hopital de la Pitie-Salpetriere, Assistance Publique Hopitaux de Paris, Université Paris VI – Pierre & Marie Curie, 75013 Paris, France, [email protected] Javier S. Castresana Unidad de Biologia de Tumores Cerebrales, Universidad de Navarra, 31008 Pamplona, Spain, [email protected] Steven D. Chang Department of Neurosurgery, Stanford University School of Medicine, Stanford, CA 94305-5484, USA, [email protected] Al Charest Department of Neurosurgery, Tufts Medical Center, Tufts University, Boston, MA 02111, USA, [email protected] Sergio Comincini Department of Genetics and Microbiology, University of Pavia, via Abbiategrasso 207 (OR via Ferrata 1), 27100 Pavia, Italy Alfredo Conti Department of Neuroscience, University of Messina, Messina, Italy, [email protected] Xing Fan Department of Neurosurgery, University of Michigan Medical School, Ann Arbor, MI, USA Mansoor Foroughi Division of Neurosurgery, B.C.’s Children Hospital, Vancouver, BC, Canada Kiyotaka Fujii Department of Neurosurgery, Kitasato University School of Medicine, 1-15-1, Kitasato, Minami, Sagamihara, Kanagawa 252-0374, Japan Takeo Fukushima Department of Neurosurgery, Fukuoka University Faculty of Medicine, Jonan-ku, Fukuoka 814-0180, Japan Timo Gaiser University of Massachusetts, Amherst, MA LRB 460 E, USA; Pathology Mannheim, University Medical Center Mannheim, Theodor-Kutzer Ufer 1-3, 68167 Mannheim, Germany, [email protected] Johannes M. Giesinger Department of Psychiatry and Psychotherapy, Innsbruck Medical University, Anichstr.35, 6020 Innsbruck, Austria Aliesha González-Arenas Facultad de Quimica, Universidad Nacional Autonoma de Mexico, Coyoacan, Mexico 04510, D.F. Mexico Loyola V. Gressot Department of Neurosurgery, Baylor College of Medicine, Houston, TX, USA Carlo Gulì Departments of Neuroscience and Clinical Oncology, University of Messina, Messina, Italy Hannu Haapasalo Department of Pathology, Tampere University Hospital, FI-33521 Tampere, Finland, [email protected]

Contributors

Contributors

xxiii

Joonas Haapasalo Department of Pathology, Tampere University Hospital, FI-33521 Tampere, Finland Valeria Hansberg-Pastor Facultad de Quimica, Universidad Nacional Autonoma de Mexico, Coyoacan, Mexico 04510, D.F. Mexico Devon Haydon Department of Neurosurgery, Washington University School of Medicine, St. Louis, MO 63110, USA Bernhard Holzner Department of Psychiatry and Psychotherapy, Innsbruck Medical University, Anichstr.35, 6020 Innsbruck, Austria, [email protected] Steven W. Hwang Department of Neurosurgery, Tufts Medical Center, Boston, MA, USA, [email protected] María del Mar Inda Ludwig Institute for Cancer Research, San Diego, CA, USA Tooru Inoue Department of Neurosurgery, Fukuoka University Faculty of Medicine, Jonan-ku, Fukuoka 814-0180, Japan Bowen Jiang Department of Neurosurgery, Stanford University School of Medicine, Stanford, CA 94305-5484, USA, [email protected] Ming-Chung Jiang Section of Haematology-Oncology, Department of Medicine, Taipei Medical University and Hospital, Taipei 110, Taiwan, [email protected] Donna Johnston Division of Neurology, Department of Pediatrics, Children’s Hospital of Eastern Ontario, Ottawa, ON, Canada K1H 8 David T.W. Jones Molecular Genetics of Pediatric Brain Tumors (B062), German Cancer Research Centre (DKFZ), Im Neuenheimer Feld 280, D-69120 Heidelberg, Germany Sergiusz Jó´zwiak Klinika Neurologii I Epileptologii, Instytut Pomnik “Centrum Zdrowia Dziecka”, 04-730 Warszawa, Poland Bozena Kaminska Laboratory of Transcription Regulation, The Nencki Institute of Experimental Biology, University of Warsaw, Warsaw, Poland Abdullah Kandil Department of Neurosurgery, Tufts Medical Center, Tufts University, Boston, MA 02111, USA Daniel L. Keene Division of Neurology, Department of Pediatrics, Children’s Hospital of Eastern Ontario, Ottawa, ON, Canada K1H 8, [email protected] Fuminari Komatsu Department of Neurosurgery, Fukuoka University Faculty of Medicine, Jonan-ku, Fukuoka 814-0180, Japan Katarzyna Kotulska Klinika Neurologii I Epileptologii, Instytut Pomnik “Centrum Zdrowia Dziecka”, 04-730 Warszawa, Poland, [email protected] Sally R. Lambert Department of Pathology, University of Cambridge, Addenbrooke’s Hospital Box 231, Cambridge, CB2 0QQ, UK, [email protected] Jeffrey Leonard Department of Neurosurgery, Washington University School of Medicine, St. Louis, MO 63110, USA, [email protected]

xxiv

Ching-Fong Liao Section of Haematology-Oncology, Department of Medicine, Taipei Medical University and Hospital, Taipei 110, Taiwan Adam N. Mamelak Department of Biomedical Engineering, University of Southern California, Los Angeles, CA 90089, USA Roger E. McLendon Section(s) of Surgical Pathology, Duke University Medical Center, Durham, NC 27710, USA, [email protected] Pitt Niehusmann Department of Neuropathology, University of Bonn, Medical Center, Sigmund-Freud-Str. 25, 53105 Bonn, Germany, [email protected] Yuko Nomura Department of Neurosurgery, Fukuoka University Faculty of Medicine, Jonan-ku, Fukuoka 814-0180, Japan Hidehiro Oka Department of Neurosurgery, Kitasato University School of Medicine, 1-15-1, Kitasato, Minami, Sagamihara, Kanagawa 252-0374, Japan Kirstie S. Opstad Division of Clinical Sciences, St. George’s, University of London, Cranmer Terrace, London SW17 0RE, UK, [email protected] Shinya Oshiro Department of Neurosurgery, Fukuoka University Faculty of Medicine, Jonan-ku, Fukuoka 814-0180, Japan Yihang Pan Department of Cellular Biology and Anatomy, Louisiana State University Medical Center, Shreveport, LA 71103, USA Seppo Parkkila Department of Pathology, Tampere University Hospital, FI-33521 Tampere, Finland Shibu Pillai Division of Neurosurgery, B.C.’s Children Hospital, Vancouver, BC, Canada Quincy Quick Department of Cellular Biology and Anatomy, Louisiana State University Medical Center, Shreveport, LA 71103, USA Juan A. Rey Research Unit, La Paz University Hospital, Madrid, Spain Hideo Saji Department of Patho-Functional Bioanalysis, Graduate School of Pharmaceutical Sciences, Kyoto University, Sakyo-ku, Kyoto 606-8501, Japan, [email protected] Gangadhara Reddy Sareddy Department of Biotechnology, School of Life Sciences, University of Hyderabad, Hyderabad 500 046, India Gabriele Schauer-Maurer Department of Psychiatry and Psychotherapy, Innsbruck Medical University, Anichstr.35, 6020 Innsbruck, Austria Giuseppe J. Sciarrone Departments of Neuroscience and Clinical Oncology, University of Messina, Messina, Italy Markus D. Siegelin Department of Pathology & Cell Biology, Columbia University College of Physicians & Surgeons, 630 W. 168th Street, New York, NY 10032, USA, [email protected] Omar Skalli Department of Cellular Biology and Anatomy, Louisiana State University Medical Center, Shreveport, LA 71103, USA, [email protected]

Contributors

Contributors

xxv

Paul Steinbok Division of Neurosurgery, B.C.’s Children Hospital, Vancouver, BC, Canada, [email protected] Olivia Tania Hernández-Hernández Facultad de Quimica, Universidad Nacional Autonoma de Mexico, Coyoacan, Mexico 04510, D.F. Mexico Chiara Tomasello Departments of Neuroscience and Clinical Oncology, University of Messina, Messina, Italy Tang-Yi Tsao Section of Haematology-Oncology, Department of Medicine, Taipei Medical University and Hospital, Taipei 110, Taiwan Hitoshi Tsugu Department of Neurosurgery, Fukuoka University Faculty of Medicine, Jonan-ku, Fukuoka 814-0180, Japan, [email protected] Jai-Nien Tung Section of Haematology-Oncology, Department of Medicine, Taipei Medical University and Hospital, Taipei 110, Taiwan Magdalena Ewa Tyburczy Translational Medicine Division, Brigham and Women’s Hospital, Boston, MA, [email protected] Masashi Ueda Radioisotopes Research Laboratory, Kyoto University Hospital, Sakyo-ku, Kyoto, 606-8507, Japan, [email protected] Satoshi Utsuki Department of Neurosurgery, Kitasato University School of Medicine, 1-15-1, Kitasato, Minami, Sagamihara, Kanagawa 252-0374, Japan, [email protected] Anand Veeravagu Department of Neurosurgery, Stanford University School of Medicine, Stanford, CA 94305-5484, USA Carol J. Wikstrand Department of Microbiology, Saba University School of Medicine, Saba, Dutch Caribbean Lisa M. Wintner Department of Psychiatry and Psychotherapy, Innsbruck Medical University, Anichstr.35, 6020 Innsbruck, Austria Fumio Yanai Department of Neurosurgery, Fukuoka University Faculty of Medicine, Jonan-ku, Fukuoka 814-0180, Japan Kun-Tu Yeh Section of Haematology-Oncology, Department of Medicine, Taipei Medical University and Hospital, Taipei 110, Taiwan

Part I

Astrocytomas: Diagnosis and Biomarkers

Chapter 1

Methylation in Malignant Astrocytomas María del Mar Inda, Juan A. Rey, Xing Fan, and Javier S. Castresana

Abstract The term epigenetics is used to describe the study of stable and heritable alterations in gene expression potential that arise during development and cell proliferation. Two epigenetic mechanisms have been thoroughly investigated in the past few years: DNA methylation and histone modifications. The failure of the maintenance of these heritable epigenetic marks can lead to inappropriate activation or inactivation of signaling pathways and result in disease, such as cancer. Promoters of tumor suppressor genes have been assessed for hypermethylation with a variety of techniques, both at specific loci or genome wide. Methylation of the MGMT gene, which favors treatment results with temozolomide, is a clear example of the influence of methylation in a specific gene in astrocytomas. At the clinical level, the emphasis is now on combining inhibitors of DNA methyl transferases and of histone deacetylases. Keywords DNA methylation · CpG islands · DNMT · O6 -Methylguanine · MGMT methylation

and are retained throughout mitosis. They do not involve mutations of the DNA itself and are referred to as epigenetic alterations. Originally, the term epigenetics, which literally means outside conventional genetics, was defined as the casual interactions between genes and their products, which bring the phenotype into being (Waddington, 1942). Nowadays, the term epigenetics is used to describe the study of stable and heritable alterations in gene expression potential that arise during development and cell proliferation (Jaenisch and Bird, 2003). Two epigenetic mechanisms have been thoroughly investigated in the past few years: DNA methylation and histone modifications. Epigenetic mechanisms are essential for development and differentiation, but they can also arise in adults, either by random change or under the influence of the environment, allowing the organism to respond to the environment by modulating gene expression. The failure of the maintenance of these heritable epigenetic marks can lead to inappropriate activation or inactivation of signaling pathways and result in disease, such as cancer.

Understanding the Word Epigenetics DNA Methylation Even though they are genetically identical, cells from a multicellular organism present differential gene expression and are structurally and functionally heterogeneous. These differences occur during development

J.S. Castresana () Unidad de Biologia de Tumores Cerebrales, Universidad de Navarra, 31008 Pamplona, Spain e-mail: [email protected]

Methylation might be responsible for the stable maintenance of a particular gene expression pattern through mitotic cell division. Ample support to this hypothesis has been provided and now, DNA methylation is recognized as an important mechanism for establishing a silent chromatin state by collaborating with proteins that modify nucleosomes. These epigenetic modifications can be copied after DNA synthesis, resulting in heritable changes in chromatin structure. Genes

M.A. Hayat (ed.), Tumors of the Central Nervous System, Volume 5, DOI 10.1007/978-94-007-2019-0_1, © Springer Science+Business Media B.V. 2012

3

4

can be transcribed from methylation-free promoters even though adjacent transcribed and non-transcribed regions are extensively methylated. In mammals, DNA methylation is predominantly found in cytosines of the dinucleotide sequence CpG and consists in the addition of a methyl group to the 5 -position of cytosines, altering the appearance of the major grove of DNA to which DNA binding proteins bind. CpG dinucleotides are not evenly distributed in the genome but rather are concentrated in short CpGrich DNA stretches called CpG islands, defined as regions of DNA greater than 200 bp, with a C + G content >50%, and an observed/expected presence of CpG >60%. In non-embryonic cells, methylation is found in approximately 80% of CpG dinucleotides. An exception for this global methylation of the genome are the CpG islands. The majority of CpG islands are associated with genes unmethylated in the germline and often located within promoter regions of genes. Approximately 60% of the human gene promoters contain CpG islands at the 5 end. How CpG islands in non-embryonic cells remain unmethylated is still unknown, but it is known that in cancer cells, methylation of CpG islands contributes to gene silencing of tumor suppressor genes. Methylation of certain CpG island promoters during development, resulting in long-term transcriptional silencing, has been observed.

Relevance of DNA Methylation in Normal Cells The relevance of DNA methylation in mammal development has been demonstrated by targeted mutagenesis of the different DNA methyltransferases (DNMT) genes in mice (Bestor, 2000). Genes involved in the establishment, maintenance or interpretation of genomic methylation pattern are essential for normal development. The first Dnmt to be discovered was Dnmt1 and it seems to act as a maintenance methyltransferase. Dnmt1 knock-out mice resulted in global demethylation and embryonic lethality (Li et al., 1992). In contrast, Dnmt3a and Dnmt3b are highly expressed in mouse embryo and are responsible for global de novo methylation after implantation. No obvious phenotype has been observed in mice after deletion of Dnmt2, but this gene is highly expressed

M. del Mar Inda et al.

during oogenesis and lacks biochemically detectable methyltransferase activity, but it seems to be responsible for the small amount of non-CpG methylation observed in the fly embryo (Lyko et al., 2000). CpG islands can normally be methylated in four cases: imprinted genes, X-chromosome inactivation in women, germline-specific genes, and tissue specific genes. X-chromosome inactivation in women is a well-characterized developmental phenomenon associated with DNA methylation in CpG islands assuring monoallelic gene expression (Jaenisch and Bird, 2003). The X-inactivation process and the genomic imprinting share some epigenetic mechanisms. The choice of the inactive X-chromosome and the initiation of the inactivation depends on Xist RNA, a noncoding transcript that originates at the X inactivation center (Xic) and coats the inactive X chromosome. Dnmt1 activity is needed for the maintenance of imprinting as well as for the X inactivation. Some studies suggest that Dnmt3L, which has no detectable methyltransferase activity, is required to establish maternal imprinting through the cooperation with de novo methyltransferase Dnmt3a. Some tissue-specific gene silencing through CpG island methylation has been reported in a variety of somatic tissues to silence these tissue-specific genes in tissues that should not express them, a well characterized example are the methionine adenosyltransferases 1A and 2A in rodents. A similar case are the germline-specific genes to restrict the expression of these genes to the male or the female germline and that later in the adult tissues will not be expressed, such as MAGE and LAGE gene families. Another interesting function for the normal DNA methylation is its role in repressing parasitic sequences. The methylation of the parasitic promoters inactivates them; over time, and thanks to the promutagenicity of the methylated cytosine, cytosines can be substituted by thymidine and destroy many transposons.

Epigenetics, Environment, Diet and Aging Epigenetic states are reversible and can be modified by environmental factors, diet and ageing and may contribute to the development of abnormal phenotypes. In addition, normal response to certain environmental

1

Methylation in Malignant Astrocytomas

stimuli may be mediated by epigenetic mechanisms. In mammals, hypo- and hypermethylation have been associated with ageing; however, the functional significance remains to be determined. It is known that age is a major risk factor for cancer development, probably through the methylation of CpG islands and silencing of tumor suppressor genes. Some examples of genes hypermethylated in ageing individuals are estrogen receptor, IGF2 and MYOD (Jaenisch and Bird, 2003). In addition, some dietary supplements, such as folate or vitamins, can affect the activity of enzymes supplying methyl groups for the methylation processes and influence the rate of disease manifestation. A methyl-deficient diet has been shown to induce liver cancer associated with both hypomethylation and the enhanced expression of oncogenes such as c-ras, c-myc or c-fos (Dizik et al., 1991).

DNA Methylation in Human Disease and Cancer The failure of the maintenance of the DNA methylation or disruption of its machinery can be the cause of disease or cancer. Mutations in the DNMT3B gene, the human homolog of Dnmt3b, cause the ICF syndrome (immunodeficiency, centromeric region instability, and facial abnormalities), a human heritable genetic disease with deficient methylation of the pericentromeric repetitive DNA and at CpG islands of the X chromosome. Another X-linked neurological disorder, the Rett’s syndrome, is due to a failure in DNA methylation-related system; more specifically, it is due to mutations in the methyl binding protein MeCP2, responsible for recruiting histone deacetylases (HDACS) and other chromatin factors to methylated DNA. These two diseases suggest that DNA methylation is not only needed to complete embryonic development, but it is also required for development after birth. DNA methylation plays a critical role in the development and differentiation of mammalian cells, and its deregulation has been involved in oncogenesis. The alteration of the DNA methylation pattern results in global dysregulation of gene expression profiles, leading to the development and progression of cancer. Since these alterations are hereditable, cells with epigenetic alterations conferring a growth advantage

5

are rapidly selected and result in uncontrolled tumor growth. Cancer can be considered to be a genetic disease at the same level as an epigenetic disease, and DNA methylation can be an excellent candidate to explain how certain environmental factors or ageing can increase the risk of cancer. In fact, DNA methylation plays an essential role in all three mechanisms by which cancer cells eliminate tumor suppressor gene function: point mutation, silencing by promoter hypermethylation and deletion by LOH due to genomic instability (Gronbaek et al., 2007). CpG sites have been considered to be mutation hotspots in the human germline and recently, it has become apparent that they are also hotspots for inactivating mutations of tumor suppressor genes such as p53 which is mutated in CpG sites in 25% of the cases. That CpG dinucleotides constitute hotspots for point mutations is due to the fact that methylated cytosines can be spontaneously deaminated to thymine and result in a C–T transition. If C–T transitions are not repaired and occur in the coding region of genes, they may activate an oncogene or suppress a tumor suppressor gene (Gronbaek et al., 2007). More than 30% of the point mutations in the germline related to disease occur at CpG dinucleotides. For example, in colorectal cancer, 44% of the mutations are C–T transitions. In addition, methylated cytosines also favor the formation of adducts on the neighboring G in the presence of some carcinogens, such as the benzo(a)pyrene present in tobacco smoke, resulting in a G–T transversion (Gronbaek et al., 2007). Commonly, cancer cells are characterized by global genomic hypomethylation and hypermethylation of CpG islands that are generally unmethylated in normal cells. DNA hypomethylation plays a critical role in tumorigenesis and may lead to the upregulation and activation of oncogenes, such as R-Ras and MAPSIN in gastric cancer, or MAGE in melanoma. The mechanisms by which DNA methylation can contribute to tumorigenesis can be summarized in three: reactivation of retrotransposons, increasing chromosomal instability, and loss of imprinting. DNA hypomethylation can allow the transcription and/or translocation of retrotransposons, increasing the genomic instability, or lead to the upregulation of oncogenic microRNAs. Loss of methylation has been observed in Alu repeats and in LINES in cancer cells, and some imprinted genes, such as H19 or IGF-2, present loss of methylation in pediatric tumors. Hypomethylation may allow

6

the formation of chromosomal breaks, translocations and/or allelic loss by illegitimate mitotic recombination, and the demethylation in pericentromeric regions of chromosomes plays a role in aneuploidy. In contrast to DNA hypomethylation which can lead to the activation of proto-oncogenes or increase genomic instability, hypermethylation of CpG islands that are unmethylated in normal cells leads to inactivation of tumor suppressor genes by silencing their expression, and several reports have shown a correlation between expression and loss of DNA methylation. How these genes are targeted for hypermethylation still remains unclear, and in some tumors, silencing by promoter hypermethylation occurs at a very high frequency. Many genes of key pathways in cancer are affected by promoter hypermethylation; however, methylation of the downstream gene sequences usually has no effect on gene expression (Jones, 1999). Examples of tumor suppressor genes silenced by hypermethylation were found in cancer and include: MGMT, Rb, p16Ink4a, BRCA1, p14ARF, APC, retinoic acid receptor-β2, RASFF1, etc (Gronbaek et al., 2007). Recently, experimental data has provided support to the idea that genes can be transcriptionally activated by removing DNA methylation (Baylin et al., 1998, 2001; Lorente et al., 2009), providing an attractive target for cancer therapeutics.

Methods to Detect Methylation Aberrant methylation is the most common alteration found in cancer cells, while silencing of tumor suppressor genes by CpG island promoter hypermethylation is the change of DNA methylation most studied in neoplasms. The detection of methylation in clinical samples (Table 1.1) may be useful in the early detection of cancer screening; therefore, it has become the focus of research in many clinical and translational laboratories. The reason for this is partially due to the early occurrence of alterations in the methylation pattern (hypo or hypermethylation) in carcinogenesis. Furthermore, since they are DNA markers, they are more stable than RNA or proteins, and studies can be performed in formalin-fixed and paraffin-embedded tissues (Fan et al., 2002). It has been demonstrated that DNA methylation can be detected in blood, sputa, ductal lavage fluids, urine, saliva, mammary aspiration

M. del Mar Inda et al.

fluid, stool, and biopsy specimens by using highly sensitive PCR-based methods after bisulfite modification. In addition to being a tumor-specific change, different tumor types have different DNA methylation profiles that are helpful in diagnosing difficult cases (Shames et al., 2007). In glioblastoma multiforme, the detection of methylation in the promoter of the MGMT gene (O6 -methylguanine-DNA methyltransferase) predicts a favorable outcome in patients treated with alkylating agents (Hegi et al., 2005). The initial studies of DNA methylation relied on the use of methylation-sensitive restriction enzymes that were able to distinguish between unmethylated and methylated recognition sites and Southern blot hybridization. This approach has many drawbacks: the limitation of the sites that can be analyzed, the problem of incomplete restriction cutting, the necessity of using high-molecular weight and elevated amounts of DNA to perform the Southern blot analysis, and the fact that the method is labor-intensive. In addition, only CpGs located within sequences recognized by methylation-sensitive enzymes can be analyzed. The majority of the methods used to detect DNA methylation are based on the chemical modification of DNA with sodium bisulfite followed by PCR with primers specific for methylated sequences. These methods, especially the ones that use primers designed specifically to amplify the methylated sequence, provide a very sensitive and specific analytical tool for detecting methylation at single loci. The treatment of DNA with sodium bisulfite deaminates cytosines to uracil, and because deamination of 5-methylcytosine is much slower, it is generally assumed that only unmethylated cytosines are transformed. There are three processes in the DNA modification by the bisulphite reaction: the reversible cytosine sulphonation, the irreversible hydrolytic deamination of the sulphonated cytosine, and the removal of the bisulfite adduct to give uracil by alkali treatment (Clark et al., 1994). It has been determined that the conversion rate under ideal conditions of unmethylated cytosines is about 99% (Taylor et al., 2007). Several groups have worked on optimizing the bisulfite treatment (Cottrell et al., 2004; Fan et al., 2002; Grunau et al., 2001). Once DNA is treated and modified with sodium bisulfite, different techniques can be used so as to make it possible for every laboratory and hospital to assess DNA methylation. Bisulfite sequencing provides a quantitative way to determine the methylation state of a genomic

1

Methylation in Malignant Astrocytomas

7

Table 1.1 Comparison among some of the different techniques to detect methylation Specimen Method treatment Application Sensitivity Quantitative

Advantages

Disadvantages Expensive and timeconsuming

No

Methylation status of individual CpG sites can be analyzed Easy to perform

Low

Yes

Reproducible

Genomewide

Low

Yes

Novel marker discovery

Bisulfite conversion

Specific locus

High

No

Cost-effective and needs small amounts of DNA

Q-MSP or Methylight

Bisulfite conversion

Specific locus

High

Yes

Easy and high throughput

Heavymethyl

Bisulfite conversion

Specific locus

High

Yes

MALDI-TOF MS

Bisulfite conversion

Genomewide/specific locus

Medium

Yes

Low false positives and high throughput Quantitative data on individual CpG sites can be obtained

Bisulfite sequencing

Bisulfite conversion

Specific locus

Low

Yes

Southern blot

Methylationsensitive enzyme

Genomewide

Low

RLGS

Methylationspecific restriction enzyme Immnunoprecipitation + array

Genomewide

MSP

ChIP-on-chip

region at a single-nucleotide resolution and is the gold standard of the methods based on the bisulfite DNA treatment. Unfortunately, this method is too expensive and time consuming to be used in a clinical setting. In this chapter we will discuss the methods most often used for detecting methylation at a single locus or multiple loci, as well as genome-wide (Table 1.1).

Limited sites available, needs high amounts of high quality DNA and is labor intensive Needs high quality DNA

No correlation with expression False positives and does not allow discrimination between unmethylated and partially methylated Does not allow discrimination between unmethylated and partially methylated Many oligonucleotides are used Expensive equipment required

The most widely used assay for sensitively detecting methylation is called methylation-specific PCR (MSP) (Herman et al., 1996). Before PCR amplification, genomic DNA is modified by sodium bisulfite treatment in order to convert all unmethylated cytosines to uracil which, after amplification, will be transformed into thymidine. Two sets of primers are designed for

8

PCR amplification: one set is designed to hybridize with the methylated sequence (M), while the other set of primers is designed to match with the nonmethylated sequence (U). After PCR, products are visualized in agarose gels by ethidium bromide staining. The sensitivity of this technique has been estimated to be one in 1000 (detection of methylated DNA in 1000-fold excess of unmethylated DNA). This method is simple, inexpensive, highly specific and sensitive, and does not require special equipment. However, MSP also presents some drawbacks. For example, it does not allow discrimination between partial levels of methylation and complete lack of methylation, leading to false positive results if the PCR conditions and sodium bisulfite modification are not optimized. The fact that MSP is gel-based makes this method unsuitable for a clinical setting because of the need to be high-throughput and homogeneous and not be quantitative. To address some of the drawbacks of the conventional MSP, a method based on Taqman technology, known as quantitative MSP (QMSP) or Methylight, has been developed (Eads et al., 2000). Each round of PCR leads to an increase in fluorescence proportional to the amount of target in the sample, and the signal is only observed when the probe has hybridized between the primers, thus eliminating the nonspecific amplification, such as primer dimer formation. This method is more suitable for routine clinical use because it does not need a secondary electrophoresis step eliminating cross-contamination problems, and it is quantitative. Other methods derived from the MSP are also very helpful in the methylation analysis of specific loci, overcoming some of the disadvantages of the MSP. This is the case of the quantitative MSP-Sybr green based, sensitive melting analysis after real-time MSP (SMART-MSP), methylation specific amplicon generation (MS-FLAG), multiplex Q-MSP (QM-MSP), methylation specific nested PCR (MSnested-PCR), etc. (Cottrell et al., 2004; Fackler et al., 2004). MALDI-TOF mass spectrometry (MALDI-TOF MS) is a novel strategy for high-throughput DNA which is based on a base-specific cleavage reaction combined with mass spectrometric analysis (Coolen et al., 2007). Briefly, DNA is converted with sodium bisulfite and amplified using a T7 promoter tagged primer. A single strand RNA molecule is then formed and it is base-specific cleaved by RNase A. Fragments originated by cleavage are analyzed by MALDI-TOF

M. del Mar Inda et al.

MS. The differences in methylation status are reflected in differences in fragment size, and quantification of abundance of each fragment represents the amount of DNA methylation in the sample. This technique is quantitatively accurate, relatively sensitive, with possible high-throughputs, but it is very complex and requires expensive equipment. Another method that uses primers which are nonspecific for the methylated or unmethylated sequence and is highly sensitive and specific is the denominated Heavy Methyl. The primers are designed to hybridize close to a region that contains a CpGrich sequence. During the reaction, there are also blockers that are designed to hybridize only with the unmethylated sequence; thus, if DNA is methylated, the blockers cannot hybridize, and amplification will occur. However, if the sequence is unmethylated, the blockers will bind to the sequence and block amplification. Amplification is detected with a probe that contains CpG sites, a quencher and a fluorophore label. Fluorophore is released from the quencher when the exonuclease activity of the polymerase cleaves the probe and light is emitted. The emitted light is proportional to the amount of PCR product, allowing an accurate quantification of methylation level. The use of the blocker increases the sensitivity and decreases the number of false positives. Other advantages of this method are that it is close tube-based, so it eliminates cross-contamination and allows highthroughputs, making it appropriate for a clinical setting. It is also possible to analyze DNA methylation globally. Global approaches to methylation analysis are high-throughput, but they are relatively expensive and labor-intensive, including microarray expression profiling, restriction landmark genomic scanning (RLGS) and immunoprecipitation of methylated DNA followed by array-based comparative genome hybridization (ChIP on chip) analysis. The RLGS has been defined as a method which provides quantitative genetic and epigenetic assessment of thousands of CPG islands in a single gel without prior knowledge of gene sequence (Costello et al., 2002). The basis of the RLGS method is the two-dimensional separation of radiolabeled DNA fragments obtained after successive digestions of genomic DNA with different restriction enzymes. DNA is first digested with an infrequently cutting enzyme sensitive to methylation, such as NotI or AscI. It is then radiolabeled, digested with a second

1

Methylation in Malignant Astrocytomas

enzyme and electrophoresed through a narrow tubeshaped gel. The DNA in the tube-gel is digested again with a more frequent cutting restriction enzyme and electrophoresed perpendicularly in a non-denaturing polyacrylamide gel. Approximately 2000 fragments originate, with a high probability of containing gene and/or promoter sequences, and making it possible to detect new hypermethylated sequences in the genome. The resulting gel autoradiography gives a determined RLGS profile that is highly reproducible, and comparisons between different samples can be performed. DNA quality is a critical parameter for generating high quality RLGS profiles. One of the disadvantages of this technique is that the loss of a fragment in the RLGS profile can be due to deletion or methylation. Recently, another method to detect global methylation has been developed. The ChIP-on-chip method is based on immunoprecipitation of methylated DNA with a monoclonal antibody to 5-methylcytosine after sonication or restriction digestion. The DNA is then labeled and hybridized to a DNA microarray with probes of the regions of interest. Methylated sequences are detected by comparing the fluorescent signal for each probe (Shames et al., 2007). As RLGS, ChIP-onchip is a genome-wide, high resolution and quantitative method, but both methods are very expensive and labor-intensive.

DNA Methylation in Astrocytic Tumors: Genes Frequently Methylated, Relevance for Diagnosis and Prognosis DNA methylation plays a critical role in mammalian central nervous system development and function, and global DNA methylation levels are dynamic during brain development varying among different brain regions (Nagarajan and Costello, 2009). As we previously discussed, genes involved in DNA methylation machinery such as DNA methyltransferases are important for normal development and several human neurodevelopmental disorders, such as Rett or human ICF syndromes, have been associated with alterations in genes involved in DNA methylation, such as MECP2 or DNMT3B. Attention has been focused on studying epigenetic alterations in glioblastoma multiforme because of the

9

possibility of being used as potential prognostic factors and as response factors for treatment. One of the most widely studied genes is O 6 -methylguanineDNA-methyltransferase (MGMT) due to the relationship between MGMT methylation status and response to alkylating agents, such as temozolomide. MGMT is involved in DNA repair, removing alkyl groups from O6 -methylguanine nucleotides and not allowing them to pair with thymine or to chemically react with other bases (Esteller et al., 1999) and protecting the cells from carcinogens. Approximately 56–68% of GBM present MGMT promoter hypermethylation (Nakamura et al., 2001), and is negatively correlated with expression. MGMT promoter hypermethylation is considered to be a biomarker of poor prognosis in GBM due to the fact that it has a critical role in DNA repair system and MGMT silencing has been correlated with increased TP53 mutations (Nakamura et al., 2001), in particular G to A transitions. On the other hand, it has been demonstrated that patients with MGMT hypermethylation respond better to treatment with alkylating agents, such as temozolomide (Hegi et al., 2005). Moreover, methylation of the pro-apoptotic gene TMS1/ASC coincides with MGMT methylation, suggesting that different types of glioma presenting differences in survival and response to treatment might have different epigenetic marks (Martinez et al., 2007). In GBM, promoter hypermethylation occurs in genes involved in different functions related to tumorigenesis and tumor progression, including apoptosis, DNA repair, drug resistance, invasion, cell cycle regulation, or angiogenesis. Methylation of cell cycle regulatory genes, such as p14, p15 and p16, has been observed in primary GBM, although at low frequencies, being more frequently methylated in low-grade astrocytomas and secondary GBM. PTEN tumor suppressor gene is mutated in 20–40% of GBM, and inactivation of PTEN expression by promoter methylation has also been found frequently in gliomas and glioma cell lines (Wiencke et al., 2007). Several groups have studied methylation patterns of different grades of glioma and found differences in the frequency of methylation of some tumor suppressor genes between different grades of glioma, primary GBM or secondary GBM or between gliomas and their relapses (Gonzalez-Gomez et al., 2003; Kunitz et al., 2007; Martinez et al., 2007; Nakamura et al., 2001; Uhlmann et al., 2003).

10

A large study of 139 tissues samples (Uhlmann et al., 2003), including 33 control tissues and 106 gliomas of different grades (33 pilocytic astrocytomas; 34 diffuse astrocytomas, 11 anaplastic astrocytomas, 7 secondary GBM and 3 primary GBM, among other oligoastrocytomas and oligodendrogliomas) assessed methylation at 15 loci including RB1, ARF, CDKN2B, APC or TIMP3. They defined an epigenotype of those tumors, showing that 7 of 15 loci analyzed presented tumor specific methylation. Notoriously, no hypermethylation was detected in the lower grade of gliomas or pilocytic astrocytomas. However, they were significantly hypomethylated at MYODI compared to control tissues. Grade II astrocytomas presented the most significant changes in methylation compared to normal controls. Some new genes were found to be methylated in gliomas such as CALCA, CDH1 or PTGS2, and the authors proposed that methylation fingerprinting for gliomas appears to be possible and could provide an additional diagnostic method for the future. In another study, 13 loci of genes involved in DNA repair, apoptosis, cell cycle regulation or tumor suppression were analyzed for DNA methylation in 32 paired tumor samples of GBM and relapses (Martinez et al., 2007). They found that the hypermethylation profile of GBM relapses was different in 62.5% of the patients, with CASP8 being the gene that was more significantly hypermethylated in GBM relapses, suggesting that a significant epigenetic silencing of this gene occurs during progression of primary to recurrent GBM. In contrast, other pro-apoptotic genes, such as CASP3 and CASP9, were unmethylated in both GBM and relapses. Characterization of diffuse astrocytomas (grade II) that underwent recurrence or progression revealed that tumors with p14ARF methylation at first biopsy were associated with shorter patient survival. Furthermore, they found that methylation of p14ARF and MGMT were frequent events in diffuse astrocytomas and were mutually exclusive. While p14ARF methylation was associated with a shorter survival, MGMT methylation was indicative of better clinical outcome after chemotherapy (Kleihues et al., 1994). No methylation of p21Waf/Cip1, p27Kip1 or p73 was observed in this study. It has been shown that primary or de novo GBM and secondary GBM present different genetic abnormalities. While primary GBM are characterized by LOH

M. del Mar Inda et al.

10q (70%), EGFR amplification (36–40%), p16INK4a deletion (31%) and PTEN mutations (25–40%), the main genetic abnormalities found in secondary GBM are LOH 10q (63%) and TP53 mutations (60%), which present a higher proportion in secondary GBM (Furnari et al., 2007; Ohgaki and Kleihues, 2007). Secondary and primary GBM also differ significantly in their pattern of DNA methylation and RNA and protein expression. In general, the promoter hypermethylation frequency is higher for secondary GBM than for primary GBM, maybe due to slower progression of the disease and the accumulation of alterations. RB1, p16INK4a, p14ARF, MGMT or TIMP-3 are methylated in a higher frequency in secondary GBM (Ohgaki and Kleihues, 2007), and for example, p14ARF is already methylated in a third of low-grade astrocytomas. Loss of MGMT expression by promoter hypermethylation was found in 75% of secondary GBM versus 36% of primary GBM. The correlation found between MGMT methylation and TP53 mutation might explain the higher frequency of TP53 mutations in secondary GBM (60–65% versus 28% in primary GBM) (Nakamura et al., 2001). The epithelial membrane 3 protein gene (EMP3), whose methylation is considered to be an unfavorable prognostic marker in neuroblastoma, is frequently methylated in low grade astrocytomas as well as secondary GBM (80%), but no in primary GBM (17%) suggesting that EMP3 methylation might be an early alteration in astrocytomas (Kunitz et al., 2007). In addition, some studies have suggested that methylation does not play a critical role in primary high grade gliomas and low frequency of RB1, p14ARF, p15INK4b, p16INK4a, p21Waf/Cip1, p27Kip1 or p73 were found in grade III and grade IV gliomas. Other genes frequently found to be methylated in gliomas are RASSF1A, BLU, Death receptor 4 (DR4), estrogen receptor (ER), or RARbeta.

Relevance of Methylation in the Clinic The importance of methylation in the clinic has been increasing due to the development of sensitive techniques for detecting methylation; methylation is a tumor-specific change that occurs early in tumorigenesis and the correlation between methylation and prognosis. In many neoplasms, therapeutic advances have

1

Methylation in Malignant Astrocytomas

been related to the availability of plasma biomarkers with prognostic and therapeutic significance. It is known that tumors released substantial amounts of DNA, containing the genetic and epigenetic alterations present in the primary tumor, into the systemic circulation, probably through apoptosis and cellular necrosis. It has been demonstrated that it is possible to detect promoter methylation in total plasma from glioma patients (Weaver et al., 2006): they analyzed promoter methylation in 4 gene locus in tumor tissue and total plasma from patients with different glioma grades, and demonstrated that 90% of the patients presented methylation of at least one locus in the primary brain tumor and that in 67% of these patients, methylation was also found in blood. As we mentioned previously, MGMT hypermethylation is associated with significant longer survival in patients with GBM and low-grade gliomas treated with alkylating agents such as temozolomide. Therefore, the detection of MGMT methylation can be used as a prognostic factor (Hegi et al., 2005). On the contrary, the detection of methylation of p14ARF promoter is associated with malignant progression and shorter survival; methylation of the pro-apoptotic gene Caspase-8 is frequently associated with relapsed GBM (Martinez et al., 2007).

Epigenetic Therapy The DNMT inhibitor Decitabine (5-aza-2 deoxycytidine) and the HDAC inhibitor Vorinostat (SAHA: suberoylanilide hydroxamic acid) are currently in use in multiple cancers, although only SAHA is in clinical trials in GBM. The advantage of epigenetic mutations is their reversibility compared to genetic mutations, but the principal problem of the epigenetic therapy is the target specificity. Even though the use of demethylating agents can reactivate silenced tumor suppressor genes, they can also activate oncogenes through hypomethylation. Another caveat of the use of 5-aza-2 -deoxycytidine is its toxicity. To overcome this problem, the combination of 5-aza-2 deoxycytidine and drugs that inhibit HDAC activity reduce the effective drug concentration and systemic toxicity while resulting in a more effective reactivation of tumor suppressor genes.

11

Future Directions Epigenetic studies of gliomas will provide further understanding of glioma biology and might identify new therapeutic targets. Consortiums, such as The Cancer Genome Atlas (TCGA; http://cancergenome. nih.gov), are helping to unravel the genetic and epigenetic alterations in GBM using high-throughput genomic and epigenomic approaches. The causes and consequences of DNA methylation in glioma are not entirely known and why some genes or pathways are preferentially targeted for methylation still remains unclear. Acknowledgements We are grateful to Laura Stokes for helping with editing the manuscript. This research was supported in part by grants from the Departmento de Salud del Gobierno de Navarra (9/07), Caja Navarra (08/13912), and Fundación Universitaria de Navarra, Pamplona; and Fondo de Investigación Sanitaria (PI081849), Madrid, Spain.

References Baylin SB, Herman JG, Graff JR, Vertino PM, Issa JP (1998) Alterations in DNA methylation: a fundamental aspect of neoplasia. Adv Cancer Res 72:141–196 Baylin SB, Esteller M, Rountree MR, Bachman KE, Schuebel K, Herman JG (2001) Aberrant patterns of DNA methylation, chromatin formation and gene expression in cancer. Hum Mol Genet 10:687–692 Bestor TH (2000) The DNA methyltransferases of mammals. Hum Mol Genet 9:2395–2402 Clark SJ, Harrison J, Paul CL, Frommer M (1994) High sensitivity mapping of methylated cytosines. Nucleic Acids Res 22:2990–2997 Coolen MW, Statham AL, Gardiner-Garden M, Clark SJ (2007) Genomic profiling of cpg methylation and allelic specificity using quantitative high-throughput mass spectrometry: Critical evaluation and improvements. Nucleic Acids Res 35:e119 Costello JF, Plass C, Cavenee WK (2002) Restriction landmark genome scanning. Methods Mol Biol 200:53–70 Cottrell SE, Distler J, Goodman NS, Mooney SH, Kluth A, Olek A, Schwope I, Tetzner R, Ziebarth H, Berlin K (2004) A real-time pcr assay for DNA-methylation using methylationspecific blockers. Nucleic Acids Res 32:e10 Dizik M, Christman JK, Wainfan E (1991) Alterations in expression and methylation of specific genes in livers of rats fed a cancer promoting methyl-deficient diet. Carcinogenesis 12:1307–1312

12 Eads CA, Danenberg KD, Kawakami K, Saltz LB, Blake C, Shibata D, Danenberg PV, Laird PW (2000) Methylight: a high-throughput assay to measure DNA methylation. Nucleic Acids Res 28:E32 Esteller M, Hamilton SR, Burger PC, Baylin SB, Herman JG (1999) Inactivation of the DNA repair gene O6methylguanine-DNA methyltransferase by promoter hypermethylation is a common event in primary human neoplasia. Cancer Res 59:793–797 Fackler MJ, McVeigh M, Mehrotra J, Blum MA, Lange J, Lapides A, Garrett E, Argani P, Sukumar S (2004) Quantitative multiplex methylation-specific pcr assay for the detection of promoter hypermethylation in multiple genes in breast cancer. Cancer Res 64:4442–4452 Fan X, Inda MM, Tunon T, Castresana JS (2002) Improvement of the methylation specific pcr technical conditions for the detection of p16 promoter hypermethylation in small amounts of tumor DNA. Oncol Rep 9:181–183 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 Gonzalez-Gomez P, Bello MJ, Arjona D, Lomas J, Alonso ME, De Campos JM, Vaquero J, Isla A, Gutierrez M, Rey JA (2003) Promoter hypermethylation of multiple genes in astrocytic gliomas. Int J Oncol 22:601–608 Gronbaek K, Hother C, Jones PA (2007) Epigenetic changes in cancer. APMIS 115:1039–1059 Grunau C, Clark SJ, Rosenthal A (2001) Bisulfite genomic sequencing: systematic investigation of critical experimental parameters. Nucleic Acids Res 29:E65–65 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 Herman JG, Graff JR, Myohanen S, Nelkin BD, Baylin SB (1996) Methylation-specific pcr: a novel pcr assay for methylation status of cpg islands. Proc Natl Acad Sci USA 93:9821–9826 Jaenisch R, Bird A (2003) Epigenetic regulation of gene expression: how the genome integrates intrinsic and environmental signals. Nat Genet 33 Suppl:245–254 Jones PA (1999) The DNA methylation paradox. Trends Genet 15:34–37 Kleihues P, Lubbe J, Watanabe K, von Ammon K, Ohgaki H (1994) Genetic alterations associated with glioma progression. Verh Dtsch Ges Pathol 78:43–47

M. del Mar Inda et al. Kunitz A, Wolter M, van den Boom J, Felsberg J, Tews B, Hahn M, Benner A, Sabel M, Lichter P, Reifenberger G, von Deimling A, Hartmann C (2007) DNA hypermethylation and aberrant expression of the emp3 gene at 19q13.3 in human gliomas. Brain Pathol 17:363–370 Li E, Bestor TH, Jaenisch R (1992) Targeted mutation of the DNA methyltransferase gene results in embryonic lethality. Cell 69:915–926 Lorente A, Mueller W, Urdangarin E, Lazcoz P, Lass U, von Deimling A, Castresana JS (2009) Rassf1a, blu, nore1a, pten and mgmt expression and promoter methylation in gliomas and glioma cell lines and evidence of deregulated expression of de novo dnmts. Brain Pathol 19:279–292 Lyko F, Ramsahoye BH, Jaenisch R (2000) DNA methylation in drosophila melanogaster. Nature 408:538–540 Martinez R, Schackert G, Esteller M (2007) Hypermethylation of the proapoptotic gene tms1/asc: prognostic importance in glioblastoma multiforme. J Neurooncol 82:133–139 Nagarajan RP, Costello JF (2009) Epigenetic mechanisms in glioblastoma multiforme. Semin Cancer Biol 19:188–197 Nakamura M, Watanabe T, Yonekawa Y, Kleihues P, Ohgaki H (2001) Promoter methylation of the DNA repair gene mgmt in astrocytomas is frequently associated with g:C –> a:T mutations of the tp53 tumor suppressor gene. Carcinogenesis 22:1715–1719 Ohgaki H, Kleihues P (2007) Genetic pathways to primary and secondary glioblastoma. Am J Pathol 170:1445–1453 Shames DS, Minna JD, Gazdar AF (2007) Methods for detecting DNA methylation in tumors: From bench to bedside. Cancer Lett 251:187–198 Taylor KH, Kramer RS, Davis JW, Guo J, Duff DJ, Xu D, Caldwell CW, Shi H (2007) Ultradeep bisulfite sequencing analysis of DNA methylation patterns in multiple gene promoters by 454 sequencing. Cancer Res 67:8511–8518 Uhlmann K, Rohde K, Zeller C, Szymas J, Vogel S, Marczinek K, Thiel G, Nurnberg P, Laird PW (2003) Distinct methylation profiles of glioma subtypes. Int J Cancer 106:52–59 Waddington CH (1942) The epigenotype. Endeavour 1:18–20 Weaver KD, Grossman SA, Herman JG (2006) Methylated tumor-specific DNA as a plasma biomarker in patients with glioma. Cancer Invest 24:35–40 Wiencke JK, Zheng S, Jelluma N, Tihan T, Vandenberg S, Tamguney T, Baumber R, Parsons R, Lamborn KR, Berger MS, Wrensch MR, Haas-Kogan DA, Stokoe D (2007) Methylation of the pten promoter defines low-grade gliomas and secondary glioblastoma. Neuro Oncol 9:271–279

Chapter 2

Deciphering the Function of Doppel Protein in Astrocytomas Alberto Azzalin and Sergio Comincini

Abstract Doppel is a newly recognized prion-like protein encoded by a novel gene locus, PRND, located on the same chromosomal region of the prion coding gene (PRNP); doppel is considered a prion paralogue and together they constitute the prion-gene family, originated through an ancestral gene duplication event. Prion and doppel have different expression patterns, suggesting that the gene products exhibit different biological functions. Actually, doppel is not involved in the etiology of the transmissible spongiform encephalopathies (TSEs) or “prion diseases” and it is highly expressed only within the testicular tissue, where its important physiological role in the process of spermatogenesis and fertilization in human and mouse has been described. Importantly, doppel is toxic when ectopically overexpressed in the central nervous system (CNS), with concomitant prion protein absence: this evidence suggests deeper investigations within particular pathological contexts, such as in Parkinson and Alzheimer’s diseases as well as in cancer. In the latter scenario several studies are showing that doppel represents a novel and attractive diagnostic molecular marker, and that it might provide insights into the regulatory pathways of tumor-cell transformation. In particular, doppel protein has been recently associated with the ability of tumor cells to migrate, as one of the most important hallmarks of cancer. In conclusion, since its discovery, the intriguing spectrum of biological and pathological functions of this new

A. Azzalin () Institute of Molecular Genetics, IGM-CNR Pavia via Abbiategrasso 207 (OR via Ferrata 1), 27100 Pavia, Italy e-mail: [email protected]

prion-like protein is constantly considered for novel investigations. Keywords Doppel · Transmissible spongiform encephalopathies · Bovine spongiform encephalopathy · CNS · GPI · Prion–doppel

Introduction Historical Background of Doppel Discovery The doppel gene was identified rather incidentally during the sequence analysis of a murine cosmid clone containing the prion gene, isolated from the I/LnJ inbred strain of mice (Moore et al., 1999): the rationale of this large scale genomic sequence project was to find regulatory elements and additional genes that might influence or contribute to the abnormal expression of the pathological isoform of the prion protein, the etiological agent of TSEs or “prion diseases” (Prusiner, 1998). In particular, it was noted that, directly adjacent to the prion gene, there was an attractive open reading frame (ORF) sequence which coded for a protein with striking similarities with the prion counterpart, as shown in Fig. 2.1. However, despite the protein structure similarity, the novel identified coding sequence shared low nucleotide similarity with the prion gene. This novel gene was therefore termed “prion–doppel” or “doppel” (for downstream of prion protein-complex or doppelgänger, i.e., a ghost-like companion). Further studies reported the evolutionary conservation of doppel gene sequences in different

M.A. Hayat (ed.), Tumors of the Central Nervous System, Volume 5, DOI 10.1007/978-94-007-2019-0_2, © Springer Science+Business Media B.V. 2012

13

14

Fig. 2.1 The prion-gene family and its gene products. The prion-gene family and its gene products. The figure shows the human Prn genomic locus and, in detail, the PRNP and PRND gene structures (a), both composed of two exons (red and blue boxes, respectively) containing single ORFs; numbers above the structure indicate sizes in base pairs. In the section b) of the figure, a schematic diagram of prion and doppel protein secondary structures are reported, showing major structural and biochemical features, such as disulfide bonds and threonine (Thr) and

A. Azzalin and S. Comincini

asparagines (Asn) glycosylation sites; aminoacid residues numbering is indicated. The two proteins show many features in common as revealed by NMR analysis (c): in both cases the proteins are composed of three α-elixes and two β-sheets, in blue and red respectively. Between square brackets the accession numbers to NCBI databases are reported (http://www.ncbi.nlm.nih.gov/). GPI, glycosyl-phosphatidyl-inositol; H, helix; OR, octarepeats; S, sulphate; SP-N and -C, amino- and carboxy-signal peptide; TM, trans-membrane domain

2

Deciphering the Function of Doppel Protein in Astrocytomas

mammalian species, such as humans and ruminants (Comincini et al., 2001). It was then inferred that the doppel gene might have derived from a proto-prion gene duplication event, thus originating the Prn-gene family, composed of prion (PRNP) and doppel (PRND) genes, sharing a common architecture (Mastrangelo and Westaway, 2001). At present, two other genes belonging to this small gene family, PRNT and SPRN, respectively, have been described, although their biological functions are still unclear. Further studies, that mirrored the studies by Moore et al. (1999) in murine species, confirmed that doppel was nearly undetectable in the CNS, markedly different from the prion protein. Doppel expression pattern is more temporary and spatially restricted compared to prion, being particularly expressed in mammal testicular tissues (Peoc’h et al., 2002). Of note, once discovered, doppel provided important insights into the puzzling phenotypes observed in Prnp knock-out mice: in particular, when doppel is overexpressed in CNS derived cells or tissues (where it is usually not expressed), together with the absence of the prion gene expression, doppel causes severe ataxic phenotype, roughly similar to that observed in prion diseases. Strikingly, however, the reintroduction of the Prnp transgene restores the normal phenotype (Weissmann and Aguzzi, 1999). Therefore different studies in mice highlighted that doppel might represent a toxic product, particularly at the CNS level and that its expression is strictly regulated in these tissues, with the exception of a limited window of expression within a week postnatal interval. Furthermore, different molecular models have been proposed in order to decipher a possible mechanism of antagonistic interaction between the prion and doppel gene products in CNS (Sakaguchi, 2008). Doppel discovery, followed by its biochemical and structural characterization (Lührs et al., 2003), provided new hypothetical avenues on the molecular pathogenesis of the prion diseases, in particular in the aetiology of the ataxic correlated phenotype. Unfortunately, extensive DNA sequencing surveys on patients affected by Creutzfeldt–Jakob (CJ) disease, Gerstmann–Sträussler–Scheinker (GSS) disease, familiar fatal insomnia (FFI) (Mead et al., 2000), as well as on scrapie and bovine spongiform encephalopathy (BSE)-affected animals (Comincini et al., 2001) failed to reveal doppel-genetic determinants associated with these pathologies. Moreover,

15

the expression of doppel protein in the CNS does not modulate TSEs transmission in mouse models, excluding a direct involvement of this protein in the disease progression (Rossi et al., 2001). The testisrestricted expression pattern of doppel, mainly identified in sperms and in Sertoli cells, confirmed in different mammalian species, prompted to investigate if this gene might have a possible function in the male reproductive tract: different studies reported that doppel gene knock-out mice gave rise to infertility in mice, because their mutated spermatozoa were unable to perform the acrosome reaction (Paisley et al., 2004). Analogously in humans, doppel DNA polymorphisms were associated with fertility defects (Peoc’h et al., 2003). Moreover, it was reported that an interaction between doppel and P34H (also knows as dicarbonyl/L-xylulose reductase), a glycosylphosphatidyl-inositol (GPI)-anchor protein with testisspecific expression, localized particularly on the acrosomal cap of spermatozoa (Azzalin et al., 2006). Interestingly, P34H protein, similarly to doppel, has been demonstrated to be involved in one of the prerequisites of human fertilization, i.e., the binding of spermatozoa to the zona pellucida. In recent years, different contributions enlarged the spectrum of interest regarding the doppel gene functions in pathological context. Because of the structural homology with the prion protein, the analysis of doppel in this field moved directly to the CNS diseases. Therefore, the examination of doppel expression and its possible role in human neurodegenerative pathologies different from prion diseases (Alzheimer’s, Pick’s, Parkinson’s and diffuse Lewy body diseases) and brain tumors was undertaken, as discussed here.

Doppel Gene Expression Analysis in Astrocytomas: A Novel Potential Tumor Marker Doppel possesses similar exon–intron architecture to that of the prion gene, as shown in Fig. 2.1a; additionally, the two genes share a chromosomal synteny in different mammalian species (Comincini et al., 2006a). As previously stated, these evidences reinforced the concept that an ancestral gene duplication event originated the prion-family gene chromosomal locus, designated Prn. Similarly to the prion gene,

16

the doppel coding region is contained within a single exon. However, the two genes exhibited marked different expression patterns, with prion gene being mostly expressed within the CNS tissues, while doppel is uniquely expressed in the adult testis tissue but, notably, at very low levels in nervous system. Therefore, while prion gene expression is relevantly related to the prion disease onset and epidemiology, doppel related studies were targeted to male gametogenesis and fertilization processes. In addition to neurodegenerative and reproductive investigations, a novel doppel-related branch of research was proposed (Comincini et al., 2004). The rationale was to investigate doppel expression perturbation within CNS-related diseases such as brain cancers, where it is expected to be physiologically not expressed. These studies led to the identification and characterization of a novel expression marker within the human glial tumors and to the association with the malignant grading progression. Glial tumors are histo-pathologically divided into four grades of malignancy, according to the WHO classification (Louis et al., 2007): pilocytic (WHO I), low-grade (WHO II), anaplastic astrocytomas (WHO III) and glioblastoma multiforme (WHO IV). In contrast to the longstanding and well defined histo-pathological criteria, the underlying molecular and genetic basis emerged only recently. In particular, several genes and pathways have been identified as being associated with tumorigenesis and with the anaplastic progression (Comincini, 2001). Adopting a high sensitive Real-time PCR based approach, doppel gene expression was investigated in large cohorts of patients with glial tumors. As a result, doppel expression was directly related to the malignancy of the tumor: highest in glioblastoma multiforme, lower in anaplastic astrocytomas, and even lower in low grade astrocytoma specimens, as graphically reported in Fig. 2.2 (Comincini et al., 2004). Extensive differences in doppel gene expression were also found within each grade of malignancy, suggesting that the quantification of doppel expression might be useful to distinguish astrocytoma subtypes. In addition, it was demonstrated that its expression is helpful in disease stratification and in the identification of patient subsets with specific molecular signatures (Comincini et al., 2007). Further molecular doppel gene expression investigations were then performed within human glial tumor specimens and in derived cell lines. These studies revealed that the

A. Azzalin and S. Comincini

upregulated doppel transcripts underwent a significant nuclear retention process within the glial tumor cells, as shown in the image of Fig. 2.2 (Comincini et al., 2006b): this alternative post-transcriptional pathway might directly regulate the excess of potentially deleterious transcripts. In addition, this aberrant nuclear retention may have a functional meaning, as other genes with this trademark were described. In fact, nuclear mRNA retention is increasingly recognized as an important mechanism to regulate the activity of transcription related proteins and to modulate cell growth and death. For this reason, the export of nuclear mRNA is constantly challenged by the opposing force of mRNA retention from one side, and its decay from the other. This balance ensures that only perfect transcripts persist and that nonfunctional and potentially deleterious transcripts are differently regulated in their biogenesis. In detail, doppel human mRNA underwent alternative maturation processes within glial tumor cells, thus originating an alternative shorter transcript of 1.9 kb, significantly different from that originated in testis tissue (Comincini et al., 2006b). Altogether, these data might suggest that glial tumor cells overproduce doppel transcripts in primis, and that these transcripts are then subjected to an initial quality/quantity control through a nuclear retention process. In a similar manner, the expression pattern and the distribution of doppel were investigated in tumors with a non-glial origin and, in a former study, high levels of doppel transcripts were detected in gastric adenocarcinoma and in anaplastic meningioma specimens (Comincini et al., 2004). Additionally, Travaglino et al. (2005) investigated doppel expression in bone marrowderived cells of patients with acute myeloid leukaemia (AML) and with myelo-dysplastic syndrome (MDS). As a result, while doppel transcripts were barely detectable in normal bone marrow samples, AML and MDS cases exhibited a marked increase in doppel expression, particularly localized in blast-like cells with a significant phenomenon of nuclear retention of its transcripts. As a consequence of the ectopically expression of doppel gene in different tumor biopsies, one may hypothesize that the corresponding gene product belongs to the group of the cancer-testis antigens (CTA), that recently captured considerable interest (Zendman et al., 2003). In fact, along with its physiological expression restricted to germ cells of the testis that exclusively reappears in neoplastictransformed cells, the doppel gene reflects other typical

2

Deciphering the Function of Doppel Protein in Astrocytomas

17

Fig. 2.2 Doppel molecular and cellular signatures in human astrocytoma. Doppel molecular and cellular signatures in human astrocytoma. As described, doppel gene and protein expression levels increase following the glioma malignancy grading; a significant nuclear retention of the doppel transcripts has been previously reported (inset a), as well as an increase of

the complexity in the glycan moiety composition (Comincini et al., 2006b). Additionally, as illustrated (inset b), the cellular localization of the doppel protein shifts from plasma membrane to the cytosol, particularly within lysosome organelles (Sbalchiero et al., 2008)

features that characterise CTA, such as its belonging to a gene family and the single-exon ORF gene structure. To date, the molecular mechanisms responsible for doppel over-expression in transformed cells remain summarily defined. Doppel altered expression could merely be an epiphenomenon because of the widespread change in gene methylation patterns observed within different tumor types (Travaglino et al., 2005). To better delve into doppel gene regulatory mechanisms, investigations in different species were performed; in general, these experimental and bioinformatics computational analysis supported the concept that doppel gene expression is tightly regulated through an interplay of positive and negative cisacting factors that specifically recognize activating and

inhibiting elements in the promoter and in its surrounding sequence (Del Vecchio et al., 2005). Furthermore, doppel expression is affected by the methylation status of its promoter sequence, differently from the prion paralogue gene (Comincini, unpublished data). It is therefore conceivable that doppel is positively regulated in testis and negatively regulated in CNS. Interestingly, as a conclusive remark, the high expression profiles of doppel gene, physiologically in the first stage of the brain development and ectopically in glial tumor specimens, may indicate the neoplastic reacquisition of tumor cells of a primitive expression behaviour (Comincini et al., 2004).

18

Doppel Protein in Astrocytomas Biochemical Features and Cellular Localization Doppel is a membrane protein that was initially identified as the first prion-like protein, due to its significant sequence homology with the cellular prion protein, as previously mentioned (Moore et al., 1999). In particular, the sequence analysis revealed that doppel is a truncated version of the prion protein, lacking the prion-typical octarepeat motifs, as graphically shown in the comparison between the secondary structures of Fig. 2.1b. Besides the protein sequence similarity, the two proteins share a number of structural and biochemical similarities: they covalently link a GPI anchor molecule for location to the outer leaflet of the cell membrane and, at the C-terminus, each protein has a super-imposable three-dimensional structure, characterized by three α-helices and by two short antiparallel β-sheets (Mo et al., 2001). Furthermore, the proteins were reported to bind copper ions with different affinity, but the physiological relevance of this biochemical property has not been clearly defined yet. Because of the already stated extensive similarities, doppel was primarily investigated as an alternative element to explore the physiological and pathological functions of the prion protein. As a consequence of these investigations, it was suggested that, instead of sharing similar activities, the two evolutionary-derived proteins appear to show different and possibly opposite activities, according to the “paralogue compensation process”. In particular, unlike the prion protein, doppel is not required for prion replication, and it is most likely unable to originate a pathogenic proteaseresistant isoform (Mo et al., 2001). Other functional differences between the two proteins were related to the adverse effect on neuronal viability and the proapoptotic behaviour of doppel (Qin et al., 2006), compared to the importance of prion protein expression in neuronal protection to oxidative stress, and in cell growth and maturation (Aguzzi and Polymenidou, 2004). In the glial tumor context, doppel protein showed peculiar features, graphically summarized in Fig. 2.2, compared to the physiological conditions, as derived from comparative testicular tissue examinations. Biochemical studies primarily highlighted that the

A. Azzalin and S. Comincini

levels of astrocytoma doppel protein seem to increase with the tumor grade, showing a similar trend to that observed for the doppel transcripts (Chiarelli, unpublished data). In addition, the examination of astrocytoma surgical samples showed that doppel was present in the soluble fraction, whereas the corresponding protein in the testis tissue was detected in the microsomial fraction, as expected for a GPI-anchored membrane molecule. These differences in doppel localization between glial tumors and normal testis might reflect a complex recycling machinery of the astrocytoma membrane molecules, where GPI-linked proteins might have remarkably different traffic within the cells. It is known that GPI-proteins contribute to the overall organization of other membrane-bound proteins and they also play a critical role in a variety of receptor mediating signal transduction pathways. Another significant biochemical feature of the doppel protein in glioma cells was the presence of different glycoforms as revealed by Comincini and collaborators (unpublished data) in astrocytoma cell lines (2006) and further confirmed in tumor specimens. In particular, an abnormal post-translational maturation process, specifically an hyper-glycosylation, with a reduced content in sialic acids, resulted in a significant increase of the doppel protein molecular mass. It is known that protein-linked oligosaccharide moieties are crucial to serve diverse functions: they stabilize the proteins against denaturation and proteolysis, enhance solubility, modulate immune responses, facilitate orientation of proteins, confer structural rigidity to proteins, and regulate their turnover. Furthermore, it has long been predicted that the carbohydrate moieties of cell surface glycoproteins play important roles in the physical function and in the structural stability of the proteins. In the case of doppel, it has been demonstrated that the protein is subjected to post-translational modifications and that the doppel membrane localization in glioma cells seems not efficient and/or not well tolerated (Sbalchiero et al., 2008). Therefore, it is possible to suppose that altered glycosylation processes, common in cancers, can cause the altered localization of doppel, resulting in a shift from the membrane to the cytosol; therefore, this highly glycosylated protein was probably targeted directly to lysosomes where it accumulates for degradation. This parsimonious explanation of the catabolism of a potentially cytotoxic protein in the lysosomal compartments would not take into account that doppel-linked oligosaccharide moieties might be

2

Deciphering the Function of Doppel Protein in Astrocytomas

additionally important in conferring a tumor phenotype. For example, the absence of doppel expression at membranes or its rapid turnover within glioma cells might also contribute to a complex re-assortment of plasma membrane proteins, likely plausible in such cancers.

Functional Pathways and Interaction Analysis Recently the first interactome characterization map of the prion-protein family members was reported (Watts et al., 2009). This includes a cell-based quantitative analysis of neuroblastoma-derived interactors of prion, and of the two known paralogues, i.e., doppel and shadoo proteins. The results highlighted, in particular, that all prion-family members shared similar molecular and subcellular microenvironments (i.e., endoplasmic reticulum, Golgi and raft-like membranes), due to cointeraction with similar protein candidates. In detail, a reciprocal prion–doppel interaction has been reproposed, confirming the sharing of prion and doppel in common plasma membrane microdomains and in internalization pathways in neuroblastoma cells, as suggested by Massimino et al. (2004). The functional interplay between doppel and prion proteins was originally raised in prion-deficient mice, overexpressing doppel, after the rescue of the doppel-dependent ataxic phenotype, by the reintroduction of the prion transgene (Moore et al., 1999). In fact, the demonstration of the prion and doppel proteins interaction was of primary importance, considering the relevance of the cell membrane microdomains in the pathogenesis of prion and other neurodegenerative diseases. However, in literature, contrasting results on the putative interaction between the two proteins have been reported: whereas in neuronal cells the results supported an interaction, in testis this was not revealed, probably because the association of these proteins to different sub-cellular localizations could account for their different functions in each tissue. Similarly, in astrocytoma-derived cells a direct and detectable interaction between prion and doppel proteins was not documented; furthermore, doppel protein, even abundantly and ectopically expressed at the cytoplasmic level in such tumor cells, failed to physically interact with some prion-interacting proteins, such as the glial

19

fibrillary acidic (GFAP) and the growth factor receptorbound 2 (Grb2) proteins (Azzalin et al., 2005). Because of the soluble form of doppel in glioma, one could hypothesize that if the prion–doppel complex gets disrupted, and/or malfunctions, the latter does not interact with prion thus reversing the beneficial effect of a neuro-protective signal of the prion protein towards doppel; as a consequence, doppel is sorted at the lysosomes, as described. However, the doppel–prion functional interplays, whether the derived proteins might physically interact and coparticipate or antagonize in biological processes, need to be finally delineated. Many results confirmed, at least in the CNS derived cells, the potential toxicity of the doppel gene product and a plausible cellular adaptive response in activating cell-death circuits. Within the brain, in particular in neuroblastoma N2a and in primary rat astrocytes, Qin et al. (2006) demonstrated that doppel caused apoptosis through a caspase-10 mediated mechanism, in a mitochondrion-independent manner, probably through a direct interaction with death cellular receptors. However, doppel itself showed a reduced apoptotic effect in neuronal cells expressing the prion protein, but a significant increase in cell death induction rate was reported in the same cells devoid of prion protein expression. In this scenario, other studies pointed to the direct involvement of the proapoptotic Bax protein in promoting the doppel induced apoptosis in transgenic murine Purkinje cells, deficient in prion protein expression. As a counterpart, further data supported the rescuing of the neuronal survival through the involvement of the anti-apoptotic Bcl2-dependent pathways, suggesting that the Bcl2-like property of the prion protein might impair doppel-induced neurotoxic effects (Heitz et al., 2008). To additionally complicate the functional networks involving doppel, more recent studies indicated that the expression of this protein in a neuronal context, in the absence of the prion protein counterpart, might also trigger autophagic cell death processes. It was therefore speculated that, as observed in amyloid neurodegenerative diseases, the doppel-induced upregulation of autophagic markers, resulted in extensive accumulation of autophagosomes, might likely reflect a progressive dysfunction of neuronal cells that finally lead to cell death (Heitz et al., 2010). Although doppel possesses an intrinsic toxic effect if not counteracted by another effector (for example the prion protein), astrocytoma cells acquired the

20

ability to protect themselves from the doppel potential toxicity, delocalizing the protein from the plasma membrane towards the cytosol and to the lysosomes. For this reason, the peculiar localization of doppel in tumor astrocytic cells, focused the research of potential interactants within the cytosol as well as into specific organelles. In this regard, RACK1 (receptor for activated C-kinase), a well-characterized cytosolic adaptor molecule, was recently documented as a doppel interacting protein (Azzalin et al., 2006). This partner was identified among a group of doppel interactor candidates isolated from a glioblastoma-derived expression library, by means of a yeast two-hybrid assay. This interaction has also been subsequently verified by coimmunoprecipitation experiments and, interestingly, bioinformatic analysis underlined that doppel protein sequence shows conserved amino acids with the sequence of the PH (Pleckstrin protein) domain, an important domain contained in many cell signaling proteins that mediates the interaction with RACK1. In addition, RACK1 is homologous to the hetero-trimeric G protein ß-subunit, that regulates cell signaling via Src- and PKC (protein kinase C)-dependent pathways. In particular, it has been described to modulate the integrin-mediated cellular adhesion and migration (Kiely et al., 2008). RACK1 has a peculiar secondary structure composed of seven WD (tryptophan-aspartic acid) domains and, due to this conformation, it can manage contemporaneously many protein partners and coordinate different cell signaling inputs. In particular, the doppel-RACK1 interaction region was mapped between the C-terminal portion of doppel and the WD(1-4) domains of RACK1, respectively. It was, therefore, suggested that doppel might participate with RACK1 in a common molecular pathway involved in migration within astrocytic tumor context, where both proteins showed altered expression patterns. In this branch of research, the involvement of doppel in cell migration, as a typical hallmark of cancer, was investigated. The movement of cells is a sophisticated mechanism that is rigidly controlled by a spectrum of different genes during the various stages of embryonic and adult development. During carcinogenesis, the motile behaviour of cells is manifested without tight cellular controls and regulations and this causes tumor cells to spread throughout the healthy tissue; consequently, cancer cells motility plays a pivotal role in tumor invasion and in metastasis formation. In particular, in astrocytomas, the infiltration ability

A. Azzalin and S. Comincini

of tumor cells constitutes one of the main causes of the unfortunate prognosis for these malignancies, and it explains the intensive efforts in aiming the development of novel therapies. Several key proteins are described as effectors of migratory ability of astrocytoma cells and deeper analysis of the signaling pathway involved in this process are assiduously carried out to discover novel molecular-based targets (Teodorczyk and Martin-Villalba, 2010). In this context, doppel protein was demonstrated to influence the cell migration, suggesting that this ability may be an intrinsic feature of the protein, independently from the biological scenario, as it was documented in different tumor cell lines, namely glioblastoma- and uterus carcinoma-derived cells (Azzalin et al., 2008). Of note, prion and doppel seemed to differently contribute to the migratory phenotype of tumor cells: in fact, it was demonstrated that doppel expression is related to migration and, in general, to the cell morphology condition, while the expression of the prion counterpart did not affect these phenotypes. In addition, an inverted correlation between migration and growth rates was reported in this study, as first described by Giese et al. (1996) as the paradigm of the “go or grow” dichotomy; however, the molecular and cellular mechanisms involved in these mutually exclusive phenotypes, are up to now not completely clarified. In the case of human astrocytomas, it might also be hypothesized that doppel is expressed mostly in the highly migratory cells of the external edge of the tumor area. In this context, doppel might influence cell migration by means of its membrane localization, promoting cell-to-cell contacts. However, doppel localization in these tumor cells could not directly sustain this speculation because of the prevalent cytoplasmic localization of the protein. For this reason, other protein partners and pathways should be considered to clarify the contribution of doppel in the cell migration process.

Conclusion Since its discovery in 1999, important data have emerged in the biology of the first prion-like protein, doppel. As the function of the cellular prion protein is still rather unclear, with the exception of the involvement of its pathological isoform within the TSEs, the

2

Deciphering the Function of Doppel Protein in Astrocytomas

study of the antagonistic roles of doppel and prion proteins in the neuronal cell survival remains of pivotal interest to directly discern the functions of doppel. Although these proteins share biochemical properties, doppel is unlikely to play a major role in prion diseases. The physiological function of doppel seems to be restricted to the male reproductive apparatus, in particular regulating male fertility. The functional involvement of doppel in a novel identified pathological scenario, such as tumors, in particular those deriving from glial cell transformation, would contribute in the near future to gain further insights into the molecular biology of such dramatic pathologies, possibly with the improvement of the molecular-assisted diagnosis and treatment.

References Aguzzi A, Polymenidou M (2004) Mammalian prion biology: one century of evolving concepts. Cell 116:331–327 Azzalin A, Del Vecchio I, Chiarelli LR, Valentini G, Comincini S, Ferretti L (2005) Absence of interaction between doppel and GFAP, Grb2, PrPC proteins in human tumor astrocytic cells. Anticancer Res 25:4369–4374 Azzalin A, Del Vecchio I, Ferretti L, Comincini S (2006) The prion-like protein doppel (Dpl) interacts with the human receptor for activated C-kinase 1 (RACK1) protein. Anticancer Res 26:4539–4548 Azzalin A, Sbalchiero E, Barbieri G, Palumbo S, Muzzini C, Comincini S (2008) The doppel (Dpl) protein influences in vitro migration capability in astrocytoma-derived cells. Cell Oncol 30:491–501 Comincini S (2001) Searching for molecular markers of human gliomas. Funct Neurol 16:291–298 Comincini S, Foti MG, Tranulis MA, Hills D, Di Guardo G, Vaccari G, Williams JL, Harbitz I, Ferretti L (2001) Genomic organization, comparative analysis, and genetic polymorphism of the bovine and ovine prion Doppel genes (PRND). Mamm Genome 12:729–733 Comincini S, Facoetti A, Del Vecchio I, Peoc’h K, Laplanche JL, Magrassi L, Ceroni M, Ferretti L, Nano R (2004) Differential expression of the prion-like protein doppel gene (PRND) in astrocytomas: a new molecular marker potentially involved in tumor progression. Anticancer Res 24:1507–1517 Comincini S, Del Vecchio I, Azzalin A (2006a) The doppel gene biology: a scientific journey from brain to testis, and return. Central Eur J Biol 1:494–505 Comincini S, Chiarelli LR, Zelini P, Del Vecchio I, Azzalin A, Arias A, Ferrara V, Rognoni P, Dipoto A, Nano R, Valentini G, Ferretti L (2006b) Nuclear mRNA retention and aberrant doppel protein expression in human astrocytic tumor cells. Oncol Rep 16:1325–1332 Comincini S, Ferrara V, Arias A, Malovini A, Azzalin A, Ferretti L, Benericetti E, Cardarelli M, Gerosa M, Passarin MG,

21 Turazzi S, Bellazzi R (2007) Diagnostic value of PRND gene expression profiles in astrocytomas: relationship to tumor grades of malignancy. Oncol Rep 17:989–996 Del Vecchio I, Azzalin A, Guidi E, Amati G, Caramori T, Uboldi C, Comincini S, Ferretti L (2005) Functional mapping of the bovine doppel gene promoter region. Gene 356: 101–108 Giese A, Loo MA, Tran N, Haskett D, Coons SW, Berens ME (1996) Dichotomy of astrocytoma migration and proliferation. Int J Cancer 67:275–282 Heitz S, Gautheron V, Lutz Y, Rodeau JL, Zanjani HS, Sugihara I, Bombarde G, Richard F, Fuchs JP, Vogel MW, Mariani J, Bailly Y (2008) BCL-2 counteracts Doppel-induced apoptosis of prion-protein-deficient Purkinje cells in the Ngsk Prnp(0/0) mouse. Dev Neurobiol 68:332–348 Heitz S, Grant NJ, Leschiera R, Haeberlé AM, Demais V, Bombarde G, Bailly Y (2010) Autophagy and cell death of Purkinje cells overexpressing Doppel in Ngsk Prnp-deficient mice. Brain Pathol 20:119–132 Kiely PA, Baillie GS, Lynch MJ, Houslay MD, O‘Connor R (2008) Tyrosine 302 in RACK1 is essential for insulin-like growth factor-I-mediated competitive binding of PP2A and beta1 integrin and for tumor cell proliferation and migration. J Biol Chem 283:22952–22961 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; Erratum in: Acta Neuropathol 2007. 114:547 Lührs T, Riek R, Güntert P, Wüthrich K (2003) NMR structure of the human doppel protein. J Mol Biol 326:1549–1557 Massimino ML, Ballarin C, Bertoli A, Casonato S, Genovesi S, Negro A, Sorgato MC (2004) Human Doppel and prion protein share common membrane microdomains and internalization pathways. Int J Biochem Cell Biol 36:2016–2031 Mastrangelo P, Westaway D (2001) The prion gene complex encoding PrPC and Doppel: insights from mutational analysis. Gene 275:1–18 Mead S, Beck J, Dickinson A, Fisher EM, Collinge J (2000) Examination of the human prion protein-like gene Doppel for genetic susceptibility to sporadic and variant CreutzfeldtJakob disease. Neurosci Lett 290:117–120 Mo H, Moore RC, Cohen FE, Westaway D, Prusiner SB, Wright PE, Dyson HJ (2001) Two different neurodegenerative diseases caused by proteins with similar structures. Proc Natl Acad Sci USA 98:2352–2357 Moore RC, Lee IY, Silverman GL, Harrison PM, Strome R, Heinrich C, Karunaratne A, Pasternak SH, Chishti MA, Liang Y, Mastrangelo P, Wang K, Smit AF, Katamine S, Carlson GA, Cohen FE, Prusiner SB, Melton DW, Tremblay P, Hood LE, Westaway D (1999) Ataxia in prion protein (PrPC )-deficient mice is associated with up-regulation of the novel PrPC -like protein doppel. J Mol Biol 292:797–817 Paisley D, Banks S, Selfridge J, McLennan NF, Ritchie AM, McEwan C, Irvine DS, Saunders PT, Manson JC, Melton DW (2004) Male infertility and DNA damage in Doppel knockout and prion protein/Doppel double-knockout mice. Am J Pathol 164:2279–2288 Peoc’h K, Serres C, Frobert Y, Martin C, Lehmann S, Chasseigneaux S, Sazdovitch V, Grassi J, Jouannet P, Launay JM, Laplanche JL (2002) The human “prion-like” protein

22 Doppel is expressed in both Sertoli cells and spermatozoa. J Biol Chem 277:43071–43078 Peoc’h K, Volland H, De Gassart A, Beaudry P, Sazdovitch V, Sorgato MC, Creminon C, Laplanche JL, and Lehmann S (2003) Prion-like protein Doppel expression is not modified in scrapie-infected cells and in the brains of patients with Creutzfeldt-Jakob disease. FEBS Lett 11:61–65 Prusiner SB (1998) Prions. Proc Natl Acad Sci USA 95: 13363–13383 Qin K, Zhao L, Tang Y, Bhatta S, Simard JM, Zhao RY (2006) Doppel-induced apoptosis and counteraction by cellular prion protein in neuroblastoma and astrocytes. Neuroscience 141:1375–1388 Rossi D, Cozzio A, Flechsig E, Klein MA, Rülicke T, Aguzzi A, Weissmann C (2001) Onset of ataxia and Purkinje cell loss in PrP null mice inversely correlated with Dpl level in brain. EMBO J 20:694–702 Sakaguchi S (2008) Antagonistic roles of the N-terminal domain of prion protein to doppel. Prion 3:107–111 Sbalchiero E, Azzalin A, Palumbo S, Barbieri G, Arias A, Simonelli L, Ferretti L, Comincini S (2008) Altered cellular

A. Azzalin and S. Comincini distribution and sub-cellular sorting of doppel (Dpl) protein in human astrocytoma cell lines. Cell Oncol 30:337–347 Teodorczyk M, Martin-Villalba A (2010) Sensing invasion: cell surface receptors driving spreading of glioblastoma. J Cell Physiol 222:1–10 Travaglino E, Comincini S, Benatti C, Azzalin A, Nano R, Rosti V, Ferretti L, Invernizzi R (2005) Overexpression of the Doppel protein in acute myeloid leukaemias and myelodysplastic syndromes. Br J Haematol 128:877–884 Watts JC, Huo H, Bai Y, Ehsani S, Jeon AH, Shi T, Daude N, Lau A, Young R, Xu L, Carlson GA, Williams D, Westaway D, Schmitt-Ulms G (2009) Interactome analyses identify ties of PrP and its mammalian paralogs to oligomannosidic N-glycans and endoplasmic reticulum-derived chaperones. PLoS Pathog 5:e1000608 Weissmann C, Aguzzi A (1999) Perspectives: neurobiology. PrPC’s double causes trouble. Science 286: 914–915 Zendman AJ, Ruiter DJ, Van Muijen GN (2003) Cancer/testisassociated genes: identification, expression profile, and putative function. J Cell Physiol 194:272–288

Chapter 3

Astrocytic Tumors: Role of Antiapoptotic Proteins Alfredo Conti, Carlo Gulì, Giuseppe J. Sciarrone, and Chiara Tomasello

Abstract Apoptosis is a fundamental anti-neoplastic mechanism to prevent tumorigenesis. Nearly all neoplastic changes during the development of a normal cell to a cancer cell, such as DNA-damage, oncogene activation or cell cycle deregulation, are potent inducers of the programmed cell death pathway. Therefore, overcoming the apoptotic failsafe is a key mechanism in the genesis and progression of tumors. In astrocytic brain tumors, the apoptotic failure has been documented and involves both the intrinsic or mitochondrial and extrinsic or receptor pathways of apoptosis. This breakdown may be caused by an imbalance of pro- and antiapoptotic members of the Bcl-2 protein family, inhibition of the activity of caspases by specific factors, changes in the p53 system. Detailed molecular knowledge of the anti-apoptotic mechanisms of astrocytic tumor cells is essential for the improvement of conventional chemotherapies and the development of new potent targeted therapies. In this chapter, authors describe the multiple antiapoptotic signals that have been demonstrated to be active in astrocytomas. Keywords Apoptosis · Tumorigenesis · Programmed cell death · FasL · TRAIL · TNF

Introduction The term programmed cell death was introduced by Lockshin and Williams (1964) to describe the destiny of some cells to die as driven by a cell-intrinsic program. Later, Kerr et al. (1972) introduced the term “apoptosis” to describe a series of common morphological features associated with this form of cell death, including cytoplasm shrinkage, membrane blebbing, nuclear fragmentation, intranucleosomal DNA fragmentation, phosphatidylserine exposure, and fragmentation into membrane-enclosed apoptotic bodies sequestered by macrophages or other engulfing cells. Apoptosis has a widespread biological significance, being involved in development, differentiation, proliferation/homeostasis, regulation and function of the immune system, and in the removal of defect and, therefore, harmful cells. Apoptosis is a fundamental anti-neoplastic mechanism in normal cells to prevent tumorigenesis. Nearly all neoplastic changes during the development of a normal cell to a cancer cell, such as DNA-damage, oncogene activation or cell cycle deregulation, are potent inducers of the programmed cell death pathway. Therefore, overcoming the apoptotic failsafe is observed in many cancers.

Death Ligands, Receptors and Messengers

A. Conti () Department of Neuroscience, University of Messina, Messina, Italy e-mail: [email protected]

The extrinsic apoptosis signalling is mediated by the activation of so-called death receptors which are cell surface molecules that transmit apoptotic signals after ligation with specific ligands (Fig. 3.1). The death ligands are members of the TNF superfamily, including: TNFα, CD95 ligand (CD95L), also known as Fas

M.A. Hayat (ed.), Tumors of the Central Nervous System, Volume 5, DOI 10.1007/978-94-007-2019-0_3, © Springer Science+Business Media B.V. 2012

23

24

Fig. 3.1 Apoptosis is activated through two major signalling pathways. The first pathway is the intrinsic or mitochondrial pathway, because the mitochondria take the key position by initiating apoptosis. Initiation by different apoptotic stimuli is still not entirely clear, but likely involves an imbalance of proand antiapoptotic members of the Bcl-2 protein family. This imbalance finally leads to the activation of the proapoptotic Bcl-2 family members BAX and/or BAK and the perturbance of the integrity of the outer mitochondrial membrane. This induces the release of cytochrome c and other apoptotic regulators, like apoptosis-inducing factor (AIF), Smac (second mitochondria-derived activator of apoptosis)/DIABLO (direct inhibitor of apoptosis protein (IAP)-binding protein with low PI), endonuclease G or Omi/HtrA2 from the intermembraneous space of mitochondria. In the cytosol, cytochrome c binds to monomeric APAF-1 which then, in a dATP-dependent conformational change, oligomerizes to assemble the apoptosome, a complex of wheel-like structure with 7-fold symmetry that triggers the activation of the initiator procaspase-9. Caspase9 provokes the cleavage of the executioner caspases, such as caspase-3. Furthermore, the potent endogenous inhibitors of caspases, the inhibitor of apoptosis proteins (IAPs), are neutralized by Smac/DIABLO or Omi/HtrA2. The second pathway is the extrinsic pathway. Extrinsic apoptosis signaling is mediated by the activation of so-called “death receptors”, which are cell surface receptors that transmit apoptotic signals after ligation with specific ligands. Death receptors belong to the tumor necrosis factor receptor (TNFR) gene superfamily, including TNFR-1, Fas/CD95, and the TRAIL receptors DR-4 and DR-5. All members of the TNFR family consist of cysteine rich extracellular

A. Conti et al.

subdomains which allow them to recognize their ligands with specificity, resulting in the trimerization and activation of the respective death receptor. Subsequent signalling is mediated by the cytoplasmic part of the death receptor which contains a conserved sequence termed death domain (DD). Adapter molecules like FADD or TRADD themselves possess their own DDs by which they are recruited to the DDs of the activated death receptor, thereby forming the so-called death inducing signaling complex (DISC). In addition to its DD, the adaptor FADD also contains a death effector domain (DED), which through homotypic DED-DED, interaction sequesters procaspase-8 to the DISC. The local concentration of several procaspase-8 molecules at the DISC leads to their autocatalytic activation and release of active caspase-8. Active caspase-8 then processes downstream effector caspases which subsequently cleave specific substrates resulting in cell death. Cells harboring the capacity to induce such direct and mainly caspase-dependent apoptosis pathways were classified to belong to the so-called type I cells. In type II cells, the signal coming from the activated receptor does not generate a caspase signalling cascade strong enough for execution of cell death on its own. In this case, the signal needs to be amplified via mitochondria-dependent apoptotic pathways. The link between the caspase signalling cascade and the mitochondria is provided by the Bcl-2 family member Bid. Bid is cleaved by caspase-8 and in its truncated form (tBID) translocates to the mitochondria where it acts in concert with the proapoptotic Bcl-2 family members BAX and BAK to induce the release of cytochrome c and other mitochondrial proapoptotic factors into the cytosol (modified from Ashkenazi, 2002)

3

Astrocytic Tumors: Role of Antiapoptotic Proteins

ligand (FasL) or Apo-1 ligand, and Apo-2 ligand/TNFrelated apoptosis-inducing ligand (Apo2L/TRAIL). The FasL and Apo2L/TRAIL act on target cells through binding their specific receptors CD95 and death receptors (DR)4/DR5, respectively. Between ligands and effector caspases, there are a number of factors that suppress apoptosis. The most proximal step to suppress a death receptor pathway is inhibition of ligand binding. This could be achieved by the lack of or mutations in death receptors or the presence of antagonistic (decoy) receptors. In glioma cells, the existence of both agonistic and decoy receptors for TNF has been demonstrated: agonistic Fas receptors are present on the glioma cells (Rieger et al., 1998), but a decoy soluble receptor (DcR3) for FasL is also released by glioma cells, likely protecting them from the FasL-induced apoptosis. The efficacy of TNF-induced apoptosis is enhanced by protein synthesis inhibition, which points to the role of expression of factors with specific antiapoptotic function. The TNF-to-TNFR ligation provokes the formation of the DISC, initiating the apoptotic cascade (Fig. 3.1). However, such activation corresponds to a concurrent and parallel activation of the transcription factor nuclear factor (NF)-κB. The nuclear factor-κB is a dimeric transcription factor controlling the expression of several regulators of immune, inflammatory, and acute phase responses (Conti et al., 2007). A role of NF-κB in the genesis and progression of cancer has also been demonstrated; in particular a constitutive NF-κB activation has been described in a variety of epithelial and lymphoid cancers. As seen in other cell systems, TNF-induced NF-κB activation in astrocytoma cells may be mediated by the TNFR-associated factor (TRAF) family, which consists of a group of six adapter proteins (TRAF1–TRAF6) that participate in the intracellular signalling activity of several members of the TNFR superfamily. Through appropriate ligand stimulation of TNFRs found on the surface of these cells, TRAF proteins can induce activation of NF-κB, resulting in both cytokine secretion and resistance to apoptosis (Conti et al., 2005). The signal transduction mechanism emanating from the TNFR is thought to be mediated by TRAF2, a signalling intermediate that has been shown to be recruited to the cytoplasmic tail of TNFR through a TNFR/TRADD/TRAF2 interaction. On the basis of this hypothesis, TNF can either induce apoptosis through FADD (FAS associating protein with death domain) and caspase recruitment or promote survival

25

through TRAF2 recruitment and NF-κB induction. As the cytosolic NF-κB concentration rises, the expression of several antiapoptotic genes is amplified. Candidate antiapoptotic genes for NF-κB induction include, cIAP (inhibitor of apoptosis) 1 and 2, Bcl-2, Bcl-X L, XIAP, and survivin. Characteristics and the role of these antiapoptotic factors will discussed in specific sections in this chapter. There are other possible anti-apoptotic mechanisms related to the death ligands and receptors. The U373MG cell line appears to be resistant to death receptor mediated apoptosis due to lack of crucial signalling components. Expression of caspase-8 sensitizes this cell line to FasL and TRAIL mediated apoptosis. Furthermore, recent studies have reported methylation of the caspase-8 gene as a mechanism for decreased levels of protein expression in neuroblastomas, rendering cells resistant to apoptosis (Teitz et al., 2000). Recent progress in the understanding of the varying susceptibility of glioma cell lines to Apo2L/TRAILinduced apoptosis has revealed that resistant cell lines expressed 2-fold higher levels of the apoptosis inhibitor phosphoprotein enriched in diabetes/phosphoprotein enriched in astrocytes-15 kDa (PED/PEA-15). This phosphoprotein protects astrocytes from TNF-α-induced apoptosis through interruption of FADD-Caspase-8 binding (Condorelli et al., 1999). Preclinical studies have established the potential for using TNF factors as a therapy in gliomas. In the field of death ligands and receptors, inducing cancer cell apoptosis via local or systemic application of Apo2L/TRAIL is one of the most promising strategies. Local injection of TRAIL exerted strong antitumor activity on intracranial human malignant glioma xenografts in athymic mice without neurotoxicity (Roth et al., 1999). However, a significant number of glioma cell lines remain resistant to TRAIL when it is used as monotherapy. In combination with conventional DNA-damaging chemotherapy, TRAIL showed synergistic cytotoxicity for human gliomas in vivo and in vitro. Of particular concern is the fact that TRAIL and FasL administration have been shown to have profound toxicity toward normal human hepatocytes, resulting in a massive and rapid induction of cell death. The local application of adenoviral vectors expressing FasL may be a strategy to circumvent systemic side effects in gliomas. Transferring the gene encoding FADD into glioma cells also inhibits glioma growth in vitro and in vivo. Finally, even more down-stream

26

effectors of death receptor-mediated apoptosis, the caspases, have been successfully employed to promote glioma cell death.

The p53 The p53 is a nuclear phosphoprotein which acts as a tumor suppressor. The gene for p53 is located on the short arm of chromosome 17 at 17p13.105-p12. The open reading frame of p53 encodes for a protein with 393 amino-acids (53 kDa). The central region of the protein contains the DNA-binding domain. The structure of p53 consists of a large beta-sandwich which encompasses three loop-based elements and is composed of two anti-parallel beta-sheets encompassing four and five beta-strands, respectively. The first loop binds to DNA within the major groove and the second loop binds to DNA within the minor groove. The function of the third loop is stabilization of the second loop. p53 has been considered “the guardian of the genome” because it plays a key role in several processes of cellular physiology including control of cell cycle, genome stability, senescence, angiogenesis, and induction of apoptosis. Those activities are promoted through pathways that are both dependent and independent by transcription regulation. Mutations involving the p53 gene are the most common among those occurring in cancer, with more than one half of human cancers expressing a mutant p53. Particularly, it is the DNA-binding domain of the protein to be frequently modified in the neoplastic cells. p53 exerts its function only after the constitution of homo-oligomeric complexes (homo-tetramers). A single mutation for each tetramer is sufficient to compromise the possibility of DNA-binding and interaction. Wild-type p53 has a shorter half-life (∼20 min) than mutated p53 (3–7 h). This means that in cancer, even in heterozygosis (a normal TP53 allele and a mutated TP53 allele) p53 homotetramers are formed often by mutant monomers. The physiological role of p53 in normal cells depends on correct expression of regulator proteins with activation and inhibition functions.

A. Conti et al.

including gliomas. MDM2 encodes for p90MDM2 , which negatively regulate p53 by promoting its degradation through translocation from cell nucleus into cytoplasm, where p53 is a target of the proteasome 26S. Proteasomes 26S are large catalytic complexes responsible for extra-lisosomial endocellular proteolysis. In senescent cells, p90MDM2 expression is downregulated, and the inhibition of p53 is also depressed. Conversely, in gliomas amplified MDM2 inhibits p53 activity. With an opposite mechanism the p53 activity is stabilized by other factors. When a DNA damage occurs due to radiation or others toxic agents, it activates DNA-dependent protein kinase (DNAPK), ataxia telangiectasia mutated (ATM), ATM- and Rad3-related (ATR) kinases. DNA-PK is a nuclear serine/threonine kinase composed of a catalytic subunit and a DNA binding subunit. It has close interaction with p53, because they form a protein complex that preferentially binds to abnormal DNA structures. Also ATM and ATR can be activated by DNA damage, although the mechanism is not exactly known. These proteins can phosphorylate p53 on a specific serine residue in 15 position avoiding its degradation by the product of MDM2, p90MDM2 . This stabilization increases p53 half-life and its intranuclear concentration, activating p21, also known as cyclin-dependent kinase inhibitor 1A (CKN1A). The p21 is a gene located on chromosome 6 (6p21.2), and encodes a potent cyclin-dependent kinase inhibitor that is able to bind to and inhibit the activity of cyclin-CDK2 or -CDK4 complexes, major regulators of cell cycle progression at G1. In this way, raised p21 activity arrests the cell cycle in G1 phase (Fig. 3.2). Stabilization of p53 also leads to activation of transcription of several genes involved in apoptosis such as GADD45, PCNA, WAF1/CIP1, MDM2, BAX, NOXA, PUMA, KILLER/DR5, PIG, Caspase-1, and inhibition of others such as Bcl-2 and survivin. Particularly, the key role in apoptosis is repression of Bcl-2 and survivin, and raising the activity of BAX in a manner discussed in other sections.

P53 and Cell Cycle Progression Regulation of p53 Activity p53 activates transcription of the Murine Double Minute 2 (MDM2) gene. This is amplified in many tumors

Blockage of cyclin-cdk2 -cdk4 modulated by increased p21 level is also involved in regulating the activity of Rb protein. This is a tumor suppressor that is

3

Astrocytic Tumors: Role of Antiapoptotic Proteins

27

Fig. 3.2 Mechanism of p53 control of cell cycle progression. Amplification of MDM2 occurs in many tumors including glioma and is the most important inhibiting factors of p53. On the other hand, DNA-pk stabilizes p53 increasing its half-life and its intranuclear concentration. P53 activates p21. p21 is a potent cyclin-dependent kinase inhibitor that is able to bind to and inhibit the activity of cyclin-CDK2 or -CDK4 complexes, major regulators of cell cycle progression at G1. The complex cyclin D1-CDK4 can cause the hyper-phosphorylation of

the retinoblastoma protein (Rb) causing the progression of the cell cycle. When Rb is hypophosphorylated, it keeps locked the E2F-DP complex, and the cell remains in the G1 stage. Hyperphosphorylation of Rb determines dissociation of E2F-DP from Rb and its activation. Unbound E2F activates factors like cyclins (e.g. cyclin E and A), which push the cell through the cell cycle by activating cyclin-dependent kinases

dysfunctional in many cancer types. The name Rb derives from the retinoblastoma, the tumor in which this protein was identified. It is encoded by the RB1 gene that is located on chromosome 13 (13q14.1q14.2). Because it is a tumor suppressor gene, both alleles must be mutated for the development of cancer. Rb can be hyper-phosphorylated or hypophosphorylated. Hyper-phosphorylation, due to raised activity of cyclin-cdk4, inhibits Rb. On the other hand, hypophosphorylated Rb represents the active form, and it is able to inhibit cell cycle progression by binding and inhibiting the transcription factors of the E2F family. E2F is a transcription factor widely believed to integrate cell-cycle progression with the transcription apparatus through its cyclical interactions with important regulators of the cell cycle, such as Rb, cyclins and cyclin-dependent kinases. E2F exerts its role by binding specific DNA sequences to the promoter of the target genes. E2F family protein has a complex structure that presents some domains. A specific domain is responsible for binding with a protein called DP, that functions as an inhibitor of E2F. As long as Rb is hypophosphorylated, it keeps locked E2FDP complex, and the cell remains in the G1 stage. Hyper-phosphorylation of Rb determines dissociation of E2F-DP from Rb and its activation. Unbound E2F

activates factors such as cyclins (e.g. cyclins E and A), which push the cell through the cell cycle by activating cyclin-dependent kinases, and a molecule called proliferating cell nuclear antigen (PCNA) which favors the DNA replication.

P53 and p16/INKa4 In glioblastoma and anaplastic astrocytomas, a correlation between p53, MDM2, p16/INKa4, PTEN, and EGFR and survival rates seem to be present. Loss of p16/INKa4 is essential for maintenance of the transformed neoplastic phenotype (Shapiro et al., 1995). This property of p16/INKa4 protein suggests that it is a tumor suppressor gene product that exerts its function in association with p53, both of which inhibit cyclindependent kinases involved in the Rb pathway. It is probable that cells unable to produce p16/INKa4 protein may be vulnerable to neoplastic transformation. p16/INKa4 protein is a potent inhibitor of cdk-4 that blocks cdk4-mediated phosphorylation of the tumor suppressor Rb protein, allowing Rb-mediated growth suppression (Fig. 3.2). The cdk4/cyclin D1 complex phosphorylates the Rb protein, thereby inducing release of the E2F transcription factor that activates

28

genes involved in the G1 to S transition. p16/INKa4 binds to CDK4, inhibits the cdk4/cyclin D1 complex, and thus inhibits the G1 to S transition. Thus, loss of normal RB1 function may result from altered expression of any of the RB1, p16/INKa4, or Cdk4 gene. This means that pathways of both p53 and p16/INKa4 have a convergence of activity on the pRB protein.

P53 and PTEN A broad relationship between p53 and PTEN (phosphatase and tensin homolog deleted on chromosome 10) has been described. PTEN is a tumor suppressor gene whose mutation was found in various types of sporadic tumors, and was originally described in malignant gliomas. The promoter of PTEN has a p53 binding site. In fact, wild-type p53 can promote PTEN transcription, whereas TGF-β down-regulates it. The protein encoded by PTEN is involved in the regulation of many important cellular functions, such as cellcycle progression, cell migration and spreading, cell growth, and apoptosis. As p53 modulates transcription of PTEN gene, the protein protects p53 from MDM2mediated degradation through a PI3K/Akt pathway. In fact, there is a positive feedback between p53 and PTEN which aims to control the cellular response to stress, DNA damage, and cancer. PTEN has a phosphatase activity against phosphoinositide substrates; it dephosphorylates, with high activity, the 3 -OH position of the inositol ring of phosphatidylinositol phosphates, in particular of phosphatidylinositol 3-phosphate (PIP3), thereby acting as the counterpart of phosphoinositide-3-kinase (PI3K). Actually, PI3K phosphorylates phosphatidynositol4,5-bisphosphate to the respective 3-phosphate (PIP3), which functions as a second messenger molecule that is important for the activation of protein kinase B/Akt (for reviews, see Vanhaesebroeck and Alessi, 2000). In addition, PIP3 facilitates the translocation of Akt to the plasma membrane and activates PDK1, which in turn phosphorylates Akt on threonine 308 within the kinase domain. Activated Akt in turn can phosphorylate a variety of substrates and thereby regulates important cellular processes, including cell-cycle progression, cell growth, cell survival, cell motility and adhesion, translation of mRNA into protein, glucose metabolism, and angiogenesis. PTEN may influence the activity

A. Conti et al.

of several cellular signalling pathways other than the PI3K/Akt pathway. For example, the lipid phosphatase activity of PTEN may also contribute to the inhibition of Ras and MAPK pathway activation by EGF. In gliomas, PTEN mutations are preferentially found in glioblastomas, with reported frequencies of up to 40%. PTEN mutations are frequent in primary (de novo) glioblastomas, less frequent (<10% of the cases) in secondary glioblastomas, namely in glioblastomas that have progressed from pre-existing lower grade gliomas. PTEN showed to play an important role in maintaining the cellular susceptibility to apoptotic stimuli. Gene transfer into U-87 MG glioma cells that lack wild-type PTEN renders the cells susceptible to apoptosis. In addition, PTEN gene transfer sensitizes glioma cells for apoptosis when irradiated or treated with FasL. The regulatory function of PTEN on apoptosis is dependent on PI3K/Akt signaling. PTEN wild-type cells overexpressing mutant (constitutively active) Akt are resistant to multiple apoptotic stimuli, as are PTEN-mutant cells. Akt itself is a serine/threonine kinase that can phosphorylate several apoptosis-associated proteins, including the proapoptotic factor BAD. Upon phosphorylation, BAD can no longer function as a proapoptotic molecule and dissociates from BclX L or Bcl-2, leading to a relative increase in the level of anti-apoptotic proteins. Although the role of BAD in Akt-mediated inhibition of apoptosis is well characterized, it is likely that PTEN/PI3K/Akt signaling may additionally influence apoptosis using other targets, such as glycogen synthase kinase 3, p70S6k, or IkB/NF-κB. There are also recent evidence of the involvement of the PTEN/PI3K/Akt signaling pathway in the control of death receptor- and drug-induced apoptosis. Recently, Opel et al. (2008) demonstrated that inhibition of PI3K by LY294002 sensitizes wild-type and mutant PTEN glioblastoma cells to both death receptorand chemotherapy-induced apoptosis. LY294002 significantly enhances apoptosis triggered by TRAIL, agonistic anti-CD95 antibodies, or several anticancer drugs (i.e., doxorubicin, etoposide, and vincristine) in a highly synergistic manner. In addition, LY294002 cooperates with TRAIL or doxorubicin to suppress colony formation, thus also showing a strong effect on long-term survival. Similarly, genetic knockdown of PI3K subunits p110α and/or p110β by RNA interference (RNAi) primes glioblastoma cells for TRAIL- or doxorubicin-mediated apoptosis. Analysis of apoptosis

3

Astrocytic Tumors: Role of Antiapoptotic Proteins

pathways revealed that PI3K inhibition acts in concert with TRAIL or doxorubicin to trigger mitochondrial membrane permeabilization, caspase activation, and caspase-dependent apoptosis.

The Bcl-2 Family of Proteins Bcl-2 is the second member of a family of proteins initially described as an expression of a reciprocal gene translocation in chromosomes 14 and 18 in follicular lymphomas; the name stands for B-cell lymphoma 2. The Bcl-2 proteins comprise a large family of proteins which share a Bcl-2 homology (BH) region that interacts by heterodimerization, and either inhibits or promotes apoptosis (Table 3.1). Bcl-2 controls apoptosis through the mitochondrial pathway and through the control of outer mitochondrial membrane integrity and function. The molecular mechanism of activity and how members of this family of proteins interacts with one another are not completely understood. Among members of this family, Bcl-2 is the best known. Bcl-2 is a 25 kDa protein located in mitochondrial membrane, in endoplasmatic reticulum, and in perinuclear envelopment; it is an anti-apoptotic factor. It is likely that Bcl-2 exerts its function through two different mechanisms. It functions like a mitochondrial ionic channel that prevents the opening of permeability transition pores and the outgoing of caspases activating factors such as cytochrome c and Smac/Diablo. On the other hand, it is a molecular Table 3.1 Members of the Bcl family and their functions Genes/molecules Effects Bcl-2 subfamily Bcl-2 Bcl-xL Bcl-w Bcl-xs Bax subfamily Bax Bak Bok BH3 subfamily Bad Bik Bid Blk HRK

Promotes survival Promotes survival Promotes survival Promotes death Promotes death Promotes death Promotes death Promotes death Promotes death Promotes death Promotes death Promotes death

29

adapter able to inhibit activity of some cytoplasmatic apoptogenic factors such as Apaf-1, through an endocellular membrane linkage mechanism. The threedimensional structure of Bcl-2 consists of two central, predominantly hydrophobic alpha-helices surrounded by six or seven amphipathic alpha-helices of varying lengths. A long, unstructured loop is present between the first two alpha-helices. The structure of the Bcl2 proteins show a striking similarity to the overall fold of the pore-forming domains of bacterial toxins (Petros et al., 2004). The Bcl-2 family is divided into three different groups based on Bcl-2 homology (BH) domains and function. The antiapoptotic members, such as Bcl2 and Bcl-X L, typically have BH1 through BH4 domains. The proapoptotic members can be divided into two groups: those with BH1, BH2 and BH3 domains, such as BAX and BAK, and those with only BH3 domains, such as Bad, and Bim. It has been described that cellular susceptibility to apoptotic stimuli depends on expression levels of different kinds of Bcl-2 family members. For example, the lymphocytes in the germinative centre of lymphoid follicles and other cells with short life have higher expression of BAX gene, whereas memory B lymphocytes and other cell with prolonged life have higher expression of Bcl-2 gene. Bcl-2 family members exert their function through three probable mechanisms of action: (A) Bcl-2 and Bcl-X L binding to Apaf-1 is an essential checkpoint of cellular apoptosis, which controls activity of cytoplasmic caspase-9 by mitochondria. Bcl-2 and Bcl-X L inhibit linkage of caspase9 to Apaf-1, avoiding the apoptotic fall. (B) Bcl-2 family members prevent release of mitochondrial apoptogenic factors such as cytochrome c and AIF (apoptosis-inducing factor) into the cytoplasm, maintaining organelle integrity, while the proapoptotic BH3 member BID is able to facilitate release of cytochrome c. (C) BAX is probably involved in caspase-independent death through a channel forming activity, which could promote the mitochondrial permeability transition. Even though the exact mechanism of action of Bcl-2 family is not completely understood, it is known that p53 protein interacts with both its proapoptotic

30

and anti-apoptotic members. The binding of p53 to Bcl-2 and Bcl-X L inhibits their anti-apoptotic functions. Furthermore, the interaction between p53 and the proapoptotic proteins BAX or BAK activates them, inducing conformational changes required to exert their proapoptotic function. There are also posttranslational mechanisms of regulation: in IL-3 stimulated blood cells, the BH3 subfamily member BAD, for example, is phosphorylated by Akt and the product is sequestered in the cytosol by proteins, preventing the BAD-induced death (Datta et al., 1997). In the case of pro-survival members, phosphorylation may both augment and suppress activity: a candidate region for phosphorylation and activity-regulation is an unstructured loop of ∼60 amino acids in Bcl-X L and Bcl-2 (the loop sequence is not conserved between those two Bcl-2 members). Loop-deletion mutants of BclX L and Bcl-2 demonstrate an enhanced anti-apoptotic function. Interestingly, heterodimerization with BAX is unaffected in the case of loop deletion. But the loop domain influences the phosphorylation status of Bcl2: wildtype Bcl-2 with loop domain is susceptible to phosphorylation while loop-deleted Bcl-2 is not phosphorylated. Accordingly, it is possible that there are phosphorylation sites (serine and/or threonine) within the loop itself, or the loop may represent a recognition site for a kinase that phosphorylates the protein at different sites. Dysregulation of Bcl-2 family members is an important oncogenic mechanism in gliomas. In contrast to the expectation that the expression of antiapoptotic members of the Bcl-2 family increases with malignancy, stronger staining in astrocytoma and anaplastic astrocytoma compared with GBM has been observed in most studies (Angileri et al., 2008; Steinbach and Weller, 2004). This is probably due to the major role played by apoptosis in low grade as compared with high grade gliomas. In GBM cells, a mutant form of Bcl-2 protein, called Bcl-2L12 has been described in relation to apoptosis resistance and propensity for necrosis. Levels of this abnormal oncoprotein are very high in these cells. Recent studies have shown as Bcl2L12 is able to arrest apoptosis at post-mitochondrial level acting on effector caspase-3 and 7 via distinct mechanisms. A direct physical interaction seems to underlies the Bcl-2L12 inhibition process of caspase-7, whereas Bcl-2L12-induced transcriptional upregulation of the small heat shock protein alpha B-crystallin is needed to neutralize caspase-3 activation (Stegh

A. Conti et al.

et al., 2008). On the other hand, Bcl-2L12 expression does not affect cytochrome c release or apoptosomedriven caspase-9 activation. Another significant finding suggesting the mechanism by which Bcl-2 protein is involved in antiapoptotic process in glioma cell lines, is represented by the evidence that pharmacological inhibition of Bcl-2 family members, by BH3-mimetics for example, reactivates TRAIL-induced apoptosis. In fact, treatment with TRAIL in combination with the specific Bcl2 inhibitor HA14-1, and the Bcl-2/Bcl-X L inhibitor BH3I-2 , potently enhanced apoptosis in some glioma cell lines. These data indicate that Bcl-2 and Bcl-X L play fundamental roles in TRAIL-mediated apoptosis resistance of malignant gliomas and suggest that using TRAIL or agonistic TRAIL receptor antibodies in combination with BH3 mimetics may represent a promising approach to reactivate apoptosis in therapyresistant high grade gliomas (Hetschko et al., 2008).

Inhibitor of Apoptosis Proteins As previously noted, caspases play a crucial role in the apoptotic process. There are two ways to terminate the activity of a caspase: 1) remove it from the cell by using the ubiquitin-targeted proteasome degradation machinery; 2) directly inhibit its enzymatic activity. Intriguingly, there is evidence that members of a family of proteins called “inhibitor of apoptosis protein” (IAP) are capable of both functions (Riedl and Shi, 2004; Salvesen and Duckett, 2002). The IAPs are a family of caspase inhibitors that specifically inhibits caspases 3, 7, and 9, thereby preventing apoptosis. To date, eight human IAP family members have been identified (NAIP, c-IAP1, c-IAP2, XIAP, ILP2, MLIAP, Apollon, and Survivin). Inhibitor of apoptosis proteins are grouped into this family based on the presence of one to three baculovirus IAP repeat (BIR) domains, a zinc-binding region of about 70 aminoacids. Those proteins may also contain a RING or caspase activation recruitment domain (CARD) (reviewed by Salvesen and Duckett, 2002) (Fig. 3.3). The IAP proteins can be classified into three classes based on the presence or absence of a RING finger and the homology of their BIR domains. Class 1 IAPs contain homologous BIR domains and a RING finger motif. X-linked IAP [XIAP (also known as hILP,

3

Astrocytic Tumors: Role of Antiapoptotic Proteins

31

Fig. 3.3 Proteins of the inhibitor of apoptosis (IAP) family include XIAP (X-linked IAP), c-IAP1, c-IAP2, ILP2 (IAPlike protein-2), ML-IAP (melanoma IAP)/Livin, NAIP (neuronal apoptosis-inhibitory protein) and survivin. A conserved linker peptide that precedes the BIR2 (baculoviral IAP repeat-2)

domain of XIAP, c-IAP1 or c-IAP2 (shown in red) is responsible for inhibiting caspases-3 and -7 in mammals. The BIR3 domain of XIAP can potently inhibit caspase-9 (modified from Riedl and Shi, 2004)

MIHA, and BIRC4)] has three BIR domains and a RING finger. It was the first IAP in this class to be identified, and remains the best characterized. It binds and inhibits caspases 3, 7, and 9 with nanomolar affinity, but it does not bind or inhibit caspase-8. Cellular IAP1 (c-IAP1) (also known as MIHB, hiap2, and BIRC2) and c-IAP2 (also known as MIHC, hiap2, and BIRC3) are structurally related to XIAP with three BIR domains and a RING finger. ML-IAP (also known as livin, KIAP, and BIRC7) and ILP-2 have a RING finger and only one BIR domain, but their BIR domain is most homologous to the BIR3 domain of XIAP, cIAP1, and cIAP2. The class 2 IAP family member NAIP has three BIR domains but no RING finger motif. Its BIR domains are more distantly related to the BIR domains of the class 1 IAPs. It inhibits caspases 3 and 7, but not caspases 1, 4, 5, or 8. Class 3 IAP members, that include Survivin (also known as TIAP, and BIRC5) and Apollon (also known as BIRC6 or BRUCE), contain only a single BIR domain and no RING finger. Biochemical and structural analyses have mapped the elements in XIAP that are devoted to the inhibition of caspase activity: a region encompassing the second BIR domain (BIR2) inhibits caspase-3 and caspase-7, whereas the third BIR domain (BIR3) inhibits caspase-9 (Fig. 3.3). Caspases are proteolytically processed between their large and small subunits

as a result of either autocatalytic processing following dimerization or the activity of upstream proteases. There are also evidence that XIAP might participate in caspase polyubiquitination and monoubiquitination in vitro; however, this function is yet to be confirmed in vivo with endogenous proteins. Ubiquitination is a post-translational protein modification procedure that plays important roles in apoptosis and signal transduction. By operating the processes of ubiquitin activating enzyme (E1), ubiquitin conjugating enzyme (E2), and ubiquitin protein ligase (E3), target proteins are attached by ubiquitins. They are in turn recognized and degraded by some proteasomes.

Activation of IAPs in Gliomas The transcription of IAPs involves activation of the transcription factor NF-κB and IAPs may be the predominant anti-apoptotic factor induced by NF-κB. Two of the first cellular IAPs (c-IAP1 and c-IAP2) identified, were isolated by virtue of their interaction with the TRAF2 (Rothe et al., 1995). TNFα has been shown in Jurkat cells to upregulate c-IAP1 through NF-κB. Moreover, c-IAP1 protection from TNFα killing was found to be dependent upon the presence of NF-κB. Conversely c-IAP1 and XIAP activate NF-κB thus leading to a positive feedback loop. Accordingly,

32

upregulation of cIAP1, cIAP2 and XIAP following exposure to TNFα appears to be downstream of NF-κB given the fact that overexpression of the NF-κB inhibitor, IkB, totally suppresses IAPs induction and sensitizes cells to apoptotic stimuli and, conversely, overexpression of XIAP reconstitutes cytoprotection, abrogating the IkB inhibition. In this regard, TNFα and TNF receptors can trigger both apoptotic and anti-apoptotic responses in the cell; the latter being mediated chiefly by the transcription factor NF-κB (Beg and Baltimore, 1996). The activation NF-κB may involve a number of second messengers including TRAF2 (Conti et al., 2005) and others. XIAP and NAIP, for example, form a complex with the TAK1 kinase and its cofactor, TAB1, that leads to activation of c-Jun-NH2-terminal kinase 1 (Sanna et al., 2002). Activated c-Jun-NH2-terminal kinase 1 subsequently activates NF-κB through the mitogen activated protein kinase (MAPK) phosphorylation cascade. In addition, XIAP promotes the translocation of the NF-κB p65 subunit to the nucleus, which is a prerequisite for NF-κB activity. Finally, XIAP promotes the degradation of the NF-κB inhibitor IkB.

Survivin and Cell Cycle Progression Among the members of IAP family, survivin is getting increasing attention in the field of oncology research. Its expression has been detected in a number of different cancers, including glioma. Furthermore, a previous analysis of 3.5 million human transcriptomes identified survivin as among the top four transcripts uniformly upregulated in cancers (Velculescu et al., 1999). Thus, survivin expression is one of the most tumor-specific of all human gene products. Recently, several studies using RT-PCR, immunohistochemistry, and Western blot have shown that the mRNA and protein of survivin are overexpressed in gliomas. Moreover, overexpression of survivin is significantly associated with tumorigenesis and progression of gliomas, as well as the poor prognosis of patients with gliomas. Initially, survivin was described as an inhibitor of caspase-9. However, over the last years, research studies have shown that the role of Survivin in cancer pathogenesis is not limited to apoptosis inhibition but also involves the regulation of the mitotic spindle checkpoint and the promotion of angiogenesis and chemoresistance.

A. Conti et al.

The survivin gene spans 14.7 kb on the telomeric position of chromosome 17 and is transcribed from a TATA-less, GC-rich promoter to generate the wild-type transcript and four different splice variant mRNAs. Survivin is a 16.5 kDa protein of 142 amino acids and is composed of a single BIR domain and an extended C-terminal α-helical coiled-coil domain; it does not contain the RING-finger domain found in other IAPs (Fig. 3.3). Survivin functions as chromosomal passenger protein that localizes to kinetochores at metaphase, transfers to the central spindle mid-zone at anaphase and accumulates in mid-bodies at telophase. Physical interactions with the inner centromere protein, Aurora B and Borealin/Dasra B (Pennati et al., 2007) are required to target the complex to the kinetochore, properly form the bipolar spindle and complete cytokinesis. Such a function of preservation of genome fidelity and regulation of microtubule dynamics requires a sharp cell-cycle-dependent transcription of the survivin gene during the mitotic phase as well as post-translational modifications of the protein including phosphorylation by the p34cdc2 and Aurora B kinases and monoubiquitination through Lys48 and Lys63 linkages (Pennati et al., 2007). It has been recently suggested that this pathway is dominant in normal cells and constitutes the primary function of survivin in adult tissues (Altieri, 2006). However, an up-regulation of survivin in G2 /M cell compartments has been observed in various cancer cell lines. These findings, together with the evidence that knockout mice are characterized by a catastrophic defect of microtubule assembly, with absence of mitotic spindle, formation of multinucleated cells and 100% embryonic lethality are consistent with a critical role of survivin in mitosis to preserve the mitotic apparatus and to allow normal mitotic progression. The fraction of survivin produced through non-cellcycle-dependent mechanisms mediates apoptosis inhibition through intermolecular cooperation with cofactors including the hepatitis B virus X-interacting protein, a target of the oncogenic viral hepatitis B virus X protein and XIAP (Dohi et al., 2004), leading to the formation of complexes that inhibit caspase-9 processing. Moreover, subcellular compartmentalization of survivin in mitochondria seems to play a role in the anti-apoptotic function of the protein. Specifically the existence of a mitochondrial pool of survivin, which is able to orchestrate a novel pathway of apoptosis inhibition in tumor cells, has recently been reported.

3

Astrocytic Tumors: Role of Antiapoptotic Proteins

In conclusion, deregulation of the apoptosis processes and evasion of apoptosis is a general mechanism in gliomas. Since nearly every step during carcinogenesis activates apoptosis, the development of antiapoptotic strategies must be an early and essential event. Contribution of molecules like Bcl-2, c-IAPs or Survivin to early steps in the carcinogenesis of glioma should be further investigated. Also therapeutic resistance is mediated by the profound deregulation of the apoptotic machinery in glioma cells. Detailed molecular knowledge of the anti-apoptotic mechanisms of glioma cells is essential for the improvement of conventional chemotherapies and the development of new potent targeted therapies. Hopefully, this will lead to an improvement of the prognosis of this dismal disease.

References Altieri DC (2006) The case for survivin as a regulator of microtubule dynamics and cell-death decisions. Curr Opin Cell Biol 18:609–615 Angileri FF, Aguennouz M, Conti A, La Torre D, Cardali S, Crupi R, Tomasello C, Germanò A, Vita G, Tomasello F (2008) Nuclear factor-kappaB activation and differential expression of survivin and Bcl-2 in human grade 2-4 astrocytomas. Cancer 15:2258–2266 Ashkenazi A (2002) Targeting death and decoy receptors of the tumour-necrosis factor superfamily. Nat Rev Cancer 2: 420–430 Beg AA, Baltimore D (1996) An essential role for NF-kappaB in preventing TNF-alpha-induced cell death. Science 274: 782–784 Condorelli G, Vigliotta G, Cafieri A, Trencia A, Andalo P, Oriente F, Miele C, Caruso M, Formisano P, Beguinot F (1999) PED/PEA-15: an anti-apoptotic molecule that regulates FAS/TNFR1-induced apoptosis. Oncogene 18: 4409–4415 Conti A, Ageunnouz M, La Torre D, Cardali S, Angileri FF, Buemi C, Tomasello C, Iacopino DG, D’Avella D, Vita G, Tomasello F (2005) Expression of the tumor necrosis factor receptor-associated factors 1 and 2 and regulation of the nuclear factor-kappaB antiapoptotic activity in human gliomas. J Neurosurg 103:873–881 Conti A, Miscusi M, Cardali S, Germanò A, Suzuki H, Cuzzocrea S, Tomasello F (2007) Nitric oxide in the injured spinal cord: synthases cross-talk, oxidative stress and inflammation. Brain Res Rev 54:205–218 Datta SR, Dudek H, Tao X, Masters S, Fu H, Gotoh Y, Greenberg ME (1997) Akt phosphorylation of BAD couples survival signals to the cell-intrinsic death machinery. Cell 91:231–241 Dohi T, Beltrami E, Wall NR, Plescia J, Altieri DC (2004) Mitochondrial survivin inhibits apoptosis and promotes tumorigenesis. J Clin Invest 114:1117–1127 Hetschko H, Voss V, Senft C, Seifert V, Prehn JH, Kögel D (2008) BH3 mimetics reactivate autophagic cell death in anoxia-resistant malignant glioma cells. Neoplasia 10: 873–885

33 Kerr JF, Wyllie AH, Currie AR (1972) Apoptosis: a basic biological phenomenon with wide-ranging implications in tissue kinetics. Br J Cancer 26:239–257 Lockshin RA, Williams CM (1964) Programmed cell death. II. Endocrine potentiation of the breakdown of the intersegmental muscles of silkmoths. J Insect Physiol 10:643–649 Opel D, Westhoff MA, Bender A, Braun V, Debatin KM, Fulda S (2008) Phosphatidylinositol 3-kinase inhibition broadly sensitizes glioblastoma cells to death receptor- and drug-induced apoptosis. Cancer Res 68:6271–6280 Pennati M, Folini M, Zaffaroni N (2007) Targeting survivin in cancer therapy: fulfilled promises and open questions. Carcinogenesis 28:1133–1139 Petros AM, Olejniczak ET, Fesik SW (2004) Structural biology of the Bcl-2 family of proteins. Biochim Biophys Acta 1644:83–94 Riedl SJ, Shi Y (2004) Molecular mechanisms of caspase regulation during apoptosis. Nat Rev Mol Cell Biol 5:897–907 Rieger J, Naumann U, Glaser T, Ashkenazi A, Weller M (1998) APO2 ligand: a novel lethal weapon against malignant glioma? FEBS Lett 427:124–128 Roth W, Isenmann S, Naumann U, Kügler S, Bähr M, Dichgans J, Ashkenazi A, Weller M (1999) Locoregional Apo2L/TRAIL eradicates intracranial human malignant glioma xenografts in athymic mice in the absence of neurotoxicity. Biochem Biophys Res Commun 265(2):479–483 Rothe M, Pan MG, Henzel WJ, Ayres TM, Goeddel DV (1995) The TNFR2-TRAF signaling complex contains two novel proteins related to baculoviral inhibitor of apoptosis proteins. Cell 83:1243–1252 Salvesen GS, Duckett CS (2002) IAP proteins: blocking the road to death’s door. Nat Rev Mol Cell Biol 3:401–410 Sanna MG, da Silva Correia J, Ducrey O, Lee J, Nomoto K, Schrantz N, Deveraux QL, Ulevitch RJ (2002) IAP suppression of apoptosis involves distinct mechanisms: the TAK1/JNK1 signaling cascade and caspase inhibition. Mol Cell Biol 22:1754–1766 Shapiro GI, Park JE, Edwards CD, Mao L, Merlo A, Sidransky D, Ewen ME, Rollins BJ (1995) Multiple mechanisms of p16INK4A inactivation in non-small cell lung cancer cell lines. Cancer Res 55:6200–6209 Stegh AH, Kesari S, Mahoney JE, Jenq HT, Forloney KL, Protopopov A, Louis DN, Chin L, De Pinho RA (2008) Bcl2L12-mediated inhibition of effector caspase-3 and caspase-7 via distinct mechanisms in glioblastoma. Proc Natl Acad Sci USA 105:10703–10708 Steinbach JP, Weller M (2004) Apoptosis in gliomas: molecular mechanisms and therapeutic implications. J Neurooncol 70:245–254 Teitz T, Wei T, Valentine MB, Vanin EF, Grenet J, Valentine VA, Behm FG, Look AT, Lahti JM, Kidd VJ (2000) Caspase 8 is deleted or silenced preferentially in childhood neuroblastomas with amplification of MYCN. Nat Med 6:529–535 Vanhaesebroeck B, Alessi DR (2000) The PI3K-PDK1 connection: more than just a road to PKB. Biochem J 346:561–576 Velculescu VE, Madden SL, Zhang L, Lash AE, Yu J, Rago C, Lal A, Wang CJ, Beaudry GA, Ciriello KM, Cook BP, Dufault MR, Ferguson AT, Gao Y, He TC, Hermeking H, Hiraldo SK, Hwang PM, Lopez MA, Luderer HF, Mathews B, Petroziello JM, Polyak K, Zawel L, Kinzler KW (1999) Analysis of human transcriptomes. Nat Genet 23:387–388

Chapter 4

Deregulation of the Wnt/β-Catenin/Tcf Signaling Pathway in Astrocytomas Gangadhara Reddy Sareddy and Phanithi Prakash Babu

Abstract Astrocytomas are the most common and deadliest primary brain tumors. Despite tremendous improvement in the understanding of molecular aspects of astrocytomas, these tumors have dismal prognosis. Astrocytomas develop as a result of stepwise accumulation of genetic alterations and the consequent disruption and augmentation of the apoptotic pathways and of survival signaling. Recent evidences suggest that neural stem cells and progenitor cells acts as source of brain tumors. The self-renewal and maintenance of neural progenitor cells is tightly regulated by the developmental pathways, particularly by Wnt signaling pathway. The aberrant operation of Wnt signaling may leads to the tumor development and its oncogenic role was evident in various human cancers. Recently, much interest is focused on the understanding of the role of Wnt signaling in astrocytomas and recent investigations provided the evidence that Wnt/βcatenin/Tcf signaling pathway is deregulated in astrocytomas and contributed to malignant progression. The extracellular inhibitors of Wnt signaling such as sFRPs and Dickkopf family proteins were downregulated and hypermethylated in astrocytomas. Several Wnt proteins and their cognate receptors were activated and their overexpression promotes the proliferation of glioma cell lines. β-Catenin is overexpressed in astrocytomas and progressively increased from low-grade to higher grades. Knockdown of β-catenin resulted in

P.P. Babu () Department of Biotechnology, School of Life Sciences, University of Hyderabad, Hyderabad 500 046, India e-mail: [email protected]

the inhibition of proliferation and induction of apoptosis in glioblastoma cell lines. β-Catenin transcriptional partners Lef1 and Tcf4 and their target genes were upregulated in astrocytomas and correlated with the histological grading of astrocytomas. In essence, Wnt/β-catenin/Tcf is abnormally activated in astrocytomas and serves as a candidate therapeutic target for astrocytomas. Keywords Deregulation · Wnt signaling pathway · Astro-cytomas · β-Catenin

Introduction Gliomas are the most common and devastating primary central nervous system neoplasms that accounts for 77% of brain tumors. The annual incidence of gliomas is approximately 5–10 cases per 100,000 in western population and is a leading cause of death among children and adults diagnosed with a neoplasia of the brain. Gliomas are subdivided into oligodendrogliomas, ependymomas, astrocytomas and oligoastrocytomas based on the resemblance of tumor cells to the original parental cells. According to World Health Organization (WHO) astrocytomas are further classified into four clinical grades which includes pilocytic astrocytomas (WHO grade I), diffuse astrocytomas (WHO grade II) anaplastic astrocytomas (WHO grade III) and glioblastoma multiforme (WHO grade IV). Glioblastoma multiforme (GBM) is the most frequent and most malignant glioma representing about 50% of all brain tumors. Despite optimal treatment the median survival of glioblastoma patients is up to 12–15 months, and 2–5 years for patients with

M.A. Hayat (ed.), Tumors of the Central Nervous System, Volume 5, DOI 10.1007/978-94-007-2019-0_4, © Springer Science+Business Media B.V. 2012

35

36

anaplastic astrocytomas, and those with diffuse astrocytoma can survive for as long as 10–15 years (Ohgaki and Kleihues, 2009; Louis et al., 2007). Pilocytic astrocytomas are well differentiated astrocytomas, most frequently occur in children and young adults. Most of the pilocytic astrocytomas are benign tumors and associated with a favorable outcome. Diffuse astrocytomas are well differentiated slowgrowing tumors, typically arise in young adults (∼34 years). These tumors tend to grow into surrounding normal brain structures and further transform into anaplastic astrocytoma and eventually to GBM. Characteristic genetic alterations of diffuse astrocytomas are point mutations in the TP53 tumor suppressor gene (50–80% of cases) and overexpression of PDGF-A and PDGFR-α (60%), but gene amplification was only detected in a small subset (<10%) of secondary GBM (Nupponen and Joensuu, 2006). Anaplastic astrocytoma arise de novo or from less malignant diffuse astrocytomas characterized by increased cellularity, nuclear atypica and mitotic activity. Males are frequently affected and the mean age is 41 years. Anaplastic astrocytoma (III) has a higher frequency of TP53 mutations, p16 and p19 deletion, RB alterations, and LOH on chromosome 19q (50%) (Nupponen and Joensuu, 2006). Glioblastomas are highly invasive and most malignant astrocytic tumors affecting mainly elder population. Histologically glioblastomas are heterogeneous in nature characterized by cellular polymorphism, nuclear atypia, mitotic activity, and contains the areas of microvascular proliferations, vascular thrombosis and necrosis (Kleihues and Ohgaki, 1999). Glioblastomas can be divided into two different subtypes based on their genetic and biologic differences. Majority of the glioblastomas are primary glioblastomas which develops rapidly without the evidence of a less malignant precursor lesion. Secondary glioblastomas develop slowly over a period of time by progression from less malignant precursor lesion. However, primary glioblastomas develop de novo with short clinical history usually less than 3 months and affect mostly elder population (∼55 years) and more common in men than females. The secondary glioblastomas occur in younger age group (∼39 years), shows a slightly more favorable outcome and develop far less often than primary glioblastomas. The time interval for progression from diffuse low-grade astrocytoma to secondary glioblastoma

G.R. Sareddy and P.P. Babu

varies considerably (∼4–5 years). Although, primary and secondary glioblastomas are morphologically and histologically similar, they likely involve different genetic alterations and signaling pathways (Kleihues and Ohgaki, 1999). Comparative gene expression analysis showed that primary and secondary glioblastomas have distinct expression profiles, however, half of the clinically determined primary glioblastomas has a similar pattern to the secondary glioblastomas. These suggest that primary glioblastomas may originate from a clinically undiagnosed lower grade lesion.

Molecular Pathogenesis of Glioblastomas Molecular and cytogenetic studies as well as highthroughput array-based assays suggest the distinct molecular features in primary and secondary glioblastomas. Chromosomal aberrations of gliomas have been studied using karyotyping, comparative genomic hybridization (CGH), or chromosome painting. The chromosomal aberrations identified in glioblastomas are scattered across the entire genome and affecting almost all the chromosomes. Although, primary and secondary glioblastomas are distinct from each other, these subtypes also shared chromosomal aberrations. The chromosomal aberrations associated with primary glioblastomas are amplifications and gains of 7p, 12q13–21, and chromosome 19. The main chromosomal regions showing losses are 10q, 9p, 13q, and 22q and may harbors additional aberrations such as losses of 18q, 16p, and 19q, and gains of 20q, and 12q. Diffuse astrocytomas associated with gains of 3q, 4q, 7q, 12p, and 19p, and losses of Xp, 1p, and 19q. Anaplastic astrocytomas associated with losses of 9p, 10q, 13q and gains of 1q, 7p, 11q, and Xq. The frequent genetic alteration of primary and secondary glioblastomas is the loss of heterozygosity (LOH) of 10q. The genetic hallmark of primary glioblastomas that typically lack a TP53 mutation is MDM2 amplification/overexpression (50%). EGFR amplification (40%) and/or overexpression (60%), CDKN2-A, CDKN2-B and PTEN mutations (30%), RB alteration and p16 deletion (30–40%) are also commonly observed in primary glioblastomas. LOH 10p and complete loss of entire chromosome 10 (50–80%) are exclusively present in primary glioblastomas. The

4

Deregulation of the Wnt/β-Catenin/Tcf Signaling Pathway in Astrocytomas

sequence in which gene alterations are acquired is not known since these neoplasms develop very rapidly, without a clinically or histopathologically identifiable precursor lesion. The TP53 mutations are the most commonly observed genetic alterations in secondary glioblastomas and major proportion carried form their low-grade diffuse astrocytomas and anaplastic astrocytomas. However, the TP53 mutations are less common in primary GBM (<10%). Most likely, the TP53 mutation is the initial gatekeeper lesion in astrocytic tumors, which then, through genetic instability undergoes malignant progression. The pathway to secondary glioblastoma is further characterized by LOH on chromosomes 19q and 10q. Recently it was pointed out that genomic alterations of LOH 1p and 19q, which are observed in the majority of oligodendrogliomas, may be observed in GBM. However, in contrast to oligodendrogliomas, in GBM loss of 19q is more likely to be partial than complete and loss of 1p is uncommon (∼10%). It was suggested that combined losses of chromosome arms 1p and 19q may indicate better prognosis and potential sensitivity to chemotherapy in GBM patients, while isolated loss of either 1p or 19q is of no prognostic significance. The IDH1 gene at 2q33 encodes isocitrate dehydrogenase 1 (IDH1) catalyzes the oxidative carboxylation of isocitrate to α-ketoglutarate, predominantly located in the cytosol. Recent studies demonstrated that IDH1 mutations are very early and frequent genetic alteration in low grade astrocytomas (80%), anaplastic astrocytomas and secondary glioblastomas. In contrast, IDH1 mutations are very rare (<5%) or absent in pilocytic astrocytomas and primary glioblastomas. Almost 60% of the low grade astrocytomas have both TP53 and IDH1 mutations (Ohgaki and Kleihues, 2009). The progression of astrocytomas to more malignant forms resulted from the stepwise accumulation of genetic alterations and the consequent disruption of apoptotic pathway and augmentation of survival signaling. These genetic alterations in astrocytomas ultimately resulted in the abnormal activation of signal transduction pathways or disruption of cell cycle arrest pathways. Amplification or activating mutations of EGFR, over-expression of FGF, FGFR, PDGF and PDGFR due to either gene amplification or other epigenetic mechanisms, all lead to constitutive activation of corresponding receptor tyrosine kinase signaling. Subsequently, a number of down-stream signal transduction pathways

37

are activated, including the PI3 kinase/AKT pathway, RAS/MAP kinase pathway, C-MYC pathway, protein kinase C pathway, STAT pathway and TGF-β pathway (Ohgaki and Kleihues, 2007).

Wnt Signaling Pathway During the development of nervous system neural precursor or progenitor cells (NPC) act as a source of various types of specialized cells in the brain. Several studies have suggested that these cells are able to selfrenew, a hall mark of stem cells. Wnt signaling is a candidate pathway in controlling neural stem cells selfrenewal and differentiation. Wnt signaling is required at several stages of central nervous system development (Caricasole et al., 2005). So, dramatic alterations in this pathway may leads to development of CNS malignancies, and its oncogenic role was well studied in medulloblastoma development. The Wnts comprise a large family of protein ligands that affect diverse biological processes such as embryonic induction, generation of cell polarity, and the specification of cell fate (Logan and Nusse, 2004), tissue homeostasis and cancer. The Wnt pathway activation in oncogenesis and its consequences has been extensively reviewed (Logan and Nusse, 2004; Reya and Clevers, 2005). The Wnt proteins (the name derived from mouse Int-1 and Drosophila wingless) are a family of secreted glycoproteins characterized by several conserved cysteine residues. To date 19 mammalian Wnt homologues are well characterized. A number of Wnt genes, including Wnt2, Wnt7b and Wnt 5a, have been associated with abnormal proliferation of human breast tissue and other tumors. Wnt10b and Wnt13 have been suggested to direct cell- growth regulation during development. To date, 10 human Frizzled receptors were characterized which are vary in length (537–706 amino acids) and all of them are seven pass transmembrane receptors with an extra cellular N-terminal domain and an intracellular C-terminal domain. The Wnt receptor complex that activates the canonical pathway contains two components: a member of the Frizzled (Fzd) family and either one of two single span transmembrane proteins, low density-lipoprotein receptor related proteins 5 and 6 (LRP5 and LRP6). Once bound by their cognate ligands, the Fzd/LRP co-receptor complex activates the canonical signaling pathway. In the

38

absence of Wnt, β-catenin the key molecule of the canonical signaling pathway is constantly degraded by the “destruction complex” which is composed of the scaffolding protein Axin, the tumor suppressor adenomatous polyposis coli (APC), casein kinase 1α (CK1α) and glycogen synthase kinase 3β (GSK3β). Within this complex β-catenin is sequentially phosphorylated by CK1α and GSK3β resulting in its recognition by ubiquitin ligases and subsequent degradation in proteasomes. This continuous degradation of β-catenin prevents its nuclear translocation, which allows the DNA binding Tcf/Lef proteins to interact with transcriptional co-repressors to block β-catenin target gene expression. The Wnt/β-catenin/Tcf pathway is activated when Wnt ligands binds to its their cognate specific receptor complex containing a Frizzled family member and LRP5 or LRP6. The formation of WntFzd-LRP complex triggers the recruitment of cytosolic protein dishevelled (Dvl). The formation of Dvl-Fzd complex and the phosphorylation of LRP by CK1γ facilitating relocation of Axin to the membrane and inactivation of the destruction box. These events allows the β-catenin stabilization and its nuclear translocation. In the nucleus β-catenin converts the Tcf/Lef proteins into potent transcriptional activators there by activation of its target genes (Moon et al., 2004).

Cellular Origin-Neural Stem Cells-Wnt Signaling Although the genetic alterations and signaling pathways involved in the initiation and progression of astrocytomas are well studied, the cellular origin of astrocytomas is unknown. The identification of cell of origin may improve the understanding of the behavior and pathology. Neural stem cells are immature progenitor cells with a long life span, multipotent and self-renewing, and have been isolated from the subventricular zone, hippocampus, the lining of the lateral ventricles, and the dentate gyrus of an adult (Sanai et al., 2005). These neural stem cells and glial progenitor cells commonly possess features associated with tumors of the CNS including a robust proliferative potential and diversity of a progeny. The tumor initiating cells in astrocytoma shares many similarities with normal neural stem cells such as expression of

G.R. Sareddy and P.P. Babu

CD133, and glioblastomas, in particular may be seen as neoplasms of neural progenitor cells. In addition neural stem cells are regulated by the same cellular pathways that are active in glial neoplasms and exhibiting the tumor characteristics such as high motility, association with blood vessels and white matter tracts, and activation of developmental signaling pathways. This suggest that the cell of origin to be a transformed neural stem cell. Recently it was reported that cell of origin for glioma may be a committed glial progenitor cell (Lindberg et al., 2009). Germinal zones in particular, subventricular zone have long been proposed as source of gliomas, suggested by the fact that most of the gliomas are either periventricular or contiguous with the subventricular zone and express the markers of progenitor cells. All forms of the neoplasms arise from alterations of critical cellular functions such as proliferation, apoptosis, and tissue invasion. The neural stem cells and progenitor cells in the brain are at a risk of malignant transformation, presumably because of the fact that most of the oncogenic and developmental pathways that responsible for tumor formation are critical for the functions like cell survival, self-renewal, proliferation and differentiation and neural stem cells exhibit least resistance in tumorigenesis, since they already have the ability to bypass apoptosis and senescence. The developmental pathways in particular Wnt signaling pathway critically regulate the self-renewal, proliferation and differentiation of neural stem cells and other progenitor cells in the brain (Kalani et al., 2008). Deregulation of or abnormal operation of this pathway potentially acts as source of brain tumors.

Activation of Wnt/β-Catenin/Tcf Signaling Pathway Components in Astrocytomas Although, the oncogenic role of Wnt/β-catenin/Tcf pathway is well established in colon, breast and other types of cancers, little is known about its role in astrocytomas. Recently, several investigations were carried out to understand the role of Wnt signaling in astrocytomas. These studies demonstrated that the components of canonical Wnt pathway plays significant role in the astrocytomas. Here, investigations on the possible role of Wnt signaling pathway components were described in detail.

4

Deregulation of the Wnt/β-Catenin/Tcf Signaling Pathway in Astrocytomas

Wnt Signaling Antagonists Wnt signaling pathway is regulated extracellularly by its antagonists. Extracellular Wnt antagonists are divided into two functional classes which includes the sFRP (secreted Frizzled Related Protein) class and the Dickkopf class. sFRP class includes sFRP family (sFRP1, sFRP2, sFRP3, sFRP4, and sFRP5), sizzled, sizzled2, crescent, WIF-1 (Wnt inhibitory factor-1) and cerebrus, which inhibits the Wnt signaling by direct binding to Wnt proteins. This binding impairs the ability of Wnt proteins to bind to the receptor complex. Several investigations demonstrated the role of Wnt signaling antagonists in astrocytoma malignancy. Roth et al. (2000) investigated the role of sFRPs in glioma cell growth and motility. sFRP-1 and sFRP-2 are produced by the majority of malignant glioma cell lines and ectopic expression of sFRPs increased clonogenecity and enhanced resistance to serum starvation. In contrast, sFRPs do not modulate glioma cell susceptibility to apoptosis induced by the cytotoxic cytokines and cytotoxic drugs. Further, sFRP-2 promoted the intracranial glioma tumors in nude mice. In contrast, overexpression of sFRPs inhibited the glioma cell motility in vitro. sFRP-mediated effects on glioma cells are accompanied by decrease in the expression and activity of matrix metalloproteinase-2 (MMP-2) and decreased tyrosine phosphorylation of β-catenin. This suggested that sFRPs promote survival under non-supportive conditions and inhibit the migration of glioma cells. WIF-1 generally thought to be as tumor suppressor plays essential role in astrocytomas. Yang et al. (2010) demonstrated that the expression levels of WIF1 were significantly decreased in human astrocytomas compared to normal brain tissues and negatively correlated with the histological malignancy astrocytomas. They also reported the hypermethylation of WIF-1 gene promoter and its association with reduced WIF-1 expression. However, no hypermethylation of WIF1 was observed in normal brain tissues. Gotze et al. (2010) studied the status of promoter hypermethylation of the SFRP1, SFRP2, SFRP4, SFRP5, DKK1, DKK3 and NKD1, NKD2 in different grades of glioma tissues. Hypermethylation of SFRP1, SFRP2 and NKD2 each occurred in more than 40% of the primary glioblastomas, while DKK1 hypermethylation was found in 50% of secondary glioblastomas. Further, treatment

39

of SFRP1-, SFRP5-, DKK1-, DKK3-, NKD1- and NKD2-hypermethylated glioma cells with demethylating agent 5-aza-2-deoxycytidine and with histone deacetylase inhibitor Trichostatin A resulted in increased expression of each gene. Further, SFRP1hypermethylated gliomas showed significantly lower expression of the respective transcripts when compared with unmethylated tumors. Foltz et al. (2010) reported that DKK1, SFRP1 and WIF-1 genes were epigenetically silenced in glioblastomas and their expression levels were reduced in glioblastoma samples compared to normal brain tissue samples. Expression of all the three genes was restored in T98G glioblastoma cells by treatment Trichostatin A, but only DKK1 expression was restored by treatment with the 5-azacytidine. In addition, overexpression of DKK1 significantly reduced colony formation and increased chemotherapy-induced apoptosis in T98G glioblastoma cells. However, ectopic expression of WIF1 and SFRP1 shows a relative lack of response. Chronic Wnt3a stimulation only partially reversed the growth suppression after DKK1 reexpression, whereas a specific inhibitor of the JNK pathway significantly reversed the effect of DKK1 reexpression on colony formation and apoptosis in T98G cells. This support the potential growth-suppressive function for epigenetically silenced DKK1 in GBM and suggests that DKK1 restoration could modulate Wnt signaling through both canonical and noncanonical pathways. These studies demonstrated an important role of epigenetic silencing of Wnt pathway inhibitor genes in astrocytic gliomas, in particular, in glioblastomas, with distinct patterns of hypermethylated genes distinguishing primary from secondary glioblastomas. The other class of Wnt antagonists, Dickkopf family comprises four members (Dkk-1 to Dkk-4) and a unique Dkk-3-related protein named Soggy (sgy). These molecules curtail the Wnt signaling by binding to co-receptors LRP5 and LRP6. The human Dkk-1 (hDkk-1) gene is a transcriptional target of the p53 tumor suppressor, encodes a powerful inhibitor of the Wnt signaling pathway which regulates the spatial patterning/morphogenesis of the mammalian central nervous system. The Dkk-1 expression was induced greatly following DNA damage and in response to other chemotherapeutic agents through p53 dependent mechanism. It was demonstrated that overexpression of DKK-1 augmented the Wnt2 dependent β-catenin expression and sensitizes the glioblastoma cell lines

40

to apoptosis in response to various chemotherapeutic agents that cause DNA alkylation and DNA damage (Shou et al., 2002). In another study, DKK1 expression was analyzed in a series of brain tumors for structural alterations in the entire coding sequence by single-strand conformation polymorphism and direct sequencing (Muller et al., 2005). Several sequence variants of DKK-1 but no obvious mutations were detected in these tumors. The role of REIC/Dkk-3 in glioma, which generally acts as tumor suppressor in several cancers was studied by Mizobuchi et al. (2008). The expression levels of Dkk-3 were found to be lower in human malignant glioma tissues compared to normal brain tissues and also Dkk-3 expression levels were lower in glioma cell lines compared to normal human astrocytes. Further, Dkk-3 levels were negatively correlated with the clinical grade of gliomas. Knockdown of Dkk-3 resulted in increased survival cell index compared to control cells and reciprocally, overexpression of Dkk-3 decreased the survival index in a time dependent manner. In contrast forced expression of Dkk-3 does not alter the survival cell index in normal astrocytes. Overexpression of Dkk-3 induces apoptosis in GBM cell lines through the activation of phosphor-JUN, caspase-9 and caspase-3 and also resulted in the reduction and degradation of β-catenin.

G.R. Sareddy and P.P. Babu

of cell migration and invasion. Howng et al. (2002) demonstrated that the mRNA expression of Wnt5a, Wnt10b and Wnt13 was overexpressed in most of the brain tumors, whereas Wnt1 was less expressed. Yu et al. (2007) showed that Wnt5a was upregulated in glioblastomas compared to low-grade tumors and normal brain samples and overexpression of Wnt5a in glioblastoma cell lines increased the proliferation. Knockdown of Wnt5a reduced the proliferation of glioblastoma cell lines and decreased the tumor growth in vivo. Pu et al. (2009) reported that the mRNA and protein expression levels of Wnt2, Wnt5a and Fzd2 mRNA were overexpressed in astrocytomas compared to normal brain tissues. Whereas, Wnt1 and Fzd5 were not expressed in gliomas and Wnt3, Wnt4, Wnt10b, Wnt13 expression were almost equally expressed in tumor and normal brain tissues. Further, Knockdown of Wnt2 in glioma cell lines significantly down regulated the expression of Wnt2, β-catenin as well as expression of Fzd2, p-GSK3β, cyclin D1, PI3K, and p-Akt with concomitant reduction in cell viability. Treatment of xenograft tumors with Wnt2 siRNA slowdown the tumor growth and decreased expression Wnt2, fzd2, β-catenin and p-GSK3β were observed. Zhang et al. (2006) demonstrated that the expression levels of Fzd9 were up regulated in in astrocytoma samples compared to normal brain samples.

Wnt Ligands and Frizzled Receptors Dishevelled and FRAT-1 Wnt proteins are secreted glycoproteins which are modified through post-translational modifications such as palmitoylation and N-linked glycosylation required for correct secretion and stability of Wnts respectively. Wnt ligands are categorized in to canonical and noncanonical based on the pathway activated by Wnts. Several Wnt proteins, including Wnt1, Wnt2, Wnt3a, Wnt5a, Wnt7a, Wnt7b, Wnt 10b and Wnt13 have been associated with tumor development. Kamino et al. (2011) reported that the overexpression of mRNA levels of Wnt-5a and -7b and frizzled-2, -6 and -7 in glioma cells and also demonstrated the high expression of Wnt-5a in human glioma tumor samples. The positivity of Wnt-5a expression was correlated with the clinical grade. In addition, knockdown of Wnt-5a expression suppressed migration, invasion and expression of matrix metalloproteinase-2 of glioma cells. Wnt-5a ligand supplementation resulted in stimulation

Dishevelled (Dvl) interacts with Axin and dissociates the destruction complex by relocating the Axin to the cytoplasmic tail of LRP and inhibits the GSK3β activity by interacting with FRAT1/GBP leading to the stabilization of β-catenin. Three variants of Dvl (Dvl-1, Dvl-2 and Dvl-3) proteins were identified and overexpression or constitutive activation of Dvl proteins promotes neoplastic transformation and its activation has been reported in several malignancies. We reported (Sareddy et al., 2009a) the mRNA levels of Dvl variants in astrocytoma tissues and observed that Dvl-3 was upregulated in tumor tissues compared to normal brain samples. The expression levels were progressively increased from low-grade to high-grade and correlated with the clinical grade. FRAT1 (frequently arranged in advanced T-cell lymphomas-1) was identified as a positive regulator of Wnt/β-catenin

4

Deregulation of the Wnt/β-Catenin/Tcf Signaling Pathway in Astrocytomas

pathway by inhibiting the GSK3β activity. In response to Wnt signaling activation Dvl recruits FRAT1 into the destruction complex which facilitates the dissociation of GSK3β from destruction complex, leading to the stabilization of β-catenin. Guo et al. (2010) reported that FRAT1 was overexpressed in human astrocytoma samples and showed the significant positive correlation with the pathological grading of astrocytomas. Further its immunoreactivity was positively correlated with the β-catenin immunoreactivity.

Axin-APC-GSK3β Axin was originally identified as an inhibitor of Wnt signaling and is central to the down regulation of β-catenin. Axin is a core component of destruction complex that binds directly to β-catenin, APC, GSK3β, CK1α and Dvl and facilitates the β-catenin phosphorylation by GSK3β and CK1α which resulted in the β-catenin degradation. Axin is often mutated in several human malignancies. LOH of AXIN1 was reported in 11.1% of brain tumors and majority proportion was distributed to glioblastomas (6.3%) (Nikuseva Martic et al., 2010). The relative expression of Axin was down regulated in 65.5% of brain tumors compared to normal brain tissues, of which majority of them are astrocytomas. This study also demonstrated that there was no difference in Axin protein levels in patients with AXIN1 LOH and patients without it. In contrast, relative β-catenin levels were significantly higher in patients with AXIN LOH in comparison to without it. Comparison of relative levels of β-catenin and AXIN1 revealed that they were significantly reversely proportional. APC is often mutated in several human malignancies. It was demonstrated that most of the colon and other human malignancies having highly constitutive β-catenin due to mutations in APC or β-catenin itself. Mutations in APC and β-catenin genes were noticed in the childhood brain tumor medulloblastoma but till to date no mutations were found in astrocytomas. However, recent studies reported that APC mutations were found in secondary brain metastasis rather than in primary tumors (Pecina-Slaus et al., 2010). In addition, lack of protein expression of APC and β-catenin nuclear localization was observed in these metastases. The mutation in APC is a silent mutation that might have consequences in the creation of a

41

new splice site. To date no prominent mutations were observed in β-catenin but it was reported that β-catenin is mutated at S33 as sporadic event in a brain metastasis (Lee et al., 2009). Glycogen synthase kinase 3 is a serine/threonine kinase that regulates diverse signaling pathways involved in proliferation, apoptosis and cell cycle control. There are 2 mammalian GSK3 isoforms include GSK3α, GSK3β which are functionally independent. Korur et al. (2009) analyzed the role of GSK3β in malignant gliomas and reported that mRNA and protein levels of GSK-3β were overexpressed in human glioblastoma tissues compared to normal brain tissues. Further, inhibition of GSK3β activity promoted tumor cell differentiation, tumor cell apoptosis and reduced the clonogenecity.

β-Catenin β-Catenin is the critical mediator of the canonical Wnt signaling pathway which is first identified as a member of Wnt signaling pathway. The β-catenin gene CTNNb1 was localized at 3p2.1. β-Catenin is available in two different pools: one part is a component of cell adhesion and other was in the cytoplasm or/and nucleus which is mainly involved in the regulation of Wnt pathway. The β-catenin gene is often mutated in a wide variety of human cancers. It was noticed that colorectal tumors that contain APC mutations harbors low β-catenin mutations where as in tumors lacking APC mutations in CTNNb1 was greatly increased. Dissociation of cytoplasmic inhibitory complex owing to the upstream activity of Wnt signals or mutations in APC and Axin leads to the cytosolic accumulation of β-catenin. The Oncogenic potential of β-catenin was extensively studied in wide variety of human cancers. We studied the expression levels of β-catenin mRNA and protein in 32 human astrocytoma samples and found that the mRNA and protein levels of β-catenin were elevated in astrocytomas samples compared to normal brain tissues and these levels were progressively increased form low-grade to higher grades and positively correlating with the histological malignancy of astrocytomas. Primarily its localization was observed in cytosol and nucleus which is crucial in mediating Wnt pathway activity (Sareddy et al., 2009a). Further, we also studied the role of β-catenin in ENU-induced glioma rat model. We observed that the

42

levels of β-catenin were increased as tumor grows and correlated with the progression of gliomas (Sareddy et al., 2009b). Pu et al. (2009) also demonstrated the β-catenin overexpression in astrocytomas and knockdown of β-catenin resulted in the downregulation of βcatenin, p-GSK3β, cyclin D1, PI3K, and pAKT. They also found that the reduced cell viability, cell cycle arrest at G0/G1 phase, increased apoptotic cell population and decreased invasion of glioma cell lines as a consequence of β-catenin knockdown. Also, β-catenin knockdown reduced and slowdown the tumor size and growth respectively in in vivo xenograft tumors. In these tumors β-catenin siRNA treatment causes reduction in the levels of β-catenin, pGSK3β, PI3K, Wnt2 and Fzd2 and induce tumor cell apoptosis. Similar results were also observed by Liu et al. (2010). They stated that the β-catenin mRNA and protein were elevated in astrocytomas compared to normal brain tissues and exhibited the positive correlation with the histological grading of astrocytomas. Zhang et al. (2009) studied the correlation between β-catenin distribution, tumor grade and patient 2 year survival. No correlation was observed between tumor grade and distribution of β-catenin. However, significant positive correlation was observed between levels of β-catenin and 2-year patient survival. The expression of β-catenin was not correlated with other clinicopathological characteristics such as tumor age, tumor size, and sex and tumor location. Further, survival analysis showed that patients with astrocytoma showing less expression of β-catenin tend to be associated with good prognosis, whereas, astrocytoma patients with high β-catenin expression associated with poor prognosis. It has also been demonstrated that β-catenin knockdown reduced the expression of EGFR, STAT3 and AKT1 and pAKT levels. In addition, reduction in the levels of cyclin D1, MMP2 and MMP9 which are downstream genes of the EGFR pathway was observed, implicating that the cross regulation of β-catenin pathway and EGFR pathway in the regulation of glioma cell proliferation and invasion (Yue et al., 2010).

Lef-1 and Tcf-4 Lef/Tcf transcription factors belong to high mobility group (HMG) domain proteins that recognize the same DNA consensus motif through HMG box

G.R. Sareddy and P.P. Babu

DNA-binding domain. Lef/Tcf family transcription factors consist of four members: Lef-1, Tcf-1, Tcf-3 and Tcf-4. Tcf-1 is predominantly expressed in cells of T cell lineage and Lef-1 is expressed in pre-B and T cells. Tcf-3 is expressed in somatic epithelium, keratinocytes of the skin and Tcf-4 is expressed in midbrain and in epithelium of intestine and mammary glands. In the absence of β-catenin, Tcf/Lef factors suppress the Wnt target gene expression by binding with the members of the Groucho (Grg/TLE) family of transcriptional co-repressors. β-Catenin translocation to the nucleus converts Tcf family proteins into potent transcriptional activators by displacing co repressors and recruiting an array of co-activator proteins (Barker and Clevers, 2006). β-Catenin does not have a DNA binding domain, but it has a potent transcription activation domain. In general, Lef/Tcf transcription factors do not have a strong transcription activation domain, but they have a good DNA binding/bending domain. Thus, when β-catenin binds to Lef/Tcf proteins, a potent transcription regulatory complex is formed. We examined the expression levels of Lef1 and Tcf4 in human astrocytic samples and found that compared to normal brain tissues Lef1 and Tcf-4 expression levels were significantly elevated in astrocytomas and these levels were correlating with the pathological grading of astrocytomas. We also demonstrated the interaction of β-catenin with Tcf4 in the nuclear fractions of GBM tissues and in cell lines. Further, the Lef-1 and tcf-4 levels were elevated in ENU-induced rat gliomas and protein levels of these molecules were progressively increased form initial stage to advanced stages. Zhang et al. (2011) also reported the elevation of Tcf4 levels in high-grade gliomas and inhibition of β-catenin/Tcf4 activity resulted in the inhibition of proliferation and invasion of glioma cells in vitro and in vivo. Oncogenic activity of Wnt signaling is mediated through overexpression of their target genes cyclin D1, c-Myc, c-Jun, MMP2, N-Myc etc. which play essential roles in the cell cycle progression, cell proliferation and cell survival. We observed the elevated levels of cyclin D1, c-Myc, c-jun, N-Myc in astrocytomas of different clinical grade in comparison with normal brain samples and their significant correlation with histological grading of astrocytomas. Wang et al. (2010) examined the pygopus 2 expression, which is coactivator of β-catenin in human astrocytoma samples. Pygopus 2 levels were overexpressed in astrocytoma samples compared to controls and exhibited a positive

4

Deregulation of the Wnt/β-Catenin/Tcf Signaling Pathway in Astrocytomas

correlation with tumor grade and knockdown of pygopus 2 resulted in inhibition of cell proliferation, cell cycle arrest, colony forming ability, BrdU incorporation and invasiveness of human and rat glioma cell lines. Knockdown of pygopus 2 also leads to the decreased expression of Wnt target gene cyclin D1 without altering the β-catenin levels and its nuclear translocation. Further, Chen et al. (2010) reported that overexpression of pygopus significantly enhances the cell proliferation and cell cycle progression from G1 to S and the elevation of cyclin D1 levels without altering the β-catenin levels. Furthermore, pygopus 2 showed the positive correlation with the expression of cyclin D1 in human glioma samples. In summary, several investigations that deal with the study of Wnt/β-catenin/Tcf signaling pathway in astrocytomas indicate that this pathway is aberrantly activated in human astrocytomas and involved in the progression leading to metastasis. This pathway attracted as a potential therapeutic target and inhibition of this pathway may inhibit the malignant progression of astrocytomas. Acknowledgements We acknowledge Department of Biotechnology (DBT), Department of Science & Technology (DST), Indian Council of Medical Research (ICMR), Council of Scientific & Industrial Research (CSIR), DST – Nano UoH project, DBT-CREB, Government of India, New Delhi, India for funding the lab.

References Barker N, Clevers H (2006) Mining the Wnt pathway for cancer therapeutics. Nat Rev Drug Discov 5:997–1014 Caricasole A, Bakker A, Copani A, Nicoletti F, Gaviraghi G, Terstappen GC (2005) Two sides of the same coin: Wnt signaling in neurodegeneration and neuro-oncology. Biosci Rep 25:309–327 Chen YY, Li BA, Wang HD, Liu XY, Shen SH, Zhu HW, Wang HD (2010) The role of pygopus 2 in rat glioma cell growth. Med Oncol 28:631–640 Foltz G, Yoon JG, Lee H, Ma L, Tian Q, Hood L, Madan A (2010) Epigenetic regulation of Wnt pathway antagonists in human glioblastoma multiforme. Genes Cancer 1:81–90 Gotze S, Wolter M, Reifenberger G, Muller O, Sievers S (2010) Frequent promoter hypermethylation of Wnt pathway inhibitor genes in malignant astrocytic gliomas. Int J Cancer 126:2584–2593 Guo G, Mao X, Wang P, Liu B, Zhang X, Jiang X, Zhong C, Huo J, Jin J, Zhuo Y (2010) The expression profile of FRAT1 in human gliomas. Brain Res 1320:152–158

43

Howng SL, Wu CH, Cheng TS, Sy WD, Lin PC, Wang C, Hong YR (2002) Differential expression of Wnt genes, betacatenin and E-cadherin in human brain tumors. Cancer Lett 183:95–101 Kalani MY, Cheshier SH, Cord BJ, Bababeygy SR, Vogel H, Weissman IL, Palmer TD, Nusse R (2008) Wnt-mediated self-renewal of neural stem/progenitor cells. Proc Natl Acad Sci USA 105:16970–16975 Kamino M, Kishida M, Kibe T, Ikoma K, Iijima M, Hirano H, Tokudome M, Chen L, Koriyama C, Yamada K, Arita K, Kishida S (2011) Wnt-5a signaling is correlated with infiltrative activity in human glioma by inducing cellular migration and MMP-2. Cancer Sci 102:540–548 Kleihues P, Ohgaki H (1999) Primary and secondary glioblastomas: from concept to clinical diagnosis. Neuro Oncol 1:44–51 Korur S, Huber RM, Sivasankaran B, Petrich M, Morin P Jr, Hemmings BA, Merlo A, Lino MM (2009) GSK3beta regulates differentiation and growth arrest in glioblastoma. PLoS One 4:e7443 Lee CI, Hsu MY, Chou CH, Wang C, Lo YS, Loh JK, Howng SL, Hong YR (2009) CTNNB1 (beta-catenin) mutation is rare in brain tumours but involved as a sporadic event in a brain metastasis. Acta Neurochir 151:1107–1111 Lindberg N, Kastemar M, Olofsson T, Smits A, Uhrbom L (2009) Oligodendrocyte progenitor cells can act as cell of origin for experimental glioma. Oncogene 28:2266–2275 Liu X, Wang L, Zhao S, Ji X, Luo Y, Ling F (2010) betaCatenin overexpression in malignant glioma and its role in proliferation and apotosis in glioblastoma cells. Med Oncol 16:75–79 Logan CY, Nusse R (2004) The Wnt signaling pathway in development and disease. Annu Rev Cell Dev Biol 20:781–810 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 Mizobuchi Y, Matsuzaki K, Kuwayama K, Kitazato K, Mure H, Kageji T, Nagahiro S (2008) REIC/Dkk-3 induces cell death in human malignant glioma. Neuro Oncol 10:244–253 Moon RT, Kohn AD, De Ferrari GV, Kaykas A (2004) WNT and beta-catenin signalling: diseases and therapies. Nat Rev Genet 5:691–701 Muller W, Lass U, Wellmann S, Kunitz F, von Deimling A (2005) Mutation analysis of DKK1 and in vivo evidence of predominant p53-independent DKK1 function in gliomas. Acta Neuropathol 109:314–320 Nikuseva Martic T, Pecina-Slaus N, Kusec V, Kokotovic T, Musinovic H, Tomas D, Zeljko M (2010) Changes of AXIN1 and beta-catenin in neuroepithelial brain tumors. Pathol Oncol Res 16:75–79 Nupponen NN, Joensuu H (2006) Molecular pathology of gliomas. Curr Diagn Pathol 12:394–402 Ohgaki H, Kleihues P (2007) Genetic pathways to primary and secondary glioblastoma. Am J Pathol 170:1445–1453 Ohgaki H, Kleihues P (2009) Genetic alterations and signaling pathways in the evolution of gliomas. Cancer Sci 100:2235– 2241 Pecina-Slaus N, Majic Z, Musani V, Zeljko M, Cupic H (2010) Report on mutation in exon 15 of the APC gene in a case of brain metastasis. J Neurooncol 97:143–148

44 Pu P, Zhang Z, Kang C, Jiang R, Jia Z, Wang G, Jiang H (2009) Downregulation of Wnt2 and beta-catenin by siRNA suppresses malignant glioma cell growth. Cancer Gene Ther 16:351–361 Reya T, Clevers H (2005) Wnt signaling in stem cells and cancer. Nature 434:843–850 Roth W, Wild-Bose C, Platten M, Grimmel C, Melkonyan HS, Dichgans J, Weller M (2000) Secreted Frizzled-related proteins inhibit motility and promote growth of human malignant glioma cells. Oncogene 19:4210–4220 Sanai N, Alvarez-Buylla A, Berger MS (2005) Neural stem cells and the origin of gliomas. N Engl J Med 353:811–822 Sareddy GR, Panigrahi M, Challa S, Mahadevan A, Babu PP (2009a) Activation of Wnt/beta-catenin/Tcf signaling pathway in human astrocytomas. Neurochem Int 55: 307–317 Sareddy GR, Challa S, Panigrahi M, Babu PP (2009b) Wnt/betacatenin/Tcf signaling pathway activation in malignant progression of rat gliomas induced by transplacental N-ethyl-Nnitrosourea exposure. Neurochem Res 34:1278–1288 Shou J, Ali-Osman F, Muttani AS, Pathak S, Fedi P, Srivenugopal KS (2002) Human Dkk-1, a gene encoding a Wnt antagonist, responds to DNA damage and its overexpression sensitizes brain tumor cells to apoptosis following alkylation damage of DNA. Oncogene 21:878–889 Wang ZX, Chen YY, Li BA, Tan GW, Liu XY, Shen SH, Zhu HW, Wang HD (2010) Decreased pygopus 2 expression

G.R. Sareddy and P.P. Babu suppresses glioblastoma U251 cell growth. J Neurooncol 100:31–41 Yang Z, Wang Y, Fang J, Chen F, Liu J, Wu J, Wang Y (2010) Expression and aberrant promoter mehtylation of Wnt inhibitory factor-1 in human astrocytomas. J Exp Clin Cancer Res 29:26 Yu JM, Jun ES, Jung JS, Suh SY, Han JY, Kim JY, Kim KW, Jung JS (2007) Role of Wnt5a in the proliferation of human glioblastoma cell lines. Cancer Lett 257:172–181 Yue X, Lan F, Yang W, Yang Y, Han L, Zhang A, Liu J, Zeng H, Jiang T, Pu P, Kang C (2010) Interruption of β-catenin suppresses the EGFR pathway by blocking multiple oncogenin targets in human glioma cells. Brain Res 1366: 27–37 Zhang Z, Schittenhelm J, Guo K, Buhring HJ, Trautmann K, Meyermann R, Schluesener HJ (2006) Upregulation of frizzled 9 in astrocytomas. Neuropathol Appl Neurobiol 32:615–624 Zhang LY, Jiang LN, Li FF, Li H, Liu F, Gu Y, Song Y, Zhang F, Ye J, Li Q (2009) Reduced beta-catenin expression is associated with good prognosis in astrocytoma. Pathol Oncol Res 16:253–257 Zhang J, Huang K, Shi Z, Zou J, Wang Y, Jia Z, Zhang A, Han L, Yue X, Liu N, Jiang T, You Y, Pu P, Kang C (2011) High β-catenin/Tcf-4 activity confers glioma progression via direct regulation of AKT2 gene expression. Neuro Oncol 13:600–609

Chapter 5

Subependymal Giant Cell Astrocytoma: Role of mTOR Pathway and Its Inhibitors Katarzyna Kotulska and Sergiusz Jó´zwiak

Abstract Subependymal giant cell astrocytoma is a distinct brain tumor growing near the foramen of Monro. Although histologically benign, these tumors may produce hydrocephalus and are associated with significant morbidity and mortality. Until recently, the only recommended treatment for these tumors was surgical resection. Subependymal giant cell astrocytoma is almost exclusively associated with Tuberous Sclerosis Complex. Tuberous Sclerosis Complex is a relatively frequent neurocutaneous disease, caused by the mutation in either of two genes: TSC1 or TSC2. Both genes act as tumor suppressors, and their products were shown to inhibit mTOR pathway. Recently, the recognition of the role of mTOR pathway in the pathogenesis of Tuberous Sclerosis Complex and subependymal giant cell astrocytoma led to clinical trials of mTOR inhibitors. Everolimus was shown to reduce the size of subependymal giant cell astrocytoma and is currently approved by FDA to treat patients with SEGA associated with TSC, who cannot be treated with surgery. Keywords SEGA · TSC · mTOR · GTPase-activating protein · Hypoxia-inducible factors

K. Kotulska () Klinika Neurologii I Epileptologii, Instytut Pomnik “Centrum Zdrowia Dziecka”, 04-730 Warszawa, Poland e-mail: [email protected]

Introduction Subependymal giant cell astrocytoma (SEGA) is a rare low grade brain tumor occurring almost exclusively in the settings of tuberous sclerosis complex (TSC) (Grajkowska et al., 2010). Tuberous sclerosis complex is an autosomal dominant neurocutaneous disorder characterized by benign, highly vascular, hamartomatic growth developing in various tissues and organs, including the brain, kidneys, heart, liver, lungs, retina, and skin. Its prevalence approaches 1 in 6,000 live births (Curatolo et al., 2008). The incidence of SEGA in TSC varies from 5 to 20% (Shepherd et al., 1991; Curatolo et al., 2008). In children and adolescents with TSC, SEGA present the major cause of morbidity and mortality. Although histologically benign, these neoplasms tend to grow and may obstruct cerebrospinal fluid pathways, causing hydrocephalus (Kim et al., 2001). Surgical resection is the only currently recommended treatment for SEGA producing the clinical symptoms. However, by the time the symptoms are recognized they are often irreversible even by immediate surgical intervention. Moreover, SEGA surgery is associated with distinct morbidity and mortality up to 30% in different pediatric patients series (de Ribaupierre et al., 2007). Recently, the discoveries of mammalian Target Of Rapamycin (mTOR) pathway upregulation in TSCassociated tumors have resulted in promising therapeutic trials on pharmacological treatment in TSC (Crino, 2008). Rapamycin, known also as sirolimus, and its derivate, everolimus, appeared to be effective in many TSC-related tumors, including SEGA (Krueger et al., 2010; Lam et al., 2009). In 2010, the US Food and Drug Administration (FDA) approved everolimus to

M.A. Hayat (ed.), Tumors of the Central Nervous System, Volume 5, DOI 10.1007/978-94-007-2019-0_5, © Springer Science+Business Media B.V. 2012

45

46

treat patients with SEGA associated with TSC, who cannot be treated with surgery.

mTOR Pathway in Tuberous Sclerosis Complex Tuberous sclerosis complex is caused by inactivating mutations in either of two genes: the TSC1 or TSC2 gene (Curatolo et al., 2008). TSC1 is located on 9q34 and TSC2 on 16p13 (Kwiatkowski and Manning, 2005). More than 1500 different mutations have been described for both genes. TSC1 mutations account for 26% of all TSC mutations and TSC2 mutations account for 74%. Mutational analysis performed by means of several techniques reaches up to 90% mutation detection rate in TSC patients. In about 10–15% of TSC patients, no mutation is identified (Dabora et al., 2001). This might be caused by mosaicism for mutations in TSC1 or TSC2 genes, or by detection failure, or, less likely, by the existence of a third “TSC gene”. Familial cases with at least two patients with TSC diagnosis, represent about one third of all TSC cases. Two thirds of TSC patients are recognized as sporadic. In familial cases, TSC1 and TSC2 mutation are identified with the same frequency. In sporadic patients, TSC2 mutations predominate, accounting for about 80% of cases (Dabora et al., 2001). TSC1 gene encodes the protein called hamartin and TSC2 genes encodes tuberin. Both proteins form an intracellular complex, acting as a GTPase-activating protein for a small G protein, Ras-homoloque-enriched in brain (Rheb). It is established that within the TSC1/TSC2 heterodimer, hamartin stabilizes tuberin and tuberin, via GTPase-activating protein (GAP) domain at its C-terminus, exerts GTPase activity towards Rheb (Kwiatkowski, 2003; Zhang et al., 2003). Rheb is the direct target of the TSC1/TSC2 complex, leading to the inactivation of mTOR (Kwiatkowski and Manning, 2005). Rheb, while being in its active GTP-bound state, stimulates mTOR, and promotes protein synthesis and cell growth through phosphorylation of two downstream target proteins, S6K1 and 4E-BP1 (eukaryotic translation initiation factor 4E binding protein 1). Once active, S6K1 activates ribosomal subunit protein S6, leading to ribosome recruitment and protein translation. 4E-BP1 represses activity of eIF4E (eukaryotic translation initiation factor 4E) and, when phosphorylated by

K. Kotulska and S. Jó´zwiak

mTOR, releases eIF4E from its control, allowing ribosome recruitment preceding translation initiation (Kwiatkowski, 2003). mTOR is a 289-kDa, ubiquitously expressed protein which regulates cell growth, proliferation and survival. It is a serine-threonine kinase receiving inputs from many various signaling pathways to activate ribosome biogenesis and cap-dependent translation. It is well established that mTOR is sensitive to the presence of nutrients, growth factors, cytokines, availability of cellular energy, and conditions of cellular stress (Fingar et al., 2002). mTOR can be found in two functionally and structurally different complexes – mTORC1 and mTORC2 (Loewith et al., 2002). The assembly mTORC1 is rapamycin sensitive and contains, besides mTOR, raptor, and mLST8. Some reports indicate that this complex contains also PRAS40, but other studies suggest that PRAS40 is a downstream substrate of mTORC1. Contrary to mTORC1, the complex mTORC2 is relatively insensitive to rapamycin. It contains mTOR, rictor, SIN1, mLST8, and PRP5. The main role of mTORC1 is to phosphorylate S6K, which phosphorylates S6 to induce ribosomal translation. However, mTORC1 has presumably many other targets influencing cell survival and proliferation. In TSC, loss of tuberin/hamartin complex function is associated with high activity of mTORC1 (Chan et al., 2004; Crino, 2008). As a consequence, in TSC-associated lesions, the phosphorylation of S6K on T389 is elevated and unregulated. Activated S6K together with inhibited 4E-BP1 allows translation, and it is important to note that activated mTORC1 particularly potentiates the translation of translation factors and ribosomal proteins, including Hypoxia-inducible Factors (HIFs). HIFs facilitate both oxygen delivery and adaptation to oxygen deprivation by regulating the expression of genes that are involved in many cellular processes, including glucose uptake and metabolism, angiogenesis, erythropoiesis, cell proliferation, and apoptosis. The genes regulated by HIFs include: angiopoetin-1, angiopoetin-2, Vascular Endothelial Growth Factor (VEGF), erythropoietin, Glucose Transporter-1 (GLUT-1), Transforming Growth Factor alpha (TGFalpha), MatrixMetalloproteinase-2, Matrix Metalloproteinase-9, D-type cyclins, and many others. These pathways contribute to the involvement of mTOR in the production of pro-angiogenic factors and of endothelial cell growth and proliferation (Crino, 2008).

5

Subependymal Giant Cell Astrocytoma: Role of mTOR Pathway and Its Inhibitors

The role of mTORC2 (mTOR/rictor complex) is less known. It is established that mTORC2 phosphorylates and activates serine/ threonine kinase Akt in the hydrophobic Ser473 site. Akt, a downstream effector of PI3K (phosphoinositide-3-OH kinase) is known to activate mTORC1-dependent translation. It activates also protein kinase C, serum- and glucocorticoidinduced protein kinase-1. TORC2 was originally known as rapamycin insensitive, but recent studies indicate that it is also targeted by rapamycin (Loewith et al., 2002). Rapamycin or its analog, everolimus disrupt rictor/mTOR association and suppress Akt signaling. On the other hand, mTORC2 kinase activity towards Akt was found to be significantly attenuated in cells lacking TSC1 or TSC2. This effect does not depend on GAP activity of tuberin and its impact on Rheb. These findings indicate the complex mechanisms of mTORc1 and mTORC2 interactions with tuberin/hamartin complex and PI3K/Akt signaling (Rosner and Hengstschlaeger, 2008). Another significant misregulation caused by TSC1/TSC2 loss is the inhibition of autophagy (Jung et al., 2010). Autophagy is regulated both via mTORC1 and mTORC2. The autophagy induction by mTORC2 inhibition is mediated mainly by FoxO3, a transcription factor downstream of Akt, that is involved in autophagy gene expression. mTOR pathway regulates also cell survival and apoptosis (Freilinger et al., 2006). S6K was shown to phosphorylate the pro-apoptotic protein Bad. Phosphorylated Bad is then inactive and anti-apoptotic proteins dominate to increase cell survival. On the other hand, mTOR may induce pro-apoptotic protein Bax translation, thus activating the programmed cell death. Freilinger et al. (2006) showed that tuberin phosphorylates proapoptotic protein Bad on Ser136, activating it to interact with antiapoptotic proteins Bcl-2 and Bcl-xl. As a result, apoptosis is triggered. An inactivating mutation of TSC2, as seen in TSC, attenuates the negative effect of tuberin on cell survival. Marked decrease in tuberin’s potential to activate Bad protein was observed also as a result of Akt phosphorylation. TSC2-Rheb-mTOR pathway was also shown to control axon guidance and growth in central nervous system. As shown by Choi et al. (2008), overexpression of Tsc1/Tsc2 suppresses axon formation. On the other hand, Tsc1 or Tsc2 deficiency in hippocampal neurons in mice results in the formation of ectopic neurons, associated with an increase in SAD kinase

47

level. This kinase is strongly involved in regulation of neuronal polarity and was shown to be overexpressed in cortical tubers in TSC patients.

Subependymal Giant Cell Astrocytoma in Tuberous Sclerosis Complex Subependymal giant cell astrocytomas are the most common brain tumors occurring in about 5–20% of the TSC patients (Grajkowska et al., 2010). They usually develop in the first two decades of life, and can be found even in fetuses and newborns. SEGAs represent 1–2% of all pediatric brain tumors. They can be revealed on neuroimaging as tumors located on the surface of lateral, or, rarely, third ventricle. The progressive growth of gadolinium-enhancing lesions at the foramen of Monro in a TSC patient strongly suggests SEGA (Torres et al., 1998). No spontaneous regression has been reported. SEGAs generally exceed 1 cm in diameter but can reach a greater size, even exceeding 10 cm. SEGAs usually extend into lateral ventricle and can obstruct the foramen of Monro and flow of CSF, causing hydrocephalus. SEGA usually grow slowly, but significant increase in tumor size can be observed in 1–2-year long follow-up. It was shown that in patients with TSC2 mutation SEGA develop significantly more frequently and grow more rapidly in comparison to TSC1 mutation. Therefore, neuroimaging study is recommended in individuals with TSC2 mutation every 2 years and every 3 years in patients with TSC1 mutations (Jó´zwiak et al., 2000). SEGA surgery is associated with significant risk of morbidity and mortality. In addition, SEGAs typically do not respond to radiation therapy or chemotherapy (Franz et al., 2006). Macroscopically, SEGAs are gray to pinkish red well-circumscribed tumors, and sometimes develop massive haemorrhages and calcifications. Microscopically, SEGAs consist of a mixed cell population with dysmorphic glial cells and giant cells. The glial cells are polygonal, epithelioid, gemistocytic or spindle shaped with marked pleomorphism, dense eosinophilic cytoplasm, eccentric nuclei, prominent nucleoli, and nuclear inclusions. They are arranged in sheets, clusters, or pseudorosettes. The giant cells contain regular nuclei and abundant, eosinophilic cytoplasm (Fig. 5.1.). SEGAs are rich in vascular stroma and may contain calcifications, as well as foci of necrosis. Some mitotic figures can also be found.

48

Fig. 5.1 Subependymal giant cell astrocytoma. Pleomorphic multinucleated tumor cells with abundant eosinophilic cytoplasm. HE staining. Courtesy Dr. Wieslawa Grajkowska, Department of Pathology, The Children’s Memorial Health Institute, Warsaw, Poland

SEGAs show low labeling index of Ki67 about 2.9%. Immunohistochemical studies have demonstrated that SEGAs are positive for GFAP, neurofilament, S-100, neuron-specific enolase, and synaptophysin proteins suggesting the dual neuronal and glial origin. Hence, the name of subependymal giant cell tumor (SEGT) is suggested (Grajkowska et al., 2010). SEGAs express also some cellular markers characteristic for progenitor cells derived from subventricular zone, like doublecortin or collapsing response mediator protein-4 (CRMP-4). Therefore, it has been suggested that giant cells and dysmorphic neurons in SEGA derive from the subventricular zone (Kim et al., 2001). In contrast to other TSC-associated lesions, like renal angiomyolipoma and pulmonary lymphangioleiomyomatosis, SEGAs do not stain with HMB45. There was minimal expression of p53 and high expression of Bax together with low Bcl-2 proteins in these tumors (Kim et al., 2001).

mTOR Pathway in Subependymal Giant Cell Astrocytoma There is a limited set of data on the expression of mTOR pathway in SEGAs. Hamartin was found to be lost in all nine tumors, including those obtained

K. Kotulska and S. Jó´zwiak

from patients with TSC1 and TSC2 mutations, analyzed in the study of Jó´zwiak et al. (2004). In this work, tuberin was lost or weakly expressed regardless the mutation found in the patient. Variable and irrespective of mutational status, reduction of tuberin and hamartin expression was found in the study of Johnson et al. (2002). Phosphorylation of Akt was elevated in half of SEGA samples studies by Jó´zwiak et al. (2007). In the same study, SEGAs have been shown to be highly immunoreactive for the phosphorylated isoforms of Mek1/2 and Erk1/2, suggesting abnormally hyperactive MAP kinase signaling. The level of phosphorylated S6K was increased in all SEGA samples in this study and in primary SEGA cell cultures, as reported by Tyburczy et al. (2010). Tumorigenesis in TSC is believed to fulfill Knudson’s two-hit model of tumor development. In keeping with this model, the tumor formation requires the inactivation of both alleles of either TSC1 or TSC2 genes (Kwiatkowski, 2003; Chan et al., 2004; Kwiatkowski and Manning, 2005). Heterozygosity at either locus is not sufficient to tumor development. A germline mutation, which can be either spontaneous or inherited, inactivates one allele of TSC1 or TSC2 gene. A so-called “second hit”, somatic mutation affecting heterozygous loci is referred to as loss of heterozygosity (LOH). In TSC, LOH is commonly found in kidney angiomyolipomas or cardiac rhabdomyomas, but rarely in brain lesions, including SEGAs. Henske et al. (1997) examined 11 SEGAs and found LOH on 16p13 in only one tumor. They did not find any case of LOH on 9q34. In the other study of Henske et al. (1996) LOH on 16p13 was found in one out of five SEGAs. No LOH on 9q34 was detected. Loss of heterozygosity was much more frequently fund in the study of Chan et al. who analyzed six SEGA samples and found biallelic inactivation of either TSC1 or TSC gene in five of them. Two samples showed LOH on 9q34 and three showed LOH on 16p13. On the other hand, SEGA seem to develop only on the basis of TSC1/TSC2 disruption. Ichikawa et al. (2005) reported a case of young woman with SEGA and no other clinical signs and symptoms of TSC. Molecular analysis of the tumor revealed biallelic inactivation of TSC2. However, TSC2 mutation was not found in the patient’s peripheral blood, buccal mucosa, urinary sediment nail, and hair.

5

Subependymal Giant Cell Astrocytoma: Role of mTOR Pathway and Its Inhibitors

Genes Regulating mTOR Activity in Subependymal Giant Cell Astrocytoma As LOH is not a common finding in SEGA, other possibilities of TSC1/TSC2 complex inactivation have been mooted to explain brain lesion pathogenesis in TSC. Many reports indicate that post-translational inactivation of TSC1/TSC2 complex arise through Akt kinase activation. Inoki et al. (2002) showed that Akt directly phosphorylates tuberin and this phosphorylation inactivates TSC2. This mechanism is likely responsible for the promotion of S6K activation by Akt. The activation of Akt was found in may TSCassociated tumors, but the results concerning Akt activation in SEGA are conflicting. The other mechanism of tumorigenesis in TSC is inactivation of the hamatrin-tuberin complex after phosphorylation by various kinases, such as extracellular signal-regulated kinase (ERK), which was found in subependymal giant cell astrocytomas and cardiac rhabdomyomas (Jó´zwiak et al., 2007). TSC1-TSC2 complex stability may be also influenced by death associated protein kinase (DAPK). DAPK functions in a wide range of biological pathways, including TNF-regulated cell death, stressinduced apoptosis, and autophagy (Lin et al., 2011). DAPK can mediate an inactivating phosphorylation of TSC2 and reduce its bioactivity after growth factor signaling. It has been shown in cell cultures that DAPK acts as a positive mediator of the mTORC1 translation pathway in response to growth factor stimulation. Recently, gene expression profiling of SEGAs showed other genes possibly involved in these tumors development. Tyburczy et al. (2010) found 50 genes either up-or down-regulated in SEGAs as compared to control brain tissue. Fourteen of these genes are known to be involved in tumorigenesis and 23 are known to play role in nervous system development. The latter were all down-regulated in SEGAs. Some genes implicated in tumorigenesis and nervous system development, like ANXA1 (annexin 1), GPNMB (glycoprotein nmb), LTF (lactotransferrin), RND3 (Rho family GTPase 3), S100A11 (S100 calcium binding protein a11), SFRP4 (secreted frizzled-related protein4), and NPTX1 (neuronal pentraxin I) are likely to be mTOR effector genes in SEGA, as their expression was modulated by an mTOR inhibitor, rapamycin, in SEGA-derived cells. Primary SEGA cell cultures were treated with rapamycin at concentration resulting

49

in inhibition of mTOR kinase activity and reduction in S6K and S6 phosphorylation. The levels of phosphorylation of these kinases was not influenced by rapamycin in control cultures of normal human astrocytes. Rapamycin treatment caused marked decrease in mRNA level of six genes highly up-regulated in SEGA (ANXA1, GPNMB, LTF, RND3, S100A11, and SFRP4) and significant increase in mRNA level of one gene down-regulated in SEGA (NPTX1).

mTOR Inhibitors Sirolimus and its congeners: everolimus and temsirolimus, are the known mTOR inhibitors already used in clinical practice (Gibbons et al., 2009). Sirolimus is a macrolide antibiotic first discovered as a product of the bacterium Streptomyces hygroscopicus in a soil from Rapa Nui island and hence the other name of this drug – rapamycin. Sirolimus is a powerful anti-proliferative and immunosuppressant agent, widely used in patients with renal transplants and in drug-eluting stents implanted into coronary arteries for prevention of restenosis. Sirolimus binds to the cytosolic protein FK-binding protein 12 (FK506-Binding Protein 12) and the sirolimus/FKBP12 complex binds with very high affinity to mTORC1, inhibiting its signaling capacity (Fig. 5.2.). The binding of sirolimus/FKBP12 complex to mTOR affects the phosphorylation of S6K1 and 4EBP1 (Loewith et al., 2002). Disruption of S6K1 and 4E-BP1 function, as a consequence of mTORC1 inhibition, influences the translation of mRNAs encoding many proteins crucial for cell cycle regulation. This inhibits tumor growth, reducing cellular protein synthesis and hampering cell progression from G1 to S phase. Dephosphorylation of S6K1 is also associated with reduced expression of factors involved in angiogenesis (mainly VEGF) and activation of proapoptotic protein Bad. Therefore, sirolimus inhibits not only cell proliferation, but also acts also as an inhibitor of cell survival. Moreover, as showed by Guba et al. (2005) it exerts strong antiangiogenic effects, reducing production of VEGF and inhibiting the response of vascular endothelial cells to stimulation by VEGF. It is important to note, that, as shown by Ruegg et al. (2007) rapamycin is not toxic for normal cultured hippocampal neurons. There were no effects of

50

K. Kotulska and S. Jó´zwiak Akt TSC1

Akt TSC2

TSC1

Erk

GTP

GDP Rheb

4eBP1

Sirolimus/Everolimus

elF4B

GDP Rheb

FKBP12

S6K1

elF4G

Erk

GTP

Rheb

mTORC1

TSC2

mTOR

4eBP1

S6K1 elF4G elF4G

S6

scaffolding scaffolding protein protein

scaffolding protein

Rheb

elF4B

S6

elF4E

elF4E

Translation

Proliferation Cell growth

Translation

Proliferation

Fig. 5.2 Rapamycin and its analogs bind to the immunophilin FK506-binding protein-12 (FKBP12). The complex of rapamycin and FKBP12 binds to mammalian Target of Rapamycin (mTOR), inhibiting its kinase activity and inhibiting the phosphorylation of the two main downstream mTOR targets: S6K1 and 4E-BP1 (eukaryotic translation initiation factor 4E binding protein 1). 4E-BP1 represses activity of eIF4E (eukaryotic translation initiation factor 4E) and, when phosphorylated by mTOR, releases eIF4E from its control, allowing

ribosome recruitment preceding translation initiation. mTOR is activated by a number of different factors. The tuberin (TSC2) and hamartin (TSC1) complex inhibits mTOR activity exerting GTPase activity towards Rheb (Ras-homoloque-enriched in brain). The inactivation of TSC1/TSC2 complex by a mutation or activation of ERK kinase or PI3K/AKT decreases the GAP activity toward Rheb allowing mTOR to increase ribosome biogenesis and mRNA translation

rapamycin on cell size or dendrite length, as well as neuronal gene transcription in vitro in this particular paradigm. Moreover, rapamycin did not significantly change voltage-gated sodium and potassium currents in vitro. Although rapamycin does not bind to preformed mTORC2, it was recently showed to suppress its assembly and function. Rosner and Hengstschlaeger (2008) reported that long-term rapamycin treatment results inhibition of mTORC2 assembly. This effect is mediated by complete dephosphorylation and cytoplasmic translocation of nuclear rector and sin1. In contrast, short treatment with rapamycin did not affect mTORC2 signaling. Everolimus (RAD001) is a selective mTOR inhibitor, specifically targeting the mTOR-raptor signal transduction complex (mTORC1). Everolimus, similarly to sirolimus, binds to the intracellular receptor protein FKBP12. Long-term treatment with everolimus was, similarly to sirolimus, shown to disrupt mTORC2 assembly and function (Gibbons et al.,

2009). It is orally bioavailable and was under clinical trials in renal cell cancer and endometrial cancer. In combination with imatinib mesylate it was studied in gastrointestinal stromal tumors. There are also studies on everolimus in combination with erlotinib in metastatic breast cancer. Everolimus is also used in drug-eluting coronary stents used for restenosis prevention. Moreover, it has shown a potential to reduce the risk of chronic allograft vasculopathy after heart transplantation. Anti-angiogenic effect of everolimus was evidenced both in vitro and in vivo. Everolimus reduced the proliferation of the cultured human umbilical vein endothelial cells, acting against VEGF signaling. In vivo, everolimus selectively inhibited VEGFdependent angiogenic response in mice with primary and metastatic tumors, causing a significant reduction in blood vessel density when compared to controls (Lane et al., 2009). Temsirolimus (CCI-779) is a derivate of sirolimus. The mode of action of temsirolimus is the same

5

Subependymal Giant Cell Astrocytoma: Role of mTOR Pathway and Its Inhibitors

as described for sirolimus. Long-term treatment with temsirolimus was, similarly to sirolimus, shown to disrupt mTORC2 assembly and function (Gibbons et al., 2009). Temsirolimus was also shown to arrest the cell cycle in G1 phase and to inhibit angiogenesis. Similarly to sirolimus and everolimus, temsirolimus inhibits VEGF production in the tumor cells. It is approved both in US and EU for the treatment of renal cell carcinoma and mantle cell lymphoma. An important aspect of the antitumor effect of rapamycin and its congeners, everolimus and temsirolimus, is their potential to act on both tumor cells directly, by inhibiting growth and indirectly, by inhibiting angiogenesis.

mTOR Inhibitors in Tuberous Sclerosis: Lessons from Preclinical Studies Sirolimus, temsirolimus, and everolimus were shown to inhibit the proliferation of a range of human tumor cell lines in vitro, including lines originating from lung, breast, prostate, colon, melanoma and glioblastoma (Gibbons et al., 2009). Given the role of TSC1/TSC2 in the mTOR pathway and the proven mTOR upregulation in TSC lesions, mTOR inhibitors should offer the effective therapeutic strategy for TSC patients. The effects of mTOR inhibitors have been studied in both cell cultures and animal models of TSC. Many studies have shown that rapamycin and its congeners act as signal transduction inhibitors. In in vitro models of TSC mTOR inhibitors reduced the proliferative potential of cultured cells as well as inhibit the phosphorylation of mTOR downstream targets: S6K and 4E-BP1. In SEGA cell cultures, rapamycin treatment resulted in marked reduction in S6K phosphorylation and significant decrease in tumor cell size (Jó´zwiak et al., 2007; Tyburczy et al., 2010). In the culture of mouse hippocampal neurons, Choi et al. (2008) showed that rapamycin prevents axonal, but not dendritic growth. Moreover, rapamycin treatment of Tsc2 deficient neurons resulted in the decreased number of neurons with multiple axons. The significance of these observations in humans with TSC warrants further studies. Recent studies, however, indicate that inhibition of mTORC1 with everolimus or sirolimus may lead to

51

rebound activation of Akt, which can protect cells from drug-induced death. Inhibition of both mTOR and Akt pathway, as shown by Polizzi et al. (2009) suppresses tumor growth in Tsc2+/− mice. However, as documented in the study of Jó´zwiak et al. (2007) Akt activation is not constantly observed in SEGAs; instead, Erk activation seems to play an important role in the pathogenesis of these tumors. Mi et al. (2009) showed in mice Tsc2 knockout cells, that combinatorial application of rapamycin and MAPK inhibitor attenuated the disruption of feedback suppression of Erk/MAPK signaling from mTOR caused by rapamycin alone. Moreover, the dual inhibition of mTORC1 and Erk/MAPK pathways exerted more potent negative effect on the proliferation of Tsc2 deficient cells than inhibition of either of these pathways rendered singly. Rapamycin treatment of human SEGA primary cultures resulted in significant decrease of cell size but did not affect their proliferation, morphology, or migration (Tyburczy et al., 2010). In contrast, treatment with an inhibitor of Erk kinase activity, U0126, resulted in inhibition of proliferation and migration together with change in cell morphology (the cells became more round and actin filaments were arranged) in primary SEGA cultures. Inhibition of Erk activity did not change the size of SEGA cells in culture. The most striking effect, however, was observed when both mTOR and ERK signaling pathway were inhibited. Treatment of primary SEGA cell cultures with combination of U0126 and rapamycin resulted in significant decrease of cells viability, proliferation, migration and profound changes in cell morphology. Lee et al. (2009) examined the effect of combination of rapamycin with various agents, including interferon gamma, statins, doxycycline, and sorafenib. All these combinations, except the last one, failed to be superior to rapamycin alone in the treatment of kidney tumors in Tsc2+/− mice. Moreover, in this study the effect of rapamycin or any of the tested agents given alone was not different from control. Sorafenib is multi-targeted kinase inhibitor that inhibits Erk/MAPK pathway as well as VEGFR, and PDGFR. While applied together with rapamycin, sorafenib markedly increased mice survival and decreased tumor volume. The animal models of TSC include Eker rats, and several knock-out mice models. The Eker rat is a spontaneous model of TSC2 mutation (Kwiatkowski, 2003; Kwiatkowski and Manning, 2005). Homozygous

52

Tsc2−/− animal die prenatally, and heterozygous TSC2+/− develop kidney cystadenomas and carcinomas, pituitary adenomas, uterine leiomyomas and leiomyosarcomas, splenic haemangiosarcomas, as well as brain hamartomas resembling human subependymal nodules. Pituitary adenomas represent the major cause of mortality in Eker rats in many studies. In this model, rapamycin was shown to decrease the size of kidney tumors and reduce the mortality related to pituitary adenoma. Several TSC1 and TSC2 knockout models were elaborated in mice (Kwiatkowski, 2003; Kwiatkowski and Manning, 2005). Generally, in TSC2+/− mice the kidney tumors, liver heamangiomas and heamangiosarcomas of various localizations are observed. Homozygous knockout mice die at midgestation. TSC1 knockout mice are characterized by the kidney tumors, liver heamangiomas and heamangiosarcomas of various localizations, however, the development of tumors is sex dependent. Females develop liver haemangiomas more frequently than males and the tumors are more severe in them. Homozygous TSC2−/− mice die at midgestation. Neither TSC1+/− nor TSC2+/− mice do not demonstrate any neuropathological features characteristic for TSC (Kwiatkowski, 2003; Kwiatkowski and Manning, 2005). There are several models of conditional deletion of TSC1 or TSC2 in mice (Lee et al., 2009; Polizzi et al., 2009). The mice presenting astrocyte-restricted TSC1 knockout show several astrocyte abnormalities and develop seizures. The mice with neuron-specific TSC1 knockout demonstrate many enlarged and dysplastic neurons, as well as cortical dysplasia. In both models, rapamycin treatment resulted in reduced neuronal enlargement, reduced astrocyte proliferation and correction of mTORC1 signaling. However, it should be noted that none of the models present with tumors resembling human SEGAs.

Targeting mTOR in Subependymal Giant Cell Astrocytoma Recognition of signaling pathways involved in TSC pathogenesis as well as the encouraging results of inhibiting mTOR in animal models of TSC lead to clinical trials. In 2006, Franz et al. (2006) reported R ) the effects of rapamycin (sirolimus, Rapamune

K. Kotulska and S. Jó´zwiak

treatment in four TSC patients with SEGA. Three children and one adult received rapamycin due to symptomatic and growing SEGA. Serum level of rapamycin ranged from 7.7 to 10.9 ng/ml. In all patients the medication resulted in significant tumor size decrease. It is important to note, that reduction of SEGA volume was observed as soon as two weeks after the onset of treatment. The side effects included: lipid elevation, aphthous ulcers, acneiform rash, and ankle oedema. Rapamycin appeared to be effective also in three pediatric TSC patients with SEGA reported by Lam et al. (2009). In this study, rapamycin caused 50–65% regression of SEGA after three months of treatment. Serum levels of rapamycin ranged from 10 to 15 ng/ml. However, discontinuation of rapamycin treatment in one patient resulted in tumor regrowth. Resumption of therapy was again associated with tumor regression. In all rapamycin studies in SEGA patients, the medication was well tolerated and no unsuspected or major adverse events were noted. The most frequent adverse events included: oral ulcers and transient hypercholesterolaemia. Jó´zwiak et al. (2011) reported a case of a child with giant, inoperable, life-threatening SEGA treated R ). This child underwent with everolimus (Certican several subtotal surgeries, and, due to intraoperative cardiac arrest, was excluded from further neurosurgical interventions. His history is notable for high CSF protein levels which have required external drainage multiple times. Everolimus treatment at 4.5 mg/m2 per day resulted in dramatic clinical improvement and significant decrease in tumor size. At the same time, CSF protein decreased, allowing implantation of a peritoneal shunt (Fig. 5.3.). No side effects were noted. An open-label, prospective study of everolimus R ) for SEGA in TSC patients was pub(Certican lished in 2010, by Krueger et al. (2010). Twenty-eight patients included in this study were 3 years of age or older and presented with serial growth of SEGA on at least two successive MRI scans. Everolimus was administered at a starting dose of 3.5 mg/m2 of body surface area. The dose was than adjusted to achieve a whole-blood trough concentration of 5–15 ng/ml. The primary endpoint of this study was the change in the volume of SEGA after 6 months of treatment. In 75% of patients the tumor reduction by at least 30% was observed. In 32% of patients, the reduction was 50% or more. Tumor shrinkage was most rapid during the initial 3 months of treatment with evidence

5

Subependymal Giant Cell Astrocytoma: Role of mTOR Pathway and Its Inhibitors

Fig. 5.3 Giant SEGA In 10-year-old boy with TSC (a). After several subtotal resections of the tumor, the child was excluded from further neurosurgical interventions. Everolimus treatment at 4.5 mg/m2 per day resulted in dramatic clinical improvement and significant decrease in tumor size, as well as regression of hydrocephalus in 6 months (b). Courtesy Prof. Elzbieta Jurkiewicz, Department of Radiology, The Children’s Memorial Health Institute, Warsaw, Poland

a

of a sustained response during treatment lasting for 21.5 months on average. No patient developed new lesions, worsening hydrocephalus, increased intracranial pressure, and none required surgical resection or other therapy for SEGA. Moreover, authors showed that everolimus treatment was associated with a clinically relevant reduction in seizure frequency. The treatment was generally well tolerated. One patient discontinued everolimus due to hyperkinesis and one due to infection. The most common adverse events were mild and included: oral mucositis, upper respiratory tract infection, sinusitis, otitis media, fever, gastric infection, acneiform rash, and skin infection. Laboratory abnormalities were seen in 30% of patients and included transaminase elevations, hypercholesterolemia and hypertriglyceridemia, leukopenia, anemia, and hyperglycemia. Currently, a large, multicenter, randomized, placebo-controlled, phase III clinical trial (CRAD001M2301, EXIST-1) of everolimus is ongoing in patients of any age with SEGA. This study will evaluate the antitumor activity and clinical benefits of everolimus in growing SEGA associated with TSC. The impact of everolimus treatment on epilepsy and cognition will also be examined. On October 29, 2010, the US Food and Drug Administration granted accelerated approval to R ), an mTOR inhibitor, for everolimus (Afinitor patients with SEGA associated with TSC, who require therapy but are not candidates for surgical resection.

53

b

Conclusions Subependymal giant cell astrocytoma is a rare brain tumor, but represents an important possible cause of mortality and morbidity among children with TSC. Until recently, the only possible treatment of SEGA was the surgical intervention. Now, thanks to the recognition of mTOR pathway control disruption in TSC, mTOR inhibitors can be offered to patients with growing SEGA. The initial clinical results are tremendously encouraging, however, still many questions remain to be answered and many issues to be studied. First, the safety data of these drugs are limited. Of utmost importance, the possible negative impact of mTOR inhibitors in very young children was not explored yet. Second, the long-term efficacy of mTOR inhibiting is not known. And finally, many data indicate that mTOR inhibitors should be administered continuously, otherwise the tumors regrow.

References Chan JA, Zhang H, Roberts PS, Jozwiak S, Grajkowska W, Lewin-Kowalik J, Kotulska K, Kwiatkowski DJ (2004) Pathogenesis of tuberous sclerosis subependymal giant cell astrocytomas: biallelic inactivation of TSC1 or TSC2 leads to mTOR activation. J Neuropathol Exp Neurol 63: 1236–1242

54 Choi YJ, Di Nardo A, Kramvis I, Meikle L, Kwiatkowski DJ, Sahin M, He X (2008) Tuberous sclerosis complex proteins control axon formation. Genes Dev 22(18):2485–2495 Crino PB (2008) Rapamycin and tuberous sclerosis complex: from Easter Island to epilepsy. Ann Neurol 63(4): 415–417 Curatolo P, Bombardieri R, Jozwiak S (2008) Tuberous sclerosis. Lancet 372:657–668 Dabora SL, Jozwiak S, Franz DN, Roberts PS, Nieto A, Chung J, Choy YS, Reeve MP, Thiele E, Egelhoff JC, Kasprzyk-Obara J, Domanska-Pakiela D, Kwiatkowski DJ (2001) Mutational analysis in a cohort of 224 tuberous sclerosis patients indicates increased severity of TSC2, compared with TSC1, disease in multiple organs. Am J Hum Genet 68:64–80 de Ribaupierre S, Dorfmüller G, Bulteau C, Fohlen M, Pinard JM, Chiron C, Delalande O (2007) Subependymal giant-cell astrocytomas in pediatric tuberous sclerosis disease: when should we operate? Neurosurgery 60(1):83–89 Fingar DC, Salama S, Tsou C, Harlow E, Blenis J (2002) Mammalian cell size is controlled by mTOR and its downstream targets S6K1 and 4EBP1/eIF4E. Genes Dev 16: 1472–1487 Franz DN, Leonard J, Tudor C, Chuck G, Care M, Sethuraman G, Dinopoulos A, Thomas G, Crone KR (2006) Rapamycine causes regression of astrocytomas in tuberous sclerosis complex. Ann Neurol 59:490–498 Freilinger A, Rosner M, Krupitza G, Nishino M, Lubec G, Korsmeyer SJ, Hengstschlãger M (2006) Tuberin activates the proapoptotic molecule BAD. Oncogene 25(49): 6467–6479 Gibbons JJ, Abraham RT, Yu K (2009) Mammalian target of rapamycin: discovery of rapamycin reveals a signaling pathway important for normal and cancer cell growth. Semin Oncol 36(Suppl 3):3–17 Grajkowska W, Kotulska K, Jurkiewicz E, Matyja E (2010) Brain lesions in tuberous sclerosis complex. Review. Folia Neuropathol 48(3):139–149 Guba M, Yezhelyev M, Eichhorn ME, Schmid G, Ischenko I, Papyan A, Graeb C, Seeliger H, Geissler EK, Jauch KW, Bruns CJ (2005) Rapamycin induces tumor-specific thrombosis via tissue factor in the presence of VEGF. Blood 105(11):4463–4469 Henske EP, Scheithauer BW, Short MP, Wollmann R, Nahmias J, Hornigold N, van Slegtenhorst M, Welsh CT, Kwiatkowski DJ (1996) Allelic loss is frequent in tuberous sclerosis kidney lesions but rare in brain lesions. Am J Hum Genet 59(2): 400–406 Henske EP, Wessner LL, Golden J, Scheithauer BW, Vortmeyer AO, Zhuang Z, Klein-Szanto AJ, Kwiatkowski DJ, Yeung RS (1997) Loss of tuberin in both subependymal giant cell astrocytomas and angiomyolipomas supports a two-hit model for the pathogenesis of tuberous sclerosis tumors. Am J Pathol 151(6):1639–1647 Ichikawa T, Wakisaka A, Daido S, Takao S, Tamiya T, Date I, Koizumi S, Niida Y (2005) A case of solitary subependymal giant cell astrocytoma. Two somatic hits of TSC2 in the tumor, without evidence of somatic mosaicism. J Mol Diagn 7(4):544–549 Inoki K, Li Y, Zhu T, Wu J, Guan K (2002) Tsc2 is phosphorylated and inhibited by Akt and suppresses mTOR signaling. Nat Cell Biol 4:648–657

K. Kotulska and S. Jó´zwiak Johnson MW, Miyata H, Vinters HV (2002) Ezrin and moesin expression within the developing human cerebrum and tuberous sclerosis-associated cortical tubers. Acta Neuropathol 104(2):188–96 Jozwiak J, Grajkowska W, Kotulska K, Jozwiak S, Zalewski W, Zajaczkowska A, Roszkowski M, Slupianek A, Wlodarski P (2007) Brain tumor formation in tuberous sclerosis depends on Erk activation. Neuromol Med 9(2):117–127 Jozwiak S, Schwartz RA, Janniger CK, Bielicka-Cymerman J (2000) Usefulness of diagnostic criteria of tuberous sclerosis complex in pediatric patients. J Child Neurol 15:652–659 Jó´zwiak S, Kwiatkowski D, Kotulska K, Larysz-Brysz M, Lewin-Kowalik J, Grajkowska W, Roszkowski M (2004) Tuberin and hamartin expression is reduced in the majority of subependymal giant cell astrocytomas in tuberous sclerosis complex consistent with a two-hit model of pathogenesis. J Child Neurol 19(2):102–106 Jó´zwiak S, Perek-Polnik M, Kotulska K, Jurkiewicz E, Roszkowski M, Perek D (2011) Effective everolimus treatment of inoperable, life-threatening SEGA in a patient with tuberous sclerosis. Abstracts for 9th European Pediatric Neurology Society Congress, Cavtat, Croatia, 2011. Jung CH, Ro SH, Cao J, Otto NM, Kim DH (2010) mTOR regulation of autophagy. FEBS Lett 584(7):1287–1295 Kim SK, Wang KC, Cho BK, Jung HW, Lee YJ, Chung YS, Lee JY, Park SH, Kim YM, Choe G, Chi JG (2001) Biological behavior and tumorigenesis of subependymal giant cell astrocytomas. J Neurooncol 52(3):217–225 Krueger DA, Care MM, Holland K, Agricola K, Tudor C, Mangeshkar P, Wilson KA, Byars A, Sahmoud T, Franz DN (2010) Everolimus for subependymal giant-cell astrocytomas in tuberous sclerosis. N Engl J Med 363(19):1801–1811 Kwiatkowski DJ (2003) Rhebbing up mTOR: new insights on TSC1 and TSC2, and the pathogenesis of tuberous sclerosis. Cancer Biol Ther 2(5):471–3476 Kwiatkowski DJ, Manning BD (2005) Tuberous sclerosis: a GAP at the crossroads of multiple signaling pathways. Hum Mol Genet 14:251–258 Lam C, Bouffet E, Tabori U, Mabbott D, Taylor M, Bartels U (2009) Rapamycin (sirolimus) in tuberous sclerosis associated pediatric central nervous system tumors. Pediatr Blood Cancer 54(3):476–479 Lane HA, Wood JM, McSheehy PM, Allegrini PR, Boulay A, Brueggen J, Littlewood-Evans A, Maira SM, Martiny-Baron G, Schnell CR, Sini P, O’Reilly T (2009) mTOR inhibitor RAD001 (everolimus) has antiangiogenic/vascular properties distinct from a VEGFR tyrosine kinase inhibitor. Clin Cancer Res 15(5):1612–1622 Lee N, Woodrum L, Nobil A, Rauktys A, Messina M, Dabora S (2009) Rapamycin weekly maintenance dosing and the potential efficacy of combination sorafenib plus rapamycin but not atorvastatin or doxycycline in tuberous sclerosis preclinical models. BMC Pharmacol 9:8–23 Lin Y, Henderson P, Pettersson S, Satsangi J, Hupp T, Stevens C (2011) Tuberous sclerosis-2 (TSC2) regulates the stability of death-associated protein kinase-1 (DAPK) through a lysosome-dependent degradation pathway.. FEBS J 278(2):354–370 Loewith R, Jacinto E, Wullschleger S, Lorberg A, Crespo JL, Boenefant D, Oppliger W, Jenoe P, Hall MN (2002) Two

5

Subependymal Giant Cell Astrocytoma: Role of mTOR Pathway and Its Inhibitors

TOR complexes, only one of which is rapamycin sensitive, have distinct roles in growth control. Mol Cell 10:457–468 Mi R, Ma J, Zhang D, Li L, Zhang H (2009) Efficacy of combined inhibition of mTOR and ERK/MAPK pathways in treating a tuberous sclerosis complex cell model. J Genet Genomics 36:355–361 Pollizzi K, Malinowska-Kolodziej I, Stumm M, Lane H, Kwiatkowski D (2009) Equivalent benefit of mTORC1 blockade and combined PI3K-mTOR blockade in a mouse model of tuberous sclerosis. Mol Cancer 15(8):38 Rosner M, Hengstschlaeger M (2008) Cytoplasmic and nuclear distribution of the protein complexes mTORC1 and mTORC2: rapamycin triggers dephosphorylation and delocalization of the mTORC2 components rector and sin1. Hum Mol Genet 17(19):2934–2948 Rüegg S, Baybis M, Juul H, Dichter M, Crino PB (2007) Effects of rapamycin on gene expression, morphology, and electrophysiological properties of rat hippocampal neurons. Epilepsy Res 77(2–3):85–92

55

Shepherd CW, Gomez MR, Lie JT, Crowson CS (1991) Causes of death in patients with tuberous sclerosis. Mayo Clin Proc 66(8):792–796 Torres OA, Roach ES, Delgado MR, Sparagana SP, Sheffield E, Swift D, Bruce D (1998) Early diagnosis of subependymal giant cell astrocytoma in patients with tuberous sclerosis. J Child Neurol 13(4):173–177 Tyburczy ME, Kotulska K, Pokarowski P, Mieczkowski J, Kucharska J, Grajkowska W, Roszkowski M, Jozwiak S, Kaminska B (2010) Novel proteins regulated by mTOR in subependymal giant cell astrocytomas of patients with tuberous sclerosis complex and new therapeutic implications. Am J Pathol 176(4):1878–1890 Zhang Y, Gao X, Saucedo LJ, Ru B, Edgar BA, Pan D (2003) Rheb is a direct target of the tuberous sclerosis tumour suppressor proteins. Nat Cell Biol 5:578–581

Chapter 6

Role of Progesterone Receptor Isoforms in Human Astrocytomas Growth Ignacio Camacho-Arroyo, Valeria Hansberg-Pastor, Edith Cabrera-Muñoz, Olivia Tania Hernández-Hernández, and Aliesha González-Arenas

Abstract Progesterone (P) regulates several functions in cells through the interaction with its intracellular receptor (PR), which is a ligand-activated transcription factor that modifies the expression of genes involved in the control of cell growth and proliferation, such as vascular endothelial growth factor and epidermal growth factor receptor. Two PR isoforms have been reported: PR-B and PR-A, encoded by the same gene but with different function and regulation. It has been shown that PR isoforms are expressed in U373 and D54 cell lines, which are derived from grades III and IV of human astrocytomas, respectively. Our group has recently reported that P increases cell growth in both cells lines. The PR antagonist, RU486, blocked P effects and its treatment alone significantly reduced human astrocytomas cell growth in vitro. The over-expression of PR-A in U373 cells significantly reduced P effects. These data suggest that P regulates human astrocytomas cell growth through the interaction with PR and that PR-B/PR-A expression ratio is determinant in P functions in astrocytomas. Keywords Progesterone · Brain tumor · Glioma · PR isoforms · Ubiquitin–proteasome system

I. Camacho-Arroyo () Facultad de Quimica, Universidad Nacional Autonoma de Mexico, Coyoacan, Mexico 04510, D.F. Mexico e-mail: [email protected]

Introduction Progesterone (P) participates in the regulation of several physiological and pathological processes in the brain of mammals. This hormone has been involved in the growth of brain tumors such as chordomas, meningiomas and astrocytomas. The latter are the most frequent primary brain tumors and constitute a leading cause of cancer related deaths. P mainly elicits its effects by interaction with its intracellular receptor (PR), which is a ligand-activated transcription factor. Two isoforms of PR, PR-A and PR-B, differentially regulate gene transcription, recognize different promoters and present distinct functions. PR activity is fundamental in development, growth and proliferation of tumors. In this chapter we present an overview of the research about the role of P and PR isoforms in cell growth of human astrocytomas.

Astrocytomas Astrocytomas arise from astrocytes and are the most common primary intracerebral neoplasms in humans. Astrocytic tumors constitute 65–70% of all gliomas, and four malignancy grades are recognized by the World Health Organisation (WHO). As described by Rousseau et al. (2008), Grade I applies to lesions with low proliferative potential, minimal variation in shape and size of nuclei and the possibility of cure following surgical resection. Grade II tumors are generally infiltrative, present low proliferative activity, nuclear atypia and some grade II tumors tend to progress to higher grades of malignancy. The designation of grade III tumor is generally reserved for lesions with

M.A. Hayat (ed.), Tumors of the Central Nervous System, Volume 5, DOI 10.1007/978-94-007-2019-0_6, © Springer Science+Business Media B.V. 2012

57

58

histological evidence of malignancy, including nuclear atypia and brisk mitotic activity, patients with grade III tumors receive adjuvant radiation and/or chemotherapy. Grade IV tumors exhibit more advanced features of malignancy, including vascular proliferation, mitosis and necrosis, typically associated with rapid pre- and postoperative disease evolution and a fatal outcome. Hulleman and Helin (2005), mentioned that average survival of patients with astrocytomas grade I could be long after surgery and treatment, while grade II is around 7 years and patients with astrocytomas grade III have a median survival half that time, around 3.5 years. Astrocytomas grade IV (glioblastomas) patients have a very poor prognosis with average survival reported between 9 and 11 months. Ichimura et al. (2004), classified the glioblastomas into those that develop from a previously diagnosed astrocytoma and those that appear to develop de novo. The treatment given to patients with astrocytomas depends on many factors, including the tumor size and localization, its growth rate, and the symptoms the patient is experiencing. Various strategies have been used to treat astrocytomas including extensive surgical resection, fractionated and focused radiation and intracavitary and/or intra-arterial chemotherapy that result in prolonged and non-always significant survival for patients and compromise brain function. An alternative treatment for astrocytomas is hormonal therapy based on sex steroid hormones such as P, which participates in the regulation of cell proliferation of several tumors such as breast cancer and meningiomas (intracranial or intraspinal from the arachnoidal layer of meninges).

Genomic Mechanism of Action of Progesterone (P) Many P actions are mediated by PR which is a ligand-activated transcription factor that regulates the expression of several genes involved in metabolism, development and reproduction as well as cell cycle progression. Camacho-Arroyo (2003), mentioned that the genomic mechanism of action involves the interaction of P with PR, located in cytoplasm or nucleus. In absence of ligand PR is associated with heat shock proteins (HSP70 and HSP90). When the hormone

I. Camacho-Arroyo et al.

interacts with PR, it induces conformational changes that allow the dissociation with heat shock proteins, followed by phosphorylation and dimerization of the receptor. The resulting structure possesses high affinity for specific sequences in the DNA, known as P response elements (PRE) which are present in the promoter region of P target genes. Once bound to PRE, PR is able to recruit coregulators proteins (coactivators or corepressors) regulating gene transcription (Fig. 6.1). Nuclear receptors coregulators are required by the receptors for efficient transcriptional regulation. Coactivators interact with nuclear receptors in a ligand-dependent manner and enhance their transcriptional activity, whereas, corepressors interact with nuclear receptors, either in the absence of hormone or in the presence of antagonists and diminish the transcription rate of their target genes. Rosenfeld et al. (2006), have reported that the steroid receptor coactivator (SRC) family is a wellstudied group of coregulator proteins that have histone acetyltransferase (HAT) activity. Some of the bestcharacterized nuclear receptor coactivators belong to the SRC-family. McKenna et al. (1999), mentioned that the members of SRC family interact with steroid receptors including PR, and enhance their transcriptional activation in a ligand-dependent manner. Lange et al. (2000) and Camacho-Arroyo et al. (2002), observed that P induces PR phosphorylation, which signals it to degradation by the ubiquitin–proteasome system.

Transcriptional Activity of PR Isoforms Kastner et al. (1990), have reported two main PR isoforms in humans: a full-length form (PR-B, 114 kDa) and an N-terminal truncated one (PR-A, 94 kDa). PR isoforms are encoded by the same gene, but their expression is regulated by distinct promoters. Leonhardt et al. (2003), have seen in general that PR-B is a much stronger transcriptional activator than PR-A and the latter exhibits a dominant negative inhibitory effect on the activity of PR-B. Richer et al. (2002), observed in human breast cancer cells that PR-A and PR-B are functionally unique transcriptional factors that differentially regulate gene transcription within the same promoter context, and are capable of recognizing different promoters.

6

Role of Progesterone Receptor Isoforms in Human Astrocytomas Growth

59

Fig. 6.1 Genomic mechanism of P action. In the unligated form, PR is associated with chaperone proteins such as heat shock proteins HSP70 and HSP90. After P cell entrance and binding, PR dissociates from chaperone proteins and dimerizes with another PR. Then, PR is phosphorylated (yellow triangles) and forms a complex with coregulators such as the coactivator protein SRC-1. PR interacts with progesterone responsive elements (PRE) in the promoter region of steroid target genes, the basal transcription machinery is recruited and the gene is transcribed. P, progesterone; A/B, C, D and E are PR domains, TIFIIA/B, transcription initiation factors, CBP/p300, histone acetyltransferase

Richer et al. (2002), reported that in human breast cancer cells of 94 genes regulated by P: 65 were uniquely regulated by PR-B, 4 uniquely by PR-A, and only 25 by both PR isoforms, e.g. genes involved in cell cycle and apoptosis such as BIRC3 or PCNA were up regulated by PR-B meanwhile Bcl-X was upregulated by PR-A. TCF8 and DSIPI transcription factors are up regulated by PR-B and PR-A respectively. Mulac-Jericevic et al. (2000), demonstrated that PR-B is required for normal mammary gland development, while PR-A is essential for uterine development and reproductive function. Mulac-Jericevic et al. (2000), found that P up-regulation of calcitonin and amphiregulin genes depends on PR-A. All these studies suggest that each PR isoform has a different transcriptional function that regulates different cell processes.

Regulation of PR Isoforms Expression The human PR gene is located in chromosome 11 (11q22-q23) and consists of eight exons of variable sizes separated by seven introns. Kastner et al. (1990),

identified two distinct promoters and two translational start sites that produce the two PR isoforms, A and B. Promoter A (+464 to +1105) has three putative transcription start sites (+751, +761 and +842) and a translational start site ATG1 at +1236. Promoter B (–711 to +31) has two transcription start sites at +1 and +15 and a translational start site ATG2 at +744. It is known that PR isoforms are differentially regulated by E and P in different cells and tissues. GuerraAraiza et al. (2003), mentioned that in several cell types PR isoforms are up-regulated by E. According to Kraus et al. (1994), PR up-regulation by E is mediated by half estrogen-responsive elements located in the PR promoter whereas PR downregulation by its own ligand (P) is associated with ligand-dependent proteolysis; Camacho-Arroyo et al. (2002) and Lange et al. (2000), informed that P induces receptor phosphorylation, which signals PR to degradation by the ubiquitin–proteasome pathway. Estrogens effects are mediated through its interaction with two intracellular receptors (ERs), ERα and ERβ. Although the classical model of estrogen action has proposed that ER induces gene expression by binding first to the ligand and then to the estrogen

60

response element (ERE) in DNA, neither of the PR promoters contains a consensus palindromic ERE, however, any of ERs can bind to these promoters. Ellmann et al. (2009), reviewed that estrogen responsiveness can also be mediated by the interaction of ER with DNA-bound transcription factors such as AP-1 and Sp1 proteins. Numerous transcription factor binding sites can be far apart from the proximal promoter and still play a role in regulating gene expression. Recently, BonéyMontoya et al. (2010), identified eight regions associated with ERα, located 48–311 kb upstream of the PR-B transcription start site which possesses one or more EREs and each of them comprise one consensus ERE half-site. DNA methylation regulates gene transcription by modulating chromatin conformation. Momparler and Bovenzi (2000), found that hypermethylated DNA in promoter regions of many genes is usually associated with downregulated or silenced gene expression. Works from Liu et al. (2003), assumed that in many cancer cell lines one or both PR isoforms are silenced by methylation.

Role of P in Astrocytomas Proliferation Sager et al. (2003), account for many reports about proliferative and anti-proliferative effects of P in different types of cancer. In brain tumors P has different effects. Olson et al. (1986), described that in cell cultures of meningiomas different doses of P (1–100 nM) stimulate cell growth, but in prolactinomas, Piroli et al. (1998), observed that P inhibits cell growth. Pinski et al. (1993), published that a high dose of the PR antagonist RU486 (0.5 mg/day, during 4 days) was more effective to reduce the tumor volume as compared with the lower dose (0.1 mg/day) in nude mice bearing xenografts of the human malignant glioma U87MG cell line. In contrast, Altinoz et al. (2001), reported that the progestin medroxyprogesterone (6 μM), inhibits S-phase of C6 rat glioma cells by 41 and 73% at 48 and 96 h, respectively. However, P effects on human astrocytomas growth have not been completely characterized, GonzalezAguero et al. (2007), have observed that P induces cell proliferation in U373 and D54 cell lines derived from human astrocytomas grades III and IV, respectively. In

I. Camacho-Arroyo et al.

a time-course study over a 5-day period with different doses of P (1 nM to 10 μM). In both cell lines P (10 nM) induced an increase in cell growth from the second day of culture in D54 cells and from the third day in the case of U373 cells. In both cell lines P (10 nM) effect persisted until day 5. It is important to mention that the concentration of P that induces a significant increase in the number of astrocytoma cells was found by Stening et al. (2007), in the luteal phase of the woman menstrual cycle. The treatment with the PR antagonist, RU486 (10 μM) without P for 5 days significantly decreased the number of U373 and D54 cells as compared with vehicle treatment from the second day of the experiment. RU486 co-administered with P significantly blocked the effects of the latter on days 2 and 4 in D54 and U373 cells, respectively, suggesting that PR is involved in P regulation of astrocytomas growth. Hernandez-Hernandez et al. (2010), determined SRC-1 and SRC-3 expression regulation by P in U373 and D54 cell lines. In D54 cell line SRC-1 mRNA expression and protein content were induced by P. SRC-3 expression was not modified by P in any cell line. Recently, we determined the effect of P over the mRNA expression and protein content of three different molecules involved in cell proliferation and metastasis: epidermic growth factor receptor (EGFR), vascular endothelial growth factor (VEGF) and cyclin D1. We found that P did not regulate mRNA expression of EGFR or VEGF in U373 cells. However, in D54 cells this hormone increased EGFR and VEGF mRNA expression and protein content; this effect was blocked by RU486. Cyclin D1 mRNA was up-regulated by P in U373 cells and the effect was also blocked by RU486. Our data suggest that P effects on EGFR, VEGF and cyclin D1 expression are mediated by its PR.

PR Isoforms Regulation and Function in Astrocytomas Khalid et al. (1997), found that PR expression assessed by immunohistochemistry directly correlates with histologic grades of human astrocytomas. The percentage of PR expressing in cells high-grade astrocytic tumor biopsies was higher than that of low-grade ones, suggesting that PR-positive tumors possess a high proliferative potential. Carroll et al. (1995), and

6

Role of Progesterone Receptor Isoforms in Human Astrocytomas Growth

61

Assimakopoulou et al. (1998), detected a strong PR nuclear immunopositivity in glioblastomas of higher malignancy compared with anaplastic astrocytomas (grade III) and lower grade ones. Gonzalez-Aguero et al. (2001), have observed that PR gene was expressed in 83 and 100% of biopsies from human astrocytomas grades III and IV, respectively. At mRNA level, PR-B expression was higher than of PR-A in human astrocytomas grades III and IV. PR immunostaining was detected in 85% of astrocytomas (grades III and IV) and, as in the case of mRNA level, PR-B isoform was the predominant one in most tumors (73%). Gonzalez-Aguero et al. (2007), have detected PR isoforms in U373 and D54 human astrocytoma cell lines. In U373 cell line PR-B was the predominant isoform (PR-B:PR-A ratio 3:1) whereas in D54 cells PR-A was the predominant one (PR-B:PR-A ratio 0.66:1). As Richer et al. (2002), reported, PR isoforms rate is important because P can exert different functions in a cell, depending on the expression pattern of its isoforms. In the mammary gland of PRA knockout mice (PR-A KO) Mulac-Jericevic et al. (2000), observed a lower expression of lactoferrin as compared with PR KO after treatment with estradiol (E) and P, and the contrary effect was observed for

calcitonin and histidine carboxylase expression. Thus, this differential PR isoform expression should be involved in P effects in U373 and D54 cell growth. In U373 and D54 astrocytoma cell lines CabreraMunoz et al. (2009), found that PR isoforms are regulated by E (10 nM) and P (10 nM). In U373 cells PR isoforms content was increased by E, whereas in D54 cells E had no significant effects. In both cell lines P alone did not modify PR isoforms content but this was down-regulated when P was administered after E treatment. PR-A isoform was more sensitive to E + P treatment than PR-B since a greater diminution was observed in PR-A content after the combined treatment, and PR down-regulation was blocked with PR antagonist, RU 486, in U373 and D54 cells. Cabrera-Munoz et al. (2009), have evaluated the effects of PR-A over-expression on cell growth of U373 cells (which endogenously express a low amount of PR-A) and observed an increase of 60% in PRA content in transfected U373 cells. In a time-course study over a 6-day period PR-A over-expression significantly diminished the cell number of U373 cells treated with P (from day 4 to day 6), suggesting that PR-A has an inhibitory effect on cell growth when it is activated by its ligand (Fig. 6.2).

Fig. 6.2 Effects of PR-A transfection on U373 human astrocytomas cell growth. Upper panel, U373 wild type (WT) and U373 cells transfected with PR-A (PR-A) were lysed, proteins (70 μg) were separated by electrophoresis, and gels were electrotransferred for Western blot detection of both PR isoforms (PR-A and PR-B). Lower panel, U373 cells WT () or transfected with PR-A () were treated with hormone vehicle (0.02%

cyclodextrin) (left) or with 10 nM of P (right) (day 0). Each experiment was performed in three independent cultures, each one by duplicate, during 6 days. Every day cells were removed from incubation and the number of cells was measured by trypan blue dye exclusion. Data are mean ± S.E.M. ∗ p < 0.05 vs. WT. Published with permission of Elsevier, licence number: 2394291286863

62

Conclusion and Perspectives Astrocytomas are the most frequent primary brain tumors and constitute a leading cause of cancer related deaths. An alternative treatment for astrocytomas is hormonal therapy based on sex steroid hormones such as P that increases cell proliferation in U373 and D54 human astrocytoma cell lines by interacting with PR. The expression of PR isoforms (PR-A and PR-B) directly correlates with malignancy grades of human astrocytomas, being PR-B the predominant isoform in high-grade tumors. The increase in astrocytoma cell growth in vitro by P should involve changes in the expression of several factors participating in the control of angiogenesis, growth and proliferation, processes altered in cancer. The overexpression of PR-A decreases astrocytoma grade III cell growth in vitro when it is activated by its ligand, suggesting that PR-A/PR-B ratio is determinant in P effects on human astrocytomas growth. Phosphorylation increases transcriptional activity of PR and also induces its degradation by the ubiquitinproteasome pathway; therefore a basic aspect in the study of PR isoforms function in human astrocytomas is its regulation by phosphorylation and the kinases involved in this event. Transcriptional regulation of the PR gene requires the combined participation of various cis and transelements. The study of PR expression has demonstrated that methylation of one or both of its promoter regions plays an important role in cancer. Therefore, studying the methylation pattern of the PR promoter region in astrocytomas is important to understand the relation between PR content and tumor progression. Another key aspect is the analysis of genes regulated by P in these brain tumors which would give relevant information about molecules involved in proliferation or metastasis in human astrocytomas.

References Altinoz MA, Bilir A, Ozar E, Onar FD, Sav A (2001) Medroxyprogesterone acetate alone or synergistic with chemotherapy suppresses colony formation and DNA synthesis in C6 glioma in vitro. Int J Dev Neurosci 19:541–547 Assimakopoulou M, Sotiropoulou-Bonikou G, Maraziotis T, Varakis J (1998) Does sex steroid receptor status have any prognostic or predictive significance in brain astrocytic tumors? Clin Neuropathol 17:27–34

I. Camacho-Arroyo et al. Bonéy-Montoya J, Ziegler YS, Curtis CD, Montoya JA, Nardulli AM (2010) Long-range transcriptional control of progesterone receptor gene expression. Mol Endocrinol 24: 346–358 Cabrera-Munoz E, Gonzalez-Arenas A, Saqui-Salces M, Camacho J, Larrea F, Garcia-Becerra R, Camacho-Arroyo I (2009) Regulation of progesterone receptor isoforms content in human astrocytoma cell lines. J Steroid Biochem Mol Biol 113:80–84 Camacho-Arroyo I (2003) Progesterone receptor isoforms expression and progesterone actions in the brain. In: Pandalai SG (ed) Recent Research Developments in Life Sciences. Research Signpost, Kerala, India, 221–242 Camacho-Arroyo I, Villamar-Cruz O, Gonzalez-Arenas A, Guerra-Araiza C (2002) Participation of the 26S proteasome in the regulation of progesterone receptor concentrations in the rat brain. Neuroendocrinology 76:267–271 Carroll RS, Zhang J, Dashner K, Sar M, Black PM (1995) Steroid hormone receptors in astrocytic neoplasms. Neurosurgery 37:496–503 Ellmann S, Sticht H, Thiel F, Beckmann MW, Strick R, Strissel PL (2009) Estrogen and progesterone receptors: from molecular structures to clinical targets. Cell Mol Life Sci 66:2405– 2426 Gonzalez-Aguero G, Gutierrez AA, Gonzalez-Espinosa D, Solano JD, Morales R, Gonzalez-Arenas A, Cabrera-Munoz E, Camacho-Arroyo I (2007) Progesterone effects on cell growth of U373 and D54 human astrocytoma cell lines. Endocrine 32:129–135 Gonzalez-Aguero G, Ondarza R, Gamboa-Dominguez A, Cerbon MA, Camacho-Arroyo I (2001) Progesterone receptor isoforms expression pattern in human astrocytomas. Brain Res Bull 56:43–48 Guerra-Araiza C, Villamar-Cruz O, Gonzalez-Arenas A, Chavira R, Camacho-Arroyo I (2003) Changes in progesterone receptor isoforms content in the rat brain during the oestrous cycle and after oestradiol and progesterone treatments. J Neuroendocrinol 15:984–990 Hernandez-Hernandez T, Rodrıguez-Dorantes M, GonzalezArenas A, Camacho-Arroyo I (2010) Progesterone and estradiol effects on SRC-1 and SRC-3 expression in human astrocytoma cell lines. Endocrine 37:194–200 Hulleman E, Helin K (2005) Molecular mechanisms in gliomagenesis. Adv Cancer Res 94:1–27 Ichimura K, Ohgaki H, Kleihues P, Collins VP (2004) Molecular pathogenesis of astrocytic tumours. J Neurooncol 70: 137–160 Kastner P, Krust A, Turcotte B, Stropp U, Tora L, Gronemeyer H, Chambon P (1990) Two distinct estrogen-regulated promoters generate transcripts encoding the two functionally different human progesterone receptor forms A and B. EMBO J 9:1603–1614 Khalid H, Shibata S, Kishikawa M, Yasunaga A, Iseki M, Hiura T (1997) Immunohistochemical analysis of progesterone receptor and Ki-67 labeling index in astrocytic tumors. Cancer 80:2133–2140 Kraus WL, Montano MM, Katzenellenbogen BS (1994) Identification of multiple, widely spaced estrogen-responsive regions in the rat progesterone receptor gene. Mol Endocrinol 8:952–969

6

Role of Progesterone Receptor Isoforms in Human Astrocytomas Growth

Lange CA, Shen T, Horwitz KB (2000) Phosphorylation of human progesterone receptors at serine-294 by mitogenactivated protein kinase signals their degradation by the 26S proteasome. Proc Natl Acad Sci USA 97:1032–1037 Leonhardt SA, Boonyaratanakornkit V, Edwards DP (2003) Progesterone receptor transcription and non-transcription signaling mechanisms. Steroids 68:761–770 Liu ZJ, Maekawa M, Horii T, Morita M (2003) The multiple promoter methylation profile of PR gene and ER alpha gene in tumor cell lines. Life Sci 73:1963–1972 McKenna NJ, Lanz RB, O’Malley BW (1999) Nuclear receptor coregulators: cellular and molecular biology. Endocr Rev 20:321–344 Momparler RL, Bovenzi V (2000) DNA methylation and cancer. J Cell Physiol 183:145–154 Mulac-Jericevic B, Mullinax RA, DeMayo FJ, Lydon JP, Conneely OM (2000) Subgroup of reproductive functions of progesterone mediated by progesterone receptor-B isoform. Science 289:1751–1754 Olson JJ, Beck DW, Schlechte J, Loh PM (1986) Hormonal manipulation of meningiomas in vitro. J Neurosurg 65: 99–107 Pinski J, Halmos G, Shirahige Y, Wittliff JL, Schally AV (1993) Inhibition of growth of the human malignant glioma cell line (U87MG) by the steroid hormone antagonist RU486. J Clin Endocrinol Metab 77:1388–1392

63

Piroli G, Torres A, Grillo C, Lux-Lantos V, Aoki A, De Nicola AF (1998) Mechanisms in progestin antagonism of pituitary tumorigenesis. J Steroid Biochem Mol Biol 64:59–67 Richer JK, Jacobsen BM, Manning NG, Abel MG, Wolf DM, Horwitz KB (2002) Differential gene regulation by the two progesterone receptor isoforms in human breast cancer cells. J Biol Chem 277:5209–5218 Rosenfeld MG, Lunyak VV, Glass CK (2006) Sensors and signals: a coactivator/corepressor/epigenetic code for integrating signal-dependent programs of transcriptional response. Genes Dev 20:1405–1428 Rousseau A, Mokhtari K, Duyckaerts C (2008) The 2007 WHO classification of tumors of the central nervous system – what has changed? Curr Opin Neurol 21:720–727 Sager G, Orbo A, Jaeger R, Engstrom C (2003) Non-genomic effects of progestins–inhibition of cell growth and increased intracellular levels of cyclic nucleotides. J Steroid Biochem Mol Biol 84:1–8 Stening K, Eriksson O, Wahren L, Berg G, Hammar M, Blomqvist A (2007) Pain sensations to the cold pressor test in normally menstruating women: comparison with men and relation to menstrual phase and serum sex steroid levels. Am J Physiol Regul Integr Comp Physiol 293: R1711–R1716

Chapter 7

Astrocytic Tumors: Role of Carbonic Anhydrase IX Joonas Haapasalo, Hannu Haapasalo, and Seppo Parkkila

Abstract The alpha-carbonic anhydrase family consists of thirteen active isoenzymes in mammals. Their function is essential in many physiological processes, e.g., in formation of gastric juice and cerebrospinal fluid, gluconeogenesis, lipogenesis, ureagenesis, and bone resorption. CA IX is a highly active isoenzyme which is present in few normal tissues. The normal choroid plexus expresses CA IX but the other regions of the human brain show only slight expression. In contrast, CA IX is ectopically expressed in several human tumors, e.g. in colorectal cancer and breast cancer. It may increase extracellular acidification and promote the survival and migration of malignant cells. In addition, it has been proposed that CA IX is involved in achieving the resistance to chemotherapy. Here, we describe the expression of CA IX in human diffusely infiltrating astrocytomas. We also report that CA IX could serve as a useful biomarker for predicting poor prognosis of astrocytomas, and the immunostaining could be used in the diagnostic evaluation. Furthermore, CA IX could be used in future as a target molecule for therapeutic interventions. Keywords Carbonic anhydrase · CA IX expression · Choroid plexus · Astrocytoma · Astrocytic tumor

Introduction Carbonic anhydrases (CAs) are zinc-containing metalloenzymes that catalyze reversible hydration of carbon dioxide in the reaction CO2 + H2 O ⇔ + HCO− 3 + H . The carbonic anhydrase enzyme family consists of thirteen active isozymes in mammals. These include five cytoplasmic (CA I, CA II, CA III, CA VII, and CA XIII), five membrane-associated (CA IV, CA IX, CA XII, CA XIV, and CA XV), two mitochondrial (CA VA and CA VB) and one secreted (CA VI) form. The CAs have different structural and catalytic properties, and their functions are involved in many physiological processes. In addition to the classical function in the regulation of pH homeostasis, they participate, e.g., in gluconeogenesis, lipogenesis, ureagenesis, bone resorption, and formation of gastric juice and cerebrospinal fluid (Pastorekova et al., 2004). CA IX is present in few normal tissues, whereas it is ectopically expressed in many human tumors. It has been suggested that CA IX increases the capability of tumor cells to survive, invade, and achieve resistance to chemotherapy. The normal human brain tissue shows only a slight or no expression of CA IX except for the choroid plexus producing cerebrospinal fluid. In this chapter, we describe the expression of CA IX in human diffusely infiltrating astrocytomas, which are highly malignant glial tumors of the central nervous system.

Carbonic Anhydrase IX H. Haapasalo () Department of Pathology, Tampere University Hospital, FI-33521 Tampere, Finland e-mail: [email protected]

CA9 gene was first cloned by Pastorek et al. (1994) and found to encode a 466 amino acid-long protein composed of a proteoglycan domain, a central catalytic

M.A. Hayat (ed.), Tumors of the Central Nervous System, Volume 5, DOI 10.1007/978-94-007-2019-0_7, © Springer Science+Business Media B.V. 2012

65

66

CA domain, a transmembrane part, and a short COOHterminal cytoplasmic tail. Recently, Hilvo et al. (2008) reported a detailed characterization of human CA IX protein. Stopped-flow spectrophotometry experiments demonstrated that, in the excess of certain metal ions, the recombinant CA IX protein exhibits the highest catalytic activity ever measured for any CA isozyme. Investigations on the oligomerization revealed that CA IX protein forms dimers that are stabilized by intermolecular disulfide bond(s).

Expression of Carbonic Anhydrase IX in Normal Tissue When the expression of CA IX in mouse tissue was investigated, the highest immunoreactivity for CA IX was described in the gastric mucosa, moderate signals were seen in the colon and brain, and low expression was detectable in the pancreas and various segments of the small intestine (Hilvo et al., 2004). Similarly, high expression of human CA IX has been reported in the gastrointestinal tract, in which the epithelia of the gastric and gallbladder mucosa were strongly stained (Pastorek et al., 1994; Pastoreková et al., 1997; Saarnio et al., 1998b). Intensive signals have also been reported in the intestinal epithelium, the duodenum and jejunum, having stronger expression than in the more distal segments (Saarnio et al., 1998b). Moreover, CA IX expression has been shown in the mesothelium, and in the epithelial cells of the esophagus, pancreatic and biliary ducts (Turner et al., 1997; Pastoreková et al., 1997; Ivanov et al., 2001). CA IX has not been detected in the normal human brain tissue except for the weak immunostaining in the epithelial cells of the choroid plexus (Ivanov et al., 2001; Proescholdt et al., 2005). This is in accordance with the recent observations that CA IX mRNA levels are low in a normal brain compared to malignant tissues (Said et al., 2007).

Expression of Carbonic Anhydrase IX in Neoplastic Tissue CA IX is clearly the most predominant CA isozyme in various tumors. Its expression has been frequently observed in tumors derived from different segments

J. Haapasalo et al.

of the gastrointestinal tract, beginning from oral and ending to colorectal cancers (Pastoreková et al., 2006). The expression and localisation of CA IX have been examined in head and neck squamous cell carcinoma (HNSCC) (Beasley et al., 2001). The enzyme was related to the location of tumor microvessels, angiogenesis, necrosis, and tumor stage, and was considered a potential target for future therapy in HNSCC. Turner et al. (1997) suggested that tumor-associated CA IX may play a role in the proliferation and regeneration in esophageal squamous epithelium, and loss of its expression may be related to cancer progression in Barrett’s-associated adenocarcinomas. Ectopic expression is a characteristic feature for CA IX. Based on this phenomenon, CA IX is usually most highly expressed in tumors that originate from CA IX-negative tissues. Because the normal gastric mucosa contains the highest levels of CA IX among normal tissues, it was not surprising that relatively low CA IX expression was reported in gastric carcinomas (Chen et al., 2005). Interestingly, a subgroup of gastric cancers retain CA IX expression in malignant cells at the invasion front (Chen et al., 2005), suggesting that increased CA IX expression may contribute to a more advanced disease and tumor progression in a subset of gastric cancers. Many colorectal tumors also overexpress CA IX (Saarnio et al., 1998a). It seems that CA IX expression is quite heterogenous among different categories of colorectal cancer. The recent results have shown, however, that CA IX expression is most prominent in hereditary non-polyposis colorectal cancer (HNPCC) (Niemelä et al., 2007). Furthermore, its high expression in premalignant lesions has suggested that it might be a useful marker in early diagnosis of colorectal tumors (Saarnio et al., 1998a). Some CA IX positivity has been reported in both biliary and pancreatic tumors. Saarnio et al. (2001) showed that immunostaining for CA IX was mainly localised on the basolateral surface of the epithelial cells in the biliary epithelial tumors, similar to the normal biliary mucosa. The presence of CA IX in neoplastic hepatobiliary cells and its absence in hepatocellular carcinomas suggests that CA IX could be used as a marker for biliary differentiation in hepatobiliary neoplasms. The expression of CA IX in a relatively low number of malignant pancreatic tumor specimens suggests that it may have a limited value in diagnostic evaluation of pancreatic carcinoma.

7

Astrocytic Tumors: Role of Carbonic Anhydrase IX

However, the increased expression of CA IX in hyperplastic ductal epithelium may contribute to pancreatic tumorigenesis. CA IX may also serve as a valuable marker to predict the prognosis of certain cancers. Its expression, for instance, has predicted poor survival rate in lung cancer (Kim et al., 2005). The presence of CA IX has specifically been linked to the expression of proteins that are involved in angiogenesis, apoptosis inhibition, and cell-cell adhesion disruption, which explains the strong association of this enzyme with poor clinical outcome of lung cancer (Giatromanolaki et al., 2001). There are also several other publications demonstrating the association between the CA IX expression and disease prognosis. In cervical cancer, Loncaster et al. (2001) showed clinical evidence that CA IX expression correlates with the levels of tumor hypoxia, and also associates with a poor prognosis of the disease. Interestingly, it was suggested that the level of CA IX expression may be used to select patients who would benefit most from hypoxia-modification therapies or bio-reductive drugs. Breast cancer is a very common tumor type in which CA IX might be useful as a tumor biomarker. The main conclusion from several expression studies is that CA IX correlates with a poor prognosis, even though Span et al. (2007) recently pointed out that CA IX is more predictive than prognostic in this cancer type. Some human renal cancer cell lines and renal cancers have shown high CA IX expression at mRNA or protein level (McKiernan et al., 1997). Recently, Sandlund et al. (2007) assessed CA IX expression in different subtypes of renal cell cancer. They found that expression is higher in conventional cancer types rather than in other renal cell cancers, and patients with both conventional renal cell cancer and low CA IX expression had a less favourable prognosis. According to the extensive literature available on CA IX in renal cancer, the enzyme may indeed represent an ideal marker for clear cell renal cancer, as well as a promising therapeutic target for novel oncological applications, including immunotherapy and radioisotopic methods (Pastorekova et al., 2006). Recent studies have also clearly demonstrated that CA IX is highly expressed in some malignant gliomas (Haapasalo et al., 2006). The immunohistochemical results showing positive signal for CA IX in glioma cells have been confirmed at mRNA level. When CA9

67

mRNA was studied in human malignant glioma cell lines, distinct patterns of hypoxic expression of CA IX were observed (Said et al., 2007), suggesting that the low oxygen concentration is probably the driving force for the increased CA IX expression due to the presence of a hypoxia responsive element (HRE) in the CA9 promoter (Pastorekova et al., 2006). The role and prognostic significance of CA IX in malignant gliomas is also discussed in detail in another chapter of this series.

Role of Carbonic Anhydrase IX in Carcinogenesis The distribution of CA IX in tumors is related to a transcriptional activation of the CA9 gene by a hypoxia-inducible factor 1 (HIF-1), which binds to HRE as a response to low oxygen supply (Pastorekova et al., 2006). This is a common feature in solid tumors in which irregular and functionally defective tumor vasculature results in hypoxic regions. HIF-1 plays a central role by activating genes that change the expression profile of tumor cells suffering from hypoxia; thus, either leading to adaptation to the hypoxic stress or resulting in cell death. Furthermore, the surviving tumor cell population is associated with worse prognosis and resistance to anti-cancer treatment due to increasingly aggressive behaviour involving invasion and metastases. This mechanism is supported by various immunohistochemical studies in which the CA IX expression is located in perinecrotic regions of solid tumors (Haapasalo et al., 2006). CA IX isoform is also induced in cells with an inactivating mutation of the von Hippel–Lindau (VHL) tumor suppressor gene, which is a common feature e.g., in renal carcinomas. Loss of functional VHL protein causes stabilization of HIF-1, leading to concomitant up-regulation of CAs with loss of regulation by hypoxia (Pastorekova et al., 2004, 2006). Many reports have been published to assess the physiology of CA IX in neoplastic tissue. CA IX increases extracellular acidification by shifting the site of CO2 hydration from intra- to extracellular, and this process can be disturbed by inhibiting CA IX with selective sulfonamide inhibitors (Pastorekova et al., 2004, 2006). The authors suggest that this could lead to increased capability of tumor cells to survive and

68

invade. Furthermore, CA IX has also been proposed to diminish the intracellular pH gradient in the hypoxic core of three-dimensional tumor spheroids grown from cancer-derived cell lines (Swietach et al., 2008). Another potential mechanism for the function of CA IX was proposed by the finding that CA IX modulates E-cadherin mediated cell adhesion by decreasing the binding of this cell adhesion molecule to beta-catenin (Pastorekova et al., 2004, 2006). The disruption of the link between these adhesion molecules would possibly promote cell motility and invasion. Acetazolamide, a potent CA inhibitor, suppresses invasion of renal cancer cells in vitro. However, when the inhibition of CA IX in RNAi-treated breast cancer cells was studied, the invasion capacity of the cells appeared to be slightly reduced compared with the control, but this difference was not considered significant (Robertson et al., 2004).

Immunohistochemical and Analytical Methods Immunohistochemical staining for CA IX can be performed using several protocols and techniques. Several anti-human CA IX antibodies are currently available from commercial sources. However, our group has mainly used the monoclonal M75 anti-human CA IX antibody which was originally described by Pastorekova et al. (1992). This antibody has faithfully worked in both manual and automated immunostaining systems, and the signal has been easily detectable in routine formalin-fixed paraffin-embedded tissue sections. Additionally, this antibody has been excellent in rat tissue sections as well as in western blotting (Pastorekova et al., 1997; Hilvo et al., 2008). Based on our experience with CA IX immunostaining, we recommend automated staining method mainly because of an increased sensitivity, and significantly lower concentration of the antibody needed for optimal staining results. The detailed immunostaining protocols are available in our previous publications (Pastorekova et al., 1997; Saarnio et al., 1998a, b; Haapasalo et al., 2006; Niemelä et al., 2007). Briefly, the manual CA IX immunostaining has been performed according to the following procedure: (1) pretreatment of the sections with undiluted cow colostral whey (Biotop Oy, Oulu, Finland) for 30 min and wash in phosphate-buffered saline

J. Haapasalo et al.

(PBS), (2) incubation for 1 h with M75 antibody (monoclonal mouse-anti human CA IX) (1:10) in 1% bovine serum albumin (BSA)-PBS. (3) incubation for 1 h in biotinylated goat anti-mouse IgG (Zymed Laboratories Inc., South San Francisco, CA) diluted 1:300 in 1% BSA-PBS and then (4) incubation for 30 min with peroxidase-conjugated streptavidin (Zymed) diluted 1:500 in PBS. Thereafter (5) incubation for 2 min in DAB solution, containing 9 mg 3,3 diaminobenzidine tetrahydrochloride (Fluka, Buchs, Switzerland) in 15 ml PBS and 5 μl 30% H2 O2 . The sections are rinsed three times for 10 min in PBS after incubation steps 2 and 3, and four times for 5 min in PBS after step 4. The tumor sections are counterstained with haematoxylin after immunostaining.

Evaluation of Carbonic Anhydrase IX in Astrocytic Tumors Immunohistochemical evaluation of CA IX has been performed under a light microscope. Gastric mucosa can be used as a positive control. A typical CA IX immunostaining pattern includes stronger plasma membrane staining and weaker intracellular signal. In terms of the staining intensity, the scores can be evaluated as follows: 0, no reaction; +, weak reaction; ++, moderate reaction; and +++, strong reaction. A fourstep evaluation can be used in the estimation of the extent of the highest staining intensity: 0, no positive cells; +, <25 % positive cells; ++, 25–50% positive cells; and +++, >50% positive cells.

Prognostic Significance and Role of Carbonic Anhydrase IX in Astrocytomas The prognostic significance of CA IX in astrocytic tumors has been clearly documented. In tissue microarrays, CA IX immunopositivity was observed in 284 cases out of 362 (78%) astrocytic tumors (Haapasalo et al., 2006). The positive areas were often located in close proximity to necrotic regions of glioblastomas (Fig. 7.1). Weak cytoplasmic staining was occasionally seen in the neoplastic cells in the infiltrative zone of lower grade tumors. Some nuclear staining was seen

7

Astrocytic Tumors: Role of Carbonic Anhydrase IX

Fig. 7.1 CA IX immunostaining of a glioblastoma. Neoplastic cells surrounding capillaries (arrows) are negative, whereas tumor cells in border of necrotic area are strongly immunopositive (asterisk). Moderate immunopositivity is seen in some non-necrotic cells (arrow heads). (a) Magnification ×100; (b) Magnification ×200; haematoxylin counter staining

in the gliomas with higher differentiation. The CA IX cytoplasmic immunoreactivity showed a strong association with increasing tumor WHO grade. In grade 2 astrocytomas, 65% of the tumors were identified as CA IX-positive cases, whereas 73% of grade 3 astrocytomas and 82% of grade 4 astrocytomas (glioblastomas) were positive. When CA IX intensity was compared to important clinicopathologic and molecular factors, CA IX did not correlate to p53 expression, epidermal growth factor receptor amplification, apoptosis, or cell proliferation by MIB-1. However, others have described an association between CA IX expression and proliferation (Proescholdt et al., 2005; Korkolopoulou et al., 2007).

69

CA IX intensity had significant prognostic value in univariate analysis, and the immunopositive patients had a worse prognosis. The survival difference was also significant when grade 3 and grade 4 astrocytomas were evaluated separately. Most importantly, Cox multivariate survival analysis revealed that CA IX intensity, along with patient age and WHO tumor grade, was an independent prognostic factor. Our finding on the prognostic significance of CA IX in astrocytic tumors has been confirmed by Korkolopoulou et al. (2007). Increasing CA IX immunopositivity was associated with shortened survival in the entire cohort of 84 adult patients in univariate analysis. Furthermore, multivariate analysis revealed that CA IX with the tumor grade and patient age were the only parameters independently affecting the survival. CA IX expression predicted significantly the survival of patients with grades 2 and 3 tumors. They also demonstrated a perinecrotic distribution of CA IX immunostaining which increased in parallel with the extent of necrosis and histological grade. Sathornsumetee et al. (2008) reported that high expression of CA IX was associated with poor survival outcome in a patient cohort treated with bevacizumab and irinotecan. Additionally, when both CA IX and HIF-2alpha were simultaneously included in a Cox model as two separate factors, only CA IX remained as a statistically significant factor. Hypoxia-regulated protein expression, including CA IX, has been recently studied in a cohort of glioblastomas (Flynn et al., 2008). The authors found no correlation between CA IX expression and patient survival nor did they report correlation between CA IX expression and tumor grade when low-grade astrocytomas were included in the analysis. In fact, the patients with CA IX-positive tumors seemed to have worse prognosis, although this finding was not statistically significant. The reported discrepancy between this and the other studies may be related to different immunostaining methods and to a smaller number of patients investigated. The M75 antibody, used e.g., by us and Korkolopoulou et al. (2007) against human CA IX has been characterized for specificity, and it has shown no cross-reactivity with other CAs (Saarnio et al., 1998b), representing the most reliable CA IX antibody thus far. The study of Ivanov et al. (2001) showed high expression of CA IX in a large series of cancer cell lines, and also in human tumor tissues. Strong CA IX

70

immunoreactivity was reported in glioblastomas, especially in necrotic/hypoxic regions, whereas low-grade gliomas were negative. The results of Proescholdt et al. (2005) were comparable to the study described above. The strongest and most consistent staining was observed in glioblastomas and predominantly near necrotic areas. The statistical analyses exhibited a positive correlation between the CA IX expression and increasing tumor grade. In some cases, the staining was seen in most tumor cells, including those located near the blood vessels. This finding led the authors to suggest that CA IX induction in gliomas may involve hypoxia-independent mechanisms. Interestingly, it has been shown that acidosis also induces CA IX independently of pericellular hypoxia in glioblastoma cell lines (Ihnatko et al., 2006). Because CA IX is a hypoxia/necrosis marker, it can be used also as a diagnostic tool in the grading of astrocytomas. Of these tumors, necrotic neoplasms correspond to WHO grade 4 (glioblastoma). Small tumor biopsies, which contain only a small amount of tissue for diagnosis, are especially good targets for the evaluation of necrosis by CA IX immunostaining. In conclusion, CA IX is expressed abundantly in astrocytic tumors and especially in high-grade astrocytomas. This might be due to both hypoxia and high cell density, because glioblastomas represent a category of highly hypoxic and cellular tumors. CA IX may increase extracellular acidification and promote the survival and migration of malignant cells. As a result, CA IX has been shown to be a useful biomarker for predicting poor prognosis of astrocytomas and the immunostaining can also be used in the diagnostic evaluation. Furthermore, CA IX could be used as a target molecule for therapeutic interventions.

References Beasley NJ, Wykoff CC, Watson PH, Leek R, Turley H, Gatter K, Pastorek J, Cox GJ, Ratcliffe P, Harris AL (2001) Carbonic anhydrase IX, an endogenous hypoxia marker, expression in head and neck squamous cell carcinoma and its relationship to hypoxia, necrosis, and microvessel density. Cancer Res 61:5262–5267 Chen J, Röcken C, Hoffmann J, Krüger S, Lendeckel U, Rocco A, Pastorekova S, Malfertheiner P, Ebert MP (2005) Expression of carbonic anhydrase 9 at the invasion front of gastric cancers. Gut 54:920–927

J. Haapasalo et al. Flynn JR, Wang L, Gillespie DL, Stoddard GJ, Reid JK, Owens J, Ellsworth GB, Salzman KL, Kinney AY, Jensen RL (2008) Hypoxia-regulated protein expression, patient characteristics, and preoperative imaging as predictors of survival in adults with glioblastoma multiforme. Cancer 113:1032–1042 Giatromanolaki A, Koukourakis MI, Sivridis E, Pastorek J, Wykoff CC, Gatter KC, Harris AL (2001) Expression of hypoxia-inducible carbonic anhydrase-9 relates to angiogenic pathways and independently to poor outcome in nonsmall cell lung cancer. Cancer Res 61:7992–7998 Haapasalo J, Nordfors K, Hilvo M, Rantala I, Soini Y, Parkkila A-K, Pastorekova S, Pastorek J, Parkkila S, Haapasalo H (2006) Expression of carbonic anhydrase IX in astrocytic tumors predicts poor prognosis. Clin Cancer Res 12:473–477 Hilvo M, Baranauskiene L, Salzano AM, Scaloni A, Matulis D, Innocenti A, Scozzafava A, Monti SM, Di Fiore A, De Simone G, Lindfors M, Jänis J, Valjakka J, Pastoreková S, Pastorek J, Kulomaa MS, Nordlund HR, Supuran CT, Parkkila S (2008) Biochemical characterization of CA IX: one of the most active carbonic anhydrase isozymes. J Biol Chem 283:2799–2809 Hilvo M, Rafajova M, Pastorekova S, Pastorek J, Parkkila S (2004) Expression of carbonic anhydrase IX in mouse tissues. J Histochem Cytochem 52:1313–1322 Ihnatko R, Kubes M, Takacova M, Sedlakova O, Sedlak J, Pastorek J, Kopacek J, Pastorekova S (2006) Extracellular acidosis elevates carbonic anhydrase IX in human glioblastoma cells via transcriptional modulation that does not depend on hypoxia. Int J Oncol 29:1025–1033 Ivanov S, Liao SY, Ivanova A, Danilkovitch-Miagkova A, Tarasova N, Weirich G, Merrill MJ, Proescholdt MA, Oldfield EH, Lee J, Zavada J, Waheed A, Sly W, Lerman MI, Stanbridge EJ (2001) Expression of hypoxia-inducible cell-surface transmembrane carbonic anhydrases in human cancer. Am J Pathol 158:905–911 Kim SJ, Rabbani ZN, Dewhirst MW, Vujaskovic Z, Vollmer RT, Schreiber EG, Oosterwijk E, Kelley MJ (2005) Expression of HIF-1alpha, CAIX, VEGF, and MMP-9 in surgically resected non-small cell lung cancer. Lung Cancer 49: 325–335 Korkolopoulou P, Perdiki M, Thymara I, Boviatsis E, Agrogiannis G, Kotsiakis X, Angelidakis D, Rologis D, Diamantopoulou K, Thomas-Tsagli E, Kaklamanis L, Gatter K, Patsouris E (2007) Expression of hypoxia-related tissue factors in astrocytic gliomas. A multivariate survival study with emphasis upon carbonic anhydrase IX. Hum Pathol 38:629–638 Loncaster JA, Harris AL, Davidson SE, Logue JP, Hunter RD, Wycoff CC, Pastorek J, Ratcliffe PJ, Stratford IJ, West CM (2001) Carbonic anhydrase (CA IX) expression, a potential new intrinsic marker of hypoxia: correlations with tumor oxygen measurements and prognosis in locally advanced carcinoma of the cervix. Cancer Res 61:6394–6399 McKiernan JM, Buttyan R, Bander NH, Stifelman MD, Katz AE, Chen MW, Olsson CA, Sawczuk IS (1997) Expression of the tumor-associated gene MN: a potential biomarker for human renal cell carcinoma. Cancer Res 57:2362–2365 Niemelä AM, Hynninen P, Mecklin J-P, Kuopio T, Kokko A, Aaltonen L, Parkkila A-K, Pastorekova S, Pastorek J, Waheed A, Sly WS, Ørntoftm TF, Kruhøfferm M, Haapasalo H, Parkkila S, Kivelä AJ (2007) Carbonic anhydrase IX

7

Astrocytic Tumors: Role of Carbonic Anhydrase IX

is highly expressed in hereditary non-polyposis colorectal cancer. Cancer Epidemiol Biomarkers Prev 16:1760–1766 Pastorek J, Pastoreková S, Callebaut I, Mornon JP, Zelník V, Opavsky R, Zat’ovicová M, Liao S, Portetelle D, Stanbridge EJ, Závada J, Burny A, Kettman R (1994) Cloning and characterization of MN, a human tumor-associated protein with a domain homologous to carbonic anhydrase and a putative helix-loop-helix DNA binding segment. Oncogene 9:2877–2888 Pastoreková S, Parkkila S, Parkkila A-K, Opavský R, Zelnik V, Saarnio J, Pastorek J (1997) Carbonic anhydrase IX, MN/CA IX: Analysis of stomach complementary DNA sequence and expression in human and rat alimentary tracts. Gastroenterology 112:398–408 Pastoreková S, Parkkila S, Pastorek J, Supuran CT (2004) Carbonic anhydrases: current state of the art, therapeutic applications and future prospects. J Enzyme Inhib Med Chem 19:199–229 Pastoreková S, Parkkila S, Zavada J (2006) Tumor-associated carbonic anhydrases and their clinical significance. Adv Clin Chem 42:167–216 Pastoreková S, Závadová Z, Kost’ál M, Babusiková O, Závada J (1992) A novel quasi-viral agent, MaTu, is a two-component system. Virology 187:620–626 Proescholdt MA, Mayer C, Kubitza M, Schubert T, Liao SY, Stanbridge EJ, Ivanov S, Oldfield EH, Brawanski A, Merrill MJ (2005) Expression of hypoxia-inducible carbonic anhydrases in brain tumors. Neuro Oncol 7:465–475 Robertson N, Potter C, Harris AL (2004) Role of carbonic anhydrase IX in human tumor cell growth, survival, and invasion. Cancer Res 64:6160–5 Saarnio J, Parkkila S, Parkkila A-K, Haukipuro K, Pastoreková S, Pastorek J, Kairaluoma MI, Karttunen TJ (1998a) Immunohistochemical study of colorectal tumors for expression of a novel transmembrane carbonic anhydrase, MN/CA IX, with potential value as a marker of cell proliferation. Am J Pathol 153:279–285

71 Saarnio J, Parkkila S, Parkkila A-K, Pastoreková S, Haukipuro K, Pastorek J, Juvonen T, Karttunen TJ (2001) Transmembrane carbonic anhydrase, MN/CA IX, is a potential biomarker for biliary tumours. J Hepatol 35:643–649 Saarnio J, Parkkila S, Parkkila A-K, Waheed A, Casey MC, Zhou ZY, Pastoreková S, Pastorek J, Karttunen T, Haukipuro K, Kairaluoma MI, Sly WS (1998b) Immunohistochemistry of carbonic anhydrase isozyme IX (MN/CAIX) in human gut reveals polarized expression in the epithelial cells with the highest proliferative capacity. J Histochem Cytochem 46:497–504 Said HM, Staab A, Hagemann C, Vince GH, Katzer A, Flentje M, Vordermark D (2007) Distinct patterns of hypoxic expression of carbonic anhydrase IX (CA IX) in human malignant glioma cell lines. J Neurooncol 81:27–38 Sandlund J, Oosterwijk E, Grankvist K, Oosterwijk-Wakka J, Ljungberg B, Rasmuson T (2007) Prognostic impact of carbonic anhydrase IX expression in human renal cell carcinoma. BJU Int 100:556–60 Sathornsumetee S, Cao Y, Marcello JE, Herndon JE 2nd, McLendon RE, Desjardins A, Friedman HS, Dewhirst MW, Vredenburgh JJ, Rich JN (2008) Tumor angiogenic and hypoxic profiles predict radiographic response and survival in malignant astrocytoma patients treated with bevacizumab and irinotecan. J Clin Oncol 26:271–8 Span PN, Bussink J, De Mulder PH, Sweep FC (2007) Carbonic anhydrase IX expression is more predictive than prognostic in breast cancer. Br J Cancer 96:1309 Swietach P, Wigfield S, Supuran CT, Harris AL, Vaughan-Jones RD (2008) Cancer-associated, hypoxia-inducible carbonic anhydrase IX facilitates CO2 diffusion. BJU Int 101:22–4 Turner JR, Odze RD, Crum CP, Resnick MB (1997) MN antigen expression in normal, preneoplastic, and neoplastic esophagus: a clinicopathological study of a new cancer-associated biomarker. Hum Pathol 28:740–744

Chapter 8

Development of Cysts in Pilocytic Astrocytomas: Role of Eosinophilic Granular Bodies (Method) Jai-Nien Tung, Tang-Yi Tsao, Kun-Tu Yeh, Ching-Fong Liao, and Ming-Chung Jiang

Abstract Pilocytic astrocytomas are the major cystic astrocytoma found in childhood central nervous system tumor. Eosinophilic granular bodies (EGBs) are ground eosinophilic hyaline bodies observed in microcysts in pilocytic astrocytomas. EGBs have long been thought to be related to the lysosomal system. Lysosomes and their proteases play important roles in the progression of various cancers including the brain tumors. However, the presence of lysosomal proteins in EGBs in pilocytic astrocytomas and their relationship to disease progression are not fully studied. We recently showed that EGBs can be observed by immunohistochemistry with antibodies against the lysosomal membrane proteins, LAMP-1 and LAMP2, and the major lysosomal protease, cathepsin D. Eosinophilic granular bodies, together with other factors, may be involved in the development of cysts in pilocytic astrocytomas. Here we discuss the possible role of EGBs in cyst development in pilocytic astrocytomas. EGBs observation facilitates brain tumor discrimination; we also provide protocol for EGBs observation using immunohistochemistry with antibodies against LAMP-1 and LAMP-2. LAMP-1 and LAMP-2 can show distinct ring-like pattern in EGBs staining and this facilitates EGBs observation. Keywords Pilocytic astrocytoma · EGBs · LAMP-1 · LAMP-2 · lysosomal protease · GFAP

M.-C. Jiang () Section of Haematology-Oncology, Department of Medicine, Taipei Medical University and Hospital, Taipei 110, Taiwan e-mail: [email protected]

Introduction Pilocytic astrocytomas are the most common pediatric tumors of the central nervous system (CNS). Pilocytic astrocytomas are cystic tumors and cyst formation, either microcystic or macrocystic or combined, was frequently observed in the tumors (Lee et al., 1989). Cyst development in the tumors may induce increased intracranial pressure due to mass effect or ventricular obstruction and contributes to cerebral damage (Kleihues and Cavenee, 2000). On histopathology, pilocytic astrocytomas are biphasic astrocytic tumors contain alternating densely compacted bipolar cells with Rosenthal fibers and loosely textured multipolar cells with microcysts and eosinophilic granular bodies (EGBs) (Kleihues and Cavenee, 2000). Eosinophilic granular bodies, also regarded as hyaline granular bodies or eosinophilic hyaline droplets, are brightly eosinophilic round bodies of variable size in pilocytic astrocytomas and in other brain tumors such as ganglioglioma and pleomorphic xanthoastrocytoma (Takei et al., 1976; Barnard and Scott, 1980; Teo and Ng, 1998; Katsetos et al., 1994; Liberski and Kordek, 1997; Tibbetts et al., 2009). How are the EGBs generated, and what is their role in disease progression are not fully known. EGBs have long been thought to be related to the lysosomal system (Hitotsumatsu et al., 1994; Katsetos et al., 1994). However, presence of lysosomal proteins in EGBs in pilocytic astrocytomas and their relationship to the progression of the tumor are not fully studied. Lysosomes are cellular membrane bound vesicles containing various hydrolytic enzymes responsible for the degradation of most

M.A. Hayat (ed.), Tumors of the Central Nervous System, Volume 5, DOI 10.1007/978-94-007-2019-0_8, © Springer Science+Business Media B.V. 2012

73

74

J.-N. Tung et al.

cellular macromolecules. Besides degradation of cellular macromolecules, lysosomal proteases are implicated in a variety of pathological processes such as the progression of degenerative neurological disorders (Dumontel et al., 1999; Stahl et al., 2007). Cathepsin D is a major lysosomal proteinase that plays significant roles in the progression of various cancers including the brain cancers (Sivaparvathi et al., 1996; Rochefort and Liaudet-Coopman, 1999; Rochefort et al., 2000). We recently studied the contents of EGBs and analyzed their morphologies in tumor specimens from pilocytic astrocytomas. Our results show that EGBs can be observed by immunostaining with antibodies against LAMP-1, LAMP-2, and cathepsin D (Tung et al., 2010). Our studies also indicate that EGBs, together with other factors, may be involved in cyst development in pilocytic astrocytomas.

Eosinophilic Granular Bodies Staining with Antibodies Against LAMP-1, LAMP-2 and Cathepsin D We studied the presence LAMP-1, LAMP-2, and cathepsin D in the EGBs of pilocytic astrocytoma with formalin-fixed/paraffin-embedded pilocytic astrocytoma sections (Tung et al., 2010). LAMP-1 and LAMP-2 stained positively and showed peripheral rim and granular staining patterns in EGBs in pilocytic astrocytomas (Fig. 8.1a, b). In the granular staining patterns, the LAMP-1 and LAMP-2 immuno-reactive products localized either in the form of diffuse floccular densities or larger conglomerates of amorphous globular materials in EGBs (Fig. 8.1a, b). LAMP1 and LAMP-2 also stained positively in the cytoplasm of astrocytes in pilocytic astrocytomas, and the staining are much weaker in the cytoplasm of astrocytes than in EGBs (Fig. 8.1a, b). Cathepsin D staining shows a fine granular pattern in the whole bodies of EGBs in pilocytic astrocytomas (Fig. 8.1c). Cathepsin D also positively stained in the astrocytes in pilocytic astrocytomas specimens and the staining are stronger in astrocytes than in EGBs (Fig. 8.1c).

Fig. 8.1 Positive staining of eosinophilic granular bodies in pilocytic astrocytomas with antibodies against LAMP-1, LAMP-2 and cathepsin D. Immunohistochemistry with formalin-fixed/paraffin-embedded pilocytic astrocytoma sections using anti-LAMP-1 (a) and anti-LAMP-2 (b) antibodies show both peripheral rim and granular staining patterns and the staining are more pronounced in EGBs than in the cytoplasm of astrocytes in pilocytic astrocytomas. Cathepsin D (c) shows a fine granular pattern in the whole bodies of EGBs and its staining is stronger in astrocytes than in EGBs in pilocytic astrocytomas. Scale = 50 μm

8

Development of Cysts in Pilocytic Astrocytomas: Role of Eosinophilic Granular Bodies (Method)

Eosinophilic Granular Bodies Staining with PAS, H&E and Antibodies Against GFAP and CSE1L EGBs show homogeneous and finely red granular in PAS (periodic acid-Schiff) staining (Fig. 8.2a). PAS stained well in EGBs in pilocytic astrocytomas; however it may also stain in other tissues such as the loosely textured structures in pilocytic astrocytomas (Fig. 8.2a). H&E (hematoxylin and eosin) shows weak staining in granular materials distributed in areas close to the membrane of EGBs (Fig. 8.2b). H&E also positively stained in astrocytes and other structures in pilocytic astrocytomas specimens (Fig. 8.2b). Antiglial fibrillary acidic protein (GFAP) antibodies show dissimilar staining patterns in EGBs (Fig. 8.2c). Some EGBs showed homogenous patterns for GFAP staining in the whole bodies, while some EGBs showed a peripheral rim-staining pattern with anti-GFAP antibodies (Fig. 8.2c). Although GFAP is also positive for EGBs staining; the staining is weak. Moreover, GFAP also stained strongly in astrocytes and the loosely textured structures in pilocytic astrocytomas specimens;

Fig. 8.2 The patterns of eosinophilic granular bodies in pilocytic astrocytomas stained with PAS, H&E, anti-GFAP, and anti-CSE1L antibodies. PAS (a) stained homogeneous and fine granular materials in EGBs as well as in the ruffled membrane of EGBs. PAS also positively stained in the loosely textured structures in pilocytic astrocytomas. H&E (b) stained weakly

75

this makes EGBs observation difficult. CSE1L, the cellular apoptosis susceptibility protein, is also a secretory protein (Tung et al., 2009; Tsai et al., 2010). EGBs are also positive for CSE1L staining. The antiCSE1L antibodies reactive materials show a fine granular staining pattern in the whole inclusions of EGBs in pilocytic astrocytomas (Fig. 8.2d). CSE1L staining is weaker than staining with antibodies against LAMP-1 and LAMP-2. LAMP-1 and LAMP-2 can show distinct ring-like pattern in EGBs staining and this facilitates EGBs observation.

The Origin of Eosinophilic Granular Bodies and Possible Role of Eosinophilic Granular Bodies in Cyst Development in Pilocytic Astrocytomas Cell nuclei are blue in color when stained with Mayer’s hematoxylin (Gille et al., 2002). Using of Mayer’s hematoxylin to stain the EGBs in pilocytic astrocytoma specimens has shown that most EGBs have distinct nuclei (Fig. 8.3a). Some EGBs show condensed nucleus (Fig. 8.3a). Some EGBs show nucleus with

in granular materials in areas close to the membrane of EGBs. GFAP (c) shows mainly homogenous fine granular staining in the whole bodies of EGBs. GFAP also heavily stained in astrocytes in pilocytic astrocytomas. CSE1L (d) shows a fine granular staining pattern in the whole inclusions of the EGBs. Scale bar = 50 μm

76

J.-N. Tung et al.

Fig. 8.3 Immunohistochemistry with formalin-fixed/paraffinembedded pilocytic astrocytoma sections using anti-LAMP-2 antibodies show EGBs have distinct nuclei (a) and EGBs scattered in microcysts in pilocytic astrocytoma (b). (a) A representative image shows EGB connected with a condensed nucleus

(asterisk), EGB connected with a nucleus by rope-like cytoplasm remnants (arrow), and the membrane of EGB struck with a fragmented nucleus (arrowhead). (b) A representative image shows EGBs scattered in microcyst in pilocytic astrocytoma. Scale bars = 100 μm

rope-like cytoplasmic remnants (Fig. 8.3a). Some EGB membranes are stuck with a fragmented nucleus (Fig. 8.3a). There is no indication of EGBs secreted by intact astrocytes suggesting that, if not all, at least some EGBs may be derived from anucleated apoptotic astrocytoma cells. LAMP-1, LAMP-2, and cathepsin D are heavily stained in EGBs in pilocytic astrocytomas. LAMP-1 and LAMP-2 play important roles in lysosomal biogenesis and in maintaining the structural integrity of the lysosomal compartment (Eskelinen et al., 2003). Cathepsin D plays an important role in mediating cell apoptosis (Deiss et al., 1996; Roberg and Ollinger, 1998). High expression of LAMP-1 and LAMP-2 in astrocytoma cells may induce abnormally high amount of lysosome production in astrocytoma cells in pilocytic astrocytomas. The high concentration of cathepsin D and other lysosomal protease may induce apoptosis of astrocytoma cells that produced abnormally high number of lysosomes. Finally, the high expression of LAMP-1 and LAMP-2 in the apoptotic astrocytoma cells may contribute to maintain the integrity of the cytoplasmic membrane, and thus EGBs formation. Therefore, LAMP-1, LAMP-2, and cathepsin D may, to some extent, be involved in EGBs formation in pilocytic astrocytomas.

Neurological deficit and increased intracranial pressure due to mass effect or ventricular obstruction occurred frequently in patients with pilocytic astrocytoma (Kleihues and Cavenee, 2000; Buschmann et al., 2003). Pilocytic astrocytomas are generally slow-growing tumors (Haapasalo et al., 1999; Kleihues and Cavenee, 2000). Since the growth rates of the tumors are slow, there must be a relation between the growing and extension of cysts in pilocytic astrocytomas and the clinical presentations of the tumor. Pilocytic astrocytomas are frequently biphasic tumors with densely fibrillated areas rich in Rosenthal fibers and loosely arranged, often microcystic areas, in which EGBs can be found (Rao et al., 2005; Malik et al., 2006; White et al., 2008). EGBs are often presented in microcysts in pilocytic astrocytomas; Barnard and Scott reported that microcysts in pilocytic astrocytomas are frequently lined by intensely eosinophilic granular bodies of irregular but generally globular outline (Barnard and Scott, 1980). We have also noted that EGBs are scattered widely across cysts in pilocytic astrocytomas and in many areas which EGBs aggregated in clusters are close to areas of fluid in cysts (Fig. 8.3b). Lysosomes and lysosomal proteinases including cathepsin D are capable of destroying the

8

Development of Cysts in Pilocytic Astrocytomas: Role of Eosinophilic Granular Bodies (Method)

extracellular matrix of connective tissue and participating in cyst development during the progression of various diseases (Birek et al., 1980; Gerasimov et al., 1991; Hazama et al., 1995; Cooper, 2002). Since cathepsin D, and possibly other lysosomal proteinases, is present in EGBs in the microcysts of pilocytic astrocytomas, the extracellular matrix-digestive abilities of cathepsin D and other lysosomal proteinases may participate in extracellular matrix digestion in the surrounding brain tissue. Therefore, the presence of eosinophilic granular bodies in the cysts of pilocytic astrocytomas may, to some extent, be involved in cyst development of the tumor.

Method of Eosinophilic Granular Bodies Observation with Immunohistochemistry Using Anti-LAMP-1 and Anti-LAMP-2 Antibodies EGBs observation facilitates brain tumor discrimination. LAMP-1 and LAMP-2 can show distinct ring-like pattern in EGBs staining and this facilitates EGBs observation. We provide method of EGBs observation with immunohistochemistry using anti-LAMP-1 and anti-LAMP-2 antibodies.

Protocol 1. Fix tumor tissues in formalin and embed in paraffin blocks according to standard procedures. 2. Cut 6-μm thick tumor sections with a microtome and apply tumor sections to glass slides. 3. Place the slides in a 56–60◦ C oven for 15 min. 4. Deparaffinization the slides with xylenes for 5 min three times. 5. Rehydration with graded ethanol by washing the slides in 100% ethanol for 5 min two times, then in 90% ethanol for 5 min two times, and then in 80% ethanol for 5 min two times. Finally wash the slides with deionized water for 30 s. 6. Place the slides in PBS (pH 7.4) for 30 min for further rehydration.

77

7. Immerse the slides in a microwave-resistant plastic containing 10 mM citrate buffer (pH 6.0) and operate the microwave oven for 10 min on high power for antigen retrieval. 8. Allow slides to cool in the buffer for at least 20 min. Wash slides in deionized H2 O three times for 2 min each. Aspirate excess liquid from slides. 9. The following steps are carried out at room temperature in a humidified chamber. Do not allow the sections to dry out at any time. 10. Add enough drops of 3% hydrogen peroxide to cover the whole section and incubate for 5 min at room temperature to quench endogenous peroxidase activity. Wash slides in PBS twice for 5 min each. 11. Incubate the specimens with 1.5% normal blocking serum in PBS or with 5% BSA in PBS for 1 h. 12. Drain excess fluid from the slides, covering each tumor section with 100 μl anti-LAMP-1 or antiLAMP-2 antibodies and incubated for 1 h at room temperature or overnight at 4◦ C. Optimal antibody concentration should be determined by titration. Anti-LAMP-1 (H228) and anti-LAMP-2 (H207) from Santa Cruz (Santa Cruz Biotechnology, Santa Cruz, CA, USA) at a dilution of 1:50 in PBS work well in our hands. 13. Wash with three changes of PBS for 5 min each. 14. Incubate for 30 min with biotin-conjugated secondary antibodies. The secondary antibodies are proper diluted with 1.5% normal blocking serum in PBS with. Wash with three changes of PBS for 5 min each. 15. Incubation for 30 min with avidin biotin enzyme reagent. Wash with three changes of PBS for 5 min each. 16. Incubation in peroxidase substrate for 1–10 min or until desired stain intensity develops. Wash slides with deionized H2 O for 5 min. 17. Counterstained the slides with Mayer’s hematoxylin for 30 s to 5 min. Wash slides with several changes of deionized H2 O. 18. Dehydrate through alcohols and xylenes according to standard procedures. Cover tumor section with a glass coverslip and observe by light microscopy.

78

Discussion Observation for the presence of EGBs is useful for diagnosing the CNS tumors and this may be beneficial for making therapeutic decisions. For example, pilomyxoid astrocytomas share similar features with pilocytic astrocytomas, yet display subtle histological differences. The pilomyxoid astrocytomas lack biphasic appearance, Rosenthal fibers, and eosinophilic granular bodies that are commonly observed pilocytic astrocytomas (Tihan et al., 1999; Buccoliero et al., 2008; Amatya et al., 2009; Azad et al., 2010). Pilomyxoid astrocytomas are thought to be more aggressive with more frequent local recurrence as well as cerebrospinal spread (Tihan et al., 1999; Buccoliero et al., 2008; Amatya et al., 2009; Azad et al., 2010). Therefore, recognition of pilomyxoid astrocytoma and its distinction from pilocytic astrocytoma is very important. Observation for the presence of EGBs is therefore useful for pilomyxoid astrocytoma and pilocytic astrocytoma discrimination and being valuable for therapeutic strategy determination. Cyst development in pilocytic astrocytomas may contribute to increased intracranial pressure due to mass effect and thereby leading to cerebral damage and neurological disorder of the patients. The finding that eosinophilic granular bodies may be involved in the development of cysts in pilocytic astrocytomas may provide new insights into the role of eosinophilic granular bodies in pilocytic astrocytomas, and this may in turn be valuable for the treatment of pilocytic astrocytomas.

References Amatya VJ, Akazawa R, Sumimoto Y, Takeshima Y, Inai K (2009) Clinicopathological and immunohistochemical features of three pilomyxoid astrocytomas: comparative study with 11 pilocytic astrocytomas. Pathol Int 59:80–85 Azad S, Kudesia S, Chawla N, Azad R, Singhal M, Rai SM, Arora P (2010) Pilomyxoid astrocytoma. Indian J Pathol Microbiol 53:294–296 Barnard RO, Scott T (1980) A note on the nature of eosinophilic granular bodies in astrocytic gliomas. Acta Neuropathol 50:245–247 Birek P, Wang HM, Brunette DM, Melcher AH (1980) Epithelial rests of Malassez in vitro. Phagocytosis of collagen and the possible role of their lysosomal enzymes in collagen degradation. Lab Invest 43:61–72

J.-N. Tung et al. Buccoliero AM, Gheri CF, Maio V, Moncini D, Castiglione F, Garbini F, Sanzo M, Taddei A, Genitori L, Taddei GL (2008) Occipital pilomyxoid astrocytoma in a 14-year-old girl-case report. Clin Neuropathol 27:373–377 Buschmann U, Gers B, Hildebrandt G (2003) Pilocytic astrocytomas with leptomeningeal dissemination: biological behavior, clinical course, and therapeutical options. Childs Nerv Syst 19:298–304 Cooper JB (2002) Aspartic proteinases in disease: a structural perspective. Curr Drug Targets 3:155–173 Deiss LP, Galinka H, Berissi H, Cohen O, Kimchi A (1996) Cathepsin D protease mediates programmed cell death induced by interferon-gamma, Fas/APO-1 and TNF-alpha. EMBO J 15:3861–3870 Dumontel C, Rousselle C, Guigard MP, Trouillas J (1999) Angulate lysosomes in skin biopsies of patients with degenerative neurological disorders: high frequency in neuronal ceroid lipofuscinosis. Acta Neuropathol 98:91–96 Eskelinen EL, Tanaka Y, Saftig P (2003) At the acidic edge: emerging functions for lysosomal membrane proteins. Trends Cell Biol 13:137–145 Gerasimov AM, Toporova SM, Furtseva LN, Berezhnoy AP, Vilensky EV, Alekseeva RI (1991) The role of lysosomes in the pathogenesis of unicameral bone cysts. Clin Orthop Relat Res 266:53–63 Gille J, Ehlers EM, Okroi M, Russlies M, Behrens P (2002) Apoptotic chondrocyte death in cell-matrix biocomposites used in autologous chondrocyte transplantation. Ann Anat 184:325–332 Haapasalo H, Sallinen S, Sallinen P, Helén P, Jääskeläinen J, Salmi TT, Paetau A, Paljärvi L, Visakorpi T, Kalimo H (1999) Clinicopathological correlation of cell proliferation, apoptosis and p53 in cerebellar pilocytic astrocytomas. Neuropathol Appl Neurobiol 25:134–142 Hazama F, Chue CH, Kataoka H, Sasahara M, Amano S (1995) Pathogenesis of lacuna-like cyst formation and diffuse degeneration of the white matter in the brain of strokeprone spontaneously hypertensive rats. Clin Exp Pharmacol Physiol Suppl 22:S260–S261 Hitotsumatsu T, Iwaki T, Fukui M, Tateishi J (1994) Cytoplasmic inclusions of astrocytic elements of glial tumors: special reference to round granulated body and eosinophilic hyaline droplets. Acta Neuropathol 88:501–510 Katsetos CD, Krishna L, Friedberg E, Reidy J, Karkavelas G, Savory J (1994) Lobar pilocytic astrocytomas of the cerebral hemispheres: II. Pathobiology – morphogenesis of the eosinophilic granular bodies. Clin Neuropathol 13:306–314 Kleihues P, Cavenee WK (2000) Pathology and genetics of tumors of the nervous system (WHO). International Agency for Research on Cancer (IARC) Press, Lyon, pp 45–51 Lee YY, Van Tassel P, Bruner JM, Moser RP, Share JC (1989) Juvenile pilocytic astrocytomas: CT and MR characteristics. Am J Roentgenol 152:1263–1270 Liberski PP, Kordek R (1997) Ultrastructural pathology of glial brain tumors revisited: a review. Ultrastruct Pathol 21:1–31 Malik A, Deb P, Sharma MC, Sarkar C (2006) Neuropathological spectrum of pilocytic astrocytoma: an Indian series of 120 cases. Pathol Oncol Res 12:164–171 Rao RD, Brown PD, Giannini C, Maher CO, Meyer FB, Galanis E, Erickson BJ, Bucker JC (2005) Central nervous system tumors. In: Chang AE, Ganz PA, Hayes DF, Kinsella T, Pass

8

Development of Cysts in Pilocytic Astrocytomas: Role of Eosinophilic Granular Bodies (Method)

HI, Schiller JH, Stone RM, Strecher V (eds) Oncology: an evidence-based approach. Springer, New York, NY, p 495 Roberg K, Ollinger K (1998) Oxidative stress causes relocation of the lysosomal enzyme cathepsin D with ensuing apoptosis in neonatal rat cardiomyocytes. Am J Pathol 152:1151–1156 Rochefort H, Garcia M, Glondu M, Laurent V, Liaudet E, Rey JM, Roger P (2000) Cathepsin D in breast cancer: mechanisms and clinical applications, a 1999 overview. Clin Chim Acta 291:157–170 Rochefort H, Liaudet-Coopman E (1999) Cathepsin D in cancer metastasis: a protease and a ligand. APMIS 107:86–95 Sivaparvathi M, Sawaya R, Chintala SK, Go Y, Gokaslan ZL, Rao JS (1996) Expression of cathepsin D during the progression of human gliomas. Neurosci Lett 208:171–174 Stahl S, Reinders Y, Asan E, Mothes W, Conzelmann E, Sickmann A, Felbor U (2007) Proteomic analysis of cathepsin B- and L-deficient mouse brain lysosomes. Biochim Biophys Acta 1774:1237–1246 Takei Y, Mirra SS, Miles ML (1976) Eosinophilic granular ceels in oligodendrogliomas: an ultrastructural study. Cancer 38:1968–1976 Teo JG, Ng HK (1998) Cytodiagnosis of pilocytic astrocytoma in smear preparations. Acta Cytol 42:673–678

79

Tibbetts KM, Emnett RJ, Gao F, Perry A, Gutmann DH, Leonard JR (2009) Histopathologic predictors of pilocytic astrocytoma event-free survival. Acta Neuropathol 117:657–665 Tihan T, Fisher PG, Kepner JL, Godfraind C, McComb RD, Goldthwaite PT, Burger PC (1999) Pediatric astrocytomas with monomorphous pilomyxoid features and a less favorable outcome. J Neuropathol Exp Neurol 58:1061–1068 Tsai CS, Chen HC, Tung JN, Tsou SS, Tsao TY, Liao CF, Chen YC, Yeh CY, Yeh KT, Jiang MC (2010) Serum cellular apoptosis susceptibility protein is a potential prognostic marker for metastatic colorectal cancer. Am J Pathol 176:1619–1628 Tung MC, Tsai CS, Tung JN, Tsao TY, Chen HC, Yeh KT, Liao CF, Jiang MC (2009) Higher prevalence of secretory CSE1L/CAS in sera of patients with metastatic cancer. Cancer Epidemiol Biomarkers Prev 18:1570–1577 Tung JN, Tsao TY, Tai CJ, Yeh KT, Cheng YW, Jiang MC (2010) Distribution of LAMP-1, LAMP-2, and cathepsin D in eosinophilic granular bodies: possible relationship to cyst development in pilocytic astrocytomas. J Int Med Res 38:1354–1364 White JB, Piepgras DG, Scheithauer BW, Parisi JE (2008) Rate of spontaneous hemorrhage in histologically proven cases of pilocytic astrocytoma. J Neurosurg 108:223–226

Chapter 9

Role of Synemin in Astrocytoma Cell Migration Quincy Quick, Yihang Pan, and Omar Skalli

Abstract Synemin is an intermediate filament (IF) protein with the unique property to incorporate into IF networks while also binding to actin-associated proteins such as α-actinin, vinculin, zyxin, and α-dystrobrevin. Interestingly, synemin is present in astrocytomas of all grades but not in normal, mature astrocytes. Synemin contribution to the malignant behavior of astrocytoma cells was explored through RNAi experiments that established that synemin is a positive regulator of astrocytoma cell motility. In support of this role is the abundance of synemin in cytoplasmic domains important for motility, such as leading edges and lamellipodias. Synemin down-regulation also disrupted actin organization and increased the proportion of unpolymerized actin. In addition, antagonizing synemin decreased the proportion of α-actinin associated with actin filaments, providing evidence that synemin interaction with α-actinin may influence actin dynamics upon which motility ultimately depends. In addition to synemin, astrocytoma cells express two other IF proteins, vimentin and nestin. These latter two proteins contribute to the motility of carcinoma and/or fibroblastic cells, raising the possibility that they function similarly in astrocytoma cells. However, in contrast to synemin, vimentin and nestin are not present in the leading edge of astrocytoma cells, suggesting that they influence motility through mechanism(s) distinct from that of synemin. Thus, antagonizing synemin function appears to diminish the motility of astrocytoma cells

O. Skalli () Department of Cellular Biology and Anatomy, Louisiana State University Medical Center, Shreveport, LA 71103, USA e-mail: [email protected]

and future studies will determine whether antagonizing synemin can be accomplished with small molecules such as those being developed to target other IF proteins. Keywords Synemin · Cytoskeletal Vimentin · Nestin · GFAP

protein

·

Introduction The high rate of recurrence of astrocytomas results from two major factors, namely the invasion of normal brain tissue by stray astrocytoma cells prior to surgical removal of the main tumor mass and the refractoriness of these invading cells to conventional therapeutic modalities performed after the surgery. It is therefore important to unravel the mechanisms underlying the high motility and invasive properties of astrocytoma cells. Since these two cellular behaviors are ultimately determined by the cytoskeleton, the identification of cytoskeletal proteins differentially expressed in astrocytoma cells versus normal glial cells may provide crucial clues regarding the mechanisms responsible for the high motility of astrocytoma cells. Recently, our laboratory has demonstrated that synemin is such a cytoskeletal protein. In this chapter, we will review the current knowledge about how synemin and related cytoskeletal proteins contribute to the motility and other malignant properties of astrocytoma cells. Synemin is one of ∼70 intermediate filament (IF) proteins. These proteins share a tripartite structure consisting of a central α-helical rod domain of 310 amino-acids flanked by N- and C-terminal non-helical domains of variable size (Omary, 2009). Six types of IF proteins have been distinguished based on structural

M.A. Hayat (ed.), Tumors of the Central Nervous System, Volume 5, DOI 10.1007/978-94-007-2019-0_9, © Springer Science+Business Media B.V. 2012

81

82

features of the rod domain and on the intron/exon organization of the encoding genes. Generally, each class of IF proteins is specific for a type of tissue or cell (Omary, 2009). For instance, type I and II IF proteins are confined to epithelia, type III IF proteins are mostly present in mesenchymal cells, with the important exception of the astrocyte-specific GFAP while type IV IF proteins are neuron-specific. IFs perform classical cytoskeletal functions, such as maintenance of cell shape and determination of the mechanical properties of cells and tissues (Omary, 2009). In addition, it is increasingly recognized that IF proteins participate in signal transduction by interacting with signaling proteins in a phosphorylation-dependent manner. For instance, it has been particularly well established that keratins reversibly bind to 14-3-3 proteins to modulate the proliferation and survival of hepatocytes and keratinocytes (Kim and Coulombe, 2007; Omary, 2009). In the case of keratinocytes, the genetic ablation of keratin 17 also abolished the interaction between keratin 17 and 14-3-3σ (Kim and Coulombe, 2007). As a consequence, keratin 17 deficient keratinocytes exhibited reduced levels of the pro-survival protein Akt and of its down-stream effector, the mammalian target of rapamycin and this resulted in decreased protein synthesis (Kim and Coulombe, 2007). A further example of the ability of IF proteins to facilitate signal propagation was observed in injured axons, where vimentin bound to and promoted the spatial translocation of activated MAP kinases, such as Erk1 and Erk2 (Perlson et al., 2005). Altogether, these findings reveal that IF proteins not only impact cellular activities and properties dominated by the cytoskeleton, such as motility and shape, but also participate in signaling pathways affecting cell proliferation and survival. Therefore, changes in the expression of IF proteins during the malignant transformation, such as those occurring for synemin in astrocytomas, have the potential to considerably contribute to malignant behavior.

Synemin: An Unusual IF Protein Present in Astrocytoma Cells Synemin, from the Greek syn (with) and nema (filament), was initially characterized as a high molecular weight (∼220 kDa) protein associated with IFs

Q. Quick et al.

in muscle cells (Granger and Lazarides, 1980). DNA sequencing eventually revealed that synemin belongs to the IF protein family because it comprises a α-helical central domain of 310 amino-acids with structural features common to all IF proteins (Bellin et al., 1999). However, synemin N- and C-terminal domains are ∼10 times shorter and ∼10 times larger than the ∼100 amino-acid long N- and C-terminal domains of most IF proteins, respectively. Sequence comparison and analysis of the exon/intron synemin gene organization support a model in which synemin, together with nestin and its non-mammalian orthologs, evolved from neurofilament proteins (Guérette et al., 2007). Consequently, synemin is usually classified together with nestin as a type VI-IF protein, although it is also sometimes considered a type IV IF protein along with neurofilament proteins. The human synemin gene comprises five exons and four introns and is located on chromosome 15q26 (Xue et al., 2004). Synemin is one of the few IF proteins to exhibit alternative splice variants. The largest isoform, α-synemin, uses all exons and intron IV to generate a 220 kDa protein, while β-synemin (170 kDa), initially called desmuslin (Mizuno et al., 2001), is translated from all five exons and does not contain any intronic sequences. The smallest isoform, L-synemin, arises from alternative splicing via removal of exon 4 and the four intronic sequences (Xue et al., 2004), consequently merging exons 3 and 5. This creates a new reading frame that produces an L-synemin specific sequence of 36 and 4 aminoacids in humans and mice, respectively. In addition, because synemin exon 4 encodes most of the Cterminal domain, L-synemin is much smaller (41 kDa) than α- and β-synemin and consists mostly of the rod domain. Another characteristic distinguishing synemin from most IF proteins, which are distributed in a cell or tissue type-specific manner, is its diverse expression in a variety of cells and tissues, including smooth and striated muscle cells (Granger and Lazarides, 1980), lens tissue, endothelial cells and hepatic stellate cells (Schmitt-Graeff et al., 2006). In addition, synemin is developmentally regulated in the central nervous system, where it is present in GFAP-positive astrocyte precursor cells in the embryonic and post-natal rat brain (Sultana et al., 2000). However, in the adult rat and human brain, astrocytes are negative for all synemin isoforms (Sultana et al., 2000; Jing et al., 2005),

9

Role of Synemin in Astrocytoma Cell Migration

albeit L-synemin is present in specialized neurons (Izmiryan et al., 2010). Synemin is also regulated in several pathologies of the central nervous system, as seen in the induction of α- and β-synemin in reactive astrocytes that appear in response to physical or metabolic trauma (Jing et al., 2005, 2007). Furthermore, α- and β-synemin are expressed in astrocytoma cells in grade I to IV tumors (Jing et al., 2005) (Fig. 9.1) and in established human astrocytoma cell lines (Jing et al., 2005). Considering the expression profile of synemin during glial development, the presence of synemin in astrocytoma cells may indicate the reversion of these cells to a stage of lesser differentiation which might contribute to a role for synemin in the development of this neoplasm.

Importance of Synemin for Astrocytoma Cell Motility The occurrence of synemin in astrocytoma cells, but not in normal astrocytes, also suggests that synemin contributes to the malignant properties of astrocytoma cells, such as invasion and motility. This notion was tested with Boyden chamber assays on three human astrocytoma cell lines and the results demonstrated that synemin down-regulation with shRNAs inhibited astrocytoma cell motility by 60–80% (Pan et al., 2008). Further demonstration that synemin is a positive regulator of astrocytoma cell motility was provided with siRNAs and scrape wound assays, which revealed that astrocytoma cells treated with synemin siRNAs were significantly less apt at closing the wound edges when compared to cells incubated with non-target siRNAs (Fig. 9.2).

Fig. 9.1 Immunohistochemistry with antibodies against α- and β-synemin demonstrates that in the normal human brain synemin is present only in vascular smooth muscle cells (arrows) but that it is expressed by most tumors cells in glioblastoma multiforme tissue. Bar = 10 μm

83

Of note, the role of synemin in human cancers is not limited to astrocytomas, as demonstrated by studies showing that synemin is expressed in a subset of human carcinomas (Pan et al., 2008; Schmitt-Graeff et al., 2006; Sun et al., 2010) and by RNAi experiments that demonstrated that synemin down-regulation impeded the motility of HeLa cervical carcinoma cells (Sun et al., 2010). Altogether, these results establish that synemin is not a ubiquitous component of the cytoskeleton of malignant cells but that when present it promotes the motility of these cells. The involvement of synemin in the motility of astrocytoma cells may be attributed to its subcellular distribution (Jing et al., 2005). In these cells, a pool of synemin distributes along the vimentin and/or GFAP IF network that appears to radiate from the nucleus through the cytoplasm (Fig. 9.3a). Synemin incorporates into this IF network by copolymerizing with vimentin (Bellin et al., 1999; Jing et al., 2007). However, unlike GFAP and vimentin, synemin is also included in cytoplasmic domains important for motility such as leading edges and ruffled membranes (Jing et al., 2005) (Fig. 9.3a). The targeting of synemin to these regions may be accomplished through the interaction of synemin with the actin binding protein, α-actinin, an established player in cell motility. Synemin and α-actinin co-localize at the leading edge and form a complex that can be immunoprecipitated from the cytoplasm of astrocytoma cells (Jing et al., 2005). In addition, experiments with purified proteins demonstrated direct binding of synemin to α-actinin (Bellin et al., 1999). Collectively, synemin’s interaction with α-actinin may provide a foundation for the molecular mechanism by which synemin participates in astrocytoma

84

Q. Quick et al.

Fig. 9.2 Scrape-wound (a) and Boyden chamber (b) motility assays demonstrate the impact of synemin down-regulation with siRNAs on the motility of U-373 MG human glioblastoma cells. In both types of assays, the cells were stained by Coomassie blue and visualized by light microscopy. (a) In the scrape wound assay, synemin siRNAs significantly impaired the ability of U373 MG cells to fill the wound gap. (b) Micrographs showing

cells that migrated through 8 μm size pores onto the lower aspect of a Boyden chamber. Note that the number of migrating cells is clearly lower after treatment with synemin siRNAs compared to control siRNAs. Experimental conditions were similar to those detailed in Pan et al. (2008) for cells treated with synemin shRNAs. Bar = 100 μm

cell motility. Significantly, synemin down-regulation decreases the association of α-actinin with actin filaments, suggesting that synemin binding to α-actinin stabilizes this association (Pan et al., 2008). By stabilizing the association of α-actinin with actin filaments, synemin may in turn affect actin dynamics and cell migration. This is suggested by the finding that synemin down-regulation increases the amount of depolymerized actin in astrocytoma cells (Pan et al., 2008) and alters the overall organization of the actin cytoskeleton (Fig. 9.3b). Synemin interactions with actin-associated proteins are not limited to α-actinin but also include the focal contact proteins vinculin, zyxin, and talin, as well as α-dystrobrevin and dystrophin, which are instrumental in regulating cell adhesion and migration (e.g. Mizuno et al., 2001; Sun et al., 2010). Thus, synemin interactions with focal contact proteins further represents

a potential means by which synemin participates in motility. However, synemin down-regulation did not affect the adhesion of astrocytoma cells to the substratum (Pan et al., 2008). This contrasts with observations in HeLa cervical carcinoma cells, as in these cells adhesion to the substratum was decreased after synemin RNAi, possibly through a zyxin-dependent mechanism (Sun et al., 2010). Altogether, these findings strongly support the notion that, when expressed in cancer cells, synemin substantially contributes to their motility by affecting actin dynamics through the interaction of synemin with either α-actinin and/or focal contact proteins. RNAi studies also revealed that, in addition to participating in astrocytoma cell motility, synemin is a potent positive regulator of the proliferative capacity of astrocytoma cells (Pan et al., 2008). Thus, synemin expression favors two key aspects of the malignant

9

Role of Synemin in Astrocytoma Cell Migration

85

Fig. 9.3 Staining of U-373 MG human glioblastoma cells with anti-synemin and anti-vimentin (a) or with fluorescent phalloidin to label actin filaments (b). (a) Staining of U-373 MG cells demonstrates that synemin has a filamentous distribution that coincides with that of the vimentin IF network. However, unlike vimentin, synemin also stains leading edges and ruffled

membranes (arrows). (b) In U-373 MG cells treated with control siRNAs, actin concentrates at the leading edge which brightly stains with fluorescent phalloidin (arrowheads). In contrast, after treatment with synemin siRNAs, submembranous actin staining is minimal (arrows). Bar = 10 μm

behavior of astrocytoma cells: motility and proliferation. How synemin impacts the proliferation of astrocytoma cells remains to be determined. However, the finding that in muscle cells synemin binds to the RII regulatory subunit of kinase A (Russell et al., 2006) raises the possibility that synemin influences cell proliferation by interacting with components of signaling cascades.

to favor the motility of carcinoma cells, which may explain why their presence in carcinomas is associated with a poorer prognosis when compared to vimentin and/or nestin negative carcinomas. The connection between nestin and cell motility was demonstrated by RNAi in nestin-positive prostate carcinoma cells. In these cells, nestin down-regulation markedly inhibited in vitro migration and invasion and led to a 5-fold reduction in metastasis in animal models (Kleeberger et al., 2007). A positive link between vimentin and motility has been established in a number of systems, including scrape wound closure by vimentin-positive human mammary epithelial cells (Gilles et al., 1999), directional migration towards chemo-attractive stimuli in vimentin–/– fibroblasts and skin wound healing in vimentin–/– mice (Eckes et al., 2000). This latter model revealed that vimentin is important for fibroblasts to

Role of IF Proteins Other Than Synemin in Cell Motility Synemin is one of the four IF proteins that may be present in astrocytoma cells, the other ones being GFAP, vimentin, and nestin. The role of nestin and vimentin in astrocytoma cell motility remains to be investigated. These two IF proteins, however, appear

86

generate the traction forces that are essential for cell migration. Vimentin–/– mice were also instrumental in outlining vimentin contribution in the transendothelial migration of lymphocytes (Nieminen et al., 2006). Finally, vimentin was recently demonstrated to be essential for the epithelial–mesenchymal transition, the process by which epithelial cells acquire a mesenchymal shape as they become increasingly motile (Mendez et al., 2010). Several studies have suggested that vimentin participates in cell motility by affecting the dynamics of focal contacts. In endothelial cells, vimentin regulated the size of focal contacts as the cells adapted to shear stress (Tsuruta and Jones, 2003). At the molecular level, vimentin affects focal contact dynamics by contributing to the anchorage of adhesion proteins to these sites (Nieminen et al., 2006) and/or by modulating integrin trafficking (Ivaska et al., 2005; Bhattacharya et al., 2009). In several cell types, including glioblastoma cells (Fortin et al., 2010), vimentin regulates integrin β1 trafficking by interacting with protein kinase Cε (Ivaska et al., 2005) and also possibly with the actin associated protein fimbrin (Kim et al., 2010). In addition to vimentin, nestin, and synemin, astrocytoma cells may also express a fourth IF protein, GFAP. In contrast to vimentin, nestin, and synemin, which are not present in normal astrocytes, GFAP is abundant in normal astrocytes and its protein level is comparatively lower in astrocytoma cells (for review see: Rutka et al., 1997). Several experimental studies have demonstrated that high GFAP levels correlate with low motility (for review see: Rutka et al., 1997). Therefore, the low GFAP protein level in astrocytoma cells relative to normal astrocytes are considered to release this inhibitory effect, thereby promoting the motility of astrocytoma cells relative to normal astrocytes (for review see: Rutka et al., 1997). This further demonstrates that the differences in the IF protein composition of normal astrocytes versus astrocytoma cells represents an important factor in the acquisition of a highly motile phenotype.

Prospects for Synemin and Other IF Proteins in Astrocytoma Therapy The usefulness of cytoskeletal proteins as targets for cancer treatment is evidenced by molecules that disrupt microtubules dynamics because these drugs are used

Q. Quick et al.

for the clinical treatment of human cancers, including invasive astrocytomas. However, the clinical effectiveness of microtubules disrupting compounds such as vincristine, colchicine, and taxol is limited by their high neuro- and hemato-toxicity and by their affinity for the resistance factor, P-glycoprotein (Katsetos et al., 2009). These caveats highlight the need to investigate whether other cytoskeletal proteins may represent therapeutic targets for the treatment of astrocytomas. The evidence reviewed above supports the notion that synemin may be such a target because it is intricately involved with the migration and proliferation of astrocytoma cells and because it is expressed in astrocytomas but not in normal brain tissue. The development of drugs targeting IFs has lagged behind that of drugs directed against microtubules or microfilaments. Recently, however, it has been reported that the active polyphenol in green tea, epigallocatechin gallate (EGCG), and the natural product derivative, withaferin A, bind vimentin and GFAP IFs (Ermakova et al., 2005; Bargagna-Mohan et al., 2007). Importantly, these two compounds display antitumorigenic effects on several human cancers, including astrocytomas (Shah et al., 2009; Das et al., 2010), via the promotion of tumor cell death and the inhibition of metastasis and cell proliferation. Interestingly, the mode of action of these agents may be IF protein specific. This is supported by evidence that withaferin A enhanced GFAP expression in glioblastoma cells (Shah et al., 2009), while invoking vimentin cleavage in soft tissue sarcoma cells (Lahat et al., 2010). The identification of EGCG and withaferin A as agents that bind IFs and antagonize tumor cell behavior provides a foundation for the development of synemin specific inhibitors that could obstruct synemin function in the malignant properties of astrocytomas. It should be mentioned that several phosphatase inhibitors such as calyculin A, fostriescin, and okadaic acid promote the disruption of the IF network due to the hyperphosphorylation of IF proteins (Hyder et al., 2008). Furthermore, the phosphatase inhibitor orthovanadate was shown to cause the breakdown of the keratin network and to negatively affect the ability of 14-3-3 proteins to bind keratins (Strnad et al., 2002), suggesting that phosphatase inhibition antagonizes IFs involvement in pro-survival signaling cascades. Phosphatase inhibitors, however, cannot be considered true IF drugs as they do not bind to IF

9

Role of Synemin in Astrocytoma Cell Migration

proteins and have effects resulting from the hyperphosphorylation of many non-IF proteins. Nonetheless, the potent impact that IF protein hyperphosphorylation has on the assembly of IF networks suggest that compounds that may protect from dephosphorylation specific IF phosphorylation sites could be used to selectively manipulate the function of IF proteins.

Conclusion The potential of IF proteins to influence the malignant behavior of astrocytoma cells is indicated by the substantial differences existing between the IF protein composition of normal astrocytes and astrocytoma cells. While the former cells express primarily GFAP, the latter also express synemin, vimentin and nestin; in addition the GFAP content of astrocytoma cells is lower than that of normal astrocytes. Collectively, studies on the IF proteins present in astrocytoma cells suggest that each of these proteins may uniquely contribute to the acquisition of the highly motile properties characteristic of these cells. In this respect, however, synemin stands out when compared to nestin, vimentin and GFAP because it is the only IF protein expressed in astrocytoma cells to be present in leading edges and ruffled membranes. This suggests that synemin plays a direct role in the motility of astrocytoma cells. Together with the binding sites that synemin possesses for several actin associated proteins, this supports a mechanism by which synemin positively impacts the motility of astrocytoma cells by modulating actin dynamics, a process which is fundamental to motility. Further work is needed to understand at the molecular level the unique crosstalk between the actin cytoskeleton and the IF protein synemin. Synemin is also remarkable compared to other IF proteins expressed in astrocytomas because it profoundly influences the proliferation of astrocytoma cells, through mechanisms yet to be identified. The supportive roles of synemin in astrocytoma cell motility and proliferation coupled with synemin expression in astrocytoma cells but not in normal glia suggest that synemin is a valuable target for drug development. Although the development of drugs specific for IF proteins is at its infancy, the recent identification of compounds affecting vimentin and GFAP raises the exciting prospect that in the near future small

87

molecules abrogating synemin function in astrocytoma cells may be identified and evaluated therapeutically.

References Bargagna-Mohan P, Hamza A, Kim YE, Khuan A, Ho Y, MorVaknin N, Wendschlag N, Liu J, Evans RM, Markovitz DM, Zhan CG, Kim KB, Mohan R (2007) The tumor inhibitor and antiangiogenic agent withaferin A targets the intermediate filament protein vimentin. Chem Biol 14:623–634 Bellin RM, Sernett SW, Becker B, Ip W, Huiatt TW, Robson RM (1999) Molecular characteristics and interactions of the intermediate filament protein synemin. Interactions with alpha-actinin may anchor synemin-containing heterofilaments. J. Biol. Chem. 274:29493–29499 Bhattacharya R, Gonzalez AM, Debiase PJ, Trejo HE, Goldman RD, Flitney FW, Jones JC (2009) Recruitment of vimentin to the cell surface by beta3 integrin and plectin mediates adhesion strength. J Cell Sci 122:1390–1400 Das A, Banik NL, Ray SK (2010) Flavonoids activated caspases for apoptosis in human glioblastoma T98G and U87MG cells but not in human normal astrocytes. Cancer 116:164–176 Eckes B, Colucci-Guyon E, Smola H, Nodder S, Babinet C, Krieg T, Martin P (2000) Impaired wound healing in embryonic and adult mice lacking vimentin. J Cell Sci 113: 2455–2462 Ermakova S, Choi BY, Choi HS, Kang BS, Bode AM, Dong Z (2005) The intermediate filament protein vimentin is a new target for epigallocatechin gallate. J Biol Chem 280: 16882–16890 Fortin S, Le Mercier M, Camby I, Spiegl-Kreinecker S, Berger W, Lefranc F, Kiss R (2010) Galectin-1 is implicated in the protein kinase C epsilon/vimentin-controlled trafficking of integrin-beta1 in glioblastoma cells. Brain Pathol 20:39–49 Gilles C, Polette M, Zahm JM, Tournier JM, Volders L, Foidart JM, Birembaut P (1999) Vimentin contributes to human mammary epithelial cell migration. J Cell Sci 112: 4615–4625 Granger BL, Lazarides E (1980) Synemin: a new high molecular weight protein associated with desmin and vimentin filaments in muscle. Cell 22:727–738 Guérette D, Khan PA, Savard PE, Vincent M (2007) Molecular evolution of type VI intermediate filament proteins. BMC Evol Biol 7:164 Hyder CL, Pallari HM, Kochin V, Eriksson JE (2008) Providing cellular signposts–post-translational modifications of intermediate filaments. FEBS Lett 582:2140–2148 Ivaska J, Vuoriluoto K, Huovinen T, Izawa I, Inagaki M, Parker PJ (2005) PKCepsilon-mediated phosphorylation of vimentin controls integrin recycling and motility. EMBO J 24:3834–3845 Izmiryan A, Peltekian E, Paulin D, Li ZL, Xue ZG (2010) Synemin isoforms in astroglial and neuronal cells from human central nervous system. Neurochem Res 35: 881–887 Jing R, Pizzolato G, Robson RM, Gabbiani G, Skalli O (2005) Intermediate filament protein synemin is present in human reactive and malignant astrocytes and associates with ruffled membranes in astrocytoma cells. Glia 50:107–120

88 Jing R, Wilhelmsson U, Goodwill W, Li L, Pan Y, Pekny M, Skalli O (2007) Synemin is expressed in reactive astrocytes in neurotrauma and interacts differentially with vimentin and GFAP intermediate filament networks. J Cell Sci 120: 1267–1277 Katsetos CD, Dráberová E, Legido A, Dumontet C, Dráber P (2009) Tubulin targets in the pathobiology and therapy of glioblastoma multiforme. I. Class III beta-tubulin. J Cell Physiol. 221:505–513 Kim S, Coulombe PA (2007) Intermediate filament scaffolds fulfill mechanical, organizational, and signaling functions in the cytoplasm. Genes Dev 21:1581–1597 Kim H, Nakamura F, Lee W, Hong C, Pérez-Sala D, McCulloch CA (2010) Regulation of cell adhesion to collagen via beta1 integrins is dependent on interactions of filamin A with vimentin and protein kinase C epsilon. Exp Cell Res 316:1829–1844 Kleeberger W, Bova GS, Nielsen ME, Herawi M, Chuang AY, Epstein JI, Berman DM (2007) Roles for the stem cell associated intermediate filament Nestin in prostate cancer migration and metastasis. Cancer Res 67:9199– 9206 Lahat G, Zhu QS, Huang KL, Wang S, Bolshakov S, Liu J, Torres K, Langley RR, Lazar AJ, Hung MC, Lev D (2010) Vimentin is a novel anti-cancer therapeutic target; insights from in vitro and in vivo mice xenograft studies. PLoS One 5:e10105 Mendez MG, Kojima S-I, Goldman RD (2010) Vimentin induces changes in cell shape, motility, and adhesion during the epithelial to mesenchymal transition. FASEB J 24: 1838–1851 Mizuno Y, Thompson TG, Guyon JR, Lidov HG, Brosius M, Imamura M, Ozawa E, Watkins SC, Kunkel LM (2001) Desmuslin, an intermediate filament protein that interacts with alpha-dystrobrevin and desmin. Proc Natl Acad Sci USA 98:6156–6161 Nieminen N, Henttinen T, Merinen M, Marttila–Ichihara F, Eriksson JE, Sirpa J (2006) Vimentin function in lymphocyte adhesion and transcellular migration. Nat Cell Biol 8:156–162 Omary MB (2009) “IF-pathies”: a broad spectrum of intermediate filament-associated diseases. J Clin Invest 119: 1756–1762

Q. Quick et al. Pan Y, Jing R, Pitre A, Williams BJ, Skalli O (2008) Intermediate filament protein synemin contributes to the migratory properties of astrocytoma cells by influencing the dynamics of the actin cytoskeleton. FASEB J 22:3196–3206 Perlson E, Hanz S, Ben-Yaakov K, Segal-Ruder Y, Seger R, Fainzilber M (2005) Vimentin-dependent spatial translocation of an activated MAP kinase in injured nerve. Neuron 45:715–726 Russell MA, Lund LM, Haber R, McKeegan K, Cianciola N, Bond M (2006) The intermediate filament protein, synemin, is an AKAP in the heart. Arch Biochem Biophys 456: 204–215 Rutka JT, Murakami M, Dirks PB, Hubbard SL, Becker LE, Fukuyama K, Jung S, Tsugu A, Matsuzawa K (1997) Role of glial filaments in cells and tumors of glial origin: a review. J Neurosurg 87:420–430 Schmitt-Graeff A, Jing R, Desmoulieres A, Skalli O (2006) Synemin expression is widespread in liver fibrosis and is induced in proliferating and malignant biliary epithelial duct cells. Hum Pathol 37:1200–1210 Shah N, Kataria H, Kaul SC, Ishii T, Kaur G, Wadhwa. R (2009) Effect of the alcoholic extract of Ashwagandha leaves and its components on proliferation, migration, and differentiation of glioblastoma cells: combinational approach for enhanced differentiation. Cancer Sci 100:1740–1747 Strnad P, Windoffer R, Leube RE (2002) Induction of rapid and reversible cytokeratin filament network remodeling by inhibition of tyrosine phosphatases. J Cell Sci 115:4133–4148 Sultana S, Sernett SW, Bellin RM, Robson RM, Skalli O (2000) The intermediate filament protein synemin is transiently expressed in a subpopulation of astrocytes during development. Glia 30:143–153 Sun N, Huiatt TW, Paulin D, Li Z, Robson RM (2010) Synemin interacts with the LIM domain protein zyxin and is essential for cell adhesion and migration. Exp Cell Res 316:491–505 Tsuruta D, Jones JC (2003) The vimentin cytoskeleton regulates focal contact size and adhesion of endothelial cells subjected to shear stress. J Cell Sci 15:4977–4984 Xue ZG, Cheraud Y, Brocheriou V, Izmiryan A, Titeux M, Paulin D, Li Z (2004) The mouse synemin gene encodes three intermediate filament proteins generated by alternative exon usage and different open reading frame. Exp Cell Res 298:431–444

Chapter 10

Diffuse Astrocytomas: Immunohistochemistry of MGMT Expression David Capper

Abstract O6-Methylguanine-DNA methyltransferase (MGMT) represents a major mechanism of resistance of diffuse gliomas against alkylating agents. In glioblastoma, the most frequent and most malignant diffuse astrocytoma, hypermethylation of the MGMT promoter is found in 40–50% of cases and has been linked to improved survival after alkylating chemotherapy. MGMT promoter hypermethylation is expected to result in an inhibition of MGMT protein synthesis. The immunohistochemical investigation of MGMT protein in surgical specimens has been suggested to represent an easy approach to assess response to alkylating agents and several initial studies found strong associations of MGMT immunohistochemistry and patient outcome. Subsequent studies with concordant metholodical approaches could not confirm these associations. As possible reasons for these discordant results, technical aspects of MGMT immunohistochemistry, as well as admixed MGMT positive reactive astrocytes, lymphocytes and macrophages/microglia have been proposed. Despite some indication of an association of MGMT immunohistochemistry and survival, the method can currently not be recommended for evaluation of patient outcome or possible response to alkylating agents in diffuse astrocytomas. Keywords MGMT · CpG island · DNA · IHC markers

D. Capper () Department of Neuropathology, Institute of Pathology, Ruprecht-Karls-University, 69120 Heidelberg, Germany e-mail: [email protected]

Introduction Alkylating agents damage DNA by transferring a methyl group to the O6-position of guanine, thereby facilitating DNA interstrand crosslinking that may subsequently lead to inhibition of cell growth or cell death. O6-methylguanine-DNA methyltransferase (MGMT, also called AGT) is a DNA repair protein that mediates the removal of such cytotoxic mono-adducts and thereby averts the formation of lethal DNA crosslinks (reviewed in Pegg, 1990). These observations have led to the concept that MGMT activity represents a major mechanism of resistance of tumors against alkylating agents. The transfer of a single alkyl group of O6-guanine to MGMT leads to the irreversible inactivation and subsequent degradation of the MGMT protein. Consequently the number of damaged DNA sites that can be repaired by a cell is determined by the number of functional MGMT molecules present (Pegg, 1990). MGMT expression levels and activity in non-neoplastic cells are regulated by various stimuli. In rodents a modest induction of MGMT was observed by alkylating agents, ionizating radiation and other genotoxins, while data of human cultured cells suggest that many human cells might be incapable of an MGMT upregulation by such stimuli (reviewed in Margison et al., 2003). In mice MGMT activation after irradiation was dependent on functional p53. Transcription factor AP-1 and glucocorticoid hormone may induce MGMT expression while MGMT activity is further regulated on the protein level by MGMT phosphorylation (Margison et al., 2003). The designation “diffuse astrocytomas” covers three main tumor entities, namely diffuse astrocytoma World Health Organization (WHO) grade II, anaplastic astrocytoma WHO grade III and glioblastoma WHO

M.A. Hayat (ed.), Tumors of the Central Nervous System, Volume 5, DOI 10.1007/978-94-007-2019-0_10, © Springer Science+Business Media B.V. 2012

89

90

grade IV. Because of their extensive diffuse infiltration of the brain complete surgical removal is not possible and adjuvant treatment by radio- and chemotherapy is essential. For this reason the assessment of MGMT status as a possible predictor of response to alkylating agents is especially important. For Glioblastomas and other diffuse astrocytomas, hypermethylation of the normally unmethylated promoter CpG island in the 5 position of the MGMT gene has been proposed as a frequent mechanism of MGMT inactivation (Esteller et al., 1999). Following the lines that MGMT protein is directly involved in the resistance of diffuse astrocytomas towards chemotherapy, several studies have investigated the predictive and prognostic properties of MGMT protein detection by immunohistochemistry (IHC). IHC of MGMT offers several advantages such as functioning in formalin fixed paraffin embedded tissue and the possibility of detecting MGMT expression in individual cells. Unfortunately the results of these studies are rather inconsistent. Several factors restrain standardization of MGMT IHC and the determination of cut-off levels for prognostic evaluation. The following chapter gives an overview of commonly used antibodies, technical aspects of MGMT IHC, and results of studies investigating MGMT IHC in diffuse astrocytomas.

MGMT Immunohistochemistry – Technical Considerations The majority of published investigations analyzing MGMT IHC in diffuse astrocytomas have relied on the commercially available mouse monoclonal antibody

Fig. 10.1 MGMT automated immunohistochemistry with clone MT23.3. MGMT promoter methylated glioblastoma stained at antibody dilution 1:25, 1:100 and 1:1000. Strong nuclear staining of all cells is observed at low dilutions while only vascular cells (lower left) and single round cells are stained at high dilution. Magnification 400-fold

D. Capper

clone MT3.1 (Mclendon et al., 1998) or less frequently clone MT23.2 (Pollack et al., 2006). Earlier studies used other clones such as mouse monoclonal clone 3B8 (Belanich et al., 1996) or polyclonal rabbit antisera (Yuan et al., 2003), both currently not commercially available. For clones MT3.1 and MT23.2 denatured full length MGMT protein was used as immunogen (Brent et al., 1990), the exact binding site has not been characterized. A standardized comparison of MT3.1 or MT23.2 immunobinding has also not been performed for diffuse astrocytomas. The majority of published manual IHC protocols include antigen retrieval with heating in citrate buffer (pH 6.0). In our IHC experience with clones MT3.1 and MT23.2 both can be used for manual immunostaining and yield comparable results. We could not establish MT3.1 for automated immunostaining on Ventana Benchmark immunohistochemistry system, whereas MT23.2 robustly works under such conditions. The epitope detected by clone MT23.2 is highly sensitive to various artificial tissue alterations. Slight thermal alteration during neurosurgical resection may disrupt immunogeneity of MGMT protein, with the danger of a false negative IHC interpretation. Such areas can be identified by the complete loss of immunoreaction in neoplastic and non-neoplastic cells, whereas cases with negative tumor cell staining generally show at least weak staining of vessels or inflammatory cells. Other IHC markers commonly retain immunogeneity in such foci, demonstrating the high sensitivity of the MT23.2 epitope. The dilution of MGMT antibodies has drastic effects on the results of MGMT immunostaining (Fig. 10.1). Even at low dilutions MGMT staining remains predominantly nuclear, giving the false impression of a specific immunoreaction. In our experience, MGMT antibodies should be carefully titrated

10 Diffuse Astrocytomas: Immunohistochemistry of MGMT Expression

down to the point that the majority of normal vessel endothelia remain moderately positive.

MGMT Immunohistochemistry Compared to MGMT Promoter Methylation and MGMT Activity The application and the limitations of MGMT promoter methylation assays such as methylation specific polymerase chain reaction are discussed elsewhere in this volume. While some studies imply a close association of MGMT promoter methylation status and IHC detection of MGMT protein in gliomas (Maxwell et al., 2006; Metellus et al., 2009), a large number of investigations demonstrated no or only a rather loose association (Brell et al., 2005; Cao et al., 2009; Grasbon-Frodl et al., 2007; Preusser et al., 2008; Rodriguez et al., 2008; Spiegl-Kreinecker et al., 2010; Yachi et al., 2008). Thus analysis of MGMT promoter methylation and MGMT IHC cannot be used interchangeably. A possible explanation for this discrepancy may be the additional regulatory mechanisms of MGMT expression by exogenous and endogenous stimuli described above (Margison et al., 2003). Further, Sasai could demonstrate that while demethylating agents increased MGMT mRNA expression this did not lead to an increase of MGMT protein, indicating additional regulatory mechanisms of MGMT translation (Sasai et al., 2007). Better correlations of MGMT IHC have been observed with MGMT protein activity (Christmann et al., 2010; Maxwell et al., 2006) with a moderate correlation for glioblastoma and a slightly better correlation for anaplastic astrocytoma (Christmann et al., 2010).

MGMT Immunohistochemistry in Glioma Cells and Non-neoplastic Cells In both astrocytoma cells and non-neoplastic cells most of the MGMT protein is located in the nucleus, resulting in a diffuse or finely granular nuclear immunostaining. In rare cases, an additional droplet-like cytoplasmic reaction has been observed in glioma cells (Capper et al., 2008; Yuan et al., 2003). This represents a nonspecific reaction in the majority of cases

91

(Capper et al., 2008). The percentage of positively labeled tumor cells is highest in low-grade diffuse astrocytomas and decreases in high-grade gliomas (Capper et al., 2008; Yuan et al., 2003). This observation is supported by Mineura et al. (1996) who reported a similar decrease of MGMT mRNA levels from low to high grade astrocytomas. As observed by Maxwell et al. (2006) the amount of MGMT varies from tumor cell to tumor cell with both positive and negative tumor cells in the same specimen. While Grasbon-Frodl et al. (2007) described a relatively even distribution of the number of MGMT positive cells in glioblastoma tissue, others have observed considerable variation (Cao et al., 2009). As stated in the technical considerations above, areas with artificial loss of MGMT immunogeneity may falsely be interpreted as variation of MGMT expression, possibly accounting in part for this discrepancy. A study comparing MGMT protein expression in the tumor center and infiltration zone observed strong variation between these two regions and concluded that the infiltration zone of gliomas is not suited for IHC MGMT assessment, as analysis likely results in an overestimation of MGMT protein expression (Capper et al., 2008). Other brain cells show nuclear immunoreaction for MGMT to a varying degree. Besides endothelial cells, ependymal cells, astrocytes and oligodendrocytes frequently show a nuclear labeling in non-neoplastic brain, while cortical neurons are consistently negative (Nakasu et al., 2004). When present, reactive astrocytes, lymphocytes and macrophages/microglia often show a positive nuclear staining (Nakasu et al., 2004). The latter cell types contribute significantly to the actual tumor mass of gliomas and several studies have implicated that these cells impede the standardization of IHC MGMT assessment (Sasai et al., 2008; Nakasu et al., 2004). With histology alone discrimination of these cells from tumor cells is often not possible. Double IHC may help to differentiate astrocytoma cells from inflammatory cells (Fig. 10.2). Cut-off values of 10–15% positive cells have been proposed for MGMT negative cases to account for these non-neoplastic cells (Capper et al., 2008; Nakasu et al., 2004). Besides non-neoplastic components of gliomas, interpretation of MGMT IHC is hampered by the variable extent of nuclear staining intensity, ranging from very weak to strong labeling intensity. This variation introduces a possible risk for a significant observer

92

Fig. 10.2 Double Immunohistochemistry of MGMT (brown) and CD68 (red) of same cases as Fig. 10.1. Nuclear MGMT immunoreaction is observed in vessel cells and in microglia/macrophages double stained by CD68. Glioma tumor cells are negative for MGMT and CD68. Magnification 400-fold

bias. As stated above, the dilution of MGMT antibody also has strong influence on the staining intensity and may contribute to the lack of consistent data (Fig. 10.1). In a comprehensive study of interobserver agreement monoclonal MGMT antibody clone MT23.2 showed a significantly higher interobserver agreement than monoclonal MGMT antibody clone MT3.1 (Preusser et al., 2008). Interobserver agreement on MGMT IHC using clone MT23.2 ranged from moderate to almost perfect and was better than agreement on other tumor relevant questions in this study such as the presence of assessable tumor tissue (Preusser et al., 2008). Clone MT3.1 demonstrated poorer agreement and may thus be less useful for assessment of MGMT protein expression by IHC. In the same study only poor to slight agreement was observed for the presence of MGMT positive hematogenous cells. This observation underscores the difficulty of assessing the presence and number of inflammatory cells in diffuse gliomas and that this factor strongly restrains standardization of MGMT IHC.

MGMT Immunohistochemistry and Patient Outcome For diagnostic neuropathology the main potential of MGMT IHC is in the identification of tumors with loss of MGMT expression and expected better response

D. Capper

to alkylating agents. For glioblastoma, several studies have demonstrated significant associations of IHC MGMT expression and overall survival (Anda et al., 2003; Belanich et al., 1996; Capper et al., 2008; Chinot et al., 2007; Jaeckle et al., 1998; Metellus et al., 2009). Cut-off values in these series to differentiate low expressing from high expressing glioblastomas was either median (35%, Chinot et al., 2007), arbitrary set at 10% (Metellus et al., 2009) or 20% (Anda et al., 2003) or calculated to best separate prognostic groups (15%, Capper et al., 2008; 60,000 molecules/nucleus in quantitative immunofluorescence microscopy Belanich et al., 1996; Jaeckle et al., 1998). In these studies 36–56% of investigated glioblastomas were scored as low MGMT expressing cases. Other studies with concordant metholodical approaches and cut-offs did not observe significant associations of survival and MGMT IHC (Cahill et al., 2007; KarayanTapon et al., 2009; Preusser et al., 2008; Rodriguez et al., 2008). In these series 44–72% of analyzed glioblastomas were scored as low expressing tumors. In previous discussions, variations of cut-off values and the combination of various tumor grades have been discussed as possible causes for lack of association (Capper et al., 2008). Preusser et al. (2008) and Karayan-Tapon et al. (2009) both retrospectively investigated a large cohort of homogenously treated glioblastoma patients, demonstrating that this is not the cause of the discrepant results. Further, the numbers of negatively scored cases are also in a comparable range, making differences of antibody dilution or staining procedure unlikely causes. The reasons for the discordant results remain unclear. Despite some indication of an association of MGMT IHC and survival, the method can currently not be recommended for evaluation of patient outcome or possible response to alkylating agents in glioblastoma. Only little data exists concerning MGMT IHC of anaplastic astrocytomas. In an analysis of 24 anaplastic astrocytomas treated with bis-chloroethylnitrosourea overall median survival for cases with high MGMT levels was 14 months versus 62 months when MGMT expression was low (Jaeckle et al., 1998), while Capper et al. (2008) observed no association of MGMT IHC and outcome for anaplastic astrocytomas. Brell et al. (2005) reported an association of high IHC MGMT expression and worse survival of patients with anaplastic glioma (also including anaplastic oligodendroglioma and oligoastrocytoma) after

10 Diffuse Astrocytomas: Immunohistochemistry of MGMT Expression

treatment with alkylating agents (Brell et al., 2005). For diffuse astrocytomas grade II, MGMT IHC also did not demonstrate satisfying associations with patient outcome. While Capper et al. (2008) observed an association of high MGMT with worse survival; this was strongly dependent on patient age. Nakasu et al. (2007) could not detect an association of overall survival and IHC MGMT expression, while a significant association was observed with tumor progression: Tumors scored MGMT negative showed a more rapid tumor progression (Nakasu et al., 2007). Pediatric high grade gliomas have been investigated for MGMT IHC by two major studies (Pollack et al., 2006; Schlosser et al., 2010) yielding partly conflicting results. Polack et al. (2006) demonstrated that overexpression of MGMT compared to normal brain was associated with worse prognosis in 12 of 109 malignant gliomas treated with alkylating agents (Pollack et al., 2006). In a series of 24 relapsed high grade gliomas Schlosser et al. (2010) observed an association of survival with MGMT promoter methylation but not with IHC protein expression (Schlosser et al., 2010). As with glioblastoma, while there is some indication for prognostic properteis of MGMT IHC for anaplastic gliomas, diffuse astrocytomas WHO grade II and pediatric diffuse gliomas, published data is conflicting. At the current stage MGMT IHC can not be recommended as a reliable marker of patient outcome or response to alkylating agents. Cao et al. (2009) have recently proposed combining analysis of MGMT promoter methylation with MGMT IHC for prognostic evaluation. Indeed, the combination of MGMT promoter methylation with negative MGMT IHC expression could reduce the shortcomings observed for both methods (Cao et al., 2009).

References Anda T, Shabani HK, Tsunoda K, Tokunaga Y, Kaminogo M, Shibata S, Hayashi T, Iseki M (2003) Relationship between expression of O6-methylguanine-DNA methyltransferase, glutathione-S-transferase pi in glioblastoma and the survival of the patients treated with nimustine hydrochloride: an immunohistochemical analysis. Neurol Res 25:241–248 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

93

of brain tumor patients treated with carmustine. Cancer Res 56:783–788 Brell M, Tortosa A, Verger E, Gil JM, Viñolas N, Villá S, Acebes JJ, Caral L, Pujol T, Ferrer I, Ribalta T, Graus F (2005) Prognostic significance of O6-methylguanine-DNA methyltransferase determined by promoter hypermethylation and immunohistochemical expression in anaplastic gliomas. Clin Cancer Res 11:5167–5174 Brent TP, von Wronski M, Pegram CN, Bigner DD (1990) Immunoaffinity purification of human O6-alkylguanineDNA alkyltransferase using newly developed monoclonal antibodies. Cancer Res 50:58–61 Cahill DP, Levine KK, Betensky RA, Codd PJ, Romany CA, Reavie LB, Batchelor TT, Futreal PA, Stratton MR, Curry WT, Iafrate AJ, Louis DN (2007) Loss of the mismatch repair protein MSH6 in human glioblastomas is associated with tumor progression during temozolomide treatment. Clin Cancer Res 13:2038–2045 Cao VT, Jung TY, Jung S, Jin SG, Moon KS, Kim IY, Kang SS, Park CS, Lee KH, Chae HJ (2009) The correlation and prognostic significance of MGMT promoter methylation and MGMT protein in glioblastomas. Neurosurgery 65: 866–875 Capper D, Mittelbronn M, Meyermann R, Schittenhelm J (2008) Pitfalls in the assessment of MGMT expression and in its correlation with survival in diffuse astrocytomas: proposal of a feasible immunohistochemical approach. Acta Neuropathol 115:249–259 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 O6methylguanine-DNA methyltransferase and survival in inoperable newly diagnosed glioblastoma patients treated with neoadjuvant temozolomide. J Clin Oncol 25:1470–1475 Christmann M, Nagel G, Horn S, Krahn U, Wiewrodt D, Sommer C, Kaina B (2010) MGMT activity, promoter methylation and immunohistochemistry of pre-treatment and recurrent malignant gliomas: a comparative study on astrocytoma and glioblastoma. Int J Cancer 127:2106–2118 Esteller M, Hamilton SR, Burger PC, Baylin SB, Herman JG (1999) Inactivation of the DNA repair gene O6methylguanine-DNA methyltransferase by promoter hypermethylation is a common event in primary human neoplasia. Cancer Res 59:793–797 Grasbon-Frodl EM, Kreth FW, Ruiter M, Schnell O, Bise K, Felsberg J, Reifenberger G, Tonn JC, Kretzschmar HA (2007) Intratumoral homogeneity of MGMT promoter hypermethylation as demonstrated in serial stereotactic specimens from anaplastic astrocytomas and glioblastomas. Int J Cancer 121:2458–2464 Jaeckle KA, Eyre HJ, Townsend JJ, Schulman S, Knudson HM, Belanich M, Yarosh DB, Bearman SI, Giroux DJ, Schold SC (1998) Correlation of tumor O6 methylguanineDNA methyltransferase levels with survival of malignant astrocytoma patients treated with bis-chloroethylnitrosourea: a southwest oncology group study. J Clin Oncol 16: 3310–3315 Karayan-Tapon L, Quillien V, Guilhot J, Wager M, Fromont G, Saikali S, Etcheverry A, Hamlat A, Loussouarn D, Campion L, Campone M, Vallette FM, Gratas-Rabbia-Ré C (2009) Prognostic value of O(6)-methylguanine-DNA

94 methyltransferase status in glioblastoma patients, assessed by five different methods. J Neurooncol 97:311–322 Margison GP, Povey AC, Kaina B, Santibáñez Koref MF (2003) Variability and regulation of O6-alkylguanine-DNA alkyltransferase. Carcinogenesis 24:625–635 Maxwell JA, Johnson SP, Quinn JA, McLendon RE, Ali-Osman F, Friedman AH, Herndon JE 2nd, Bierau K, Bigley J, Bigner DD, Friedman HS (2006) Quantitative analysis of O6-alkylguanine-DNA alkyltransferase in malignant glioma. Mol Cancer Ther 5:2531–2539 McLendon RE, Cleveland L, Pegram C, Bigner SH, Bigner DD, Friedman HS (1998) Immunohistochemical detection of the DNA repair enzyme O6-methylguanine-DNA methyltransferase in formalin-fixed, paraffin-embedded astrocytomas. Lab Invest 78:643–644 Metellus P, Coulibaly B, Nanni I, Fina F, Eudes N, Giorgi R, Barrie M, Chinot O, Fuentes S, Dufour H, Ouafik L, Figarella-Branger D (2009) Prognostic impact of O6methylguanine-DNA methyltransferase silencing in patients with recurrent glioblastoma multiforme who undergo surgery and carmustine wafer implantation: a prospective patient cohort. Cancer 115:4783–4794 Mineura K, Yanagisawa T, Watanabe K, Kowada M, Yasui N (1996) Human brain tumor O(6)-methylguanine-DNA methyltransferase mRNA and its significance as an indicator of selective chloroethylnitrosourea chemotherapy. Int J Cancer 69:420–425 Nakasu S, Fukami T, Baba K, Matsuda M (2004) Immunohistochemical study for O6-methylguanine-DNA methyltransferase in the non-neoplastic and neoplastic components of gliomas. J Neurooncol 70:333–340 Nakasu S, Fukami T, Jito J, Matsuda M (2007) Prognostic significance of loss of O6-methylguanine-DNA methyltransferase expression in supratentorial diffuse low-grade astrocytoma. Surg Neurol 68:603–608 Pegg AE (1990) Mammalian O6-alkylguanine-DNA alkyltransferase: regulation and importance in response to alkylating carcinogenic and therapeutic agents. Cancer Res 50: 6119–6129 Pollack IF, Hamilton RL, Sobol RW, Burnham J, Yates AJ, Holmes EJ, Zhou T, Finlay JL (2006) O6-methylguanineDNA methyltransferase expression strongly correlates with outcome in childhood malignant gliomas: results from the CCG-945 Cohort. J Clin Oncol 24:3431–3437

D. Capper Preusser M, Charles Janzer R, Felsberg J, Reifenberger G, Hamou MF, Diserens AC, Stupp R, Gorlia T, Marosi C, Heinzl H, Hainfellner JA, Hegi M (2008) Anti-O6methylguanine-methyltransferase (MGMT) immunohistochemistry in glioblastoma multiforme: observer variability and lack of association with patient survival impede its use as clinical biomarker. Brain Pathol 18:520–532 Rodriguez FJ, Thibodeau SN, Jenkins RB, Schowalter KV, Caron BL, O’neill BP, James CD, Passe S, Slezak J, Giannini C (2008) MGMT immunohistochemical expression and promoter methylation in human glioblastoma. Appl Immunohistochem Mol Morphol 16:59–65 Sasai K, Akagi T, Aoyanagi E, Tabu K, Kaneko S, Tanaka S (2007) O6-methylguanine-DNA methyltransferase is downregulated in transformed astrocyte cells: implications for anti-glioma therapies. Mol Cancer 5(6):36 Sasai K, Nodagashira M, Nishihara H, Aoyanagi E, Wang L, Katoh M, Murata J, Ozaki Y, Ito T, Fujimoto S, Kaneko S, Nagashima K, Tanaka S (2008) Careful exclusion of non-neoplastic brain components is required for an appropriate evaluation of O6-methylguanine-DNA methyltransferase status in glioma: relationship between immunohistochemistry and methylation analysis. Am J Surg Pathol 32: 1220–1227 Schlosser S, Wagner S, Mühlisch J, Hasselblatt M, Gerss J, Wolff JE, Frühwald MC (2010) MGMT as a potential stratification marker in relapsed high-grade glioma of children: the HIT-GBM experience. Pediatr Blood Cancer 54:228–237 Spiegl-Kreinecker S, Pirker C, Filipits M, Lötsch D, Buchroithner J, Pichler J, Silye R, Weis S, Micksche M, Fischer J, Berger W (2010) O6-Methylguanine DNA methyltransferase protein expression in tumor cells predicts outcome of temozolomide therapy in glioblastoma patients. Neuro Oncol 12:28–36 Yachi K, Watanabe T, Ohta T, Fukushima T, Yoshino A, Ogino A, Katayama Y, Nagase H (2008) Relevance of MSP assay for the detection of MGMT promoter hypermethylation in glioblastomas. Int J Oncol 33:469–475 Yuan Q, Matsumoto K, Nakabeppu Y, Iwaki T (2003) A comparative immunohistochemistry of O6-methylguanine-DNA methyltransferase and p53 in diffusely infiltrating astrocytomas. Neuropathology 23:203–209

Chapter 11

Central Nervous System Germ Cell Tumors: An Epidemiology Review Daniel L. Keene and Donna Johnston

Abstract The objective of this chapter is to review the histological characteristics, clinical presentation and the incidence of central nervous system (CNS) germ cell tumors. This group of tumors account for approximately 1–3% of the total tumor population. Based on histological examination, the group can be subdivided into germinoma and non-germinoma germ cell tumors with the germinonas being significantly more common. Though germ cell tumors can be located anywhere along the midline neuro-axis, they are significantly more frequent located in the pineal region. The clinical presentation varies according to the tumor location. The tumor tends to have a predilection for males around time of puberty who are of Asian background. The reason for this remains unclear. Keywords Central nervous system · Germ cell tumor · Teratoma · Yolk-sac tumor

Introduction Germ cell tumors occur in various locations throughout the body, but almost always occur in the midline (i.e., mediastinum, sacrococygeal region, retroperitoneum, and brain diecephalon). It has been postulated that germ cell tumors arise from primitive germ cells. Normally these primitive germ cells appear in the yolk

D.L. Keene () Division of Neurology, Department of Pediatrics, Children’s Hospital of Eastern Ontario, Ottawa, ON, Canada K1H 8 e-mail: [email protected]

sac of the embryo around third to fourth week of gestation and normal migrates to the ovaries or testis; however, as a result of misplaced migration of the primitive germ cell it migrates to different parts of the body during embryonic life. These cells then develop into germ cell tumors. An alternate explanation is that the tumor cells arise from pleuropotential embryonic cells that have escaped the influence of a primary organizer during embryonic development. Primary CNS germ cell tumors are relatively uncommon accounting for ∼1–3% of all CNS tumors diagnosed under the age of 18 years (Goodwin et al., 2009). This paper reviews the histological classification, clinical presentation, and descriptive epidemiology of CNS germ cell tumors.

Histological Classification Germ cell tumors of the CNS are divided by the World Health Organization into benign and malignant tumor. The benign group consists of the teratoma and the malignant intracranial germ cell tumors consist of the germinoma and non-germinoma group. Teratomas may consist of cystic and solid component; however, by definition, a teratoma must contain components of the three embryonic germ cell layers (ectoderm, mesoderm and endoderm). If it contains fetal or immature tissue, it is classified as an immature teratoma. If malignant transformation of a carcinomatous or sarcomatous nature is seen in the solid component, the teratoma is classified as a malignant teratoma (De Girolami et al., 2008). The histology of the teratoma is reflected in is magnetic resonance image where it has a strikingly heterogenous honeycomb appearance with

M.A. Hayat (ed.), Tumors of the Central Nervous System, Volume 5, DOI 10.1007/978-94-007-2019-0_11, © Springer Science+Business Media B.V. 2012

95

96

multiple cysts and areas of calcification within it, with high intensity signal on T1W1 and T2W2 images. This signal change probably represents the high protein liquid contain. Fat can often be detected on fat saturation sequences (Korogi et al., 2001). On histological examination, the germinona is similar to the testicular seminona and the ovarian dysgerminoma. It is somewhat circumscribed, and grows by expansion and extension into the subarachnoid space. This can result in seeding along the neuroaxis. Microscopically, the tumor consists of two cells types, which can be somewhat intermixed. One cell type consists of round tumor cells with vesticular nuclei with prominent nucleoli and abdundant clear cytoplasm. The cytoplasm often stains positive for PAS+ material, placenta alkaline phosphatase (PLAP) antibodies, alpha feto-protein (AFP), CD117. Positive staining for beta human chorionic gonadotrophin (β-HCG) may be present in scattered synctio-troplast-like cells.. Small T-lymphocytes make up the second cell type. These lymphocytes can be in clusters and can form germinalcentre-like nodules. Mib-1 index is often high in keeping with the active, proliferate nature of these tumors (De Girolami et al., 2008). Though this tumor is not classified as a secreting germ cell tumor, mild elevation in serum and/or cerebrospinal fluid levels of β-HGC has been reported (Keene et al., 2007). On magnetic resonance imaging, the germinona may have a cystic component: however the solid component is usually mildly hypointense on T1 weighted images and mildly hyperintense on T2 weighted images. The tumor markedly enhances in a heterogeneous fashion after Gd-DTPA administration (Korogi et al., 2001). Germinonas make up >50% of the CNS germ cell tumors reported in series (Keene et al., 2007). The non-germinoma germ cell tumor group consists of the yolk-sac tumors (endodermal sinus tumor), embryonal carcinoma, choriocarcinoma, and the teratomas. The yolk-sac tumors is a papillary primitive epithelial tumor with a myxoid matrix in organoid structure containing a blood vessel surrounded by an epithelial-lined frond. Often some cells stain positive for AFP, as well as elevated serum and/or cerebrospinal fluid levels of AFP. The embryonal carcinoma is characterized of primitive carcinoma cells in sheets or gland-like aggregates. These cells stain positive for PLAP and cytokeratins. Choriocarcinoma is a hemorrhagic tumor containing cytotrophoblastic and syncytiophoblastic giant cells within sheets of primitive

D.L. Keene and D. Johnston

epithelial cells. These tumor cells often stain positive for β-HCG (De Girolami et al., 2008) and can have elevated serum and/or cerebrospinal fluid levels of β-HCG. Similar to the germinona, on magnetic imaging, the non-germinoma group of tumors can have both cystic and solid component with the solid component being iso- to slightly hypo-intense relative to grey matter on T1-weighted image and mixed iso- and hyper-intense on T2-weighted images. After GD-DTPA administration, the tumor heterogeneously enhances. In addition, choriocarcinoma can have a hemorrhagic component (Korogi et al., 2001).

Discriptive Epidemiology Based on Tumor Location and Clinical Presentation The clinical presentation is related to the site in which the tumor can be found. As these tumors can be found in various locations within the CNS, the clinical presentation varies. The commonest location is in the pineal region, a region consisting of pineal gland, posterior aspect of the third ventricle, quadrigeminal plate, ambient cisterns, velum interpositium and tela choroidea. Tumors in this region have been reported to make up ∼75% of the cases of the CNS germ cell tumors (Keene et al., 2007; Shotelersuk et al., 2003). The presenting symptoms for tumors in this location is often vomiting, headache as a result of increased intracranial pressure secondary to tumor expansion into the posterior third ventricle blocking the aqueduct. As well tumor expansion may lead to abnormalities of upward eye movements (Matsutami et al., 1997). Tumors in the suprasellar region account for between 10 and 20% of the cases (Keene et al., 2007; Shotelersuk et al., 2003). Tumor extension into this region can result in diabetes insipidus, and precocious puberty. Compression of the optic pathways by the tumor can result in visual field deficits (Matsutami et al., 1997). Isolated progressive hemiparesis or plegia is the presenting symptom/sign in about in ∼2.5–5% of the reported cases. This is secondary to present of the germ cell tumor in the region of the basal ganglia or supra-tentorial white periventricular white matter (Keene et al., 2007; Matsutami et al., 1997).

11 Central Nervous System Germ Cell Tumors: An Epidemiology Review

In addition, there have been reports of discrete germ cell tumors in more than one location in the brain parenchyma at the time of presentation. Rarely, germ cell tumors have been reported in the cerebellum and the cerebellar pontine angle (Matsutami et al., 1997). Though the majority of persons with CNS germ cell tumors tend to present around the time of adolescents (10–12 years), on occasion it can occur into the third decade (Goodwin et al., 2009). Germ cell tumors occur approximately four times more often in males than in females (Keene et al., 2007; Goodwin et al., 2009) with germinomas occurring about twice as often and nongerminonas about eleven times more often (Keene et al., 2007). Tumors occur in the pineal region more frequently in males than females, particularly for germinomas (Goodwin et al., 2009). In females, no clear pattern of tumor location has been reported (Goodwin et al., 2009; Keene et al., 2007).

Incidence It has been suggested that germ cell tumors occur more frequently in the Asian population than North America. These studies have been based on differences in the proportional of CNS germ cell tumors in comparison to other tumor types. Most have been based on single institute studies (Araki and Matsunoto, 1969; Nomura, 2001; Oi, 1998). Based on the results of a survey of patients diagnosed as having a primary brain tumors in Kumamoto prefecture in Japan between 1989 and 1994, Kuratsu and Ushio (1996) reported that the overall incidence of germ cells tumors was reported to be 0.17 per 100,000 population years with a rate of 0.3 per 100,000 population years for males and 0.07 per 100,000 population years for females (4.29 male to female ratio). In contrast, Keene et al. (2007) reported the incidence of CNS germ cell tumors in Canada to be 1.06 per million children (0.7 per million children for germinona and 0.3 per million children for non-germinona). The tumors were also reported to have been more predominant in males (2.4:1 for germinoma and 11:1 for non-germinoma). These results were based on a retrospective national Canadian survey of patients diagnosed, between the years of 1990 and 2004, as having a CNS germ cell tumor.. An attempt

97

to investigate the possibility of a genetic predominance or population susceptibility, Keene et al. (2007) examined the ethnic origin of the Canadian cases. The relative risk was 1.86 times higher in the persons of Asian background compared to Caucasians. However, a definite conclusion could not be obtained as the ethnic background was missing in a significant number of the patients. Using Surveillance, Epidemiology and End Results (SEER-17) registry data, Goodwin et al. (2009) also tried to answer this question. They reported that the incidence of germ cell tumors was higher in persons of Asian background (2.6 per million persons) compared to 1.29 per million persons with Caucasian background, and 0.33 per million person with African American background. As SEER registry information does not immigration status to and from the United States and the rate found was still lower than reported incidence in other Asian countries, the question as to whether the predilection among Asian person was genetic or environmental in origin remained unanswered. Smith et al. (1998) reported that the overall incidence of tumors of the CNS had increased by 35% between 1973 and1994. It has been a matter of debate as to the reason for this increase. Was it due to an etiological phenomenon or the result of technological advancements in imaging? Smith et al. (1998) demonstrated that the incidence actually jumped around 1985. This was the time that there was a marked increase in the access to MRI scanning; thus supporting the hypothesis that jump hypothesis. It is not known whether a similar phenomenon occurred with germ cell tumors. However, between 1990 and 2004, Keene et al. (2007) reported that there was a non-significant increase in the annual incidence of both germinonas and non-germinoma germ cell tumors in their series. This suggests, that in the Canadian population, the occurrence rate has remained relatively stable over time. In conclusion, CNS germ cell tumors are rare tumors accounting for between 1 and 3% of total tumor population. Based on the tumor histology, the tumor is divided into germinona and non-germinoma germ cell tumors. This group of tumors tends to occur more frequently in the pineal region, though it can be found through the midline neuro-axis. It has predilection for males of Asian background near the time of puberty.

98

References Araki C, Matsumoto S (1969) Statistical reevaluation of pinealoma and related tumors in Japan. J Neurosurg 30: 146–149 De Girolami U, Févre-Montange M, Seilhean D, Jouvet A (2008) Pathology of tumors in the pineal region. Rev Neurologique 164:882–895 Goodwin T, Sainami K, Fisher P (2009) Incidence patterns of CNS germ cell tumors. J Pediatr Hematol Oncol 1:541–544 Keene D, Johnston D, Strother D, Fryer C, Carret AS, Crooks B, Eisenstat D, Moghrabi A, Wilson B, Brossard J, Mpofu C, Odame I, Zelcer S, Silva M, Samson Y, Hand J, Bouffet E (2007) Epidemiological survey of CNS germ cell tumors in Canadian children. J Neurooncol 82:289–95 Korogi Y, Takahushi M, Ushio Y (2001) MRI of pineal region tumors. J Neurooncol 54:251–261 Kuratsu J, Ushio Y (1996) Epidemiology study of primary intracranial tumors: a regional survey of Kumamoto prefecture in the southern part of Japan. J Neurosurg 84:946–950

D.L. Keene and D. Johnston Matsutani M, Sano K, Takakura K, Fujimaki T, Nakamura O, Funata N, Seto T (1997) Primary intracranial germ cell tumors: a clinical analysis of 15 histologically verified cases. J Neurosurg 86(4):46–455 Normura K (2001) Epidemiology of germ cell tumors in Asia of pineal region tumor. J Neurooncol 54:211–217 Oi S (1998) Recent advances and racial differences in therapeutic strategy to the pineal region tumor. Child’s Nerv Syst 14:3–35 Shotelersuk K, Rojpornpradit P, Chottetanaprasit T, Lerbutsayanukul C, Lertsanguansinchai P, Khorprasert C, Asavametha N, Suriyappe S, Jumpangern C (2003) Intracranial germ cell tumors: experience in king Chulalongkorn memorial hospital. J Med Assoc Thai 86:60–611 Smith M, Freidlin B, Ries G, Simon R (1998) Trends in reporting incidence of primary malignant brain tumors in children in the United States. J Natl Cancer Inst 90: 1267–1277

Chapter 12

RAF Fusion Genes and MAPK Activation in Pilocytic Astrocytomas Sally R. Lambert and David T. W. Jones

Abstract Pilocytic astrocytomas (PA) are the most common childhood tumor of the central nervous system (CNS). Recent research has highlighted a key role for the mitogen activated protein kinase (MAPK) pathway in the development of these tumors. Whilst several mechanisms of activation of this pathway have been described in PA, it is most commonly caused by the formation of a BRAF fusion gene that encodes a protein with a constitutively active kinase domain. This fusion gene is highly specific to PAs when compared to other brain tumour subtypes, with direct implications for improved diagnostic procedures. Furthermore, the discovery of such a frequent molecular change underlying the development of these tumors affords a potential opportunity for targeted therapy of PAs. Keywords Pilocytic astrocytoma · CNS · MAPK · NF1 gene · BRAF

Introduction Pilocytic astrocytomas (PA) account for 18% of all pediatric central nervous system (CNS) tumors in the US (Central Brain Tumor Registry of the United States, 2011). They usually present as well-circumscribed, slow-growing lesions and are classified as grade I by the World Health Organisation (WHO; Louis et al., 2007). Due to their slow-growth and typically

S.R. Lambert () Department of Pathology, University of Cambridge, Addenbrooke’s Hospital Box 231, Cambridge, CB2 0QQ, UK e-mail: [email protected]

non-invasive behaviour they are considered benign, with 96% of patients surviving 10 years or more (Ohgaki and Kleihues, 2005). The cerebellum is the most frequent site for PAs, but they can also arise in other regions of the CNS such as the optic nerve, hypothalamus and brain stem. Surgery is the most common treatment option and is not usually combined with adjuvant therapies such as radiotherapy or chemotherapy, due to the potentially damaging effects of these on the developing brain of a child. Despite their generally benign course, recurrence occurs in up to 19% of patients and is associated with significant morbidity, as are PAs that arise in regions of the brain that are inaccessible to surgery, such as the optic nerve (Dirven et al., 1997). Furthermore, many patients experience long-term health and cognitive effects related to their tumor and/or therapy, even where gross total resection is curative of the primary lesion (Armstrong et al., 2011). Unlike WHO grade II or III astrocytomas however, malignant progression to a higher grade is extremely rare. Pilocytic astrocytomas usually exhibit a biphasic architecture, with regions of compact fibrillary tissue contrasting with microcystic areas. The tumour cells are typically elongated and bipolar with piloid (hairlike) extensions, giving the tumor its name. Rosenthal fibres and eosinophilic granular bodies are also common features, although they are not always observed and are found in other diseases of the brain, so are not solely diagnostic of a PA. Histopathologically, PAs show a wide variety in morphology and can share features with higher grade malignant astrocytomas, oligodendrogliomas and ependymomas (Louis et al., 2007). There are also several variants of PA including pilomyxoid astrocytoma and anaplastic PA, which are both associated with a less favourable prognosis

M.A. Hayat (ed.), Tumors of the Central Nervous System, Volume 5, DOI 10.1007/978-94-007-2019-0_12, © Springer Science+Business Media B.V. 2012

99

100

(Tihan et al., 1999; Tomlinson et al., 1994). Accurate diagnosis of these tumours can therefore be difficult, impacting directly upon the treatment and prognosis of the patient. A molecular diagnostic test for use in addition to current histopathological and clinical criteria is highly desirable. The first indication of the importance of MAPK signalling to PA tumorigenesis was the observation that patients with Neurofibromatosis type 1, a disease arising from mutations in the neurofibromatosis 1 (NF1) gene, are at an increased risk of developing glioma. Pilocytic astroctyomas account for about 50% of NF1associated gliomas and approximately 15% of NF1 patients develop a PA, predominantly in the optic pathway (Listernick et al., 1999; von Deimling et al., 1995). Mutations in NF1 abrogate its function, resulting in the activation of Ras and downstream ERK/MAPK signaling. A number of further studies have since implicated alterations in the MAPK pathway in the development of sporadic PAs, as outlined below.

Molecular Genetic Development of Pilocytic Astrocytoma Until recently, the molecular genetic changes underlying PA development were mostly uncharacterized. Large regions of genomic imbalance were only identified in a small subset of cases by cytogenetic and comparative genomic hybridization (CGH) studies, with 64–86% of PAs displaying balanced karyotypes (Agamanolis and Malone, 1995; Bigner et al., 1997; Jones et al., 2006). Of those cases with genomic imbalance, gains involving chromosome 7 were reported most frequently – in approximately 25% of cases – and whole chromosomal gains were observed to be more common in adult patients (Jones et al., 2006; Schrock et al., 1996; White et al., 1995). In addition, the analysis of candidate genes that are frequently altered in higher grade astrocytomas (such as TP53 and PTEN), revealed only occasional mutation of these genes in PAs (Walker et al., 2001). Recently however, several reports have described genetic aberrations affecting the MAPK pathway in the majority of PAs. These advances have highlighted the key role of this pathway in the development of PA and have led to speculation that PA could represent a single-pathway disease. The

S.R. Lambert and D.T.W. Jones

remainder of this chapter will focus on these recent studies of the molecular genetic basis of PA tumorigenesis and their implications for improved diagnosis and targeted molecular therapeutics.

BRAF:KIAA1549 Fusion Genes in Pilocytic Astrocytoma The identification of RAF fusion genes in a significant proportion of PAs has markedly advanced our understanding of the biology underlying PA development. The most frequently identified fusion gene occurs via a mechanism of tandem duplication of approximately 2 Mb at 7q34. This fuses the 5 end of the uncharacterised gene KIAA1549 to the 3 end of v-RAF murine sarcoma viral oncogene homolog B1 (BRAF), creating a fusion gene under the control of the KIAA1549 promoter. This aberration has been reported in 66–100% of PAs, making it by far the most frequently identified change in this tumor type (Forshew et al., 2009; Jones et al., 2008; Schiffman et al., 2010; Sievert et al., 2009). Furthermore, the fusion gene is highly specific to PAs when compared to other types of brain tumor (including higher grade astrocytoma, oligodendroglioma and ependymoma), highlighting its potential as a diagnostic marker for PAs (Jones et al., 2008; Lawson et al., 2010). Multiple breakpoints have been identified in both KIAA1549 and BRAF, creating fusion gene variants with different in-frame exon combinations. The most frequently identified breakpoint region is located in intron 16 of KIAA1549 and intron 8 of BRAF (creating a KIAA1549ex16 -BRAFex9 transcript) and accounts for 63% of all cases with an identified BRAF fusion gene. Other reported variants of the fusion gene include KIAA1549ex15 -BRAFex9 which has been found in 27% of all BRAF fusion cases, KIAA1549ex16 BRAFex11 which accounts for 9% of BRAF fusions and KIAA1549ex18 -BRAFex10 and KIAA1549ex19 -BRAFex9 , both of which have been reported in single cases only. Detailed breakpoint mapping has revealed the genomic breakpoints within each tumor are unique and do not cluster to a hotspot location within the KIAA1549 or BRAF introns (Lawson et al., 2011). Furthermore, sequence microhomology at the breakpoint or flanking genomic sequence has been demonstrated in 85% of

12 RAF Fusion Genes and MAPK Activation in Pilocytic Astrocytomas

PAs with simple genomic rearrangements, implicating non-homologous end joining or microhomologymediated break induced replication as potential mechanisms for tandem duplication and resulting fusion gene formation (Cin et al., 2011; Lawson et al., 2011). These have yet to be confirmed however and further studies are warranted to establish the precise mechanism involved. In addition, a BRAF fusion gene arising from an interstitial deletion between BRAF and the FAM131B gene (which is located distally to BRAF) has been reported in 3 cases of PA, suggesting multiple mechanisms may result in fusion of BRAF in PA (Cin et al., 2011). Notably, all reported fusion variants produce an inframe protein that retains the kinase domain of BRAF, but lacks the auto-regulatory N-terminal domain of the wildtype protein. This would be predicted to result in constitutive activation of the remaining kinase domain. When compared with control samples of normal foetal brain, tumor samples with a BRAF fusion gene showed a significant upregulation of MEK/ERK phosphorylation, which is consistent with a constitutively active BRAF kinase domain (Forshew et al., 2009). Furthermore, a fusion gene construct with the most commonly cited breakpoint (KIAA1549ex16 -BRAFex9 ) was shown to have elevated kinase activity compared to the wildtype protein in-vitro and to transform NIH3T3 cells in a colony forming assay, providing evidence that the fusion gene has direct transforming properties (Jones et al., 2008). Besides the contribution of the KIAA1549 promoter to the fusion gene, the normal cellular role of the KIAA1549 protein is unknown. There are two alternative promoters that produce a long protein of ∼210 kDa and a shorter, ∼80 kDa protein, as well as additional splice variants. Transcripts from both promoters are expressed in the brain and encode a putative transmembrane domain that is retained in the fusion protein and may affect protein localisation.

MAPK Activation via RAF1 Fusion Genes Although the KIAA1549-BRAF fusion gene has been identified as the most frequent genetic change in PA, alternative mechanisms of MAPK upregulation have also been discovered. With significant similarity to

101

the KIAA1549-BRAF fusion, a fusion between v-raf-1 murine leukemia viral oncogene homolog 1 (RAF1) and the SLIT-ROBO Rho GTPase-activating protein 3 (SRGAP3) gene has been reported in two individual PA cases (Forshew et al., 2009; Jones et al., 2009). This fusion gene also arises via a mechanism of tandem duplication at 3p25, with different breakpoints reported in each case. As with the common BRAF fusion, the two transcripts – SRGAP3ex11 -RAF1ex8 and SRGAP3ex12 -RAF1ex10 – both result in an in-frame protein that lacks the auto-regulatory domain of RAF1 and is under the control of the SRGAP3 promoter. The SRGAP3 gene is highly expressed in normal foetal and adult brain and regulates neuronal migration and axonal branching (Endris et al., 2002). In-vitro assessment of the SRGAP3ex12 -RAF1ex10 variant revealed that the fusion protein also has transforming activity in NIH3T3 cells and demonstrated elevated phosphorylation of MEK, both of which are consistent with a constitutively active kinase domain (Jones et al., 2009). Elevated phosphorylation of MEK/ERK was also demonstrated in protein lysates from one tumor sample with the SRGAP3ex11 -RAF1ex8 fusion, confirming activation of this pathway in-vivo (Forshew et al., 2009). The creation of fusion genes involving either BRAF or RAF1 have been reported in other solid cancers, but at a substantially lower frequency than described in PAs. They have been identified in 1–2% of melanoma, prostate, gastric and papillary thyroid carcinomas (Ciampi et al., 2005; Palanisamy et al., 2010). Interestingly, the frequency of fusion gene formation rose to 11% in papillary thyroid carcinoma from children and young adolescents exposed to radiation in the Chernobyl nuclear disaster, suggesting radiation may play a causative role in the formation of fusion genes in these cases (Ciampi et al., 2005). However, there is no indication for a role of radiation exposure in the formation of PA-related fusion genes. In all the other cancer types, the fusion gene partner was distinct to those reported in PA and tandem duplication was not identified as the mechanism by which the fusion was formed. The high frequency of RAF fusion genes in PA, consistency of the fusion partner and common mechanism of formation all strongly suggest these aberrations fundamentally underpin the aetiology of PA development, in contrast to other solid tumors.

102

Alternative Mechanisms of MAPK Activation in Pilocytic Astrocytoma In total, fusion genes involving the RAF kinases account for activation of the MAPK pathway in 68– 100% of PA cases, depending on the specific cohort. However, those PAs that do not display a RAF fusion gene have been demonstrated to show upregulation of the MAPK pathway, suggesting that alternative mechanisms of activation may occur in the subset of cases without a characterised fusion gene (Jacob et al., 2009). Several alternative mechanisms of MAPK pathway activation have been identified in PA, albeit at a considerably lower frequency than the BRAF fusion gene. These include mutation of KRAS and BRAF, which are reported in less than 1% and up to 9% of cases respectively (Eisenhardt et al., 2010; Forshew et al., 2009; Jones et al., 2008). The most frequently cited mutation of BRAF in PA is the V600E mutation, which is commonly observed in other cancer types including melanoma, colorectal cancer and papillary thyroid carcinoma (Davies et al., 2002). Intriguingly, this mutation was shown to be strongly associated with tumors arising outside the cerebellum, suggesting that although MAPK activation appears to be intrinsic to most – if not all – PAs, tumors that arise in different locations may be subject to distinct molecular aberrations (Schindler et al., 2011). In addition to the V600E mutation, a 3 bp insertion in close proximity to this frequently mutated hotspot codon has also been reported (Eisenhardt et al., 2010; Jones et al., 2009). This mutation inserts a threonine at position 598 or 599 (the sequence context does not distinguish between either mutation) and confers a constitutively active kinase domain, as demonstrated by elevated kinase activity and transforming properties in-vitro (Jones et al., 2009). Furthermore, a subset of patients who develop PA are diagnosed clinically with NF1. This disease is characterised by cafe-au-lait spots, neurofibromas, Lisch nodules, optic glioma (commonly PA) and mild cognitive impairment. Mutations in the NF1 gene are causative of the disease and whilst it can be inherited in an autosomal dominant fashion, approximately half of all cases represent new mutations (von Deimling et al., 1995). The NF1 gene encodes the neurofibromin protein, which is a GTPase activating protein that downregulates Ras activity. Approximately 82% of

S.R. Lambert and D.T.W. Jones

mutations in this gene are predicated to truncate the neurofibromin protein, thereby impairing its activity (Shen et al., 1996). This results in the activation of Ras and downstream ERK/MAPK signalling.

Targeted Therapy Against the MAPK Pathway BRAF and RAF1 are serine/threonine kinases that transmit signals from the cell surface to the cytoplasm and nucleus as part of the RAS-RAFMEK-ERK/MAPK signalling cascade. In brief, receptor-mediated activation of Ras leads to the recruitment and activation of Raf kinases (either BRAF, ARAF or RAF1), which in turn phosphorylate and activate MEK1/2. Activated MEK phosphorylates ERK1/2 (otherwise known as MAPK1/3), which collectively have in excess of 70 known cellular substrates (Kolch, 2005). The pathway controls a diverse range of cellular processes, including cell proliferation, survival, differentiation, apoptosis, motility and metabolism, all of which are intrinsically linked to the characteristics of a cancer cell (Kolch, 2005). The classical MAPK pathway in which the RAF kinases reside is activated by a variety of extracellular ligands, including epidermal growth factor (EGF), fibroblast growth factor (FGF) and platelet derived growth factor (PDGF). Activation of the MAPK pathway in PA has recently been suggested to result in oncogene-induced senescence, whereby an initial stage of transformation is followed by cellular senescence as a result of constitutive MAPK signalling (Jacob et al., 2011; Raabe et al., 2011). This is consistent with the generally slow-growing nature of PAs and may account for the lack of progression to higher grade tumors. The ERK/MAPK pathway is estimated to be deregulated in approximately one-third of all human cancers (Dhillon et al., 2007). Deregulation of the pathway has been reported at all levels of the signalling cascade, including the sustained expression of autocrine or paracrine ligand production, receptor tyrosine kinase activation, Ras activation, Raf activation or direct amplification or deregulation of its nuclear transcription targets such as the AP-1 transcription factor complex and the proto-oncogene c-Myc. In contrast to the RAF fusion genes identified in PAs, upregulation of the MAPK pathway in higher grade astrocytoma is most

12 RAF Fusion Genes and MAPK Activation in Pilocytic Astrocytomas

commonly achieved via alternative mechanisms, such as overexpression of the tyrosine kinase receptors for EGF and PDGF in WHO grade IV glioblastoma multiforme (Mischel and Cloughesy, 2003). This further highlights the specificity of the RAF fusion genes to PA development. Due to the high frequency of tumors displaying aberrant MAPK signalling and the multiple levels that form the signalling cascade, this pathway is an attractive target for therapeutic intervention in cancer. The identification of MAPK activation in PA suggests targeted therapy against the MAPK pathway may be a useful treatment option for these tumors, particularly in those patients who develop recurrent disease, or have tumors that are in areas inaccessible to surgery. Current drugs that target the MAPK pathway and are used in the clinic include Sorafenib, which is a broad spectrum Raf kinase inhibitor with crossreactivity against PDGF and the vascular endothelial growth factor receptor 2 (Dhillon et al., 2007). Sorafenib is efficacious in certain cancer types such as renal cell carcinoma, however its anti-tumor activity is believed to stem in part from off-target effects on vascularisation rather than disruption of the Raf kinases per se (Dhillon et al., 2007). Early-phase clinical trials investigating the efficacy of Sorafenib in paediatric cases of PA are currently being planned. With the recent advances of targeted therapies, inhibitors that act on a specific molecular change are also under development. One of the most promising candidates to date is the PLX4032 inhibitor which shows potent activity in-vitro against the BRAF V600E mutation (Bollag et al., 2010). This compound is currently in phase III clinical trials and has shown encouraging clinical results in patients with late-stage metastatic melanoma, in which the V600E mutation is a frequent event (Flaherty et al., 2010). If this compound is licensed for the clinic, it could potentially be used in the treatment of PA with the V600E mutation, particularly as this mutation correlates with PA arising in an extra-cerebellar location which can be more difficult to treat with surgery (Schindler et al., 2011). MEK inhibitors such as AZD6244 are also in preclinical assessment and would have the benefit of working on a downstream target of Raf, thereby showing potential efficacy regardless of the upstream alteration. A caveat to these exciting new approaches however, is that tumors can develop resistance to highly specific targeted drugs, particularly in complex pathways

103

such as the MAPK pathway where activation of an alternate level in the signalling cascade can bypass the original molecular event. Furthermore, the blood brain barrier is a major obstacle in the delivery of drugs to the brain and the formulation of new and existing compounds must cross this barrier for effective delivery to the tumor. The recent development of the first mouse model of PA, where expression of the V600E-mutant BRAF kinase domain in neural precursor cells gives rise to a slow-growing tumor resembling human PA, should prove a useful tool in the preclinical assessment of these novel therapeutic agents (Gronych et al., 2011).

Summary In conclusion, activation of the MAPK pathway underpins the development of PA and occurs most frequently via the formation of fusion genes involving the RAF kinases. The fusion genes are highly specific to PA when compared to other types of brain cancer, which affords a diagnostic opportunity for the improved classification of PA in clinical samples. This is particularly important given the varied morphology that PAs can exhibit and the less aggressive treatment course that is followed for PA in comparison to higher grade brain tumors, which usually necessitate adjuvant treatment to try to control the disease. In addition to the potential diagnostic utility of RAF fusion genes in PA, new therapeutic options are also indicated. In the short term, drugs already designed to target BRAF or downstream targets (such as MEK) are now entering preclinical and early clinical trials for PA. In the longer term, novel targeted agents against MAPK pathway members and/or the BRAF fusion itself should be beneficial for a substantial number of patients with PA.

References Agamanolis DP, Malone JM (1995) Chromosomal abnormalities in 47 pediatric brain tumors. Cancer Genet Cytogenet 81:125–134 Armstrong GT, Conklin HM, Huang S, Srivastava D, Sanford R, Ellison DW, Merchant TE, Hudson MM, Hoehn ME, Robison LL, Gajjar A, Morris EB (2011) Survival and longterm health and cognitive outcomes after low-grade glioma. Neuro Oncol 13:223–234

104 Bigner SH, McLendon RE, Fuchs H, McKeever PE, Friedman HS (1997) Chromosomal characteristics of childhood brain tumors. Cancer Genet Cytogenet 97:125–134 Bollag G, Hirth P, Tsai J, Zhang J, Ibrahim PN, Cho H, Spevak W, Zhang C, Zhang Y, Habets G, Burton EA, Wong B, Tsang G, West BL, Powell B, Shellooe R, Marimuthu A, Nguyen H, Zhang KY, Artis DR, Schlessinger J, Su F, Higgins B, Iyer R, D’Andrea K, Koehler A, Stumm M, Lin PS, Lee RJ, Grippo J, Puzanov I, Kim KB, Ribas A, McArthur GA, Sosman JA, Chapman PB, Flaherty KT, Xu X, Nathanson KL, Nolop K (2010) Clinical efficacy of a RAF inhibitor needs broad target blockade in BRAF-mutant melanoma. Nature 467:596–599 CBTRUS (2011) CBTRUS statistical report: primary brain and central nervous system tumors diagnosed in the United States in 2004–2007. Source: Central Brain Tumor Registry of the United States, Hinsdale, IL. website: www.cbtrus.org Ciampi R, Knauf JA, Kerler R, Gandhi M, Zhu Z, Nikiforova MN, Rabes HM, Fagin JA, Nikiforov YE (2005) Oncogenic AKAP9-BRAF fusion is a novel mechanism of MAPK pathway activation in thyroid cancer. J Clin Invest 115:94–101 Cin H, Meyer C, Herr R, Janzarik WG, Lambert SR, Jones DT, Jacob K, Benner A, Witt H, Remke M, Bender S, Falkenstein F, Van Anh TN, Olbrich H, von Deimling A, Pekrun A, Kulozik AE, Gnekow A, Scheurlen W, Witt O, Omran H, Jabado N, Collins VP, Brummer T, Marschalek R, Lichter P, Korshunov A, Pfister SM (2011) Oncogenic FAM131BBRAF fusion resulting from 7q34 deletion comprises an alternative mechanism of MAPK pathway activation in pilocytic astroctyoma. Acta Neuropathol 121:763–774 Davies H, Bignell GR, Cox C, Stephens P, Edkins S, Clegg S, Teague J, Woffendin H, Garnett MJ, Bottomley W, Davis N, Dicks E, Ewing R, Floyd Y, Gray K, Hall S, Hawes R, Hughes J, Kosmidou V, Menzies A, Mould C, Parker A, Stevens C, Watt S, Hooper S, Wilson R, Jayatilake H, Gusterson BA, Cooper C, Shipley J, Hargrave D, PritchardJones K, Maitland N, Chenevix-Trench G, Riggins GJ, Bigner DD, Palmieri G, Cossu A, Flanagan A, Nicholson A, Ho JW, Leung SY, Yuen ST, Weber BL, Seigler HF, Darrow TL, Paterson H, Marais R, Marshall CJ, Wooster R, Stratton MR, Futreal PA (2002) Mutations of the BRAF gene in human cancer. Nature 417:949–954 Dhillon AS, Hagan S, Rath O, Kolch W (2007) MAP kinase signalling pathways in cancer. Oncogene 26:3279–3290 Dirven CM, Mooij JJ, Molenaar WM (1997) Cerebellar pilocytic astrocytoma: a treatment protocol based upon analysis of 73 cases and a review of the literature. Childs Nerv Syst 13: 17–23 Eisenhardt AE, Olbrich H, Roring M, Janzarik W, Van Anh TN, Cin H, Remke M, Witt H, Korshunov A, Pfister SM, Omran H, Brummer T (2010) Functional characterization of a BRAF insertion mutant associated with pilocytic astrocytoma. Int J Cancer Endris V, Wogatzky B, Leimer U, Bartsch D, Zatyka M, Latif F, Maher ER, Tariverdian G, Kirsch S, Karch D, Rappold GA (2002) The novel Rho-GTPase activating gene MEGAP/ srGAP3 has a putative role in severe mental retardation. Proc Natl Acad Sci USA 99:11754–11759 Flaherty KT, Puzanov I, Kim KB, Ribas A, McArthur GA, Sosman JA, O’Dwyer PJ, Lee RJ, Grippo JF, Nolop K, Chapman PB (2010) Inhibition of mutated, activated BRAF in metastatic melanoma. N Engl J Med 363:809–819

S.R. Lambert and D.T.W. Jones Forshew T, Tatevossian RG, Lawson AR, Ma J, Neale G, Ogunkolade BW, Jones TA, Aarum J, Dalton J, Bailey S, Chaplin T, Carter RL, Gajjar A, Broniscer A, Young BD, Ellison DW, Sheer D (2009) Activation of the ERK/MAPK pathway: a signature genetic defect in posterior fossa pilocytic astrocytomas. J Pathol 218:172–181 Gronych J, Korshunov A, Bageritz A, Milde T, Jugold M, Hambardzumyan D, Remke M, Hartmann C, Witt H, Jones DT, Witt O, Heiland S, Bendszus M, Holland EC, Pfister S, Lichter P (2011) Mutated BRAF kinase domain alone is sufficient to induce pilocytic astrocytoma in mice. J Clin Invest 121:1344–1348 Jacob K, Albrecht S, Sollier C, Faury D, Sader E, Montpetit A, Serre D, Hauser P, Garami M, Bognar L, Hanzely Z, Montes JL, Atkinson J, Farmer JP, Bouffet E, Hawkins C, Tabori U, Jabado N (2009) Duplication of 7q34 is specific to juvenile pilocytic astrocytomas and a hallmark of cerebellar and optic pathway tumours. Br J Cancer 101:722–733 Jacob K, Quang-Khuong DA, Jones DT, Witt H, Lambert SR, Albrecht S, Witt O, Vezina C, Shirinian M, Faury D, Garami M, Hauser P, Klekner A, Bognar L, Farmer JP, Montes JL, Atkinson J, Hawkins C, Korshunov A, Collins VP, Pfister SM, Tabori U, Jabado N (2011) Genetic aberrations leading to MAPK pathway activation mediate oncogene-induced senescence in sporadic pilocytic astrocytomas. Clin Cancer Res 17:4650–4660 Jones DT, Ichimura K, Liu L, Pearson DM, Plant K, Collins VP (2006) Genomic analysis of pilocytic astrocytomas at 0.97 Mb resolution shows an increasing tendency toward chromosomal copy number change with age. J Neuropathol Exp Neurol 65:1049–1058 Jones DT, Kocialkowski S, Liu L, Pearson DM, Backlund LM, Ichimura K, Collins VP (2008) Tandem duplication producing a novel oncogenic BRAF fusion gene defines the majority of pilocytic astrocytomas. Cancer Res 68:8673–8677 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 fusion in activating the MAPK pathway in pilocytic astrocytoma. Oncogene 28:2119–2123 Kolch W (2005) Coordinating ERK/MAPK signalling through scaffolds and inhibitors. Nat Rev Mol Cell Biol 6:827–837 Lawson AR, Tatevossian RG, Phipps KP, Picker SR, Michalski A, Sheer D, Jacques TS, Forshew T (2010) RAF gene fusions are specific to pilocytic astrocytoma in a broad paediatric brain tumour cohort. Acta Neuropathol 120:271–273 Lawson AR, Hindley GF, Forshew T, Tatevossian RG, Jamie GA, Kelly GP, Neale GA, Ma J, Jones TA, Ellison DW, Sheer D (2011) RAF gene fusion breakpoints in pediatric brain tumors are characterized by significant enrichment of sequence microhomology. Genome Res 21:505–514 Listernick R, Charrow J, Gutmann DH (1999) Intracranial gliomas in neurofibromatosis type 1. Am J Med Genet 89:38–44 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 Mischel PS, Cloughesy TF (2003) Targeted molecular therapy of GBM. Brain Pathol 13:52–61

12 RAF Fusion Genes and MAPK Activation in Pilocytic Astrocytomas Ohgaki H, Kleihues P (2005) Epidemiology and etiology of gliomas. Acta Neuropathol 109:93–108 Palanisamy N, Ateeq B, Kalyana-Sundaram S, Pflueger D, Ramnarayanan K, Shankar S, Han B, Cao Q, Cao X, Suleman K, Kumar-Sinha C, Dhanasekaran SM, Chen YB, Esgueva R, Banerjee S, LaFargue CJ, Siddiqui J, Demichelis F, Moeller P, Bismar TA, Kuefer R, Fullen DR, Johnson TM, Greenson JK, Giordano TJ, Tan P, Tomlins SA, Varambally S, Rubin MA, Maher CA, Chinnaiyan AM (2010) Rearrangements of the RAF kinase pathway in prostate cancer, gastric cancer and melanoma. Nat Med 16:793–798 Raabe EH, Lim KS, Kim JM, Meeker A, Mao XG, Nikkhah G, Maciaczyk J, Kahlert U, Jain D, Bar E, Cohen KJ, Eberhart CG (2011) BRAF activation induces transformation and then senescence in human neural stem cells: a pilocytic astrocytoma model. Clin Cancer Res 17:3590–3599 Schiffman JD, Hodgson JG, VandenBerg SR, Flaherty P, Polley MY, Yu M, Fisher PG, Rowitch DH, Ford JM, Berger MS, Ji H, Gutmann DH, James CD (2010) Oncogenic BRAF mutation with CDKN2A inactivation is characteristic of a subset of pediatric malignant astrocytomas. Cancer Res 70: 512–519 Schindler G, Capper D, Meyer J, Janzarik W, Omran H, HeroldMende C, Schmieder K, Wesseling P, Mawrin C, Hasselblatt M, Louis DN, Korshunov A, Pfister S, Hartmann C, Paulus W, Reifenberger G, von Deimling A (2011) Analysis of BRAF V600E mutation in 1,320 nervous system tumors reveals high mutation frequencies in pleomorphic xanthoastrocytoma, ganglioglioma and extra-cerebellar pilocytic astrocytoma. Acta Neuropathol 121:397–405 Schrock E, Blume C, Meffert MC, du Manoir S, Bersch W, Kiessling M, Lozanowa T, Thiel G, Witkowski R, Ried T,

105

Cremer T (1996) Recurrent gain of chromosome arm 7q in low-grade astrocytic tumors studied by comparative genomic hybridization. Genes Chromosomes Cancer 15: 199–205 Shen MH, Harper PS, Upadhyaya M (1996) Molecular genetics of neurofibromatosis type 1 (NF1. J Med Genet 33:2–17 Sievert AJ, Jackson EM, Gai X, Hakonarson H, Judkins AR, Resnick AC, Sutton LN, Storm PB, Shaikh TH, Biegel JA (2009) Duplication of 7q34 in pediatric low-grade astrocytomas detected by high-density single-nucleotide polymorphism-based genotype arrays results in a novel BRAF fusion gene. Brain Pathol 19:449–458 Tihan T, Fisher PG, Kepner JL, Godfraind C, McComb RD, Goldthwaite PT, Burger PC (1999) Pediatric astrocytomas with monomorphous pilomyxoid features and a less favorable outcome. J Neuropathol Exp Neurol 58:1061–1068 Tomlinson FH, Scheithauer BW, Hayostek CJ, Parisi JE, Meyer FB, Shaw EG, Weiland TL, Katzmann JA, Jack CR Jr. (1994) The significance of atypia and histologic malignancy in pilocytic astrocytoma of the cerebellum: a clinicopathologic and flow cytometric study. J Child Neurol 9:301–310 von Deimling A, Krone W, Menon AG (1995) Neurofibromatosis type 1: pathology, clinical features and molecular genetics. Brain Pathol 5:153–162 Walker C, Joyce KA, Thompson-Hehir J, Davies MP, Gibbs FE, Halliwell N, Lloyd BH, Machell Y, Roebuck MM, Salisbury J, Sibson DR, Du Plessis D, Broome J, Rossi ML (2001) Characterisation of molecular alterations in microdissected archival gliomas. Acta Neuropathol 101:321–333 White FV, Anthony DC, Yunis EJ, Tarbell NJ, Scott RM, Schofield DE (1995) Nonrandom chromosomal gains in pilocytic astrocytomas of childhood. Hum Pathol 26: 979–986

Chapter 13

Biomarker Discovery in Central Nervous System Neoplasms: Past, Present and Future Anne F. Buckley, Roger E. McLendon, Carol J. Wikstrand, and Darell D. Bigner

Abstract Brain tumor biomarkers have long been used as diagnostic tools; they are now finally being brought to bear as therapeutic elements in the fight against brain cancer. Because of the heterogeneity of glial brain tumors, it is clear that no one marker will ever define or defeat them: this has been understood since early days of the search for brain tumor biomarkers. In this chapter, therefore, we provide a context for the trajectory of current research by first discussing the history of brain tumor biomarker research. This is followed by a discussion of the immunologic, cytogenetic, biochemical, and molecular approaches that are currently being used to discover, characterize, and exploit the aberrant features of brain tumor cells. Major biomarkers that have been discovered during decades of research are covered in the earlier sections: gangliosides, tenascin, EGFR and its mutant form EGFRvIII, and GLI. Later sections describe newer, high-thoughput analyses such as gene expression analysis and integrated genome analysis, and outline their role in the discovery of more recent promising biomarkers including doublecortin, osteonectin, semaphoring3B, OTX2, and IDH1 and IDH2. A combination of old and new approaches that harnesses the power of multi-institutional and international cooperation has greatly accelerated the rate of

R.E. McLendon () Section(s) of Surgical Pathology, Duke University Medical Center, Durham, NC 27710, USA e-mail: [email protected]

discovery in this field, and is bringing us closer to our goal of finding cures for brain cancers. Keywords Neurooncology · Biomarkers · EGFR · Ganglioside patterns · Monoclonal antibodies

Introduction Neurooncology has been practiced at our institution since the opening of Duke Hospital in the 1930s. Drs. Deryl Hart and Clarence Gardner, general surgeons who had been taught neurosurgery during their training at Johns Hopkins, began with surgical treatment of brain tumors. The large volume of patients and the huge amount of work required led in 1937 to the recruitment of Dr. Barnes Woodhall, who established a Division of Neurosurgery and intensified brain tumor research efforts. In 1942, while Dr. Woodhall was on military service, Dr. Guy Odom came from the Montreal Neurological Institute and established the Brain Tumor Clinic, the Brain Tumor Neuropathology Laboratory, and an active brain tumor service. Upon Dr. Woodhall’s return at the end of World War II, some of the earliest efforts at brain tumor chemotherapy were carried out based on his wartime experiences with nitrogen mustard. These efforts intensified in the 1950s as Dr. Stephen Mahaley, a Duke neurosurgical resident at the time, continued the work on regional chemotherapy. In the 1960s, Dr. Mahaley branched out into antibody localization studies in collaboration with Dr. Eugene Day. Dr. Darell Bigner joined the effort in 1964, working on viral oncogenesis in brain tumors. He also began to develop brain tumor cell cultures and xenografts as a source of brain tumor tissue and models

M.A. Hayat (ed.), Tumors of the Central Nervous System, Volume 5, DOI 10.1007/978-94-007-2019-0_13, © Springer Science+Business Media B.V. 2012

107

108

on which to do research. In the early 1980s, seeing the need for fresh and snap-frozen brain tumor tissue for neuro-oncologic studies, Dr. Bigner established the Brain Tumor Tissue Bank at Duke, which has proven to be an invaluable source of tissue for research. Dr. Bigner has also fostered numerous fruitful collaborations with cancer researchers in and outside the U.S., with the view that this would more rapidly advance the field of biomarker discovery. This chapter is organized as an approximate chronology of the techniques we have used in our discovery of various biomarkers, each of which is briefly discussed to provide the reader with a sense of how we developed our approaches (increasingly in productive collaboration with other groups) to advance the search for and use of new brain tumor markers. Space limitations preclude our citing all of the original research papers, but these can be found within articles in the reference list.

Methods of Biomarker Discovery Immunologic Methods: Initial Work on Neuroectodermal Tumors Immunology was recognized early on in the twentieth century as an attractive approach to defining antigenic biomarkers of tumors, and from that came the idea of turning antibodies into weapons to use against cancer. As far back as 1905, Paul Erlich proposed using antibodies to deliver chemotherapeutic agents. Multiple groups in the 1960s demonstrated immune responses to brain tumors arising naturally in human patients and showed that this was, to some extent, cell-mediated. At Duke, Mahaley et al. (1965) used lysates of human gliomas to produce antisera in rabbits. These antisera localized in vivo to human brain tumor cells, as assessed by autoradiography of 125 I-labeled antisera applied to tumor tissue from patients who had been treated with the antisera by intracarotid injection a few days before tumor resection. This was a powerful demonstration of the ability of antibodies to target tumors in vivo. At that time, it was not known whether glioma cells had quantitatively or qualitatively distinctive surface antigens. Westermark’s group developed antisera

A.F. Buckley et al.

in rabbits against a glioblastoma and found reactivity specific for malignant glioma cell lines (Wahlstrom et al., 1974). At Duke, we carried out a systematic examination of the surface antigen profiles of human glial tumors, using a large panel of glioma cell lines (Wikstrand et al., 1977). Antisera were raised in nonhuman primates using multiple glioblastoma tissues and cell lines, and high-titer sera were exhaustively preabsorbed with nongliomatous material including normal adult and fetal brain tissues. These antisera were then applied to human glial tumor cells. As measured by indirect live-cell membrane immunofluorescence and [14 C]nicotinamide release (a measure of complement activation by bound antibodies), the antisera bound to human glioma cells but not to normal adult brain. Various absorbed sera did demonstrate reactivity with some nonglial tumor tissues. In addition, the presence of moieties shared by fetal brain tissue and glial tumors, but not by normal adult brain, was demonstrated, presaging the description of oncofetal antigens in multiple tumor types. However, the use of antisera requires multiple adsorption steps to remove non-specific antibodies, and even when purified, these polyvalent antisera react against multiple antigens and multiple portions of each antigen, which limits their use in biomarker discovery. The development of monoclonal antibodies (MAbs) by Kohler and Milstein 1975 was a tremendous advance, making it possible to produce theoretically endless quantities of antibodies directed against not only a single molecule, but against a specific portion of that molecule. Their potential for exquisite specificity means that MAbs remain a mainstay of biomarker research today, not only in discovery, but in the development of tumor-specific therapeutics. Following the introduction of Mab technology, several groups reported the generation of Mabs directed against glioma-associated cell surface molecules. JeanPierre Mach, Stefan Carrel and Nicolas de Tribolet at the Ludwig Institute for Cancer Research in Lausanne, Hugh Coakham and John Kemshead from the Imperial Cancer Research Fund Laboratories in Bristol and London, and our group at Duke characterized Mabs generated by immunization of mice with glioma cell lines, fetal brain tissue, and neuroectodermal tissue (Schnegg et al., 1981; Carrel et al., 1982; Wikstrand et al., 1982; Allan et al., 1983). These studies culminated in a joint publication by these groups (Wikstrand et al., 1987) comparing the in vivo localization

13 Biomarker Discovery in Central Nervous System Neoplasms: Past, Present and Future

capacities of these Mabs, as a prelude to evaluating glioma-specific Mabs as immunodiagnostic and immunotherapeutic agents (Vick et al., 1987; Lee et al., 1987).

Immunologic and Biochemical Methods: Gangliosides The search for brain tumor biomarkers has always been intertwined with the identification of markers for normal brain cells, especially in the context of brain development because of parallels between the behavior of developing cells and cancer cells. Mabs were used in this context to exploit the variety of gangliosides, which are glycolipids characterized by their content of sialic acid (sialoglycolipids). These antigens are localized to the outer leaflet of the plasma membrane, so they are very accessible to antibodies. In the mid-1980s the Duke Brain Tumor Center entered into a collaboration with Lars Svennerholm, Pam Fredman and their colleagues at the University of Gothenburg and the Karolinska Hospital in Sweden, to pursue the discovery of tumor-specific gangliosides. This was an attractive approach because many of the surface antigens thus far identified by MAbs had turned out to be carbohydrate structures of glycoproteins and glycolipids, leading to the idea that carbohydrate modifications might have specific functions in growth and differentiation. Neoplastic transformation is, in fact, accompanied by changes in ganglioside patterns (Feizi, 1985). One of the first studies on fetal ganglioside antigens associated with human gliomas was published by Svennerholm’s group, who conducted a quantitative analysis of lipids extracted from glioma tissue to assess differences in ganglioside content between normal and tumor tissue (Fredman et al., 1986). They found that the major gangliosides GM1, GD1a, and GT1b were markedly reduced in tumor tissue; in contrast, there was an increase in the gangliosides GM3 and GD3. They also reported structurally-unidentified gangliosides, both mono- and oligosialylated, which could not be detected in normal brain tissue. This work suggested that these novel gangliosides were tumor-associated and might be usefully detected by MAbs. However, a problem with isolating gangliosides from primary gliomas is that the borders of infiltrative

109

tumors are so ill-defined that normal brain cells are a significant source of contaminant gangliosides. In order to reduce the contribution of normal glial cell constituents, we used a cell line, D-54 MG, which had been established as a serially-transplantable subcutaneous xenograft glioma line in athymic mice. Lipids extracted from this glioma were analyzed by methods that included gas-liquid chromatography and fast-atom bombardment mass spectrometry. This work led to further characterization of glioma-associated gangliosides in the lactotetraose series: the sialyllactotetraosyl-ceramide 3 -isoLM1 and its disialylated form 3 6 -isoLD1 (Mansson et al., 1986). Detailed studies of the distribution of tumorassociated gangliosides were made possible by the development of MAbs specific for different forms of tumor-associated gangliosides. For example, MAbs to 3 6 -isoLD1 were selected from clones raised against a teratoma cell line, and screened with ganglioside fractions of D-54 MG glioma cells (Wikstrand et al., 1993). Using these MAbs, 3 6 -isoLD1 in the CNS was found to be restricted to proliferating astroglia (fetal/neonatal, reactive and neoplastic), which indicated that specific MAbs against that biomarker could have clinical applications. Significantly, expression of 3 6 -isoLD1 was not seen in cultured CNS tumor cells, even though it was increased in tumor cells of the same origin grown in vivo. This was a useful reminder that biomarker expression can be strongly influenced by the model system used for discovery. Rolf Bjerkvig’s group showed that 3 -isoLM1 was highly expressed on invasive glioma cells far from the presenting main mass of tumor cells (Hedberg et al., 2001). In other studies, MAbs to gangliosides were used to assess the presence of fetal ganglioside antigens in newly-excised human glioma tissue and in brain adjacent to the tumors (von Holst et al., 1997). Lipids were extracted from excised human tumor tissue and separated by anion-exchange chromatography. Individual glycolipids were separated by thin-layer chromatography and analyzed with MAbs in a quantitative TLCELISA combined with densitometric scanning. The fetal gangliosides 3 -isoLM1, 3 6 -isoLD1, and a third ganglioside normally found in adult tissue, GD3, were frequently increased both in tumor and in the adjacent brain, a finding that pointed to the presence of individual, grossly-undetectable tumor cells infiltrating beyond the edge of the tumor.

110

This work suggested that gangliosides could be used as monitoring agents in combination with PET scans to determine the extent of glioma infiltration, and that they might also be targets for immunotherapy. Gangliosides, however, are not T-cell-dependent antigens, so immunization with them generally results in the induction of IgM antibodies of low or moderate affinity. Monoclonal IgM antibodies to GD3 can inhibit the proliferation of human malignant glioma cells in vitro (Hedberg et al., 2000), but the large size of the IgM molecule precludes its use in therapeutic approaches. IgG3 antibodies to gangliosides GD2 and GD3 have been produced, presumably by the activation of γδ T-cells (Honsik et al., 1986), and a phase I trial in patients with neuroblastoma and malignant melanoma using an IgG3 anti-GD2 antibody generated in mice showed some efficacy (Cheung et al., 1987). The Bigner laboratory is currently using LC3 synthase knockout mice and naïve human libraries to prepare IgG antibodies and single fragment chain antibody fragments, specifically reactive with 3 -isoLM1 and 3 6 -isoLD1 (unpublished data).

Immunologic Methods: Tenascin Biomarkers do not have to be of the tumor cells themselves: by the late 1970s it was becoming increasingly clear that the development and invasion of tumors might be influenced, and perhaps directed, by the nonneoplastic microenvironment in which the tumor develops. An example of a novel tumor-associated stromal biomarker discovered by immunologic means is tenascin, a protein discovered independently by several laboratories in the 1980s (Erickson and Bourdon, 1989). Bourdon et al. (1983) in our laboratory exploited the specificity of monoclonal antibodies to define new antigens of interest in glial tumors. They immunized mice with intact cells from the human glioblastoma line U-251 MG, generating MAbs which were tested by CS-RIA for reactivity to the glioma cell line and an osteogenic sarcoma line. Only those with restricted reaction to the glial line were selected for further screening. The MAb 81C6 was chosen for full screening because it had a high binding ratio against U251 MG, and did not react to carcinoma cell lines

A.F. Buckley et al.

in CS-RIA. A large panel of glioma cell lines and other neuroectodermal tumors, as well as other nonneuroectodermal tumors and normal human tissues, were then used to screen for specificity of the new Mab. The antibody did not bind to normal brain or tumors of non-neuroectodermal origin. Antigen localization was examined by immunofluorescence on cell monolayers and suspended layers, revealing that 81C6 does not react with the glioma cells themselves, but with a component of their extracellular matrix (ECM). Analysis of frozen tissue sections with peroxidaseantiperoxidase immunohistology showed that 81C6 localizes to basement membranes of abnormal blood vessels associated with gliomas. This staining pattern was not that of any of the known ECM antigens for which antibodies were available, and 81C6 immunoprecipitated an extracellular matrix protein of Mr 230,000. We referred to the new tumor-stromal antigen as glioma-mesenchymal extracellular matrix (GMEM) antigen. Another group (Chiquet-Ehrismann et al., 1986) used a polyclonal antibody raised against chicken myotendinous antigen (which they had discovered a few years earlier in normal chick embryos) on ENU-induced mammary adenocarcinomas in mice. They found a prominent reaction in the stroma of these tumors, and named the antigen tenascin (a combination of the Latin words tenere (to hold) and nascere (to be born)) to describe its location and developmental expression pattern. The fact that this protein is expressed in development, and again in wound healing and oncogenesis, makes it a very interesting molecule. It is expressed in many tumors, and seems to be more prominent in those that are anaplastic, or less-differentiated. For example, it is almost always found in glioblastoma multiforme (GBM), a WHO grade IV glioma, but is infrequent in more differentiated astrocytomas (Erickson and Bourdon, 1989). Currently, clinical trials are testing the safety and efficacy of antitenascin antibodies labeled with radioactive iodine in the treatment of glioblastomas (http://clinicaltrials. gov/ct2/show/NCT00615186). The MAb used in the commercial preparation is the same one developed by Bourdon et al. (1983): 81C6. Data from the phase II trials of this agent are promising, reaffirming the potential value of immunologically-defined tumorassociated stromal biomarkers.

13 Biomarker Discovery in Central Nervous System Neoplasms: Past, Present and Future

Cytogenetic Methods: Karyotypic Analysis Bigner et al. (1984), working in the Duke Brain Tumor Center, began to explore the cytogenetic abnormalities of astrocytic gliomas using karyotypic analysis. One of the first applications of this method of genetic analysis was in human malignancies, and it provided a means of identifying tumor markers that is still widely used in diagnosis and monitoring. Metaphase spreads stained with orcein demonstrated the gross structure of the chromosomal material in cells and showed abnormalities in some tumor cells, including changes in ploidy, gain or loss of whole chromosomes, or structural abnormalities of chromosomal arms, including losses or gains of chromosomal material. More subtle alterations began to emerge when the trypsin-Giemsa banding technique was developed: homogenously-staining regions (chromosomal regions that do not exhibit the usual banding pattern), translocations, insertions, inversions, extra bands, and structurally abnormal marker chromosomes. These aberrations were often characteristic of particular tumors and suggested that the changes were markers of malignant transformation.

Cytogenetics: Double-Minute Chromosomes and EGFR Amplification Double-minute (DM) chromosomes were first demonstrated in a human tumor cell by Spriggs et al. (1962). These structures resemble minute chromosomes, but they appear as paired dots without a centromere. They are unstable and segregate randomly during cell division, so one tumor cell may receive many copies and another, few. Following this observation, many groups demonstrated the presence of double-minute chromosomes in different human tumors, including neurogenic tumors (Mark 1971). Bigner et al. (1984, 1986) thus began a long-term study in collaboration with the Central Hospital in Skovde, Sweden, in which karyotypic profiling was used to determine if there were any patterns in the gross chromosomal changes of malignant gliomas. In the first part of this work, using both direct preparations and short-term culture of human gliomas

111

(9 glioblastomas, 2 anaplastic astrocytomas and a gliosarcoma), they generated G-banding, C-banding and standard metaphase spreads, and compared them to preparations made from blood cells of the same patients (Bigner et al., 1984). They found that a few nonrandom numerical changes were the earliest changes in this tumor type, and could be grouped into three subgroups: those with gonosomal (sex chromosomal) losses, gains of whole copies of chromosome 7 and losses of chromosome 10, or losses of chromosome 22. They also found that approximately half of the gliomas exhibited DM chromosomes. Other preferential chromosomal alterations were also seen, and in examination of an additional 15 gliomas (Bigner et al., 1986), they were able to confirm that deletions and translocation involving 9p were common in high-grade gliomas, as well as rearrangements of chromosomes 1, 6 and 13, and less frequently 7, 11 and 16. EGFR is encoded by the c-erbB1 proto-oncogene and is a transmembrane receptor with an extracellular ligand-binding domain and an intracellular tyrosine kinase. Kondo and Shimizu (1983) established that the locus for the human gene is on chromosome 7. Libermann et al. (1985) had shown elevated EGFR expression in human glioma biopsies as assessed by kinase activity, and demonstrated amplification of EGFR in some of these tumors using densitometric quantification of Southern blots. Following up on some early evidence that suggested a correlation of DM chromosomes with oncogenes, we began a collaboration with Bert Vogelstein in the mid-1980s in order to better characterize the genetic abnormalities noted in gliomas. Structural abnormalities of 7p seemed to be associated with some EGFR abnormalities, but polysomy of chromosome 7 (present in over half of the gliomas) seemed to be unrelated to EGFR amplification. In 33 gliomas, we found the majority of DMs contained markedly increased copy numbers of the oncogene, EGFR, and concluded that DMs are the usual location for amplified genes, usually EGFR (Bigner et al., 1987). At the same time, analyzing 63 primary gliomas using slot-blot and in situ hybridization of an EGFR cDNA probe to tumor samples, Wong et al. (1987) showed that large increases in expression of EGFR are invariably associated with alterations in gene structure, i.e., amplification. It is now known that EGFR expression is found in up to 90% of high-grade gliomas and is an indication of shorter survival time

112

in patients with gliomas (Yuan et al., 2001), so this is a biomarker useful for both diagnostic and prognostic purposes. This strong association of the EGFR gene to chromosomal abnormalities in gliomas made it of great interest and fueled further investigation.

Molecular Biology, Biochemical, and Immunologic Methods: EGFRvIII As discussed, increased expression of EGFR in gliomas is often accompanied by gene rearrangements. It was already known that the viral homolog of EGFR, v-erbB, encodes a truncated version of the receptor that lacks the extracytoplasmic domain. This led to speculation that these proteins are oncogenic because their intracytoplasmic domain is unregulated and constitutively active (Downward et al., 1984). Libermann et al. (1985) had noted a possible rearrangement of the EGFR based on novel restriction fragment polymorphisms on Southern blots and abnormal transcripts on Northern blots. In work done in collaboration with Bert Vogelstein’s laboratory at Johns Hopkins, Humphrey et al. (1988) found a structurally-altered version of EGFR using a combination of enzyme affinity assays, immunoprecipitation, western blots, and immunohistochemistry to determine expression of EGFR in xenografts from eight human gliomas. They then characterized the genetic alterations associated with amplified EGFR using DNA from human glioblastoma lines propagated as xenografts (Wong et al., 1992). The gene was cloned from a library made from a glioma without EGFR rearrangements and a restriction map was generated. Southern blot hybridization with phage clones was done using restriction enzyme-digested DNA from different glioma xenografts. RNAase protection assays were used to determine structures of mRNA transcripts, and then the altered regions were sequenced. They found that the same deletion mutation of EGFR protein (loss of extracytoplasmic domains) occurred in tumors from different patients. That the same alteration was seen in all five of the gliomas studied strongly suggested a role for this mutant receptor (later called EGFRvIII) in tumorigenesis. In order to exploit this tumor-specific alteration, we generated site-specific anti-peptide antibodies against the mutated region of EGFRvIII (Humphrey et al., 1990). A polyclonal antiserum raised against

A.F. Buckley et al.

a 14-amino acid synthetic peptide corresponding to the junction formed by this deletion was found to recognize mutant, but not normal EGFR. The production of this new polyclonal antibody demonstrated that a reagent could be generated that was absolutely specific to the altered sequence resulting from the in-frame deletion. Subsequently, monoclonal versions were developed that were also tumor specific (Wikstrand et al., 1995). The expression of EGFRvIII was also observed in breast and lung cancers, and the antibodies were also specific for these tumors (ibid). Another advantage of this biomarker is the extracellular location of the EGFRvIII mutation, making it theoretically more accessible in vivo for localization and therapeutic targeting. EGFRvIII expression has now been demonstrated in numerous other cancers, including ovarian and prostate (Kuan et al., 2001) and head and neck squamous cell carcinoma (Sok et al., 2006). It is now known that EGFRvIII is the most common mutation of this receptor in human glioblastomas, and we and other groups continue to develop therapies directed against that target, because it holds the promise of being completely specific for tumor cells. In addition to the use of MAbs to EGFRvIII as therapy, mutant peptide is also being exploited as a vaccine. Phase II clinical trials and continuation studies of a commercial peptide vaccine preparation showed that EGFRvIII-specific CD3+CD8+gamma-IFN producing T-cells were induced even with concurrent temozolomide therapy, and improved median survival. Historically-controlled Phase II data suggest that vaccination against EGFRvIII after gross total resection, chemotherapy, and radiation can approximately double progression-free survival and overall survival (http:// clinicaltrials.gov/ct2/show/NCT00458601).

Molecular Biology Methods: GLI In our research we have frequently evaluated brain tumor markers previously identified in other tumor systems, but one of the biggest challenges is to identify novel markers. A logical approach to this is to look for differences between normal and brain tumor cells, and an example of the successful use of this approach is the discovery of the GLI gene (Kinzler et al., 1987) using the precision of molecular biology.

13 Biomarker Discovery in Central Nervous System Neoplasms: Past, Present and Future

By the 1980s, only a few of the genes altered in human cancers had been precisely defined. A glioblastoma cell line, D-259 MG, had already been found to contain numerous DM chromosomes. Roninson’s (1983) denaturation-renaturation gel technique (which detects any sequence repeated more than twenty times per haploid genome) was used to confirm the amplification of genes that had been suggested by the presence of DM chromosomes. The D-259 MG cell line was analyzed and showed evidence of numerous amplifications, which showed up as extra bands when compared to DNA prepared from normal cells. To determine whether a previously-known oncogene was included in the amplified regions, Southern blots were done for twenty oncogenes: none of them showed evidence of amplification. A plasmid library was then constructed using DNA enriched for amplified sequences by the Roninson method, and plasmids bearing unique inserts were used as probes in Southern blots to confirm more than 50-fold amplification of the target gene in D-259 MG cells, as well as in the original biopsy tissue from which the tumor cell line had been derived. The next step was to determine the chromosomal location of the amplified gene by using the plasmids as probes for Southern blots of a panel of mouse-human and hamster-human hybrids; this showed segregation to chromosome 12. The same approach was taken to probe portions of chromosome 12, which indicated that the amplified sequences are normally located on 12(cen-q14.3). This location was confirmed by in situ hybridization to normal metaphase chromosomes. Then the amplification unit was cloned and a genomic library was constructed from D-259 MG cells using a cosmid vector. This was screened for cosmids containing these amplified sequences, and four validated cosmids were used to probe Northern blots of total RNA isolated from D-259 MG cells. All four detected a major 4.8-kB messenger RNA (mRNA) and a minor 1.9-kB mRNA. The new gene was also found to be amplified in another glioma line. The GLI gene was subsequently cloned (Kinzler et al., 1988) using RNA from the D-259 MG line. A plasmid containing a genomic fragment including GLI that hybridized to a mRNA transcript in D-259 MG cells was used to screen an oligo(dT) primed cDNA library prepared from mRNA isolated from the same cell line. Four cDNA clones were found

113

that represented the major GLI transcript based on S1 nuclease or RNAase A protection analysis. These clones were then sequenced, revealing a predicted protein sequence that included five zinc fingers, indicating that this was a new member of the Kruppel family of zinc finger proteins. These DNA-binding transcription factors are vital in the development and function of many different cell lineages, but this was the first demonstration that a Kruppel gene could be altered in cancer. Subsequently, six other GLI family genes were isolated by hybridization of GLI cDNA to human genomic DNA (Ruppert et al., 1988). The GLI gene is evolutionarily conserved, and is expressed in embryonal carcinoma cell lines, which suggests that it may have a role in normal development. This approach (evaluating amplified chromosomal regions using molecular biologic techniques) can thus be used to isolate single genes that have genetic alterations in human cancers, even if they are novel and of unknown function. The continued presence of an amplified gene suggests that it provides a significant growth or survival advantage to the tumor cells because amplifications are normally lost quickly without some sort of selective pressure. Using this approach, we can assess tumor cells for mutations likely to be relevant to their pathology, localize the mutations with some precision to a region of a chromosome, and rapidly isolate and sequence the gene.

Large-Scale Arrays Many advances in biomarker discovery have been based on the discovery of single proteins or genes. Now, however, new technologies allow us to screen tumors for gene mutations, changes in gene copy number, or altered gene expression levels by looking at large numbers of known genes or the entire human genome. Computer software is available to deal with the enormous amount of data that is generated. We now can rapidly develop an extremely detailed picture of individual tumors, and stratify them on the basis of alterations in hundreds of genes or dozens of pathways. We are now able to see patterns of gene expression that may point to new oncogenic pathways, or new connections between known pathways. Largescale analyses thereby promise to greatly accelerate the speed of biomarker discovery.

114

Gene Expression Arrays: Doublecortin, Osteonectin and Semaphorin3B Gene expression arrays are an efficient means of looking at large numbers of known genes. Gene expression patterns can inform the analysis of tumorigenic pathways because they can reflect epigenetic alterations that are not detected by DNA sequencing or copy number analyses. Messenger RNA (mRNA) is purified from tumor and normal tissue, and complementary DNA (cDNA) made from this mRNA is used to probe arrays of human genes to see if the expression level of any of the genes is altered in the tumor relative to normal tissue. Another advantage of gene expression arrays is that they are now quite cheap, and many institutions have core facilities that run the assays and assist in interpretation of the data by biostatisticians using specialized software. The promise of gene expression analysis is that new biologic associations can be discovered between known genes, and those relationships can be used as tumor biomarkers. Current thinking is that primary GBMs present as grade IV tumors without evidence of a preexisting lower grade lesion. They tend to occur in older patients and are more likely to have EGFR amplification, p16INK4A deletions, and deletions of phosphatase and tensin homologue (PTEN). Secondary GBMs arise from lower-grade tumors, and are thought more likely to have mutations of TP53 and RB. However, in spite of these apparent genetic differences, there is overlap in the genetic profiles of primary and secondary GBMs, and there is no significant difference in patient survival between the two when controlled for patient age. Reasoning that analysis of multiple genes would provide a more detailed picture of a tumor’s biologic behavior, we collaborated with Joseph Nevins and Mike West in using gene expression arrays to analyze GBMs from a group of patients over 50 years of age, who would be considered to have poorer prognoses, in order to identify new prognostic markers (Rich et al., 2005). We isolated RNA and DNA from fresh frozen tumor tissue from patients for whom we had good followup data. Twenty of the patients had lived less than a year after diagnosis, twenty had lived longer than 2 years, and ten had survived for 1–2 years. We used tumors only from patients over 50 years of age, in order to enrich for primary GBMs, and looked at gene

A.F. Buckley et al.

expression patterns in relation to survival. Using a gene chip with 20,000 probe sets, we looked for patterns of expression that might indicate a relationship between known genes that would be relevant to survival. Then, using computational statistical methods to define association networks, we looked for additional genes linked to those arising from the primary predictive models. We compared these results to data obtained using more traditional PCR-based molecular analysis of EGFR amplification, EGFR, PTEN, and TP53 mutation, and loss of heterozygosity at 9p, 10q, and 17p. We found that the status of these familiar markers did not associate with patient outcome. Rather, three interesting genes stood out following computational analysis: doublecortin, osteonectin and semaphorin3B. None of them served as a useful marker of survival by itself, but if all three were increased in expression, survival was reduced. The doublecortin gene underlies defects in neuronal migration and neocortical organization. Osteonectin is an extracellular protein involved in development, tissue healing and remodeling, and angiogenesis. Semaphorin 3B is a secreted molecule that regulates neuronal migration. Thus, all three of these genes are involved in regulation of cellular motility (Rich et al., 2005). Given that the most intractable problem in glial tumor treatment is the propensity of these neoplasms to spread diffusely in the CNS, the finding of a survival association with doublecortin, semaphorin3B and osteonectin indicates that tumor migration is a key factor in glioblastoma prognosis, and that these genes may be relevant targets of therapy. Therapeutic agents directed against these molecules would most likely be useful in conjunction with therapies directed against genes regulating proliferation and resistance to apoptosis.

Digital Karyotyping and Gene Expression Analysis: OTX2 The completion in recent years of the human reference genome sequence has led to the development of techniques to measure the genomic copy number (gene dosage) at the level of single genes or handfuls of genes. One of these new techniques is digital karyotyping (Yan and Bigner, 2006). This technique

13 Biomarker Discovery in Central Nervous System Neoplasms: Past, Present and Future

allows gene dosage screening of an entire genome by sequencing representative 21-base pair DNA sequence tags obtained from specific locations in the genome. Tag density can be analyzed to assess the relative genetic content of different loci, with the advantage of high resolution and unbiased gene dosage readout. Digital karyotyping protocols and software for extraction and analysis of genomic tags are available at http:// www.digitalkaryotyping.org. Using this method combined with serial analysis of gene expression (SAGE), Di et al. (2005) constructed a digital karyotyping library from the medulloblastoma cell line D458MED, and evaluated the genomic density of tags. They found subchromosomal regions of amplification and deletion already known to be associated with medulloblastomas, including amplification of the C-MYC oncogene on chromosome 8q24.21, loss of chromosome 17p, and gain of 17q. However, more subtle changes were also seen that would not be detectable using traditional methods. One of these was the 28-fold amplification of a region on chromosome 14q22.3. Examination of a public human genome database (http://genome.ucsc.edu/cgi-bin/hgGateway) identified the full-length homeobox gene OTX2 in that region. Quantitative real-time PCR was used to measure OTX2 mRNA expression in tissue from medulloblastomas and normal cerebellar tissue; this revealed that the gene was amplified more than ten-fold in three medulloblastoma cell lines, a finding confirmed by fluorescent in situ hybridization (FISH). Using SAGE data available in the National Center for Biotechnology Information Cancer Genome Anatomy Project repository (http://cgap.nci.nih.gov/SAGE), OTX2 SAGE tags were shown to be present in 157 tumors and 54 normal tissues. This group also found that OTX2 SAGE tags were most prevalent and showed highest expression in medulloblastomas, whereas they were absent or at low levels in other tumors and normal tissue. Furthermore, OTX2 transcripts were found in most anaplastic medulloblastomas, a more aggressive variant of that primitive neuroectodermal tumor, suggesting a prognostic application for this biomarker. The OTX2 gene is of great interest in medulloblastoma tumorigenesis, as it is one of the homeodomain transcription factors that control developmental programs in brain morphogenesis: OTX2 is expressed in embryonic cerebellum, and is important in its development. The gene had previously been found by other

115

groups to be expressed at high levels in medulloblastomas, but the finding of specific genetic amplifications suggested that the gene is directly involved in medulloblastoma pathogenesis and not merely associated with it. In the course of the same study, they also found that knockdown of OTX2 expression using small interfering RNA (siRNA) inhibited medulloblastoma growth in vitro, and that all-trans retinoic acid showed activity against medulloblastomas expressing OTX2. Thus, this combination of approaches not only identified a novel biomarker of anaplastic medulloblastoma, it also led to possible use of this marker as a target for therapy.

Integrated Genomic Analysis: IDH1 and IDH2 One of the newest approaches to biomarker discovery is demonstrated by Parsons et al. (2008), in which we collaborated with Bert Vogelstein, Kenneth Kinzler, and Victor Velculescu on combined analyses of gene copy number and expression using integrated data analysis. This collaboration involved genomewide sequencing of 20,661 protein-coding genes in 22 GBMs. To complement those data, the presence of focal copy number alterations, including homozygous deletions and amplifications in each tumor’s gene set, were examined using high-density oligonucleotide microarrays. Then the expression profiles of 18 of those GBMs were assessed using SAGE and nextgeneration sequencing technology. This was followed up with a similar analysis of a total of 149 GBMs to validate the findings of the discovery screen. The mutational data thereby obtained were integrated to identify candidate GBM genes that were most likely to provide a selective advantage to the tumor cells (driver mutations), as opposed to mutations that have no net effect on growth (passenger mutations). Many of these candidates were already established in glioma mutagenesis (TP53, PTEN, CDKN2A, RB1, EGFR, NF1, PIK3CA, and PIK3R1), which validated the sensitivity of our approach. However, we also identified multiple genes that were not previously known to be altered in GBMs. One of the most promising of these is the gene encoding isocitrate dehydrogenase 1, IDH1. The gene is on chromosome 2q33, and encodes an enzyme that

116

catalyzes the oxidative carboxylation of isocitrate to a-ketoglutarate to produce nicotinamide adenine dinucleotide phosphate (NADPH). Because of this, IDH1 is thought to play a substantial role in the cellular control of oxidative damage. We found that 12% of the 149 GBMs examined harbored mutations in the active site of IDH1, and that the mutations were recurrent: all mutations were amino acid substitutions at position 132, an evolutionarily-conserved residue in the isocitrate binding site. These R132 IDH1 mutations tended to occur in young patients, in most patients with secondary GBMs, and were associated with better overall survival. This comprehensive approach thus uncovered a novel marker for secondary glioblastomas, which may identify a subpopulation of GBM patients with similar tumor biology. Follow-up studies in which 939 tumor samples were sequenced (Yan et al., 2009) showed that the majority of grade II and grade III astrocytomas and oligodendrogliomas, as well as glioblastomas that develop from these lesions, have R132 mutations of IDH1 or at the analogous R172 residue of the closely-related IDH2 gene. It was found that tumors with IDH1 or IDH2 mutations had distinctive genetic and clinical characteristics, and that patients with these mutations in their tumors had a better outcome compared to those with wild-type IDH genes. These data indicate that IDH mutations play a role in the development of gliomas. It is conceivable that treatments can now be designed to exploit IDH1 and IDH2 mutations in glioma patients. There is a critical need for biomarkers that distinguish reliably between primary and secondary glioblastomas, and one of the best ways to find such markers seems to be large-scale, unbiased genomic analysis of large numbers of patient samples.

Multicenter Comprehensive Genomic Characterization The power of comprehensive genomic characterization combined with an extensive collaborative effort is evident in work published recently by the Cancer Genome Atlas Research Network (2008). This article describes the Cancer Genome Atlas pilot project, the aim of which was to evaluate large-scale, multi-dimensional analyses of multiple molecular

A.F. Buckley et al.

characteristics of glioblastomas. In this pilot study, 206 GBMs were collected from five major neurosurgical centers, sequenced by teams in three genome sequencing centers, and characterized in seven cancer genome characterization centers. The final data were collated by a team of assigned data coordinators. Almost sixty different groups and hundreds of people in various departments, institutes and centers were involved in this enormous effort. Retrospective specimen repositories were screened for newly-diagnosed glioblastomas. The specimens were assessed to ensure a minimum of 80% tumor cells and <50% necrosis. 206 specimens were qualified for copy number, expression and DNA methylation analyses. 143 of these had matched normal tissue and were, therefore, appropriate for re-sequencing. 21 of the 206 were post-treatment glioblastomas, which were used for comparative purposes; the remaining 185 were predominantly primary glioblastomas based on clinical behavior. The work published was an interim analysis of DNA copy number, gene expression and DNA methylation in the 206 GBMs. Genomic copy number alterations (CNAs) were assessed on three types of copy number microarray and analyzed with three algorithms. Initial findings included recurrent alterations not previously reported in glioblastomas, such as homozygous mutations in NF1 and PARK2, and amplifications of AKT3. Transcriptional analyses (using four expression microarrays) integrated with these data showed that ∼76% of genes in recurrent CNAs have expression patterns that correlate with copy number. Whole-genome amplified genomic DNA sequencing of 143 cases demonstrated patterns of mutations in human GBMs, in that eight genes were significantly mutated in 91 cases, including TP53, PTEN, NF1, EGFR, ERBB2, RB1, PIK3R1, and PIK3CA. Of note was the finding that mutations in TP53 occurred in ∼40% of newlydiagnosed GBMs, and that these cluster in the DNA binding domain. The frequency of mutations in NF1 was also remarkable (27% of 206 cases), proving real support for a significant role for this gene in sporadic glioblastomas. Unlike PIK3CA, PIK3R1 has rarely been reported as mutated in human cancers, so this frequent mutation in GBMs is significant, as is the finding of frequent mutations in ERBB2, a gene previously reported as mutated in only one glioblastoma. By measuring cancer-specific DNA methylation of CpG dinucleotides in the promoters of 2,305 genes

13 Biomarker Discovery in Central Nervous System Neoplasms: Past, Present and Future

compared to normal brain DNA (using the Illumina Golden Gate Assay), and integrating those data with somatic mutation results, a link was found also between MGMT methylation status and the hypermutation phenotype. The finding that MMR deficiency and MGMT methylation together in the context of therapy affect the frequency and pattern of somatic point mutations in glioblastomas is potentially of clinical importance. Finally, in order to construct an integrated picture of genetic alterations in GBMs, the unequivocal alterations were mapped onto known gliomagenic pathways. This big-picture analysis revealed a highly interconnected network of aberrations in three major tumor pathways: RTK signaling, and the p53 and RB tumor suppressor pathways. One of the stated aims of this study was to rapidly communicate the data to others in the research community, so all of the data from this study were immediately deposited for public access at the data coordinating center (http://cancergenome.nih. gov/). This is another advantage of this type of cooperative research that will speed advances in biomarker research.

Summary and Future Prospects The great advances in treatment of breast and hematologic cancers are based on agents developed through a precise understanding of the biology of the biomarker being targeted. We have yet to approach this in the field of brain cancer, but we are making progress. Most of the methods described here are still used to some extent in biomarker discovery. Each has its advantages and its drawbacks. Following the discovery of novel biomarkers using the newer techniques, older techniques will continue to be applied to specific molecules in order to fully characterize their role and to develop specific therapies. However, it is clear that the overall trend in the field of cancer biomarker discovery is towards large-scale, unbiased genomic studies. This increase in scale also refers to the inclusion of multiple research groups in diverse institutions. We can generate more statistical power for our studies by combining tumor specimens from more than one tissue bank, and by amplifying other institutional resources. This sort of collaboration will advance the field of biomarker research much more rapidly and

117

effectively than anything done by a single research group in one institution. We have always believed that more can be done if researchers in different institutions work together. The power of numbers, both computationally and humanly speaking, will be increasingly applied to the search for biomarkers in decades to come. The biologic, histologic, and clinical heterogeneity of human brain tumors, particularly those of glial origin, has been emphasized repeatedly in brain tumor studies. This means that a single gene or small groups of genes can never fully characterize their behavior. The idea that tumors are best characterized by multiple rather than single biomarkers has been discussed in relation to surface markers and genes, but the advent of large-scale genomic analysis takes this to a new level, potentially allowing the development of molecular tumor signatures that include hundreds of genes. High-throughput analysis of tumor mRNA, DNA, protein expression and epigenetic events permit the more rapidly identification of novel markers, and provide a more complete picture of individual tumors. Brain tumor biomarkers are already being used diagnostically and therapeutically; new approaches will include the use of biomarkers to reliably individualize the diagnosis and therapy of patients’ tumors, allowing us finally to overcome the problem of glioma heterogeneity.

References Allan PM, Carson JA, Harper EI, Asser U, Coakham HB, Brownell B, Hemshead JT (1983) Biological characterization and clinical applications of a monoclonal antibody recognizing an antigen restricted to neuroectodermal tissues. Int J Cancer 31:592–598 Bigner SH, Mark J, Mahaley SM, Bigner DD (1984) Patterns of the early, gross chromosomal changes in malignant human gliomas. Hereditas 101:103–113 Bigner SH, Mark J, Mahaley SM, Bullard DE, Mahaley MS, Bigner DD (1986) Chromosomal evolution in malignant human gliomas starts with specific and usually numerical deviations. Cancer Genet Cytogenet 22:121–135 Bigner SH, Wong AJ, Mark J, Muhlbaier LH, Kinzler K, Vogelstein B, Bigner DD (1987) Relationship between gene amplification and chromosomal deviations in malignant human gliomas. Cancer Genet Cytogenet 24:163–176 Bourdon MA, Wikstrand CJ, Furthmayr H, Matthews TJ, Bigner DD (1983) Human glioma-mesenchymal extracellular matrix antigen defined by monoclonal antibody. Cancer Res 43:2796–2805

118 Carrel S, Schreyer M, Schmidt-Kessen A, Mach J-P (1982) Reactivity spectrum of 30 monoclonal antimelanoma antibodies to a panel of 28 melanoma and control cell lines. Hybridoma 1:387–397 Cheung NK, Lazarus H, Miraldi FD, Abramowsky CR, Kallick S, Saarinen UM, Spitzer T, Strandjord SE, Coccia PF, Berger NA (1987) Ganglioside GD2 specific monoclonal antibody 3F8: a phase I study in patients with neuroblastoma and malignant melanoma. J Clin Oncol 5:1430–1440 Chiquet-Ehrismann R, Mackie EJ, Pearson CA, Sakakura T (1986) Tenascin: an extracellular matrix protein involved in tissue interactions during fetal development and oncogenesis. Cell 47:131–139 Di C, Liao S, Adamson DC, Parrett TJ, Broderick DK, Shi Q, Lengauer C, Cummins JM, Velculescu VE, Fults DW, McLendon RE, Bigner DD, Yan H (2005) Identification of OTX2 as a medulloblastoma oncogene whose product can be targeted by all-trans retinoic acid. Cancer Res 65:919–924 Downward J, Yarden Y, Mayes E, Scrace G, Totty N, Stockwell P, Ullrich A, Schlessinger J, Waterfield MD (1984) Close similarity of epidermal growth factor receptor and v-erb-B oncogene protein sequences. Nature 307:521–527 Erickson HP, Bourdon MA (1989) Tenascin: an extracellular matrix protein prominent in specialized embryonic tissues and tumors. Ann Rev Cell Biol 5:71–92 Feizi T (1985) Demonstration by monoclonal antibodies that carbohydrate structures of glycoproteins and glycolipids are onco-developmental antigens. Nature 314:53–57 Fredman P, von Holst H, Collins VP, Ammar A, Dellheden B, Wahren B, Granholm L, Svennerholm L (1986) Potential ganglioside antigens associated with human gliomas. Neurol Res 8:123–126 Hedberg KM, Dellheden B, Wikstrand CJ, Fredman P (2000) Monoclonal anti-GD3 antibodies selectively inhibit the proliferation of human malignant glioma cells in vitro. Glycoconj J 17:717–726 Hedberg KM, Mahesparan R, Read T-A, Tysnes BB, Thorsen F, Visted T, Bjerkvig R, Fredman P (2001) The glioma-associated gangliosides 3 isoLM1, GD3 and GM2 show selective area expression in human glioblastoma xenografts in nude rat brains. Neuropathol Appl Neurobiol 27:451–464 Honsik CJ, Jung G, Reisfeld RA (1986) Lymphokine-activated killer cells targeted by monoclonal antibodies to the disialogangliosides GD2 and GD3 specifically lyse human tumor cells of neuroectodermal origin. Proc Natl Acad Sci USA 83:7893–7897 Humphrey PA, Wong AJ, Vogelstein B, Friedman HS, Werner MH, Bigner DD, Bigner SH (1988) Amplification and expression of the epidermal growth factor receptor gene in human glioma xenografts. Cancer Res 48:2231–2238 Humphrey PA, Wong AJ, Vogelstein B, Zalutsky MR, Fuller GN, Archer GE, Friedman HS, Kwatra MM, Bigner SH, Bigner DD (1990) Anti-synthetic peptide antibody reacting at the fusion junction of deletion-mutant epidermal growth factor receptors in human glioblastoma. Proc Natl Acad Sci 87:4207–4211 Kinzler KW, Bigner SH, Bigner DD, Trent JM, Law ML, O’Brien SJ, Wong AJ, Vogelstein B (1987) Identification of an amplified, highly expressed gene in a human glioma. Science 236:70–73

A.F. Buckley et al. Kinzler KW, Ruppert JM, Bigner SH, Vogelstein B (1988) The GLI gene is a member of the Kruppel family of zinc finger proteins. Nature 332:371–374 Kohler G, Milstein C (1975) Continuous culture of fused cells secreting antibody of predefined specificity. Nature 256: 495–497 Kondo I, Shimizu N (1983) Mapping of the human gene for epidermal growth factor receptor (EGFR) on the p13 leads to q22 region of chromosome 7. Cytogenet Cell Genet 35:9–14 Kuan C-T, Wikstrand CJ, Bigner DD (2001) EGF mutant receptor vIII as a molecular target in cancer therapy. Endocr Relat Cancer 8:83–96 Lee Y, Bullard DE, Wikstrand CJ, Zalutsky MR, Muhlbaier LH, Bigner DD (1987) Comparison of monoclonal antibody delivery to intracranial gliomas xenografts by intravenous and intracarotid administration. Cancer Res 47:1941–1946 Libermann TA, Nusbaum HR, Razon N, Kris R, Lax I, Soreq H, Whittle N, Waterfield MD, Ullrich A, Schlessinger J (1985) Amplification, enhanced expression and possible rearrangement of EGF receptor gene in primary human brain tumors of glial origin. Nature 313:144–147 Mahaley SM, Mahaley JL, Day ED (1965) The localization of radioantibodies in human brain tumors II: Radioautography. Cancer Res 25:779–793 Mansson J-E, Fredman P, Bigner DD, Molin K, Rosengren B, Friedman HS, Svennerholm L (1986) Characterization of new gangliosides of the lactotetraose series in murine xenografts of a human glioma cell line. FEBS Lett 3692: 109–113 Mark J (1971) Chromosomal characteristics of neurogenic tumors in adults. Hereditas 68:61–100 Parsons DW, Jones S, Zhang X, Lin JC-H, Leary RJ, Angenendt P, Mankoo P, Carter H, Siu I-M, Gallia GL, Oliviv A, McLendon R, Rasheed BA, Keir S, Nikolskaya T, Nikolsky Y, Busam DA, Tekleab H, Diaz LA Jr, Hartigan J, Smith DR, Strausberg RL, Kazue Nagahashi Marie S, Mieko Oba Shinjo S, Yan H, Riggins GJ, Bigner DD, Karchin R, Papadopolous N, Parmigiani G, Vogelstein B, Velculescu VE, Kinzler KW (2008) An integrated genomic analysis of human glioblastoma multiforme. Science 321:1807–1812 Rich JN, Hans C, Jones B, Iversen ES, McLendon RE, Rasheed BKA, Dobra A, Dressman HK, Bigner DD, Nevins JR, West M (2005) Gene expression profiling and genetic markers in glioblastoma survival. Cancer Res 65:4051–4058 Roninson IB (1983) Detection and mapping of homologous, repeated and amplified DNA sequences by DNA renaturation in agarose gels. Nucleic Acids Res 11:5413–31 Ruppert JM, Kinzler KW, Wong AJ, Bigner SH, Kao F-T, Law ML, Seuanez HN, O’Brien SJ, Vogelstein B (1988) The GLIKruppel family of human genes. Mol Cell Biol 8:3104–3113 Schnegg JF, Diserens AC, Carrel S, Accolla RS, deTribolet N (1981) Human glioma-associated antigens detected by monoclonal antibodies. Cancer Res 41:1201–1213 Sok JC, Coppelli FM, Thomas SM, Lango MN, Xi S, Hunt JL, Freilino ML, Graner. MW, Wikstrand CJ, Bigner DD, Gooding WE, Furnari FB, Grandis JR (2006) Mutant epidermal growth factor receptor (EGFRvIII) contributes to head and neck cancer growth and resistance to EGFR targeting. Clin Cancer Res 12:5064–5073 Spriggs AI, Boddington MM, Clarke CM (1962) Chromosomes of human cancer cells. Br Med J. 2:1431–1435

13 Biomarker Discovery in Central Nervous System Neoplasms: Past, Present and Future The Cancer Genome Atlas Research Network (2008) Comprehensive genomic characterization defines human glioblastoma genes and core pathways. Nature 455:1061–1068 Vick WW, Wikstrand CJ, Bullard DE, Kemshead J, Coakham HB, Schlom J, Johnston WW, Bigner DD, Bigner SH (1987) The use of a panel of monoclonal antibodies in the evaluation of cytologic specimens from the central nervous systems. Acta Cytol 31:815–824 von Holst H, Nygren C, Bostrom K, Collins VP, Fredman P (1997) The presence of foetal ganglioside antigens 3 IsoLM1 and 3 6 -IsoLD1 in both glioma tissue and surrounding areas from human brain. Acta Neurochir 139:141–145 Wahlstrom T, Linder E, Saksela E, Westermark B (1974) Tumorspecific membrane antigens in established cell lines from gliomas. Cancer 34:274–279 Wikstrand CJ, Mahaley MS, Bigner DD (1977) Surface antigenic characteristics of human glial brain tumor cells. Cancer Res 37:4267–4275 Wikstrand CJ, Bourdon MA, Pegram CN, Bigner DD (1982) Human fetal brain antigen expression common to tumors of neuroectodermal tissue origin. J Neuroimmunol 3:43–62 Wikstrand CJ, McLendon RE, Carrel S, Kemshead JT, Mach J-P, Coakham HB, de Tribolet N, Bullard DE, Zalutsky MR, Bigner DD (1987) Comparative localization of gliomareactive monoclonal antibodies in vivo in an athymic mouse human glioma xenograft model. J Neuroimmunol 15:37–56 Wikstrand CJ, Longee DC, McLendon RE, Guller GE, Friedman HS, Fredman P, Svennerholm L, Bigner DD (1993) Lactotetraose series ganglioside 3 ,6 -isoLD1 in tumors of

119

central nervous and other systems in vitro and in vivo. Cancer Res 53:120–126 Wikstrand CJ, Hale LP, Batra SK, Hill ML, Humphrey PA, Kurpad SN, McLendon RE, Moscatello D, Pegram CN, Reist CJ, Traweek ST, Wong AJ, Zalutsky MR, Bigner DD (1995) Monoclonal antibodies against EGFRvIII are tumor specific and react with breast and lung carcinomas and malignant gliomas. Cancer Res 55:3140–3148 Wong AJ, Bigner SH, Bigner DD, Kinzler KW, Hamiltom SR, Vogelstein B (1987) Increased expression of the epidermal growth factor receptor gene in malignant gliomas is invariably associated with gene amplification. Proc Natl Acad Sci USA 84:6899–6903 Wong AJ, Ruppert JM, Bigner SH, Grzeschik CH, Humphrey PA, Bigner DD, Vogelstein B (1992) Structural alterations of the epidermal growth factor receptor gene in human gliomas.. Proc Natl Acad Sci USA 89:2965–2969 Yan H, Parsons DW, Jin G, McLendon RE, 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 gene mutations play a fundamental role in astrocytoma and oligodendroglioma development. New Engl J Med 360:765–773 Yan H, Bigner DD (2006) Digital karyotyping: a powerful tool for cancer gene discovery. In: Dunn M., Jorde L, Little P, Subramanian S (eds) Encyclopedia of genetics, genomics, proteomics and bioinformatics. Wiley, Chichester Yuan C-T, Wikstrand CJ, Bigner DD (2001) EGF mutant receptor vIII as a molecular target in cancer therapy. Endocr Relat Cancer 8:83–96

Chapter 14

Astrocytomas: Role of Taurine in Apoptosis Using Magnetic Resonance Spectroscopy Kirstie S. Opstad

Abstract New cancer therapies are being developed that trigger tumor apoptosis and in vivo methods of apoptotic detection and early treatment response would be of great value. Magnetic resonance spectroscopy (MRS) can determine a tumor’s biochemical profile in vivo and this work investigates whether a specific spectroscopic signature exists for apoptosis in human astrocytomas. High-resolution magic angle spinning (HRMAS) 1 H MRS provides detailed 1 H spectra of brain tumor biopsies for direct correlation with histopathology. Metabolites, mobile lipids and macromolecules were quantified from presaturation HRMAS 1 H spectra acquired from 41 biopsies of grades II (n = 8), III (n = 3) and IV (n = 30) astrocytomas. Subsequently, TUNEL and Haematoxylin & Eosin (H&E) staining provided quantification of apoptosis, cell density and necrosis. Taurine was found to significantly correlate with apoptotic cell density (TUNEL) in both non-necrotic (R = 0.727, p = 0.003) and necrotic (R = 0.626, p = 0.0005) biopsies. However, the ca 2.8 ppm polyunsaturated fatty acid peak, observed in other studies as a marker of apoptosis, correlated only in non-necrotic biopsies (R = 0.705, p < 0.005). Thus, the taurine 1 H MRS signal in astrocytomas may be a robust apoptotic biomarker that is independent of tumor necrotic status. Keywords Apoptosis · HRMAS · Astrocytoma · DNA · Necrotic biopsies

1H

MRS ·

K.S. Opstad () Division of Clinical Sciences, St. George’s, University of London, Cranmer Terrace, London SW17 0RE, UK e-mail: [email protected]

Introduction Apoptosis is a form of programmed cell death by which an organism can rid itself of damaged or unwanted cells. Some of the distinctive features of apoptosis are nuclear condensation, loss of mitochondrial function, DNA fragmentation and cell shrinkage (Morán et al., 2000). Whilst playing a major part in embryogenesis, apoptosis is also important throughout life in maintaining normal tissue homeostasis. In contrast, neoplastic cells accumulate by increased cellular proliferation and decreased cellular turnover, with much evidence showing that apoptosis is inhibited in cancer. Such changes are due to multiple genetic aberrations and several have been shown to promote the aggressive characteristics of high-grade tumors (Rao and James, 2004). Two major pathways of apoptosis have been identified (Hengartner, 2000): (i) the death-receptor pathway triggered by receptors such as CD95 (Fas) and tumor necrosis factor receptor 1α (TNF-1α); and (ii) the mitochondrial pathway in response to extracellular leads and internal insults, such as DNA damage, and involving p53 and the Bcl-2 family. Dysregulation of these apoptosis-regulating proteins in cancer is providing potential new targets for drug discovery and novel approaches for cancer treatments (Steinbach and Weller, 2004). However, if the overall aim of these treatments is to increase apoptosis, there is a need for early assessment of drug response. In particular, a non-invasive method is required for brain tumors. Proton magnetic resonance spectroscopy (1 H MRS) provides a non-invasive metabolic profile

M.A. Hayat (ed.), Tumors of the Central Nervous System, Volume 5, DOI 10.1007/978-94-007-2019-0_14, © Springer Science+Business Media B.V. 2012

121

122

of a tumor’s biochemistry and is easily incorporated into a routine clinical magnetic resonance imaging (MRI) scan to monitor biochemical changes within post-treatment brain tumor tissues. Previous studies investigating drug-induced apoptosis have shown correlations between apoptosis and MRS-visible mobile lipids and cytosolic lipid bodies in several human cancer cell lines (Di Vito et al., 2001; Al-Saffar et al., 2002; Iorio et al., 2003; Milkevitch et al., 2005). Therefore it has been suggested that MRS-detectable lipids accumulate in tumor cells undergoing apoptosis. This observation has also been supported by the in vivo accumulation of 1 H MRS lipids, particularly the polyunsaturated fatty acid (PUFA) resonances at ca. 5.4 and 2.8 ppm, demonstrated in drug-induced apoptosis studies of rodent tumor models (Hakumäki et al., 1999; Griffin et al., 2003; Valonen et al., 2005). However, lipid droplet production is generally believed to occur in both apoptotic and necrotic cell death (Delikatny et al., 2002). High-resolution magic angle spinning (HRMAS) 1 H MRS is increasingly being used to study the biochemical profile of pathological tissue. By spinning intact biopsy samples at the magic angle (≈ 54.7◦ ), the line broadening effects due to dipolar coupling and field inhomogeneities within the sample are reduced, allowing high-resolution spectra to be obtained directly from whole biopsy samples. Thus, using HRMAS 1 H MRS, the understanding of tissue biochemistry can be improved; we can validate and improve interpretation of lower-resolution in vivo 1 H MR spectra (Opstad et al., 2010) and the biopsy samples can be retrieved post-HRMAS for subsequent histological analysis, allowing a direct comparison to be made between the biochemical and histological profile (Wu et al., 2003; Opstad et al., 2008a). The aim of this work was to determine a biochemical correlate of apoptosis in gliomas that could provide a metabolic signature of apoptosis with in vivo 1 H MRS, thus potentially improving the future diagnostic and prognostic capabilities of MRS. The Terminal deoxynucleotidyl Transferase Biotin-dUTP Nick End Labelling (TUNEL) method was used for apoptotic cell detection (Gavrieli et al., 1992) in post-HRMAS tissue, and the number of TUNELpositive nuclei per mm2 compared with the estimated metabolite concentrations for each biopsy, across different grades of glioma.

K.S. Opstad

Methodology Biopsies were collected (and snap-frozen in liquid nitrogen) during routine surgical resection of 41 adult human gliomas; and histology revealed 8 astrocytoma grade II (AS2), 3 astrocytoma grade III (AS3) and 30 glioblastoma multiforme (GBM) tumors. 10–15 mg of biopsy tissue were used for each HRMAS analysis, with presaturation measurements performed on a 600 MHz Bruker Avance spectrometer and biopsy samples maintained at 4◦ C throughout to reduce enzymatic metabolic changes within the biopsy tissues (Opstad et al., 2008b). Each brain tumor biopsy presaturation spectrum was analyzed using LCModelTM (Version 6.1-4), with a presaturation basis set containing 23 in vitro metabolite spectra and 18 simulated lipid/macromolecule (Lip/MM) peaks (Opstad et al., 2008a, 2008b, 2010). Metabolite and Lip/MM proton concentrations were determined using the biopsy water signal as a reference, assuming a water concentration of 44 M (Cheng et al., 1998), with no corrections made for T1 or T2 relaxation. Following HRMAS, each biopsy sample was removed, re-frozen on cardice and cryostat sectioned at 10 μm. Sections were arranged so that each microscope slide had 5 sections representative of the entire length of the biopsy. The TUNEL method (Gavrieli et al., 1992) was then used to stain one slide from each biopsy. Although it has been reported that TUNEL can label necrotic cells as well as apoptotic cells (Gold et al., 1994), the necrotic areas were found to be clearly visible and showed little, if any, TUNEL-positive nuclei. TUNEL-positive nuclei were manually counted, biopsy section areas determined and results expressed as number of TUNEL-positive nuclei per mm2 for each biopsy. Haematoxylin & Eosin (H&E) staining was also performed on an adjacent slide per biopsy for morphological determination (cell density and percentage necrosis). Two of the 27 necrotic biopsies had poor H&E staining which prevented accurate measurement of the cell density and percentage of necrosis and were therefore excluded from measurements involving these factors.

14 Astrocytomas: Role of Taurine in Apoptosis Using Magnetic Resonance Spectroscopy

Results

123

Linear regression analysis (with Bonferroni correction for multiple comparisons) revealed taurine as the only 1 H MRS metabolite to significantly correlate with the TUNEL-stained positive nuclei per mm2 in both nonnecrotic (R = 0.727, p = 0.003, n = 14) and necrotic biopsies (R = 0.626, p = 0.0005, n = 27), with a similar slope of correlation between taurine and the density of apoptotic cells for both groups and both intercepting very close to the origin (Fig. 14.1a). In the non-necrotic biopsies significant positive correlations were also found between the number of apoptotic cells and the ca. 2.8 (Fig. 14.1b) and 1.3 ppm Lip/MM (data not shown). However, non-significant negative correlations were found for the necrotic biopsy group. The correlation of apoptotic cell density with the ca. 2.8 ppm Lip/MM also intercepted close to the origin (Fig. 14.1b), but not for the correlation of apoptosis with the ca. 1.3 ppm Lip/MM.

Figure 14.2 shows example HRMAS spectra of a non-necrotic (top spectrum) and necrotic (middle spectrum) biopsy, highlighting the taurine and ca. 2.8 ppm contributions. The taurine concentrations in both spectra were the same (1.041 and 1.042 mM respectively), but a greater concentration of the ca. 2.8 ppm Lip/MM in the necrotic spectrum (middle spectrum) is clearly shown. The percentage of necrosis in this biopsy was 8.8%. The apoptotic cell density was found to be independent of total cell density in the non-necrotic biopsies and although an apparent correlation of apoptotic cell density with total cell density was found in the necrotic biopsy group, this was an artefact of the effect of necrotic dilution on both the apoptotic and total cell counts per mm2 (Fig. 14.3). Similar results were found with the correlation between cell density and taurine concentrations in both biopsy groups. Although cell density did have a strong negative correlation with the percentage of necrosis in the necrotic biopsy group (R = –0.917, p < 0.0001, n = 25), no significant differences were found between the biopsy groups (non-necrotic and necrotic) in cell density, number of TUNEL-positive nuclei per mm2 or taurine concentrations. In contrast, the ca. 2.8 and 1.3 ppm Lip/MM proton concentrations were both significantly different between the two biopsy groups (p < 0.0001 for both, Mann Whitney U-test). Principal component analysis of the non-necrotic astrocytoma biopsy HRMAS metabolite concentration data has also revealed a potential metabolic signature

Fig. 14.1 Plots and linear regression analyses (±95% confidence limits) of TUNEL-positive nuclei per mm2 against (a) taurine concentration (mM) for non-necrotic (closed symbols and solid lines, R = 0.727, p = 0.003, n = 14) and necrotic astrocytoma biopsies (open symbols and dashed lines, R = 0.626,

p = 0.0005, n = 27); and (b) ca. 2.8 ppm Lip/MM proton concentration for non-necrotic (closed symbols and solid lines, R = 0.705, p = 0.005, n = 14) and necrotic astrocytoma biopsies (open symbols and dashed line showing the insignificant negative correlation, n = 27)

Histological analysis of the H&E stained sections indicated necrosis present in 27/41 biopsies and for subsequent data analysis the biopsies were divided into two groups: (i) non-necrotic; and (ii) necrotic (according to a measurable percentage of necrosis).

124

K.S. Opstad

TUNEL-positive nuclei per mm2 was R = 0.741, p = 0.002, n = 14.

Discussion The major findings of this study were:

Fig. 14.2 HRMAS 1 H MRS from a non-necrotic (top) and necrotic (middle) biopsy showing the taurine and ca. 2.8 ppm Lip/MM contributions. The bottom spectrum is an in vitro taurine HRMAS 1 H MR spectrum showing the full taurine spectral pattern. Note that the biopsy spectra are slightly shifted to the right for full view of the spectral peaks (Taurine – Tau, Creatine – Cr)

(i) the taurine concentration in glioma biopsies correlated with the number of TUNEL-stained apoptotic cells (Fig. 14.1) independently of the presence of necrosis; and (ii) the ca 2.8 ppm Lip/MM peak from PUFAs correlated with the number of TUNEL-stained apoptotic cells, but only in the non-necrotic biopsies. Thus we have two potential biomarkers for the in vivo determination of apoptosis in gliomas.

Taurine and Apoptosis

Fig. 14.3 Plots of TUNEL-positive nuclei per mm2 against cell density per mm2 for non-necrotic (closed symbols, n = 14), necrotic astrocytoma biopsies with a percentage necrosis < 25% (large open symbols, n = 19) and necrotic astrocytoma biopsies with a percentage necrosis > 50% (small open symbols, n = 6)

relating to apoptosis. The principal component loadings greater than 0.5 and in order of decreasing loading in principal component 1 (PC1) were: ca. 0.9 ppm Lip/MM, ca. 1.3 ppm Lip/MM, phosphocholine, glutathione, taurine, glutamate and ca. 2.8 ppm Lip/MM; and the correlation between PC1 and number of

Taurine is one of the most abundant amino acids found in human brain and has been shown to have several physiological functions such as a neuromodulator, a neurotransmitter (Albrecht and Schousboe, 2005) and an osmolyte (Pasantes-Morales et al., 2002). A review from 1992 into the release of taurine from ex vivo brain slices had suggested that a cellular efflux of taurine appeared to be mediated by membrane carriers in an outward direction and although taurine release, in response to K+ , was generally slow, it was much greater in the developing brain than mature brain, concluding that the release of taurine from the ex vivo brain slices could not solely be as a result of intracellular swelling alone (Saransaari and Oja, 1992). In mouse brain, apoptosis has been shown to be at its greatest during brain development, but then decreases with age; and similarly caspase-3 (a key mediator of apoptosis) also decreases between the developing, juvenile and mature brain stages (Zhu et al., 2005). Organic osmolytes, such as taurine, are important in cell volume regulation and cell swelling-induced taurine efflux is exhibited by most cells (Shennan, 2008). However, cell shrinkage is a distinctive characteristic of apoptotic cells, with changes in ion channel fluxes thought to play a major part (Gómez-Angelats and Cidlowski, 2002). While cell swelling involves

14 Astrocytomas: Role of Taurine in Apoptosis Using Magnetic Resonance Spectroscopy

the efflux of K+ and Cl– , cell shrinkage has been shown to activate systems such as a (Na+ - K+ - Cl– ) co-transporter (efflux) and Na+ (efflux)/H+ (influx) exchange, in parallel with a Cl– (efflux)/HCO− 3 (influx) exchange (Shennan, 2008). Cellular extrusion of taurine, resulting in a decrease in ionic strength, has already been implicated as a signalling mechanism for apoptosis in several cell types [cerebellar granule neurons, NIH 3T3 fibroblasts and Jurkat T-lymphocytes (Morán et al., 2000; Friis et al., 2005; Lang et al., 1998)] and evidence from the study of taurine release in apoptotic cerebellar granule neurons has suggested this taurine efflux may be part dependent on a Na+ efflux, whereas an osmosensitive efflux of taurine was Na+ -independent (Morán et al., 2000). However, the exact mechanism of taurine efflux during apoptosis is far from clear. One study has implicated p38 mitogenactivated protein kinase activation prior to an efflux of K+ and taurine from the cell, resulting in stimulation of the cysteine protease caspase-3 in NIH 3T3 fibroblasts (Friis et al., 2005). Whereas another, using Jurkat T-lymphocytes, has implicated CD95-receptor triggering leads to a caspase-dependent stimulation of cellular taurine release facilitating apoptotic DNA fragmentation (labelled by the TUNEL reaction) and cell shrinkage (Lang et al., 1998). Whether taurine accumulation occurs only in relation to apoptotic cells or if apoptosis and increased taurine content derive from a common cause such as hypoxia, resulting in a population of cells with increased taurine of which some are in the apoptotic stage, is yet to be elucidated. Further work is now required to understand the processes that give rise to the correlation between apoptosis and taurine in gliomas as shown in Fig. 14.1a. Necrosis is one of the main histopathological indicators of malignancy in astrocytomas and the presence of significant amounts of necrosis (or large variations in cell density) could produce false correlations between any two parameters whose value is dependent on the number of viable cells. However, although the number of TUNEL-positive nuclei per mm2 did correlate with the total biopsy cell density, when all the necrotic biopsies were included in the analysis, 6 biopsies had a percentage necrosis of greater than 50% and hence very low TUNEL-staining. By treating these samples as outliers, and comparing only the samples with <25% necrosis, there was no correlation between the number of TUNEL-positive cells per biopsy sample and cell density (Fig. 14.3). Thus

125

the correlation between taurine and TUNEL-staining was a true correlation independent of cell density or necrosis, when the highly-necrotic samples were excluded.

Ca. 2.8 ppm Polyunsaturated Fatty Acid and Apoptosis The ca. 2.8 ppm PUFA 1 H MRS lipid peak has been previously assigned as a biomarker of apoptotic response in a study of anticancer treatment by gene-therapy in pre-clinical rodent gliomas (Hakumäki et al., 1999; Griffin et al., 2003) and also proposed as an in vivo biomarker for apoptosis in the clinic (Hakumäki et al., 1999; Schmitz et al., 2005). Although there was a good correlation between the ca. 2.8 ppm (-CH=CHCH2 CH=CH-) PUFA proton concentrations and the number of TUNEL-stained apoptotic nuclei in human glial brain tumor biopsies (Fig. 14.1b), this was only in the non-necrotic biopsies, which also showed an intercept close to the origin, suggesting that the PUFA peak at ca. 2.8 ppm originates during the process of apoptosis prior to necrosis (Hakumäki et al., 1999; Griffin et al., 2003). Correlation between apoptosis and the ca. 1.3 ppm (-CH2 -) Lip/MM resonance in the non-necrotic biopsies showed a non-zero intercept indicating the presence of additional ca. 1.3 ppm signals that do not relate to apoptosis. It has been suggested that apoptosis produces lipid bodies that remain after cell death (Griffin et al., 2003), similar to those observed with necrosis in rodent gliomas (Rémy et al., 1997; Lahrech et al., 2001; Zoula et al., 2003). However, in the necrotic biopsies, there was a correlation between the percentage of necrosis and the ca. 2.8 ppm proton concentration (R = 0.663, p = 0.0003, n = 25), but no correlation was found between the ca. 2.8 ppm peak and the number of TUNEL-positive nuclei. This suggests that the ca. 2.8 ppm resonance in the necrotic biopsies arises primarily from hypoxic/necrotic lipid bodies rather than apoptotic lipid bodies. Overall the data are in agreement with previous findings in animal studies that the PUFA peak at ca. 2.8 ppm originates during the process of apoptosis, but the results suggest this correlation is only true for low-grade astrocytomas, which are non-necrotic.

126

Effects of Tissue Ischemia on Biopsy Metabolite Concentrations During surgical removal of brain tumor tissue there is an uncontrolled period of tissue ischemia prior to liquid nitrogen freezing that allows biochemical changes to occur in the biopsy tissue. Taurine concentrations have been shown to remain stable for up to 8 h in postmortem human brain (Perry et al., 1981) and our laboratory has also recently shown taurine to remain stable in human brain tumor biopsies (AS2 and GBM) and normal rat brain during prolonged HRMAS spinning (Opstad et al., 2008b). Thus significant post-ischemic changes in tumor biopsy taurine concentrations as a result of the biopsy excision and/or the HRMAS procedure appear unlikely. No changes in the ca. 0.9, 1.3 and 2.8 ppm Lip/MM concentrations have been observed after prolonged HRMAS spinning (4 h) in any of our experiments (unpublished data).

Principal Component Analysis – An Apoptotic Signature? A recent pattern recognition HRMAS study of cervical carcinoma biopsies (all except one being nonnecrotic) has shown that all the lipid peaks in the presaturation spectra contributed to the principal components that correlated with the apoptotic cell density, with a major contribution from the ratio of the ca. 1.3 ppm to ca. 0.9 ppm Lip/MM peaks. Taurine was found to be unrelated to apoptosis in these cervical carcinoma biopsies, but instead related to the tumor cell density and tumor fraction when combined with creatine, choline-containing metabolites, glucose and lactate (Lyng et al., 2007). The human glioma data showed no correlation between apoptosis (either the number of TUNEL-positive nuclei per mm2 or the apoptotic index, calculated as the percentage of TUNEL-positive nuclei per cell density) and the ca. 1.3 ppm to ca. 0.9 ppm Lip/MM ratios. Pattern recognition is an objective technique often used to identify meaningful patterns from 1 H MRS data and preliminary analysis of the presented non-necrotic astrocytoma biopsy HRMAS metabolite concentration data using principal component analysis has also revealed a potential metabolic signature relating to apoptosis. The major metabolites contributing to this signature

K.S. Opstad

were ca. 0.9 ppm Lip/MM, ca. 1.3 ppm Lip/MM, phosphocholine, glutathione, taurine, glutamate and ca. 2.8 ppm Lip/MM, all of which have previously been implicated in apoptosis (Franco et al., 2007; Hakumäki et al., 1999; Morán et al., 2000; Friis et al., 2005; Lang et al., 1998). However, as yet, no PC factors have been related to apoptosis in the necrotic biopsy group. From this limited comparison of tumor data, it appears that biomarkers associated with apoptosis may be cancer type dependent. Taurine is difficult to measure accurately by direct methods in in vivo 1 H MRS, due to signal overlap from the stronger signals of myo-inositol and cholines at ≈3.2 ppm and glucose at ≈3.4 ppm (Govindaraju et al., 2000). However, taurine edited MRS, such as previously applied to in vivo 1 H MRS of normal rat brain (Lei and Peeling, 1999), may provide a more robust measurement and allow the use of taurine as an in vivo biomarker of apoptosis in astrocytomas (both pre- and post-therapeutic response) in the clinic. In conclusion, this study has shown that the taurine concentration in astrocytomas correlates with the number of TUNEL-stained apoptotic nuclei independently of the presence of necrosis. The data also suggest that taurine may be a better biomarker of apoptosis in glial tumors than the ca. 2.8 ppm PUFA peak, for which a correlation with apoptosis is only found in non-necrotic biopsies. The measurement of taurine in gliomas in vivo by non-invasive MRS could be a useful technique for monitoring apoptosis in the clinic.

Declaration This work was first published in British Journal of Cancer as: Opstad KS, Bell BA, Griffiths JR, Howe FA (2009) Taurine: a potential marker of apoptosis in gliomas. Br J Cancer 100:789–794. doi:10.1038/sj.bjc.6604933 www.bjcancer.com

References Albrecht J, Schousboe A (2005) Taurine interaction with neurotransmitter receptors in the CNS: an update. Neurochem Res 30:1615–1621 Al-Saffar NM, Titley JC, Robertson D, Clarke PA, Jackson LE, Leach MO, Ronen SM (2002) Apoptosis is associated with triacylglycerol accumulation in Jurkat T-cells. Br J Cancer 86:963–997

14 Astrocytomas: Role of Taurine in Apoptosis Using Magnetic Resonance Spectroscopy Cheng LL, Chang IW, Louis DN, Gonzalez RG (1998) Correlation of high-resolution magic angle spinning proton magnetic resonance spectroscopy with histopathology of intact human brain tumor specimens. Cancer Res 58: 1825–1832 Delikatny EJ, Cooper WA, Brammah S, Sathasivam N, Rideout DC (2002) Nuclear magnetic resonance-visible lipids induced by cationic lipophilic chemotherapeutic agents are accompanied by increased lipid droplet formation and damaged mitochondria. Cancer Res 62:1394–1400 Di Vito M, Lenti L, Knijn A, Iorio E, D’Agostino F, Molinari A, Calcabrini A, Stringaro A, Meschini S, Arancia G, Bozzi A, Strom R, Podo F (2001) 1 H NMR-visible mobile lipid domains correlate with cytoplasmic lipid bodies in apoptotic T-lymphoblastoid cells. Biochim Biophys Acta 1530:47–66 Franco R, Panayoitidis MI, Cidlowski JA (2007) Glutathione depletion is necessary for apoptosis in lymphoid cells independent of reactive oxygen species formation. J Biol Chem 282:30452–30465 Friis MB, Friborg CR, Schneider L, Nielsen MB, Lambert IH, Christensen ST, Hoffmann EK (2005) Cell shrinkage as a signal to apoptosis in NIH 3T3 fibroblasts. J Physiol 567:427–443 Gavrieli Y, Sherman Y, Ben-Sasson SA (1992) Identification of programmed cell death in situ via specific labeling of nuclear DNA fragmentation. Cell Biol 119:493–501 Gold R, Schmied M, Giegerich G, Breitschopf H, Hartung HP, Toyka KV, Lassmann H (1994) Differentiation between cellular apoptosis and necrosis by the combined use of in situ tailing and nick translation techniques. Lab Invest 71: 219–225 Gómez-Angelats M, Cidlowski JA (2002) Cell volume control and signal transduction in apoptosis. Toxicol Pathol 30: 541–551 Govindaraju V, Young K, Maudsley AA (2000) Proton NMR chemical shifts and coupling constants for brain metabolites. NMR Biomed 13:129–153 Griffin JL, Lehtimäki KK, Valonen PK, Gröhn OH, Kettunen MI, Ylä-Herttuala S, Pitkänen A, Nicholson JK, Kauppinen RA (2003) Assignment of 1 H nuclear magnetic resonance visible polyunsaturated fatty acids in BT4C gliomas undergoing ganciclovir-thymidine kinase gene therapy-induced programmed cell death. Cancer Res 63:3195–3201 Hakumäki JM, Poptani H, Sandmair AM, Ylä-Herttuala S, Kauppinen RA (1999) 1 H MRS detects polyunsaturated fatty acid accumulation during gene therapy of glioma: implications for the in vivo detection of apoptosis. Nat Med 5:1323–1327 Hengartner MO (2000) The biochemistry of apoptosis. Nature 407:770–776 Iorio E, Di Vito M, Spadaro F, Ramoni C, Lococo E, Carnevale R, Lenti L, Strom R, Podo F (2003) Triacsin C inhibits the formation of 1 H NMR-visible mobile lipids and lipid bodies in HuT 78 apoptotic cells. Biochim Biophys Acta 1634:1–14 Lahrech H, Zoula S, Farion R, Rémy C, Décorps M (2001) In vivo measurement of the size of lipid droplets in an intracerebral glioma in the rat. Magn Reson Med 45:409–414 Lang F, Madlung J, Uhlemann AC, Risler T, Gulbins E (1998) Cellular taurine release triggered by stimulation of the Fas(CD95) receptor in Jurkat lymphocytes. Pflugers Arch 436:377–383

127

Lei H, Peeling J (1999) A localized double-quantum filter for in vivo detection of taurine. Magn Reson Med 42:454–460 Lyng H, Sitter B, Bathen TF, Jensen LR, Sundfør K, Kristensen GB, Gribbestad IS (2007) Metabolic mapping by use of highresolution magic angle spinning 1 H MR spectroscopy for assessment of apoptosis in cervical carcinomas. BMC Cancer 7:11 Milkevitch M, Shim H, Pilatus U, Pickup S, Wehrle JP, Samid D, Poptani H, Glickson JD, Delikatny EJ (2005) Increases in NMR-visible lipid and glycerophosphocholine during phenylbutyrate-induced apoptosis in human prostate cancer cells. Biochim Biophys Acta 1734:1–12 Morán J, Hernández-Pech X, Merchant-Larios H, PasantesMorales H (2000) Release of taurine in apoptotic cerebellar granule neurons in culture. Pflugers Arch 439:271–277 Opstad KS, Bell BA, Griffiths JR, Howe FA (2008) An investigation of human brain tumour lipids by high-resolution magic angle spinning 1 H MRS and histological analysis. NMR Biomed 21:677–685 Opstad KS, Bell BA, Griffiths JR, Howe FA (2008) An assessment of the effects of sample ischaemia and spinning time on the metabolic profile of brain tumour biopsy specimens as determined by high-resolution magic angle spinning 1 H NMR. NMR Biomed 21:1138–1147 Opstad KS, Wright AJ, Bell BA, Griffiths JR, Howe FA (2010) Correlations between in vivo 1 H MRS and ex vivo 1 H HRMAS metabolite measurements in adult human gliomas. J Magn Reson Imaging 31:289–297 Pasantes-Morales H, Franco R, Ochoa L, Ordaz B (2002) Osmosensitive release of neurotransmitter amino acids: relevance and mechanisms. Neurochem Res 27:59–65 Perry TL, Hansen S, Gandham SS (1981) Postmortem changes of amino compounds in human and rat brain. J Neurochem 36:406–410 Rao RD, James CD (2004) Altered molecular pathways in gliomas: an overview of clinically relevant issues. Semin Oncol 31:595–604 Rémy C, Fouilhé N, Barba I, Sam-Laï E, Lahrech H, Cucurella MG, Izquierdo M, Moreno A, Ziegler A, Massarelli R, Décorps M, Arús C (1997) Evidence that mobile lipids detected in rat brain glioma by 1 H nuclear magnetic resonance correspond to lipid droplets. Cancer Res 57: 407–414 Saransaari P, Oja SS (1992) Release of GABA and taurine from brain slices. Prog Neurobiol 38:455–482 Schmitz JE, Kettunen MI, Hu DE, Brindle KM (2005) 1 H MRSvisible lipids accumulate during apoptosis of lymphoma cells in vitro and in vivo. Magn Reson Med 54:43–50 Shennan DB (2008) Swelling-induced taurine transport: relationship with chloride channels, anion-exchangers and other swelling-activated transport pathways. Cell Physiol Biochem 21:15–28 Steinbach JP, Weller M (2004) Apoptosis in gliomas: molecular mechanisms and therapeutic implications. J Neurooncol 70:245–254 Valonen PK, Griffin JL, Lehtimäki KK, Liimatainen T, Nicholson JK, Gröhn OH, Kauppinen RA (2005) Highresolution magic-angle-spinning 1 H NMR spectroscopy reveals different responses in choline-containing metabolites upon gene therapy-induced programmed cell death in rat brain glioma. NMR Biomed 18:252–259

128 Wu CL, Taylor JL, He W, Zepeda AG, Halpern EF, Bielecki A, Gonzalez RG, Cheng LL (2003) Proton high-resolution magic angle spinning NMR analysis of fresh and previously frozen tissue of human prostate. Magn Reson Med 50:1307–1311 Zhu C, Wang X, Xu F, Bahr BA, Shibata M, Uchiyama Y, Hagberg H, Blomgren K (2005) The influence of age on

K.S. Opstad apoptotic and other mechanisms of cell death after cerebral hypoxia-ischemia. Cell Death Differ 12:162–176 Zoula S, Hérigault G, Ziegler A, Farion R, Décorps M, Rémy C (2003) Correlation between the occurrence of 1 H-MRS lipid signal, necrosis and lipid droplets during C6 rat glioma development. NMR Biomed 16:199–212

Chapter 15

Imaging of Hypoxia-Inducible Factor-1-Active Regions in Tumors Using a POS and 123 I-IBB Method Masashi Ueda and Hideo Saji

Abstract Hypoxic regions in solid tumors are critically important in tumor physiology and cancer treatment, and they appear to be strongly associated with malignant progression and therapeutic resistance. The transcription factor hypoxia-inducible factor-1 (HIF-1) was recently reported to be a master transcriptional activator of various genes related to malignant phenotypes; therefore, noninvasive imaging of HIF-1active tumors is important for targeted cancer therapy and predicting prognosis. The α-subunit of HIF-1 (HIF-1α) is degraded in an oxygen-dependent manner under normoxic conditions, whereas it is stable under hypoxic conditions and controls HIF-1 transcriptional activity. Thus, a probe that is degraded in a similar manner as HIF-1α can be used to evaluate HIF-1 activity in vivo. A fusion protein named POS was recently developed as such a probe. It consists of 3 domains: protein transduction, oxygen-dependent degradation, and monomeric streptavidin domains. The streptavidin moiety is labeled by ([123 I]iodobenzoyl)norbiotinamide (123 I-IBB), a radioiodinated biotin derivative, to produce 123 I-IPOS. 123 I-IPOS is stable in hypoxic cells and accumulates in tumors. Tumors could be clearly visualized 24 h after 123 I-IBB injection. The tumoral accumulation of 123 I-IPOS was positively correlated with HIF-1 transcriptional activity and coincided with the HIF1-positive areas. Thus, 123 I-IPOS is a potential probe for the imaging of HIF-1-active hypoxic regions in

H. Saji () Department of Patho-Functional Bioanalysis, Graduate School of Pharmaceutical Sciences, Kyoto University, Sakyo-ku, Kyoto 606-8501, Japan e-mail: [email protected]

tumors. However, the drawback of 123 I-IPOS was that a long time (24 h) was required to obtain a well-contrasted image. To overcome this problem, a pretargeting approach was used. In the pretargeting approach, a radionuclide is delivered separately from the tumor-seeking molecule, that is, POS is administered and allowed to undergo degradation in normal regions, after which 123 I-IBB is administered. 123 I-IBB is not retained in normal regions; however, it binds to the POS retained in HIF-1-active regions, which enables specific imaging of such regions. In fact, 123 I-IBB accumulated in tumors pretargeted with POS, whereas it did not accumulate in regions that were not pretargeted. Tumors could be clearly visualized 6 h after injection, which is one-fourth of the time required for the direct targeting approach. The tumoral accumulation of 123 I-IBB in POS-pretargeted mice was positively correlated with HIF-1 transcriptional activity and coincided with the HIF-1-positive areas. Thus, POS pretargeting with 123 I-IBB is a useful technique for the rapid imaging and detection of HIF-1-active regions in tumors. Keywords Hypoxic regions Reoxygenation · Probe · Immunohistochemistry

123 I-IBB · Radionuclides

· ·

Introduction Insufficient blood supply to a rapidly growing tumor leads to the presence of oxygen tension below physiological levels, hypoxia, in solid tumors. Tumor hypoxia is critically important in tumor physiology and cancer treatment, and it appears to be strongly associated

M.A. Hayat (ed.), Tumors of the Central Nervous System, Volume 5, DOI 10.1007/978-94-007-2019-0_15, © Springer Science+Business Media B.V. 2012

129

130

with malignant progression and therapeutic resistance (Semenza, 2010a). The transcription factor hypoxiainducible factor-1 (HIF-1) has been reported to be one of the critical components of hypoxic responses (Eltzschig and Carmeliet, 2011). HIF-1 was recently reported to be closely related to tumor malignancy and to hinder both chemotherapy and radiation therapy (Semenza, 2010b). Thus, the development of techniques to noninvasively detect HIF-1-active hypoxic tumor cells is of great interest. Such a technique may be useful in obtaining a qualitative diagnosis and establishing a therapeutic strategy for cancer.

In Vivo Molecular Imaging Recently, the field of “molecular imaging,” in which physiological functions and disease-related changes are visualized in vivo at the molecular level, has been emphasized. In this field, nuclear medicine imaging, a noninvasive and high-sensitivity procedure, is primarily employed. Nuclear medicine imaging is a procedure to visualize the distribution of radioactivity by administering a radioactive probe into the body and detecting the emitted radiation using positron emission tomography and single-photon emission computed tomography. The images obtained provide not only morphological information but also quantitative information regarding vital tissue activity/functions. These imaging techniques may become more important for clarifying the pathogeneses of diseases by visualizing the functions of genes, enzymes, proteins, and receptors that regulate physiological functions/conditions and intuitively understanding vital functions by visualizing these molecules (Sadeghi et al., 2010; Wu and Kandeel, 2010; Chitneni et al., 2011). This review will focus on the development of molecular imaging techniques to visualize the HIF-1-active hypoxic regions in tumors.

Development of an Oxygen-Dependent Degradable Probe HIF-1 is a heterodimeric protein consisting of an oxygen-sensitive alpha subunit (HIF-1α) and a constitutively expressed beta subunit. Under normoxic conditions, 2 proline residues in the oxygen-dependent

M. Ueda and H. Saji

degradation domain (ODD) of HIF-1α are hydroxylated, leading to the proteasomal degradation of HIF-1α. Under hypoxic conditions, oxygen is the ratelimiting factor for prolyl hydroxylation, resulting in decreased degradation of HIF-1α (Kizaka-Kondoh and Konse-Nagasawa, 2009). That is, the ODD of HIF1α regulates HIF-1 activity. Thus, a probe containing ODD and degrading in a similar manner as HIF-1α can likely be used to evaluate HIF-1 activity in vivo. Based on this concept, some ODD fusion proteins have been developed for tumor diagnosis and therapy (Kizaka-Kondoh et al., 2009a, b; Kuchimaru et al., 2010). Harada et al. (2002) synthesized a fusion protein with 3 domains, TAT-ODD-procaspase-3 (TOP3). TAT is derived from the protein-transduction domain (PTD) of the human immunodeficiency virus type-1 tat protein (Schwarze et al., 1999) and efficiently delivers TOP3 to any tissue in vivo. The ODD contains the VHL-mediated protein destruction motif of human HIF-1α protein and confers hypoxia-dependent stabilization to TOP3 (Harada et al., 2002). Procaspase-3 originates from human caspase-3 protein (FernandesAlnemri et al., 1994) and confers cytocidal activity to TOP3. Intraperitoneal injection of TOP3 into tumorbearing mice results in specific targeting and apoptosis of HIF-1-active tumor cells (Harada et al., 2005). Kudo et al. (2009) utilized the ODD fusion proteins to image HIF-1-active hypoxic tumors. The amino acid sequence called PTD3 that exhibits higher cellpenetrating activity than TAT (Kizaka-Kondoh et al., 2009a) was used as the PTD. The ODD was derived from HIF-1α548–603 . To label the ODD fusion protein with a radioactive reagent, PTD3-ODD was fused to monomeric streptavidin (SAV), resulting in the chimeric protein PTD3-ODD-SAV (POS). POS was then labeled by a radioiodinated biotin derivative, 3[123 I]iodobenzoyl)norbiotinamide (123 I-IBB), to produce 123 I-IBB-POS (123 I-IPOS) (Kudo et al., 2009).

Imaging of HIF-1-Active Tumors Using 123 I-IPOS Figure 15.1a describes the imaging concept for HIF-1active regions in tumors. The PTD enables 123 I-IPOS to be delivered to normoxic and hypoxic tissues. In normoxic tissue, POS is degraded in a similar manner as HIF-1α, and 123 I-IBB is cleared. In contrast, in

15 Imaging of Hypoxia-Inducible Factor-1-Active Regions in Tumors Using a POS and 123 I-IBB Method

131

Fig. 15.1 (a) Concept of hypoxia imaging using 123 I-IPOS. PTD: protein transduction domain, ODD: oxygen-dependent degradation domain (HIF-1α 548–603), SAV: monomeric streptavidin, 123 I-IBB: (3-[123 I]iodobenzoyl)norbiotinamide.

(b) Scintigraphic image of a tumor-bearing mouse 24 h after the injection of 123 I-IPOS. The tumor xenografted in the right thigh was clearly visualized (circle). The arrowhead indicates the liver

HIF-1-active tissues, POS escapes degradation, and radioactivity is retained within the cell. Thus, 123 IIPOS enables the specific imaging of HIF-1-active hypoxic regions. After a 24-h incubation of cells with 125 I-IPOS under normoxic (20% O2 ) or hypoxic (0.1% O2 ) conditions, more than 2-fold higher radioactivity was accumulated in hypoxic cells than in normoxic cells. Furthermore, the radioactivity accumulated in hypoxic cells decreased in a time-dependent manner after reoxygenation. Size-exclusion high-performance liquid chromatography (HPLC) analysis revealed that more than 80% of the intracellular radioactivity was derived from intact 125 I-IPOS and that approximately 70% of the radioactivity in the reoxygenated medium was derived from 125 I-IBB and other small molecules (Kudo et al., 2009). These results indicate oxygendependent degradation of POS followed by clearance of the released 125 I-IBB from the normoxic cells, proving the concept in vitro. Then, in vivo evaluation (biodistribution, planar imaging, metabolite analysis, comparison between probe accumulation and HIF-1 transcriptional activity, autoradiography, and immunohistochemical analysis) of 125 I-IPOS was performed. In the biodistribution study, 125 I-IPOS exhibited tumor accumulation. The tumor-to-blood ratio, which is used as an index of image contrast and calculated as the ratio of radioactivity accumulated in the tumor and blood, increased in a time-dependent manner and reached

5 at 24 h after injection (Kudo et al., 2009). The tumor was clearly visualized 24 h after injection by planar imaging (Fig. 15.1b). Size-exclusion analysis of radioactive compounds in tumors revealed that more than 77% of the intracellular radioactivity was derived from intact 125 I-IPOS. When the tumoral accumulation of 125 I-IPOS was compared with HIF1 transcriptional activity using mice bearing tumors expressing the HIF-1-dependent luciferase reporter gene (Kizaka-Kondoh et al., 2009a) positive and significant correlation was observed (Kudo et al., 2009). The autoradiographic study revealed that the intratumoral distribution of 125 I-IPOS was heterogeneous. Moreover, the 125 I-IPOS-distributed areas coincided with the HIF-1-positive areas detected by immunohistochemistry. These findings indicated that the accumulation of radioiodinated POS reflects the expression of HIF-1; thus, 123 I-IPOS may be a useful imaging probe for HIF-1-active tumors.

Pretargeting Approach The pretargeting approach is a method in which a radionuclide is delivered separately from a tumorseeking molecule (antibody or protein). This method uses a combination of tumor-seeking molecules and the prompt clearance of low-molecular-weight radiolabeled compounds (effector molecules) that are

132

cleared within minutes from the blood. Combining pretargeting with nuclear medical imaging involving the use of a protein is an excellent strategy for the selective delivery of radionuclides to tumor cells. This strategy attempts to exploit the specificity of the protein and avoid the disadvantages associated with its macromolecular size, such as slow blood clearance. Several pretargeting methods have been developed that differ in how they selectively capture the radionuclide, such as streptavidin-biotin (Sano et al., 2010), antibodyhapten (Sharkey et al., 2005), and sense-antisense morpholino oligomer (He et al., 2010). One of the advantages of the pretargeting approach is that it provides a high tumor-to-normal tissue ratio within a short time after injection. In addition, because the effector molecules used in this method are rapidly cleared from the body, radiation exposure is reduced. Moreover, some recent studies have demonstrated that the tumor uptake of the effector molecules used in the pretargeting approach was identical to or even higher than that of directly radiolabeled antibody. The images and therapeutic effects reported by these studies were significantly improved (Goldenberg et al., 2008; Sharkey et al., 2008).

Imaging of HIF-1-Active Tumors Based on the Pretargeting Approach 123 I-IPOS

is a potential probe for imaging HIF-1active tumors. However, because of its large molecular size (34 kDa), 123 I-IPOS is slowly cleared from the blood; therefore, a high tumor-to-normal tissue ratio cannot be obtained quickly after probe injection. To overcome this problem, Ueda et al., proposed a pretargeting method based on the high-affinity interaction between streptavidin and biotin. The principle of the pretargeted imaging of HIF-1-active tumor regions is outlined in Fig. 15.2a. First, POS was administered and allowed to undergo degradation in normal regions; subsequently, 123 I-IBB was administered. Because POS has been degraded and expelled, 123 I-IBB is not retained in normal regions. However, POS is retained in HIF-1-active regions, and 123 I-IBB binds to the POS in these regions; specific imaging of such regions can then be achieved. An examination of the biodistribution of 125 I-IBB alone revealed that the tumor-to-blood ratio was less

M. Ueda and H. Saji

than 1 at all time points. This indicated that 125 I-IBB did not accumulate in tumors (Ueda et al., 2010). In contrast, in the pretargeted group, the tumor-to-blood ratio was greater than 1 as early as 1 h after the injection of 125 I-IBB, and the tumor-to-blood ratio increased in a time-dependent manner. The blood clearance in both groups was comparable. In the pretargeted group, the tumor accumulation of 125 I-IBB 6 h after its injection was more than 30-fold higher than that in the case of 125 I-IBB alone (Ueda et al., 2010). Moreover, the tumoral uptake of 125 I-IBB in the POS-pretargeted group was significantly reduced by treatment with D-biotin or non-radioactive IBB. Size-exclusion analysis of radioactive compounds in tumors revealed that a major proportion of the radioactivity was attributable to macromolecules. Collectively, these findings indicate that radioactivity in the tumor is caused by the binding of 125 I-IBB to the SAV moiety of POS in vivo. Figure 15.2b illustrates the accumulation of 123 IIBB in a POS-pretargeted tumor. The tumor was clearly visualized 6 h after 123 I-IBB injection, and the image was comparable to that obtained 24 h after 123 I-IPOS injection (Fig. 15.1b). Using the pretargeting approach, the tumoral accumulation (1.6% injected dose per gram tissue [ID/g]) and the tumor-to-blood ratio (4.2) 6 h after injection (Ueda et al., 2010) were comparable to the findings 24 h after 125 I-IPOS injection (1.4% ID/g, 5.1) (Kudo et al., 2009). The images of the pretargeted mice clearly indicated that 123 I-IBB accumulated in the tumor more rapidly and cleared much more promptly from the body than 123 I-IPOS. These results indicate that the pretargeting approach shortens the time required to obtain an adequate image by 75%. A highly significant correlation was observed between 125 I-IBB accumulation and HIF-1 transcriptional activity in the same tumor in POS-pretargeted mice (Ueda et al., 2010), when the tumoral accumulation of 125 I-IBB was compared with HIF-1 transcriptional activity using mice bearing tumors expressing the HIF-1-dependent luciferase reporter gene (Kizaka-Kondoh et al., 2009a). The autoradiographic study revealed that the intratumoral distribution of 125 I-IBB was heterogeneous and significantly correlated with HIF-1α-positive regions in the POS-pretargeted tumor. These findings indicate that the accumulation of 125 I-IBB in the POS-pretargeted tumor reflects the expression of HIF-1.

15 Imaging of Hypoxia-Inducible Factor-1-Active Regions in Tumors Using a POS and 123 I-IBB Method

133

Fig. 15.2 (a) Concept of hypoxia imaging using the pretargeting approach. PTD: protein transduction domain, ODD: oxygen-dependent degradation domain (HIF1α 548–603), SAV: monomeric streptavidin, 123 I-IBB:

(3-[123 I]iodobenzoyl)norbiotinamide. (b) Scintigraphic image of a POS-pretargeted mouse at 6 h after injection of 123 I-IBB. The tumor xenografted in the right thigh was clearly visualized (circle). The arrowhead indicates the liver and intestine

In conclusion, using the pretargeting approach, clear tumor images were obtained in a shorter time than was possible with the direct labeling method. Intratumoral accumulations of 125 I-IBB in the POS-pretargeted tumors were significantly correlated with HIF-1α positivity. These findings demonstrate that the pretargeting method with POS and 123 I-IBB is effective for the rapid imaging of HIF-1-active hypoxic tumors.

testing of a 99mTc-labeled bivalent MORF. Mol Pharm 7:1118–1124. Kizaka-Kondoh S, Itasaka S, Zeng L, Tanaka S, Zhao T, Takahashi Y, Shibuya K, Hirota K, Semenza GL, Hiraoka M (2009a) Selective killing of hypoxia-inducible factor-1-active cells improves survival in a mouse model of invasive and metastatic pancreatic cancer. Clin Cancer Res 15:3433–3441. Kizaka-Kondoh, S., and Konse-Nagasawa, H (2009) Significance of nitroimidazole compounds and hypoxiainducible factor-1 for imaging tumor hypoxia. Cancer Sci. 100: 1366–1373. Kizaka-Kondoh S, Tanaka S, Harada H, Hiraoka M (2009b) The HIF-1-active microenvironment: an environmental target for cancer therapy. Adv Drug Deliv Rev 61:623–632. Kuchimaru T, Kadonosono T, Tanaka S, Ushiki T, Hiraoka M, Kizaka-Kondoh S (2010) In vivo imaging of HIF-active tumors by an oxygen-dependent degradation protein probe with an interchangeable labeling system. PloS One 5:e15736. Kudo T, Ueda M, Kuge Y, Mukai T, Tanaka S, Masutani M, Kiyono Y, Kizaka-Kondoh S, Hiraoka M, Saji H (2009) Imaging of HIF-1-active tumor hypoxia using a protein effectively delivered to and specifically stabilized in HIF-1-active tumor cells. J Nucl Med 50:942–949. Sadeghi MM, Glover DK, Lanza GM, Fayad ZA, Johnson LL (2010) Imaging atherosclerosis and vulnerable plaque. J Nucl Med 51 Suppl 1:51S–65S. Sano K, Temma T, Kuge Y, Kudo T, Kamihashi J, Zhao S, Saji H (2010) Radioimmunodetection of membrane type-1 matrix metalloproteinase relevant to tumor malignancy with a pretargeting method. Biol Pharm Bull 33:1589–1595. Schwarze SR, Ho A, Vocero-Akbani A, Dowdy SF (1999). In vivo protein transduction: delivery of a biologically active protein into the mouse. Science 285:1569–1572. Semenza GL (2010a) Defining the role of hypoxia-inducible factor 1 in cancer biology and therapeutics. Oncogene 29: 625–634.

References Chitneni SK, Palmer GM, Zalutsky MR, Dewhirst MW (2011) Molecular imaging of hypoxia. J Nucl Med 52:165–168. Eltzschig HK, Carmeliet P (2011) Hypoxia and inflammation. N Engl J Med 364:656–665. Fernandes-Alnemri T, Litwack G, Alnemri ES (1994) CPP32, a novel human apoptotic protein with homology to Caenorhabditis elegans cell death protein Ced-3 and mammalian interleukin-1 beta-converting enzyme. J Biol Chem 269:30761–30764. Goldenberg DM, Rossi EA, Sharkey RM, McBride WJ, Chang CH (2008) Multifunctional antibodies by the Dock-andLock method for improved cancer imaging and therapy by pretargeting. J Nucl Med 49: 158–163. Harada H, Hiraoka M, Kizaka-Kondoh S (2002) Antitumor effect of TAT-oxygen-dependent degradation-caspase-3 fusion protein specifically stabilized and activated in hypoxic tumor cells. Cancer Res 62:2013–2018. Harada H, Kizaka-Kondoh S, Hiraoka M (2005) Optical imaging of tumor hypoxia and evaluation of efficacy of a hypoxiatargeting drug in living animals. Mol Imaging 4:182–193. He J, Wang Y, Dou S, Liu X, Zhang S, Liu G, Hnatowich D (2010) Affinity enhancement pretargeting: synthesis and

134 Semenza GL (2010b) HIF-1: upstream and downstream of cancer metabolism. Curr Opin Genet Dev 20:51–56. Sharkey RM, Cardillo TM, Rossi EA, Chang CH, Karacay H, McBride WJ, Hansen HJ, Horak ID, Goldenberg DM (2005) Signal amplification in molecular imaging by pretargeting a multivalent, bispecific antibody. Nat Med 11: 1250–1255. Sharkey RM, Karacay H, Litwin S, Rossi EA, McBride, WJ, Chang CH, Goldenberg DM (2008). Improved therapeutic results by pretargeted radioimmunotherapy of non-Hodgkin’s lymphoma with a new recombinant, trivalent,

M. Ueda and H. Saji anti-CD20, bispecific antibody. Cancer Res 68:5282– 5290. Ueda M, Kudo T, Kuge Y, Mukai T, Tanaka S, Konishi H, Miyano A, Ono M, Kizaka-Kondoh S, Hiraoka M, Saji H (2010) Rapid detection of hypoxia-inducible factor-1-active tumours: pretargeted imaging with a protein degrading in a mechanism similar to hypoxia-inducible factor-1alpha. Eur J Nucl Med Mol Imaging 37:1566–1574. Wu Z, Kandeel F (2010) Radionuclide probes for molecular imaging of pancreatic beta-cells. Adv Drug Deliv Rev 62:1125–1138.

Chapter 16

Diffuse Low-Grade Astrocytomas: P53-Mediated Inhibition of Angiogenesis Timo Gaiser and Markus D. Siegelin

Abstract Angiogenesis is playing a crucial role in the growth and progression of astrocytomas. P53 a tumor suppressor gene located on chromosome 17p has been implicated to the regulation of cell death, particularly apoptosis, proliferation and also angiogenesis. In this in vitro study we evaluated the association of p53 gene status (wild-type or mutated) with micro vascular density in a set of astrocytomas. Immunohistochemistry for CD31, a surface marker expressed on endothelial cells, was performed on 23 diffuse astrocytomas (WHO Grade II). Mutation status of the p53 gene was identified by PCR amplification with consequent sequencing of genomic DNA extracted from each tumor tissue. Intratumoural or peritumoural microvascular hot spots were assessed and images taken at a 200× fold magnification. Microvessel count was performed with a modern automatic image analyses algorithm by using these images. P53 mutation occurred in 11 out of 23 (47%) astrocytomas. In p53 mutated gliomas the micro-vascular density and the absolute vessel number was significantly higher compared to p53 wild-type gliomas, thereby supporting the hypothesis of a p53-mediated regulation on angiogenesis in diffuse low-grade astrocytomas.

M.D. Siegelin () Department of Pathology & Cell Biology, Columbia University College of Physicians & Surgeons, 630 W. 168th Street, New York, NY 10032, USA e-mail: [email protected] T. Gaiser () University of Massachusetts, Amherst, MA, LRB 460 E, USA Pathology Mannheim, University Medical Center Mannheim, Theodor-Kutzer Ufer 1-3, 68167 Mannheim, Germany e-mail: [email protected]

To analyze a possible molecular mechanisms between these two factors, LN229, a glioma cell line, harbouring a p53 mutation, was transfected with p53 wild-type and empty vector, as a negative control. A protein array analysis provided evidence that Thrombospondin-1, Coagulation factor III, Serpin E1 and MMP-9 are potential p53 targets and important key players in regulating angiogenesis in gliomas. Our results support the hypotheses that p53 regulates angiogenesis in low grade astrocytomas. Keywords Astrocytomas · Angiogenesis · p53 Mutation · Microvessel density · CD31 · Protein array

Introduction Angiogenesis is thought to play a crucial role in the growth and progression of primary brain tumors and is therefore subject of current preclinical and clinical research studies to understand neovascularization in gliomas. During progress of gliomas from lower to higher stages remarkable biological changes occur, e.g. increased mitotic figures, marked cellular pleomorphism, necrosis (geographic and pseudopalisading necrosis) and microvascular, glomeruloid-like proliferation. One of the major hallmarks and the important diagnostic criterium discriminating anaplastic astrocytoma (WHO Grade III) from glioblastoma (WHO Grade IV) is microvascular hyperplasia (Louis et al., 2007).While the vascular structure of low-grade astrocytoma resembles that of normal brain vessels, glioblastoma shows florid and glomeruloid microvascular proliferation with the contribution of pericytes and vascular smooth muscle cells (Takeuchi et al.,

M.A. Hayat (ed.), Tumors of the Central Nervous System, Volume 5, DOI 10.1007/978-94-007-2019-0_16, © Springer Science+Business Media B.V. 2012

135

136

2009; Wesseling et al., 1995). Therefore an “angiogenic switch” is supposed to be necessary for glioblastoma (Hanahan and Weinberg, 2007). Moreover, all types of astrocytomas rely on solid blood vessels for survival and growth (Wesseling et al., 1997). In order to grow beyond 1–2 mm3 gliomas need a vascular system for nutrient delivery (Folkman, 1990). Additionally, angiogenesis was found to be correlated with aggressiveness and clinical recurrence within each astrocytoma group (Leon et al., 1996). Since the prognosis of patients with astrocytoma is still bad and no curative treatment is available, especially for patients with glioblastoma (15% 5-year survival rate), it is of high priority to investigate the mechanisms for angiogenesis in these tumors (Krex et al., 2007). In this regard, understanding the mechanisms of how glioblastomas and low-grade gliomas drive tumor progression by modulation of angiogenesis might lead to promising novel therapeutic treatments to combat this form of deadly brain cancers. P53 a tumor suppressor gene located on chromosome 17p and one of the most investigated molecules during the last 30 years, has been implicated to angiogenesis (Dameron et al., 1994a, b). Other functions verified for p53 are cell cycle arrest, apoptosis, DNA repair, inflammation and cellular differentiation (Boehme and Blattner, 2009; Haupt et al., 2003; Tanaka et al., 2000; Yamanishi et al., 2002; Lutzker and Levine, 1996). The most dramatic demonstrations of the potency of p53 as a tumor suppressor derived from animal studies, demonstrating that p53 knock-out mice developed a wide spectrum of different tumors at an early age (Donehower et al., 1992). This is also underlined by the fact that p53 is mutated in 50% of all human tumors (Greenblatt et al., 1994; Hainaut and Hollstein, 2000; Hollstein et al., 1991). With respect to low-grade gliomas, these tumors commonly harbor TP53 mutations and as recently shown, IDH-1 mutations. Over a time course of 8 to 10 years these tumors degenerate into a glioblastoma (secondary) with a rate of virtually 100%. The aim of this study was to address microvessel density (MVD) in p53 mutated and p53 wildtype astrocytomas. After observing MVD differences in these two groups the molecular background mechanism was further investigated. In vitro studies were performed with LN229, a glioblastoma cell line mutated for p53, because of the lack of

T. Gaiser and M.D. Siegelin

low-grade astrocytoma cell lines. In summary, important molecular differences in angiogenesis-related proteins after transfection LN229 with wild-type p53 were observed.

Immunohistochemical Examination of Microvessel Density and Human Angiogenesis Array (Methodology) Immunohistochemistry Fixed astrocytomas tumors were embedded in paraffin. Four-micrometer-thick sections were immunostained on the Benchmark XT automated stainer (Ventana Medical Systems, Tucson, Arizona, USA). Tumor samples were stained with an antibody against CD31 (1:10, clone M0823, DAKO, Hamburg, Germany) located on endothelial cells to identify blood vessels or against p53 (1:50, clone NCL-p53 DO1, Novocastra, Dossenheim, Germany) located in the tumor nuclei. Antigen–antibody complexes were detected with Ventana’s Enhanced V-Red Detection, which is an alkaline phosphatase that uses naphthol and fast red chromogen. Most approaches to detect MVD, such as the one used in this work, are based on the protocol of Weidner et al. (1991, 1992). Briefly, after immunohistochemical staining of endothelium or other vascular wall components, areas of highest vessel density were selected and counted either manual or computer-assisted. Consequently, statistical analysis was performed to see if MVD correlates with overall survival, disease-free survival or other measures. We evaluated MVD in astrocytoma with a modern automatic image analyses algorithm provided by S.CO (S.CO LifeScience Company, Garching, Germany). In detail: areas of highest neovascularization were selected by scanning CD31 stained tumor sections at low power (25× and 40×) and identifying the areas of infiltrating tumor with the highest number of microvessel. Vessels belonging the dura mater or the arachnoida were not considered in the vessel counts. However these vessels served as internal controls in assessing the quality of the CD31 staining. After the area of highest neovascularization was identified a picture

16 Diffuse Low-Grade Astrocytomas: P53-Mediated Inhibition of Angiogenesis

with 200× magnification was collected and sent to the analyses program. The degree of vascularization was quantified automatically and objectively with the S.CO Image Analysis System (S.CO LifeScience Company, Garching, Germany), which bases on the Cognition R by Definiens AG. In detail: Network Technology a threshold based approach identifies all red stained objects in the picture and automatically detects a vascular candidate using the criteria colour. Then an object oriented method is applied to eliminate those objects that show only a diffuse and faint red staining compared to the surrounding. The remaining objects are classified into three categories using the criteria object size: small objects = single immunopositive cells (yellow), medium size objects = capillaries (pink), large objects = arteries and veins (red) (Fig. 16.1). The number and area of immunohistochemical positive vessel structures were given as output values separate for each image. The evaluation for p53 followed a much easier approach and was done manually. For p53 immunohistochemistry two different distribution patterns of stained nuclei were found: Pattern A showed a widespread strong positivity of tumor cell nuclei involving either the entire tissue section or segments of the tumor sample. In pattern B, only single tumor cells scattered throughout the tumor sample showed an only weak but positive reaction. Only cases of pattern A were considered positive for immunohistochemistry.

P53 Mutation Analyses Characterization of p53 mutations within the tumor tissues was done by PCR amplification of genomic DNA extracted from the indicated tumor tissues. Genomic DNA was extracted using a genomic DNA Extraction Kit (Qiagen, Hilden, Germany). Primer pairs were generated to span each exon of the p53 gene, as well as flanking intronic sequence, and PCR products R PCR purification were purified using ExoSAP-IT kit. PCR products were subsequently sequenced on an ABI-Prism 3100 Genetic Analyzer and analysed for mutations.

137

Fig. 16.1 Example of an image used to estimate the MVD as a function of p53 mutation status. For both images original magnification was 200×. (a) Area with the highest density of microvessels in a diffuse low-grade astrocytoma (WHO Grade II). Vessels stained with an antibody against CD31 using fast red chromogen. (b) Same image after applying the vessel quantification algorithm. This procedure identifies all red stained objects in the picture and eliminates background. The remaining objects are classified into three categories using the criteria “object size”: small objects = single immunopositive cells (yellow), medium size objects = capillaries (pink), large objects = arteries and veins (red)

Human Angiogenesis Array The Human glioblastoma cell line LN229 (p53 mutant CCT (Pro) → CTT (Leu) mutation at codon 98) was used for evaluating the background mechanisms for the p53 dependent angiogenesis. We transfected LN229 with p53-wild-type vector and Fugene Transfection reagent (Roche Deutschland Holding GmbH, Mannheim, Germany). As a negative control

138

and to substract transfection associated effects, LN229 was also transfected with an empty vector (Clontech, USA; Mountain View, CA). For analyzing the expression profiles of angiogenesis-related proteins we used the Human Angiogenesis Array Kit (R&D Systems, Ltd., Abingdon, United Kingdom). This array consists of 55 on a nitrocellulose membrane spotted antibodies dedicated to proteins related with angiogenesis. Cell samples (1 × 107 cells from LN229 p53 WT- and empty vector-transfected) were harvested and 300 μg of protein was mixed with 15 μl of biotinylated detection antibodies. After pre-treatment, the cocktail was incubated overnight at 4◦ C on a rocking platform. Following a wash step to remove unbound material, streptavidin–horseradish and chemiluminescent detection reagents were added sequentially. The light signal intensity correlates with the amount of converted substrate. All experiments were performed in three triplicates. The data on developed X-ray film were quantified by scanning on a transmission-mode scanner and subsequently analyzed by using ImageJ analysis software (http://rsbweb.nih.gov/ij/).

Statistics Statistical analyses were used to explore the correlation between the presence or absence of p53 mutations and p53 protein expression. Microvessel densities were compared in non-mutated versus mutated tumors and in p53 IHC positive and negative tumors. Therefore, a Two-tailed Students’t-test was used by comparing the MVD between different groups. A value of p < 0.05 was considered significant.

Immunohistochemical Detected P53 Protein Does Not Correlate with Microvessel Density In a series of 23 diffuse astrocytomas (WHO grade II) we found that 9/23 (39.13%) of the astrocytomas were positive for p53 IHC. We found no correlation between a genetically detectable p53 mutations and a nuclear p53 protein accumulation. We found also no differences in MVD

T. Gaiser and M.D. Siegelin

between p53 IHC positive and p53 negative astrocytomas. We detected a mean MVD of 5.30% for the p53 IHC positive cases and a mean MVD of 5% for the p53 IHC negative cases, respectively.

Astrocytomas with P53 Mutation Have a Higher Number of Vessels We identified p53 mutations in 11/23 (47%) of the astrocytomas, in these tumors we could observe a significant higher MVD compared to astrocytomas with a wild-type p53 genotype. MVD for p53 mutated low-grade astrocytomas was 8.30% (±1.5) and 3.6% (±1.2) for p53 wild-type low-grade astrocytomas, respectively. Additionally the absolute vessel number was significantly increased in p53 mutated astrocytomas (60±4) compared to p53 wild-type low-grade astrocytomas (38±5).

Thrombospondin-1, Coagulation Factor III and Serpin E1 are Upregulated Under P53 (Human Angiogenesis Array Data) We transfected LN229 glioma cells that harboured a TP53 mutation with either an empty plasmid (pcCMVempty) or with a plasmid that encodes the p53 wild-type protein (pcCMV-p53 wild-type) to elucidate the molecular mechanisms that potentially lead to a higher MVD in p53 mutated astrocytomas The overexpression of p53 after transfection could be confirmed by western blot. The performed Human Angiogenesis Array Kit detected protein expression differences for the LN229 p53 transfected vs. pcCMVempty in the following proteins: 5.31-fold increase of thrombospondin-1, a 4.17-fold increase of coagulation factor III, a 1.76-fold increase of serpin E1 (PAI-1) and a 0.39-fold decrease of MMP-9.

Discussion Glioma is by far the most frequently examined brain neoplasm in terms of angiogenesis. Already in 1989 by Schiffer et al., an increased vascularity could be linked

16 Diffuse Low-Grade Astrocytomas: P53-Mediated Inhibition of Angiogenesis

to increased glioma growth (Schiffer et al., 1989). Beside that, this work is especially interesting because the authors counted the number of vessels per square mm, the number of nuclei per vessel, and the internal and external diameters, on a conventional histochemical staining (Luxol Fast Blue PAS Hematoxylin) thereby introducing a similar approach as done in our work. Using immunohistochemistry for quantifying vessels in gliomas was then done by Leon et al. (1996). In 93 adult astrocytomas, using FVIII-immunostaining, a negative association between increasing microvessel numbers (counting vessels in a 200× magnification) and disease-free survival by could be demonstrated. Also a qualitative approach by measuring MVD on a scale from 1 to 4 correlated with shortened survival (Leon et al., 1996). In a large retrospective study about low-grade astrocytomas, Abdulrauf et al. (1998) could identify that patients with more than seven microvessels per 400× field had a shorter survival (mean, 3.8 years) for progression to high grade astrocytoma. They concluded that diffuse astrocytomas are therefore very heterogenous and some tumors were more dependent on neovascularization then others. In this context the experiments of Wesseling et al. can be interpretated. They examined a mixed group of astrocytoma (including glioblastomas, astrocytomas and anaplastic astrocytomas) and found a wide range of vessel heterogeneity within each group. They identified areas without increased numbers of microvessel in every subgroup, suggesting that neovascularization may not be a requisite for at least some gliomas (Wesseling et al., 1998). So, to summarize the results it appears that a subgroup of gliomas is more depending on neovascularization than others. This clearly offers the theoretical background for our study because one of the major discriminating factors within astrocytomas is the mutation of the p53 gene. The mutation frequency for p53 in low grade astrocytoma is approx. 60% and mutations frequency does not significantly increase during malignant progression to secondary glioblastomas indicating that this genetic change is an early event (Okamoto et al., 2004; Reifernberger et al., 1996; Watanabe et al., 1997). Some studies found a shorter progression interval from low-grade astrocytomas to glioblastoma in

139

tumors harboring a p53 mutation (Stander et al., 2004; Watanabe et al., 1997). In our study we looked for the association of p53 mutation with MVD. We observed that p53 mutated low-grade astrocytomas exhibited a significant higher vascular density and absolute number of vessels compared to p53 wild-type astrocytomas. This underlines the important role of p53 in terms of regulation of angiogenesis but it also raises the question of the molecular background involved in that process. Beside the numerous functions already demonstrated for p53, recent papers suggest that the normal p53 protein stimulates the expression of genes that prevent the process of angiogenesis (el-Deiry 1998; Teodoro et al., 2007; Vogelstein et al., 2000). Especially in a subsequent, work done by Teodora et al. an activation of the gene encoding α(II) collagen prolyl-4-hydroxylase, which is necessary for the release of the anti-angiogenated collagen-derived peptides, such as endostatin and tumstatin, could be linked to p53 (Teodoro et al., 2006). They used H1299, a non-small cell lung carcinoma cell line; in glioblastoma cell lines however this mechanism could not be proven effective (Berger et al., 2010). At least from our data we claim that low grade astrocytomas, in which the angiogenic switch has not occurred yet, have higher likelihood to recruit new blood vessels if p53 is inactivated by mutation. This step seems to provide a critical growth advantage in tumor development. As a side note, we saw no correlation between p53 protein expressions detected by IHC and MVD. The cause of this is to be found in technical issues. IHC can fail to detect certain p53 mutant protein (Anker et al., 1993; Kupryjanczyk et al., 1993). Although only mutated p53 protein is accumulating in the nucleus of the cell, IHC can still not be used to detect all different types of mutated p53 proteins. The binding ability of the p53 antibody is dependent on the secondary structure of the protein and therefore different for each mutation. Hence, it is not surprising that a correlation between p53 IHC and MVD was not found. To analyze the important molecular background between the increased MVD and a p53 mutation we used an array specially designed for analyzing angiogenesis-related proteins (Human Angiogenesis Array Kit). The investigation was done with a

140

glioblastoma cell line, LN229 (mutated for p53) due to the lack of low-grade astrocytoma cell lines. After digital image analyses of the array we detected four proteins being expressed differentially. Thrombospondin-1 (TSP-1) was increased in p53 transfected LN229 cells (fivefold increase) compared to empty transfected cells. TSP-1 is a 450-kDa extracellular matrix glycoprotein. It has a complex structure and modulates cellular behaviours like motility, adhesion, and proliferation. Furthermore TSP-1 acts as a potential key player in angiogenesis. The increase of TSP-1 in p53 wild-type transfected LN229 glioma cells is compatible with the literature showing that decreased expression of TSP-1 is a key step in creating a pro-angiogenetic environment (Iddings et al., 2007). TSP-1 negatively regulates the growth and migration of endothelial cells both in vitro and in vivo (Hsu et al.,1996; Nor et al., 2000). The loss of p53 led to a deficiency in TSP-1 expression, and the subsequent inability to shut off angiogenesis. Restoration of p53 expression in these tumors re-established TSP-1 expression (Giuriato et al., 2006). Furthermore Kazuno et al. showed that gliomas lacking expression of the gene for thrombospondin-2, a powerful antiangiogenic molecule with many properties similar to TSP-1, are having quantitatively increased microvessel densities (Kazuno et al., 1999). In wild-type p53 transfected LN229 cells coagulation factor III (CF III) was also increased compared to empty transfected LN229 cells (4.17-fold). CF III is expressed by cells which are normally not exposed to flowing blood such as sub-endothelial cells, e.g. smooth muscle cells and cells surrounding blood vessels. This can change when the blood vessel is damaged by for example physical injury or rupture of atherosclerotic plaques. CF III is a cell surface glycoprotein and enables cells to initiate the blood coagulation cascades, and it functions as the high-affinity receptor for the coagulation factor VII. Unlike other cofactors of these protease cascades, which circulate as nonfunctional precursors, this factor is a potent initiator that is fully functional when expressed on cell surfaces. Aside from its role in the coagulation cascade binding of CF III to factor VIIa of the coagulation cascade has also been found to start signaling processes inside the cell. The signaling function of CF III/VIIa plays a role in angiogenesis and apoptosis (Belting et al., 2005).

T. Gaiser and M.D. Siegelin

An association of CF III and p53 has not been published to the best of our knowledge. Plasminogen activator inhibitor type-1 (PAI-1) was 1.76-fold over expressed in wild-type p53 transfected LN229. It has already been reported that p53 also regulate genes responsible for the proteolytic degradation of the extracellular matrix, which is a crucial feature for local invasion and metastasis of neoplastic cells. An important and highly regulated cascade of such proteolytic events involves the plasminogen activator (PA) and inhibitor (PAI) system. So our results are in line with previous reports since it has been demonstrated that wild-type but not mutant p53 specifically binds to and activates the promoter of the PAI-1 gene (Kunz et al., 1995). PAI-1 inhibits the serine proteases tissue-type plasminogen activator and urokinase-type (Kunz et al., 1995). Cellular transformation often results in a dramatic increase in the production of the plasminogen activators (PA) and in altered expression of the inhibitor, PAI-1. The balance between PAs and their inhibitors appears to be critical for the invasive phenotype of tumor cells implying that altered expression slightly in favor of plasminogen activators contributes to the malignant phenotype (Kunz et al.,1995). Matrix metalloproteinase-9 (MMP-9) was decreased by a factor of 0.39 in p53 transfected LN229 glioma cells. Our results are in line with data from the current literature showing that reintroduction of wild-type p53 into mutant p53 soft tissue tumor cells decreased MMP-9 mRNA and protein levels (Liu et al., 2006). In glioblastoma cells antisense MMP-9 revealed marked reduction in the invasiveness of the adenoviral infected cells compared with parental and vector controls (Lakka et al., 2003). In conclusion our data showed a higher MVD in p53 mutated low-grade astrocytomas and support the hypothesis of a p53-mediated regulation on angiogenesis in diffuse low-grade astrocytomas. Additionally the differential angiogenesis protein analyses for p53 transfected cell lines provides evidence that Thrombospondin-1, Coagulation factor III and Serpin E1 are potentially p53 targets and important key players in regulating angiogenesis in gliomas. For future analyses the influence of the EGFR signaling pathway, e.g. PTEN and EGFR alterations, and IDH-1 mutational status on MVD and angiogenesis might be worthwhile to explore since both pathways are closely linked to metabolism and hypoxia/ischemia.

16 Diffuse Low-Grade Astrocytomas: P53-Mediated Inhibition of Angiogenesis

References Abdulrauf SI, Edvardsen K, Ho KL, Yang XY, Rock JP, Rosenblum ML (1998) Vascular endothelial growth factor expression and vascular density as prognostic markers of survival in patients with low-grade astrocytoma. J Neurosurg 88(3):513–520 Anker L, Ohgaki H, Ludeke BI, Herrmann HD, Kleihues P, Westphal M (1993) p53 protein accumulation and gene mutations in human glioma cell lines. Int J Cancer 55(6):982–987 Belting M, Ahamed J, Ruf W (2005) Signaling of the tissue factor coagulation pathway in angiogenesis and cancer. Arterioscler Thromb Vasc Biol 25(8):1545–1550 Berger B, Capper D, Lemke D, Pfenning PN, Platten M, Weller M, von Deimling A, Wick W, Weiler M (2010) Defective p53 antiangiogenic signaling in glioblastoma. Neuro Oncol 12(9):894–907 Boehme KA, Blattner C (2009) Regulation of p53–insights into a complex process. Crit Rev Biochem Mol Biol 44(6): 367–392 Dameron KM, Volpert OV, Tainsky MA, Bouck N (1994a) Control of angiogenesis in fibroblasts by p53 regulation of thrombospondin-1. Science 265(5178):1582–1584 Dameron KM, Volpert OV, Tainsky MA, Bouck N (1994b) The p53 tumor suppressor gene inhibits angiogenesis by stimulating the production of thrombospondin. Cold Spring Harb Symp Quant Biol 59:483–489 Donehower LA, Harvey M, Slagle BL, McArthur MJ, Montgomery CA Jr., Butel JS, Bradley A (1992) Mice deficient for p53 are developmentally normal but susceptible to spontaneous tumours. Nature 356(6366):215–221 el-Deiry, W.S. (1998) Regulation of p53 downstream genes. Semin Cancer Biol 8(5):345–357 Folkman J (1990) What is the evidence that tumors are angiogenesis dependent? J Natl Cancer Inst 82(1):4–6 Giuriato S, Ryeom S, Fan AC, Bachireddy P, Lynch RC, Rioth MJ, van Riggelen J, Kopelman AM, Passegue E, Tang F, Folkman J, Felsher DW (2006) Sustained regression of tumors upon MYC inactivation requires p53 or thrombospondin-1 to reverse the angiogenic switch. Proc Natl Acad Sci USA 103(44):16266–16271 Greenblatt MS, Bennett WP, Hollstein M, Harris CC (1994) Mutations in the p53 tumor suppressor gene: clues to cancer etiology and molecular pathogenesis. Cancer Res 54(18):4855–4878 Hainaut P, Hollstein M (2000) p53 and human cancer: the first ten thousand mutations. Adv Cancer Res 77:81–137 Hanahan D, Weinberg RA (2000) The hallmarks of cancer. Cell 100(1):57–70 Haupt S, Berger M, Goldberg Z, Haupt Y (2003) Apoptosis – the p53 network. J Cell Sci 116(Pt 20):4077–4085 Hollstein M, Sidransky D, Vogelstein B, Harris CC (1991) p53 mutations in human cancers. Science 253(5015):49–53 Hsu SC, Volpert OV, Steck PA, Mikkelsen T, Polverini PJ, Rao S, Chou P, Bouck NP (1996) Inhibition of angiogenesis in human glioblastomas by chromosome 10 induction of thrombospondin-1. Cancer Res 56(24):5684–5691 Iddings DM, Koda EA, Grewal SS, Parker R, Saha S, Bilchik A (2007) Association of angiogenesis markers with lymph

141

node metastasis in early colorectal cancer. Arch Surg 142(8):738–744. discussion 744–735 Kazuno M, Tokunaga T, Oshika Y, Tanaka Y, Tsugane R, Kijima H, Yamazaki H, Ueyama Y, Nakamura M (1999) Thrombospondin-2 (TSP2) expression is inversely correlated with vascularity in glioma. Eur J Cancer 35(3): 502–506 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) Long-term survival with glioblastoma multiforme. Brain 130(Pt 10): 2596–2606 Kunz C, Pebler S, Otte J, von der Ahe D (1995) Differential regulation of plasminogen activator and inhibitor gene transcription by the tumor suppressor p53. Nucleic Acids Res 23(18):3710–3717 Kupryjanczyk J, Thor AD, Beauchamp R, Merritt V, Edgerton SM, Bell DA, Yandell DW (1993) p53 gene mutations and protein accumulation in human ovarian cancer. Proc Natl Acad Sci USA 90(11):4961–4965 Lakka SS, Gondi CS, Yanamandra N, Dinh DH, Olivero WC, Gujrati M, Rao JS (2003) Synergistic down-regulation of urokinase plasminogen activator receptor and matrix metalloproteinase-9 in SNB19 glioblastoma cells efficiently inhibits glioma cell invasion, angiogenesis, and tumor growth. Cancer Res 63(10):2454–2461 Leon SP, Folkerth RD, Black PM (1996) Microvessel density is a prognostic indicator for patients with astroglial brain tumors. Cancer 77(2):362–372 Liu J, Zhan M, Hannay JA, Das P, Bolshakov SV, Kotilingam D, Yu D, Lazar AF, Pollock RE, Lev D (2006) Wildtype p53 inhibits nuclear factor-kappaB-induced matrix metalloproteinase-9 promoter activation: implications for soft tissue sarcoma growth and metastasis. Mol Cancer Res 4(11):803–810 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(2):97–109 Lutzker SG, Levine AJ (1996) A functionally inactive p53 protein in teratocarcinoma cells is activated by either DNA damage or cellular differentiation. Nat Med 2(7): 804–810 Nor JE, Mitra RS, Sutorik MM, Mooney DJ, Castle VP, Polverini PJ (2000) Thrombospondin-1 induces endothelial cell apoptosis and inhibits angiogenesis by activating the caspase death pathway. J Vasc Res 37(3):209–218 Okamoto Y, Di Patre PL, Burkhard C, Horstmann S, Jourde B, Fahey M, Schuler D, Probst-Hensch NM, Yasargil MG, Yonekawa Y, Lutolf UM, Kleihues P, Ohgaki H (2004) Population-based study on incidence, survival rates, and genetic alterations of low-grade diffuse astrocytomas and oligodendrogliomas. Acta Neuropathol 108(1):49–56 Reifenberger J, Ring GU, Gies U, Cobbers L, Oberstrass J, An HX, Niederacher D, Wechsler W, Reifenberger G (1996) Analysis of p53 mutation and epidermal growth factor receptor amplification in recurrent gliomas with malignant progression. J Neuropathol Exp Neurol 55(7):822–831 Schiffer D, Chio A, Giordana MT, Mauro A, Migheli A, Vigliani MC (1989) The vascular response to tumor infiltration in

142 malignant gliomas. Morphometric and reconstruction study. Acta Neuropathol 77(4):369–378 Stander M, Peraud A, Leroch B, Kreth FW (2004) Prognostic impact of TP53 mutation status for adult patients with supratentorial World Health Organization Grade II astrocytoma or oligoastrocytoma: a long-term analysis. Cancer 101(5):1028–1035. doi:10.1002/cncr.20432 Takeuchi H, Hashimoto N, Kitai R, Kubota T, Kikuta KI (2009) Proliferation of vascular smooth muscle cells in glioblastoma multiforme. J Neurosurg 113(2):218–224 Tanaka H, Arakawa H, Yamaguchi T, Shiraishi K, Fukuda S, Matsui K, Takei Y, Nakamura Y (2000) A ribonucleotide reductase gene involved in a p53-dependent cell-cycle checkpoint for DNA damage. Nature 404(6773):42–49 Teodoro JG, Parker AE, Zhu X, Green MR (2006) p53-mediated inhibition of angiogenesis through up-regulation of a collagen prolyl hydroxylase. Science 313(5789):968–971 Teodoro JG, Evans SK, Green MR (2007) Inhibition of tumor angiogenesis by p53: a new role for the guardian of the genome. J Mol Med 85(11):1175–1186 Vogelstein B, Lane D, Levine AJ (2000) Surfing the p53 network. Nature 408(6810):307–310 Watanabe K, Sato K, Biernat W, Tachibana O, von Ammon K, Ogata N, Yonekawa Y, Kleihues P, Ohgaki H (1997) Incidence and timing of p53 mutations during astrocytoma progression in patients with multiple biopsies. Clin Cancer Res 3(4):523–530

T. Gaiser and M.D. Siegelin Weidner N, Semple JP, Welch WR, Folkman J (1991) Tumor angiogenesis and metastasis–correlation in invasive breast carcinoma. N Engl J Med 324(1):1–8 Weidner N, Folkman J, Pozza F, Bevilacqua P, Allred EN, Moore DH, Meli S, Gasparini G (1992) Tumor angiogenesis: a new significant and independent prognostic indicator in early-stage breast carcinoma. J Natl Cancer Inst 84(24): 1875–1887 Wesseling P, Schlingemann RO, Rietveld FJ, Link M, Burger PC, Ruiter DJ (1995) Early and extensive contribution of pericytes/vascular smooth muscle cells to microvascular proliferation in glioblastoma multiforme: an immuno-light and immuno-electron microscopic study. J Neuropathol Exp Neurol 54(3):304–310 Wesseling P, Ruiter DJ, Burger PC (1997) Angiogenesis in brain tumors; pathobiological and clinical aspects. J Neurooncol 32(3):253–265 Wesseling P, van der Laak JA, Link M, Teepen HL, Ruiter DJ (1998) Quantitative analysis of microvascular changes in diffuse astrocytic neoplasms with increasing grade of malignancy. Hum Pathol 29(4):352–358 Yamanishi Y, Boyle DL, Pinkoski MJ, Mahboubi A, Lin T, Han Z, Zvaifler NJ, Green DR, Firestein GS (2002) Regulation of joint destruction and inflammation by p53 in collageninduced arthritis. Am J Pathol 160(1):123–130

Chapter 17

Spontaneous Regression of Cerebellar Astrocytomas Mansoor Foroughi, Shibu Pillai, and Paul Steinbok

Abstract Cerebellar astrocytomas (CA) in children are associated with good long term prognosis when resected totally. However, when these tumours cannot be surgically excised, they present a management dilemma. Adjuvant radiotherapy or chemotherapy for the residual tumour is controversial because arrested growth or even regression of the astrocytoma remnants is a well recognized, albeit poorly understood phenomenon. The definition, incidence, evolution, and mechanisms of regression of CAs are discussed. Regression of CA is seen in about 16% of residuals and is more likely when the size of the residual is small (possibly less than 3.3 cm3 ). Keywords Spontaneous regression · Cerebellar astrocytomas · Tumours · radiotherapy · MRI imaging · surgery

Introduction “The art of medicine is in amusing a patient while nature affects the cure.” Voltaire “Nature is the curer of disease.” Hippocrates, (the father of medicine)

Spontaneous regression of a tumour may be defined as a clear decrease in the mass of a tumour without any

P. Steinbok () Division of Neurosurgery, B.C.’s Children Hospital, Vancouver, BC, Canada e-mail: [email protected]

medical intervention. Such a phenomenon has been described in many tumours including haemangioma, sarcoma, adenocarcinoma, gastric cancer, neuroblastoma, retinoblastoma, renal cell carcinoma, malignant melanoma, lymphomas/leukaemia, gestational trophoblastic disease e.g. choriocarcinoma, germinoma, infantile fibrosarcoma, and clear cell carcinoma of endometrium. Regarding intracranial tumours it has been well described in vestibular schwannoma (Flint et al., 2005), glioma (Rozen et al., 2008) and also cerebellar astrocytoma (CA) (Davis and Joglekar, 1981; Dirven et al., 1997; Due-Tonnessen et al., 2002; Gunny et al., 2005; Smoots et al., 1998; Steinbok et al., 2006). Astrocytoma regression has been mostly described in tumours occurring in children, following subtotal resection or biopsy. This has been mainly for tumours involving the optic chiasm, and especially in those associated with neurofibromatosis type 1 (NF1). The first description of such a case was by Borit and Richardson, 1982, who reported the first convincing case in a 1 year old child presenting with a large optic chiasmatic/hypothalamic tumour and obstructive hydrocephalus. After biopsy and confirmation of a pilocytic astrocytoma no further treatment was given. The hydrocephalus resolved spontaneously and following death of the child from pneumonia 12 years later, autopsy showed that there was a frontal lobe cyst, but no tumour. Since then several cases or small case series have been reported documenting partial regression of biopsied or unbiopsied optic pathway tumours associated with NF1 (Rossi et al., 1999; Schmandt et al., 2000). A similar phenomenon has been reported for optic hypothalamic gliomas in the absence of known NF1 (Parsa et al., 2001; Takeuchi et al., 1997). There have also been reports of regression of Non-NF1 associated low grade astrocytoma

M.A. Hayat (ed.), Tumors of the Central Nervous System, Volume 5, DOI 10.1007/978-94-007-2019-0_17, © Springer Science+Business Media B.V. 2012

143

144

at varied locations, such as, the cerebellum, (Gunny et al., 2005; Palma et al., 2004; Saunders et al., 2005; Steinbok, 1994; Steinbok et al., 2006), thalamus (Balkhoyor and Bernstein, 2000) and temporal lobe (Rozen et al., 2008). Although referred to as spontaneous, the cause for such a phenomenon may involve a combination factors in any individual case. However such regression without medical intervention (i.e. repeat surgery, radiotherapy or chemotherapy) is important to recognize and acknowledge. Recognition of this phenomenon may encourage a more conservative approach after subtotal resection of the tumour, involving surveillance during follow-up of patients, if the tumor is asymptomatic. The first line therapy for CA is surgical resection, which if complete, is generally curative. The aims of surgery for intracranial CA are to obtain tissue diagnosis, treat the mass effect and any hydrocephalus by restoring CSF flow, and also to achieve maximum cyto-reduction without causing new morbidity. In some patients only a subtotal resection can be achieved safely and it is in this population where the issue of spontaneous regression becomes important.

Criteria for Regression There are no consistent and clear criteria in the literature regarding identification of regression of intracranial tumours, including CAs. Regression of CA can be said to occur following histological confirmation of the tumour via biopsy/excision, if a radiologically proven nodular residual, as detected on a 24–48 h/3

Fig. 17.1 CT scans showing spontaneous regression of residual cerebellar astrocytoma- pre-operative, residual enhancing tumour 48 h after resection, and regression of tumour after 12 years without any further treatment

M. Foroughi et al.

month post-resection CT or MRI, shows unequivocal radiological regression. The nodular residual tumour should not be confused with the linear enhancement which is sometimes seen along the tumour resection margin after resection and is a manifestation of postsurgical change (Cairncross et al., 1985). Scans done 3–5 days after surgery can sometimes show nodular enhancement mimicking tumour residual and hence, the need to document tumour residual within 48 h postresection or delay this for about 3 months by which time the acute post-operative changes will have had time to subside (Rollins et al., 1998). The larger the residue, the easier it is to note and document such regression. Figure 17.1 shows an example of regression of CA.

Incidence and Time Course of Regression There is controversy regarding the incidence of regression of CAs. Most agree that progression of residual CAs is more common than regression and this is often anticipated during surveillance. Abdollahzadeh et al. reported on CA progression in all 5 children with residual CA following surgery who were followed up for 1–8 years post resection, while Dirven found that only one out of 22 children with residual CA had regression of tumour (Abdollahzadeh et al., 1994; Dirven et al., 1997). Regression of CA residuals is quoted in one review to occur in 14% of cases with cerebellar astrocytoma residual following surgery (Palma et al., 2004). Due-Tønnessen et al., reporting on the long term outcome of 110 consecutive cases of surgically resected CAs concluded that among the 40 patients

17 Spontaneous Regression of Cerebellar Astrocytomas

with serial MRI imaging, regression occurred spontaneously in 3 of 7 patients with residual tumour 11.9– 12.5 years after surgery (Due-Tonnessen et al., 2002). Saunders et al. in a series of 84 CAs treated surgically, described 14 cases with residual disease. Eleven of these did not receive adjuvant therapy, and five of the 11 tumours (45%) regressed during surveillance period ranging from 7 to 66 months (mean of 32 months) (Saunders et al., 2005). Smoots et al. reported spontaneous tumour regression in 4 (24%) of 17 patients with residual CAs without NF1 following surgery (tumour volume ranging from 0.1 to 2.5 cm3 ) at a mean of 21 months (range 3–49 months) (Smoots et al., 1998). Davis et al. found that among 5 patients out of 43 with CA who had serial CT scans only one showed regression of the residual tumour (Davis and Joglekar, 1981). Thus, from the above case series with radiographically documented residuals, we estimate that the incidence of regression of residual CA is about 16% (16 out of 92). The time course of CA regression ranges widely from 3 months (Smoots et al., 1998) to 11–12 years (Due-Tonnessen et al., 2002; Steinbok et al., 2006).

Imaging Surveillance It is important to document re-growth, stability or regression of any residual disease following surgery for CA, and note the response to any adjuvant radiotherapy and/or chemotherapy that may have been indicated for significant residual or recurrent disease. However, it is not clear how long imaging surveillance is required. In a review aiming to clarify recommendations for surveillance frequency and duration, Saunders et al. followed up 84 children with CA for a mean period of 73 months (range 2–159 months). Of the 11 children who did not receive any further therapy following an incomplete resection, tumor progression leading to further treatment was detected in five at 7, 9, 12, 13, and 20 months from initial surgery. Tumour regression was observed in 5 children at 7, 16, 29, 42, and 66 months, respectively. Further imaging for a maximum of 3 years and 10 months in two of the latter group did not reveal any further progression. They recommended that for follow up of residual tumor, 6-month interval imaging for at least 3 years, yearly images for another 2 years, and subsequent

145

2-year imaging should be done. No clear recommendations regarding follow-up imaging after regression of a residual CA can be made at present but 2-year imaging for another 4 years after regression seems appropriate.

Mechanisms of Regression “Drugs can only repress symptoms: they cannot eradicate disease. The true remedy for all diseases is Nature’s remedy . . . There is at bottom only one genuinely scientific treatment for all diseases, and that is to stimulate the phagocytes. Stimulate the phagocytes. Drugs are a delusion”. Thus said the physician Sir Bloomfield Bonington in George Bernard Shaw’s 1906 play, The Doctor’s Dilemma. Multiple factors such as genetics, environment, diet, and lifestyle are known to be involved in various degrees and combinations in the development of different tumours in different individuals. Therefore, similarly there must be various combinations of factors, each with various degrees of influence which cause tumour regression. Such multiple factors include immunological mechanisms, the effects of growth factors or cytokines, induction of differentiation, hormonal factors, elimination of a carcinogen, ischemic tumour necrosis following blood supply occlusion (possibly secondary to inhibition of angiogenesis), psychological mechanisms, and apoptosis (Bodey, 2002; Papac, 1998; Steinbok et al., 2006). In such tumour regression, the rate of cell death must exceed cellular production. Cellular necrosis following vascular damage to tumour during surgery is easy to envisage, and this may also occur following inhibition of angiogenesis. For apoptosis to occur, it may require the confluence of multiple apoptosis-inducing phenomena (Hercbergs, 1999). These could include general lowering of the threshold of tumour cells to undergo apoptosis, as might occur with hypothyroidism or other hormonal changes; an apoptosis-promoting intercurrent condition, such as pyrexia or infection as implicated for the anti-tumour effects of Coley’s vaccine and induced infections in sarcomas (Hoption Cann et al., 2002); or increased ability of the host T-cells to kill tumour cells, as might occur with increased autoimmunity to the tumour. It is easy to appreciate a combination of such factors causing frequent

146

tumour regression, when the rate of cellular division and growth in such relatively benign tumours as CA is so low, as compared to more malignant tumours. Activation of the host immune system following an infection may be a factor but there are no such reports associated with these regressions. Exposure of the hidden antigenic sites on the tumour to the immune system following surgery may be a factor but this is not likely as a major factor since such regressions can occur after several years of stability (Steinbok et al., 2006). Almost certainly the reduction in the critical tumour mass by surgery and any subsequent avascular necrosis is a major factor in facilitating tumour regression, particularly when this regression occurs within a few months after the primary surgery. An attractive theory recently proposed relates to the reported finding that, in brain tumours, there maybe relatively small populations of cancer stem cells capable of proliferation, and a larger population of other tumour cells originally derived from cancer stem cells but incapable of proliferation (Singh et al., 2004). Following surgery and mass reduction, it is conceivable that a small amount of residual tumour might contain no cancer stem cells and would, therefore be incapable of proliferation. The remaining cells would die due to apoptosis and this explanation would fit with the time course seen in spontaneous regression of some CA (Steinbok et al., 2006).

Associated Factors Influencing Regression Size In the series by Smoots et al. describing spontaneous tumour regression in 4 (24%) of 17 patients with residual disease (tumour volume ranging from 0.1 to 2.5 cm3 ), it was concluded that residual tumours with a smaller post-operative volume had a higher chance of regression. Gunny et al. in a series of 83 lowgrade CAs, including 13 with incomplete resection, did not find statistically significant independent variables (symptomatology, age, gender, histological grade or the Ki-67 fraction) as predictors of spontaneous regression (Gunny et al., 2005). The numbers involved were too small for meaningful statistical significance of the independent variables, but interestingly tumour

M. Foroughi et al.

progression did not occur in children with residual tumour volumes below 10 cm3 . However, we have personally seen progression occurring in CA tumour residuals with initial volumes lower than 10 cm3 , which is in reality a substantial tumour residual volume to begin with.

Hormones The onset of menarche and rise in female hormones causing tumour regression may be an important factor. There is evidence regarding the protective role of estrogens and other steroid hormones against gliomas, with 1.5 times higher incidence rates of gliomas in men compared to women (Huang et al., 2004; Wigertz et al., 2008), changes in incidence around menarche and menopause (Inskip et al., 1995; McKinley et al., 2000), and presence of hormone receptors in glial tumours (Felini et al., 2009). A case-control study showed decreased incidence amongst women who use hormones during menopause (Huang et al., 2004), and another showed lower risk of gliomas in women that had ever been pregnant compared to never pregnant, with the decreased risk with increasing number of pregnancies (Wigertz et al., 2008).

Use of Dietary Supplements, Natural Remedies and Herbs Patients and families afflicted with various cancers and tumours often resort to the use of natural foods, herbs and remedies to help their immune system battle with cancer and tumours. There are a myriad of dietary supplements that are said and reported to have beneficial anti-cancer and anti-tumour effects. Examples include Broccoli, cauliflower and brussles sprouts rich in 3,3-diindolylmethane (DIM) implicated in prostate (Ahmad et al., 2009), breast (Hong et al., 2002) and pancreatic (Abdelrahim et al., 2006) cancer effects. Garlic is well used for many benefits including its anticancer effects (Ariga and Seki, 2006), with some of its constituents being implicated such as Ajoene,Alk(en)yl sulfides, and Gamma-glutamyl-Semethylselenocysteine (GGMSC). The Chinese herbs containing Camptothecin and hydroxycamptothecin; harringtonine and homoharringtonine; colchicine and

17 Spontaneous Regression of Cerebellar Astrocytomas

colchicinamide; curzerenone; monocrotaline; lycobetaine; oridonin; indirubin; cantharidinare thought to have a positive anti-cancer effect (Hsu, 1980). However the degree of impact of the myriad possible factors in causing regression of low grade glioamas is almost impossible to ascertain, and maybe it is simply best accepted that diet, lifestyle, environment and psychological factors can all, through the immune system, exert a variable degree of influence in encouraging tumour regression.

Conclusion CAs whether treated surgically (residual) or not, are well recognised to undergo regression without conventional medical intervention in the form of surgery, radiotherapy and chemotherapy. The frequency of such regression is debated but is estimated to be about 16%. What is more certain is that such regression seems to occur in small tumour remnants (up to 3.3 cm3 ) between 3 months and 11–12 years post primary surgical intervention. The etiology of such regression is likely to be multi factorial and differs from case to case, although there is likely to be influence of multiple mechanisms such as the immune system, cyto-reduction below a critical mass or critical cell type, hormonal influence, apoptosis, and devascularisation. This regression can even follow initial growth of the surgical remnant as we have observed in cases of PA outside the cerebellum. Therefore caution should be exercised during surveillance of CA surgical remnants, not to rush into further surgery, radiotherapy or chemotherapy.

References Abdelrahim M, Newman K, Vanderlaag K, Samudio I, Safe S (2006) 3,3 -diindolylmethane (DIM) and its derivatives induce apoptosis in pancreatic cancer cells through endoplasmic reticulum stress-dependent upregulation of DR5. Carcinogenesis 27:717–728 Abdollahzadeh M, Hoffman HJ, Blazer SI, Becker LE, Humphreys RP, Drake JM, Rutka JT (1994) Benign cerebellar astrocytoma in childhood: experience at the Hospital for Sick Children 1980–1992. Childs Nerv Syst 10: 380–383

147 Ahmad A, Kong D, Sarkar SH, Wang Z, Banerjee S, Sarkar FH (2009) Inactivation of uPA and its receptor uPAR by 3,3 diindolylmethane (DIM) leads to the inhibition of prostate cancer cell growth and migration. J Cell Biochem 107: 516–527 Ariga T, Seki T (2006) Antithrombotic and anticancer effects of garlic-derived sulfur compounds: a review. Biofactors 26: 93–103 Balkhoyor KB, Bernstein M (2000) Involution of diencephalic pilocytic astrocytoma after partial resection. Report of two cases in adults. J Neurosurg 93:484–486 Bodey B (2002) Spontaneous regressionof neoplasms: new possibilities for immunotherapy. Expert Opin Biol Ther 2: 459–476 Borit A, Richardson EP Jr (1982) The biological and clinical behaviour of pilocytic astrocytomas of the optic pathways. Brain 195:167–187 Cairncross JG, Pexman JH, Rathbone MP (1985) Post-surgical contrast enhancement mimicking residual brain tumour. Can J Neurol Sci 12:75 Davis CH, Joglekar VM (1981) Cerebellar astrocytomas in children and young adults. J Neurol Neurosurg Psychiatry 44:820–828 Dirven CM, Mooij JJ, Molenaar WM (1997) Cerebellar pilocytic astrocytoma: a treatment protocol based upon analysis of 73 cases and a review of the literature. Childs Nerv Syst 13: 17–23 Due-Tonnessen BJ, Helseth E, Scheie D, Skullerud K, Aamodt G, Lundar T (2002) Long-term outcome after resection of benign cerebellar astrocytomas in children and young adults (0–19 years): report of 110 consecutive cases. Pediatr Neurosurg 37:71–80 Felini MJ, Olshan AF, Schroeder JC, Carozza SE, Miike R, Rice T, Wrensch M (2009) Reproductive factors and hormone use and risk of adult gliomas. Cancer Causes Control 20: 87–96 Flint D, Fagan P, Panarese A (2005) Conservative management of sporadic unilateral acoustic neuromas. J Laryngol Otol 119:424–428 Gunny RS, Hayward RD, Phipps KP, Harding BN, Saunders DE (2005) Spontaneous regression of residual low-grade cerebellar pilocytic astrocytomas in children. Pediatr Radiol 35:1086–1091 Hercbergs A (1999) Spontaneous remission of cancer—a thyroid hormone dependent phenomenon? Anticancer Res 19: 4839–4844 Hong C, Firestone GL, Bjeldanes LF (2002) Bcl-2 familymediated apoptotic effects of 3,3 -diindolylmethane (DIM) in human breast cancer cells. Biochem Pharmacol 63: 1085–1097 Hoption Cann SA, van Netten JP, van Netten C, Glover DW (2002) Spontaneous regression: a hidden treasure buried in time. Med Hypotheses 58:115–119 Hsu B (1980) The use of herbs as anticancer agents. Am J Chin Med 8:301–306 Huang K, Whelan EA, Ruder AM, Ward EM, Deddens JA, Davis-King KE, Carreon T, Waters MA, Butler MA, Calvert GM, Schulte PA, Zivkovich Z, Heineman EF, Mandel JS, Morton RF, Reding DJ, Rosenman KD (2004) Reproductive factors and risk of glioma in women. Cancer Epidemiol Biomarkers Prev 13:1583–1588

148 Inskip PD, Linet MS, Heineman EF (1995) Etiology of brain tumors in adults. Epidemiol Rev 17:382–414 McKinley BP, Michalek AM, Fenstermaker RA, Plunkett RJ (2000) The impact of age and sex on the incidence of glial tumors in New York state from 1976 to 1995. J Neurosurg 93:932–939 Palma L, Celli P, Mariottini A (2004) Long-term follow-up of childhood cerebellar astrocytomas after incomplete resection with particular reference to arrested growth or spontaneo.us tumour regression. Acta Neurochir (Wien) 146:581–588, discussion 588 Papac RJ (1998) Spontaneous regression of cancer: possible mechanisms. In Vivo 12:571–578 Parsa CF, Hoyt CS, Lesser RL, Weinstein JM, Strother CM, Muci-Mendoza R, Ramella M, Manor RS, Fletcher WA, Repka MX, Garrity JA, Ebner RN, Monteiro ML, McFadzean RM, Rubtsova IV, Hoyt WF (2001) Spontaneous regression of optic gliomas: thirteen cases documented by serial neuroimaging. Arch Ophthalmol 119:516–529 Rollins NK, Nisen P, Shapiro KN (1998) The use of early postoperative MR in detecting residual juvenile cerebellar pilocytic astrocytoma. Am J Neuroradiol 19:151–156 Rossi LN, Triulzi F, Parazzini C, Maninetti MM (1999) Spontaneous improvement of optic pathway lesions in children with neurofibromatosis type 1. Neuropaediatrics 30:205–209 Rozen WM, Joseph S, Lo PA (2008) Spontaneous regression of low-grade gliomas in pediatric patients without neurofibromatosis. Pediatr Neurosurg 44:324–328

M. Foroughi et al. Saunders DE, Phipps KP, Wade AM, Hayward RD (2005) Surveillance imaging strategies following surgery and/or radiotherapy for childhood cerebellar low-grade astrocytoma. J Neurosurg 102:172–178 Schmandt SM, Packer RJ, Vezina LG, Jane J (2000) Spontaneous regression of low-grade astrocytomas in childhood. Pediatr Neurosurg 32:132–136 Singh SK, Clarke ID, Hide T, Dirks PB (2004) Cancer stem cells in nervous system tumors. Oncogene 23:7267–7273 Smoots DW, Geyer JR, Lieberman DM, Berger MS (1998) Predicting disease progression in childhood cerebellar astrocytoma. Childs Nerv Syst 14:636–648 Steinbok P (1994) Management and outcome of low-grade astrocytomas of the midline in children: A retrospective review. Neurosurgery 35:342–343 Steinbok P, Poskitt K, Hendson G (2006) Spontaneous regression of cerebellar astrocytoma after subtotal resection. Childs Nerv Syst 22:572–576 Takeuchi H, Sato KM, Kubota K, Tattersall M (1997) Chiasmal gliomas with spontaneous regression: proliferation and apoptosis. Childs Nerv Syst 13:229–233 Wigertz A, Lonn S, Hall P, Auvinen A, Christensen HC, Johansen C, Klaeboe L, Salminen T, Schoemaker MJ, Swerdlow AJ, Tynes T, Feychting M (2008) Reproductive factors and risk of meningioma and glioma. Cancer Epidemiol Biomarkers Prev 17:2663–2670

Chapter 18

Subependymal Giant Cell Astrocytoma: Gene Expression Profiling Magdalena Ewa Tyburczy and Bozena Kaminska

Abstract Subependymal giant cell astrocytomas (SEGAs) are rare brain tumors occurring in patients with tuberous sclerosis complex (TSC). TSC is characterized by formation of benign tumors in many organs and neurologic disorders (epilepsy, mental retardation, and autism). The disease is caused by mutations in TSC1 (hamartin) or TSC2 (tuberin) which encode tumor suppressors. Inactivation of TSC1/TSC2 leads to enhanced activity of mTOR kinase that regulates i.a. transcription, translation, and cell growth. Data from experiments on lower eukaryotes showed several hundred newly identified genes modulated by TOR at the transcriptional level. However, very little was known about such genes in the human brain. Therefore, we performed global gene expression profiling in SEGAs with use of Affymetrix microarrays, which revealed a considerable number of differentially expressed genes. Identified genes differentially expressed in SEGAs when compared to normal human brain were mainly involved in tumorigenesis (ANXA1, GPNMB, S100A11, LTF, RND3, and SFRP4; up-regulated in SEGA), and the nervous system development and differentiation. Moreover, down-regulation of genes associated with functioning of the nervous system (e.g. CADPS2, CNDP1, ERMIN, GABRA1, MBP, MOBP, NEUROD1, NPTX1, RELN, SHANK3, and TF) in SEGA may be a cause of neurologic dysfunctions in TSC patients. Furthermore, up-regulation of MBP, ERMIN (markers of oligodendrocytes), NEUROD1, and NPTX1 (markers of immature neurons) in SEGA

M.E. Tyburczy () Translational Medicine Division, Brigham and Women’s Hospital, Boston, MA e-mail: [email protected]

cells compared with normal human astrocytes suggested their mixed, glio-neuronal nature and inability to undergo terminal differentiation. Since SEGAs in patients with TSC appear to be of a mixed glioneuronal lineage, the current practice of classifying these tumors as astrocytomas merits revision. Finally, it was shown that expression of ANXA1, GPNMB, S100A11, LTF, RND3, SFRP4, and NPTX1 was regulated by mTOR in SEGA cells. The presented data demonstrate new insights into transcriptional regulation induced by mTOR signaling dysfunctions. We anticipate that these results may influence future preclinical and clinical trials for TSC. Keywords Subependymal giant cell · Intratumoral · Transcriptional regulation · Autism · Tsc1 · SEGA

Introduction Subependymal giant cell astrocytomas (SEGAs) are primary brain tumors that occur, in most cases, in the first two decades of life near the foramen of Monro and grow inside the lateral ventricle (Buccoliero et al., 2009). SEGAs may derive from subependymal nodules (SENs) which arise in fetal age, are asymptomatic, and about 10% of them grow and differentiate into SEGAs. SEGAs are larger than SENs (lesions greater than 12 mm are classified as SEGAs) and, in contrast to SENs, enlarge progressively. SEGAs are well circumscribed and slowly growing, thus they were categorized as benign (World Health Organization [WHO] Grade I). In spite of the benign nature, due to their localization, the tumors may cause obstruction of cerebrospinal fluid flow, hydrocephalus, and an increase

M.A. Hayat (ed.), Tumors of the Central Nervous System, Volume 5, DOI 10.1007/978-94-007-2019-0_18, © Springer Science+Business Media B.V. 2012

149

150

of intracranial pressure leading to intratumoral hemorrhages, blurred vision, and death. SEGAs are diagnosed almost exclusively in patients with tuberous sclerosis complex (TSC) and appear in 10–20% of TSC cases (Jozwiak and Jozwiak, 2007). Tuberous sclerosis complex (TSC) is a disease affecting about one in 6000 people and characterized by formation of benign tumors and hamartomas (benign tumor-like growths composed of normal mature cells in abnormal number and distribution) in many organs, mainly in the skin, brain, kidneys, lungs, heart, and liver (Jozwiak et al., 2008). Neurologic disorders, such as epilepsy, mental retardation, and neurobehavioral abnormalities, like autism, occur in over 85% of TSC patients. They are assumed to be caused by structural alterations in the brain, namely aforementioned SEGAs and SENs, as well as cerebral cortical tubers. Tubers develop in fetal age due to aberrant proliferation of neurons, and are characterized histologically by a loss of the normal six-layered structure of the cortex (Jozwiak and Jozwiak, 2007). Epilepsy is the most prevalent clinical symptom of TSC – it affects at least 90% of patients. It was shown that seizures mostly originate in the regions of tubers location in the brain, hence it is believed that tubers serve as the epileptogenic foci. Indeed, surgical resection of tubers often reduces seizures in patients with intractable epilepsy. However, some patients continue to seize after tuberectomy. Moreover, occurrence of seizures was observed in TSC patients in the absence of cortical tubers. Animal studies have demonstrated that mice with conditional inactivation of Tsc1 in glial cells (Tsc1-GFAP-CKO mice) develop progressive seizures, suggesting that glial dysfunction may be involved in epileptogenesis in TSC (Orlova and Crino, 2010). Mental retardation, from mild to severe, appears in about 50% of TSC patients. Occurrence of its profound forms is highly related to early development of epilepsy and its intractability, as well as location and a higher number of cortical tubers. Furthermore, early occurrence of seizures was also shown to be associated with the incidence of autism (Curatolo et al., 2010).

Molecular Pathophysiology of TSC Tuberous sclerosis complex is caused by inactivating mutations in either TSC1 (tuberous sclerosis complex 1), located on chromosome 9 (9q34), or

M.E. Tyburczy and B. Kaminska

TSC2 (tuberous sclerosis complex 2), located on chromosome 16 (16p13.3). TSC can be inherited as an autosomal dominant disorder, though about 70% of cases result from spontaneous germline mutations in TSC1 or TSC2 (Kwiatkowski, 2003). TSC1 encodes a 130-kDa protein TSC1/hamartin and TSC2 encodes a 200-kDa protein TSC2/tuberin. The proteins bind to each other to form functional heterodimers TSC1/TSC2 which regulate multitude of cellular processes, including cell size, proliferation, and cell cycle. The TSC1/TSC2 complex is the main cellular inhibitor of mTOR (mammalian target of rapamycin). TSC2 contains a C-terminal GTPase activating protein domain (GAP) and acts as a GTPase activating protein toward Rheb (Ras homolog enriched in brain), a Ras family GTPase. Rheb-GTP activates mTOR by preventing its binding to its endogenous inhibitor FKBP38 (FK506binding protein 38). FKBP38 is a member of the FKBP family that is structurally related to FKBP12. TSC1/TSC2 inhibits mTOR activity by stimulating the conversion of active Rheb-GTP into inactive Rheb-GDP. mTOR – a 289-kDa serine-threonine kinase is a central component of two distinct multiprotein complexes: mTORC1 (mTOR complex 1) and mTORC2 (mTOR complex 2). mTORC1 is mainly involved in regulation of transcription, translation, cell cycle, autophagy, and microtubule dynamics, whereas mTORC2 controls the actin cytoskeleton organization by promoting PKCα (protein kinase Cα) phosphorylation. It was also shown to activate Akt and SGK1 (serum and glucocorticoid inducible kinase 1). The best-characterized effectors of mTOR are S6K1 (p70 ribosomal S6 kinase 1) and 4E-BP1 (eukaryotic initiation factor 4E (eIF4E)-binding protein 1) which are phosphorylated by mTORC1 to promote translation (Fig. 18.1). Gene deletion studies in mice and fruit fly have shown that S6 kinases, dS6K in fruit fly, and S6K1 and S6K2 in mice, are important regulators of cell growth in response to insulin stimulation and nutrition availability. mTORC1 functions can be inhibited by the bacterial macrolide rapamycin which is a specific inhibitor of mTOR kinase activity. Rapamycin and its analogs (rapalogs) form a complex with the intracellular receptor FK506 binding protein 12 (FKBP12) and then bind to the FKBP12-rapamycin binding domain (FRB) of mTOR, inhibiting mTORC1 functions. On the other hand, mTORC2 was shown to be insensitive to

18 Subependymal Giant Cell Astrocytoma: Gene Expression Profiling

151

hundred genes regulated by TOR kinase at the transcriptional level. However, very little was known about such genes in the human brain. Microarray technology is widely employed for studying the molecular mechanisms underlying complex diseases. The search for relevant gene expression changes is challenged by the identification of very large datasets of differentially expressed genes with uncertain relationships to disease pathogenesis. In our previous study, we performed gene expression profiling in SEGA samples from TSC patients to identify genes that are associated with the clinical symptoms of TSC in the brain and to find out which of them are regulated by mTOR in SEGAs (Tyburczy et al., 2010). Several newly identified genes were found to be regulated by mTOR kinase in cultured SEGA-derived cells. In this chapter, we focus on the potential roles of newly identified genes in SEGA tumorigenesis and SEGA giant cells differentiation, as well as in neurological disorders characteristic for TSC patients.

Fig. 18.1 TSC1/TSC2-mTOR signaling pathway. Abbreviations: 4E-BP1, 4E-binding protein 1; Akt/PKB, protein kinase B; IRS, insulin receptor substrate; mTORC1, mammalian target of rapamycin complex 1; mTORC2, mammalian target of rapamycin complex 2; PDK1, 3-phosphoinositide-dependent protein kinase 1; PI3K, phosphatidylinositol 3-kinase; PTEN, phosphatase and tensin homolog; Rheb, ras homolog enriched in brain; S6K1, S6 kinase 1; TSC1, tuberous sclerosis complex 1; TSC2, tuberous sclerosis complex 2

rapamycin. However, chronic rapamycin treatment can inhibit mTORC2 activity in some cells by blocking its assembly (Laplante and Sabatini, 2009). Rapamycin has been used in clinical trials to treat SEGAs and a significant tumor shrinkage has been observed (Lam et al., 2010).

Global Gene Expression Profiling and Analysis Despite the discovery of a role of mutations in TSC1 and TSC2 in the pathogenesis of tuberous sclerosis complex, molecular mechanisms involved in this process are still poorly understood. Experiments made on yeast and fruit fly led to identification of several

Genes Up-regulated in SEGA are Linked to Progression or Inhibition of Tumorigenesis Analysis of gene expression profiling in SEGAs with use of Affymetrix microarrays resulted in the identification of functional groups of genes with the biggest difference in expression levels between SEGAs and normal human brain. The newly identified genes are mainly associated with tumorigenesis (most of them up-regulated in SEGA) or with the development and functioning of the nervous system (all of them downregulated in SEGA) (Table 18.1). An extensive literature search revealed that identified genes up-regulated in SEGA are associated with progression or suppression of various tumors. Annexin A1 (lipocortin 1) is a calcium- and phospholipidsbinding protein with anti-inflammatory properties. It was implicated in regulating cell proliferation, differentiation, apoptosis and endocytosis. A role of ANXA1 in tumorigenesis is unclear and varies in various tumor types. Strong expression of annexin A1 was detected in normal breast tissue and benign tumors compared to breast cancer, and low protein levels correlated with worse overall patients survival. However, it was

152

M.E. Tyburczy and B. Kaminska

Table 18.1 Genes with highest up- or down-regulation scores in SEGAs compared to normal human brains Gene Unigene symbol Gene name Fold change over controls Tumorigenesis Hs.658169 SFRP4 Secreted frizzled-related protein Hs.610567 LTF Lactotransferrin Hs.190495 GPNMB Glycoprotein (transmembrane) nmb Hs.494173 ANXA1 Annexin A1 Hs.489142 COL1A2 Collagen, type I, alpha 2 Hs.6838 RND3 Rho family GTPase 3 Hs.148641 CTSH Cathepsin H Hs.593414 S100A11 s100 calcium binding protein a11 Hs.702229 APOD Apolipoprotein D Hs.525205 NDRG2 NDRG family member 2 Hs.444212 VSNL1 Visinin-like 1 Hs.287518 SEPT4 SEPTIN 4 Hs.591255 FAT2 FAT tumor suppressor homolog 2 (Drosophila) Hs.655499 ST18 Suppression of tumorigenicity 18 Nervous system development and functioning Hs.551713 MBP Myelin basic protein Hs.75061 MARCKSL1 MARCKS-like 1 Hs.167317 SNAP25 Synaptosomal-associated protein, 25 kDa Hs.149035 SHANK3 SH3 and multiple ankyrin repeat domains 3 Hs.270055 SH3GL3 SH3-domain GRB2-like 3 Hs.478153 SERPINI1 Neuroserpin Hs.518267 TF Transferring Hs.702002 NPTX1 Neuronal pentraxin I Hs.80395 MAL mal, T-cell differentiation protein Hs.708214 CADPS2 Ca2+-dependent activator protein for secretion 2 Hs.79361 KLK6 Kallikrein-6 Hs.175934 GABRA1 Gamma-aminobutyric acid (GABA) A receptor, alpha 1 Hs.647962 ZIC1 Zic family member 1 (odd-paired homolog, Drosophila) Hs.655654 RELN Reelin Hs.121333 MOBP Myelin-associated oligodendrocyte basic protein Hs.653700 ZIC2 Zic family member 2 (odd-paired homolog, Drosophila) Hs.400613 CNDP1 Carnosine dipeptidase 1 (metallopeptidase M20 family) Hs.443894 ERMN Ermin, ERM-like protein Hs.591255 FAT2 FAT tumor suppressor homolog 2 (Drosophila) Hs.295449 PVALB Parvalbumin Hs.458423 CBLN1 Cerebellin 1 precursor Hs.709709 NEUROD1 Neurogenic differentiation 1 Hs.90791 GABRA6 Gamma-aminobutyric acid (GABA) A receptor, alpha 6 Others and unknown Hs.334629 SLN Sarcolipin Hs.643005 SLC40A1 Solute carrier family 40, member 1 Hs.628678 VIM Vimentin Hs.155247 ALDOC Aldolase C, fructose-bisphosphate Hs.40808 TMEM178 Transmembrane protein 178 Hs.528335 FAM123A Family with sequence similarity

2.514 2.197 2.139 1.791 1.757 1.690 1.451 1.390 0.686 0.668 0.632 0.580 0.423 0.386 0.757 0.682 0.660 0.632 0.608 0.607 0.573 0.540 0.525 0.521 0.514 0.506 0.506 0.504 0.463 0.451 0.444 0.431 0.423 0.412 0.388 0.342 0.311

2.021 1.719 1.314 0.708 0.634 0.555

18 Subependymal Giant Cell Astrocytoma: Gene Expression Profiling Table 18.1 (continued) Gene Unigene symbol

Gene name

Fold change over controls

Hs.592182 AMPH Amphiphysin Hs.208544 KCNK1 Potassium channel, subfamily K, member 1 Hs.201083 MAL2 mal, T-cell differentiation protein 2 Hs.207603 CBLN3 Cerebellin 3 precursor Hs.106576 AGXT2L1 Alanine-glyoxylate aminotransferase 2-like 1 Hs.591707 FSTL5 fstl5 Hs.159523 CRTAM Cytotoxic and regulatory T cell molecule Hs.511757 GJB6 Gap junction beta-6 protein Genes mentioned in the text are printed in bold type

also shown that ANXA1 promoted invasiveness of breast cancer cells. Up-regulation of annexin A1 was observed in astrocytomas, and down-regulation in head and neck cancer (Lim and Pervaiz, 2007; de Graauw et al., 2010; Wang et al., 2010). Annexin A1 interacts with S100A11 (S100C) which belongs to a family of S100 calcium ion-binding proteins. S100A11 participates in regulating inflammation, endocytosis, enzymes activity, cell growth, and apoptosis. Strong expression of S100A11 was detected in uterine smooth muscle, thyroid, stomach and colorectal cancers, and anaplastic large cell lymphoma. The widely observed up-regulation of S100A11 in tumors indicates that S100A11 may be involved in growth enhancement and malignant progression of cancer cells. On the other hand, this protein acted as a potential tumor suppressor in pancreatic and bladder cancerogenesis and inhibited growth of normal human keratinocytes (He et al., 2009). Another identified gene – GPNMB (glycoprotein nonmetastatic melanoma protein b), encodes a membrane glycoprotein. Its higher expression was observed in glioblastoma multiforme and it was shown that GPNMB stimulated migration and invasiveness of breast cancer cells (Kuan et al., 2006; Rose et al., 2010). However, up-regulation of GPNMB was detected in lowly invasive melanoma cell lines. In melanocytes GPNMB contributed to the adhesion of melanocytes with keratinocytes. (Weterman et al., 1995). Another differentially expressed gene encodes lactotransferrin (lactoferrin, LTF), a multifunctional glycoprotein with anti-bacterial and anti-inflammatory properties, which regulates iron homeostasis as well as cell growth and differentiation. It was suggested

153

0.538 0.519 0.473 0.463 0.422 0.408 0.407 0.381

that LTF serves as a tumor suppressor because its enhanced expression inhibited growth and proliferation of nasopharyngeal carcinoma cells and epithelial cancer HeLa cells. Moreover, its down-regulation was detected in prostate cancer (Shaheduzzaman et al., 2007; Zhou et al., 2008). Another identified gene associated with tumorigenesis encodes RND3 (RhoE), a member of Rho GTPase family, involved in regulation of actin cytoskeleton dynamics. Overexpression of RND3 led to cell cycle arrest and induction of apoptosis in glioblastoma cell lines (Poch et al., 2007). On the other hand, it was also shown that silencing RND3 expression decreased migration and invasiveness of melanoma cells, which suggested the involvement of RND3 in stimulation of cell invasiveness (Klein and Aplin, 2009). Furthermore, SFRP4, the next identified gene with higher expression in SEGA, encodes a negative regulator of Wnt signaling pathway, with pro-apoptotic and anti-proliferative properties. Strong expression of SFRP4 was detected in esophageal squamous cell carcinoma and colorectal cancer compared to neighboring healthy tissue (Huang et al., 2010; Zinovyeva et al., 2010). On the other hand, it was shown that overexpression of SFRP4 in endometrial cancer cells caused inhibition of their proliferation (Carmon and Loose, 2008). Identified genes – ANXA1, S100A11, GPNMB, LTF, RND3, and SFRP4, up-regulated in SEGA and acting as regulators of cellular signaling, are potentially involved in development of SEGAs. Concomitant expression and complex interplay between proteins with pro-tumorigenic and anti-tumor properties may be responsible for a benign phenotype of SEGA.

154

M.E. Tyburczy and B. Kaminska

that NeuroD-deficient mice fail to develop a granule Down-Regulation of Identified Genes in SEGA May Be Associated with Neurologic cell layer within the dentate gyrus, and exhibited spontaneous seizures associated with electrophysiological Dysfunctions Occurring in TSC Patients Among genes down-regulated in SEGA there are several genes encoding proteins typical for oligodendrocytes, involved in myelin formation and stability. These are: MBP (myelin basic protein), and MOBP (myelin-associated oligodendrocyte basic protein), transferrin which participates in differentiation of oligodendrocytes, and ermin – a marker of myelinating oligodendrocytes. Lower expression of these genes may lead to diminished myelination observed in TSC patients. It is assumed that neurologic dysfunctions characteristic for TSC patients are associated with the disturbance of white matter maturation (Arulrajah et al., 2009). Next, we identified genes, which down-regulation in SEGA may be involved in the occurrence of epilepsy, mental retardation, and autism in patients with TSC. mTOR signaling pathway was shown to be associated with epileptogenesis. It was observed that early treatment with rapamycin prevented the development of epilepsy in mice with Tsc1 knockout in astrocytes, and late treatment suppressed seizures in these mice (Zeng et al., 2008). Furthermore, the correlation between occurrence of epilepsy and mutations in genes encoding γ-aminobutyric acid receptor subunits – GABAA , which is a main mediator of inhibitory synaptic transmission in the central nervous system. Moreover, the mechanism of action of some antiepileptic drugs (benzodiazepines and barbiturates) is to increase GABAA receptor currents. Mutations causing loss of the receptor functions were identified in patients with epilepsy in GABRA1 (γ-aminobutyric acid (GABA) A receptor, α1), which down-regulation was detected in SEGA (Macdonald et al., 2010). RELN, another gene potentially engaged in epileptogenesis, encodes matrix metalloproteinase reelin that controls migration and laminar arrangement of neurons in the developing brain, and modulates synaptic plasticity in the mature brain. Decreased expression of reelin was detected in the hippocampus of some patients with epilepsy, together with granule cell dispersion (Forster et al., 2010). Next identified gene involved in regulation of neurons functioning is NEUROD1, encoding a transcription factor that induces terminal maturation of neurons. It was shown

evidence of seizure activity in the hippocampus and cortex (Liu et al., 2000). The latest data indicate that autism may develop due to enhanced mTORC1 activity leading to increased protein synthesis at synapses and their malfunctioning, as well as to disturbances in the balance of excitatory and inhibitory synaptic connectivity. Among genes down-regulated in SEGA there is SHANK3 (SH3 and multiple ankyrin repeat domains 3), encoding a scaffolding protein of the postsynaptic density that regulates the size and shape of dendritic spines. Inactivating mutations in this gene have been observed in patients with autism (Bourgeron, 2009). Another differentially expressed gene – CADPS2 (Ca2+ -dependent activator protein for secretion 2), encodes a secretory vesicleassociated protein involved in the release of BDNF (brain-derived neurotrophic factor) and neurotrophin 3 that are essential for the development of cerebellum. It was shown that Cadps2-knockout mice revealed impairments in cerebellar development and functions (such as disturbances of synapses formation), and autistic-like cellular and behavioral phenotypes. Moreover, mutations in CADPS2 were detected in autistic patients, probably leading to reduction of synaptic BDNF release (Sadakata et al., 2007). The expression of CNDP1, which encodes carnosine dipeptidase 1 (carnosinase) degrading carnosine into L-histidine and β-alanine, was found to be down-regulated in SEGA. It was shown that carnosinase deficiency was associated with the occurrence of mental retardation, seizures, hearing loss, and progressive childhood dementia (Cohen et al., 1985).

SEGA Exhibit a Mixed-Lineage Phenotype with Restricted Ability to Differentiate into Glial Cells or Neurons SEGAs are classified as astrocytomas, although they are formed by at least three cell populations, namely astrocytes, dysmorphic neurons, and giant cells of unknown origin, which are 5–10 times larger than normal astrocytes. First ultrastructural studies on giant cells suggested their astrocytic origin due to the

18 Subependymal Giant Cell Astrocytoma: Gene Expression Profiling

observed presence of glial filaments, glycogen, and formation of hemidesmosomes with pia and vascular basement membranes. Later studies showed that subpopulations of giant cells expressed GFAP (glial fibrillary acidic protein) – a specific marker of astrocytes. On the other hand, the presence of numerous microtubules and rough endoplasmic reticulum was detected in giant cells, which indicated their neuronal features. Moreover, giant cells also express proteins typical for immature neurons and neuroepithelial precursor cells, such as nestin, MAP2C (microtubule-associated protein 2C), and N-methyl-D-aspartate (NMDA) 2D receptor subunit. Therefore, the cells were classified into neuron-like and indeterminate giant cells (Jozwiak and Jozwiak, 2007). The immunohistochemical staining of 9 SEGAs revealed expression of GFAP and NSE (neuron-specific enolase) in all tumors, and neurofilaments and synaptophysin in 8 specimens. Based on these findings, the authors postulated a mixed, glioneuronal nature of SEGAs (Buccoliero et al., 2009). Cellular markers detected in giant cells suggest their inability to terminally differentiate before migration into cortex. This hypothesis is supported by the data showing expression of doublecortin (DCX) in giant cells. DCX participates in migration of neurons during cortical development. DCX regulates the fetal migration of neuroblasts from the proliferative ventricular zone near the pial surface and is transiently expressed in proliferating progenitor cells and newly generated neuroblasts. When the cells begin to express mature neuronal markers, DCX expression decreases suddenly. The expression of DCX detected in giant cells may suggest their restricted differentiation (Jozwiak and Jozwiak, 2007). Differentiation of neural progenitor cells into giant cells was demonstrated in a murine model. It was observed that Tsc2 null neuroepithelial progenitor (NEP) cells grew without growth factors and expressed high levels of GFAP, contrary to wild type NEP cells. Tsc2 null NEP cells differentiated into cells morphologically resembling human giant cells, while wild type NEP cells exhibited differentiation into astrocytes or neurons. Moreover, Tsc2 null NEP cells expressed GFAP, a neuronal marker III, β-tubulin, and nestin at the same time (Onda et al., 2002). We observed significantly increased expression of MBP in SEGA cells compared to normal human astrocytes, together with less significant up-regulation of another oligodendrocytic marker – ERMIN, as well

155

as genes encoding proteins typical for immature neurons – NEUROD1 and NPTX1 (neuronal pentraxin 1). Above-mentioned data allow to make a hypothesis that giant cells constitute a heterogeneous line of cells derived from neuroepithelial progenitor cells, and having limited ability to differentiate into glial cells or neurons.

mTOR Kinase Regulates Expression of ANXA1, GPNMB, S100A11, LTF, RND3, SFRP4, and NPTX1 in SEGA Cells Tuberous sclerosis complex is a natural model to study the TSC-mTOR signaling pathway due to the occurrence of mutations in TSC1 or TSC2, and consequentially increased mTORC1 activity. We have successfully derived primary cultures of SEGA cells and treated them with rapamycin which completely inhibited kinase activity of mTORC1 to determine if newly identified genes are truly mTORdependent. We observed significant down-regulation of genes involved in tumorigenesis (ANXA1, GPNMB, S100A11, LTF, RND3, and SFRP4), and increase of NPTX1 expression, which was down-regulated in SEGA compared with control brain. It was demonstrated that mTOR controls morphology of dendritic spines and protein synthesis at synapses, participating in this way in regulating synaptic plasticity, learning, and memory formation. NPTX1, which expression is regulated by mTOR, encodes neuronal pentraxin I that controls synapses formation and rearrangement (Kirkpatrick et al., 2000). It was shown that Nptx1/2 knockout mice exhibited a delay in functional maturation of glutamatergic synapses in retinal ganglion cells, as well as reduction in AMPAR-mediated transmission at developing visual system synapses (Bjartmar et al., 2006; Koch and Ullian, 2010). Down-regulation of Nptx1 together with upregulation of Anxa1 and Gpnmb were also observed in Tsc2 null NEP cells, which confirms that mTOR regulates expression of these genes at the transcriptional level (Onda et al., 2002). In conclusion, we identified many genes differentially expressed in SEGA compared with normal human brain. Several identified genes (ANXA1,

156

GPNMB, S100A11, LTF, RND3, SFRP4, and NPTX1) are regulated by mTOR at the transcriptional level in SEGA cells, showing a link between their expression pattern and pathophysiology of SEGA. Genes up-regulated in SEGA are involved in tumorigenesis, and genes down-regulated in SEGA may be associated with a failure of SEGA giant cell differentiation and the occurrence of neurologic disorders in TSC patients.

References Arulrajah S, Ertan G, Jordan L, Tekes A, Khaykin E, Izbudak I, Huisman TA (2009) Magnetic resonance imaging and diffusion-weighted imaging of normal-appearing white matter in children and young adults with tuberous sclerosis complex. Neuroradiology 51:781–786 Bjartmar L, Huberman AD, Ullian EM, Renteria RC, Liu X, Xu W, Prezioso J, Susman MW, Stellwagen D, Stokes CC, Cho R, Worley P, Malenka RC, Ball S, Peachey NS, Copenhagen D, Chapman B, Nakamoto M, Barres BA, Perin MS (2006) Neuronal pentraxins mediate synaptic refinement in the developing visual system. J Neurosci 26:6269–6281 Bourgeron T (2009) A synaptic trek to autism. Curr Opin Neurobiol 19:231–234 Buccoliero AM, Franchi A, Castiglione F, Gheri CF, Mussa F, Giordano F, Genitori L, Taddei GL (2009) Subependymal giant cell astrocytoma (SEGA): Is it an astrocytoma? Morphological, immunohistochemical and ultrastructural study. Neuropathology 29:25–30 Carmon KS, Loose DS (2008) Secreted frizzled-related protein 4 regulates two Wnt7a signaling pathways and inhibits proliferation in endometrial cancer cells. Mol Cancer Res 6:1017–1028 Cohen M, Hartlage PL, Krawiecki N, Roesel RA, Carter AL, Hommes FA (1985) Serum carnosinase deficiency: a nondisabling phenotype?. J Ment Defic Res 29(Pt 4):383–389 Curatolo P, Napolioni V, Moavero R (2010) Autism spectrum disorders in tuberous sclerosis: pathogenetic pathways and implications for treatment. J Child Neurol 25: 873–880 de Graauw M, van Miltenburg MH, Schmidt MK, Pont C, Lalai R, Kartopawiro J, Pardali E, Le Devedec SE, Smit VT, van der Wal A, Van’t Veer LJ, Cleton-Jansen AM, ten Dijke P, van de Water B (2010) Annexin A1 regulates TGF-beta signaling and promotes metastasis formation of basal-like breast cancer cells. Proc Natl Acad Sci USA 107:6340–6345 Forster E, Bock HH, Herz J, Chai X, Frotscher M, Zhao S (2010) Emerging topics in Reelin function. Eur J Neurosci 31: 1511–1518 He H, Li J, Weng S, Li M, Yu Y (2009) S100A11: diverse function and pathology corresponding to different target proteins. Cell Biochem Biophys 55:117–126 Huang D, Yu B, Deng Y, Sheng W, Peng Z, Qin W, Du X (2010) SFRP4 was overexpressed in colorectal carcinoma. J Cancer Res Clin Oncol 136:395–401

M.E. Tyburczy and B. Kaminska Jozwiak J, Jozwiak S (2007) Giant cells: contradiction to two-hit model of tuber formation?. Cell Mol Neurobiol 27:251–261 Jozwiak J, Jozwiak S, Wlodarski P (2008) Possible mechanisms of disease development in tuberous sclerosis. Lancet Oncol 9:73–79 Kirkpatrick LL, Matzuk MM, Dodds DC, Perin MS (2000) Biochemical interactions of the neuronal pentraxins. Neuronal pentraxin (NP) receptor binds to taipoxin and taipoxin-associated calcium-binding protein 49 via NP1 and NP2. J Biol Chem 275:17786–17792 Klein RM, Aplin AE (2009) Rnd3 regulation of the actin cytoskeleton promotes melanoma migration and invasive outgrowth in three dimensions. Cancer Res 69:2224–2233 Koch SM, Ullian EM (2010) Neuronal pentraxins mediate silent synapse conversion in the developing visual system. J Neurosci 30:5404–5414 Kuan CT, Wakiya K, Dowell JM, Herndon JE 2nd, Reardon DA, Graner MW, Riggins GJ, Wikstrand CJ, Bigner DD (2006) Glycoprotein nonmetastatic melanoma protein B, a potential molecular therapeutic target in patients with glioblastoma multiforme. Clin Cancer Res 12:1970–1982 Kwiatkowski DJ (2003) Tuberous sclerosis: from tubers to mTOR. Ann Hum Genet 67:87–96 Lam C, Bouffet E, Tabori U, Mabbott D, Taylor M, Bartels U (2010) Rapamycin (sirolimus) in tuberous sclerosis associated pediatric central nervous system tumors. Pediatr Blood Cancer 54:476–479 Laplante M, Sabatini DM (2009) mTOR signaling at a glance. J Cell Sci 122:3589–3594 Lim LH, Pervaiz S (2007) Annexin 1: the new face of an old molecule. FASEB J 21:968–975 Liu M, Pleasure SJ, Collins AE, Noebels JL, Naya FJ, Tsai MJ, Lowenstein DH (2000) Loss of BETA2/NeuroD leads to malformation of the dentate gyrus and epilepsy. Proc Natl Acad Sci USA 97:865–870 Macdonald RL, Kang JQ, Gallagher MJ (2010) Mutations in GABAA receptor subunits associated with genetic epilepsies. J Physiol 588:1861–1869 Onda H, Crino PB, Zhang H, Murphey RD, Rastelli L, Gould Rothberg BE, Kwiatkowski DJ (2002) Tsc2 null murine neuroepithelial cells are a model for human tuber giant cells, and show activation of an mTOR pathway. Mol Cell Neurosci 21:561–574 Orlova KA, Crino PB (2010) The tuberous sclerosis complex. Ann N Y Acad Sci 1184:87–105 Poch E, Minambres R, Mocholi E, Ivorra C, Perez-Arago A, Guerri C, Perez-Roger I, Guasch RM (2007) RhoE interferes with Rb inactivation and regulates the proliferation and survival of the U87 human glioblastoma cell line. Exp Cell Res 313:719–731 Rose AA, Annis MG, Dong Z, Pepin F, Hallett M, Park M, Siegel PM (2010) ADAM10 releases a soluble form of the GPNMB/Osteoactivin extracellular domain with angiogenic properties. PLoS One 5:e12093 Sadakata T, Washida M, Iwayama Y, Shoji S, Sato Y, Ohkura T, Katoh-Semba R, Nakajima M, Sekine Y, Tanaka M, Nakamura K, Iwata Y, Tsuchiya KJ, Mori N, DeteraWadleigh SD, Ichikawa H, Itohara S, Yoshikawa T, Furuichi T (2007) Autistic-like phenotypes in Cadps2-knockout mice and aberrant CADPS2 splicing in autistic patients. J Clin Invest 117:931–943

18 Subependymal Giant Cell Astrocytoma: Gene Expression Profiling Shaheduzzaman S, Vishwanath A, Furusato B, Cullen J, Chen Y, Banez L, Nau M, Ravindranath L, Kim KH, Mohammed A, Ehrich M, Srikantan V, Sesterhenn IA, McLeod D, Vahey M, Petrovics G, Dobi A, Srivastava S (2007) Silencing of Lactotransferrin expression by methylation in prostate cancer progression. Cancer Biol Ther 6:1088–1095 Tyburczy ME, Kotulska K, Pokarowski P, Mieczkowski J, Kucharska J, Grajkowska W, Roszkowski M, Jozwiak S, Kaminska B (2010) Novel proteins regulated by mTOR in subependymal giant cell astrocytomas of patients with tuberous sclerosis complex and new therapeutic implications. Am J Pathol 176:1878–1890 Wang LP, Bi J, Yao C, Xu XD, Li XX, Wang SM, Li ZL, Zhang DY, Wang M, Chang GQ (2010) Annexin A1 expression and its prognostic significance in human breast cancer. Neoplasma 57:253–259 Weterman MA, Ajubi N, van Dinter IM, Degen WG, van Muijen GN, Ruitter DJ, Bloemers HP (1995) nmb, a novel gene, is expressed in low-metastatic human melanoma cell lines and xenografts. Int J Cancer 60:73–81

157

Zeng LH, Xu L, Gutmann DH, Wong M (2008) Rapamycin prevents epilepsy in a mouse model of tuberous sclerosis complex. Ann Neurol 63:444–453 Zhou Y, Zeng Z, Zhang W, Xiong W, Wu M, Tan Y, Yi W, Xiao L, Li X, Huang C, Cao L, Tang K, Shen S, Li G (2008) Lactotransferrin: a candidate tumor suppressorDeficient expression in human nasopharyngeal carcinoma and inhibition of NPC cell proliferation by modulating the mitogen-activated protein kinase pathway. Int J Cancer 123:2065–2072 Zinovyeva MV, Monastyrskaya GS, Kopantzev EP, Vinogradova TV, Kostina MB, Sass AV, Filyukova OB, Uspenskaya NY, Sukhikh GT, Sverdlov ED (2010) Identification of some human genes oppositely regulated during esophageal squamous cell carcinoma formation and human embryonic esophagus development. Dis Esophagus 23:260–270

Part II

Astrocytomas: Therapy

Chapter 19

Time-Resolved Laser Induced Fluorescence Spectroscopy (TRLIFS): A Tool for Intra-operative Diagnosis of Brain Tumors and Maximizing Extent of Surgical Resection Pramod Butte and Adam N. Mamelak

Abstract Time-resolved laser induced fluorescence spectroscopy (TR-LIFS) can be used to differentiate various tissue types in vivo, based on the intrinsic fluorescent time-decay properties of these tissues. We have developed TR-LIFS as an intra-operative tool for the delineation of gliomas from normal brain, and established an algorithm for real-time differentiation of normal cortex, normal white matter, low grade gliomas, and high grade glioma. Forty two patients who were undergoing glioma (WHO grade IIV) surgery were enrolled in the study. A TR-LIFS prototype apparatus (gated detection, fast digitizer) was used to induce in vivo fluorescence using a pulsed N2 laser (337 nm excitation, 0.7 ns pulse width) and to record the time-resolved spectrum (360–550 nm range, 10 nm interval). The sites of TR-LIFS measurement were biopsied, and histology confirmed by conventional pathology (H&E staining). Comparison of the parameters derived from the TR-LIFS data, including intensity values and time-resolved intensity decay features (average fluorescence lifetime and Laguerre coefficients values), were used for to define the most common TR-LIFS “signature” for each tissue type. Seventy one biopsy samples of tumor and normal brain were analyzed. A linear discriminant algorithm classified low-grade gliomas with 100% sensitivity and 98% specificity. High-grade glioma demonstrated a high degree of heterogeneity reducing the discrimination accuracy of these tumors to 47% sensitivity and 94% specificity. These results indicate the potential

P. Butte () Department of Neurosurgery, Cedars-Sinai Medical Center, Los Angeles, CA 90048, USA e-mail: [email protected]

of TR-LIFS for real-time, non-invasive intra-operative diagnosis of brain tumors. Validation studies are underway. If these validation studies confirm the findings, TR-LIFS is likely to become a rapid, safe, and easy to use means to maximize extent of tumor resection while minimizing damage to surrounding normal tissues. Keywords Fluorescence spectroscopy Glioblastoma · TR-LIFS · Fluorophores Astrocytoma · Discriminant function analysis

· ·

Background Current 18 month survival rate of patients with glioblastoma (WHO Grade IV glioma) ranges from 15 to 34%, making glioma one of the most lethal tumors (Brandes et al., 2008). There are multiple options available for treatment of gliomas. Surgery is most commonly employed treatment option for glioma, followed by chemotherapy and radiation therapy (Ducray et al., 2010; Robins et al., 2009). The long term survival for patients with gliomas often depends on the extent of surgical resection. In fact, extent of tumor resection is a single most important factor in determining survival (Berger, 1994; Byar et al., 1983; Sanai and Berger, 2008). However, surgical resection of gliomas poses a challenge due to their tendency to infiltrate the surrounding normal brain. Overly aggressive resections can result in neurological injury, while less aggressive resections may leave significant residual disease. An intra-operative tool that provides real time pathological diagnosis of the tissue prior to its removal would be the ideal means to determine which

M.A. Hayat (ed.), Tumors of the Central Nervous System, Volume 5, DOI 10.1007/978-94-007-2019-0_19, © Springer Science+Business Media B.V. 2012

161

162

tissue could be safely removed and which tissue represented normal brain that should be preserved. This is particularly true at the edges of the tumor resection, where tumor infiltration into normal brain can be hard to determine with direct vision or MRI. The actual pathological diagnosis can only be provided by biopsy and frozen section which is limited in capacity due to time constraints and cost. Thus, in order to enhance the ability of surgeon to achieve a near completely excision of brain tumors without sacrificing safety, newer technologies are being investigated to accurately distinguish between the tumor and normal brain and guide tumor resection in near-real time.

Fluorescence Spectroscopy Laser-Induced Fluorescence Spectroscopy (LIFS) represents a promising new adjunctive technique for in vivo tissue diagnosis. Fluorescence is the emission of light from any substance which is electronically excited in singlet states when an electron in the excited orbital is paired to the second electron in the ground state orbital. The emitted light can be detected and analyzed to provide valuable information regarding the compound it was emitted from. This phenomenon of photon release is influenced by many factors such as solvents, pH, molecular orientation, presence of a second fluorophore, optical properties etc. LIFS has advantages such as wavelength tunability, narrow bandwidth excitation and short pulse excitation. Because of these features LIFS can selectively excite fluorophores and is widely used in many industrial and scientific settings. Fluorescence is broadly classified into two types of measurements: 1) static (steady-state) and; 2) dynamic (time-resolved). Steady state fluorescence spectroscopy is performed by excitation of the fluorescent substance with a continuous beam of light and recording the intensity of the emission spectrum. Steady state fluorescence measurement provides information regarding intensity (e.g. concentration) and spectral distribution (e.g. type of fluorophore, number of emitting components). Steady state spectroscopy is relatively simple and does not require complex

P. Butte and A.N. Mamelak

and expensive instrumentation, but the spectral information from a complex system like biological tissues which may have various fluorophores with overlapping emission spectra can modify the final spectra thus making it difficult to interpret. Time-resolved (dynamic) measurement, is used for measuring intensity decay. The sample is illuminated by a pulse of light and the intensity decay is recorded with a high speed detection system on a subnanosecond scale. Time-resolved fluorescence measurements provide additional information such as lifetime (e.g. multiple contributions by different fluorescent components), and anisotropy (e.g. reorientation of molecules). The measured decay is indicative of relative concentrations and decay profiles of constituent fluorophores. It has been demonstrated that laser-induced fluorescence spectroscopy of endogenous fluorophores (autofluorescence) is a useful tool for characterization of biological tissues and offers potential for in vivo diagnosis of diseased tissues and the optimization of therapeutic interventions. Earlier work has demonstrated that steady-state LIFS of endogenous fluorophores (auto-fluorescence) has potential for diagnosis of neoplasms including brain cancer (Bottiroli et al., 1995, 1998; Chung et al., 1997; Croce et al., 2003; Poon et al., 1992; Wagnieres et al., 1998). These include studies of glioblastoma (Butte et al., 2010; Croce et al., 2003; Lin et al., 2001; Marcu et al., 2004), astrocytoma (Lin et al., 2001), oligodendroglioma (Lin et al., 2001), and metastatic carcinoma (Lin et al., 2001). Several groups have reported results on the utility of time-resolved LIFS (TR-LIFS) for tissue diagnosis (Andersson-Engels et al., 1990; Cubeddu et al., 1995; Elson et al., 2004; Marcu et al., 2004; Siegel et al., 2003; Sud et al., 2006; Svensson et al., 2007; Wagnieres et al., 1998). We reported the TRLIFS of several types of brain tumors measured both ex-vivo in freshly excised tissue specimens (Butte et al., 2005; Marcu et al., 2004; Yong et al., 2006) and in vivo in patients (Butte et al., 2010). These results pointed to the potential of TR-LIFS to distinguish brain tumors from cerebral cortex and white matter. We subsequently wished to develop an algorithmic means to rapidly differentiate these tissue subtypes during surgery, as a practical tool to guide glioma resection.

19 Time-Resolved Laser Induced Fluorescence Spectroscopy (TRLIFS)

Materials and Methods Clinical Methods The study was carried out with approval of the CedarsSinai Medical Center Institutional Review Board, with informed consent obtained from each patient. Forty two patients diagnosed with glioma were recruited for the study. The patients underwent craniotomy for surgical removal of brain tumor based on clinical indications. During the surgery, as the brain surface and tumor were removed, the neurosurgeon placed a TR-LIFS fiber optic probe (see below for details) on areas of interest (Fig. 19.1) and the brain tissue was spectroscopically investigated. Areas with distinct pathologic features were selected based on the gross visual evaluation by the neurosurgeon. The goal of the measurements was to record from multiple sites in both the tumor and surrounding normal tissue in each patient. To establish fluorescent signals from “normal” brain as controls, regions of the exposed brain furthest from the visible tumor and apparently tumor-free based on the MRI utilized for intra-operative neuronavigation were recorded as well. No biopsy was obtained from the normal brain tissues due to ethical limitations. Areas identified as tumors based on conventional diagnostic methods (e.g. preoperative MRI) and surgeon experience were TR-LIFS interrogated

Fig. 19.1 Schematic of prototype the Time-Resolved Laser Induced Fluorescence Spectroscopy (TR-LIFS) apparatus. The probe is held over normal cortex to record the Time-resolved fluorescence

163

during the surgical resection. A small biopsy was performed at each spectroscopically investigated spot, except at the areas not considered suitable for physical biopsy due to risks posed to patient. Tumor samples with relatively homogenous morphology were categorized as solid tumors, whereas infiltrated normal tissues at the boundary between the tumor and normal brain tissue was categorized as margins.

Instrumentation Experiments were conducted with an instrumental setup, which allowed for spectrally-resolved fluorescence lifetime measurements (Fig. 19.1). A detailed account of this apparatus and its performance has been previously reported (Fang et al., 2004) and will be briefly reviewed here. Delivery Catheter Light delivery and collection were implemented with a custom made bifurcated sterilizable probe. It had a central excitation fiber of 600 μm core diameter, surrounded by a collection ring of twelve 200 μm core diameter fibers. The probe was flexible throughout its entire length (3 m) except of a 7 cm distal part consisting of a rigid stainless steel tube. After tissue

164

excitation, the emitted fluorescence light was collected and directed into the entrance slit of the spectrometer via the collection channel of the probe. The signal was then detected, amplified, and finally digitized at 8 bits resolution by a digital oscilloscope. The overall time resolution of the systems was approximately 300 ps. Based on the optical properties of human brain and tumor tissue, we have estimated a penetration depth between ∼250–400 μm for astrocytoma and normal cortex respectively at a wavelength of 337 nm.

Collection of Fluorescent Data – Technical Details The fiber optic probe was positioned 3 mm above the exposed brain tissue specimen with the help of a spacer to optimize the collection efficiency of the probe as previously reported (Papaioannou et al., 2004), the spacer also steadied the probe over the tissue, thus avoiding artifacts in the fluorescence emission due to pulsation of brain. Time-resolved emission of each sample was recorded in the 360–550 nm spectral range and scanned at 10 nm intervals. The energy output of the laser (at the tip of the fiber) for sample excitation was adjusted to 3.0 μJ/pulse. The area illuminated by the probe was 2.654 mm2 , thus the total fluence per pulse received by the tissue is 1.39 μJ/mm2 well within safety limits.

P. Butte and A.N. Mamelak

TR-LIFS Data Analysis In the context of TR-LIFS, the intrinsic fluorescence impulse response functions (IRF), h(n), describes the real dynamics of the fluorescence decay. The IRF were recovered by numerical deconvolution of the measured input laser pulse from the measured fluorescence response transients. The Laguerre expansion technique (Jo et al., 2004a,b) was used for deconvolution due to several technical considerations (Butte et al., 2010). This method allows a direct recovery of the intrinsic properties of a dynamic system from the experimental input-output data. The technique uses the orthonormal Laguerre functions to expand the IRF and to estimate the Laguerre expansion coefficients (LEC). The fluorescence decay profile describes the biochemical and morphological characteristics of the tissue. Each Laguerre coefficient describes various dynamics of a complex fluorescence decay curve, i.e. lower order functions describe slower decay characteristics and higher order Laguerre functions describe the faster decay characteristics. The coefficients of the Laguerre function describe the relative contribution of each Laguerre function to the observed fluorescence decay. In order to characterize and model complete temporal dynamics, all the Laguerre coefficients were computed and used as parameters along with the average lifetime which indicates a single value on a complex fluorescence decay curve.

Histopathological Analysis Parameters Selection Each biopsy sample was fixed with 10% buffered formalin and stained with Hemotoxylin and Eosin (H&E). Each specimen was independently interpreted by a neuropathologist with no knowledge of the spectroscopic data. Pathology was the correlated with fluorescence spectroscopy measurements for subsequent development of the classification algorithm. For the purpose of spectroscopic classification gliomas were grouped as low grade glioma (LGG-WHO Grade I & II) and high grade glioma (HGG-WHO Grade III & IV). LGG were further subclassified into distinct pathological entities including oligodendroglioma, oligodendroastrocytoma, and diffuse astrocytoma while high grade gliomas were subcalssified into anaplastic astrocytoma, anaplastic oligodendroglioma, anaplastic oligoastrocytoma and glioblastoma multiforme (Louis et al., 2007).

Once the fluorescence IRF’s were estimated for each emission wavelength, the steady-state spectrum (Iλ ), was computed by integrating each intensity decay curve as a function of time. In order to derive parameters from the intensity values, normalized fluorescence spectra was obtained by dividing the discrete intensity values with the intensity value at the peak emission. It has been shown that blood absorption affects all the wavelengths equally and using intensity ratios negated the requirement for spectral correction (AnderssonEngels et al., 1990). Thus, instead of using normalized intensity values, we used the intensity ratios, which only describe the relative intensity between various spectral bands. Further, to characterize the temporal dynamics of the fluorescence decay, two sets of parameters were used: (1) the average lifetime (τλ ) computed

19 Time-Resolved Laser Induced Fluorescence Spectroscopy (TRLIFS)

as the interpolated time at which the IRF decays to 1/eth of its maximum value; and (2) the normalized value of the corresponding LECs. Thus, a complete description of fluorescence from each sample as a function of emission wavelength, Eλ , was given by the variation of a set of spectroscopic parameters at distinct wavelengths (emission intensity – Iλ, average lifetime of fluorescence emission – τ λ, and Laguerre coefficients LECλ ).

Data Reduction and Statistical Analysis To identify a set of spectroscopic parameters that best discriminate between various tissue types, a univariate statistical analysis (one-way ANOVA) was used to compare the spectroscopic parameters (Iλ , τ λ , and LECs) at every Eλ for each type of tissue defined by histology. A p-value of <0.05 was assumed to indicate statistical significance. The variables were tested for statistical significance amongst various combinations of two tissue types (Fig. 19.2). The significant variables are grouped in six separate sets based on their ability to discriminate between these two types. Second, as these parameters were used in linear discriminant analysis, Lilliefor’s test was performed to ensure the parameters used in the classification were normally distributed. Although, discarding the parameters based on their non-normality reduces the number of parameters and adds a confounding factor, using nonnormal parameters with linear discriminant functions analysis can lead to misclassification (Baron, 1991; Bello, 1992).

Classification In order to classify and predict the fluorescence signal acquired we adopted a classification algorithm based on the binary separation of various tissue types using linear discriminant function analysis (DFA). This approach is a novel application designed to identify a single sample intra-operatively in near real-time (Fig. 19.2). Training phase: Fluorescence emission data was divided in four groups: normal cortex (NC), normal white matter (NWM), low grade glioma (LGG) and high grade gioma (HGG). Since no specific parameter was found useful in discriminating all tissues types,

165

the discriminant model was divided in six models of binary sets, each parsing and classifying all the data set in two tissue types either (normal NC vs. LGG or, NC vs. HGG or, LGG vs. HGG, etc.). For every binary set, the parameters were selected independently. ANOVA was performed to choose the parameters, which were statistically significant between the two types of tissues. In order to test for normal distribution, Lilliefor’s test was performed and the parameters with nonnormal distribution were discarded. Once the parameters for each set of classifying models were determined, DFA coefficients were calculated and the value of the centroid noted. Test Phase: A step-wise linear discriminant analysis was used to classify tumor tissue (NC, NW, LGG, and HGG) into one of two-group models (e.g. NC versus NWM) based on a set of features (parameters) that describe the tissue and that were derived in the training phase. In general, this approach determined the combination of predictor variables that account for most of the differences observed between the groups. For instance, the larger the difference between the means of two tissue types relative to the variability within each tissue group, the better the discrimination between the two groups. Assuming that the groups are linearly separable, that is, the groups can be separated by a linear (or higher dimensional) combination of features (parameters) that describe the tissues being compared. The classification criterion is to assign a tissue sample to the group with the highest conditional probability (e.g., Bayes’ Rule); note that this rule also minimizes total error of classification. In general, we are interested in the probability P(Gi /x) that a tissue type belongs to group i, given a set of measurements x. By applying Bayes’ Theorem we can describe the posterior distribution as follows: P(Gi /x) = P(x/Gi )P(Gi ), where P(x/Gi ) describes the probability of getting a particular set of measurements x given that the tissue comes from group i. Prior probability P(Gi ) is the probability about the group i known without making any measurement. Here, we assumed that the prior probability is equal for all. The posterior distribution was used to determine group membership for all six two-group tissues being compared. The total number of posterior probabilities obtained from these models for each tissue type is n–1, where n is the number of tissue types being evaluated. The posterior probabilities were then averaged and compared to determine the group to which the sample belongs. In order to

166

P. Butte and A.N. Mamelak

TRAINING PHASE Normal White Matter (NWM)

Normal Cortex (NC)

Low Grade Glioma (LGG)

High Grade Glioma (HGG)

Tissue Types

NC vs NWM

NC vs LGG

NC vs HGG

NWM vs LGG

NWM vs HGG

LGG vs HGG

Iλ , τf , LECs

Iλ , τf , LECs

Iλ , τf , LECs

Iλ , τf , LECs

Iλ , τf , LECs

Iλ , τf , LECs

Binary Sets Extracted Spectroscopic Parameters

ANOVA

ANOVA

ANOVA

ANOVA

ANOVA

ANOVA

Determine significant parameters within a set

Normal Distribution

Normal Distribution

Normal Distribution

Normal Distribution

Normal Distribution

Normal Distribution

Lilliefors test for Normality

TEST PHASE

SAVE SIX SETS OF TRANING PARAMETERS Test Data

Cross Validation

Newly Acquired data (x )

NC vs NWM

P(GNC⏐x)

P(GNWM⏐x)

NC vs LGG

P(GNC⏐x)

P(GLGG⏐x)

Training Set

NC vs HGG

P(GNC⏐x)

P(GHGG⏐x)

NWM vs LGG

P(GNWM⏐x)

P(GLGG⏐x)

NWM vs HGG

P(GNWM⏐x)

P(GHGG⏐x)

LGG vs HGG

P(GLGG⏐x)

P(GHGG⏐x)

P(G ⏐x) P(G ⏐x) P(GNC⏐x) P(G ⏐x) Σ NWM Σ LGG Σ HGG Σ n n n n Assign tissue to the group with the maximum Discriminant Function (in each of the six group comparisons)

Linear Discriminant Analysis Posterior Probabilities Estimate Group Membership

Fig. 19.2 Elements of classification algorithm to identify a single sample time-resolved fluorescence spectroscopy data acquired intra-operatively. The posterior probability of the

sample belonging to the groups being analyzed is calculated using linear discriminant analysis. All the posterior probabilities are averaged for each type of tissue to determine the diagnosis

eliminate bias in sensitivity and specificity due to the data driven predictions, a leave-one- out cross validation approach was used to calculate the classification (discriminant) score and thereby predict group membership. In the leave-one-out cross-validation, a single set of spectroscopic data is removed from the study population and used in the validation process. The cutoff value was evaluated in a data-driven way from the remaining n–1 parameters. Thereafter, the resulting cutoff value was applied to the newly acquired data. This tissue was then classified as a true positive, a false positive, a false negative, or a true negative, depending on whether this tissue is classified as belonging or not to the group membership being assessed. This process was repeated for all parameters in the dataset, and the results based on all of the parameters then used to evaluate the sensitivity and specificity corresponding to the cutoff value that was derived in the n–1 parameters.

Results Fluorescence emission data were collected from a total of 186 samples. Pathological analysis of the tissue corresponding to these measurements was available in 71 samples. These include normal cortex (N = 35); normal white matter (N = 12); low grade glioma (N = 7) including mixed oligoastrocytoma and oligodendroglioma (N = 4) and diffuse astrocytoma (N = 3); and high grade glioma (N = 17).

Time-Resolved Fluorescence Characteristics All samples (Fig. 19.3a, b) showed relatively broad emission spectrum with a varying degree of reduced emission at 415 nm which corresponds to the band

19 Time-Resolved Laser Induced Fluorescence Spectroscopy (TRLIFS)

167

Fig. 19.3 Comparison of average fluorescence values mean ± SE of the spectroscopic parameters, emission spectra intensity, average lifetime, average Laguerre coefficients (LEC-0, LEC-1)

across the emission wavelengths for distinct brain tissue types (a) normal cortex, normal white matter and low grade glioma, (b) normal cortex, normal white matter and high grade glioma

of hemoglobin absorption (Beychok et al., 1967). Variability of the signal at 415 nm was due to different amounts of blood interfering with the optical probe during surgery. The τ λ values follows a similar trend for all tissue types; lifetime at blue-shifted wavelengths (∼ 390 nm) was typically longer when compared with red-shifted wavelengths (> 440 nm). LECs as a function of wavelengths provided additional information to fluorescence spectra and lifetime for comparing and classifying tissue samples. The zeroth-order coefficient (LEC-0) and the first order (LEC-1) were used to classify the tissues. LEC-0 closely followed the average lifetime in data, whereas LEC-1 provided information regarding faster dynamics in the fluorescence decay curve and thus was used as an additional parameter for tissue characterization. Normal Cortex (NC): NC fluorescence was characterized by a broad fluorescence emission with two distinct peaks centered at 390 and 440 nm, with the emission slightly higher at 440 nm. The τ λ values at 390 nm (τ390 = 2.12 ± 0.10 ns) were longer than those at 440 nm (τ460 = 1.16 ± 0.08 ns). The LEC-1 values at 390 nm (LEC-1390 = –0.032 ± 0.019) were lower

when compared with LEC-1 values at 440 nm (LEC1440 = 0.095 ± 0.010), a trend that mirrored the τ λ values. Normal White Matter (NWM): The NWM spectra showed a main peak emission shifted to 450 nm when compared to NC. The emission at 390 nm (I390 = 0.63 ± 0.07) was also lower when compared to NC. The average lifetimes at 390 nm (τ390 = 1.933 ± 0.15 ns) and 440 nm (1.193 ± 0.11 ns) were similar to those of NC. The LEG-0 had similar values at 390 nm (LEC0390 = 0.766 ± 0.02) and 440 nm (LEC-0440 = 0.7269 ± 0.018) to those on NC. The LEG-1 at 390 nm (LEC1390 = –0.00885 ± 0.002) also similar to LGG while at 440 nm (LEC-1440 = 0.105 ± 0.020) was similar to NC. Low Grade Glioma (LGG): LGG demonstrated a main peak emission centered at 440–450 nm. A secondary peak was present around 390 nm, however, this was lower (∼30%) when compared with the main peak emission (Fig. 19.3C). The τ λ values appeared slighter longer than those observed in NC, both at 390 nm (τ390 = 1.81 ± 0.35 ns) and 440 nm (τ440 = 1.38 ± 0.31) but this difference was not statistically

168

P. Butte and A.N. Mamelak

significant. The LEC-0 at 390 nm (LEC-0390 = 0.7749 ± 0.34) was similar to 440 nm (LEC-0440 = 0.775 ± 0.33). The LEG-1 at 390 nm (LEC-1390 = –0.0077 ± 0.0045) was higher than the LEC-1 values observed at 440 were lowest in (LEC-1440 = –0.068 ± 0.0041). The LEC-1440 for LGG was the lowest of all observed for all tissue types. Notably, we determine differences within the subsets of LLG. Diffuse astrocytoma was found to have higher emission intensity and longer τ λ at 390 nm than oligodendroglioma. When the fluorescence from oligoastrocytoma and oligodendroglioma were studied separately, the τ λ at 390 nm was shorter (τ390 = 1.57± 0.2 ns) compared with diffuse astrocytoma (τ390 = 2.58 ± 0.5 ns). High Grade Glioma (HGG): HGG demonstrated a main peak emission centered at 450 nm. A secondary peak was present around 390 nm. This was lower (∼ 40%) when compared with the 450 nm emission. Average lifetime observed at 390 nm (τ390 = 1.93 ± 0.18 ns) was longer than at 440 nm (τ440 = 1.138 ± 0.12 ns). The LEG-0 at 390 nm (LEC-0390 = 0.761 ± 0.023) was higher than at 440 nm (LEC-0440 = 0.0394 ± 0.016). The values of the LEG-1 observed at 390 nm and 440 nm were (LEC-1390 = –0.0183 ± 0.34, (LEC-1440 = 0.0875 ± 0.0223) respectively. As observed in LGG tumors, HGG demonstrated variation in the fluorescence emission characteristics based on histopathological sub-classification. It was noted that the fluorescence emission data collected from HGG demonstrated a great degree of variability. Glioblastoma multiforme (GBM) fluorescence emission characteristics were similar to low-grade oligodendroglioma with a shorter τλ at 390 nm (τ390 = 2.145 ± 0.25 ns) compared to anaplastic oligodendroglioma (τ390 = 2.66 ± 0.5 ns) with longer τλ at 390 nm. A single sample of recurrent glioma with necrotic changes was characterized by a single emission peak at 390 nm of wavelength and overall faster average lifetime. (τ390 = 1.11 ns) fig. 4b. In addition, we observed difference in the fluorescence parameters

of HGG core compared with margins of the tumor within the same patient (data not shown). The fluorescence parameters observed for the tumor margins resembled those of LLG.

Statistical Analysis and Classification Table 19.1 depicts the classification results using both spectral intensities ratios and time-resolved parameters, computed based on the algorithm that allows for binary separation of distinct tissue groups using LDA. All LGG were correctly identified, except for one measurement in which NC was classified as LGG. The accuracy dropped when classifying HGG , with a sensitivity of 47% and specificity of 95%. We attribute this lower sensitivity value for HHG to high variability in the fluorescence signature of HGG.

Discussion TR-LIFS can accurately differentiate LGG from normal tissues (100% sensitivity and 98% specificity). These finding are in agreement to those reported previously by our group (Butte et al., 2010). Differentiating between LGG and NWM is of importance when trying to achieve a complete excision at the margins, as the tumor will be surrounded by NWM. Detection of LGG represents a major challenge to current tumor resection. LGG are often visually bland. As such they are far more difficult to differentiate from surrounding normal brain than HGG. Further, these tumors typically infiltrate into surrounding tissues that maintain function, often resulting in surgeons performing only biopsies or limited resections. Several recent studies have demonstrated that the extent of resection of LGG correlates with long term survival (Sanai and Berger, 2008; Smith et al., 2008). Consequently, one of the

Table 19.1 Classification results using TR-LIFS TR-LIFS Histopathology

NC

NWM

LGG

HGG

Sensitivity (%)

Specificity (%)

NC (n = 35) NWM (n = 12) LGG (n = 7) HGG (n = 17)

27 4 0 6

3 8 0 3

1 0 7 0

3 0 0 8

79.41 66.67 100.00 47.06

78.26 90.63 98.44 94.64

19 Time-Resolved Laser Induced Fluorescence Spectroscopy (TRLIFS)

great potential applications of TR-LIFS is the ability to differentiate LGG from normal brain as an adjunctive tool for increasing the extent of resection. The HHG were classified with high specificity (95%) but very low sensitivity (47%). Although, high specificity reduces the risk of resecting normal brain, we attribute this to the high variability in the TR-LIFS signals obtained from various subclasses of HGG. We noted these tumors exhibited highly heterogeneous fluorescence spectroscopic features. Such differences may be attributed to the large variance in the protein expression (Umesh et al., 2009) within the same HGG tumor. Generally, HGG cells are more pleomorphic when compared with LGG. Thus, future studies will need to determine how well heterogeneities in HHG can be distinguished using TR-LIFS derived parameters. We anticipate that a more comprehensive classification based on the biochemical and immunohistochemical features will improve the classification accuracy. The small sample size available for this study most likely also contributed to the low sensitivity values. In order to obtain an unbiased estimation of the classification accuracy, a larger set of data needs to be obtained and the training set and the test set used have to be completely independent. Consequently, a higher number of HHG need to be investigated along with the biochemical heterogeneities within the same tumor for a more comprehensive assessment of the ability of TR-LIFS to distinguish HGG tissue from normal brain tissue. The NC measurements were obtained from areas distal to the tumor site and, where the arachnoid and pia covering the cortex were generally preserved. Arachnoid and pia mater consists predominately of collagen fibers. Collagen has peak fluorescence at 390 nm of wavelength with an average lifetime of ∼3 ns (Marcu et al., 2001). The change in collagen amount between optical probe and the actual brain tissue may have contributed to the variation in the averaged lifetime values. Due to our inability to perform a biopsy on the normal healthy brain, we were unable to confirm this hypothesis. We note that the average lifetime values at 390 nm (1.45 ± 0.4 ns) of LGG in our previous study (Butte et al., 2010) were found shorter than those observed in the current study (1.58 ± 0.2 ns). To understand this difference in average lifetime value at 390 nm from LGG samples, we subdivided the LGG tumors

169

into mixed astrocytoma, oligodendroglioma and diffuse astrocytomas. It was observed that the oligodendral tumors had faster lifetime at 390 nm compared to tumors of astrocytic origin such as diffuse astrocytomas. This is a significant finding as diffuse astrocytomas have a tendency to convert to GBM We hypothesize that this difference may be attributed to the IDH-1 mutation in oligodendral tumors (Hartmann et al., 2009; Yan et al., 2009), which affects the NADP+ dependant isocitrate dehydrogenase leading to an up-regulation of glutamate decarboxylase (GAD) which has peak fluorescence emission at 390 nm of wavelength and average lifetime of 1.8–2.1 ns (Rosato et al., 1989). We are in the process of designing new experiments to confirm presence of both collagen and Glutamate decarboxylase as the fluorophores responsible for this variation. The classification using only the spectral intensity ratios demonstrated a significant drop in the sensitivity and specificity of normal white matter and high grade glioma, whereas, time-resolved fluorescence data without any correction for spectral response yielded superior accuracy. This emphasizes the potential of time resolved fluorescence in the intra-operative setting in which factors which affect the spectral intensity values such as hydration, temperature, and absorption by blood cannot be controlled. The fluorescence decay characteristics are immune to these factors. Limitations of classification method: While we were able to acquire large fluorescence data that allows for the formation of two data sets training and test, we were not able to include all data in the classification stage. We were limited by constrains on the number of biopsies performed on the brain tissue. Only a limited number of spectroscopic data sets can be correlated with the histopathological diagnosis. Consequently, we used a cross-validation method (leave-one-out) to test our algorithm. Although cross-validation method can remove the bias in such analyses, there is a danger in overestimation of the data by selecting the best possible outcome. In addition, we used known pathology to determine the best set of parameters to be used in the classification. Such assumption in the analysis can lead to distorted and predetermined results (Kriegeskorte et al., 2009). In order to avoid this issue we anticipate acquiring more data with histopathological analysis in the future to create a separate training set and test set.

170

In Vivo Measurements, Challenges and Opportunities for Glioma Diagnosis Brain tumors demonstrate broad biochemical, molecular, and metabolic diversity. Consequently, these tumors are likely to have a broad range of spectroscopic signatures. This poses a challenge as conventional histopathology does not describe the heterogeneity in terms of metabolic states. The intrinsic fluorophores potentially being responsible for the distinct fluorescence emission spectrum include NADH/NAD(P)H at 460 nm and glutamate decarboxylase GAD or pyridoxamine-5-phosphate (PMP) at 390 nm. The emission characteristic of such fluorophores most likely is modulated by distinct tumor pathologies and metabolic states. For example we observed that astrocytoma demonstrated longer τ λ at 390 nm when compared with both oligoastrocytoma and oligodendroglioma. It is also important to note that the fluorescence signature of diffuse astrocytoma closely resembled HGG, When compared with oligodendrogial tumors, astrocytomas are more likely to rapidly transform to HGG. Thus a similar fluorescence signature may help in determining the longterm prognosis and the need for more aggressive initial therapies. Similarly, when HGG samples were histopathologically sub-classified in glioblastoma multiforme, anaplastic oligodendroglioma and recurrent glioma with radiation necrosis, TR-LIFS characteristics were distinct for each group Lastly, recurrent glioma with radiation necrosis did not have significant emission intensity at 440–460 nm of wavelength. This is an expected result as amount of viable cells in the necrotic tissue are fewer leading to absence or deficiency of NADH/NAD(P)H. Taken together these results suggest that TR-LIFS may be a useful tool for predicting response to therapy and long-term prognosis independent of histopathology. A close analysis of spectroscopic parameters derived from both spectral and time demonstrated that a set of parameters combining spectral and temporal features are needed in order to accurately discriminate between distinct tissue types. By determining the optimal spectral bands for discrimination of pathologic vs. normal conditions, the TR-LIFS data acquisition can be reduced to a limited number of spectral bands. This will enable development of specialized TR-LIFS systems that collect fluorescence pulse transients at

P. Butte and A.N. Mamelak

a limited number of wavelength bands as recently reported (Sun et al., 2008). In turn this will result in shorter data acquisition time (< 1 s per point measurement). The main goal of developing this technique is not to replace histopathology, but to provide a guide to the neurosurgeon during tumor resection. Acquiring additional data in order to differentiate the subclasses of tumors will allow us to explore the potential of TR-LIFS technique further. The most important advantage of TR-LIFS is the speed at which it can provide the diagnosis. The speed should allow the surgeon to inspect multiple sites during the surgery thus ensuring more complete resection. Currently, the technique is used as a single point spectroscopy, but newer techniques are being developed to obtain a two dimensional image (Sun et al., 2008; 2009). This is anticipated to provide a more robust and practical anatomical data.

Conclusion We have tested a prototype TR-LIFS system in the operating room environment to study the fluorescence emission characteristics of low and high grade glioma tumors and normal brain tissue. Current results show that intra-operative measurement of the intrinsic fluorescence of brain tumors and normal brain tissue has the potential to distinguish LGG from normal tissue with high accuracy and specificity. HGG demonstrated a high degree of heterogeneity, thus reducing the discrimination accuracy of these tumors as a broad category. The classification accuracy of TR-LIFS can be improved by obtaining more samples and using finer classification of brain tumors based on metabolic state and biochemical information. TR-LIFS holds the potential to characterize distinct types of glioma tumors, and thus can be a valuable tool for aiding the surgeon in near complete resection of the brain tumors.

References Andersson-Engels S, Johansson J, Stenram U, Svanberg K, Svanberg S (1990) Time-resolved laser-induced fluorescence spectroscopy for enhanced demarcation of human atherosclerotic plaques. J Photochem Photobiol B 4:363–369

19 Time-Resolved Laser Induced Fluorescence Spectroscopy (TRLIFS) Baron AE (1991) Misclassification among methods used for multiple group discrimination–the effects of distributional properties. Stat Med 10:757–766 Bello AL (1992) Misclassification among methods used for multiple group discrimination–the effects of distributional properties. Stat Med 11:1623–1624 Berger MS (1994) Malignant astrocytomas: surgical aspects. Semin Oncol 21:172–185 Beychok S, Tyuma I, Benesch RE, Benesch R (1967) Optically active absorption bands of hemoglobin and its subunits. J Biol Chem 242:2460–2462 Bottiroli G, Croce AC, Locatelli D, Marchesini R, Pignoli E, Tomatis S, Cuzzoni C, Di Palma S, Dalfante M, Spinelli P (1995) Natural fluorescence of normal and neoplastic human colon: a comprehensive “ex vivo” study. Lasers Surg Med 16:48–60 Bottiroli G, Croce AC, Locatelli D, Nano R, Giombelli E, Messina A, Benericetti E (1998) Brain tissue autofluorescence: an aid for intraoperative delineation of tumor resection margins. Cancer Detect Prev 22:330–339 Brandes AA, Tosoni A, Franceschi E, Reni M, Gatta G, Vecht C (2008) Glioblastoma in adults. Crit Rev Oncol Hematol 67:139–152 Butte PV, Pikul BK, Hever A, Yong WH, Black KL, Marcu L (2005) Diagnosis of meningioma by time-resolved fluorescence spectroscopy. J Biomed Opt 10:064026 Butte PV, Fang Q, Jo JA, Yong WH, Pikul BK, Black KL, Marcu L (2010) Intra-operative delineation of primary brain tumors using time-resolved fluorescence spectroscopy. J Biomed Opt:15(2) Byar DP, Strike TA, Walker MD, Green SB (1983) Prognostic factors for malignant glioma. Oncology of the nervous system, 379–395. Boston: Marinus Nijhoff Chung YG, Schwartz JA, Gardner CM, Sawaya RE, Jacques SL (1997) Diagnostic potential of laser-induced autofluorescence emission in brain tissue. J Korean Med Sci 12:135–142 Croce AC, Fiorani S, Locatelli D, Nano R, Ceroni M, Tancioni F, Giombelli E, Benericetti E, Bottiroli G (2003) Diagnostic potential of autofluorescence for an assisted intraoperative delineation of glioblastoma resection margins. Photochem Photobiol 77:309–318 Cubeddu R, Pifferi A, Taroni P, Valentini G, Canti G (1995) Tumor detection in mice by measurement of fluorescence decay time matrices. Opt Lett 20:2553 Ducray F, Dutertre G, Ricard D, Gontier E, Idbaih A, Massard C (2010) Advances in adults’ gliomas biology, imaging and treatment. Bull Cancer 97:17–36 Elson D, Requejo-Isidro J, Munro I, Reavell F, Siegel J, Suhling K, Tadrous P, Benninger R, Lanigan P, McGinty J, Talbot C, Treanor B, Webb S, Sandison A, Wallace A, Davis D, Lever J, Neil M, Phillips D, Stamp G, French P (2004) Timedomain fluorescence lifetime imaging applied to biological tissue. Photochem Photobiol Sci 3:795–801 Fang Q, Papaioannou T, Jo JA, Vaitha K, Marcu L (2004) Timedomain laser-induced fluorescence spectroscopy apparatus for clinical diagnostics. Rev Sci Instrum 75:151–162 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

171

oligodendroglial differentiation and age: a study of 1,010 diffuse gliomas. Acta Neuropathol 118:469–474 Jo JA, Fang Q, Papaioannou T, Marcu L (2004a) Fast modelfree deconvolution of fluorescence decay for analysis of biological systems. J Biomed Opt 9:743–752 Jo JA, Fang Q, Papaioannou T, Qiao JH, Fishbein MC, Dorafshar A, Reil T, Baker D, Freischlag J, Marcu L (2004b) Novel methods of time-resolved fluorescence data analysis for invivo tissue characterization: application to atherosclerosis. Conf Proc IEEE Eng Med Biol Soc 2:1372–1375 Kriegeskorte N, Simmons WK, Bellgowan PS, Baker CI (2009) Circular analysis in systems neuroscience: the dangers of double dipping. Nat Neurosci 12:535–540 Lin WC, Toms SA, Johnson M, Jansen ED, Mahadevan-Jansen A (2001) In vivo brain tumor demarcation using optical spectroscopy. Photochem Photobiol 73:396–402 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 Marcu L, Fishbein MC, Maarek JM, Grundfest WS (2001) Discrimination of human coronary artery atherosclerotic lipid-rich lesions by time-resolved laser-induced fluorescence spectroscopy. Arterioscler Thromb Vasc Biol 21: 1244–1250 Marcu L, Jo JA, Butte PV, Yong WH, Pikul BK, Black KL, Thompson RC (2004) Fluorescence lifetime spectroscopy of glioblastoma multiforme. Photochem Photobiol 80:98–103 Papaioannou T, Preyer N, Fang Q, Carnohan M, Ross R, Brightwell A, Cottone G, Jones L, Marcu L (2004) Effects of fiber-optic probe design and probe-to-target distance on diffuse reflectance measurements of turbid media: an experimental and computational study at 337 nm. Appl Opt 43:2846–2860 Poon WS, Schomacker KT, Deutsch TF, Martuza RL (1992) Laser-induced fluorescence: experimental intraoperative delineation of tumor resection margins. J Neurosurg 76: 679–686 Robins HI, Lassman AB, Khuntia D (2009) Therapeutic advances in malignant glioma: current status and future prospects. Neuroimag Clin N Am 19:647–656 Rosato N, Mei G, Finazzi-Agro A, Tancini B, BorriVoltattorni C (1989) Time-resolved extrinsic fluorescence of aromatic-L-amino-acid decarboxylase. Biochim Biophys Acta 996:195–198 Sanai N, Berger MS (2008) Glioma extent of resection and its impact on patient outcome. Neurosurgery 62:753–764, discussion 264–756 Siegel J, Elson DS, Webb SE, Lee KC, Vlandas A, Gambaruto GL, Leveque-Fort S, Lever MJ, Tadrous PJ, Stamp GW, Wallace AL, Sandison A, Watson TF, Alvarez F, French PM (2003) Studying biological tissue with fluorescence lifetime imaging: microscopy, endoscopy, and complex decay profiles. Appl Opt 42:2995–3004 Smith JS, Chang EF, Lamborn KR, Chang SM, Prados MD, Cha S, Tihan T, Vandenberg S, McDermott MW, Berger MS (2008) Role of extent of resection in the long-term outcome of low-grade hemispheric gliomas. J Clin Oncol 26:1338–1345 Sud D, Zhong W, Beer DG, Mycek MA (2006) Time-resolved optical imaging provides a molecular snapshot of altered

172 metabolic function in living human cancer cell models. Opt Express 14:4412–4426 Sun Y, Liu R, Elson DS, Hollars CW, Jo JA, Park J, Marcu L (2008) Simultaneous time- and wavelength-resolved fluorescence spectroscopy for near real-time tissue diagnosis. Opt Lett 33:630–632 Sun Y, Phipps J, Elson DS, Stoy H, Tinling S, Meier J, Poirier B, Chuang FS, Farwell DG, Marcu L (2009) Fluorescence lifetime imaging microscopy: in vivo application to diagnosis of oral carcinoma. Opt Lett 34:2081–2083 Svensson T, Andersson-Engels S, Einarsdottir M, Svanberg K (2007) In vivo optical characterization of human prostate tissue using near-infrared time-resolved spectroscopy. J Biomed Opt 12:014022

P. Butte and A.N. Mamelak Umesh S, Tandon A, Santosh V, Anandh B, Sampath S, Chandramouli BA, Sastry Kolluri VR (2009) Clinical and immunohistochemical prognostic factors in adult glioblastoma patients. Clin Neuropathol 28:362–372 Wagnieres GA, Star WM, Wilson BC (1998) In vivo fluorescence spectroscopy and imaging for oncological applications. Photochem Photobiol 68:603–632 Yan H, Bigner DD, Velculescu V, Parsons DW (2009) Mutant metabolic enzymes are at the origin of gliomas. Cancer Res 69:9157–9159 Yong WH, Butte PV, Pikul BK, Jo JA, Fang Q, Papaioannou T, Black K, Marcu L (2006) Distinction of brain tissue, low grade and high grade glioma with time-resolved fluorescence spectroscopy. Front Biosci 11:1255–1263

Chapter 20

Magnetic Resonance-Guided Laser Interstitial Thermal Therapy for Brain Tumors Kevin Beccaria, Michael S. Canney, and Alexandre C. Carpentier

Abstract Lasers were invented in the middle of the 20th century, from the previous theoretical work of Einstein on the process of amplification of stimulated emission, and quickly became used in various devices for industrial, telecommunication and military applications. In the medical field, neurosurgeons were some of the first researchers to investigate the use of lasers for therapy in the 1960s, but were limited by cumbersome designs and the availability of lasers that could only operate in pulsed modes. As improved laser designs were introduced that allowed for continuous mode operation and other wavelengths, new medical techniques were developed and improved. Subsequently, in the early 1970s, Sutton introduced the concept of interstitial hyperthermia, which led to the development of a new technique called laser interstitial thermotherapy (LITT). LITT treatments consist of heating tumors with prolonged and moderate temperature elevations, inducing alteration of cell membranes and enzyme denaturation, leading to the formation of a selective zone of coagulation necrosis in the heated tissue without vaporization. The Nd-YAG and diode laser are the most frequent lasers used for LITT due to their optimum wavelength for absorption and heating in tissue. With the recent development and widespread availability of magnetic resonance imaging (MRI), laser fibers can be inserted stereotactically within the tumor and the treatment can be controlled in real-time. MR imaging allows for monitoring of the temperature elevation

A.C. Carpentier () Department of Neurosurgery and Advanced Surgical Technologies Research Team, Hopital de la Pitie-Salpetriere, Assistance Publique Hopitaux de Paris, Université Paris VI – Pierre & Marie Curie, 75013 Paris, France e-mail: [email protected]

in the treated area and for prediction of the extent of induced necrosis based on the temperature history. Several clinical studies have been performed that show that the technique is now safe and allows for control of brain tumors metastases for which traditional treatments including radiosurgery have failed. Additional technical progress is now being performed on methods to treat tumors with larger diameters and with more complex shapes. Lastly, additional clinical studies are needed in larger patient populations to confirm the findings of the initial clinical trials. Keywords LITT · Laser · Thermal dosimetry · Neurosurgery · Histology · Radiosurgery

Introduction Lasers are now commonly used in various applications in telecommunications, defense, and medicine. In medicine, they are used predominantly in specialized fields such as ophthalmology, dermatology and gynecology. The use of lasers in neurosurgery has been reported since shortly after they were invented in the late 1950s, especially for the treatment of brain and spinal tumors; however, their use has been limited in medicine until the last decade due to substantial technical developments needed to properly control treatments. Furthermore, the introduction of the concept of interstitial hyperthermia by Sutton (1971), and the proposition to use laser as a heat source by Bown (1983) led to renewed interest in lasers in medicine. Since then, laser interstitial thermotherapy (LITT) has been investigated as a potential minimally invasive technique for treating tumors of the central nervous

M.A. Hayat (ed.), Tumors of the Central Nervous System, Volume 5, DOI 10.1007/978-94-007-2019-0_20, © Springer Science+Business Media B.V. 2012

173

174

system. Furthermore, progress in medical imaging modalities, particularly in MRI, has made real-time control of LITT treatments possible. Several clinical studies have now been completed and have shown that LITT with real-time MRI can be used to treat deepseated tumors with few side-effects. This chapter aims at describing the physical and biological principles of LITT as well as summarizing the results of recent LITT procedures on patients suffering from brain metastases.

History of the Laser In the early 20th century, Einstein began theoretical work which led to the development of the laser. In his study entitled “ZurQuantentheorie der Strahlung”, the physicist outlined the possibility to obtain stimulated emission of photons and then provided the theoretical basis for laser physics. Using the process of amplification of stimulated emission, Schawlow and Townes described in 1958 the Maser (Microwaves Amplification by Stimulated Emission of Radiation), an apparatus for producing coherent microwaves. Several years later, Maiman generated electromagnetic waves many orders of magnitude shorter than that of microwaves for the generation of coherent light, and then created the first Laser (Light Amplification by Stimulated Emission of Radiation) with a ruby crystal as the active medium. Maiman’s invention generated immense excitement in the scientific and technical community and laser technology developed rapidly. Parallel to this technical evolution, potential applications in medicine led researchers as Stellar (1968) to experiment on the effects of pulsed ruby lasers on biological tissues, in particular on central nervous system tissues and tumors. These first studies proved that lasers could be used to damage malignant cells and tumors. However, pulsed-wave mode lasers were associated with considerable damage to non-malignant brain tissue, and these early devices had no cutting or haemostatic effects, so that their interest in neurosurgery was short-lived. As new, more powerful continuous mode lasers were developed, and lasers were demonstrated to be efficient at incising tissues cleanly, rapidly, and bloodlessly. During the same time, the first Argon and Neodimium-YAG lasers were developed, the latter of which proved to have better

K. Beccaria et al.

applicability for thermal ablation than CO2 lasers. However, early laser apparatus were cumbersome, indications in neurosurgery were not well defined, and the initial excitement over the use of lasers in medicine slowly faded. At the end of the 1970s, other wavelength lasers were developed, and more compact carbon dioxide lasers with better maneuverability and the ability to induce minimal damage to healthy nervous tissue were introduced in Germany, Austria and Japan. These “non-contact” devices was once again considered as a potential neurosurgical instrument, as they were accurate, haemostatic, and safe for vital and functional structures such as brain stem, cranial nerves or functional areas of cerebral cortex. The photothermal effects, particularly coagulation and vaporization, were used in common indications such as intracerebral, extra-axial, and spinal tumor resection, or in vascular malformation coagulation and resection. In the early 1970s, Sutton (1971) introduced the possibility to treat tumors, notably brain tumors, with a new concept: interstitial hyperthermia. The technique consisted of introducing a heat source within the tumor and was based on the higher thermal sensitivity of hypoxic and low-pH tissues to cause selective tissue necrosis in tumors. Several experimental studies were performed to treat tumors with interstitial hyperthermia, by using different types of heat sources: radiofrequency, microwave, or ultrasound. Bown (1983) was one of the first to use laser as a heat source in interstitial hyperthermia. The first applications of interstitial hyperthermia in neurosurgery were reported in 1990. Sugiyama et al. (1990) used LITT in a clinical study to treat five deep-seated tumors (three glioma and two brain metastases) in five patients with a Nd-YAG laser in combination with systemic drug therapy. No difficulties or side-effects were described during the procedures. One patient died of primary lung cancer 11 months after the treatment, another patient died of recurrent glioma at the original site (different from the irradiated site); in the three other cases, tumors disappeared on CT images without any recurrence after a follow-up of 9–31 months. This innovative experiment showed that Nd-YAG lasers were safe to use and could be effective for brain tumor treatments. After this initial clinical success, laser interstitial hyperthermia, also known as laser interstitial thermotherapy (LITT), was then developed to treat deepseated tumors. But, while laser vaporization could be performed under visual control, there was no direct

20 Magnetic Resonance-Guided Laser Interstitial Thermal Therapy for Brain Tumors

visual control of the treated volume in interstitial laser therapy, leading to considerable difficulties in accurately controlling the heat distribution and necrosed region in the targeted tissue, without damaging surrounding normal parenchyma. To overcome this limitation and control the extent of the heated region, research was performed to develop imaging techniques for more precisely controlling LITT procedures. MRI was demonstrated to be the best imaging modality for treatment monitoring because of its high soft tissue contrast and its ability to accurately measure temperature elevation. MRI was first used to control the evolution of thermal lesions during LITT procedures with Tracz et al. (1992), but the development of the tissue-independent, proton-resonancefrequency (PRF) reliable methods for temperature measurements established by De Porter (1995) revolutionized the technique. Recently, a new technology used by Carpentier et al. (2008) allows for real-time temperature imaging and prediction of the extent of thermal damage, improving the safety and efficiency of LITT in treatment of brain tumors.

Physical and Technical Parameters (a) Fundamental Principles of Laser Technology “LASER” is the acronym for Light Amplification by Stimulated Emission of Radiation. The physical basis comes from Einstein’s work in the early 20th century and the principle of stimulated emission of photons. A very brief overview of laser physics is described here. Electrons in the nuclei of atoms are at different energy states. When atoms absorb energy, the absorption yields an electron transition to a higher energy state. Subsequently, the electron of the energized atom can spontaneously drop to a lower energy state and produce a photon during the transition. These atomic events occur continuously in the world around us. In the case of stimulated emission, a cascade effect is created where photons stimulate the release of other photons and a single frequency light beam is created. Of course, this is a schematic view of a complex physics phenomenon, in which we do not take into account the complexity of energy states, molecules, and interactions between spontaneous and stimulated reactions.

175

But, it allows for an understanding of how lasers can produce light of high energy, characterized by three properties: monochromaticity, coherence and collimation. Monochromatic laser light is composed of a single wavelength (1.06 μm for Nd-YAG laser, 10.6 μm for carbon dioxide laser), in contrast to ambient light that contains wavelengths from a broad electromagnetic spectrum. Lasers produce intense light because of the coherence of the beam: all the monochromatic waves travel from their origin temporally and spatially in phase. Lastly, laser light does not diverge in space and travels in a narrow beam, so that it can travel long distances in a parallel beam.

(b) Functioning Principles A laser is composed of three main elements: – an active medium (lasing medium) which can be stimulated and excited to emit photons to generate the laser beam – an optical resonator in which the active medium is housed – a stimulation source that can excite and activate the lasing medium The stimulated emission occurs in the active medium. This medium can be solid (ruby, neodymium: YttriumAluminium-Grenat), gaseous (carbon dioxide, heliumneon, argon, krypton) or liquid (organic dyes). The medium is housed in a glass chamber called an optical generator at the two extremities of which are two mirrors. One mirror is completely reflective and the other is partially reflective. The stimulation source is generally a voltage source, but it can be a chemical reaction (chemical lasers), a flash lamp or another laser (solid lasers, dye lasers. . .). When the stimulation source is on, its energy is absorbed by the active medium. This leads to electron transitions in the atoms of the medium. When the majority of these atoms attain the energized state, spontaneous and stimulated emissions of photons occur that permit a chain reaction previously described to occur. The totally reflected mirror reflects the photons to the partially reflected mirror through which the highly concentrated parallel laser beam escapes from the optical resonator. Photons emitted non parallel to

176

the beam axis are dissipated as heat on the side of the optical resonator. The transmission of the laser beam is ideally provided by a fiber optic, generally 400–600 μm in diameter. This optical fiber consists of a silica-based core surrounded by a thin cladding also made of silica or a hard polymer material. The high difference in refractive index of the core and the cladding of the fiber results in total internal reflection along the fiber, so that large amounts of energy can be transmitted over a long distance with minimal losses and good efficiency. However it can only transmit continuous or long pulsed (micro- or milliseconds) emissions of near-ultraviolet, visible or near-infrared light. Therefore, it is not suitable for wavelengths in regions of the far-ultraviolet (excimer lasers), the far-infrared (carbon dioxide lasers), or for very short pulses (nanoseconds). In order to generate spherical or ellipsoidal tissue lesions, a diffusing tip is used at the end of the optical fiber in LITT applications. These fiber tips can irradiate in a circumferential and homogenous area over 0.5–2.5 cm. Diffusing fibers often incorporate small particles such as gold or titanium dioxide, in an elastomer mixture such as a silicone epoxy, and can resist high temperatures as reported Stafford et al. (2010). The first laser fibers used were bare fibers, without any cooling system, and were restricted in use to an output power of about 2–5 W and application times of 15–20 min. Higher powers or application times could lead to vaporization or charring of the irradiated tissue, and could destroy the fiber. The consequence was that LITT procedures were limited to the generation of lesions of up to 2 cm in diameter. The generation of larger lesions required the use of several fibers, successively or at the same time. The development of internally-cooled lasers now allows for the treatment of larger volumes with additional thermal control. In internally-cooled lasers, fiber optics are inserted inside specific double-lumen probes (or envelope, or applicator) in which circulates a roomtemperature saline cooling fluid. This protects the optical fiber from burning, and allows for laser powers of up to 30 W, with consequent generation of larger and more controllable volumes of coagulation necrosis. Thus, lesions up to 3 cm in diameter can be obtained; furthermore, complex and irregular volume lesions can be generated by moving the fiber within the applicator.

K. Beccaria et al.

(c) Lasers Used in LITT Several hundred different kinds of lasers have been developed since the first ruby-based laser in 1958, each one with particular properties and uses depending on their physical characteristics (active medium, wavelength, emission mode. . .). However, only a few are used for medical procedures, in particular for neurosurgical procedures. Since the technique has been first described, LITT has conventionally been performed with Nd-YAG lasers. In some recent studies on animals by Kangasniemi et al. (2004) and on humans by Carpentier et al. (2008), the diode laser has also been tested for LITT. The YAG laser is a laser in which the active medium is a complex crystal formed by three components: Yttrium, Aluminum and Garnet. These lasers emit infrared radiation, and their wavelength can be doped by another substance. In Nd-YAG lasers, this substance is Neodymium, and the wavelength obtained is generally about 1064 nm (from 939 to 1440 nm). The effects on tissue of exposure to Nd-YAG lasers depend on the wavelength used and in the composition of the irradiated tissue (water, pigmentation. . .). An NdYAG laser (1064 nm) has a low absorption in water (10 times less than carbon dioxide lasers), and consequently, the penetration and extinction length (distance at which 90% of the incident laser energy is totally absorbed) in cerebral tissue is higher. This induces a dominant and diffuse scatter effect and deep heating within the volume of tissue concerned. While it is poorly absorbed by tissue, Devaux and Roux (1996) showed that Nd-YAG energy is preferentially absorbed by heme pigmented tissue, so that 1064 nm wavelength Nd-YAG laser allows good hemostasis, but poor vaporization. Its deep tissue scatter and heating effect make 1064 nm wavelength Nd-YAG laser anideal device for LITT procedures. The Diode laser is also easy to use, since they are small in size and only require a standard power plug. Moreover, these lasers are much cheaper than Nd-YAG lasers. The energy conversion to light from diode lasers is also high (30–40%). Kangasniemi et al. (2004) have used 980 nm diode lasers for LITT procedures. Kou et al. (1993) showed that the absorption coefficient of water is about four times higher at 980 nm than it is at 1064 nm. Thus, energy is quickly absorbed in brain tissue, even with low power. Since the absorption is

20 Magnetic Resonance-Guided Laser Interstitial Thermal Therapy for Brain Tumors

higher, the heating is faster than for Nd-YAG lasers, meaning that it requires less exposure time and faster treatments can be performed.

Interactions Between Laser and Biological Tissues: Mechanisms and Histology (a) Immediate Mechanisms of Laser Interactions with Tissues Immediate Laser interactions with biological tissues can be divided in four mechanisms well described by Mordon and Brunetaud (1992). The first one is the photothermal effect due to conversion of light into heat. This effect is used in medical applications of lasers. When the laser beam interacts with the target tissue, the beam is reflected, scattered, or absorbed by the tissue. When the light is absorbed by the tissue, it is converted into heat. The depth of penetration and the absorption depend on the wavelength used and on the molecular structure and the composition of the tissue (cellular composition, water content, pigmentation. . .). As the energy is absorbed, the tissue heats up and heat is transported to adjacent tissue through conduction and convection by blood flow. When the tissue reaches a critical temperature, cell death occurs. The precise mechanisms of cell death that occur with heating depend on the tissue being treated and on the intensity and the duration of heating. In general, three different heating regimes can be defined and lead to different effects at the cellular level: – Hyperthermia is a moderate elevation of temperature of only several degrees from body temperature to 40–45◦ C. As shown by Salcman and Samaras (1981) prolonged and moderate temperature elevation leads to alteration of cell membranes and enzyme denaturation responsible for a selective necrosis of the heated tissue without vaporization. Hypoxic and low-pH tissues may have higher thermal sensitivity and be particularly sensitive to this effect. – Coagulation corresponds to an irreversible necrosis without substance loss, and is due to high (50–95◦ C) and rapid heating of tissue. Tissue is

177

dessicated and retracted. Healing appears after destruction of the necrotic tissue. – Vaporization corresponds to an immediate substance loss due to very high (more than 100◦ C) and rapid (tenth of a second) temperature elevation. The boundaries of the lesion are characterized by coagulation necrosis. The second effect of tissues exposed to lasers is the photochemical effect, which is based on the reaction between tissue and a photosensitizing agent activated by light. The photosensitizing agent generally has tropism for tumorous tissue. The activation of the agent leads to phototoxic reactions, or fluorescent radiation so that it can be used for diagnostic and treatment. The third effect is photomechanical effect. High intensity beams concentrated on small surfaces with very shorts pulses can create local shock waves and mechanical stress. This effect is used to treat urinary and biliary lithiasis. Lastly, the fourth mechanism is the photoablative effect. It corresponds to rupture of molecular links within the tissue subjected to light. There is material loss without temperature elevation.

(b) Secondary Mechanisms of Laser Interactions with Tissues Immune stimulation: Increased HSP70 immunoreactivity has been observed by Ivarson et al. (2003) in a model of induced liver adenocarcinoma treated by LITT, with HSP70 shifts from cytoplasm to nucleus. Furthermore, LITT resulted in an increased number of tumour-infiltrating macrophages with increased presence of HSP70 in their membrane and cytoplasm. Growth factor: In another study by Isbert et al. (2007), perfomed on the large clinical hepatic database of Vogl et al. (1998), LITT was shown to decrease a residual liver adenocarcinoma growth in comparison to hepatic surgical resection. The phenomenon was explained by a different expression of mRNA of specific growth factors (HGF and CTGF) after LITT (low expression) and hepatic surgical resection (high expression). Blood brain barrier opening: Disruption of the blood brain barrier (BBB) has been observed in immuno histochemical studies after cerebral tissue irradiation with laser. Roux et al. showed that the disruption

178

was confined to the lesion area, predominantly in the peripheral zone of edema. After a 10 min exposition to temperatures ranging from 42.5 and 43.5◦ C, Sugiyama et al. (1990) noticed that the blood brain barrier remained opened for 6 days following the procedure.

(c) Histology of LITT-Induced Lesions in Brain Tissue Histological changes in cerebral tissue treated by LITT have been studied in vivo and ex vivo in cats by Tracz et al. (1992), rabbits by Schatz et al. (1992), rats by Schober et al. (1993), pigs by Schulze et al. (1998) and Higuchi et al. (1992) and dogs by Kangasniemi et al. (2004). The shape of acute lesions due to LITT procedure depends strongly on the distal diffuser system at the end of the optic fiber. The induced necrotic lesions are typically ellipsoidal or spherical in shape and have a typical architectural structure, divided into three to four different zones. The immediate vicinity of the fiber track is the site of incipient vaporization where the lesion is produced with high power (>2–3 W) and with induced temperatures higher than 100◦ C. This corresponds to a pseudo-cavity with charred and carbonized tissue, hemoglobin decomposition products, fibrin and debris at its margins. No evidence of central vaporization is seen at low power. This vaporization phenomenon has to be avoided in clinical applications. The only way to prevent such phenomenon is to use real time MRI thermal imaging sequences with a automatic computerized controlled laser emission. Beyond vaporization, when it exists, appears a typical lesion due to laser application: coagulation necrosis. This zone of coagulation necrosis is not homogenous and consists of two different rims. The inner rim is composed by dense coagulation necrosis, in which tissue structure is well distinguishable, but not viable. Cell elements are shrunken and blurred. Cellular, nuclear and mitochondrial membranes of nerve cells, glial cells, and endothelial cells either display local defects or are broken up into fragments. Blood vessels have thickened walls, and are full of erythrocyte “ghosts” devoid of hemoglobin. Myelin sheaths have a granular appearance and axonal structures are fragmented and partially curled. The outer rim is composed of disperse coagulation necrosis in which fine tissue structure is better conserved, but infiltrated by interstitial fluid. Cells are necrotic, but have preserved

K. Beccaria et al.

outlines. Blood vessels are not thrombosed, and axons do not show apparent changes. This distinction between dense and disperse coagulation necrosis is not described by all authors. Coagulation necrosis is sharply separated from a surrounding zone of edema. In this zone of viable brain, microvacuolisation of the neurophil is visible, as well as neuronal shrinkage and axonal swelling. There is no thrombosis. Some rare polymorph neutrophils and mononuclear cells can also be observed. The lesion is characterized by a blood-brain-barrier opening which is stronger in this last zone of edema. In cat and rabbit brains, the size of the acute lesion is directly dependent on both power and energy deposited. The size increases with the applied power. However, in their clinical study in 8 patients, Kahn et al. (1994) did not find any relation between lesion size and applied laser energy on intracerebral tumors. Lesion size also evolves in time after the treatment. Lesions typically increase in size from 50 to 200% in volume within 24–48 h after laser irradiation. Then, they decrease progressively to return back to their initial size in 7–10 days. The initial expansion of lesion size could be attributed to several mechanisms: vascular damage and development of micro hemorrhages at the margins of the thermal lesions, delayed additional hyperthermic cell death at the periphery of the lesion, or intrafocal edema. During the first week after irradiation, the central zone of coagulation becomes necrotic with progressive blurring of cell borders and a loss of histologic staining qualities (necrosis is completely unstained by day 3–4). A resorptive reaction begins from the periphery with in-growth of capillaries and macrophage transforming to gitter cells. Edema spreads to the adjacent normal brain tissue and reaches a maximum 3–6 days after treatment. At 1 week, a circular rim of granulation tissue surrounds the necrosis, proceeding from the periphery towards the center. This can be observed as a characteristic healing response to necrosis. In some cases, the center of necrosis is progressively liquefying with formation of a central cyst defect, well-defined by a thin fibrous capsule in 2–3 months. With their canine tumor model, Kangasniemi et al. (2004) showed that the LITT procedure was able to destroy induced brain transmissible venereal tumors (TVT). In three out of seven dogs, a peripheral region around the thermally coagulated tissue showed persistence of tumor cells. In these three cases, surviving tumor cells were located out of the

20 Magnetic Resonance-Guided Laser Interstitial Thermal Therapy for Brain Tumors

irradiation field of the laser fiber tip that was suitable for round or oval lesions not larger than approximately 10 mm.

(d) Thermal Dosimetry Optimal and safe use of LITT depends on the ability to carefully control the heat distribution and to predict the extent of coagulation necrosis during the procedure. Modern MRI sequences allow accurate temperature mapping in real-time during the procedure. To predict the induced thermal damage, the temperature elevation in the irradiated volume is used to predict a probability of necrosis of the tissue. This is possible thanks to mathematical thermal dose models which estimate thermal damage to tissue from the spatiotemporal temperature history. Three thermal models are currently used. Yung et al. (2010) have shown that all three thermal dose models can be considered reliable. In the Arrhenius rate analysis model, damage is modeled as a change in the state of the tissue, knowing that coagulation begins to be apparent from 54 to 60◦ C, with denaturation of proteins and cellular components. The CEM43 model developed by Sapareto et al. (1984) is based on the Arrhenius model and empirical data from hyperthermia observations: damage is quantified by relating the temporal temperature history to a reference constant temperature of 43◦ C. A temperature of 43◦ C has been arbitrarily chosen to convert all thermal exposures to “equivalent-minutes” at this temperature: CEM43 is the Cumulative Equivalent Minutes at the reference temperature of 43◦ C. Lastly, the threshold temperature model is based on the fact that tissue is damaged nearly instantaneously once it reaches a certain temperature. This model does not take into account the temperature history, and is particularly adapted to rapid tissue ablation.

MRI Imaging and LITT Medical imaging is essential during the different steps of LITT procedures, from target definition to followup of induced lesions, through trajectory selection, fiber tracking, and prediction of final lesion size during the treatment. Monitoring of the thermal damage is essential to ensure that the all the lesion is treated, and

179

to prevent extension of damage to vital or functional structures. In the previous trials by Sugiyama et al. (1990) and Roux et al. (1992a, b), laser probes were stereotactically inserted within tumors, and thermocouple probes were stereotactically inserted a few millimeters away to control the temperature rise during procedures. This monitoring was obviously invasive and not precise, and there was a need for an in vivo method to test, monitor, and control the effects of medical lasers before, during, and after the LITT procedure. Monitoring of LITT with ultrasound imaging has been studied in pigs by Schulze et al. (1998), and showed that temperature elevation could be monitored during the procedure. However, ultrasound-based monitoring techniques are not feasible for neurosurgery, since the intact skull obstructs penetration of acoustic waves. Computed tomography (CT) is also not an appropriate imaging modality as it has low soft-tissue contrast resolution, and its sensitivity to early tissue changes following laser irradiation is poor as shown by Menovsky et al. (1996). On the contrary, MRI has high soft tissue contrast, multiplanar imaging capabilities, high spatial resolution and temperature sensitivity, so that it is currently the best imaging modality to monitor LITT with good correlation with histology and to follow the evolution of treated lesions.

(a) MRI Thermal Imaging Sequences Usefulness of MRI in LITT monitoring has been first described by Jolesz et al. (1988). In their in vitro and in vivo study, they showed the potential of MRI to map the spatial and temporal distribution of the effects of Nd:YAG lasers on cerebral tissue. During rabbit brain irradiation, they observed a complete loss of signal intensity at the fiber optic tip and a decrease in signal intensity around it; the halo of decreased signal intensity surrounding the lesion at the fiber tip returned to normal intensity after irradiation. Therefore, they were able to use MR imaging to locate the target tissue to be treated and test the position of the laser, and they showed that MR images could demonstrate the spatial extent of both reversible and irreversible thermally induced tissue changes. The authors interpreted the irreversible and complete signal loss as a combination of tissue water loss and altered tissue

180

K. Beccaria et al.

t = 0s

t = 18s

t = 54s

t = 103s

t = 139s

t = 187s

Fig. 20.1 Real time MRI thermal imaging (upper row) and necrosis prediction superimposed on anatomic images (from Carpentier et al., 2008)

water mobility, and the surrounding reversible signal loss as an effect of a local temperature rise. Tracz et al. (1992) also observed the appearance of a darkto-hypointense region around the fiber tip during irradiation of normal brains of anesthetized cats. However, these signal changes are not appropriate to monitor LITT. First, there is no possibility to distinguish signal changes caused by coagulation and necrosis, and caused by temperature elevation. Moreover, as seen before, the induced necrosis evolves during the first 24–48 h after the treatment so that the acute image does not correspond to the final lesion. To overcome these initial problems with MR lesion monitoring, temperature mapping was proposed to predict real-time thermal damage and control the LITT procedure. In phantom materials and normal rabbit brain tissue in vivo, Bleier et al. (1991) investigated fast diffusion imaging with every 2 s acquisitions: they were able to show the dynamics of temperaturerelated signal intensity changes in the regions irradiated by an Nd-YAG laser. Meanwhile, the protonresonance-frequency (PRF) method was used by De Porter (1995), which can provide reliable temperature quantification in vivo independently from the tissue type. Schulze et al. (1998) tested the PRF method in pig brains from cadavers. With temperature maps acquired every minute, they could visualize a nearly circular distribution of temperature around the fiber tip and expansion of the heated region during the course of LITT. Temperature at the vicinity of the probe reached more than 80◦ C after 15 min, and they observed a steep temperature gradient with increasing distance to the laser fiber tip. Histological analysis showed that the border of the lesions corresponded

to 60–65◦ C isotherms on the MRI maps. This led Kahn et al. (1998) to perform the first LITT procedure with MR PRF thermometry in a patient with an astrocytoma WHO II located adjacent to the left pre-central gyrus. During the irradiating procedure, they monitored the 60–65◦ C isotherm to visually predict the limits of the induced tissue damage (see Fig. 20.1). In addition, a good correlation was observed between this isotherm and an enhancing rim around the lesion on post-irradiation contrast-enhanced MRI. No side effects were noted, the patient’s condition improved, and follow-up studies until 120 days after LITT showed a continuing regression of the lesion size. Kickhefel et al. (2010) showed that single shot echoplanar imaging (ss EPI) currently used sequences are faster and more precise for thermometry compared to gradient echo (GRE) or segmented echoplanar imaging (seg EPI) sequences.

(b) Real Time Computation Recent progress in computer hardware has led to the development of software which can provide rapid processing of MR data to produce real-time quantitative temperature maps, estimates of thermal ablation zones, and computer-controlled feedback during treatments. These methods are now currently used routinely in both preclinical and clinical studies. For example, in the clinical study by Carpentier et al. (2008), MR temperature mapping was obtained in a single image plane in the brain centered on the laser fiber every 6 s during ablation. Improved MR sequences can be further used to obtain 3D acquisitions through the treatment

20 Magnetic Resonance-Guided Laser Interstitial Thermal Therapy for Brain Tumors

181

volume, allowing for precise control of heat distribution and necrosis prediction. Thus, adjacent normal tissue structures can be preserved and the safety of the procedure can be further improved. Furthermore, security processes built-in to the control software can allow for the determination of “security points” before the procedure (see Fig. 20.2). Fiber damage or lesion of vital or functional structures can be avoided by positioning high-temperature limit points to limit temperatures below a pre-definite value. Other points can be used to confirm temperature elevation above a critical value in order to obtain induced necrosis. Thanks to automatized control of laser delivery, laser irradiation can then be either stopped manually by the surgeon when the predicted thermal ablation zone is sufficient or automatically if any of the temperature limits is exceeded.

(c) Evolution of Brain Thermal Lesions on MRI

Fig. 20.2 Laser control delivery software with real time MRI thermometry analysis allows security processes: hightemperature limit security points to limit temperatures, Temperature follow-up points, desired target volume. The laser

irradiation can then be either stopped manually by the surgeon when the predicted thermal ablation zone is sufficient or automatically if any of the temperature limits is exceeded. (Courtesy of BioTex Visualase Inc.)

Many studies have described the appearance of thermally-induced lesions and their evolution on MRI. Schawabe et al. (1997) and Kahn et al. (1994) described their findings in patients treated by LITT for primary or secondary brain tumors. Kangasniemi et al. (2004) studied the evolution of LITT lesions after treatment of induced tumors in dog brains. In these studies, the results were similar and complementary. The typical lesion architecture observed immediately after LITT is comprised of five concentric zones: the light guide, the central zone, the peripheral zone, a thin rim at the outer border of the peripheral zone, and perifocal edema. In T1-weighted images, the light

182

K. Beccaria et al.

J0

J7

J15

M1

M3

Fig. 20.3 Contrast T1 MRI follow-up after a LITT treatment of a radiosurgery resistant metastasis

guide and peripheral zones are hypointense while the central zone is hyperintense. The perifocal edema is slightly hypointense. The thin rim at the border of the peripheral zone is hypointense with enhancement after gadolinium injection. Signal intensities of the different zones are opposite on T2-weighted images. The total size of lesions can increase by 0–45% in diameter during the first 10 days after LITT, essentially due to expansion of the peripheral zone. After the initial increase, the total lesion size decreases exponentially to reach 50% of the initial lesion size within a mean of 93 days (see Fig. 20.3). Kangasniemi et al. (2004) observed an increase of lesion size up to 250% and explained this phenomenon by the different absorption characteristics of tissues at 980 nm. Perifocal edema is generally not apparent immediately following LITT treatments, but appears between 1 and 3 days after therapy with a maximum extent after 4–27 days (mean 6 days). The regression of edema can last 15–45 days. The severity of edema generally does not coincide with the tumor grade or the applied laser energy. During the evolution of the lesion, the signal intensity of the central zone decreases progressively in T1-weighted images, while the signal intensity of the peripheral zone increases, resulting in more homogenous lesion, without differentiation into different zones. Moreover, the enhancing rim after gadolinium injection shows a continuous reduction in diameter and enhancement, and is visible even in late controls after treatment. According to the authors, this zonal division in MRI corresponds to the zonal architecture described in histological studies. High signal intensity of the central zone in T1-weighted images may correspond to heat-induced methemoglobin conversion from deoxyhemoglobin and high protein content fluid collections. Low signal intensity in T1-weighted images and high

signal intensity in T2-weighted images of the peripheral zone correspond to more or less important edema.

Clinical Studies Sugiyama et al. (1990) were the first to publish the results of LITT applied to brain tumors in patients, who were treated with a combination of LITT and drug therapy. The procedure was performed under general anesthesia, and thermocouples placed during CT-guided stereotactic procedures at the periphery and in the center of the tumor maintained temperatures between 42.5 and 43.5◦ C during the treatment. A 30– 40 min hyperthermia treatment was performed with 2–3 W of applied power with a Nd-YAG laser. Five patients were treated for metastatic lung carcinomas (2) and gliomas (3). There was no surgical difficulty during the procedure and all metastases treated disappeared, with no recurrence at either 11 or 29 months of follow-up. One patient died 11 months after treatment from primary lung cancer; the other patient was still alive 29 months after treatment. Roux et al. (1992a,b) reported LITT for brain tumors, in which one patient was treated for a 15 × 15 × 15 mm thalamic metastasis of melanoma. The procedure was performed in two steps: first, location and biopsy was performed under stereotactic conditions and several days later, LITT was used. A thermocouple was placed at the tumor periphery to assure a temperature between 41 and 43◦ C and LITT was performed during 17.5 min at 5–6 W. There was no peri-operative morbidity or mortality and the patient was discharged one week after the treatment. However, the tumor eventually grew, the patient’s condition deteriorated, and the patient died 4 months later.

20 Magnetic Resonance-Guided Laser Interstitial Thermal Therapy for Brain Tumors

Kahn et al. (1994) and Schwabe et al. (1997) treated two cases of brain metastasis. LITT was performed immediately following stereotactic biopsy and histology analysis. One metastasis was 27 mm in diameter, and the other, an adenocarcinoma metastasis, was 20 mm in diameter. A Nd-YAG laser (1064 nm wavelength and continuous wave) was used to emit 4 W during 16 and 20 min treatments, respectively. Monitoring was assured by MRI morphologic control and LITT was stopped when the diameter of the peripheral zone determined on images during LITT was similar to the size of the neoplasm. During therapy, the patient treated for adenocarcinoma metastasis developed an increase of preexisting hemiparesis and an incomplete motor aphasia; he died 1 week after the procedure. In the second patient, the lesion could not be segmented due to its inhomogeneous structure and a preexisting rim-like Gd-DTPA enhancement caused by the disruption of the blood-brain barrier that could not be differentiated from the Gd-DTPA-enhancing rim at the outer border of the irreversibly damaged lesion. In their review, Schulze et al. (2004) reported the results of seven LITT treatments for metastatic brain tumors. The procedures were still performed under general anesthesia and localization of the tumor was determined by MRI imaging. After biopsy and histological analysis, the light guide was inserted under MRI guidance. Laser energy was applied using a 12 W pulsed laser for 10–16 min. The heating process and temperature elevation was monitored in real-time by a gradient echo MR sequence and the PRF method. Custom software allowed interactive monitoring of temperature and adjustment of therapy if necessary. Therapy was adjusted to the temperature within the lesion and the size of the tumor and its neighboring structures. The mean time interval between the initiation and the end of anesthesia was 296.1 min (min: 130 min, max: 350 min) and the total time of stereotactic interstitial laser procedure was 224.6 min (min: 105 min, max: 330 min). There was no intra-operative complication, but follow up was not described. Recently, Carpentier et al. (2008) reported a pilot clinical trial using a new real-time magnetic resonanceguided LITT system for 15 radiosurgery-resistant focal metastatic intracranial tumors in 6 patients. All the procedures were performed under local anesthesia. The laser fiber was connected to a 15-W 980-nm diode laser and placed in the treatment zone under stereotactic MRI guidance. A low-power, subtherapeutic, test

183

pulse (3–4 W, 30–60 s) was first applied to allow thermal visualization and to check the probe position. Subsequently, a treatment dose (7.5–15 W, 30–180 s) was applied and the PRF method was used to obtain temperature maps during the treatment every six seconds. Several high-temperature limit points were set near the applicator to limit temperatures below 90◦ C, and one to two low-temperature limits were set either near the margin of the desired thermal ablation zone or at the border of critical structures to maintain temperatures below 50◦ C. Automatic control of laser delivery and a security process was used during all procedures. In all cases, post-treatment MRI findings matched with estimated lesions. The immediate treatment result was complete ablation in 3 cases and partial ablation in the 3 others (decided for safety reasons at the beginning of the cases). There was no peri-operative mortality; morbidity was limited to one probe misplacement on a bone free posterior fossa lesion, one transient increase in cerebellar syndrome resolving after 3 weeks, and one transient aphasia resolving after 2 weeks. With the exception of one patient who had been hospitalized before the procedure, all patients were discharged home within 14 h after treatment. All patients except one had improvement of Karnofsky Performance Index after treatment. Mass effects and edema did not increase after treatment, and lesions showed a typical time course with a maximal increase in apparent diameter at approximately 7 days, followed by a steady decrease in size. No peripheral recurrence was observed at 12 months after treatment in patients with complete ablation; for patients in whom treatment was partial, peripheral recurrence was slightly visible at 3 months. Long-term follow-up shows an estimated median survival of 17.4 ± 3.5 months, exceeding the median prognosis survival prediction at the time of enrollment (7 months). This survival gain was accompanied by sustained or improved (2 patients in better neurological condition) quality of life, and patient neurological deterioration was delayed. Two patients are still alive at 30 and 19 months follow-up. Since these trials were completed, the LITT procedure has been performed as an out-patient procedure at several other clinics. On-going clinical trials are now performed at the MD Anderson Cancer Center for brain tumour metastasis and at the Cleveland Clinic for glioblastomas. Several devices, manufactured for example by BiotexVisualise (Tx) or Monteris Medical

184

(Canada) have received FDA clearance, so that many centers are beginning to use this optimal technology process almost in a clinical routine mode.

Conclusion With successful demonstration of laser interstitial thermotherapy treatments in several clinical studies, there has been a recent revival of interest in their application in neurosurgery. Several clinical studies have shown that LITT was able to control brain metastasis with few side effects in patients for whom standard medical treatments had failed. The improved safety and efficiency of modern LITT procedures is due to technical developments in laser technology, medical imaging, and computerized real-time control of laser emission. Today’s lasers are easier to use and laser irradiation can be better controlled. Development of new MRI sequences allows for the control of heat deposition in tissue in real-time and for the prediction of the extent of induced thermal lesions, so that clinicians can now generate lesions in malignant tissue with minimal damage to the healthy surrounding brain tissue. However, limits of this technique do exist such as the limited diameter of treatable lesions and the volume increase effect of the treated region in the days following treatments. Currently, these effects limit, by now, LITT indications to 35 mm maximum diameter brain tumors with a regular and elementary volume. Moreover, the first clinical results are promising, but larger patient samples and long-term efficacy studies are necessary to confirm these results.

References Bleier AR, Jolesz FA, Cohen MS, Weisskoff RM, Dalcanton JJ, Higuchi N, Feinberg DA, Rosen BR, McKinstry RC, Hushek SG (1991) Real-time magnetic resonance imaging of laser heat deposition in tissue. Magn Reson Med 21: 132–137 Bown SG (1983) Phototherapy of tumours. World J Surg 7: 700–701 Carpentier A, McNichols RJ, Stafford RJ, Itzcovitz J, Guichard JP, Reizine D, Delaloge S, Vicaut E, Payen D, Gowda A, George B (2008) Real-time magnetic resonance-guided laser thermal therapy for focal metastatic brain tumors. Neurosurgery 63(1 Suppl 1):ONS21-8, discussion ONS28–9 De Poorter J (1995) Noninvasive MRI thermometry with the proton resonance frequency method: study of susceptibility effects. Magn Reson Med 34:359–367

K. Beccaria et al. Devaux BC, Roux FX (1996) Experimental and clinical standards, and evolution of lasers in neurosurgery. Acta Neurochir (Wien) 138:1135–1147 Higuchi N, Bleier AR, Jolesz FA, Colucci VM, Morris JH (1992) Magnetic resonance imaging of the acute effects of interstitial neodymium:YAG laser irradiation on tissues. Invest Radiol 27:814–821 Isbert C, Ritz JP, Roggan A, Schuppan D, Ajubi N, Buhr HJ, Hohenberger W, Germer CT (2007) Laser-induced thermotherapy (LITT) elevates mRNA expression of connective tissue growth factor (CTGF) associated with reduced tumor growth of liver metastases compared to hepatic resection. Lasers Surg Med 39:42–50 Ivarsson K, Myllymäki L, Jansner K, Bruun A, Stenram U, Tranberg KG (2003) Heat shock protein 70 (HSP70) after laser thermotherapy of an adenocarcinoma transplanted into rat liver. Anticancer Res 23 (5A):3703–3712 Jolesz FA, Bleier AR, Jakab P, Ruenzel PW, Huttl K, Jako GJ (1988) MR imaging of laser-tissue interactions. Radiology 168:249–253 Kahn T, Bettag M, Ulrich F, Schwarzmaier HJ, Schober R, Fürst G, Mödder U (1994) MRI-guided laser-induced interstitial thermotherapy of cerebral neoplasms. J Comput Assist Tomogr 18 (4):519–532 Kahn T, Harth T, Kiwit JCW, Schwarzmaier HJ, Wald C, Mödder U (1998) In vivo MRI thermometry using a phase-sensitive sequence: preliminary experience during MRI-guided laserinduced interstitial thermotherapy of brain tumors. J Magn Reson Imaging 8:160–164 Kangasniemi M, McNichols R, Bankson JA, Gowda A, Price RE, Hazle JD (2004) Thermal therapy of canine cerebral tumors using a 980 nm diode laser with MR temperaturesensitive imaging feedback. Lasers Surg Med 35:41–50 Kickhefel A, Roland J, Weiss C, Schick F (2010) Accuracy of real-time MR temperature mapping in the brain: a comparison of fast sequences. Phys Med XX:1–10 Kou L, Labrie D, Chylek P (1993) Refractive indices of water and ice in the 0.65-2.5 μm spectral range. Appl Opt 32: 3531–3540 Menovsky T, Beek JF, Van Gemert MJC, Roux FX, Bown SG (1996) Interstitial laser thermotherapy in neurosurgery: a review. Acta Neurochir (Wien) 138:1019–1026 Mordon S, Brunetaud JM (1992) Bases physiques des applications thérapeutiques des lasers. Neurochirurgie 38:203–207 Roux FX, Merienne L, Devaux B, Leriche B, Ciocola C (1992a) Les lasers YAG en neurochirurgie. Neurochirurgie 38: 229–234 Roux FX, Merienne L, Fallet-Bianco C, Beuvon F, Devaux B, Leriche B, Cioloca C (1992b) La thermothérapie interstitielle laser stéréotaxique. Une alternative dans la prise en charge thérapeutique de certaines tumeurs cérébrales. Neurochirurgie 38:238–244 Salcman M, Samaras GM (1981) Hyperthermia for braintumors:biophysical rationale. Neurosurgery 9:327–335 Sapareto SA, Dewey WC (1984) Thermal dose determination in cancer therapy. Int J Radiat Oncol Biol Phys 10 (6):787–800 Schatz SW, Bown SG, Wyman DR, Groves JT, Wilson BC (1992) Low power interstitial Nd:YAG laser photocoagulation in normal rabbit brain. Lasers Med Sci 7:433–439 Schober R, Bettag M, Sabel M, Ulrich F, Hessel S (1993) Fine structure of zonal changes in experimental Nd:YAG

20 Magnetic Resonance-Guided Laser Interstitial Thermal Therapy for Brain Tumors laser-induced interstitial hyperthermia. Lasers Surg Med 13: 234–241 Schulze CP, Kahn T, Harth T, Schwurzmaier HJ, Schober R (1998) Correlation of neuropathologic findings and phasebased MRI temperature maps in experimental laser-induced interstitial thermotherapy. J Magn Reson Imaging 8: 115–120 Schulze PC, Vitzthum HE, Goldammer A, Schneider JP, Schober R (2004) Laser-induced thermotherapy of neoplastic lesions in brain – underlying tissue alterations, MRI-monitoring and clinical applicability. Acta Neurochir (Wien) 146:803–812, Review article Schwabe B, Kahn T, Harth T, Ulrich F, Schwarzmaier HJ (1997) Laser-induced thermal lesions in the human brain: short- and long-term appearance on MRI. J Comput Assist Tomogr 21 (5):818–825 Stafford RJ, Fuentes D, Elliott AA, Weinberg JS, Ahrar K (2010) Laser-induced thermal therapy for tumor ablation. Crit Rev Biomed Eng 38(1):79–100 Stellar S (1968) Laser studies on nervous system tissue, neoplasms and related biological systems. Proc Virchow Med Soc 26 (Suppl):416–442

185

Sugiyama K, Sakai T, Fujishima I, Ryu H, Uemura K, Yokoyama T (1990) Stereotactic interstitial laser-hyperthermia using Nd-YAG laser. Stereotact Funct Neurosurg 54–55: 501–505 Sutton C (1971) Tumor hyperthermia in the treatment of malignant gliomas of the brain. Trans Ann Neurol Assoc 96: 195–199 Tracz RA, Wyman DR, Little PB, Towner RA, Stewart WA, Schatz SW, Pennock PW, Wilson BC (1992) Magnetic resonance imaging of interstitial laser photocoagulation in brain. Lasers Surg Med 12:165–173 Vogl, Mack, Roggan, Straub, Eichler, Müller, Knappe, Felix, Vogl TJ, Mack MG, Roggan A, Straub R, Eichler KC, Müller PK, Knappe V, Felix R (1998) Internally cooled power laser for MR-guided interstitial laser-induced thermotherapy of liver lesions: initial clinical results. Radiology 209(2):381–385 Yung JP, Shetty A, Elliott A, Weinberg JS, McNichols RJ, Gowda A, Hazle JD, Stafford RJ (2010) Quantitative comparison of thermal dose models in normal canine brain. Med Phys 37(10):5313–5321

Chapter 21

Nanotechnology-Based Therapy for Malignant Tumors of the Central Nervous System Abraham Boskovitz, Abdullah Kandil, and Al Charest

Abstract Glioblastoma multiforme (GBM) is an insidious cancer for which there are currently no efficient treatments. The major impediments to successful treatment of GBM are the tumor cells’ inherent ability to develop resistance to chemotherapeutic agents and ionizing radiation and to migrate and invade normal surrounding tissues. These disastrous characteristics lead to post-treatment tumor recurrence uniformly and result in a high mortality rate. Approaches designed to target and eliminate post-surgical GBMs and their invading cells would result in significant clinical improvements for this cancer. Success in the clinic will come from personalization of cancer therapies. This will first require an in-depth understanding of gliomas at the molecular level to uncover weaknesses that can be therapeutically exploited along with novel approaches that are designed to deliver those therapeutic agents. Biological nanotechnology is a growing field that offers significant improvements in methods of drug delivery for cancer in general and in malignant gliomas. Nanooncology represents one of the most significant applications of nanotechnology to medicine and is poised to produce momentous changes in cancer treatment protocols. For instance, nanoparticles have recently been shown to offer multifunctional platforms combining diagnostics, therapeutics delivery and monitoring of treatment response, demonstrating how nanooncology is paradigm shifting. Here, we review these technologies and forecast improvement in

A. Charest () Department of Neurosurgery, Tufts Medical Center, Tufts University, Boston, MA 02111, USA e-mail: [email protected]

the delivery of therapeutic agents to malignant brain tumors. Keywords Central nervous system · Gliomas · Xenograft · Nanotechnology · Brain Cancer · Phospholipid

Introduction Malignant gliomas comprise a heterogeneous group of neoplasms that widely differ in location, histological features, tendency for progression and invasion and response to therapeutic interventions (for a review on gliomas see Furnari et al. (2007)). Gliomas are histologically graded according to a system established by the World Health Organization (WHO), classifying gliomas into grades I-IV. In this system, low-grade gliomas (WHO grade I-II) behave in a benign fashion and can usually be successfully resected surgically without recurrence. Gliomas of grades III-IV are diffusely infiltrating and are considered malignant. Grade IV gliomas, also known as GBM, are the most aggressive subtype. They can either develop de novo without prior clinical symptoms (primary GBM) or through increasing malignancy from lower grade gliomas (socalled secondary GBM). Regardless of its etiology, malignant glioma is an incurable cancer, demonstrating a clear need for the development of new therapeutic approaches. The treatment of malignant gliomas is one of the most challenging protocols in oncology. It is well acknowledged that the failure of chemotherapy to treat high-grade gliomas is due to the inability of systemically administered anticancer drugs to reach the brain parenchyma. In the central nervous system

M.A. Hayat (ed.), Tumors of the Central Nervous System, Volume 5, DOI 10.1007/978-94-007-2019-0_21, © Springer Science+Business Media B.V. 2012

187

188

(CNS), the architecture and genetic makeup of the vasculature is such that it protects the brain from noxious substances circulating in the bloodstream. The endothelial cells lining the circulation system found within the CNS evolved a physical barrier that is characterized by tight intercellular junctions and by the absence of fenestrations, which would normally allow therapeutic drugs to be fully permeable. This blood brain barrier (BBB) separates the vascular compartment from the cerebral parenchyma and prevents the uptake of all large-molecule and more than 98% of drugs into the CNS. Malignant gliomas typically cause a local disruption of the BBB, allowing for anticancer drugs to penetrate tumors. However, this disruption of the BBB does not occur in the distantly located healthy CNS tissues surrounding the main tumor where tumor cells characteristically invade and colonize. Tumor cells present within these inaccessible locations represent the source of disease recurrence and are virtually responsible for all fatalities. Clinical success for GBM will require novel technologies capable of bypassing the BBB and delivering therapeutic agents to distantly located tumor cells. There is a need for robust model systems capable of recapitulating these salient features of GBM. Typically, glioma cell culture studies have inherent limitations in modeling cancer. Those include the inability to recapitulate basic principles of invasion, migration, angiogenesis and the roles that the tumor microenvironment plays during those events. Animal model systems are a superior paradigm to study tumor biology and for use in preclinical trials to develop novel treatment modalities for this disease. These systems comprise two basic categories: those that implant tumor cells into recipient mice and those that create tumors in mice de novo in genetically engineered mouse strains. Allograft implantation models are characterized by tumor cells derived from a given species being implanted back into the same species. On the other hand, xenografts models are defined as tumor cells from a different species are implanted into immunosuppressed mice. Xenograft models can either be orthotopic (in original site) or heterotypic (nonautochthonous site). The traditional xenograft models use cell lines that have been cultured in serum condition for decades, and when injected and grown in mice, tend to not show the classical histologic appearance of human gliomas, including a vasculature that is

A. Boskovitz et al.

functionally and architecturally reminiscent of human GBMs. More importantly, they are not predictive of clinical response. The major criticisms associated with these models are the absence of a functioning immune system and the methods by which these tumors are seeded. It is clear that cancer formation and progression engage immune surveillance-escaping mechanisms, which, in an immunocompetent host, select for cancer developmental processes that are not represented in a xenograft system. In addition, injection of a large number of cells into the recipient animals differs considerably from how a spontaneous tumor develops and evolves in situ. Most if not all nanotechnologies tested so far employ xenograft systems. In light of these criticisms and with improved genetic tools at hand, many have turned towards the creation and utilization of model systems that are thought to better represent key aspects of tumor biology, the use of genetically engineered mouse models (GEMMs). These serve as relevant model systems to more accurately study the biological effects of novel bio-nano materials in vivo. Importantly, tumors form within a microenvironment that is identical to that of human tumors and thus develop a vasculature that parallels that of human patients. The development of structures with dimensions ranging in the nanometer scale (10–9 m) that are assembled from various materials has recently generated a revolution in their applications in various fields, including cancer biology. Nanotechnology is a growing field that comprises the development and characterization of man-made materials that are less than 1000 nm in diameter. This size range exceeds that of standard organic molecules but its lower range (<100 nm) approaches that of many proteins and biological macromolecules. Nanotechnology has now found a role in the imaging, diagnosis and treatment of many tumors. In neuro-oncology, advances in nanotechnology have improved the prospects of therapeutic drug delivery to GBM. Nanoparticles (NPs) are designed to overcome some of the drawbacks experienced with conventional methods. Ideally, there are desirable features necessary for a successful nanotechnology-based delivery of therapeutics to GBM. NPs should be designed to 1) preferentially target tumor over normal brain 2) readily cross the BBB and be physically stable in circulation if administered systemically 3) demonstrate a powerful penetration of

21 Nanotechnology-Based Therapy for Malignant Tumors of the Central Nervous System

the parenchyma if administered locally 4) have no toxicity to normal brain tissue, 5) be biodegradable and biocompatible, and 6) be non-immunogenic. In this chapter, we provide a critical review of nanoparticles that are, have been, or could be used with brain tumors. We also discuss the benefits and limitations as well as future directions for these particles in treating GBM.

Nanotechnology for Brain Cancer Nanoparticle Formulations: State of the Art Technology Nanotechnology is a fast growing field that strongly appeals to scientists across broad disciplines because of the potential to engineer various properties that might otherwise be incompatible on a single device. For example, one can engineer attachments such as biologically active molecules, targeting sequences, fluorescent or other imaging components and biocompatible coatings (to list a few) on an iron oxide based core particle. In addition, the design of the particle core structure, the size and its shape provides yet an extra dimension of physical control that can be exploited toward a specific function. NPs are thus varied in their compositions and they are typically categorized on the basis of the physico-chemical properties of their core components.

Inorganic Materials Inorganic nanoparticles can be composed of metals, metal sulfides or oxides, or, more frequently, alumina or silica (Faraji and Wipf, 2009). Nanoparticles of various shape and size have been developed with variation also in their porous characteristics (Faraji and Wipf, 2009). Silica nanocarriers provide a great example: “mesoporous” silica forms an internal complex of channels throughout the particle, whereas silica “nanoshells” contain a central cavity connected to the outside through multiple pores (Faraji and Wipf, 2009). The distribution and size of the pores of the silica-based nanoparticles determine the release kinetics of the incorporated drug (Faraji and Wipf, 2009). Surface pores can be capped with a third agent that will

189

be removed once the particle is in the targeted environment (Faraji and Wipf, 2009). Chemical and magnetic caps have been described, providing additional control of the drug release mechanism (Faraji and Wipf, 2009; Giri et al., 2005; Lai et al., 2003; Radu et al., 2004). Inorganic nanoparticles are overall stable structures, a double sword feature as long-term toxicity of the particle itself in the targeted tissue may result from the absence or extremely slow biodegradation of these carriers (Faraji and Wipf, 2009). The bulk of inorganicbased NPs research for glioma application has focused on the development of contrasting and imaging agent for improvement of MRI modalities. Recently, Pt/TiO2 CNS biocompatible inorganic NPs have shown efficacy in a preclinical rodent model of GBM (Lopez et al., 2008).

Organic Materials Polymeric Nanoparticles Polymeric nanoparticles typically range from 10 to 1000 nm in size. Unlike their inorganic counterparts, most are biocompatible and biodegradable, making them a popular choice of starting building material. Polylactic acid (PLA) and polylactic-coglycolic acid (PLGA) are frequently selected polymers and have been approved by the FDA for human use (Faraji and Wipf, 2009). Other hydrophilic polymers such as chitosan (Calvo et al., 1997), maltodextrin (Major et al., 1997) or gelatin have also been used (Beduneau et al., 2007a; Faraji and Wipf, 2009). Choosing the appropriate polymer and adjusting the drug-to-polymer ratio (Faraji and Wipf, 2009) can modulate drug entrapment and release. Their surface characteristics are strongly compatible with the necessary targeting properties of nanocarriers. Their limiting factor resides in difficult large-scale manufacturing of these polymeric nanoparticles (Faraji and Wipf, 2009). Interestingly, polymeric coating, such as polyethylene glycol (PEG) can be applied onto other polymeric and other nanocarriers in order to improve their biodistribution characteristics (Gref et al., 1994). An example of a polymeric NPs application is the use of PLGA-based polymeric NPs as a vector for gene therapy in breast cancer cells in vitro (Faraji and Wipf, 2009; Prabha and Labhasetwar, 2004).

190

Polymeric Micelles Polymeric micelles assemble from a spontaneous association of amphiphilic copolymers in an aqueous phase. They are typically characterized by a diameter that does not exceed 100 nm. The hydrophobic component of the amphiphillic copolymer is usually made of a polyester (PLA), a poly-amino acid (poly-aspartic acid (PAsp), poly-glutamic acid) (PGlu)), or a phospholipid (phosphatidylethanolamine (PE)) (Beduneau et al., 2007a). The hydrophilic moiety is generally composed of polyethers such as poly(propylene oxide) or poly(ethylene glycol) (PEG) or a molecular mass of <15 kDa. Most of them are biocompatible and biodegradable. Various anticancer agents such as paclitaxel, camptothecin and doxorubicin have been successfully incorporated into the hydrophobic core of polymeric micelles (Nakanishi et al., 2001; Nakayama et al., 2006; Watanabe et al., 2006).

A. Boskovitz et al.

They can hold hydrophilic compounds within their core, or hydrophobic ones out on the hydrophobic aspect of the phospholipid layer (Faraji and Wipf, 2009). They will attach to cell membranes and/or be internalized by endocytosis (Faraji and Wipf, 2009). Initially designed in the 1960s, several have been approved by the FDA since the early 1990s for use as drug delivering agents in indications ranging from metastatic breast cancer, acute lymphocytic leukemia, to wet macular degeneration, hepatitis C or fungal infections. However, they have yet to demonstrate an actual significant impact on the treatment of those diseases, which is impaired by a relatively rapid elimination from the circulation (Faraji and Wipf, 2009). This can be at least partially circumvented by pegylation of the nano liposomal carriers (Faraji and Wipf, 2009). An early study using pegylated liposomal doxorubicin in patients with high-grade gliomas has shown results suggesting it is a safe drug administration method, but significant improvement in the treatment of the disease is yet to be proven (Hau et al., 2004)

Lipid Nanoparticles Lipid NPs fall into two categories, solid lipid NPs and lipid nanocapsules. Solid lipid NPs consist of a solid matrix of hydrophobic lipids, such as triglycerides or other complex glycerides that are stabilized by an outer layer of surfactants (Wissing et al., 2004). They are typically stable at both room and body temperatures and remain in solid form at those temperatures. Conversely, lipid nanocapsules are composed of a liquid to oily core surrounded by hydrophilic (PEG) and lipophilic (phosphatidylethanolamine and phosphatidylcholine) surfactants (Heurtault et al., 2002). Various anticancer drugs (etoposide, docetaxel and paclitaxel) have been encapsulated into lipid nanocapsules and drug-release pharmacokinetics is adjustable (Beduneau et al., 2007b). Assessing the potential of solid lipid nanoparticles for the treatment of brain tumors, pegylated solid lipid NPs loaded with paclitaxel were administered into the brain of naïve rats, showing a significant increase in brain tissue uptake of paclitaxel (Koziara et al., 2004).

Liposomes Liposomes are made of concentric bilayers of phospholipids (unlike micelles, made of a lipid monolayer).

Delivery of Nanomaterials: Systemic Versus Local Systemic Delivery Following systemic administration of drugs, the body normally distributes drugs via the vascular system. Systemic intravenous delivery of NPs however triggers their elimination by the reticuloendothelial system (RES). This system is part of the general immune response responsible for absorbing and eliminating circulating foreign particles. It is mainly composed of monocytes and macrophages that readily recognize and absorb nanomaterials and then accumulate in lymph nodes and spleen for further processing. This process is facilitated by surface deposition of opsonic factors and complement proteins on the NPs themselves. Interestingly, both clearance and opsonization are influenced by the size and surface characteristics of the NPs. For examples, NPs greater than 200 nm activate the complement system more efficiently and are cleared more rapidly than smaller NPs. The geometry, charge, and functional groups on the surface of NPs are thought to mediate the binding to plasma proteins and blood poisonings. NPs must therefore evade the RES to effectively function as drug delivery agents.

21 Nanotechnology-Based Therapy for Malignant Tumors of the Central Nervous System

Pegylation of NPs with a hydrophilic surfactant, for the most part using glycol PEG derivatives, dramatically decreases their immunogenicity as well as their removal from the circulation by the RES, increasing their circulating half-life and therefore bioavailability for transport across the BBB (Beduneau et al., 2007a). Lipid nanocapsules coated with hydrophilic PEG660 hydroxystearate had an extended circulating half-life following intravenous injection as a result of the antiopsonization effect of the surfactant (Beduneau et al., 2007a; Hoarau et al., 2004). It has become clear that increased circulation time is necessary for proper NP tissue distribution. Systemic treatments targeting brain tumors are faced by an additional and unique challenges: crossing the BBB. The latter will prevent the passage of 98% of pharmaceutical small molecule drugs, allowing only lipophilic, electrically neutral small molecules (<5000 Da) to travel through it.

Methods to Increase Targeting Specificity of the Particles Unlike pegylation, which results in non-specific increased drug uptake by the brain parenchyma, attachment of tissue-specific ligands to the outer surface of a NP will help direct it to receptors or antigens located on the luminal surface of the endothelial cells’ of the brain capillaries and optimize transport through the BBB. Nutrients and peptides have been used to that extent (Beduneau et al., 2007a). When contemplating intraparenchymal or intratumoral administration, an entirely different selection of possible ligands is available, including tumor-specific monoclonal antibodies. An example is the synthetic low-density lipoproteins (LDL). Maletinska et al. have shown that GBM cells overexpress LDL receptors (LDLR) as compared to normal brain cells (Maletinska et al., 2000; Pitas et al., 1987). LDL, the primary ligand for LDLR, is a 22–27 nm particle composed of a surface containing phospholipids, unesterified cholesterol and the apolipoprotein B-100. The core is made of hydrophobic lipids and can contain hydrophobic drugs (Nikanjam et al., 2007a, b). The apolipoprotein B-100 binds LDLR, triggering endocytosis of the LDL particle, which is then degraded within the lysosomal compartment and the content of the lipophilic core is released into the cytoplasm (Goldstein et al.,

191

1985). In recent in vitro studies, LDL nanoparticles loaded with Paclitaxel-oleate (the lipophilic version of Paclitaxel) were uptaken by GBM cells and demonstrated a dose-dependent cell killing capability (Nikanjam et al., 2007a, b). Others studied polysorbate 80-coated poly(butyl cyanoacrylate) (PBCA) nanoparticles loaded with doxorubicin in naïve rats, showing significantly increased accumulation of the drug within the brain parenchyma following intravenous administration and in comparison to free doxorubicin or doxorubicin loaded within uncoated PBCA nanoparticles (Beduneau et al., 2007a; Gulyaev et al., 1999). In further studies, authors reported an increase in median survival of GBM xenograft-bearing rats following intravenous administration of polysorbate 80-coated, doxorubicin-loaded PBCA nanoparticles (Beduneau et al., 2007a; Steiniger et al., 2004). Apolipoprotein Apia-I, found in considerable amounts on the surface of both nanoparticles described above, could be the key factor in this increase in therapeutic efficiency, possibly by increasing transportation of the nanoparticles through the BBB. Although the mechanism is not fully elucidated yet, it is interesting that drug-free surfactant-coated PBCA nanoparticles showed a different pattern of transport and uptake, suggesting that, not only the surfactant but also the loaded drug will influence surface characteristics of the nanoparticles and thus their interaction with the BBB or targeted cells (Petri et al., 2007). Other authors studied Doxorubicin-loaded, PEGcoated poly(hexadecylcyanoacrylate) (PHDCA) nanoparticles in intracranial 9-L gliosarcoma xenograft-bearing rats (Brigger et al., 2004). Prior biodistribution studies of the unloaded nanoparticles had demonstrated significant intratumoral accumulation (Brigger et al., 2002). Encapsulation of Doxorubicin reduced systemic toxicity of the drug but failed to translate into any therapeutic advantage as the drug-loaded nanoparticles accumulated in the lungs and spleen rather than into the intracranial xenografts. These results were attributed to the positive surface charge conferred on the nanoparticles by Doxorubicin, resulting in aggregation with circulating plasma proteins and formation of larger compounds retained in the lung capillaries and eventually filtered in the spleen (Brigger et al., 2004). This further reinforces the concept that pharmacokinetics of drug-loaded nanoparticles depend not only on the nature of the nanoparticle or its coating, but also on the drug it

192

carries. Successful research has also been carried out with non-viral gene therapy used on human glial brain tumors implanted in mice using immunoliposomes conjugated to two different antibodies and carrying nonviral expression plasmid encoding antisense mRNA against the human epidermal growth factor receptor gene (EGFR) (Zhang et al., 2002).

Strategies to Overcome the Blood Brain Barrier Approaches to improve the delivery of a drug beyond the BBB include lipidization of the molecule, though a lipidized drug or pro-drug may face increased nonspecific binding by plasma proteins or uptake by the RES. Mannitol and other hyperosmolar agents have been used to increase the BBB permeability by temporary disruption of the endothelial cells’ tight junctions. NPs have a potential role as carriers to transport and deliver the drug across the BBB, either by lipophilic properties, and/or by using ligand-mediated transport through the BBB, a mechanism yet only partially studied and thought to occur through either adsorptivemediated endocytosis (AME), carrier-mediated transport (CMT) or receptor-mediated endocytosis (RME). CMT systems are functioning on the surface of endothelial cells to allow for the delivery of nutrients to the brain parenchyma. These nutrients are vital for the survival of the brain and it is not surprising that CMT systems are highly expressed on CNS endothelial cells. This offers unique opportunities to camouflage NPs with substrates for these transporters. For example, the glucose transporter GLUT1 is mainly expressed on the luminal surface of brain capilaries and promotes the transport of D-glucose from the circulation to the brain. It also mediates the passage through the BBB of substance that exhibit similar structures (e.g. 2-deoxyglucose, galaxies, mannose, or glucose analogs). Liposomes were synthesized with mannose-derivatives on their surface and shown to reach the brain in a mouse model by crossing the BBB via the glucose transporter (Umezawa and Eto, 1988). RME systems are defined as the binding of a ligand to a cognate receptor on the luminal surface of the BBB. Upon binding, the receptor-ligand complexes are internalized and undergo further processing including lysosomal degradation, endosomal recycling or exocytosis. Over the years, transcytosis has been used by many groups that have used endogenous serum/blood

A. Boskovitz et al.

ligands (insulin, transferrin or folate) to target NPs through the BBB using their high affinity receptors on the surface of capillary endothelial cells.

Local Administration Through Convection Enhanced Delivery Bypassing the BBB by administering the drug compound directly into the brain parenchyma or the tumor core eliminates the restrictions imposed by the BBB permeability. Nonetheless, it is more invasive than a systemic intravenous injection and it is met with other challenges such as local toxicity, pattern of diffusion through the targeted tissue, half-life and elimination kinetics, as well as systemic toxicity upon resorption into the systemic circulation. Convection-enhanced delivery (CED) is a pressure-driven method of direct, intraparenchymal or intratumoral infusion. The administered products diffuse throughout the interstitium at a rate and distance that will depend on the characteristics of the compound and the nature of the targeted tissue. Liposomal nanoparticles have been administered via CED (Jain, 2007): liposomes containing gadolinium and doxorubicin where infused into brain tumorbearing rats, allowing for MRI-based in vivo monitoring of the diffusion and distribution of the drug-loaded nanoparticles (Jain, 2007). Liposomal nanoparticles containing CPT-11 (Irinotecan) have been administered using CED in naïve and glioma xenograftbearing rats. The half-life of CPT-11 in the normal brain parenchyma was significantly increased from 0.3 days (free CPT-11) to 6.7–19.7 days, depending on the initial dosage. CNS toxicity was found following low dose infusion of free CPT-11, whereas nanoparticles loaded with up to 4-times that dose failed to reveal any toxicity. U87 glioma xenografts-bearing animals had an extended median survival of more than 100 days when treated with CPT-11-loaded nanoparticles. In comparison, animals treated with control nanoparticles or free CPT-11 had a median survival of 19.5 and 28.5 days, respectively (Jain, 2007; Noble et al., 2006). Recently, we have successfully delivered a siRNA against the EGF receptor using an iron oxide corebased NP coated with G4 dendrimers in a genetically engineered mouse model of GBM (Agrawal et al., 2009). These proof of concept experiments demonstrated that RNA interference-mediated gene knock down are feasible in vivo and that nanocarriers offer an

21 Nanotechnology-Based Therapy for Malignant Tumors of the Central Nervous System

unprecedented opportunity to push forward this unique and utmost targeted type of therapeutic agent. The local delivery of drug-loaded NPs to brain tumors by CED offers a better drug distribution as compared to other strategies that are governed by diffusion only and appears to be very promising for the treatment of brain malignancies.

Discussion It is becoming increasingly evident that targeted nanocarrier platforms are a promising tool to deliver nucleic acid-based molecules and drugs to CNS tumors. By modifying the surface of NPs with various targeting molecules or charge and chemical composition, NPs can be engineered to evade the immune system and stealthily target specifically to tumor cells. Systemic delivery of NPs for CNS targeting necessitates modifications to coax endogenous natural transport systems to uptake NPs through the BBB. The strategy of conjugating site-specific ligands to NPs typically leads to a significant improvement in the level of CNS accumulation. Although numerous reports describe the potential of NPs for the treatment of malignant glioma, only a few studies have clearly demonstrated therapeutic effects in pre-clinical models (Steiniger et al., 2004; Zhang et al., 2002). These nanoplatforms need to be further optimized before their clinical utilization. There is also a need to better define the influence of some parameters such as size, polymer type, charge and lipophilicity/hydrophilicity character of the NP surface on the behavior of the NP in vivo. There is also a pressing lack in our understanding of the internalization of NPs, their passage through the brain capillary endothelial cells and their movement within the brain parenchyma. These parameters need to be established in accurate and relevant animal models before they can be applied clinically.

References Agrawal A, Min DH, Singh N, Zhu H, Birjiniuk A, von Maltzahn G, Harris TJ, Xing D, Woolfenden SD, Sharp PA, Charest A, Bhatia S (2009) Functional delivery of siRNA in mice using dendriworms. ACS Nano 3:2495–2504

193

Beduneau A, Saulnier P, Benoit JP (2007a) Active targeting of brain tumors using nanocarriers. Biomaterials 28:4947–4967 Beduneau A, Saulnier P, Hindre F, Clavreul A, Leroux JC, Benoit JP (2007b) Design of targeted lipid nanocapsules by conjugation of whole antibodies and antibody Fab’ fragments. Biomaterials 28:4978–4990 Brigger I, Morizet J, Aubert G, Chacun H, Terrier-Lacombe MJ, Couvreur P, Vassal G (2002) Poly(ethylene glycol)coated hexadecylcyanoacrylate nanospheres display a combined effect for brain tumor targeting. J Pharmacol Exp Ther 303:928–936 Brigger I, Morizet J, Laudani L, Aubert G, Appel M, Velasco V, Terrier-Lacombe MJ, Desmaele D, d’Angelo J, Couvreur P, Vassal G (2004) Negative preclinical results with stealth nanospheres-encapsulated Doxorubicin in an orthotopic murine brain tumor model. J Control Release 100:29–40 Calvo P, Remunan-Lopez C, Vila-Jato JL, Alonso MJ (1997) Chitosan and chitosan/ethylene oxide-propylene oxide block copolymer nanoparticles as novel carriers for proteins and vaccines. Pharm Res 14:1431–1436 Faraji AH, Wipf P (2009) Nanoparticles in cellular drug delivery. Bioorg Med Chem 17:2950–2962 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 Giri S, Trewyn BG, Stellmaker MP, Lin VS (2005) Stimuliresponsive controlled-release delivery system based on mesoporous silica nanorods capped with magnetic nanoparticles. Angew Chem Int Ed Engl 44:5038–5044 Goldstein JL, Brown MS, Anderson RG, Russell DW, Schneider WJ (1985) Receptor-mediated endocytosis: concepts emerging from the LDL receptor system. Annu Rev Cell Biol 1:1–39 Gref R, Minamitake Y, Peracchia MT, Trubetskoy V, Torchilin V, Langer R (1994) Biodegradable long-circulating polymeric nanospheres. Science 263:1600–1603 Gulyaev AE, Gelperina SE, Skidan IN, Antropov AS, Kivman GY, Kreuter J (1999) Significant transport of doxorubicin into the brain with polysorbate 80-coated nanoparticles. Pharm Res 16:1564–1569 Hau P, Fabel K, Baumgart U, Rummele P, Grauer O, Bock A, Dietmaier C, Dietmaier W, Dietrich J, Dudel C, Hubner F, Jauch T, Drechsel E, Kleiter I, Wismeth C, Zellner A, Brawanski A, Steinbrecher A, Marienhagen J, Bogdahn U (2004) Pegylated liposomal doxorubicin-efficacy in patients with recurrent high-grade glioma. Cancer 100:1199–1207 Heurtault B, Saulnier P, Pech B, Proust JE, Benoit JP (2002) A novel phase inversion-based process for the preparation of lipid nanocarriers. Pharm Res 19:875–880 Hoarau D, Delmas P, David S, Roux E, Leroux JC (2004) Novel long-circulating lipid nanocapsules. Pharm Res 21: 1783–1789 Jain KK (2007) Use of nanoparticles for drug delivery in glioblastoma multiforme. Expert Rev Neurother 7:363–372 Koziara JM, Lockman PR, Allen DD, Mumper RJ (2004) Paclitaxel nanoparticles for the potential treatment of brain tumors. J Control Release 99:259–269 Lai CY, Trewyn BG, Jeftinija DM, Jeftinija K, Xu S, Jeftinija S, Lin VS (2003) A mesoporous silica nanosphere-based carrier

194 system with chemically removable CdS nanoparticle caps for stimuli-responsive controlled release of neurotransmitters and drug molecules. J Am Chem Soc 125:4451–4459 Lopez T, Recillas S, Guevara P, Sotelo J, Alvarez M, Odriozola JA (2008) Pt/TiO2 brain biocompatible nanoparticles: GBM treatment using the C6 model in Wistar rats. Acta Biomater 4:2037–2044 Major M, Prieur E, Tocanne JF, Betbeder D, Sautereau AM (1997) Characterization and phase behaviour of phospholipid bilayers adsorbed on spherical polysaccharidic nanoparticles. Biochim Biophys Acta 1327:32–40 Maletinska L, Blakely EA, Bjornstad KA, Deen DF, Knoff LJ, Forte TM (2000) Human glioblastoma cell lines: levels of low-density lipoprotein receptor and low-density lipoprotein receptor-related protein. Cancer Res 60:2300–2303 Nakanishi T, Fukushima S, Okamoto K, Suzuki M, Matsumura Y, Yokoyama M, Okano T, Sakurai Y, Kataoka K (2001) Development of the polymer micelle carrier system for doxorubicin. J Control Release 74:295–302 Nakayama M, Okano T, Miyazaki T, Kohori F, Sakai K, Yokoyama M (2006) Molecular design of biodegradable polymeric micelles for temperature-responsive drug release. J Control Release 115:46–56 Nikanjam M, Blakely EA, Bjornstad KA, Shu X, Budinger TF, Forte TM (2007a) Synthetic nano-low density lipoprotein as targeted drug delivery vehicle for glioblastoma multiforme. Int J Pharm 328:86–94 Nikanjam M, Gibbs AR, Hunt CA, Budinger TF, Forte TM (2007b) Synthetic nano-LDL with paclitaxel oleate as a targeted drug delivery vehicle for glioblastoma multiforme. J Control Release 124:163–171 Noble CO, Krauze MT, Drummond DC, Yamashita Y, Saito R, Berger MS, Kirpotin DB, Bankiewicz KS, Park JW (2006) Novel nanoliposomal CPT-11 infused by convectionenhanced delivery in intracranial tumors: pharmacology and efficacy. Cancer Res 66:2801–2806

A. Boskovitz et al. Petri B, Bootz A, Khalansky A, Hekmatara T, Muller R, Uhl R, Kreuter J, Gelperina S (2007) Chemotherapy of brain tumour using doxorubicin bound to surfactant-coated poly(butyl cyanoacrylate) nanoparticles: revisiting the role of surfactants. J Control Release 117:51–58 Pitas RE, Boyles JK, Lee SH, Hui D, Weisgraber KH (1987) Lipoproteins and their receptors in the central nervous system. Characterization of the lipoproteins in cerebrospinal fluid and identification of apolipoprotein B,E(LDL) receptors in the brain. J Biol Chem 262:14352–14360 Prabha S, Labhasetwar V (2004) Nanoparticle-mediated wildtype p53 gene delivery results in sustained antiproliferative activity in breast cancer cells. Mol Pharm 1:211–219 Radu DR, Lai CY, Jeftinija K, Rowe EW, Jeftinija S, Lin VS (2004) A polyamidoamine dendrimer-capped mesoporous silica nanosphere-based gene transfection reagent. J Am Chem Soc 126:13216–13217 Steiniger SC, Kreuter J, Khalansky AS, Skidan IN, Bobruskin AI, Smirnova ZS, Severin SE, Uhl R, Kock M, Geiger KD, Gelperina SE (2004) Chemotherapy of glioblastoma in rats using doxorubicin-loaded nanoparticles. Int J Cancer 109:759–767 Umezawa F, Eto Y (1988) Liposome targeting to mouse brain: mannose as a recognition marker. Biochem Biophys Res Commun 153:1038–1044 Watanabe M, Kawano K, Yokoyama M, Opanasopit P, Okano T, Maitani Y (2006) Preparation of camptothecin-loaded polymeric micelles and evaluation of their incorporation and circulation stability. Int J Pharm 308:183–189 Wissing SA, Kayser O, Muller RH (2004) Solid lipid nanoparticles for parenteral drug delivery. Adv Drug Deliv Rev 56:1257–1272 Zhang Y, Zhu C, Pardridge WM (2002) Antisense gene therapy of brain cancer with an artificial virus gene delivery system. Mol Ther 6:67–72

Chapter 22

Pilocytic Astrocytoma: Pathological and Immunohistochemical Factors Affecting Surgical Treatment and Surveillance Devon Haydon and Jeffrey Leonard

Abstract Tumors of the central nervous system are the most prevalent solid neoplasm in the pediatric population and represent a considerable burden to children, physicians, and the healthcare system. Highgrade tumors frequently require surgical resection, followed by aggressive adjuvant therapy. Low-grade lesions, in contrast, are often controlled with surgical resection alone, though adjuvant treatment modalities may be necessary if the location of the tumor prevents complete surgical removal. Pilocytic astrocytomas (PAs) are low-grade central nervous system neoplasms that typically affect children. Many PAs are cured with resection alone; others frequently recur despite aggressive management. The ability to distinguish these two groups would be valuable for the treating clinician. Age and extent of resection have long been recognized as two important prognostic variables concerning low-grade, pediatric gliomas (Garcia et al. (J Neurosurg 71:661–664, 1989); Pollack et al. (J Neurosurg 82:536–547, 1995); Gajjar et al. (J Clin Oncol 15:2792–2799, 1997); Fisher et al. (Pediatr Blood Cancer 51:245–250, 2008)). However, additional prognostic markers are necessary in the ongoing effort to correctly identify tumors at particular risk for recurrence or progression. Radiographic characteristics do not adequately predict tumor behavior. Therefore, increasing efforts have been directed toward tumor histology and immunohistochemistry in order to uncover features predictive of aggressive clinical behavior.

J. Leonard () Department of Neurosurgery, Washington University School of Medicine, St. Louis, MO 63110, USA e-mail: [email protected]

Keywords Astrocytomas · Brain tumor · Pediatric

Introduction Pilocytic astrocytomas are predominantly benign, slow-growing, well-circumscribed neoplasms that represent the single most common brain tumor in children under fifteen. As WHO grade I tumors, PAs have a low proliferative potential with long-term survival being the rule rather than the exception. According to the Central Brain Tumor Registry of the United States (CBTRUS; http://www.cbtrus.org), the incidence rate for PAs is 0.33 per 100,000 person years. Although PAs may occur at any point during an individual’s lifetime, they are primarily diagnosed during the first and second decades of life and effect males and females equally. Pilocytic astrocytomas are most commonly found in the cerebellum, optic pathways, and hypothalamic/3rd ventricular region. They can also occur in the thalamus, basal ganglia, brainstem, and spinal cord. Symptoms at clinical presentation of a PA depend on the tumor’s location. Cerebellar lesions often present from symptoms of increased intracranial pressure and hydrocephalus, specifically headache and nausea/vomiting. Disequilibrium and incoordination may also be apparent from cerebellar dysfunction. PAs of the optic pathways typically involve vision loss, but more indolent signs like painless proptosis may be the only presenting sign. Hypothalamic/3rd ventricular involvement commonly causes obstructive hydrocephalus. Additionally, lesions in this location can cause various endocrinopathies from pituitary compression and

M.A. Hayat (ed.), Tumors of the Central Nervous System, Volume 5, DOI 10.1007/978-94-007-2019-0_22, © Springer Science+Business Media B.V. 2012

195

196

dysfunction. In some cases, diencephalic syndrome can develop which is marked by cachexia, mood disturbance, and an overall failure to thrive (Huber et al., 2007). Focal neurological signs are more common with thalamic and basal ganglia PAs. Seizures, although possible, do not usually occur with PAs unless they are cortically-based.

Treatment Strategies Once an intracranial lesion concerning for PA has been identified, surgery is the first necessary treatment modality. The primary surgical objective is accurate pathological diagnosis. When present, obstructive hydrocephalus is also alleviated via removal of the obstructing mass or cerebrospinal fluid diversion. For PAs, gross total resection should be the goal of any surgical resection whenever anatomically feasible, as complete removal offers the greatest chance for cure (Fisher et al., 2001). In cases where gross total resection is accomplished, no further adjuvant therapy is necessary, and only follow-up serial imaging is required. The complexity in PA management arises when lesions are incompletely resected and when pathology reveals a higher-grade variant. Not every PA is amenable to gross total resection due to proximity to eloquent areas of the brain, such as the hypothalamus. Yet, even a percentage of completely resected, classic PAs recur overtime thereby necessitating additional treatment (Krieger et al., 1997; Benesch et al., 2006). If a recurrent/progressive PA is surgically accessible, an attempt at repeat resection is offered. However, residual or recurrent disease and higher-grade pathology introduce the need for adjuvant treatment such as chemotherapy and/or radiation. One of the main challenges faced by clinicians treating PAs is the ability to identify those lesions that are more likely to progress. Numerous researchers have examined both imaging and pathological characteristics of PAs in an attempt to further predict their clinical behavior.

Imaging Characteristics Because of its wide availability, computed tomography (CT) remains an early imaging modality used to assess intracranial pathology. Cystic components of PAs are

D. Haydon and J. Leonard

easily delineated on CT. The solid tissue component is typically iso- or hypo-dense compared to the surrounding normal parenchyma. Calcifications are less common but readily apparent on CT. Often, obstructive hydrocephalus is the presenting CT finding. Pilocytic astrocytomas are usually avidly enhancing following intravenous contrast administration as a result of their rich vascularity. Magnetic resonance imaging (MRI) is the radiographic modality of choice when evaluating intracranial, neoplastic disease. Pilocytic astrocytomas are well-circumscribed, iso- or hypo-intense lesions compared to adjacent brain tissue. Extensive peri-lesional edema, assessed best by T2 -weighted sequences, is not typically seen with PAs. Analogous to their appearance on CT, PAs show significant enhancement on MRI following gadolinium contrast administration (Fig. 22.1). Even an associated cyst’s wall can display bright contrast enhancement. Cyst wall resection is generally not recommended because the enhancing area is felt to be reactive rather than neoplastic tissue (BeniAdani et al., 2000). Cerebrospinal fluid pathway dissemination is also seen in rare cases (Faria et al., 2006). Although clearly useful when formulating a differential diagnosis and preoperative surgical plan, imaging studies alone cannot diagnose PAs with clinically acceptable specificity. Even in the setting of known pathology, CT and MRI cannot reliably predict tumor behavior. Strong et al. (1993) compared the CT and MRI characteristics of biopsy proven PA in an attempt to predict clinically aggressive lesions. Location, size, calcification, tumor morphology, and degree of contrast enhancement were all poor predictors of clinical outcome (Strong et al., 1993). Therefore, increasing work has focused on the pathological characteristics of PAs in an attempt to predict tumor behavior and overall prognosis.

Classic Pilocytic Astrocytoma Grossly, PAs are soft, gray tumors with well-defined margins bordering adjacent parenchyma. Cysts are frequently associated with PAs. In fact, the size of the actual neoplastic tissue may be quite small compared to the overall size of the associated cyst. This size differential often means that the mass effect of the cyst is responsible for clinical presentation.

22 Pathological and Immunohistochemical Factors Affecting Surgical Treatment and Surveillance

A

197

B

C

Fig. 22.1 Pilocytic astrocytomas arise in both the infratentorial and supratentorial regions. They usually enhance on MRI following contrast and are frequently associated with a cyst. Radiographic features cannot accurately identify specific pilocytic variants or high risk lesions given sizeable overlap in

appearance. Panels illustrate a classic pilocytic astrocytoma (a), anaplastic pilocytic astrocytoma (b), and pilomyxoid astrocytoma (c). Note the unusual example of disseminated disease at presentation with the pilomyxoid astrocytoma (arrows)

Pilocytic astrocytomas consist of relatively dense regions of bipolar astrocytes (Fig. 22.2a) intermixed with loosely-arranged, multipolar cells and microcystic changes (Louis et al., 2007). Rosenthal fibers are often associated with the bipolar cell populations while eosinophilic granular bodies typically occur in the multipolar cell regions. Rosenthal fibers are eosinophilic, intracytoplasmic inclusions with characteristic tapered-whorl morphology. Although they are commonly seen in PAs, their presence is not necessary for diagnosis. In fact, Rosenthal fibers occur with other tumors as well as non-neoplastic conditions where astrocytes undergo reactive gliosis. Similarly, eosinophilic granular bodies, another intracytoplasmic inclusion, are present in multiple pathologic conditions and are helpful, but not critical, for diagnosis.

Pilocytic Astrocytoma Variants Two important pilocytic variants, anaplastic pilocytic astrocytoma (Fig. 22.2b) and pilomyxoid astrocytoma (Fig. 22.2c), deserve specific consideration given their markedly aggressive clinical behavior compared to classic PA. Anaplasia has long been reported in the context of PAs (Dirks et al., 1994; Tomlinson et al., 1994). However, anaplasia has been variously defined in multiple studies with limited consistency between observers. Recently, Rodriguez et al. (2010) analyzed tumors for at least 4 mitoses per 10 high power fields, hypercellularity, cytological atypia, and the presence or absence of necrosis. Patients with an anaplastic pilocytic astrocytoma tended to be older (median age at diagnosis: 35) and harbored a PA precursor lesion in

198

D. Haydon and J. Leonard

B

A

C Fig. 22.2 Dense regions of classic pilocytic astrocytomas (a) typically demonstrate hyperchromatic, elongate-oval nuclei with inconspicuous nucleoli. Eosinophilic granular bodies (arrow) are variously interspersed. Anaplastic pilocytic astrocytomas (b) display round to oval nuclei with an increased

nucleus to cytoplasm ratio. Mitotic figures (arrow) are frequently encountered. Pilomyxoid astrocytomas (c) exhibit round, vesicular nuclei and piloid, cytoplasmic processes within a myxoid background

29% of cases (Rodriguez et al., 2010). Most striking was the dramatically worse overall survival for PAs with these anaplastic features. The survival for an anaplastic pilocytic astrocytoma with necrosis is significantly worse than a WHO grade II diffuse astrocytoma, no different than a WHO grade III astrocytoma, but statistically better than a WHO grade IV glioblastoma (Rodriguez et al., 2010). Pilomyxoid astrocytomas were first described by Tihan et al. (1999) and exhibit a distinct monomorphous histology with a higher recurrence rate. These tumors are now recognized as WHO grade II lesions. Pilomyxoid astrocytomas typically occur in younger children and are more likely located in the chiasmatic/hypothalamic region. They have far fewer eosinophilic granular bodies, and Rosenthal fibers are absent. In the initial series, pilomyxoid astrocytomas showed a mere 38.7% 1 year progression-free-survival compared to 69.2% in age- and location-matched controls (Tihan et al., 1999). More recent studies corroborate the existence of the pilomyxoid astrocytoma as a distinct clinicopathologic tumor type and offer new insight into its origin. Ceppa

et al. (2007) presents intriguing evidence for a possible maturation model in which a pilomyxoid astrocytoma represents an early, undifferentiated form of classic PA. Their report follows the case of a 6-year-old girl whose pathological specimens were extensively evaluated over four serial resections/biopsies. They chronicle the tumor’s loss of a myxoid background with replacement by the characteristic biphasic pattern seen in classic PA with Rosenthal fibers and eosinophilic granular bodies.

Histological Analyses While the majority of PAs exhibit indolent behavior, some PAs progress despite maximal treatment. In an attempt to identify these more aggressive tumors, numerous studies have aimed at classifying various histological markers that confer true prognostic significance. Such markers would be especially useful since PAs are commonly located in areas where gross total resection is difficult, and known residual disease will need to be followed by serial imaging.

22 Pathological and Immunohistochemical Factors Affecting Surgical Treatment and Surveillance

The cytological and histological appearance of PA is generally quite bland, consistent with an overall benign clinical course. Yet even more alarming features like leptomeningeal infiltration are not uncommon findings in classic PA. However, such characteristics are not consistently associated with a worse clinical course, and therefore should not affect treatment decisions until studies can detail their significance. A recent study by Horbinski et al. (2010) even demonstrated a surprising protective effect for leptomeningeal invasion in cerebellar PAs. Paixão Becker et al. (2010) analyzed many histological characteristics of PAs including classic PA features as well as nuclear pleomorphism, mitotic activity, necrosis, microvascular proliferation, presence of ganglion cells, calcification, lymphocytic infiltration, and adjacent parenchymal infiltration. None of these pathological features yielded prognostic information regarding the tumor’s clinical behavior. By contrast, Tibbetts et al. (2009) recently identified four pathological features from a group of 107 sporadic PAs that are significantly associated with event-free survival. Tumors with vascular hyalinization, calcification, areas of necrosis, and oligodendroglioma-like features were all more likely to recur, progress, or result in the mortality of the patient (Tibbetts et al., 2009). Horbinski et al. (2010) did not confirm the poor prognostic implications of calcification and necrosis; however, oligodendroglial morphology was also associated with a poor outcome in their cohort of 147 PAs when specifically located in the cerebellum. Vascular hyalinization and calcification are somewhat surprising features to portend poor prognosis since these characteristics are generally regarded as markers of disease longevity. Horbinski et al. (2010) presented the alternative notion that vascular hyalinization represented a favorable trait as it was a marker of degenerative atypia. Interesting, others have shown no relationship between macroscopic calcification noted on preoperative imaging studies and clinical course (Fernandez et al., 2003). Whether or not variations between macroscopic and microscopic calcifications represent differences in tumor biology that might then affect clinical outcomes remains to be seen. Although others failed to show a prognostic value for the presence of necrosis, it is possible that a limited number of patients and/or prevalence of more aggressive pilocytic variants precluded statistically significant results (Fernandez et al., 2003; Paixão Becker

199

et al., 2010). Tibbetts et al. (2009) emphasized that only infarct-like necrosis was included in their analysis, as palisading and pseudopalisading necrosis is typically seen in higher grade astrocytic tumors, including anaplastic pilocytic astrocytomas. Oligodendroglioma-like features (Fig. 22.3d) raise important questions concerning a PA’s potential cell of origin and gliomagenesis. Like other glial neoplasms, oligodendrogliomas are diagnosed largely by histological appearance and cytological similarities to mature oligodendrocytes. However, several immunohistochemical and genetic markers of mature oligodendrocytes are absent in oligodendrogliomas, thereby pointing to a more primitive cell of origin. Bouvier et al. (2003) evaluated several glial neoplasms for the presence of OLIG1 and OLIG2 gene expression. OLIG1 and OLIG2 encode transcription factors previously identified in oligodendrocyte progenitor cells. While OLIG1/2 gene expression was observed in multiple glial tumors, their expression was highest in oligodendrogliomas and PAs (Bouvier et al., 2003). In a similar study, Bannykh et al. (2006) examined SOX10 expression in multiple glial tumors. SOX10 encodes a transcription factor essential to myelin production. As such, it is critical for the differentiation of oligodendrocytes in the central nervous system. Like the OLIG1 and OLIG2 genes, the expression of SOX10 was highest in oligodendrogliomas and PAs (Bannykh et al., 2006). Additional details concerning histological and genetic similarities between oligodendrogliomas and PAs await further study but include the possibility of a common cell of origin during gliomagenesis. Such similarities might then explain the worse clinical course of PAs with oligodendroglioma morphology compared to more classic lesions.

Immunohistochemistry A number of immunohistochemical markers have been examined in PA samples in the hopes of acquiring relevant information for both diagnosis and prognosis. To date, few markers have revealed such information. Pilocytic astrocytomas are principally diagnosed through characteristic histopathology seen under the microscope. Admittedly, this diagnosis can be challenging given a PA’s significant overlap in appearance with other glial tumors. No immunohistochemical

200

D. Haydon and J. Leonard

A

B

C

D

Fig. 22.3 Immunohistochemical staining for Ki-67 shows scant positivity for classic pilocytic astrocytomas (a). Anaplastic pilocytic astrocytomas (b) and pilomyxoid astrocytomas (c) reveal higher proliferative indices. Oligodendroglial morphology (d) is

a feature of some pilocytic astrocytomas that has been recently associated with a worse clinical course. Note the characteristic perinuclear halos (arrow)

marker has been identified to date that can specifically diagnose PAs. Following PA diagnosis, immunohistochemical analyses are useful when attempting to predict tumor behavior. Specifically, information regarding progression-free-survival can be obtained from targeted staining for markers of proliferative index and oligodendroglial differentiation. In general, few mitoses are found in PA samples, again highlighting their low proliferative potential. Multiple methods have been employed to estimate the proliferative activity of PAs in an attempt to indentify more clinically aggressive tumors. Dirven et al. (1999) examined the staining patterns of nucleolar organizer regions (NORs) in PA samples. These stretches of DNA encode rRNA whose expression is considered an indirect indicator of proliferative potential. They found that higher staining values were associated with tumor progression while lower values corresponded to stable disease (Dirven et al., 1999). The authors rightfully concede that selection bias may have confounded their data since a higher proportion of completely resected tumors occurred in the lower staining group. However, a similar relationship was noted even when analyzing

residual tumors alone. It remains to be seen whether a higher staining value identifies a more clinically aggressive subset of PAs or a more diffuse pattern of growth, thereby making gross total resection more difficult. MIB-1 labeling index has been extensively evaluated in an attempt to quantify differences in PA proliferative ability. Such studies consistently report low values for these tumors (Fig. 22.3a). Giannini et al. (1999) found a statistically significant difference in the distribution of MIB-1 labeling indices for PAs and higher grade astrocytomas. The MIB-1 labeling index of anaplastic pilocytic astrocytomas (Fig. 22.3b) appears to be dramatically higher than their precursor counterparts with values of 24.7 and 2.6%, respectively (Rodriguez et al., 2010). However, the diagnostic potential for this measurement remains poor given the fact that the MIB-1 labeling index for PAs overlaps considerably with the clinically more aggressive diffuse astrocytomas. Klein and Roggendorf (2001) showed that roughly one third of the proliferating cells in PAs labeled by MIB-1 were in fact non-neoplastic microglia, far

22 Pathological and Immunohistochemical Factors Affecting Surgical Treatment and Surveillance

more than the other glial tumors tested. Interestingly, Tibbetts et al. (2009) reported an insignificant trend toward a worse prognosis with greater than 10% CD68 positivity. CD68 is a microglia/macrophage marker. This finding raises the intriguing notion that a PAs clinical behavior is influenced not only by tumor biology but also by microenvironment. Dirven et al. (1998) reported the utility of MIB-1 labeling index in the context of residual PA monitoring. They showed that clinically and radiographically quiescent PAs after subtotal resection had a MIB-1 labeling index of 3.3% compared with 6.6% in progressive, residual lesions. Haapasalo et al. (1999) further analyzed proliferating cell nuclear antigen (PCNA), Sphase fraction, and p53 expression to quantify cellular proliferation, but failed to identify novel PA prognostic variables. Fisher et al. (2002) found only marginal statistical significance for the ability of MIB-1 labeling to predict progression-free-survival. In contrast, Bowers et al. (2003) did note that patients harboring a PA with a MIB-1 labeling index greater than 2.0 had a shorter progression-free-survival. This finding is of particular interest as it offers potential improvement in disease surveillance for a specific subset of patients at potentially increased risk for recurrence. Takei et al. (2008) also successfully stratified PAs into two clinical subsets based on the staining patterns of oligodendroglial differentiation markers as well as proliferative index. They included immunohistochemical sections for myelin basic protein (MBP), platelet-derived growth factor receptor-a (PDGFR-α), Olig-1, Olig-2, and Ki-67. An inverse relationship between MBP expression and proliferative index was observed, whereas PDGFR-α displayed a direct relationship. Takei et al. (2008) also showed a statistically significant worse progression-free-survival in patients harboring PAs with particularly high PDGFR-α expression. The early identification of patient subgroups at higher risk for tumor progression in an otherwise indolent tumor population affords great advances in appropriate surveillance strategies and ultimately future treatment regimens.

Conclusions Current scientific exploration has achieved a number of noteworthy accomplishments regarding gliomagenesis and PA tumor biology. Microarray technology

201

has revealed unique genetic profiles for hypothalamic PAs and cerebellar PAs (Tchoghandijian et al., 2009), illustrating the potential importance of tumor microenvironment. The ultimate goal of tailoring a patient’s tumor to an individual treatment plan remains for patients with a PA. In the absence of genetic markers that adequately predict tumor behavior, pathological and immunohistochemical characterization remains the most clinically feasible means of PA risk stratification.

References Bannykh SI, Stolt CC, Kim J, Perry A, Wegner M (2006) Oligodendroglial-specific transcriptional factor SOX10 is ubiquitously expressed in human gliomas. J Neurooncol 76:115–127 Benesch M, Eder HG, Sovinz P, Raith J, Lackner H, Moser A, Urban C (2006) Residual or recurrent cerebellar low-grade glioma in children after tumor resection: is re-treatment needed? A single center experience from 1983 to 2003. Pediatr Neurosurg 42:159–164 Beni-Adani L, Gomori M, Spektor S, Constantini S (2000) Cyst wall enhancement in pilocytic astrocytoma: neoplastic or reactive phenomena. Pediatr Neurosurg 32:234–239 Bouvier C, Bartoli C, Aguirre-Cruz L, Virard I, Colin C, Fernandez C, Gouvernet J, Figarella-Branger D (2003) Shared oligodendrocyte lineage gene expression in gliomas and oligodendrocyte progenitor cells. J Neurosurg 99: 344–350 Bowers DC, Gargan L, Kapur P, Reisch JS, Mulne AF, Shapiro KN, Elterman RD, Winick NJ, Margraf LR (2003) Study of the MIB-1 labeling index as a predictor of tumor progression in pilocytic astrocytomas in children and adolescents. J Clin Oncol 21:2968–2973 Ceppa EP, Bouffet E, Griebel R, Robinson C, Tihan T (2007) The pilomyxoid astrocytoma and its relationship to pilocytic astrocytoma: report of a case and a critical review of the entity. J Neurooncol 81:191–196 Dirks PB, Jay V, Becker LE, Drake JM, Humphreys RP, Hoffman HJ, Rutka JT (1994) Development of anaplastic changes in low-grade astrocytomas of childhood. Neurosurgery 34:68–78 Dirven CM, Koudstaal J, Mooij JJ, Molenaar WM (1998) The proliferative potential of the pilocytic astrocytoma: the relation between MIB-1 labeling and clinical and neuroradiological follow-up. J Neurooncol 37:9–16 Dirven CM, Koudstaal J, Mooij JA, Molenaar WM (1999) AgNOR staining may reflect the growth potential of pilocytic astrocytomas. Childs Nerv Syst 15:384–388 Faria AV, Azevedo GC, Zanardi VA, Ghizoni E, Queiroz LS (2006) Dissemination patterns of pilocytic astrocytoma. Clin Neurol Neurosurg 108:568–572 Fernandez C, Figarella-Branger D, Girard N, Bouvier-Labit C, Gouvernet J, Paz Paredes A, Lena G (2003) Pilocytic astrocytomas in children: prognostic factors-a retrospective

202 study of 80 cases. Neurosurgery 53:544-553, discussion 554–545 Fisher BJ, Leighton CC, Vujovic O, Macdonald DR, Stitt L (2001) Results of a policy of surveillance alone after surgical management of pediatric low grade gliomas. Int J Radiat Oncol Biol Phys 51:704–710 Fisher BJ, Naumova E, Leighton CC, Naumov GN, Kerklviet N, Fortin D, Macdonald DR, Cairncross JG, Bauman GS, Stitt L (2002) Ki-67: a prognostic factor for low-grade glioma?. Int J Radiat Oncol Biol Phys 52:996–1001 Fisher PG, Tihan T, Goldthwaite PT, Wharam MD, Carson BS, Weingart JD, Repka MX, Cohen KJ, Burger PC (2008) Outcome analysis of childhood low-grade astrocytomas. Pediatr Blood Cancer 51:245–250 Gajjar A, Sanford RA, Heideman R, Jenkins JJ, Walter A, Li Y, Langston JW, Muhlbauer M, Boyett JM, Kun LE (1997) Low-grade astrocytoma: a decade of experience at St. Jude Children’s Research Hospital. J Clin Oncol 15:2792–2799 Garcia DM, Latifi HR, Simpson JR, Picker S (1989) Astrocytomas of the cerebellum in children. J Neurosurg 71:661–664 Giannini C, Scheithauer BW, Burger PC, Christensen MR, Wollan PC, Sebo TJ, Forsyth PA, Hayostek CJ (1999) Cellular proliferation in pilocytic and diffuse astrocytomas. J Neuropathol Exp Neurol 58:46–53 Haapasalo H, Sallinen S, Sallinen P, Helen P, Jaaskelainen J, Salmi TT, Paetau A, Paljarvi L, Visakorpi T, Kalimo H (1999) Clinicopathological correlation of cell proliferation, apoptosis and p53 in cerebellar pilocytic astrocytomas. Neuropathol Appl Neurobiol 25:134–142 Horbinski C, Hamilton RL, Lovell C, Burnham J, Pollack IF (2010) Impact of morphology, MIB-1, p53 and MGMT on outcome in pilocytic astrocytomas. Brain Pathol 20:581–588 Huber J, Sovinz P, Lackner H, Mokry M, Eder H, Urban C (2007) Diencephalic syndrome: a frequently delayed diagnosis in failure to thrive. Klin Padiatr 219:91–94 Klein R, Roggendorf W (2001) Increased microglia proliferation separates pilocytic astrocytomas from diffuse astrocytomas: a double labeling study. Acta Neuropathol 101:245–248 Krieger MD, Gonzalez-Gomez I, Levy ML, McComb JG (1997) Recurrence patterns and anaplastic change in a long-term study of pilocytic astrocytomas. Pediatr Neurosurg 27:1–11

D. Haydon and J. Leonard 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 Paixão Becker A, de Oliveira RS, Saggioro FP, Neder L, Chimelli LM, Machado HR (2010) In pursuit of prognostic factors in children with pilocytic astrocytomas. Childs Nerv Syst 26:19–28 Pollack IF, Claassen D, al-Shboul Q, Janosky JE, Deutsch M (1995) Low-grade gliomas of the cerebral hemispheres in children: an analysis of 71 cases. J Neurosurg 82:536–547 Rodriguez FJ, Scheithauer BW, Burger PC, Jenkins S, Giannini C (2010) Anaplasia in pilocytic astrocytoma predicts aggressive behavior. Am J Surg Pathol 34:147–160 Strong JA, Hatten HP Jr., Brown MT, Debatin JF, Friedman HS, Oakes WJ, Tien R (1993) Pilocytic astrocytoma: correlation between the initial imaging features and clinical aggressiveness. AJR Am J Roentgenol 161:369–372 Takei H, Yogeswaren ST, Wong KK, Mehta V, Chintagumpala M, Dauser RC, Lau CC, Adesina AM (2008) Expression of oligodendroglial differentiation markers in pilocytic astrocytomas identifies two clinical subsets and shows a significant correlation with proliferation index and progression free survival. J Neurooncol 86:183–190 Tchoghandijian A, Fernandez C, Colin C, El Ayachi I, Voutsinos-Porche B, Fina F, Scavarda D, Piercecchi-Marti MD, Intagliata D, Ouafik L, Fraslon-Vanhulle C, FigarellaBranger D (2009) Pilocytic astrocytoma of the optic pathway: a tumour deriving from radial glia cells with a specific gene signature. Brain 132:1523–1535 Tibbetts KM, Emnett RJ, Gao F, Perry A, Gutmann DH, Leonard JR (2009) Histopathologic predictors of pilocytic astrocytoma event-free survival. Acta Neuropathol 117:657–665 Tihan T, Fisher PG, Kepner JL, Godfraind C, McComb RD, Goldthwaite PT, Burger PC (1999) Pediatric astrocytomas with monomorphous pilomyxoid features and a less favorable outcome. J Neuropathol Exp Neurol 58:1061–1068 Tomlinson FH, Scheithauer BW, Hayostek CJ, Parisi JE, Meyer FB, Shaw EG, Weiland TL, Katzmann JA, Jack CR Jr. (1994) The significance of atypia and histologic malignancy in pilocytic astrocytoma of the cerebellum: a clinicopathologic and flow cytometric study. J Child Neurol 9:301–310

Chapter 23

Pilomyxoid Astrocytomas: Chemotherapy Hitoshi Tsugu, Shinya Oshiro, Fuminari Komatsu, Hiroshi Abe, Takeo Fukushima, Tooru Inoue, Fumio Yanai, and Yuko Nomura

Abstract Pilomyxoid astrocytoma (PMA) is a variant of pilocytic astrocytoma (PA). However, PMA shows a higher rate of recurrence and dissemination in cerebrospinal fluid (CSF) than does pilocytic astrocytoma (PA). PMA occurs predominantly in the hypothalamic/chiasmic region, and thus it is usually treated with chemotherapy following surgical biopsy. We discuss the treatment of PMA. Materials and Methods: Between 1992 and 2009, the authors treated 5 patients, 2 males and 3 females, ranging in age from 3 months to 11 years. Results: Three patients showed CSF dissemination on the initial radiographic examination. All patients received chemotherapy; the most commonly used drug combination was that of cisplatin (CDDP)/carboplatin (CBDCA) and etoposide. If these drugs were unsuccessful, they were changed, or other drugs were added to the combination. After chemotherapy, four patients showed remarkable tumor regression. Nevertheless, one patient died due to tumor progression 22 months after initial diagnosis. Conclusion: Although our series comprised a small number of patients, treatment of PMA with chemotherapy appeared to be of value. Even if initial chemotherapy is ineffective, we recommend continued CDDP/CBDCA-based chemotherapy with new drug combinations. Keywords Pilomyxoid astrocytoma · Pilocytic astrocytoma · Chemotherapy · Cisplatin · Carboplatin · Leptomeningeal dissemination

Introduction In 1999, Tihan and colleagues proposed a subtype of pilocytic astrocytoma (PA), which showed histologically and clinically different features from those of PA (Tihan et al., 1999). These include a monomorphous pattern with a myxoid background, absence of Rosenthal fibers and eosinophilic granular bodies, and more aggressive clinical behavior. The 2000 World Health Organization (WHO) classification of tumors of the central nervous system refers to Tihan et al’s variant as pilomyxoid astrocytoma (PMA) (Burger et al., 2000). PMAs have the following clinical characteristics: many patients are infants or young children, and compared with PA, there is a higher rate of tumor recurrence and dissemination in cerebrospinal fluid (CSF) and the prognosis is less favorable. After Tihan et al. (1999), there was an increase in the number of PMA studies published (Arslanoglu et al., 2003; Ceppa et al., 2007; de Chadarévian et al., 2006; Chikai et al., 2004; Darwish et al., 2004; Enting et al., 2006; Fuller et al., 2001; Gottfried et al., 2003; Komakula et al., 2007; Komotar et al., 2004a, b, 2005, 2006; Melnédez et al., 2006; Petito, 2003). There were, however, few reports regarding the treatment of PMA (Ceppa et al., 2007; Komakula et al., 2007; Komotar et al., 2004b, 2005). Our intent is to detail the treatments and clinical courses of our patients and to contribute to the management of PMA.

Materials and Methods H. Tsugu () Department of Neurosurgery, Fukuoka University Faculty of Medicine, Jonan-ku, Fukuoka 814-0180, Japan e-mail: [email protected]

After the Tihan et al. (1999), we treated 3 patients with an initial diagnosis of PMA. In addition, 2 of the patients diagnosed with PA before 1999 were

M.A. Hayat (ed.), Tumors of the Central Nervous System, Volume 5, DOI 10.1007/978-94-007-2019-0_23, © Springer Science+Business Media B.V. 2012

203

204

subsequently reclassified as having a combination of PMA and PA. We reviewed the clinical courses of these 5 patients, 2 males and 3 females, ranging in age from 3 months old to 11 years. The follow-up period for all patients exceeded 1 year. We evaluated the radiological and pathological findings and our treatments on an individual patient basis. The 3 pure PMA patients showed CSF dissemination on the initial radiographic examination. The MIB-1 labeling index (LI) ranged from 4.3 to 16.3% (mean 9.2%). We administered chemotherapy to all 5 patients, of whom 4 showed remarkable tumor regression. Nevertheless, the chemotherapy was ineffective for one patient, a 3-month-old boy, who died of tumor progression 22 months after the initial diagnosis. With the exception of this case, all patients showed a Karnofsky Performance Score of over 80% during the 22- to 213-month follow-up period.

Case Reports Case 1 A 3-month-old boy presented with remarkable emaciation, weighing 3,434 g (–4.4 S.D.), and with a head circumference of 39.2 cm (–0.9 S.D.). The patient showed diencephalic syndrome. Brain magnetic resonance imaging (MRI) revealed a heterogeneously enhancing large hypothalamic-chiasmatic mass accompanied by obstructive hydrocephalus and CSF disseminations. A ventriculoperitoneal shunt and tumor biopsy were performed. Histologically, the tumor showed proliferation of piloid cells with myxoid backgrounds, monomorphous architecture, and angiocentric arrangements. The MIB-1 LI was 11.5%. Our diagnosis was PMA. Chemotherapy with carboplatin 135 mg/m2 on day 1 and etoposide 45 mg/m2 on days 1, 2, and 3 was administered. Tumor progression was still evident after two treatment courses. The chemotherapeutic regimen was changed to MCNU 70 mg/m2 on day 1 and vincristine 1.4 mg/m2 on days 1 and 8. After four courses, tumor progression continued. The third treatment (one cycle only) was cyclophosphamide 600 mg/m2 on day 1 and cisplatin 20 mg/m2 on days 1, 2, and 3. The chemotherapy was ineffective, and the patient died 22 months after the initial diagnosis.

H. Tsugu et al.

Case 2 A 7-month-old boy presented with remarkable emaciation, weighing 5,300 g (–3.8 S.D.). The patient showed diencephalic syndrome without neurological deficits. Brain MRI disclosed a homogeneously enhancing large hypothalamic-chiasmatic mass accompanied by CSF disseminations (Fig. 23.1a). This tumor was of slightly low intensity on the T1-weighted image and bright intensity on the T2-weighted image. A tumor biopsy was performed, and we made the diagnosis of PMA. he MIB-1 LI was 8.2%. Chemotherapy with carboplatin 450 mg/m2 on day 1 and etoposide 100 mg/m2 on days 1, 2, and 3 was administered. After three courses, this combination chemotherapy was ineffective and was changed to vincristine 1.5 mg/m2 and cyclophosphamide 1200 mg/m2 on day 1, and etoposide 100 mg/m2 and cisplatin 20 mg/m2 on days 1 through 5. After five courses, the tumor size was remarkably reduced (Fig. 23.1b); however, the use of cisplatin was halted because of hearing loss. After eleven courses of the second combination of regimens, the tumor and the disseminated lesions completely disappeared (Fig. 23.1c). Unfortunately, the tumor recurred in the chiasmal region 5 months later. We decided to follow this lesion carefully without additional therapy because the lesion was very small. Now, at 6 years of age, the patient is still very well, and the recurrent tumor remains stable (Fig. 23.1d).

Case 3 A 7-year-old girl complained of headache and nausea. A brain MRI revealed a homogeneously enhancing sellar mass accompanied by disseminations into the CSF at the right crural and ambient cisterns and the left middle fossa (Fig. 23.2a). A tumor biopsy was performed, and we made a diagnosis of PMA. he MIB1 LI was 16.3%.The patient received chemotherapy with cisplatin 20 mg/m2 , etoposide 60 mg/m2 , and ifosfamide 900 mg/m2 for 5 consecutive days. After three courses, the regimens was ineffective (Fig. 23.2b) and was changed to cisplatin 70 mg/m2 and MCNU 60 mg/m2 on day 1 and vincristine 1.5 mg/m2 on days 1 and 8. After six courses, the size of the tumor and its lesions disseminated in the CSF underwent extreme

23 Pilomyxoid Astrocytomas: Chemotherapy Fig. 23.1 Case 2. (a) Pre-operative gadolinium-enhanced coronal brain MR image showing a chiasmatic-hypothalamic solid mass with homogeneous enhancement. (b) Gadolinium-enhanced brain MR image shows tumor size remarkably reduced after five courses of a second chemotherapeutic treatment (cisplatin, etoposide, vincristine, and cyclophosphamide). (c) After 11 courses, the patient achieved complete remission. (d) Gadolinium-enhanced coronal brain MR image of the tumor recurrence (arrow) shows a small chiasmatic enhanced mass at 6 years of age

205

A

B

C

D

shrinkage (Fig. 23.2c). Local tumor regrowth in the suprasellar region was found, and we then performed a partial resection (75%) of the recurrent lesion at 48 months after initial diagnosis. Additional therapy was not performed after the second surgery. Presently, the patient is very well at 13 years old. However, another small local tumor regrowth at the right crural cistern has recurred (Fig. 23.2d). We plan a resection of this lesion.

Case 4 An 11-year-old girl was sent to our hospital because of tumor recurrence. She had first presented with vomiting at 2 years of age. At that time, head CT revealed hydrocephalus and a suprasellar tumor. The imaging findings were consistent with a low-grade

opticochiasmic glioma. A surgical biopsy was not performed; instead, she received a ventriculoperitoneal shunt and then local radiotherapy (30 Gy). At 4 years of age, the tumor recurred and was partially removed. After this surgery, she underwent four courses of combinational chemotherapy (cisplatin and etoposide). Her brain MRI on admission at age 11 revealed a homogeneous enhanced mass in the suprasellar region. Subtotal removal of the tumor was performed. Histologically, the tumor showed two different appearances: angiocentric architecture (PMA portions) and a solid component accompanied by Rosenthal fibers (PA portions). The MIB-1 LI was 4.3% in the PMA portions and 2.3% in the PA portions. After surgery, no adjuvant therapy was administered. Seven years after surgery, the residual tumor remains stable without further tumor progression, and we are carefully monitoring its growth.

206 Fig. 23.2 Case 3. (a) Gadolinium-enhanced axial brain MR images show small enhanced lesions in the right crural and ambient cisterns (arrow). (b) Gadolinium-enhanced axial brain MR image after three courses of the first treatment (cisplatin, ifosfamide, and etoposide) shows enlargement of enhanced lesions. (c) Gadolinium-enhanced axial brain MR image after six courses of the second treatment (cisplatin, vincristine, and ranimustine) shows remarkably reduced tumor enhancements. (d) Gadolinium-enhanced axial brain MR image shows tumor regrowth at the right crural cistern at 13 years of age (open arrow)

H. Tsugu et al.

A

B

C

D

Case 5 A 7-year-old girl complained of gait disturbance. A brain MRI revealed a slightly enhanced mass within the third ventricle accompanied by obstructive hydrocephalus. Subtotal removal of the tumor was carried out. The surgical specimen showed biphasic histological features. The loose textured portions showed proliferation of bipolar cells in a myxoid background without any Rosenthal fibers or eosinophilic granular bodies. Our diagnosis was a combination of PMA and PA. In the PMA portions the MIB-1 LI was 5.6%, which is higher than that seen in classic PA. When the patient was 10 years old, her tumor recurred. A subtotal removal of the tumor was performed, and postoperative radiotherapy (55 Gy) was administered. Four years later, after a second surgery, the tumor recurred again. Chemotherapy with carboplatin 450 mg/m2 on the first day and etoposide 100 mg/m2 on the first day and the

next 4 days was administered. After 12 courses, the tumor shrank remarkably. Six years later at 20 years of age, her health is excellent.

Discussion In 1999, Tihan et al. reported pediatric astrocytoma with monomorphous pilomyxoid features. They proposed a subtype of PA, which showed features that were histologically and clinically different from PA, including those of monomorphous pattern with a myxoid background, absence of Rosenthal fibers and eosinophilic granular bodies, and more aggressive clinical behavior. Before the Tihan et al., 1999 report, however, Cottingham et al. (1996) had already pointed out that the classical histological pattern of PA was frequently lacking in infants. However, they did not

23 Pilomyxoid Astrocytomas: Chemotherapy

report clinical differences between classic PA and infant PA. The 2000 World Health Organization (WHO) classification of tumors of the central nervous system (Burger et al., 2000) refers to Tihan et al. (1999) and Cottingham et al. (1996) variant as PA in infants. In the 2007 WHO classification, the variant is classified as PMA, a subtype of PA, which corresponds to grade II (Scheithauer et al., 2007). The 2007 WHO classification mentions that histologically, PMAs show angiocentric cell arrangement and that they show more aggressive biological behavior than that of typical PAs; PMAs often show CSF dissemination and local tumor recurrence (Scheithauer et al., 2007). Some of these reports pointed out a close relation between PMA and PA: Ceppa et al. (2007), Chikai et al. (2004), Komakula et al. (2007), and Fernandez et al. (2003) all noted a maturation or transformation of PMA to PA, and Gottfried et al. (2003) and Komotar et al. (2006) discussed hybrid histological features of PMA and PA. In our fourth and fifth cases, we considered histologically hybrid forms because surgical specimens showed histologic features of both PMA and PA. PMAs have higher recurrence rates and a poorer prognosis than do classic PAs (Fernandez et al., 2003; Komotar et al., 2004a, b, 2005) and often display CSF dissemination (Darwish et al., 2004; Enting et al., 2006; Komotar et al., 2004a; van der Wal et al., 2003). Komotar et al. (2004a) reported that 16 PMA patients (76%) experienced local recurrence, and 3 patients (14.3%) showed CSF dissemination. However, 21 PA patients (50%) experienced local recurrence, and there were no patients with CSF dissemination. In their study, 7 PMA patients (33%) and 7 PA patients (17%) died as a result of their disease. Chemotherapy was performed in 10 patients (47%), radiotherapy in 1 patient (5%), and both chemotherapy and radiotherapy in 4 patients (19%). Unfortunately, the authors did not detail the chemotherapy and radiotherapy used. There are a limited number of reports regarding therapeutic strategies for PMA patients (Ceppa et al., 2007; Komakula et al., 2007; Komotar et al., 2004b, 2005). Initial chemotherapy in a patient reported by Ceppa et al. (2007) was preceded by carboplatin and vincristine. After twelve courses, tumor volume was significantly reduced. The tumor recurred several times, however, forcing a change in their regimen: vinblastine, temozolomide, etoposide, and cisplatin

207

were administered. Finally, the tumor shrank and remained stable with no evidence of progression in the 6 years that followed the initial diagnosis. Previous to the Tihan et al. (1999) report, several authors had described the management of aggressive PAs, especially that of infantile PAs complicated by CSF dissemination (Kageji et al., 2003; Mamelak et al., 1994; McCowage et al., 1996; Petronio et al., 1991; Pollack et al., 1994; Tamura et al., 1998). These reports can be used to assist in the management of PMA patients. Mamelak et al. (1994) recommends that patients under 3 years of age be treated with multiagent chemotherapy, but they offer no suggestions for chemotherapeutic regimens. Petronio et al. (1991) suggest that nitrosourea-based cytotoxic regimens are useful for the initial treatment of infant and childhood chiasmal/hypothalamic PAs. McCowage et al. (1996) report the successful treatment of CSF dissemination of childhood PAs with high-dose cyclophosphamide. Kageji et al. (2003) also report a successful case of treatment with high-dose chemotherapy using a combination of cyclophosphamide, carboplatin, and ranimustine. In our first case, chemotherapy was ineffective and the patient died. We started chemotherapy at 30% of the ordinary treatment dose and administered 50% of the ordinary dose in the second course. During this period, the tumor rapidly progressed. Because the effectiveness of cisplatin and carboplatin in tumor suppression depends upon dosage, we conclude that these drugs were not administered in sufficient quantity. With the exception of this instance (case 1), one patient (case 4) achieved disease stabilization and three patients (cases 2, 3, and 5) experienced significant reductions in tumor volume after the chemotherapy. In particular, cases 2 and 3 presented CSF dissemination at the initial diagnosis, and it responded remarkably well to chemotherapy. Cisplatin/carboplatinbased multiagent chemotherapy was used in the cases showing good response (cases 2–5). These preliminary results suggest that cisplatin/carboplatin-based multiagent chemotherapy is useful for the initial treatment of PMAs. Recently, in adult malignant gliomas, temozolomide has become the first line adjuvant chemotherapy after publication of the Stupp et al. (2005) report. This report showed that combined therapy comprising temozolomide and radiotherapy improved the prognosis in glioblastomas.There have also been a few reports

208

of pediatric cases of PMA/PA treated by temozolomide (Aryan et al., 2005; Ceppa et al., 2007; Kuo et al., 2003). According to these reports, most of the patients had recurrent or progressive disease following failure of previous non-temozolomide chemotherapy. Subsequently, these patients received temozolomide, and half of them were reported to have stable disease or minor response to the drug. Temozolomide may become the expected treatment for PMA/aggressive PA. Long-term follow-up will be required to assess the efficacy of temozolomide. Acknowledgement This article was revised from our previously published paper: Tsugu H, Oshiro S, Yanai F, Komatsu F, Abe H, Fukushima T, Nomura Y, Matsumoto S, Nabeshima K, Takano K, Utsunomiya H. Management of pilomyxoid astrocytomas: our experience. Anticancer Res 29: 919–926, 2009.

References Arslanoglu A, Cirak B, Horska A, Okoh J, Tihan T, Aronson L, Avellino AM, Burger PC, Yousem DM (2003) MR imaging characteristics of pilomyxoid astrocytomas. AJNR Am J Neuroradiol 24:1906–1908 Aryan HE, Meltzer HS, Lu DC, Ozgur BM, Levy ML, Bruce DA (2005) Management of pilocytic astrocytoma with diffuse leptomeningeal spread: two cases and review of the literature. Childs Nerv Syst 21:477–481 Burger PC, Scheithauer BW, Paulus W, Szymas J, Giannini C, Kleihues P (2000) Pilocytic astrocytoma. Pathology and genetics of tumours of the nervous system. In: Kleihues P, Cavenee WK (eds) World Health Organization classification of tumors. IARC Press, Lyon Ceppa EP, Bouffet E, Griebel R, Robinson C, Tihan T (2007) The pilomyxoid astrocytoma and its relationship to pilocytic astrocytoma: report of a case and a critical review of the entity. J Neurooncol 81:191–196 Chikai K, Ohnishi A, Kato T, Ikeda J, Sawamura Y, Iwasaki Y, Itoh T, Sawa H, Nagashima K (2004) Clinico-pathological features of pilomyxoid astrocytoma of the optic pathway. Acta Neuropathol (Berl) 108:109–114 Cottingham SL, Boesel CP, Yates AJ (1996) Pilocytic astrocytoma in infants: A distinctive histologic pattern. abstract) J Neuropathol Exp Neurol 55:654 Darwish B, Koleda C, Lau H, Balakrishnan V, Wickremesekera A (2004) Juvenile pilocytic astrocytoma ‘pilomyxoid variant’ with spinal metastases. J Clin Neurosci 11:640–642 de Chadarévian JP, Halligan GE, Reddy G, Bertrand L, Pascasio JM, Faerber EN, Katsetos CD (2006) Glioneuronal phenotype in a diencephalic pilomyxoid astrocytoma. Pediatr Dev Pathol 9:480–487 Enting RH, van der Graaf WT, Kros JM, Heesters M, Metzemaekers J, den Dunnen W (2006) Radiotherapy plus concomitant and adjuvant temozolomide for leptomeningeal

H. Tsugu et al. pilomyxoid astrocytoma: a case study. J Neurooncol 80: 107–108 Fernandez C, Figarella-Branger D, Girard N, Bouvier-Labit C, Gouvernet J, Paz Paredes AP, Lena G (2003) Pilocytic astrocytomas in children: prognostic factors–a retrospective study of 80 cases. Neurosurgery 53:544–555 Fuller CE, Frankel B, Smith M, Rodziewitz G, Landas SK, Caruso R, Schelper R (2001) Suprasellar monomorphous pilomyxoid neoplasm: an ultastructural analysis. Clin Neuropathol 20:256–262 Gottfried ON, Fults DW, Townsend JJ, Couldwell WT (2003) Spontaneous hemorrhage associated with a pilomyxoid astrocytoma. Case report J Neurosurg 99:416–420 Kageji T, Nagahiro S, Horiguchi H, Watanabe T, Suzuya H, Okamoto Y, Kuroda Y (2003) Successful high-dose chemotherapy for widespread neuroaxis dissemination of an optico-hypothalamic juvenile pilocytic astrocytoma in an infant: a case report. J Neurooncol 62:281–287 Komakula ST, Fenton LZ, Kleinschmidt-DeMasters BK, Foreman NK (2007) Pilomyxoid astrocytoma: neuroimaging with clinicopathologic correlates in 4 cases followed over time. J Pediatr Hematol Oncol 29:465–470 Komotar RJ, Burger PC, Carson BS, Brem H, Olivi A, Goldthwaite PT, Tihan T (2004) Pilocytic and pilomyxoid hypothalamic/chiasmatic astrocytomas. Neurosurgery 54:72–80 Komotar RJ, Mocco J, Carson BS, Sughrue ME, Zacharia BE, Sisti AC, Canoll PD, Khandji AG, Tihan T, Burger PC, Bruce JN (2004) Pilomyxoid astrocytoma: a review. MedGenMed 6:42 Komotar RJ, Mocco J, Jones JE, Zacharia BE, Tihan T, Feldstein NA, Anderson RC (2005) Pilomyxoid astrocytoma: diagnosis, prognosis, and management. Neurosurg Focus 18(6A): 1–4 Komotar RJ, Mocco J, Zacharia BE, Wilson DA, Kim PY, Canoll PD, Goodman RR (2006) Astrocytoma with pilomyxoid features presenting in an adult. Neuropathology 26: 89–93 Kuo DJ, Weiner HL, Wisoff J, Miller DC, Knopp EA, Finlay JL (2003) Temozolomide is active in childhood, progressive, unresectable, low-grade gliomas. J Pediatr Hematol Oncol 25:372–378 Mamelak AN, Prados MD, Obana WG, Cogen PH, Edwards MSB (1994) Treatment options and prognosis for multicentric juvenile pilocytic astrocytoma. J Neurosurg 81:24–30 McCowage G, Tien R, McLendon R, Felsberg G, Fuchs H, Graham ML, Kurtzberg J, Moghrabi A, Ferrell L, Kerby T, Duncan-Brown M, Stewart E, Robertson PL, Colvin OM, Golembe B, Bigner DD, Friedman HS (1996) Successful treatment of childhood pilocytic astrocytomas metastatic to the leptomeninges with high-dose cyclophosphamide. Med Pediatr Oncol 27:32–39 Melnédez B, Fiaño C, Ruano Y, Hernández-Moneo JL, Mollejo M, Martinez P (2006) BCR gene disruption in a pilomyxoid astrocytoma. Neuropathology 26:442–446 Petito CK (2003) Suprasellar monomorphous pilomyxoid gliomas. AJNR Am J Neuroradiol 24:1931–1932 Petronio J, Edwards MS, Prados M, Freyberger S, Rabbitt J, Silver P, Levin VA (1991) Management of chiasmal and hypothalamic gliomas of infancy and childhood with chemotherapy. J Neurosurg 74:701–708

23 Pilomyxoid Astrocytomas: Chemotherapy Pollack IF, Hurtt M, Pang D, Albright AL (1994) Dissemination of low-grade intracranial astrocytomas in children. Cancer 73:2869–2878 Scheithauer BW, Hawkins C, Tihan T, VandenBerg SR, Burger PC (2007) Pilocytic astrocytoma. In: Louis DN, Ohgaki H, Wiestler OD, Cavenee WK (eds) WHO Classification of Tumors of the Central Nervous System. IARC Press, Lyon 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

209 Tamura M, Zama A, Kurihara H, Fujimaki H, Imai H, Kano T, Saitoh F (1998) Management of recurrent pilocytic astrocytoma with leptomeningeal dissemination in childhood. Childs Nerv Syst 14:617–622 Tihan T, Fisher PG, Kepner JL, Godfraind C, McComb RD, Goldthwaite PT, Burger PC (1999) Pediatric astrocytomas with monomorphous pilomyxoid features and a less favorable outcome. J Neuropathol Exp Neurol 58:1061–1068 van der Wal EP, Azzrelli B, Edwards-Brown M (2003) Malignant transformation of a chiasmatic pilocytic astrocytoma in a patient with diencephalic syndrome. Pediatr Radiol 33:207–210

Part III

Astrocytomas: Prognosis

Chapter 24

Astrocytomas: Predicting Survival and Recurrence Using Cerebral Blood Volume Measurements Sotirios Bisdas

Abstract Although histopathologic assessment remains the reference standard for determining the glioma grade and predicting survival, inherent shortcomings like sampling error, inter- and intrapathologist variability, and difference in biology within WHO grades of astrocytomas may result in inadequate evaluation of the entire tumor. Thus, imaging studies and specifically, cerebral blood volume (CBV) measurements may provide information in addition to the pathologic assessment and improve grading as well as have predictive value for survival and recurrence. Various studies suggest that the tumor relative CBV (rCBV) on pretreatment MR imaging is an independent stronger than the degree of enhancement predictor of survival in patients with low- and high-grade astrocytomas. Moreover, rCBV measurements have been shown to be more significant than histopathological grade for the prediction of progression-free survival even among patients with astrocytomas of the same histologic grade. Finally, the combined assessment of histopathologic and perfusion MR imaging findings seems to surpass the individual methods and may be useful to determine optimal management strategies in patients with astrocytomas. Keywords rCBV · Astrocytomas · Glioblastoma multiforme · CBV · FLAIR image · Relaxivity

S. Bisdas () Department of Diagnostic and Interventional Neuroradiology, Karls Eberhard University, Tübingen, Germany e-mail: [email protected]

Introduction An estimated 22,070 new cases of primary malignant brain and central nervous system tumors were diagnosed in the United States in 2009 (12,010 in males and 10,060 in females). This represents 1.49% of all primary malignant cancers in the United States in 2009. The worldwide incidence rate of primary malignant brain and central nervous system tumors in 2002, age–adjusted using the world standard population, is 3.7 per 100,000 person–years in males and 2.6 per 100,000 person–years in females. The incidence rates are higher in more developed countries (males: 5.8 per 100,000 person–years; females: 4.1 per 100,000 person–years) than in less developed countries (males: 3.0 per 100,000 person–years; females: 2.1 per 100,000 person–years). This is a form of cancer that affects children and adults from all socioeconomic backgrounds. Astrocytomas (WHO grade II) occur primarily in young adults between the ages of 20 and 40 years. This is a slow-growing tumor, but it is not benign, mainly because of its invasive quality and its location. Low-grade tumors have proved difficult to study, because of lack of tissue availability. The current criterion standard for tumor grading is histopathologic assessment, but this has limitations, such as inherent sampling error associated with the limited number of biopsy samples. When resections are performed, these tumors often are found to be admixtures of normal reactive cells and tumor cells. Anaplastic astrocytomas are thought to develop from low-grade astrocytomas, although there are no sharply defined histologic criteria separating astrocytomas from anaplastic astrocytomas. This tumor occurs in both young and old

M.A. Hayat (ed.), Tumors of the Central Nervous System, Volume 5, DOI 10.1007/978-94-007-2019-0_24, © Springer Science+Business Media B.V. 2012

213

214

patients, with a peak incidence in the mid-fifties. Like astrocytomas, the gain of chromosome 7 is the most frequent numerical aberration. Glioblastoma multiforme (GBM) is the most malignant of the astrocytic tumors. This tumor is highly infiltrative and possesses mitotic activity as well as areas of necrosis and vascular proliferation. While survival times vary, depending on the tumor type and tumor grade, GBM is a devastating tumor with a median survival time of 12–15 months, despite current best treatment based on surgery, radiotherapy and systemic chemotherapy. This is a statistic that has not changed for almost the past 25 years, after establishing the role of postoperative radiotherapy for malignant astrocytomas and despite advances in established treatment modalities such as radiotherapy and chemotherapy and investigations of the cellular, molecular, and genetic pathways involved in the progression of malignant astrocytomas. As a result, biomarkers have emerged as diagnostic, predictive, and prognostic tools that have the potential to transform the field of brain tumor diagnostics. Magnetic resonance imaging (MRI) is the imaging method of choice for characterization of astrocytomas prior to any treatment initiation. T1-weighted MRI, with gadolinium, is the reference standard. However, this technique only reflects biological activity of the tumor indirectly by detecting the breakdown of the blood-brain barrier and the degree of contrast enhancement is not a reliable indicator of the tumor grade. Consequently, numerous studies have suggested that more advanced MRI methods may be employed in various stages of disease to target the changes in diffusivity through the interstitial space (diffusionweighted MRI), the tumor-induced neo-vascularisation (perfusion-weighted MRI), and the changes in concentrations of metabolites (magnetic resonance spectroscopy). Perfusion-weighted MRI is used to quantitatively measure cerebral hemodynamic characteristics and along with peak height, percentage of signal intensity recovery, microvascular permeability measurements allows the determination of cerebral blood volume (CBV) and has been shown to be clinically useful in predicting astrocytoma histopathologic grade (Law et al., 2004) and disease free survival as well as recurrence in patients with low-grade astrocytomas and glioblastomas (Bisdas et al., 2009). The purpose of this chapter is to give a brief critical approach of

S. Bisdas

the technique of cerebral blood volume estimation by MRI, to elucidate the major role of data analysis, and to review the predictive role of CBV measurements in survival and recurrence appearance in patients with astrocytomas.

Cerebral Blood Volume Estimation Contrast-enhanced MR imaging has spawned two distinct methods, based on differing philosophies, for the measurement of CBV in brain tumors. Most commonly used are the first-pass techniques, based on the magnetic susceptibility contrast phenomenon (dynamic susceptibility contrast-enhanced dynamic MRI or DSC-MRI) and the resulting changes in T2∗ (Aronen et al., 1995; Knopp et al., 1999). Albeit studied widely, DSC-MRI has a number of significant disadvantages. Without a reference blood signal, only relative CBV values can be obtained. The paramagnetic recruitment effect leads to distortion of anatomy particularly where contrast lies within large vessels. Most T2- and T2∗ -weighted images will exhibit some degree of T1 sensitivity and will show signal changes resulting from relaxivity effects particularly from contrast agent that has leaked into the extravascular extracellular space. Thus, the use of contrast pre-enhancement or low T1 sensitivity-based sequences has been suggested for a reliable dynamic imaging of enhancing brain lesions with blood-brain-barrier disruption whereas other authors suggest that an optimization of the DSC perfusion sequences in terms of echo time and flip angle may counter the T1- and T2∗ -weighted changes caused by contrast material extravasation and thus, a contrast agent preload is unnecessary (Kassner et al., 2000; Paulson and Schmainda, 2008). The major problem with the preload approach is that the efficiency of the technique is dependent on the interstitial contrast concentration at the time of the bolus passage, which, in turn, is dependent on the differing contrast diffusion rates of the tumors. T1-based methods (dynamic relaxivity contrastenhanced MRI or DRC-MRI) do not suffer from these disadvantages but have their own inherent problems caused mainly by the additive effects of intravascular and extravascular contrast media on the observed signal changes. Despite this, several techniques have

24 Astrocytomas: Predicting Survival and Recurrence Using Cerebral Blood Volume Measurements

been developed to estimate CBV measurements from DRC-MRI (Li et al., 2003). Comparison of T1-CBV maps with T2-CBV maps in patients with low-grade and anaplastic astrocytomas showed similar results whereas in patients with high-grade lesions DRCMRI maps demonstrated higher CBV values than the DSC-MRI maps (Bruening et al., 1996; Hacklander et al., 1997). Other authors could also differentiate low-from high-grade astrocytomas based on T1-CBV maps (Ludemann et al., 2006). Other authors have proposed the use of a gradient echo pulse sequence with two echoes for collection of perfusion-weighted MRI data to allow simultaneous acquisition of T2∗ weighted and T1-weighted data (Uematsu et al., 2001). T2∗ is then derived from the two echoes and used to remove the transverse decay effects from the first echo. Such an acquisition yields separated T2∗ and T1 variations and subsequent removal of T1 shortening effects, which occur with conventional DSC-MRI, reduces artifactual underestimation of tumor blood volume (Uematsu et al., 2001). Further studies have also shown improved robustness of susceptibility-corrected multiecho methods compared with the single-echo approach using ultra-small iron oxide particles, which have negligible relaxivity effects (Persigehl et al., 2010). Finally, arterial spin labeling techniques provide a non-invasive alternative for CBV measurements and further studies are warranted to elucidate their role in tumor grading and treatment monitoring (Brookes et al., 2007). As the vast majority of the CBV measurements in astrocytomas for grading, response to therapy, outcome prediction as well as for detecting recurrent lesions have been performed by means of DSC-MRI, a brief description of the technique will be given below. Paramagnetic chelates, such as gadolinium DTPA alter the magnetic susceptibility of tissue and these effects monitored during the passage of a contrast bolus form the basis of T2-weighted and T2∗ -weighted perfusionweighted MRI. The signal change due to the passage of the initial contrast bolus through the vascular bed and due to the early recirculation of contrast through the brain is transformed into contrast concentration data. Contrast concentration time course data can then be used to extract parametric variables the most important of which in tumor studies is the relative cerebral blood volume (rCBV) (Aronen et al., 1995). A number of analysis techniques have been described and a

215

common feature of many is the use of a gamma variate fitting procedure to define the shape and position of the first pass bolus (Paulson and Schmainda, 2008). Parametric images of relative blood volume (rCBV) can be derived from the area under the contrast concentration time course curve either by direct calculation from the fitted gamma variate curve or by numerical integration based on contrast concentration time course data.

Methodological Considerations in Cerebral Blood Volume and Survival Data Analysis The pathological values are usually normalized to the contralateral healthy region and the further analysis may be performed either by using hot-spot or tumor regions-of-interest (ROIs). The method of deriving the maximum rCBV (hot-spot) may be based on single-pixel measurement, on multiple-pixel measurements with averaging (Lev et al., 2004) or on pre-defined ROI measurements (Law et al., 2004). Hotspot imaging may improve tumor-brain CBV contrast, and has been reported to provide high intra- and interobserver reproducibility in CBV measurements (Wetzel et al., 2002). However, choosing the hot spot is typically subjective and, thus, error prone. Also, the location of the hot spots can change with the choice of acquisition and post-processing methods. This is problematic if CBV is to be used as a guide for selecting a site for biopsy. Histogram analysis of MR imaging–derived CBV maps has been reported to differentiate high-grade from low-grade gliomas as well as low-grade oligodendroglial subtypes with high interobserver agreement (Emblem et al., 2008; Young et al., 2007). Also, an emerging and possibly primary role for CBV mapping is to monitor therapies. For this purpose, it is going to be critically important that an approach is used that gives the most accurate tumor-brain CBV contrast throughout the entire tumor. The variable aforementioned approaches for the data analysis of CBV maps have an impact on the reported optimal CBV thresholds for differentiation of the astrocytomas as well as for predicting recurrence and survival rates (Bisdas et al., 2009). It is of major

216

importance to analyze homogenous populations (e.g., low-grade astrocytomas, glioblastomas) without any contamination. The inclusion of oligodendrogliomas and mixed oligoastrocytomas is known to contaminate the correlation between histopathological grade and perfusion parameters (Law et al., 2004; Lev et al., 2004). It is important to stress out that finding CBV cut-off values in order to classify retrospectively the astrocytomas according to their histological grade may also introduce a methodological bias (Bisdas et al., 2009). Moreover, it is not always clear what criteria should be used in determining an “optimum” threshold. For example, one could choose to minimize C1 error, which minimizes the average of the falsepositive and false-negative error rates. This would be appropriate if the consequence of misclassifying lowgrade gliomas is the same as that of misclassifying high-grade gliomas and the two are equally likely to be presented to you for classification. Alternatively, one could choose to minimize the C2 error, the total number of misclassified tumors observed in the data. This adjusts for a difference in the relative frequency of low- and high-grade gliomas in the patient population. Because high-grade gliomas are much more prevalent, then a high misclassification rate of highgrade gliomas would result in a high total number of misclassified gliomas. Hence, choosing the threshold to minimize C2 error will tend to yield high sensitivity and relatively low specificity. In any case, correlation between histopathologic samples and CBV estimates with the intention to establish diagnostic thresholds may lead to false results for astrocytoma grading due to the sampling limitations and the spatial discordance. Thus, CBV estimates should be treated as an independent factor, like histopathologic grade, when examining the predictive value of the method and inclusion of previously reported thresholds should be avoided. The vast majority of investigators tend to use a binary system to categorize post-treatment contrastenhanced lesions as a tumor recurrence or a treatment-related change; therefore, contrast-enhanced lesions containing any amount of recurrent tumor are categorized as tumors, regardless of the degree of intermixed background treatment-related change. However, substratification of the tumor samples on the basis of the proportion of tumor recurrence to treatment-related change may reveal added diagnostic

S. Bisdas

or prognostic information. These subtle histologic differences are invisible in the conventional contrastenhanced MRI. A recently published report has introduced histogram analysis of rCBV values for glioma grading, which compared with the use of the local hot-spot ROI method, the histogram distribution of the rCBV also can demonstrate heterogeneous morphology of tumor vascularity, which is a major histopathologic feature of high-grade glioma as well as of post-treatment changes (Emblem et al., 2008). Other authors describe a novel post-processing method, the so-called parametric response map, in order to enhance the prognostic role of rCBV measurements during treatment in patients with high-grade gliomas (Galban et al., 2009). A further methodological issue when analyzing CBV values as predictors for the patient survival concerns the chosen end-points. Recurrence as well as progression are major end-points but, unfortunately, the definition of them used in the majority of the studies was based partly on the use of imaging criteria. With the assumption that many patients underwent radiation therapy with administration of novel anti-angiogenic agents, the occurrence of radiation necrosis, misinterpreted as disease progression, may have biased the results. The rate of patients alive and progression free at 6 months (PFS6mo) has been recognized as a valid endpoint and also includes patients who benefit from therapy by disease stabilization. However, with the frequent use of anti-angiogenic and vasculature modifying agents this endpoint may need to be revisited (Stupp and Roila, 2009). Based on these assumptions, Bisdas et al. also recently proposed the progression-free survival at 12 months (Bisdas et al., 2009). Potential limitations across the different studies include the possible effect of different treatment protocols on the time to progression for high- and low-grade astrocytomas, whereas the absence of data that indicate that any particular treatment regimen produces radically better outcomes than any other, is not expected to result in significant differences in patient survival. When the sub-populations of lowand high-grade astrocytomas are significantly different weighted special attention has to be given to the prediction of overall survival, eg when much fewer than 50% of patients die of disease, no valid estimates of median overall survival can be generated (Law et al., 2006).

24 Astrocytomas: Predicting Survival and Recurrence Using Cerebral Blood Volume Measurements

Predictive Value of Cerebral Blood Volume Measurements The management of gliomas has been broadly based on the classification according to the WHO histopathological classification and grading system following a stereotactic biopsy or surgery. However, the clinical management of the low-grade gliomas (WHO I-II) depends on the clinical behavior of these tumors and may vary from a “wait-and-see” policy to a prompt surgical treatment, whereas high-grade (WHO III-IV) gliomas typically require a resection (subtotal or near-total) and adjuvant concomitant or not, radiochemotherapy, with temozolamide being the current standard of chemotherapy agent (Stupp and Roila, 2009). This discrepancy along with the WHO-based histopathological grading pitfalls, indicates a need for another classification system as an adjunct or, potentially, replacement for histopathology in guiding treatment decisions. CBV measurements emerge as an adjunct to histopathology for astrocytoma grading and therapy planning. It seems that the optimal role of perfusion-weighted MR imaging for evaluation of astrocytomas would be twofold: (1) in patients with imaging and clinical presentation of a low-grade tumor, low CBV values are reassuring and may exclude the need for biopsy, as the patients would be followed without any therapeutic interventions; (2) in patients with histologically confirmed high-grade gliomas, low rCBV values may indicate relatively favorable prognosis and at the same time CBV values will serve as a biomarker for the treatment monitoring. Besides the classical histopathological grading, which implies the survival rates in patients with astrocytomas, it is clear that an outcome-based classification may be more useful and, thus, different prognostic factors could be of particular value in the heterogeneous population of the low-grade gliomas as well as in the anaplastic astrocytomas and glioblastomas in order to tailor the individual therapy regimen (Law et al., 2006; Pignatti et al., 2002). In a multi-institution setting patients with low-grade gliomas and astocytomas and an adverse event (progressive disease or death) had a significantly higher baseline rCBV than those without (complete response or stable disease) (p value = 0.0138) while the median time to progression among subjects with rCBV > 1.75 was 365 days

217

compared to the median time to progression of 889 days among subjects with rCBV < 1.75 (Caseiras et al., 2010). It was recently shown that CBV estimations might be able to distinguish between the stable and rapidly progressing gliomas among patients with the same histopathological grade (WHO Grade II) (Law et al., 2006). Specifically, patients with a relative CBV of less than 1.75 had a median time to progression of 3585 days, whereas patients with a relative CBV of more than 1.75 had a time to progression of 265 days irrespective of histopathologic tumor type. Lowgrade astrocytomas with relative CBV of less than 1.75 had a significantly longer time to progression than did low-grade astrocytomas with a relative CBV of more than 1.75. Similarly, high-grade astrocytomas with a relative CBV of less than 1.75 also had significantly longer time to progression than did astrocytomas with a relative CBV of more than 1.75. In the same study, the progression-free survival curves for astrocytomas of different histologic tumor type indicated that the progression-free survival was concordant with the histologic tumor type, with low-grade oligodendroglioma having the longest time to progression, followed by low-grade astrocytoma, low-grade oligoastrocytoma, anaplastic oligoastrocytoma, anaplastic astrocytoma, and GBM (Law et al., 2006). These results highlight again the need for homogenous patient populations in order to study the prognostic significance of the CBV measurements as well as the complexity of factors, besides the vascularity, that affect the survival of patients with astrocytomas; therefore, neither imaging or histopathology may act alone as reference standard for predicting survival and recurrence. Based on the results from this work (Law et al., 2006), Bisdas et al. (2009) verified in a mixed population of low-grade astrocytomas, oligoastrocytomas, and oligodendrogliomas expanded with WHO-grade III-IV astrocytomas the significance of the homogeneity in the study population. Although the rCBV was not significantly higher than in astrocytomas, like in another recent study (Dhermain et al., 2010), the inclusion of oligodendrogliomas and mixed oligoastrocytomas contaminated the correlation between histopathological grade and perfusion parameters and thus, the oligodendrogliomas and oligoastrocytomas were excluded from any further analysis. Moreover, the higher rCBV values of oligodendrogliomas were accompanied with longer progression-free survival and median survival time, which further biased the results.

218

S. Bisdas

As aforementioned, methodological issues may affect the optimal rCBV threshold for discrimination of astrocytomas with high probability of recurrence and one-year survival, which were found to be > 4.2 and ≤ 3.8 respectively (Bisdas et al., 2009) (Fig. 24.1), compared to those reported by Law et al. (2004) (1.75) and by Lev et al. (2004) (1.5). Recently, Dhermain et al. (2010) found a mean rCBV value of 1.47 and 1.78 in patients with low-grade astrocytomas and oligodendrogliomas with unfavorable prognosis [according to the European Organisation for Research and Treatment of Cancer (EORTC) criteria], respectively. Following the proposed rCBV cut-off value of 1.75, they found that on univariate (but not on multivariate) analysis high rCBV estimates were also negative prognostic factors associated with shorter progression-free survival (Dhermain et al., 2010). Notably, other authors claimed that a similar rCBV cut-off value (1.4) had no predictive value with respect to the prognosis in patients with glioblastomas as well as in mixed populations with gliomas, however, it is not clear whether they used maximum rCBV in their analysis (Mills et al., 2006; Oh et al., 2004). Closer to the cut-off values proposed by Bisdas et al. (2009), another study demonstrated that the 2-year overall survival rate was significantly higher for patients with low (≤2.3) than with high (>2.3) maximum rCBV, irrespective of whether the tumor was anaplastic astrocytoma or GBM (Hirai et al., 2008). Similarly, Fuss et al. (2001) suggested that high pretherapeutic angiogenic activity in low-grade fibrillary astrocytomas (tumor to white matter rCBV

>2.17) indicated a subgroup of tumors at higher risk for early local recurrence or malignant transformation after fractionated stereotactic radiotherapy. Saraswathy et al. (2009) evaluated various MR parameters that predict survival in patients with newly diagnosed GBM prior to adjuvant therapy and concluded that volume of tissue having rCBV >3 had a higher risk for poor outcome. An interesting finding in the study of Hirai et al. (2008) was that survival was significantly longer in patients with GBM and low rather than high maximum rCBV values. The importance of these results lies on the preliminary evidence that, unlike histopathologic findings, maximum rCBV values may make it possible to predict the survival of patients with high-grade astrocytomas. Furthermore, CBV measurements enable the detection of anaplastic astrocytomas harboring components manifesting a high maximum rCBV, which may be associated with a poor prognosis and thus, may require the same aggressive treatment as GBM. The correlation between histopathological grade and progression-free survival has been shown to be weaker and less significant than the correlations of progression-free survival with rCBV measurements (Bisdas et al., 2009). Specifically, the relative risk for having recurrence was 11.1 times higher for patients with rCBVmax > 4.2 (p value = 0.0007) while the shorter progression-free survival was 6.7 times increased in patients with WHO-grades III and IV compared to WHO grades I and II (p value = 0.05). These results support the hypothesis that MR-derived perfusion parameters may be comparable or even

Fig. 24.1 T2-weighted fluid-attenuation inversion recovery (FLAIR) MR image (a) and color-coded cerebral blood volume (CBV) map (overlaid on the corresponding FLAIR image) (b) of a male patient with a supratentorial astrocytoma (WHO grade IV) in the left hemisphere shows a markedly hyperperfused area

in the tumor. Progression-free survival curves in patients with low- and high-grade astrocytomas have shown the significantly longer time to progression by dichotomizing the patients according to a rCBVmax cut-off value of <4.2 (p value = 0.0001) (Bisdas et al., 2009)

24 Astrocytomas: Predicting Survival and Recurrence Using Cerebral Blood Volume Measurements

better than histopathological grading as survival predictors and are in agreement with a previous study (Chaskis et al., 2006) that found significant correlation between perfusion-weighted MR measurements and final outcome (p value < 0.01). In concordance with these results, other studies, even in pediatric populations, showed the significance of rCBV estimates as independent multivariate predictor of outcome and disease progression together with histopathologic grading (Hirai et al., 2008; Tzika et al., 2004). This arises an important aspect addressed by a recent study, namely, the combined use the WHO-grading system and the rCBV criterion for an improved performance of the one-year survival and recurrence prediction rates (Bisdas et al., 2009). According to the authors, four groups in the astrocytomas population were defined as follows: group 1 with tumors WHO-grade = I-II and rCBVmax ≤ 4.2, group 2 WHO-grade = I-II and rCBVmax > 4.2, group 3 WHO-grade = III-IV and rCBVmax ≤ 4.2, and group 4 WHO-grade = III-IV and rCBVmax > 4.2. The results showed that the histopathological grading-CBV joint classification provided statistically significant results for both the one-year survival and recurrence (Bisdas et al., 2009). Using this classification system, the patients in groups 1 and 3 had a significantly higher progression-free survival (761 days) compared to the progression-free survival in patients in group 4 (135 days) (p value = 0.0001). The patients with high-grade astrocytomas and rCBVmax > 4.2 had a 20 times increased relative risk for a shorter progression-free survival than the patients with astrocytomas and low rCBV, regardless of the grade (p value < 0.0001). The rCBVmax in patients stratified in groups 1 and 3 (2.11 ± 0.99) was significantly lower than the rCBVmax in patients of group 4 (6.98 ± 2.38) (p value = 0.0005) (Bisdas et al., 2009). Thus, it seems that the combination of rCBV values and histological grading appears to provide the strongest predictive value for recurrence/progression and this issue has to be elucidated in future studies. There is also initial evidence that the rCBV measurements may not only have a predictive role as a baseline parameter but the longitudinal changes in the rCBV values obtained after combined radiation and temozolamide therapy may correlate with the overall survival (Mangla et al., 2010). Specifically, increased rCBV after treatment was shown to be a strong predictor of poor survival (median survival, 235 days versus 529 days with decreased rCBV) (p value < 0.008, log-rank test). The ROC curves for 1-year

219

survival also showed a greater area under the curve (p value = 0.005) with rCBV than with tumor size. The longitudinal perfusion-weighted MR imaging has also recently received increased attention as the significant increases in rCBV up to 12 months before contrast enhancement appearance on T1-weighted images may predict the malignant transformation of low-grade astrocytomas and thus, the survival rates (Danchaivijitr et al., 2008), though other authors advocate conventional imaging criteria, like the six-month tumor growth, as more important (Brasil Caseiras et al., 2009). In conclusion, preoperative grading of gliomas based on conventional MR imaging is often unreliable, whereas histopathologic grading frequently suffers from inherent shortcomings. Independently and in combination with histopathology, rCBV measurements on pre- and postreatment MR imaging scans in patients with astrocytomas demonstrate the significance of cerebral blood volume estimates as a clinical, objective biomarker for predicting the survival and recurrence in these patients. The combined assessment of histopathologic and perfusion MR imaging findings obtained before the inception of treatment may be useful to determine optimal management strategies in patients with high-grade astrocytomas. On the other hand, the response to therapy as assessed by CBV measurements may also detect the early non-responders and thus, lead to a modification of the initial therapy regimen. Threshold values, as proposed in literature and under consideration of possible methodological flaws, may be used in a clinical setting to evaluate tumors preoperatively for histologic grade and provide a means for guiding treatment and predicting postoperative patient outcome. Ongoing data collection from longitudinal studies is crucial to determine if rCBV can, in the long run, be superior to histopathologic examination in predicting tumor behavior and patient prognosis.

References Aronen HJ, Glass J, Pardo FS, Belliveau JW, Gruber ML, Buchbinder BR, Gazit IE, Linggood RM, Fischman AJ, Rosen BR (1995) Echo-planar MR cerebral blood volume mapping of gliomas. Clinical uitility. Acta Radiol 36:520–528 Bisdas S, Kirkpatrick M, Giglio P, Welsh C, Spampinato MV, Rumboldt Z (2009) Cerebral blood volume measurements by

220 perfusion-weighted MR imaging in gliomas: ready for prime time in predicting short-term outcome and recurrent disease? AJNR Am. ANJR Am J Neuroradiol 30:681–688 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 Brookes MJ, Morris PG, Gowland PA, Francis ST (2007) Noninvasive measurement of arterial cerebral blood volume using Look-Locker EPI and arterial spin labeling. Magn Reson Med 58:41–54 Bruening R, Kwong KK, Vevea MJ, Hochberg FH, Cher L, Harsh GRt, Niemi PT, Weisskoff RM, Rosen BR (1996) Echo-planar MR determination of relative cerebral blood volume in human brain tumors: T1 versus T2 weighting. AJNR Am J Neuroradiol 17:831–840 Caseiras GB, Chheang S, Babb J, Rees JH, Pecerrelli N, Tozer DJ, Benton C, Zagzag D, Johnson G, Waldman AD, Jager HR, Law M (2010) Relative cerebral blood volume measurements of low-grade gliomas predict patient outcome in a multi-institution setting. Eur J Radiol 73:215–220 Chaskis C, Stadnik T, Michotte A, Van Rompaey K, D’Haens J (2006) Prognostic value of perfusion-weighted imaging in brain glioma: a prospective study. Acta Neurochir (Wien) 148:277–285, discussion 285 Danchaivijitr N, Waldman AD, Tozer DJ, Benton CE, Brasil Caseiras G, Tofts PS, Rees JH, Jager HR (2008) Low-grade gliomas: do changes in rCBV measurements at longitudinal perfusion-weighted MR imaging predict malignant transformation?. Radiology 247:170–178 Dhermain F, Saliou G, Parker F, Page P, Hoang-Xuan K, Lacroix C, Tournay E, Bourhis J, Ducreux D (2010) Microvascular leakage and contrast enhancement as prognostic factors for recurrence in unfavorable low-grade gliomas. J Neurooncol 97:81–88 Emblem KE, Scheie D, Due-Tonnessen P, Nedregaard B, Nome T, Hald JK, Beiske K, Meling TR, Bjornerud A (2008) Histogram analysis of MR imaging-derived cerebral blood volume maps: combined glioma grading and identification of low-grade oligodendroglial subtypes. AJNR Am J Neuroradiol 29:1664–1670 Fuss M, Wenz F, Essig M, Muenter M, Debus J, Herman TS, Wannenmacher M (2001) Tumor angiogenesis of low-grade astrocytomas measured by dynamic susceptibility contrastenhanced MRI (DSC-MRI) is predictive of local tumor control after radiation therapy. Int J Radiat Oncol Biol Phys 51:478–482 Galban CJ, Chenevert TL, Meyer CR, Tsien C, Lawrence TS, Hamstra DA, Junck L, Sundgren PC, Johnson TD, Ross DJ, Rehemtulla A, Ross BD (2009) The parametric response map is an imaging biomarker for early cancer treatment outcome. Nat Med 15:572–576 Hacklander T, Reichenbach JR, Modder U (1997) Comparison of cerebral blood volume measurements using the T1 and T2∗ methods in normal human brains and brain tumors. J Comput Assist Tomogr 21:857–866 Hirai T, Murakami R, Nakamura H, Kitajima M, Fukuoka H, Sasao A, Akter M, Hayashida Y, Toya R, Oya N, Awai K, Iyama K, Kuratsu JI, Yamashita Y (2008) Prognostic

S. Bisdas value of perfusion MR imaging of high-grade astrocytomas: long-term follow-up study. AJNR Am J Neuroradiol 29:1505–1510 Kassner A, Annesley DJ, Zhu XP, Li KL, Kamaly-Asl ID, Watson Y, Jackson A (2000) Abnormalities of the contrast re-circulation phase in cerebral tumors demonstrated using dynamic susceptibility contrast-enhanced imaging: a possible marker of vascular tortuosity. J Magn Reson Imaging 11:103–113 Knopp EA, Cha S, Johnson G, Mazumdar A, Golfinos JG, Zagzag D, Miller DC, Kelly PJ, Kricheff II (1999) Glial neoplasms: dynamic contrast-enhanced T2∗ -weighted MR imaging. Radiology 211:791–798 Law M, Oh S, Babb JS, Wang E, Inglese M, Zagzag D, Knopp EA, Johnson G (2006) Low-grade gliomas: dynamic susceptibility-weighted contrast-enhanced perfusion MR imaging–prediction of patient clinical response. Radiology 238:658–667 Law M, Yang S, Babb JS, Knopp EA, Golfinos JG, Zagzag D, Johnson G (2004) Comparison of cerebral blood volume and vascular permeability from dynamic susceptibility contrastenhanced perfusion MR imaging with glioma grade. AJNR Am J Neuroradiol 25:746–755 Lev MH, Ozsunar Y, Henson JW, Rasheed AA, Barest GD, Harsh GRt, Fitzek MM, Chiocca EA, Rabinov JD, Csavoy AN, Rosen BR, Hochberg FH, Schaefer PW, Gonzalez RG (2004) Glial tumor grading and outcome prediction using dynamic spin-echo MR susceptibility mapping compared with conventional contrastenhanced MR: confounding effect of elevated rCBV of oligodendrogliomas [corrected]. AJNR Am J Neuroradiol 25:214–221 Li KL, Zhu XP, Checkley DR, Tessier JJ, Hillier VF, Waterton JC, Jackson A (2003) Simultaneous mapping of blood volume and endothelial permeability surface area product in gliomas using iterative analysis of first-pass dynamic contrast enhanced MRI data. Br J Radiol 76:39–50 Ludemann L, Grieger W, Wurm R, Wust P, Zimmer C (2006) Glioma assessment using quantitative blood volume maps generated by T1-weighted dynamic contrast-enhanced magnetic resonance imaging: a receiver operating characteristic study. Acta Radiol 47:303–310 Mangla R, Singh G, Ziegelitz D, Milano MT, Korones DN, Zhong J, Ekholm SE (2010) Changes in relative cerebral blood volume 1 month after radiation-temozolomide therapy can help predict overall survival in patients with glioblastoma. Radiology 256:575–584 Mills SJ, Patankar TA, Haroon HA, Baleriaux D, Swindell R, Jackson A (2006) Do cerebral blood volume and contrast transfer coefficient predict prognosis in human glioma? ANJR Am J Neuroradiol 27:853–858 Oh J, Henry RG, Pirzkall A, Lu Y, Li X, Catalaa I, Chang S, Dillon WP, Nelson SJ (2004) Survival analysis in patients with glioblastoma multiforme: predictive value of cholineto-N-acetylaspartate index, apparent diffusion coefficient, and relative cerebral blood volume. J Magn Reson Imaging 19:546–554 Paulson ES, Schmainda KM (2008) Comparison of dynamic susceptibility-weighted contrast-enhanced MR methods: recommendations for measuring relative cerebral blood volume in brain tumors. Radiology 249:601–613

24 Astrocytomas: Predicting Survival and Recurrence Using Cerebral Blood Volume Measurements Persigehl T, Wall A, Kellert J, Ring J, Remmele S, Heindel W, Dahnke H, Bremer C (2010) Tumor blood volume determination by using susceptibility-corrected DeltaR2∗ multiecho MR. Radiology 255:781–789 Pignatti F, van den Bent M, Curran D, Debruyne C, Sylvester R, Therasse P, Afra D, Cornu P, Bolla M, Vecht C, Karim AB (2002) Prognostic factors for survival in adult patients with cerebral low-grade glioma. J Clin Oncol 20:2076–2084 Saraswathy S, Crawford FW, Lamborn KR, Pirzkall A, Chang S, Cha S, Nelson SJ (2009) Evaluation of MR markers that predict survival in patients with newly diagnosed GBM prior to adjuvant therapy. J Neurooncol 91:69–81 Stupp R, Roila F (2009) Malignant glioma: ESMO clinical recommendations for diagnosis, treatment and follow-up. Ann Oncol 20(Suppl 4):126–128 Tzika AA, Astrakas LG, Zarifi MK, Zurakowski D, Poussaint TY, Goumnerova L, Tarbell NJ, Black PM (2004)

221

Spectroscopic and perfusion magnetic resonance imaging predictors of progression in pediatric brain tumors. Cancer 100:1246–1256 Uematsu H, Maeda M, Sadato N, Matsuda T, Ishimori Y, Koshimoto Y, Kimura H, Yamada H, Kawamura Y, Yonekura Y, Itoh H (2001) Blood volume of gliomas determined by double-echo dynamic perfusion-weighted MR imaging: a preliminary study. AJNR Am J Neuroradiol 22:1915–1919 Wetzel SG, Cha S, Johnson G, Lee P, Law M, Kasow DL, Pierce SD, Xue X (2002) Relative cerebral blood volume measurements in intracranial mass lesions: interobserver and intraobserver reproducibility study. Radiology 224:797–803 Young R, Babb J, Law M, Pollack E, Johnson G (2007) Comparison of region-of-interest analysis with three different histogram analysis methods in the determination of perfusion metrics in patients with brain gliomas. J Magn Reson Imaging 26:1053–1063

Chapter 25

Electronic Patient-Reported Outcome Monitoring (ePROM) in Brain Tumour Patients Lisa M. Wintner, Johannes M. Giesinger, Gabriele Schauer-Maurer, and Bernhard Holzner

Abstract The diagnosis of brain tumour commonly goes along with short survival, bad outcome prognosis and strongly impaired health-related quality of life (HRQOL) due to the disease itself, anti-cancer treatment or ancillary medication. Patient-reported outcome monitoring (PROM) ensures the capture of individual problematic issues threatening patients’ HRQOL for a targeted and patient-tailored intervention. Practical barriers of PROM can be avoided by electronic data capture (ePROM), which saves time, human resources and can help to deepen patientphysician communication and to improve patients’ satisfaction with care. Furthermore, tele-monitoring incorporates usually neglected time points when patients are discharged from hospital and symptom burden is only partly communicated to physicians and nurses. For comprehensive health care of brain tumour patients the usage of ePROM and tele-monitoring could help to meet the patients’ needs. Keywords ePROM · quality of life · computerized · Brain tumours

Introduction In 2008, brain tumours and cancer of the nervous system ranged number 15 of the most often diagnosed cancer types within the European Union (Ferlay

B. Holzner () Department of Psychiatry and Psychotherapy, Innsbruck Medical University, Anichstr.35, 6020 Innsbruck, Austria e-mail: [email protected]

et al., 2010). Compared to the five most common cancer types (colorectum, breast, prostate, lung and bladder) the incidence rate of brain tumours appears to be small-sized, however, the mortality rate for brain and nervous systems cancers is in percentage terms higher than those of most high-incidencecancers. Survival expectance depends strongly on the WHO grade of the tumour and ranges from possibly no impact on life expectancy (grade I) to a survival of only 10 months (grade IV) (Reardon and Wen, 2006). Therefore, although brain cancer is only seldom diagnosed, patients experience an extremely high negative impact as survival prognosis is poor and deteriorated HRQOL is common (Taphoorn et al., 2010b). There is a large variety of brain tumours and their differentiation between primary tumours, which originate directly from brain tissue, and secondary tumours, which are brain metastasis from other malignant diseases, is often difficult to make. However, their distinction is of particular importance since the most favourable treatment modalities e.g. in glioblastoma and in brain metastasis differ strongly and inadequate intervention endangers patients’ HRQOL. The treatment possibilities of brain tumours include surgery, chemotherapy, radiotherapy and supplementary medical therapies. Not only the tumour itself, but also treatment-related side effects and adverse events caused by supportive medication (e.g. steroids and antiepileptic drugs) confront the patient with a high level of physical and psychosocial burden. Brain tumour patients are affected by a variety of tumour symptoms, which might be caused by e.g. increased intracranial pressure (e.g. headache, anorexia, nausea, fatigue, vomiting, seizures, sleeping longer at night, drowsiness, napping during the day). Other

M.A. Hayat (ed.), Tumors of the Central Nervous System, Volume 5, DOI 10.1007/978-94-007-2019-0_25, © Springer Science+Business Media B.V. 2012

223

224

physical impairments may occur due to focal neurological deficits (e.g. motor deficit, aphasia, visual field defects). Even more burdensome both for patients and caregivers might be symptoms like personality changes, mood disturbances, decrease in mental capacity and concentration, cognitive dysfunction, anxiety and depression (Heimans and Taphoorn, 2002). These symptoms were found to negatively influence both patients’ and their carers’ overall QOL if compared to general population (Janda et al., 2007). As brain tumour patients cannot be cured in most cases, the treatment has to focus on maintaining HRQOL in at least the same extent as on life prolongation.

Patient-Reported Outcomes Patient-reported outcomes (PROs) allow the comprehensive evaluation of patients’ perception of aspects of functioning and well-being in regard to their health status, disease, and its treatment. The U.S. Food and Drug Administration presents a short and concise definition of PROs: “A PRO is a measurement of any aspect of a patient’s health status that comes directly from the patient (i.e., without the interpretation of the patient’s responses by a physician or anyone else).” (U.S. FOOD AND DRUG ADMINISTRATION, 2006). In this respect, PROs comprise a variety of directly reported health-related issues: disease-related symptoms, treatment-related adverse events, functioning, well-being, HRQOL, perceptions about treatment, satisfaction with care and professional communication (Rothman et al., 2007). The assessment of HRQOL issues include numerous domains like depression, anxiety, pain, fatigue, gastrointestinal symptoms, social functioning, or perceived cognitive dysfunction. PROs are therefore versatile in application possibilities, as they may be used for adverse event detection in drugsafety reports and medical product development, for HRQOL evaluation both in clinical research and routine, and for guidance in medical decision making. PRO-data may not only support patients in choosing between available treatments, but also enhance understanding of the disease and improve its treatment and the management of treatment-related symptom burden. The usage of PROs measurement provides several benefits, as patients namely favour the discussion of

L.M. Wintner et al.

HRQOL domains with their clinician, but often do not take the initiative themselves. Concerning emotional functioning, daily activities and familial issues approximately 25–29% of patients wish their clinician to start discussion and even 37% of patients expect their clinician to address the topic of social functioning (Detmar et al., 2000). In this respect PROs can successfully contribute to satisfaction of clinicians and patients, as clinicians attain a deeper insight in patient’s needs and may properly address relevant topics without asking the patient a variety of possibly irrelevant questions during the face-to-face consultation. The usage of PROs may also sharpen clinicians’ awareness for psychological, social and spiritual functioning and diminish the domination of physical functioning in patientdoctor conversation. In clinical routine the usage of PROs already showed a number of benefits. Based on PRO-data clinicians adjusted the dosage of analgesics in a more sophisticated way than without PRO-data (Trowbridge et al., 1997). The patient-clinician communication was improved and enhanced concerning discussed domains, if clinicians took advantage of provided PRO-data. Furthermore patients felt more emotional support, clinicians became more sensitive for normally underestimated HRQOL domains (Detmar et al., 2002), patients were more satisfied with care and the building of a relationship to the clinician (Velikova et al., 2010) and reported a better QOL (Velikova et al., 2004). The literature for HRQOL in brain tumour patients grows considerably and reports a plenitude of results: they are partly consistent and matching but also puzzling and contradictory. HRQOL and its sub-domains are reported to depend on tumour location and laterality, on low- or high-grade classification, the tumour itself and its histological type. Derived from results of clinical trials the disease itself seems to have the major negative impact on HRQOL, which can be improved by treatment, whereby treatment side-effects may limit the extent of the benefits (Taphoorn et al., 2010b). Due to the seldom occurrence of brain tumour and the differentiated types of brain tumours, scientific investigations on large patient populations with a homogeneous diagnosis is very difficult to obtain. Regardless of this methodological constraint contemporary studies report brain tumour survivors to suffer from cognitive deficits and impairs HRQOL e.g. increased fatigue or depression (Taphoorn et al., 2010b). Only low baseline QOL scores are reported for brain tumour patients after

25 Electronic Patient-Reported Outcome Monitoring (ePROM) in Brain Tumour Patients

surgery and before adjuvant therapy. It is supposed that the site of disease (well operable or badly surgically accessible) influences patients’ QOL and not surgery itself (Budrukkar et al., 2009).

225

Index (FBrSI) for the FACT-G) (see Table 25.1 for details).

EORTC QLQ-C30 and BN20 Cancer-Targeted and Brain Tumour-Specific PROs For HRQOL-assessment in cancer patients usually a generic PRO core-questionnaire is administered, which can be supplemented by disease-specific questionnaire-modules. Two internationally used, validated and reliable cancer-targeted instruments are the European Organization for Research and Treatment of Cancer core questionnaire (EORTC QLQ-C30) (Aaronson et al., 1993) and the Functional Assessment of Cancer Therapy general version (FACT-G) (Cella et al., 1993). Both instruments can be expanded with a brain cancer-specific module (BN20 for the EORTC QLQ-C30 and FACT-Br and the FACT-Br Symptom

The EORTC QLQ-C30 is a validated and widely used PRO instrument, which was originally developed to evaluate QOL in cancer patients participating in clinical trials (Aaronson et al., 1993). A modular approach was used for questionnaire design. The core questionnaire, which consists of thirty questions, is the basic part of the QOL-instrument and can be expanded by additional elements, which supplementary focus diagnosis-specific symptoms and impairments. The QLQ-C30 consists of five functioning scales (physical, role, social, emotional and cognitive functioning), a scale for global QOL, three symptom scales and six single item symptoms. Except for the two questions about general QOL every question has four response options: “not at all”, “a little”, “quite a bit” and “very

Table 25.1 Commonly used PRO instruments Instrument Scales

Val.

Items

Period

FACT-G Functional Assessment of Cancer Therapy-General

– – – –

Physical well-being Social/Family well-being Emotional well-being Functional well-being

x

27

1 week

FACT-Br Functional Assessment of Cancer Therapy-Brain Module

– – – – –

Cognitive functioning Neurological functioning Sensory functioning Psychological functioning Impact of changes in functioning on daily living.

x

23

1 week

FBrSI FACT-Br Symptom Index

– Symptom list

x

15

1 week

EORTC QLQ-C30 European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire Core 30

– Functioning (physical, role, social, emotional and cognitive) – Global QOL – Pain, fatigue, nausea/vomiting – Dyspnoea, sleeping disturbances, appetite loss, constipation, diarrhoea and financial impact

x

30

1 week

EORTC QLQ-BN20 European Organisation for Research and Treatment of Cancer Quality of Life Brain Module

– – – – –

x

20

1 week

Future uncertainty Visual disorder Communication deficit Motor dysfunction) Headache, seizure, drowsiness, hair loss, itching, weakness of both legs, difficulty controlling bladder function

226

much”. The overall QOL scale asks the patient to rate how they experienced their general QOL and physical condition between 0 “very poor” and 6 “excellent”. Except the physical functioning scale, all questions shall be rated in regard to the week before the examination date. The EORTC Brain Cancer Module (BN20) contains 20 items, which focus on symptoms that particularly concern brain cancer patients under chemotherapy or radiotherapy. Therefore both disease symptoms and treatment toxicities are included in the item list. The items are grouped into four scales (future uncertainty, visual disorder, communication deficit, motor dysfunction) and seven single items (headache, seizure, drowsiness, hair loss, itching, weakness of both legs, difficulty controlling bladder function). The QLQ-BN20 proved adequate psychometric properties in multi-national and multi-lingual validation study (Taphoorn et al., 2010a).

FACT-G and FACT-Br/FBrSI The FACT-G is a cancer-specific PRO instruments provided by the Functional Assessment of Chronic Illness Therapy (FACIT) measurement system, which supplies questionnaires concerning a variety of chronic illnesses and their conditions (cancer, HIV/AIDS, multiple sclerosis) (Webster et al., 2003). Nowadays the fourth version of the FACT-G is widely used and comprises 27 items grouped into four primary QOL domains (physical well-being, social/family well-being, emotional well-being, and functional wellbeing). The patient has to rate each item on a five-point Likert-scale between “not at all”, “a little bit”, “somewhat”, “quite a bit” and “very much” in relation to symptom severity during the last week. The FACT-G was examined concerning its ease of administration, brevity, reliability, validity and responsiveness to clinical change and was found to fit all the stipulated requirements (Cella et al., 1993). The questionnaire’s validity was also tested for a variety of language versions (Sanchez et al., 2011). The brain cancer-specific FACT-Br module consists of 23 items and explores how patients perceive their cognitive, neurological, sensory and psychological functioning and the impact of changes in these domains on their daily living. The FACT-Br Symptom Index (FBrSI) is available for symptom

L.M. Wintner et al.

rating of brain tumour patients. Following symptoms are included: headaches, seizures, weakness of extremities, self-caring issues, lack of energy, difficulties with expression, trouble with coordination, frustration due to limited agency, nausea, aphasia, hopelessness, social functioning, anxiety and enjoyment of life. Both the FACT-Br and FBrSI proved to be reliable, valid and responsive to change (Nickolov et al., 2005).

Computer-Based QOL Monitoring: ePROM Although many advantages of PRO assessments can be communicated, there are still popular counterarguments circulating. Even if PRO-data is available many clinicians do not pay attention to them because of lack of time, human resources and an adequate PRO instrument and the assumption that directly from patients obtained information does not add any additional value (Luckett et al., 2009). Furthermore some clinicians argue that information on HRQOL are not of same importance as treatment decisions, equality of PRO instruments is doubtful and the methodology of PRO measurement seems to be dubious (Barlesi et al., 2006). These objections can be devitalised, though. Meanwhile a broad variety of internationally validated and widely used PRO instruments is available (especially for cancer population the EORTC QLQ-C30 and the FACT-G with their supplemental modules). The use of PRO instruments does usually not or only a few minutes prolong the clinical appointment (Frost et al., 2007). Particularly the implementation of electronic patient-reported outcome monitoring (ePROM), which means the routinely and electronic collection of PRO-data on a systematic basis, solves the problem of time and resource constraints. Data are directly entered by patients, scores are calculated automatically and immediate information processing is possible. Electronic data capture was shown to need less time for instrument completion than paper-pencil versions (Velikova et al., 1999). The usage of ePROM is well accepted by patients (Mullen et al., 2004; Velikova et al., 2010), equivalent to paper-pencil versions (Coons et al., 2009) and valid (Abernethy et al., 2010). Also in brain tumour populations the usage of ePROM was feasible and useful for both patient

25 Electronic Patient-Reported Outcome Monitoring (ePROM) in Brain Tumour Patients

and clinician (Erharter et al., 2010; Holzner et al., 2011). Furthermore, the ePROM allows the use of computer-adaptive testing (CAT), which reduces the burden of PRO completion for patients, as the questionnaire is fitted to the patients’ individual needs and health condition. Depending on the answers to preceding questions appropriate following questions are chosen from a PRO-item bank. Currently a large project funded by the US-American National Institutes of Health (NIH) is developing generic computeradaptive PRO-instruments for use in the chronical ill (www.nihpromis.org). Administrative failure is reported as the most important reason for missing data (72.2%) in PRO-data collection in patients suffering from malignant glioma (Walker et al., 2003). About 21.7% of patients did not fill out the PRO-instruments due to their very bad health condition and only 6.1% of patients refused to participate in PRO-assessments. The major problem of administrative failures in this study can therefore be traced to irregular administration time points of PROs, too little explanation of instrument completion and missing specialized staff for QOL research who checks questionnaires completeness. Such barriers can easily be overcome with the routine usage of ePROM, which avoid a high percentage of missing data, as a standardized questionnaire explanation is integrated in the procedure and further questions are only displayed after all preceding questions have been answered. Additionally, the implementation of ePROM for the assessment of HRQOL or adverse events in clinical drug evaluation studies facilitates and accelerates data flow because complicated and defective data collection is cut short. Without ePROM clinicians ask patients about their symptoms and write a construed summary of these symptoms down in patients’ charts, from which research assistants collect information and consign them into research data bases (Trotti et al., 2007). This process is highly endangered to lose and/or alter information given by patients and needs much more human and time resources than the usage of ePROM. Admittedly, at the beginning of ePROM implementation some time burden may be set on clinicians and nurses, as they need to be trained in software handling and result interpretation, but after a phase of familiarization ePROM contributes in time saving. Firstly, the completion of a PRO instrument

227

encourages the patient to reflect more detailed on his/her health status and symptom burden and facilitates in that way the communication between patient and clinician/nurse about relevant discussion areas. Secondly, ePROM is immediately accessible for the clinician, whose attention can be called to clinically relevant deteriorations or improvements. In fact, such an alert system helps clinicians to focus on issues important for the patient without long-winded enquiry of possible difficulties. Moreover, queue times of patients might meaningfully be padded by ePROM completion.

EPRO Software For obtaining the highest grade of feasibility and utility of PRO-assessments, the use of specialized software that fits perfectly the needs of both patients and clinicians is obligatory. In the last years there have been a few attempts on constructing and implementing a software solution for QOL-monitoring in clinical routine undertaken e.g. by Joerg Sigle (AnyQuest), Galina Velikova and Irma Verdonck (OncoQuest). These tools using touch screens have shown feasibility in daily clinical practice and implementation studies suggest some important benefits for the physicians, the patients and medical care. A further example for a specific software solution is the Computer-based Health Evaluation System (CHES). The CHES program has especially been developed for electronic PRO-data capture and offers a variety of useful features for clinical routine and research purposes. Any required paper-pencil questionnaire can be implemented into CHES to facilitate all steps from data collection to result calculation, interpretation and report generation as well. CHES provides a database (e.g. MySQL or Oracle) for supplementary medical and psychosocial data. This database is particularly useful for research purposes, as it improves study logistics, reduces the need for human resources and increased data quality. Database connection can be established via LAN or Wifi. Connection via LAN is more laborious as the study nurse needs to connect the tablet PC to the LAN with the purpose of preparing the patient list to who the questionnaire shall be administered. The tablet PC has to be disconnected, handed over to patients for bed-side assessment and again connected to the

228

LAN for uploading the collected information into the database. A Wifi-connection is much more comfortable, as the database can be updated anytime without special constraints. Connection via WiFi eases work for both data collectors and clinicians, as data is transferred instantly and the immediate access to patients’ data is possible. In clinical routine the use of specialized software as CHES offers several advantages: it decreases the need of personal and time resources, diminished possible error sources and missing data, accelerates the calculation of PRO-data and immediately provides easily interpretable results by means of eyecatching and comprehensible graph charts. By means of touch screen equipped tablet PCs patients can easily complete the PRO assessment. Missing data is prevented by forwarding only to the remaining questions when all prior items were answered. The possibility to fit font and button sizes to the needs of different patients groups allow that readability and handling of the assessment is practicable for e.g. elderly people as well. As no computer literacy of patients is needed for PRO completion, in principle all patients are able to participate in ePROM. Due to the instantly performed data calculation, clinicians directly can be informed after instrument completion about clinically remarkable PRO-data that need adequate intervention. Furthermore individual PRO-data can be displayed as longitudinal as well as cross-sectional illustrations. The electronic administration of PROs by means of CHES was well accepted by brain tumour patients (Erharter et al., 2010). The on average required time of ten minutes for completion of the EORTC QLQ30 and the QLQ-BN20 even decreased with repeated instrument administration. Clinicians reported the ePRO to be beneficial, as loss of bladder controlling would not have been detected adequately without ePRO. Additionally CHES was implemented for routine HRQOL monitoring at the neuro-oncological outpatient unit of the Medical University Clinic of Neurology and reported to be feasible and profitable for patient care (Holzner et al., 2011).

Pro-Tele-Monitoring Although PROs and ePROM work hard on the routinely incorporation of the patients’ individual per-

L.M. Wintner et al.

spective in clinical practice and cancer clinical trials, there is still a piece missing. The usage of ePROM should not be limited by the walls of hospitals and research institutions, but can also be extended into the domestic environment of patients. The step into the real world of patients is absolutely necessary to capture all possible time points, which can be of special importance for patients’ well-being and adequate medical assistance. Especially with regard to chemotherapy, side-effects and related symptom burden are known to be most severe a few days after application of cytostatic drugs (Hawkins and Grunberg, 2009); a time point at which most patients are already back at their home environment. This gap cannot be bridged by traditional ePROM, since periods without attendance of an in- or outpatient unit of the hospital are of interest. The missing knowledge about patients’ well-being at their home environment may lead to an underestimation of perceived symptom burden. This shortcoming can be overcome by a web-based assessment tool or phone-based home monitoring (so called tele-monitoring). Measuring HRQOL frequently may increase longitudinal information, and potentially enable clinicians to identify early signs of adverse events and thus intervene prospectively to minimize complications. Despite the early development-stage of telemonitoring, few studies document already its usage within oncological care. One study with patients who received hematopoietic stem cell transplantation stands out for its complex design and favourable results (Bush et al., 2005). They assessed QOL online on a daily basis by means of short and dynamic question-sets and monthly via a comprehensive questionnaire. Data show that not only the less sick patients were very compliant with online QOL, since only three patients discontinued study participation due to their bad health condition. The study reports high feasibility of telemonitoring, good patient compliance and high user satisfaction. The close-meshed assessment of patients’ health and well-being during their stays at home will require additional administrative resources. The internet, though, provides an inexpensive technology that allows a simple, user friendly and reliable data collection. Frequent HRQOL assessment is via tele-monitoring much easier and more inexpensive practicable than via traditional paper-pencil questionnaires. The use of tele-monitoring ensures that patients’ HRQOL outside the hospital setting is no

25 Electronic Patient-Reported Outcome Monitoring (ePROM) in Brain Tumour Patients

longer neglected, and that it can easily be integrated into symptom estimation, monitoring and treatment.

Usefulness of Proxy Ratings and Future Directions It is undoubtedly that ePROM must be restricted to patients who are able to report their experiences. Due to the nature of brain tumours the usage of ePROM is not possible in some patients over the whole course of disease, since cognitive impairments are next to a poor physical condition often occurring disease symptoms. Patients who suffer from a lack of concentration, thought disorder, communication deficits or visual disorders cannot successfully participate in ePROM. At this point, the usage of proxy-rating can be useful to maintain the assessment of the patients’ perspective for symptom management. Proxies like spouses, children, family members or close friends can bridge the gap in communication arising from patients’ restricted agency. Studies provide evidence, that agreement between patients and proxies are moderate and good, whereas proxies tend to underestimate patients perceived HRQOL. For patients with mental confusion, cognitive impairment, poor performance status and motor deficits only low agreement rates with significant others were found. Background characteristics like age, gender, culture, the relationship between patient and proxy or housing situation did not influence patient-proxy agreement. One study reveals a differentiation of HRQOL sub-domains in regard to patient-proxy agreement (Giesinger et al., 2009). More obvious issues and physical symptoms like physical functioning, sleeping disturbances, appetite loss, constipation, financial impact and taste alterations showed appropriate agreement. Less noticeable aspects like emotional functioning, cognitive functioning, fatigue, pain, dyspnoea and seizures had only low agreement between patients and proxies. To fully benefit from all advantages ePROM and tele-monitoring have to offer, the implementation of ePROM in clinical routine and telemonitoring in patient-home-assistance is one major step. Additionally, further research on ePROM in brain tumour patients has to focus on establishing reference scores for patients with different diagnoses undergoing

229

different treatments. The steadily increasing role of ePROM in medical decision-making and the constantly ongoing software development encourages the use of these measures with neurooncological patients on a regular basis. A special focus should be put on evaluation studies to investigate in detail the impact of ePROM on medical care. Clinicians’ contribution to ePROM conduction and use of PRO-data is in the same extent necessary as patients’ compliance, and further research is needed to examine how ePROM can effectively and smoothly be integrated in clinicians’ daily routine. Although studies have already shown that routine ePROM usage is feasible, physicians and nurses in the neurooncological setting need to be directly addressed to play an active role in PRO-data collection and application in clinical practice. A sophisticated knowledge about the effect of ePROM on the patientphysician communication and on the administration of medical and psychosocial interventions would help to provide patient-tailored and satisfactory intervention and improve HRQOL.

References Aaronson NK, Ahmedzai S, Bergman B, Bullinger M, Cull A, Duez NJ, Filiberti A, Flechtner H, Fleishman SB, de Haes JC (1993) The European Organization for Research and Treatment of Cancer QLQ-C30: a quality-of-life instrument for use in international clinical trials in oncology. J Natl Cancer Inst 85(5):365–376 Abernethy AP, Zafar SY, Uronis H, Wheeler JL, Coan A, Rowe K, Shelby RA, Fowler R, Herndon JE 2nd (2010) Validation of the Patient Care Monitor (Version 2.0): a review of system assessment instrument for cancer patients. J Pain Symptom Manage 40(4):545–558 Barlesi F, Tchouhadjian C, Doddoli C, Astoul P, Thomas P, Auquier P (2006) Quality of life: attitudes and perspectives of doctors in a thoracic oncology regional care network. Sante Publique 18(3):429–442 Budrukkar A, Jalali R, Dutta D, Sarin R, Devlekar R, Parab S, Kakde A (2009) Prospective assessment of quality of life in adult patients with primary brain tumors in routine neurooncology practice. J Neurooncol 95(3):413–419 Bush N, Donaldson G, Moinpour C, Haberman M, Milliken D, Markle V, Lauson J (2005) Development, feasibility and compliance of a web-based system for very frequent QOL and symptom home self-assessment after hematopoietic stem cell transplantation. Qual Life Res 14(7):77–93 Cella DF, Tulsky DS, Gray G, Sarafian B, Linn E, Bonomi A, Silberman M, Yellen SB, Winicour P, Brannon J et al. (1993) The Functional Assessment of Cancer Therapy scale: development and validation of the general measure. J Clin Oncol 11(3):570–579

230 Coons S, Gwaltney C, Hays R, Lundy J, Sloan J, Revicki D, Lenderking W, Cella D, Basch EI. e. T. Force (2009) Recommendations on evidence needed to support measurement equivalence between electronic and paper-based patient-reported outcome (PRO) measures: ISPOR ePRO Good Research Practices Task Force report. Value Health 12(4):419–429 Detmar SB, Aaronson NK, Wever LD, Muller M, Schornagel JH (2000) How are you feeling? Who wants to know? Patients’ and oncologists’ preferences for discussing health-related quality-of-life issues. J Clin Oncol 18(18):3295–3301 Detmar SB, Muller MJ, Schornagel JH, Wever LDV, Aaronson NK (2002) Health-related quality-of-life assessments and patient-physician communication: a randomized controlled trial. JAMA 228(23):3027–3034 Erharter A, Giesinger J, Kemmler G, Schauer-Maurer G, Stockhammer G, Muigg A, Hutterer M, Rumpold G, Sperner-Unterweger B, Holzner B (2010) Implementation of computer-based quality-of-life monitoring in brain tumor outpatients in routine clinical practice. J Pain Symptom Manage 39(2):219–229 Ferlay J, Shin HR, Bray F, Forman D, Mathers C, Parkin DM (2010) GLOBOCAN 2008, Cancer Incidence and Mortality Worldwide: IARC CancerBase No. 10 [Internet]. Lyon, France: International Agency for Research on Cancer. Available from: http://globocan.iarc.fr Frost M, Bonomi A, Cappelleri J, Schunemann H, Moynihan T, Aaronson N (2007) Applying quality-of-life data formally and systematically into clinical practice. Mayo Clin Proc 82(10):1214–1228 Giesinger JM, Golser M, Erharter A, Kemmler G, SchauerMaurer G, Stockhammer G, Muigg A, Hutterer M, Rumpold G, Holzner B (2009) Do neurooncological patients and their significant others agree on quality of life ratings? Health Qual Life Outcomes 7:87 Hawkins R, Grunberg S (2009) Chemotherapy-induced nausea and vomiting: challenges and opportunities for improved patient outcomes. Clin J Oncol Nurs 13(1):54–64 Heimans J, Taphoorn MJ (2002) Impact of brain tumour treatment on quality of life. J Neurol 249(8):955–960 Holzner B, Schauer-Maurer G, Stockhammer G, Muigg A, Hutterer MGJ (2011) Computergestütztes Patient-reported Outcome Monitoring in der Neuroonkologie: Lebensqualität und Rezidiv beim Glioblastom.Wien Med Wochenschr 1611213–19 Janda M, Steginga S, Langbecker D, Dunn J, Walker D, Eakin E (2007) Quality of life among patients with a brain tumor and their carers. J Psychosom Res 63(6):617–623 Luckett T, Butow PN, King MT (2009) Improving patient outcomes through the routine use of patient-reported data in cancer clinics: future directions. Psychooncology 18(11):1129– 1138 Mullen KH, Berry DL, Zierler BK (2004) Computerized symptom and quality-of-life assessment for patients with cancer part II: acceptability and usability. Oncol Nurs Forum 31(5):E84–89 Nickolov A, Beumont J, Victorson D, Peterman A, Cella D, Liepa A, HA F (2005). Validation of functional assessment of cancer therapy: brain (FACT-Br) questionnaire and FACT-Br

L.M. Wintner et al. symptom index (FBrSI) in patients with recurrent high-grade glioma. Paper presented at: Chicago Supportive Oncology Conference, 2005, Chicago, IL Reardon DA, Wen PY (2006) Therapeutic advances in the treatment of glioblastoma: rationale and potential role of targeted agents. Oncologist 11(2):152–164 Rothman ML, Beltran P, Cappelleri JC, Lipscomb J, Teschendorf B (2007) Patient-reported outcomes: conceptual issues. Value Health 10(Suppl 2):S66–75 Sanchez R, Ballesteros M, Arnold BJ (2011) Validation of the FACT-G scale for evaluating quality of life in cancer patients in Colombia. Qual Life Res 20(1):19–29 Taphoorn MJ, Claassens L, Aaronson NK, Coens C, Mauer M, Osoba D, Stupp R, Mirimanoff RO, van den Bent MJ, Bottomley A (2010a) An international validation study of the EORTC brain cancer module (EORTC QLQBN20) for assessing health-related quality of life and symptoms in brain cancer patients. Eur J Cancer 46(6): 1033–1040 Taphoorn MJ, Sizoo EM, Bottomley A (2010b) Review on quality of life issues in patients with primary brain tumors. Oncologist 15(6):618–626 Trotti A, Colevas AD, Setser A, Basch E (2007) Patient-reported outcomes and the evolution of adverse event reporting in oncology. J Clin Oncol 25(32):5121–5127 Trowbridge R, Dugan W, Jay SJ, Littrell D, Casebeer LL, Edgerton S, Anderson J, O’Toole JB (1997) Determining the effectiveness of a clinical-practice intervention in improving the control of pain in outpatients with cancer. Acad Med 72(9):798–800 U.S. FOOD AND DRUG ADMINISTRATION (2006). Guidance for industry: patient-reported outcome measures: use in medical product development to support labeling claims Velikova G, Wright EP, Smith AB, Cull A, Gould A, Forman D, Perren T, Stead M, Brown J, Selby PJ (1999) Automated collection of quality-of-life data: a comparison of paper and computer touch-screen questionnaires. J Clin Oncol 17(3):998–1007 Velikova G, Booth L, Smith AB, Brown PM, Lynch P, Brown JM, Selby PJ (2004) Measuring quality of life in routine oncology practice improves communication and patient well-being: a randomized controlled trial. J Clin Oncol 22(4):714–724 Velikova G, Keding A, Harley C, Cocks K, Booth L, Smith AB, Wright P, Selby PJ, Brown JM (2010) Patients report improvements in continuity of care when quality of life assessments are used routinely in oncology practice: secondary outcomes of a randomised controlled trial. Eur J Cancer 46(13):2381–2388 Walker M, Brown J, Brown K, Gregor A, Whittle I, Grant R (2003) Practical problems with the collection and interpretation of serial quality of life assessment in patients with malignant glioma. J Neurooncol 63(2): 179–186 Webster K, Cella D, Yost K (2003) The functional assessment of chronic illness therapy (FACIT) measurement system: properties, applications, and interpretation. Health Qual Life Outcomes 1:79

Part IV

Hemangioblastoma

Chapter 26

Intra-operative ICG Use in the Management of Hemangioblastomas Loyola V. Gressot and Steven W. Hwang

Abstract Technology that assists in obtaining a safe, complete resection of central nervous system (CNS) tumors can have a tremendous impact on patient outcomes. Indocyanine green (ICG) is a fluorescent dye that has established utility in the intraoperative visualization of intracranial vasculature during aneurysm surgery. Recently, the use of ICG as an adjunct in tumor resection has been described. Intraoperative ICG use may help visualize vascular tumors, such as hemangioblastomas, to ascertain the tumor margins and to assure that no residual tumor remains. ICG can also assist in intraoperative localization of tumors and operative planning by visualizing the tumor in real time. Intraoperative ICG videography has the potential to become a powerful adjunct in the resection of specific CNS neoplasms. Keywords von Hippel-Lindau syndrome · Hemangioblastomas · Indocyanine green · ICG · Videography

Introduction Neoplasms of the central nervous system remain a challenging management dilemma. Although significant advances have been made over the last few decades, profound limitations still remain.

S.W. Hwang () Department of Neurosurgery, Tufts Medical Center, Boston, MA, USA e-mail: [email protected]

Improvements in treatment modalities have included the development of tools and techniques for safer surgical resection, advances in radiation therapy and improved chemotherapeutic agents. Overall, advances in technology and treatment modalities as well as systemic control have prolonged survival, but with select pathologies, the importance of gross total surgical resection (GTR) has become paramount in treatment. Achieving a gross total resection can potentially be curative for low grade tumors and has been shown to effect survival outcomes in high grade lesions including glioblastoma multiforme (Stummer et al., 2008), and anaplastic astrocytoma (Keles et al., 2006; McGirt et al., 2008). Gross total resection has also been shown to increase disease free progression in spinal intramedullary hemangioblastomas, ependymomas (Graces-Ambrossi et al., 2009) and pilocytic astrocytomas (Karikari et al., 2011). Therefore adjuvant tools that improve our ability to safely achieve a gross total resection are critical to the advancement of oncologic care. Hemangioblastomas are rare vascular tumors typically located in the infratentorial space or within the spine and often associated with von HippelLindau syndrome. Complete resection can be curative, although as hemangioblastomas are highly vascular tumors, surgical resection can be complicated and limited by catastrophic intra-operative bleeding (Cristante and Herrmann, 1999; Gläsker et al., 2010). Therefore technology permitting the visualization of tumor vessels in real time and verifying the complete resection of the lesion is an invaluable tool. Furthermore, most intramedullary spinal tumors occur in the cervical region (Karikari et al., 2011) therefore increasing the potential iatrogenic risk of devastating

M.A. Hayat (ed.), Tumors of the Central Nervous System, Volume 5, DOI 10.1007/978-94-007-2019-0_26, © Springer Science+Business Media B.V. 2012

233

234

neurological injury, such as respiratory depression and quadriparesis. Thus the utilization of intraoperative technology to optimize a complete resection while minimizing trauma to neural tissue would further improve the safety of surgical intervention. Therefore, development of adjuvant tools, such as use of indocyanine green video fluoroscopy, to optimize safe and complete resection of these lesions has a significant role in oncological care.

L.V. Gressot and S.W. Hwang

desired view prior to injection of the dye. The operative field must be clear of debris, cottonoids, blood product, or other obstructive material for the desired structures to be visualized. Calcification or atherosclerosis of the intracranial vasculature may obscure the field and marginalize interpretation of the imaging (Raabe et al., 2005). The anesthesiologist and surgical team should be cognizant that, as ICG absorbs light, it may cause a transient drop in pulse oximetry readings when administered intravenously.

Background Initially, ICG was designed by Kodak labs and applied to cardiac output research, but was quickly adapted to hepatologic study given its clearance attributes (Hemming et al., 1992). It has also been used extensively in ophthalmology to diagnose choroidal hemangiomas (Arvelo et al., 2000), to differentiate ocular melanoma from nevi (Kubicka-Trzaska et al., 2002), as a stain intra-operatively for cataract extraction (Yi and Sullivan, 2002), and has been proposed as an ancillary diagnostic tool for the identification of ocular tumors (Schalenbourg et al., 2000). Indocyanine green (ICG) is a water-soluble, tricarbocyanine, near infrared fluorescent dye that is not metabolized in human physiology. It is excreted into bile by the liver and does not flow through enterohepatic recirculation. The half-life is approximately 3–4 min in plasma. ICG is typically administered as an intravenous bolus after reconstitution in sterile water and the recommended dose is 0.2–0.5 mg/kg with a maximum of 5 mg/kg per day. ICG absorbs light at 805 nm and emits it at 835 nm which is in the near infrared spectrum. Most commercial operating microscopes can be fitted with a filter to provide high resolution images that can visualize the dye both in real time, permitting the evaluation of transit speed, and recorded for playback (Raabe et al., 2003). ICG administration may be repeated multiple times during a procedure after a short latency (Woitzik et al., 2005). Intraoperative ICG videography is visualized through the filter on the monitor and therefore images are only perceived in two dimensions, without depth perception. Hence, the surgeon must optimize the

Intracranial Practical Applications of ICG The use of intraoperative ICG is becoming more prevalent in vascular neurosurgery. Intraoperative ICG is used in vascular neurosurgery to evaluate clip position, to assure adequate occlusion of an aneurysm, and to visualize filling of perforating vessels as well as patency of parent vessels (Dashti et al., 2009). Though intraoperative digital subtraction angiography remains the gold standard, ICG is arguably comparable in identifying surgically relevant details and is faster while requiring less equipment and mobilization of personnel (Raabe et al., 2005). Intraoperative ICG can also be used to assess extra-cranial- intra-cranial (EC-IC) bypass patency as well as denote the site of graft stenosis or obstruction in the event of inadequate flow (Woitzik et al., 2005). Intraoperative ICG has also been used to identify the site of residual arteriovenous malformation (AVM) nidus (Takagi et al., 2007), and as an adjunct in the resection of extracranial vascular malformations including an intramuscular arteriovenous hemangioma (Nazzi et al., 2008). ICG has even been used intraoperatively to characterize unexpected findings such as an AVM encountered during a microvascular decompression for trigeminal neuralgia (Ferroli et al., 2010) and to identify a developmental venous anomaly by observing the patterns of dynamic flow (Ferroli et al., 2008). The use of ICG has also been shown to stain glioma margins in animal models which may be of potential use to demarcate normal parenchyma brain from tumor intraoperatively (Hansen et al., 1993; Britz et al., 2002). Hansen et al. (1993) using a rat glioma

26 Intra-operative ICG Use in the Management of Hemangioblastomas

model, proved that ICG staining is reliable to identify tumor margins within 1 mm of actual parenchyma. The stain persisted for up to an hour but required doses greater than the lethal level for a rat. Britz et al. (2002) showed that adjunct administration of the bradykinin analogue RMP-7 lowers the dose of ICG required to produce margin staining to non-toxic levels. More studies are needed to further delineate if this technology will be applicable or feasible in human physiology. Ferroli et al. (2011) reported their use of ICG in the resection of 100 craniotomies of mixed pathology and attempted to delineate applications of this new surgical tool. They felt that ICG use permitted identification of arterial vasculature that supplied normal parenchyma in 7 of 71 cases and accurately identified hypervascular or hypovascular areas of the tumor reliably. In 51 of 83 tumors, ICG identified high flow arterio-venous fistulas within the tumors. These intraoperative observations influenced their decisions to sacrifice selected venous outflow (Ferroli et al., 2011). Although they subjectively noted a benefit with the use of ICG, further objective evaluation documenting potential advantages of it use such as greater surgical resection, lower complications, less blood loss or even long-term clinical outcomes is still required.

Use of ICG in Resection of Intracranial Hemangioblastomas Hemangioblastomas are slow growing, benign, highly vascular tumors that typically occur in adults and are often located in the infratentorial space. These tumors are rare and often associated with Von Hippel Lindau syndrome (Gläsker et al., 2010). The incidence is low has not been clearly defined with no gender preference associated (Aldape et al., 2007). Hemangioblastomas typically present as an enhancing nodular mass with an associated cyst or syrinx and are routinely encountered infratentorially or in the spinal cord. Although the application of ICG to intracranial tumors is still being defined, its use in vascular neurosurgery has been well documented. Hence resection of highly vascular tumors may have the greatest benefit with use of ICG. However, little clinical experience has been reported with use of ICG and hemangioblastomas

235

thus far (Murai et al., 2011; Ferroli et al., 2011). Murai et al. (2011) described their use of ICG in three cases resecting a hemanglioblastoma and Ferroli et al. (2011) applied it to the three hemangioblastomas of their tumor series. Both authors noted good visualization of the tumor and associated arterio-venous structures. Murai et al. (2011) even reported that it helped identify the location of one lesion underneath the parenchyma through inference of superficial abnormal vasculature. Intraoperative ICG videography is especially helpful in identifying regions of tumor that may be obscured from traditional microscopy by scar tissue and can be a useful adjunct in the pursuit of an en bloc resection for these tumors. The pursuit of an en bloc resection of hemangioblastomas is important not only for an oncological cure, but also to limit intraoperative hemorrhage as these tumors are highly vascular. Multiple authors have emphasized that careful development of a surgical plane between the tumor and neural tissue is paramount to avoiding massive tumor bleeding intraoperatively (Cristante and Herrmann, 1999; Gläsker et al., 2010). Identification of feeding arteries and draining veins via ICG videography can facilitate safer removal of these lesions (Murai et al., 2011). Post-resection ICG can also be utilized to confirm a gross total resection. However clinical experience using this tool is still early and its limitations have not yet been defined.

Use of ICG in Resection of Spinal Hemangioblastomas Spinal hemangioblastomas are much more likely to be associated with VHL than hemangioblastomas in other locations (Takai et al., 2010). Hemangioblastomas are the third most commonly encountered intramedullary spinal tumor comprising 5–15% of reported series (Baleriaux, 1999; Karikari et al., 2011; Cristante and Herrmann, 1999; Xu et al., 1996). Primary management of these lesions is surgical resection. Gross total resection is potentially curative, whereas subtotal resection has a higher risk of recurrence and disease progression (Takai et al., 2010). Intraoperative ICG videography has also been used for localization of intradural tumors to optimize the

236

L.V. Gressot and S.W. Hwang

location of the durotomy overlying the tumor which may have shifted to some extent during patient positioning. This technology permits the confirmation of the location of the tumor while the dura is intact avoiding epidural veins. Schubert et al. (2010) demonstrated that the tumor could be identified prior to dural opening in 93% intradural spinal cord tumors. Tumors located in a circumscribed ventral location were difficult to visualize during a posterior approach. In one case, ICG imaging revealed the need to extend the laminectomy exposure which was then accomplished prior to performing the durotomy. Also, use of intraoperative ICG videography has been useful for the identification of feeding arteries and draining veins during the resection of spinal vascular malformations (Murakami et al., 2010; Hanel et al., 2010).

Intraoperative ICG videography has similarly been applied to visualize spinal hemangioblastomas (Figs. 26.1, 26.2 and 26.3). The same potential intraoperative advantages of visualizing arterio-venous anatomy, extent of tumor, confirming complete resection, and minimizing blood loss are applicable to spinal hemangioblastomas. However, only the superficial vessels and vascular lesions are visible and the field must be clear of obstructive foreign bodies and blood to optimize the interpretation of ICG videography. Various authors have reported successful imagining using ICG doses ranging from 5 to 50 mg boluses with imaging about 1–2 min post injection (Hwang et al., 2010; Murakami et al., 2010; Schubert et al., 2010).

A

B

Fig. 26.1 (a) Sagittal T1-weighted, (b) Sagittal T1-weighted post-gadolinium pre-operative MRI demonstrating a recurrent hemangioblastoma with enhancing nodule at C1 and significant spinal cord edema

A

Fig. 26.2 (a) Intraoperative images using microscopy, (b) video ICG demonstrating a hemangioblastoma with surrounding vasculature. Note the area of tumor visible on ICG but hidden by scar tissue on plain microscopic view

B

26 Intra-operative ICG Use in the Management of Hemangioblastomas

A

237

B

C

Fig. 26.3 Postresection intra-operative images demonstrating no residual tumor visible on either (a) microscopic view or (b) ICG videography or (c) post-operative MRI (sagittal post-gadolinium T1-weighted MRI)

Conclusion

References

Intraoperative ICG videography is a relatively new technology and its application in neuro-oncology is still being investigated. Intraoperative ICG videography appears to be useful for characterization and localization of vascular tumors. However, further evaluation is needed to delineate the benefits and limitations of this tool including types of tumor enhancement, depth of tumor visualization, intra-op changes with hyperemia, and clinical outcomes or patients treated in conjunction with this technology.

Aldape KD, Plate KH, Vortmeyer AO, Zagzag D, Neumann HPH (2007). Haemangioblastoma. WHO classification of tumours of the central nervous system, 4th edn. International Agency for Research on Cancer. Lyon, pp 184–186 Arevalo JF, Shields CL, Shields JA, Hykin PG, De Potter P (2000) Circumscribed choroidal hemanigoma:characteristic features with indocyanine green videoangiography. Ophthalmology 107:344–350 Baleriaux DL (1999) Spinal cord tumors. Eur Radiol 9:1252–1258 Britz GW, Ghatan S, Spence AM, Berger MS (2002) Intracarotid RMP-7 enhanced indocyanine green staining of tumors in a rat glioma model. J Neuro Oncol 56:227–232

238 Cristante L, Herrmann HD (1999) Surgical management of intramedullary hemangioblastoma of the spinal cord. Acta Neurochir (Wien) 141:333–340 Dashti R, Laakso A, Niemela M, Porras M, Hernesniemi J (2009) Mircoscope-integrated near-infrared indocyanine green videoangiography during surgery of intracranial aneurysma: the Helsinki experience. Surg Neurol 71:543–550 Ferroli P, Tringali G, Albanese E, Broggi F (2008) Developmental venous anomaly of petrous veins: intraoperative findings and indocyanine green video angiographic study. Neurosurgery 62:418–421 Ferroli P, Acerbi F, Broggi M, Broggi G (2010) Arteriovenous micromalformation of the trigeminal root: intraoperative diagnosis with indocyanine green videoangiography: case report. Neurosurgery 67:E309–310 Ferroli P, Acerbi F, Albanese E, Tringali G, Broggi M, Franzini A, Broggi G (2011) Application of intraoperative indocyanine green angiography for CNS tumors: results on the first 100 cases. Acta Neurochir Suppl 109:251–257 Gläsker S, Klingler JH, Müller K, Würtenberger C, Hader C, Zentner J, Neumann HP, Velthoven VV (2010) Essentials and pitfalls in the treatment of CNS hemangioblastomas and von Hippel-Lindau disease. Cen Eur Neurosurg 71:80–87 Graces-Ambrossi GL, McGirt MJ, Mehta VA, Sciubba DM, Witham TF, Bydon A, Wolinksy JP, Jallo GI, Gokaslan ZL (2009) Factors associated with progression-free survival and long-term neurological outcome after resection of intramedullary spinal cord tumors: analysis of 101 consecutive cases. J Neurosurg Spine 11:591–599 Hanel RA, Nakaji P, Spetzler RF (2010) Use of microscopeintegrated near-infrared indocyanine green videoangiography in the surgical treatment of spinal dural arteriovenous fistulae. Neurosurgery 66:978–985 Hansen DA, Spence AM, Carski T, Berger MS (1993) Indocyanine green (ICG) staining and demarcation of tumor margins in a rat glioma model. Surg Neurol 40:451–456 Hemming AW, Scudamore CH, Shackleton CR, Pudek M, Erb SR (1992) Indocyanine green clearance as a predictor of successful hepatic resection in cirrhotic patients. Am J Surg 163:515–518 Hwang SW, Malek AM, Schapiro R, Wu JK (2010) Intraoperative use of indocyanine green fluorescence videography for resection of a spinal cord hemangioblastoma. Neurosurgery 67:300–303 Karikari IO, Nimjee SM, Hodges TR, Cutrell E, Hughes BD, Powers CJ, Mehta AI, Hardin C, Bagley CA, Isaacs RE, Haglund MM, Friedman AH (2011) Impact of tumor histology on resectablility and neurological outcome in primary intramedullary spinal cord tumors: A single-center experience with 102 patients. Neurosurgery 86:188–197 Keles GE, Change EF, Lambron KR, Tihan T, Chang CJ, Change SM, Berger MS (2006) Volumetric extent of resection and residual contrast enhancement on initial surgery as predictors of outcome in adult patients with hemispheric anaplastic astrocytoma. J Neurosurg 105:34–40 Kubicka-Trzaska A, Starzycka M, Romanowska-Dixon B (2002) Indocyanine green angiography in the diagnosis of small choroidal tumors. Ophthalmologica 216:316–319

L.V. Gressot and S.W. Hwang McGirt MJ, Chaichana KL, Attenello FJ, Weingart JD, Than K, Burger PC, Olivi A, Brem H, Quinones-Hinojosa A (2008) Extent of surgical resection is independently associated with survival in patients with hemispheric infiltrating low-grade gliomas. Neurosurgery 63:700–707 Murai Y, Adachi K, Matano F, Tateyama K, Teramoto A (2011) Indocyanin Green Videoangiography Study of Hemangioblastomas. Can J Neurol Sci 38:41–47 Murakami T, Koyangi I, Kaneko T, Iihoshi S, Houkin K 2010. Intraoperative indocyanine green videoangiography for spinal vascular lesions. Neurosurgery [epub ahead of print] Nazzi V, Messina G, Dones I, Ferroli P, Broggi G (2008) Surgical removal of intramuscular arterovenous hemangioma of the upper left forearm compressing radial nerve branches. J Neurosurg 108:808–811 Raabe A, Beck J, Gerlach R, Zimmermann M, Seifert V (2003) Near-infrared indocyanine green video angiography: a new method for intraoperative assessment of vascular flow. Neurosurgery 52:132–139 Raabe A, Nakaji P, Beck J, Kim LJ, Hsu FP, Kamerman JD, Seifert V, Spetzler RF (2005) Prospective evaluation of surgical microscope-integrated intraoperative near-infrared indocyanine green videoangiography during aneurysm surgery. J Neurosurg 103:982–989 Schalenbourg A, Piguet B, Zografos L (2000) Indocyanine green angiographic findings in choroidal hemangiomas: a study of 75 cases. Ophthalmologica 214:246–252 Schubert GA, Barth M, Thome C (2010) The use of indocyanine green videography for intraoperative localization of intradural spinal tumors. Spine 35:E212–E217 Stummer W, Reulen HJ, Meinel T, Pichlmeier U, Schumacher W, Toon JC, Rohde V, Oppel F, Turowski B, Woiciechowsky C, Franz K, Pietsch T, ALA-Glioma Study Group. 2008. Extent of resection and survival in glioblastoma multiforme: identification of and adjustment of bias. Neurosurgery 62: 564–576 Takagi Y, Kikuta KI, Nozaki K, Sawamura K, Hashimoto N (2007) Detection of residual nidus by surgical microscopeintegrated intraoperative near-infrared indocyanine green videoangiography in a child with a cerebral arteriovenous malformation. J Neurosurg. 107:416–418 Takai K, Taniguchi M, Takahashi H, Usui M, Saito N (2010) Comparative analysis of spinal hemangioblastomas in sporadic disease and von hippel-lindau syndrome. Neurol Med Chir (Tokyo) 50:560–567 Woitzik J, Horn P, Vajkoczy P, Schmiedek P (2005) Intraoperative control of extracranial-intracranial bypass patency by near-infrared indocyanine green videoangiography. J Neurosurg. 102:692–698 Xu QW, Bao WM, Mao RL, Yang GY (1996) Aggressive surgery for intramedullary tumor of cervical spinal cord. Surg Neurol 46:322–328 Yi DH, Sullivan BR (2002) Phacoemulsification with indocyanine green versus manual expression extracapsular cataract extraction for advanced cataract. J Cataract Refract Surg 28:2165–2169

Chapter 27

Hemangioblastoma Cysts: Diagnosis Using Fluorescence with 5-Aminolevulinic Acid Satoshi Utsuki, Hidehiro Oka, and Kiyotaka Fujii

Abstract Hemangioblastoma is a benign tumor which often emerges sporadically in adult cerebellum or is associated with von Hippel-Lindau disease. Though it can be surgically removed, in rare occasions it can reappear, necessitating a second treatment to remove the tumor. Recurrences of hemangioblastoma are located either in an area removed from the initial tumor site or locally due to an incomplete excision of the tumor. The latter recurrence can be controlled by confirming that none of the tumor remains at the time of surgery. Intraoperative fluorescence diagnosis using 5-aminolevulinic acid (5-ALA) utilizes of protoporphyrin IX (PpIX) accumulation in tumor cells, resulting in PpIX accumulation in residual tumor cells which can be seen reacting under ultraviolet light. This is a simple, non-invasive technique to detect the presence of any residual tumor while still in the operating room. Although there are no tumor cells in cysts of hemangioblastoma, there are cases where a recurrence is thought originate from a residual cyst. Because rupturing these thin-walled cysts can cause damage to the surrounding healthy cerebellum, routine extirpation of hemangioblastoma cysts is not recommended unless the cysts have been invaded by tumor cells, in which case extirpation is necessary to prevent recurrence. Intraoperative fluorescence diagnosis using 5-ALA is

S. Utsuki () Department of Neurosurgery, Kitasato University School of Medicine, 1-15-1, Kitasato, Minami, Sagamihara, Kanagawa 252-0374, Japan e-mail: [email protected]

a useful way to determine the presence of tumor cells in cyst walls associated with hemangioblastoma. Keywords 5-aminolevulinic acid · Hemangioblastoma · Intraoperative fluorescence diagnosis · Protoporphyrin IX · Residual tumor

Introduction Hemangioblastoma is a benign tumor which often emerges in adult cerebellum, approximately 62% of which occur sporadically and 38% of which occur in relation to von Hippel-Lindau (VHL) disease (Singounas, 1978). Hemangioblastoma can be extirpated microsurgically, producing mainly positive surgical results. However, even in cases where complete excision of the initial tumor was successful, longterm follow-up has established a recurrence rate of 15–27% (de la Monte and Horowitz, 1989; Niemela et al., 1999) and the risk of delayed recurrence may be high. Initial diagnosis at a young age, the presence of VHL disease, and multicentricity of CNS tumors can be cited as risk factors for recurrence. The categories of recurrence are local recurrence in the area where the initial tumor existed, and new region recurrence where no tumors were detected initially. In the latter cases, an undetected tumor may already have existed at initial surgery, a predisposition to VHL-associated tumors may have facilitated a new tumor (Niemela et al., 1999; Wanebo et al., 2003), or the recurrence may be due to a condition called hemangioblastomatosis (Weil et al., 2002). Local recurrence is caused by an incomplete excision of the tumor during initial surgery. Such incomplete excisions of tumors result

M.A. Hayat (ed.), Tumors of the Central Nervous System, Volume 5, DOI 10.1007/978-94-007-2019-0_27, © Springer Science+Business Media B.V. 2012

239

240

from either minute residual tumor infiltration or cases where the cyst is invaded with tumor cells. Cyst walls with hemangioblastoma are composed of collagen fiber, astrocytic gliosis and rosenthal fibers and is usually devoid of tumor cells (Hussein, 2007). It is therefore unnecessary to extirpate cyst walls with hemangioblastoma; Conversely, extirpation should be avoided due to the risk of additional neurological damage. However, there are reports indicating the presence of tumor cells in cyst walls (Utsuki et al., 2010) as well as recurrences resulting from residual tumors in cyst walls (Bishop et al., 2008). Accordingly, the cyst wall should be extirpated if tumor cells exist in the cyst wall to prevent recurrence of the tumor. This should be assessed during surgery. The intraoperative photodynamic diagnosis (PDD) using 5-aminolevulinic acid (5-ALA) is an efficient detection method to easily and objectively ascertain the presence of residual tumors in real time.

Mechanisms of Cyst Formation in Hemangioblastoma According to MRI studies, hemangioblastoma can be divided into four different types (Lee et al., 1989; Richard et al., 1998); tumors that are not associated with cysts, tumors that are associated with intratumoral cysts, tumors that are associated with peritumoral cysts and tumors that are associated with both peritumoral and intratumoral cysts. The frequency of each type is reported to be 46, 51, 27 and 7% respectively (Jagannathan et al., 2008). The mechanisms for the formation of intratumoral and peritumoral cysts differ from each other. While intratumoral cysts are formed by intratumoral necrosis, peritumoral cysts develop as a result of a tumor interstitial process that begins with the occurrence of edema (Lohle et al., 1998; Lonser et al., 2005). Increased tumor vascular permeability associated with hydrodynamic forces promotes fluid extravasation. Once this fluid extravasation from a tumor overcomes the capacity of the surrounding tissue to reabsorb excess fluid, edema and subsequent cyst formation occur (Lonser et al., 2005). Aquaporin-1 (AQP1), an envelope protein of the plasma membrane, has been shown to be involved in the balance between extravasation and reabsorption,

S. Utsuki et al.

operating as a constitutive channel for water transport (Chen et al., 2006). Protein identification of fluid in the cyst using two-dimensional proteomic profiling demonstrate that the proteomic patterns of intratumoral and peritumoral cyst fluid are the same (Glasker et al., 2006). Since both are considerably similar to the serum, the formation of hemangioblastoma cysts is indicated to be caused by an extravascular leakage of the serum. Accordingly, the mechanisms of both intratumoral necrosis and extravascular leakage of serum relate to intratumoral cysts associated with hemangioblastoma.

Regulation of Heme Synthesis Heme Synthesis ALA is generated in vivo from glycine and succinylCoA in mammals by ALA synthase (ALAS); an enzyme in the mitochondrial membrane. Subsequently, porphobilinogen (PBG) is produced from two ALA molecules by cytoplasmic ALA dehydratase. PBG can be transformed to uroporphyrinogen III by PBG deaminase (PBGD) and uroporphyrinogen III synthase, both of which are enzymes found in the cytoplasm, but PBGD is the rate-limiting step under normal conditions. Uroporphyrinogen decarboxylase converts uroporphyrinogen III to coproporphyrinogen III and coproporphyrinogen oxidase, which exists in the intermembrane space of the mitochondria and generates protoporphyrinogen in the cell nucleus. Protoporphyrinogen oxidase converts Protoporphyrinogen IX into Protoporphyrin IX (PpIX) and heme is synthesized by introducing iron to the tetrapyrrole structure using ferrochelatase, which is found in the inner mitochondrial membrane.

Regulation of the Heme Synthesis Pathway All enzymes of the heme pathway operate irreversibly and are partially adjusted by the feedback inhibition of ALAS. ALAS activity is the lowest after only PBGD; other enzymes have far higher activities. 5-ALA is

27 Hemangioblastoma Cysts: Diagnosis Using Fluorescence with 5-Aminolevulinic Acid

241

drawn into the cells when administered, and no further reaction occurs due to PBGD, the rate-limiting step. No excessive metabolism of 5-ALA occurs and the feedback inhibition of ALAS controls intercellular 5-ALA synthesis, while exogenous 5-ALA drawn into the cells is rapidly metabolized in the tumor cells, where PBGD is not the rate-limiting step. Furthermore, the ferrochelatase activity in tumor cells is low (Itoh et al., 2000) causing excessive PpIX to accumulate.

Detection of Tumor Cells by ALA-Derived PpIX Fluorescence (Methodology) PpIX accumulates only in tumor cells when 5-ALA is administered, and not in healthy cells due to the respective difference in enzymic activities. Because PpIX produces red fluorescence when irradiated by ultraviolet light and 5-ALA alone does not produce fluorescence, tumor areas, which accumulate PpIX, can be visually identified by this red fluorescence. Patients received 1 g of orally administered 5-ALA two hours prior to the introduction of an anesthesic. Tumor masses were extirpated under a microscope and the extirpated tumor mass, the extirpated area and the remaining cysts were irradiated with 405 nm of excitation light using a semiconductor laser device (VLD-V1 version 2; M&M Co., Ltd., Tokyo, Japan). The presence of PpIX fluorescence was observed through a low-cut filter (cut, 420 nm, M & M Co., Ltd.). This method was used to perform a PDD on the cerebellum hemangioblastoma of all patients. When we classified the cysts, there were no cases with no cysts, one case associated with intratumoral cysts, five cases associated with peritumoral cysts and one case associated with both peritumoral and intratumoral cysts. PpIX fluorescence was observed on the tumor masses in all of the cases. In addition, PpIX fluorescence was also observed on the intratumoral cyst walls in all of the cases. Although the enhancing effect of the cyst was not clear in all five cases considered to be peritumoral cysts (Fig. 27.1), PpIX fluorescence was observed on the cyst walls in two out of the five cases when the PDD was performed. In both of these cases, no obvious abnormality was seen under micrographic surgery and we believed there were no residual tumors (Fig. 27.2). However, residual tumors were suspected

Fig. 27.1 Contrast-enhanced axial T1-weighted MR image showing a homogenous enhanced mass lesion with peritumoral cyst at the right cerebellum hemisphere. The cyst wall showed no enhancement

Fig. 27.2 Intraoperative photograph showing peritumoral cyst wall after resection of a nodal lesion in microscope view. A residual tumor cannot be identified

with the PDD and an additional extirpation was performed on the cyst wall. In both cases of the extirpated samples, the cyst walls contained a thin layer of tumor cells (Fig. 27.3).

242

Fig. 27.3 Photomicrograph showing histological findings of the peritumoral cyst wall where a residual tumor cannot be identified in Fig. 27.2. The tumor cells and capillary vessels formed a thin line to border the cyst wall, and gliosis surrounds the circumference (hematoxylin and eosin staining, original magnification ×100)

Discussion Hemangioblastoma is a tumor with a clear boundary distinct from the brain and does not usually have a tumor capsule. Although brain tissue such as gliosis and Rosenthal fibres adjacent to the tumor form a boundary between the tumor and the brain, tumors do, in rare cases, invade brain parenchyma (Hussein, 2007). However, although hemangioblastoma is a benign tumor, long-term follow-up observation indicates a high incidence of recurrence. The cases of local recurrence may be due to the incomplete removal of the invading tumor cells from the brain parenchyma and the cyst wall (Bishop et al., 2008). Such residual tumor cells can be identified during surgery with a PDD using 5-ALA. This is mainly used for malignant glioma surgery (Stummer et al., 2000) and an improvement of prognosis can be achieved by extirpating areas where red fluorescence can be seen. This method assesses tumor invasions more objectively and promotes a decrease in residual tumors. Other than malignant glioma, this technique is also used to assess tumor invasions and assess residual tumors with meningioma (Kajimoto et al., 2007) and benign ependymoma (Shimizu et al., 2006). Jagannathan et al. examined all cyst walls, including peritumoral cysts, following the extirpation of hemangioblastoma and reported that the recurrence of tumors can be prevented

S. Utsuki et al.

by performing an additional extirpation on tumor-like lesions (Jagannathan et al., 2008). However, a minute traces of residual tumors may be difficult to detect during surgery. The presence of tumors in the cyst wall could not be confirmed during surgery in two of our aforementioned cases, although the presence of tumor cells was confirmed histologically. PDD using 5-ALA is an effective way to objectively detect residual tumors in hemangioblastoma. In cases such as these, intratumoral cysts develop around the tumor and may imitate peritumoral cysts with the wall becoming thinner as the cyst grows bigger (Bishop et al., 2008; Utsuki et al., 2010). Moreover, because some tumor cells were floating in the cyst fluid (Lallu et al., 2008), they could engraft to genuine peritumoral cysts and multiplied. Also, although it is rare, tumor cells not associated with either the mural nodule or with the enhancing effect of the wall may appear similar to an arachnoid cyst on the image (Vatsal et al., 2002). To prevent the recurrence of a tumor, it is necessary to extirpate the cells of such tumors completely, and PDD using 5-ALA is one of the methods which enables this.

References Bishop FS, Liu JK, Chin SS, Fults DW (2008) Recurrent cerebellar hemangioblastoma with enhancing tumor in the cyst wall: case report. Neurosurgery 62:E1378–E1379 Chen Y, Tachibana O, Oda M, Xu R, Hamada J, Yamashita J, Hashimoto N, Takahashi JA (2006) Increased expression of aquaporin 1 in human hemangioblastomas and its correlation with cyst formation. J Neurooncol 80:219–225 de la Monte SM, Horowitz SA (1989) Hemangioblastomas: clinical and histopathological factors correlated with recurrence. Neurosurgery 25:695–698 Glasker S, Vortmeyer AO, Lonser RR, Lubensky IA, Okamoto H, Xia JB, Li J, Milne E, Kowalak JA, Oldfield E H, Zhuang Z (2006) Proteomic analysis of hemangioblastoma cyst fluid. Cancer Biol Ther 5:549–553 Hussein MR (2007) Central nervous system capillary haemangioblastoma: the pathologist’s viewpoint. Int J Exp Pathol 88:311–324 Itoh Y, Henta T, Ninomiya Y, Tajima S, Ishibashi A (2000) Repeated 5-aminolevulinic acid-based photodynamic therapy following electro-curettage for pigmented basal cell carcinoma. J Dermatol 27:10–15 Jagannathan J, Lonser RR, Smith R, DeVroom HL, Oldfield EH (2008) Surgical management of cerebellar hemangioblastomas in patients with von Hippel-Lindau disease. J Neurosurg 108:210–222 Kajimoto Y, Kuroiwa T, Miyatake S, Ichioka T, Miyashita M, Tanaka H, Tsuji M (2007) Use of 5-aminolevulinic acid in

27 Hemangioblastoma Cysts: Diagnosis Using Fluorescence with 5-Aminolevulinic Acid fluorescence-guided resection of meningioma with high risk of recurrence. Case report J Neurosurg 106:1070–1074 Lallu S, Naran S, Palmer D, Bethwaite P (2008) Cyst fluid cytology of cerebellar hemangioblastoma: a case report. Diagn Cytopathol 36:341–343 Lee SR, Sanches J, Mark AS, Dillon WP, Norman D, Newton TH (1989) Posterior fossa hemangioblastomas: MR imaging. Radiology 171:463–468 Lohle PN, Wurzer HA, Seelen PJ, Kingma LM, Go KG (1998) The pathogenesis of cysts accompanying intra-axial primary and metastatic tumors of the central nervous system. J Neurooncol 40:277–285 Lonser RR, Vortmeyer AO, Butman JA, Glasker S, Finn MA, Ammerman JM, Merrill MJ, Edwards NA, Zhuang Z, Oldfield EH (2005) Edema is a precursor to central nervous system peritumoral cyst formation. Ann Neurol 58:392–399 Niemela M, Lemeta S, Summanen P, Bohling T, Sainio M, Kere J, Poussa K, Sankila R, Haapasalo H, Kaariainen H, Pukkala E, Jaaskelainen J (1999) Long-term prognosis of haemangioblastoma of the CNS: impact of von Hippel–Lindau disease. Acta Neurochir (Wien) 141:1147–1156 Richard S, Campello C, Taillandier L, Parker F, Resche F (1998) Haemangioblastoma of the central nervous system in von Hippel-Lindau disease. French VHL Study Group J Intern Med 243:547–553

243

Shimizu S, Utsuki S, Sato K, Oka H, Fujii K, Mii K (2006) Photodynamic diagnosis in surgery for spinal ependymoma. Case illustration. J Neurosurg Spine 5:380 Singounas EG (1978) Haemangioblastomas of the central nervous system. Acta Neurochir (Wien) 44:107–113 Stummer W, Novotny A, Stepp H, Goetz C, Bise K, Reulen HJ (2000) Fluorescence-guided resection of glioblastoma multiforme by using 5-aminolevulinic acid-induced porphyrins: a prospective study in 52 consecutive patients. J Neurosurg 93:1003–1013 Utsuki S, Oka H, Sato K, Shimizu S, Suzuki S, Fujii K (2010) Fluorescence diagnosis of tumor cells in hemangioblastoma cysts with 5-aminolevulinic acid. J Neurosurg 112:130–132 Vatsal DK, Husain M, Husain N, Chawla S, Roy R, Gupta RK (2002) Cerebellar hemangioblastoma simulating arachnoid cyst on imaging and surgery. Neurosurg Rev 25:107–109 Wanebo JE, Lonser RR, Glenn GM, Oldfield EH (2003) The natural history of hemangioblastomas of the central nervous system in patients with von Hippel-Lindau disease. J Neurosurg 98:82–94 Weil RJ, Vortmeyer AO, Zhuang Z, Pack SD, Theodore N, Erickson RK, Oldfield EH (2002) Clinical and molecular analysis of disseminated hemangioblastomatosis of the central nervous system in patients without von Hippel-Lindau disease. J Neurosurg 96:775–787

Chapter 28

Hemangioblastoma: Stereotactic Radiosurgery Anand Veeravagu, Bowen Jiang, and Steven D. Chang

Abstract CNS hemangioblastomas are rare, vascular neoplasms that arise primarily in the posterior cranial fossa. Prognosis is generally favorable, with a recurrence rate of fewer than 25% in multiple surgical series. Although current standard of care for CNS hemangioblastomas is surgical resection, other treatment modalities including endovascular embolization and stereotactic radiosurgery (CyberKnife, LINAC, Gamma Knife) are being applied. Increasing evidence has suggested the effectiveness of stereotactic radiosurgery in managing CNS hemangioblastomas. Herewithin, we review the indications and multiinstitutional experiences in using such a treatment modality. Keywords Radiosurgery · Hemangioblastomas · Tumor

Mast

cell

·

Introduction Central nervous system (CNS) hemangioblastomas were first described in the cerebellum by Jackson in 1872. Hemangioblastomas are usually slow growing tumors which account for 1–3% of all CNS neoplasms and 7–10% of posterior fossa tumors. These highly vascular and histologically benign (WHO I) lesions consist of a small mural nodule with an

S.D. Chang () Department of Neurosurgery, Stanford University School of Medicine, Stanford, CA 94305-5484, USA e-mail: [email protected]

associated pseudocyst in 30–80% of cases, with the remaining lesions consisting of solid tumors. The tumor tissue itself has a well defined border and has a bright red color on gross morphological examination (Fig. 28.1). Histologically, hemangioblastomas are vascular lesions containing channels lined by cuboidal epithelium and are interspersed with nests of foamy stromal cells and pericytes. Mast cells are also found and may be responsible for the production of erythropoietin which can cause erythrocytosis (Fig. 28.2). Although debate still exists, the clusters of stromal cells surrounding the vascular plexus are thought to be the neoplastic component of the lesions. These lesions typically occur in the cerebellum (63%), spinal cord (32%), and brainstem (5%), though some cases of lumbosacral nerve root and supratentorial lesions have been reported as well. There is not thought to be any sex or ethnic predominance and the mean age at diagnosis is in the late third or early fourth decade of life. CNS hemangioblastomas are most commonly treated by surgical resection, which is an effective strategy capable of achieving curative results. A number of large clinical case series have shown that when appropriately applied, surgical resection is often necessary to provide symptomatic improvement. In the case of unfavorable anatomic location or post surgical recurrence, radiosurgery is often the next line of treatment. In particular, our experience and the reported literature surrounding the use of stereotactic radiosurgery (CyberKnife, LINAC, Gamma Knife) highlight favorable outcomes in certain clinical settings. The size, morphology, location, and clinical presentation of hemangioblastomas must all be considered when choosing treatment.

M.A. Hayat (ed.), Tumors of the Central Nervous System, Volume 5, DOI 10.1007/978-94-007-2019-0_28, © Springer Science+Business Media B.V. 2012

245

246

A. Veeravagu et al.

Fig. 28.1 Histologic Appearance of hemangioblastoma. (a) low power (100×; H&E): well-demarcated proliferation of vessels and stromal cells with adjacent gliosis (b) medium power

(200×; H&): capillary network and admixed stromal cells (c) higher power (400×; H&E): vacuolated stromal cells and fine capillaries

Fig. 28.2 MRI with Gadolinium enhancement at T1/T2/FLAIR weightings. (a) Axial T1-weight post contrast images showing recurrent cerebellar hemangioblastoma associated with a large, cystic cavity. (b) Coronal T1-weight post contrast images again

demonstrating focus of recurrent hemangioblastoma and associate cystic cavity. (c) Axial T1-weight post contrast images demonstrating interval resection of mural nodule and drainage of associated cerebellar cyst with decompression of posterior fossa

Conventional Radiotherapy

a high dose of radiation. One study of 24 patients noted a 10-year survival rate of 57% for patients treated with more than 36 Gy, compared with a survival rate of 27% for patients treated with less than 36 Gy. In another review of 27 patients irradiated postoperatively (19 with gross residual disease and 6 with VHL disease), local control was achieved for 33% of patients treated with less than 50 Gy, compared with 57% of those who received more than 50 Gy. Although conventionally fractionated irradiation seems to increase the probability of hemangioblastoma control, local control even with increased radiation doses is less than optimal. Furthermore, the exposure to significant volumes of normal tissue with radiotherapy remains a concern.

Radiosurgical treatment of hemangioblastomas has become increasingly popular. Current indications for radiosurgery for hemangioblastomas include: 1) Unfavorable and inaccessible region of the CNS axis, 2) Recurrence after surgical resection (particularly in VHL patients), 3) Medical co-morbidities that preclude surgery. As with other radiosurgical targets, hemangioblastomas treated with radiosurgery are typically less than 3 cm in size. Contraindications to radiosurgery for hemangioblastomas include tumors greater than 3 cm and those inducing significant neurologic symptoms due to mass effect and edema requiring urgent decompression. The rationale for radiosurgery for hemangioblastomas comes from prior use of conventionally fractionated external beam radiation for residual or unresectable hemangioblastomas. Studies suggest that control of hemangioblastomas depends upon achieving

Stereotactic Radiosurgery Because single fractions of 20–25 Gy have been estimated to achieve the biological equivalence of 50–100 Gy administered by conventional fractionated

28 Hemangioblastoma: Stereotactic Radiosurgery

247

irradiation, stereotactic radiosurgery may provide better control than that achieved with standard radiotherapy, especially for highly vascular, benign tumors. The steep dose gradient achieved with this technique minimizes damage to eloquent structures in the posterior fossa. Because hemangioblastomas are typically small, well-defined tumors that show no histological infiltration, they are ideally suited for radiosurgical treatment. In a recent study from the NIH, 20 patients with 44 lesions were treated with Gamma Knife SRS. At a mean follow-up of 8.5 years (range 3.0–17.6 years), all 20 patients remained living. Local control was reported to be 91, 83, and 61% at 2, 5, and 10 years after Gamma Knife SRS, respectively (Asthagiri et al., 2010). This group noted that 33% of SRS treated asymptomatic, small (<1.0 cm in diameter) lesions progressed over long term follow-up, suggesting that SRS is not a viable tool for the prophylactic treatment

of asymptomatic tumors and should be reserved for treatment of surgical inoperable CNS hemangioblastomas. In another study of 32 patients who underwent Gamma Knife treatment for 74 intracranial tumors following surgical resection, survival after stereotactic radiosurgery was 100, 94.4, and 68.7% at 1, 3, and 7 years, respectively (Kano et al., 2008). A summary of key stereotactic radiosurgery studies are presented in Table 28.1. Overall, Gamma Knife, LINAC, and CyberKnife prove to be effective and safe in the management of CNS hemangioblastomas. In 1998, Chang and Adler reported on 13 patients with 29 hemangioblastoma tumors treated with LINAC based radiosurgery (including the use of CyberKnife for 2 patients with cervical spine lesions) (Chang et al., 1998). The mean patient age was 40 years, mean radiation dose to the tumor periphery was 23.2 Gy, and the mean tumor volume was 1.6 cm3 . During a follow-up period of 43 months, 3% of the

Table 28.1 Radiosurgical treatment of hemangioblastomas: key studies Mean Authors Number Number marginal Mean size and year Intervention of patients of tumors dose of tumor Karabagli et al. (2010)

Gamma Knife RS

13

34

15.8 Gy

N/A

Asthagiri et al. (2010)

Gamma Knife RS

20

44

18.9 Gy

0.50 cm3

31

92

23.4 Gy

1.8 cm3

32

74

16 Gy

0.72 cm3

Moss et al. (2009)

Kano et al. (2008)

Gamma Knife RS

Outcome At a mean radiographic follow-up of 50.2 months, growth control was achieved for all tumors (partial remission in three tumors [8.8%] and no change in 31 tumors [91.2%]). No radiation-related complications were encountered. Local control rate at 2, 5, 10, and 15 years after SRS treatment was 91, 83, 61, and 51%, respectively. All patients were alive at last follow-up (mean 8.5 years). 33% of small (<1.0 cm diameter), asymptomatic tumors progressed over a long-term follow-up. 13 (16%) of the treated hemangioblastomas progressed, 18 tumors (22%) showed radiographic regression, and 51 tumors (62%) remained unchanged in size. With median follow-up of 69 months, the actuarial local control rates at 36 and 60 months were 85 and 82%, respectively. Radiosurgery improved lesion-associated symptoms in 36 of 41 tumors. The overall survival after radiosurgery was 100, 94.4, and 68.7 at 1, 3, and 7 years, respectively. Follow-up imaging studies demonstrated tumor control in 68 tumors (91.9%). The progression-free survival after SRS at 1, 3, and 5 years was 96.9, 95.0, and 89.9%, respectively.

248

A. Veeravagu et al.

Table 28.1 (continued) Number of patients

Number of tumors

Mean marginal dose

Mean size of tumor

Matsunaga Gamma et al. Knife RS (2007)

22

67

14.0 Gy

1.69 cm3

Wang et al. (2005)

Gamma Knife RS

35

93

17.2 Gy

1.3 cm3

Tago et al. (2005)

Gamma Knife RS

13

38

19.6 Gy

7.2 cm3

Park et al. (2005)

Gamma Knife RS

9

84

16.6 Gy

N/A

Jawahar et al. (2000)

Gamma Knife RS

27

29

16 Gy

3.2 cm3

Chang et al. (1998)

LINAC and 13 CyberKnife

29

23.2 Gy

1.6 cm3

Authors and year

Intervention

Outcome The control rate for tumour growth was 83.6% at a mean follow-up of 63 months. The only factor affecting tumour growth control was the presence of a cystic component at the time of gamma knife radiosurgery. No complication such as radiation-induced peritumoural oedema or radiation necrosis occurred. A prescription dose of 12–13 Gy was used for brainstem tumors. At the most recent follow up, 29 (82.9%) patients were alive and satisfactory tumor control had been achieved in 29 (82.9%). A stable or improved neurological status was obtained in 21 patients (60%). The 1-year tumor control rate was 94%; 2 years, 85%; 3 years, 82%; 4 years, 79%; and 5 years, 71%. The actuarial 5- and 10-year survival rates were both 80.8%. The tumor control rate was 97.4% at a median radiological follow-up period of 35 months. Actuarial 5- and 10-year control rates were both 96.2%. New lesions and/or those increasing in size outside the irradiated area were discovered in five patients (38.5%). 3 of the 84 lesions (3.6%) failed to be controlled after a mean follow-up period of 4.3 years (range 8.6–141 months). One patient who had undergone two GKS treatments suffered delayed radiation-induced complications, and posterior fossa decompression and ventriculoperitoneal shunt insertion were required At a mean follow-up of 4 years, 21 patients (79%) were alive. The median survival after radiosurgery was 6.5 years (actuarial 5 year survival = 75.1 ± 11.5%). The 2-year actuarial control rate was 84.5 ± 7.1% and at 5 years, 75.2 ± 8.9%. Factors affecting good outcome indicated that smaller tumor volume and higher radiosurgical dose (>18 Gy) were significant. Only one (3%) of the treated hemangioblastomas progressed at a mean follow-up period of 43 months. Five tumors (17%) disappeared, 16 (55%) regressed, and 7 (24%) remained unchanged in size. Five of nine patients (55.6%) with symptoms referable to treated hemangioblastomas experienced symptomatic improvement.

28 Hemangioblastoma: Stereotactic Radiosurgery

249

Table 28.1 (continued) Authors and year Niemela et al. (1996)

Intervention Gamma Knife RS

Number of patients

Number of tumors

Mean marginal dose

Mean size of tumor

10

11

18.2 Gy

N/A

treated hemangioblastomas progressed, 17% disappeared, 55% regressed, and 24% remained unchanged in size (97% tumor control rate). Symptomatic improvement was observed in 55% of patients. Two symptomatic and one asymptomatic case of radiation necrosis was noted. The radiation damage seemed to be associated with tumor doses that exceeded 35 Gy or with treatment of multiple tumors resulting in additive dose to intervening brain. Most recently, Moss et al. provided a long-term retrospective evaluation of radiosurgical hemangioblastoma treatment effectiveness, with a special emphasis on the relatively recent use of frameless, imageguided stereotactic radiosurgery in the treatment of spinal lesions (Moss et al., 2009). Within a span of 16 years, 92 hemangioblastomas in 31 patients, 26 of whom with VHL, were treated with radiosurgery. Frame-based linear accelerator radiosurgery was used to treat 27 tumors whereas 67 tumors were treated with CyberKnife SRS. The radiation dose to the tumor periphery averaged 23.4 Gy and the mean tumor volume was 1.8 cm3 . In a median follow-up of 69 months, only 16% of the treated hemangioblastomas progressed, whereas 22% of the tumors showed radiographic regression and 62% remained unchanged in size. The actuarial local control rates at 36 and 60 months were 85 and 82%, respectively. Overall, treatment with stereotactic radiosurgery improved lesion-associated symptoms in 36 of 41 tumors. Others institutions have also reported similar findings. In a multi-institution study of radiosurgery for hemangioblastoma (mean dose 15.5 Gy range 12–20 Gy), an 86% actuarial freedom from

Outcome The solid part of six hemangioblastomas shrank in a median of 30 months, whereas four hemangioblastomas were unchanged at a median of 14 months. Five hemangioblastomas had an adjoining cyst and three of these cysts had to be evacuated after radiosurgery. One patient with two cerebellar hemangioblastomas (margin dose 25 Gy each) developed edema at 6 months and required a shunt and prolonged corticosteroid treatment.

progression 2 years after treatment was reported, with 18 of 38 hemangioblastomas (22 patients) followed beyond 2 years (Matsunaga et al., 2007). In this series, an association was established between mean marginal dose and tumor control. 60% (three of five) of tumors treated with 12–14 Gy progressed, 13% of those treated with 14–17 Gy progressed, but none of the 15 patients treated with greater than 17 Gy progressed. Finally, in a 2010 study by Karabagli and colleagues, tumor control was achieved for all 34 tumors (in 131 patients) at a mean radiographic follow-up of 50.2 months. This study, along with others summarized in Table 28.1, illustrate the role of stereotactic radiosurgery in tumor control or regression with low complication rates.

Complications and Management As with any surgical or radiosurgical procedure, postoperative complications can occur. However, an extensive preoperative workup may reduce many of these complications. Full diagnostic imaging of the brain and spine is necessary. Pretreatment laboratory studies and blood tests can be done to determine if the patient has VHL and/or other VHL associated lesions. Abdominal CT scans or ultrasound should reveal these associated lesions. Post craniotomy complications following hemangioblastoma resection are similar to other craniotomy procedures, including infection, hematoma, edema, and neurological deficits.

250

Radiosurgical complications do not occur at the time of treatment and are typically delayed. The risk of radiation necrosis is proportional to dose and volume treated. Patients with significant pre-existing edema surrounding their lesion may not be ideal candidates for radiosurgery as this edema may worsen with treatment, resulting in worsened symptoms. Cyst fluid production from the nodular tumor may not cease immediately after radiosurgery, and there is the possibility that the cystic portion of the tumor will increase during the post radiosurgical period. If this event occurs, stereotactic drainage of the cyst fluid may be necessary to allow further time for the solid portion of the tumor to respond to radiosurgery treatment. Finally, in patients that have had multiple radiosurgical procedures, it may be difficult to determine the composite doses for the surrounding adjacent normal brain for treatments delivered years or even a decade apart.

Conclusion CNS hemangioblastomas are usually slow growing tumors which account for 1–3% of all CNS neoplasms and 7–10% of posterior fossa tumors. These tumors remain an emblematic lesion of patients with von Hippel-Lindau (VHL) syndrome. The location of the tumor dictates presenting symptoms, with hemangioblastomas located in the cerebellar hemispheres presenting with more severe symptoms than lesions located in the spinal cord or brainstem. Treatment modalities include gross surgical resection and stereotactic radiosurgery. Surgical resection is the primary therapy and has the potential to achieve a curative result. Stereotactic radiosurgery is often implicated for tumors located in inaccessible areas or tumor recurrence post surgery. Evidence from the current literature support the use of stereotactic radiosurgery as an effective management modality for CNS hemangioblastomas. SRS treatment has a low

A. Veeravagu et al.

risk for adverse radiation effects and is associated with tumor control, tumor regression, and symptomatic improvement.

References Asthagiri AR, Mehta GU, Zach L, Li X, Butman JA, Camphausen KA et al: (2010) Prospective evaluation of radiosurgery for hemangioblastomas in von Hippel-Lindau disease. Neuro Oncol 12:80–86 Chang SD, Meisel JA, Hancock SL, Martin DP, McManus M, Adler JR Jr. (1998) Treatment of hemangioblastomas in von Hippel-Lindau disease with linear accelerator-based radiosurgery. Neurosurgery 43:28–34, discussion 34–25 Jawahar A, Kondziolka D, Garces YI, Flickinger JC, Pollock BE, Lunsford LD (2000) Stereotactic radiosurgery for hemangioblastomas of the brain. Acta Neurochir (Wien) 142:641–644, discussion 644–645 Kano H, Niranjan A, Mongia S, Kondziolka D, Flickinger JC, Lunsford LD ( 2008) The role of stereotactic radiosurgery for intracranial hemangioblastomas. Neurosurgery 63:443–450, discussion 450–441 Karabagli H, Genc A, Karabagli P, Abacioglu U, Seker A, Kilic T (2010) Outcomes of gamma knife treatment for solid intracranial hemangioblastomas. J Clin Neurosci 17:706–710 Matsunaga S, Shuto T, Inomori S, Fujino H, Yamamoto I (2007) Gamma knife radiosurgery for intracranial haemangioblastomas. Acta Neurochir (Wien) 149:1007–1013, discussion 1013 Moss JM, Choi CY, Adler JR Jr., Soltys SG, Gibbs IC, Chang SD (2009) Stereotactic radiosurgical treatment of cranial and spinal hemangioblastomas. Neurosurgery 65:79–85, discussion 85 Niemela M, Lim YJ, Soderman M, Jaaskelainen J, Lindquist C (1996) Gamma knife radiosurgery in 11 hemangioblastomas. J Neurosurg 85:591–596 Park YS, Chang JH, Chang JW, Chung SS, Park YG (2005) Gamma knife surgery for multiple hemangioblastomas. J Neurosurg 102(Suppl):97–101 Tago M, Terahara A, Shin M, Maruyama K, Kurita H, Nakagawa K et al: (2005) Gamma knife surgery for hemangioblastomas. J Neurosurg 102(Suppl):171–174 Wang EM, Pan L, Wang BJ, Zhang N, Zhou LF, Dong YF et al (2005) The long-term results of gamma knife radiosurgery for hemangioblastomas of the brain. J Neurosurg 102(Suppl):225–229

Part V

Ganglioglioma

Chapter 29

Gangliogliomas: Molecular Pathogenesis and Epileptogenesis Eleonora Aronica and Pitt Niehusmann

Abstract Gangliogliomas (GG) constitute the most frequent tumor entity in young patients undergoing surgery for intractable epilepsy. Surgery provides the best chance for curing epilepsy and preventing malignant transformation. GG consist of a mixture of dysplastic neurons and neoplastic astroglial cells. Although malignant transformation of GG is rare, it has been documented in their astroglial component. The characteristic histopathological features of GG together with the coexistence with cortical dysplasia and the expression of stem cell markers (such as CD34) suggest a developmental origin for these lesions. The malformative nature of GG is also supported by the detection of molecular alterations common to other developmental glioneuronal lesions. In particular, recent studies suggest a role for the phosphatidyl-inositol 3-kinase (PI3K)-mTOR pathway in the molecular pathogenesis of glioneuronal lesions. The cellular mechanism(s) underlying the epileptogenicity of GG is still not clearly defined. Several mechanisms are possibly involved. Recent studies support the role of alterations of the balance between excitation and inhibition, and neuron-glia interactions may also play a critical role in the generation of seizures. Astroglial cells express functional receptors for a variety of neurotransmitters and may critically modulate synaptic transmission. In addition, an increasing number of observations indicate that specific inflammatory pathways are activated in GG and may contribute to the onset of seizures and epileptogenesis.

E. Aronica () Department of Neuropathology, Academic Medical Center, 1105 AZ, Amsterdam, The Netherlands e-mail: [email protected]

Keywords Gangliogliomas · MRI · Glioneuronal · Epileptogenesis

Introduction Gangliogliomas (GG) are well differentiated, slowly growing neuroepithelial tumors. They are rare tumors, representing ∼1.3% of all brain tumors (Louis et al., 2007). However, GG appear to constitute the most frequent tumor entity in young patients undergoing surgery for intractable epilepsy (Blumcke and Wiestler, 2002). This chapter provides a brief overview of the clinical and histopathological features of GG. Particular emphasis will be given to the molecular pathogenesis. This is of major importance, since molecular-genetic studies may help to better define the pathogenic relationship between GG and other epilepsy-associated focal developmental lesions (such as focal malformations of cortical development) and may help to achieve a classification which combine histological features and pathogenetic mechanisms. We will also discuss the mechanisms of epileptogenesis in GG, reviewing the available data from human studies and discussing the role of the lesion itself, with its cellular/molecular features, as well as the role of the perilesional region. Finally, the potential clinical implications of these observations will be discussed. The emerging knowledge of the molecular pathogenesis and epileptogenesis of GG is needed as basis necessary to improve the management of refractory epilepsy or developing a target-specific therapy.

M.A. Hayat (ed.), Tumors of the Central Nervous System, Volume 5, DOI 10.1007/978-94-007-2019-0_29, © Springer Science+Business Media B.V. 2012

253

254

Epidemiological and Clinical Aspects Incidence and Age/Sex Distribution The true incidence of GG is unknown, because many of these lesions have been previously not correctly classified and there are no population-based epidemiological data. However, in large series describing the neuropathological features of surgical specimens from patients with chronic intractable focal epilepsy GG appear to be the most common tumor entity (Blumcke and Wiestler, 2002; Luyken et al., 2004). The age at occurrence for this tumor ranges from 2 months to 70 years; however GG predominantly occur in young patients. Data based on large series of patients indicate a mean/median age at diagnosis from 8.5 to 25 years (Blumcke and Wiestler, 2002; Louis et al., 2007). A trend to occur in older patients has been reported, however, for the GG with anaplastic features (WHO grade III) (Blumcke and Wiestler, 2002). Several studies suggest a prevalence of male compared to female patients with GG raging from 1.1:1 to 1.9:1.

Localization GG can be located throughout the entire central nervous system (CNS), including cerebellum, spinal cord, pituitary gland, pineal gland, hypothalamus, and brainstem. However, the temporal lobe is the most common location (> 70 %) reported for GG WHO grade I (Blumcke and Wiestler, 2002; Louis et al., 2007). In GG with anaplastic features (WHO grade III) the temporal lobe appears to be less frequently affected (Blumcke and Wiestler, 2002).

Imaging The computed tomography (CT) appearance of GG consists in a low density, well circumscribed lesion.

E. Aronica and P. Niehusmann

The solid portion of the tumor can be isodense, hypodense or mixed. Calcification of the tumor can be detected on CT. Contrast enhancement is variable. On magnetic resonance imaging (MRI) GG may display a solid or cystic appearance or be cystic with a mural nodule. On T1-weighted images the signal intensity is variable, the tumor is generally hypointense (less often, iso-intense); T2-weighted sequences generally show an hyperintense circumscribed mass; variable enhancement can be observed within the solid portion of the tumor (Louis et al., 2007); (Fig. 29.1a–c).

Symptoms The clinical history of patients with GG consists of a protracted history of focal symptoms, which vary depending on the size and location of the tumor. When the tumor is located in the region of the third ventricle and hypothalamus, symptoms related to hypothalamic dysfunction can be observed. Tumors located in the neocortex or temporal lobe are frequently associated with a history of chronic seizures, varying from months to decades before diagnosis. The seizures are often represented by drug-resistant complex partial seizures, although simple partial seizures and secondary generalization of partial seizures may also occur. Tumors involving the brainstem or the spinal cord often present with a shorter duration of symptoms before diagnosis (due to the tumor mass effect).

Neurophysiological Features On preoperative electroencephalogram (EEG) examination, patients with GG show focal, multifocal or widespread epileptiform discharges. Acute intraoperative electrocorticography (ECoG) provides an unique opportunity to assess the epileptogenicity of exposed areas of the cortex during surgery and to correlate these findings with histopathology (Fig. 29.1d). Continuous 

Fig. 29.1 (continued) interval between two subsequent spikes being 1 s at the most (frequency ≥ 1 Hz); bursts of spikes, sudden occurrence of spikes for at least 1 s with a frequency of 10 Hz

or more; (4) recruiting discharges, rhythmic spike activity characterised by an increased amplitude and a decreased frequency (electrocorticographic seizure; (Ferrier et al., 2006)

29 Gangliogliomas: Molecular Pathogenesis and Epileptogenesis

Fig. 29.1 Imaging and electrocorticography (ECoG) in ganglioglioma (GG). Panels (a)–(c): coronal MRI images revealing a focal lesion in the left temporal lobe. Panel (a) T1-weighted inversion recovery image; Panel (b) T1-weighted image with contrast; Panel (c) T2-weighted image (images kindly provided by Prof Urbach, Dep. of Neuroradiology; University of Bonn Medical Center, Germany). Panel (d) acute intraoperative ECoG provides an unique opportunity to assess the epileptogenicity of exposed areas of the cortex during surgery. Panel (d) shows silicon (5×4-) grid electrodes, with intercontact distances of 10 mm placed on the exposed surface of the cortex in a 2 year

255

old patient with temporal GG. ECoG can guide the extent of the resection, which may include a part of the macroscopically normal appearing perilesional zone (insert in d). (e) representative ECoG showing bursts of spikes recorded from the electrodes overlying the temporal cortex (arrows; ECoG kindly provided by Dr. C. Ferrier and Dr. F.S.S. Leijten, Department of Clinical Neurophysiology, University Medical Center, Utrecht, The Netherlands). (f) representative epileptiform ECoG discharge patterns. (1) sporadic spikes, spikes occurring at irregular time intervals at several sites; (2) continuous spiking, occurring rhythmically theat regular time intervals for at least 10 s, the

256

spiking, bursts, and recruiting discharges can occur in patients with GG (Fig. 29.1e and f). When continuous spiking is found, GNTs appear to be associated with a high neuronal density and with dysplastic regions surrounding the tumor (Ferrier et al., 2006). The close relationship between epileptiform discharge patterns and histopathological changes is a strong argument for electrocorticographic tailoring of the resection in GG. However, further research is required to evaluate whether tailoring resection to these discharge patterns has an impact on the surgical outcome of GG.

Pathology Macroscopy Macroscopically GG appear as a glassy, grayyellowish solid lesion with often cyst/nodular configuration; focal mineralization can be encountered on cutting the surgical specimen. Only rarely can hemorrhage and necrosis be observed (in anaplastic lesions).

Intraoperative Diagnosis It is often difficult to obtain a confident diagnosis using smear preparations, even in cases in which the diagnosis is suggest by clinical and imaging features, if the distribution of glial and neuronal elements is not preserved or the tumor smears poorly. In cases where the tumor does smear, dysplastic neurons with large nuclei and prominent nucleoli surrounded by astroglial cells

E. Aronica and P. Niehusmann

can be detected. However, when the smear preparations contain only astroglial cells the appearance can be similar to a low grade astroglial tumor (pilocytic or fibrillary astrocytomas). Frozen sections may be useful to better identify the neuronal and glial components, supporting the intraoperative diagnosis of GG.

Histopathological Features GG consist of a mixture of dysplastic neurons and glial cells (Fig. 29.2a). The glial component is represented by a large spectrum of glial cells, including cells resembling the cell type of a fibrillary or pilocytic astrocytoma, with different degrees of cellularity. The neuronal component is represented by dysplastic neurons, which vary in amount and characteristically show prominent Nissl substance, lack uniform orientation and have abnormal shapes and sizes, vesicular nuclei and prominent nucleoli (Fig. 29.2a). Bi- or multinucleate neurons may also be observed (Blumcke and Wiestler, 2002; Louis et al., 2007). The large spectrum of histopatological features observed in GG represents a diagnostic challenge to neuropathologists. In some tumors the neuronal component is relatively inconspicuous and a GG could be misdiagnosed as a low grade astrocytoma. In contrast, in cases with a predominant neuronal phenotype, cortical dysplasia and gangliocytoma should be considered in the differential diagnosis. In some GG the predominant cell type may be represented by smaller neuronal cells or cells resembling oligodendrocytes (clear cell morphology), raising the differential diagnosis of oligodendroglioma or dysembryoplastic neuroepithelial tumor (DNT). A pseudopapillary architecture can also be observed in 

Fig. 29.2 Histopathological feautures of ganglioglioma (GG). (a) Hematoxylin/Eosin (HE) staining of GG showing the mixture of neuronal cells, lacking uniform orientation (arrows and insert) and glial cells. (b) NeuN staining detects the neuronal component (nuclear staining) of GG. (c) GFAP staining detects the astroglial tumor component. (d) confocal merged image showing co-expression (yellow) of two presynaptic vesicle proteins, synaptophysin (SYN; red) with synaptic vesicle protein 2A (SV2A; green) in a dysplastic neuronal cell. Note the strong perineuronal staining. Insert in (d) shows synaptophysin staining of a large binucleate cell. Panels (e) and (f): phosphorylated ribosomal S6 protein (PS6) IR in dysplastic cells (arrows in e and high magnification in f). (g) HLA-DR detects the presence of cells of

the microglia/macrophage lineage. (h) CD3 detects the presence of T-lymphocytes around a blood vessel. (i) CD34 (precursor cell marker) IR in GG; insert a in (i) shows a CD34 immunoreactive cell with intense ramification of processes; insert b in (i) shows an AMOG positive cell with a similar morphology. (j) low magnification photograph showing CD34 positive tumor aggregates infiltrated into adjacent neocortex. (k) NeuN staining showing alterations in architectural composition (cortical dislamination and layer I hypercellularity) of the neocortex adjacent the tumor, but not infiltrated by tumor cells. Scale bar: a, g–i: 80 μm; b, c: 40 μm; d, f: 15 μm e: 60 μm; d: 1.5 mm; g: 30 μm; j–h: 160 μm

29 Gangliogliomas: Molecular Pathogenesis and Epileptogenesis

Fig. 29.2 (continued)

257

258

GG and recently, a new variant with prominent papillary appearance, designed as papillary glioneuronal tumor, has been included as a separate entity (Louis et al., 2007). Additional histopathological features commonly observed in GG include the presence of eosinophilic granular bodies, Rosenthal fibers, calcifications and a prominent capillary network. The reticulin staining can be useful to visualize the fiber network in GG. In addition, perivascular lymphocytic infiltration, as well as the presence of an abundant population of activated microglial cells represent common features of this tumor (Fig. 29.2g and h; (Louis et al., 2007; Aronica et al., 2008)). Mitotic activity is rare and the majority of cases show a low expression of proliferation markers (such as Ki67 labeling) within the glial component. The large majority of GG correspond to WHO grade I. According to the WHO classification (Louis et al., 2007) GG with anaplastic glial features are considered WHO grade III. These tumors are characterized by increased cellularity and proliferation activity within the gilal component of the tumor. The detection of microvascular proliferation and necrosis support the anaplastic transformation of the tumor. The existence of GG grade II is still a matter of debate and clear criteria for grade II are not yet established (Luyken et al., 2004; Louis et al., 2007; Majores et al., 2008).

Immunohistochemistry The panel of antibodies used for the immunohistochemical characterization of GG includes both glial and neuronal markers. A well-characterized glial marker protein, such as glial fibrillary acidic protein (GFAP) can be used to demonstrate the neoplastic gial component of the tumor (Fig. 29.2c). Several neuronal markers, such as neuronal nuclear protein (NeuN), microtubule-associated protein 2 (MAP2) and synaptophysin can be used to detect the neuronal component (Fig. 29.2b). Immunocytochemical detection of synaptophysin, as well as of the synaptic vesicle protein 2A (SV2A; both presynaptic vesicle proteins) often show a strong perikaryal and occasionally cytoplasmic immunoreactivity (IR) in dysplastic neurons within GG specimens (Fig. 29.2d). Strong perineuronal synaptophysin IR can be found in non-neocortical

E. Aronica and P. Niehusmann

areas (spinal cord, thalamus, brain stam), but is rarely observed in the human neocortex (Quinn, 1998). Thus synaptophysin- and SV2A-positive neurons represent a common feature of glioneuronal tumors. However, this staining has to be interpreted with caution, particularly in cases with extra-cortical localization or without clear evidence of neuronal differentiation (Quinn, 1998). Immunocytochemical detection of the onco-fetal protein CD34 can also be helpful in the diagnosis of GG. CD34, negative in neural cells in adult brain, has been shown to be consistently expressed in 70– 80% of GG, also detecting peritumoral satellite lesions (Fig. 29.2i and j); (Blumcke and Wiestler, 2002). Immunocytochemical detection of AMOG (adhesion molecule on glia) shows a pattern of IR similar to that observed for CD34 and AMOG colocalizes with CD34 (Fig. 29.2i(b)); (Boer et al., in press). Thus both CD34 and AMOG may help to identify a population of glioneuronal precursor cells in GG. The phosphorylated ribosomal S6 protein (PS6) represents an additional marker, which can be used to detected in the neuronal component of GG, indicating activation of the Pi3K-mTOR pathway in this tumor (Fig. 29.2e and f) (Boer et al., in press).

Differential Diagnosis Considering the broad spectrum of histopathological features displayed by GG a careful differential diagnosis with other glioneuronal and glial entities is required. Differential diagnosis is particularly difficult in stereotactic biopsy where the typical architectural features of GG are not represented and only few neurons can be detected within the specimen. Establishing the nature of these neurons, differentiating dysplastic neurons of GG from entrapped neurons of an astroglial tumor, represents a common problem in the diagnosis of GG.

Gangliocytoma Gangliocytoma is characterized by the presence of clusters of large dysplastic neurons and non-neoplastic glial cells (Louis et al., 2007). The absence of neoplastic glial cells represents the most important feature in

29 Gangliogliomas: Molecular Pathogenesis and Epileptogenesis

the differential diagnosis with GG. Whether gangliocytoma represents a clearly distinct tumor entity is still matter of discussion and requires further studies. Pilocytic Astrocytoma Many GG contain a predominant glia component, displaying a pilocytic astrocytoma pattern. The differential diagnosis is particularly difficult in tumors located in the cerebellum, where a pilocytic astrocytoma is often expected. The lack of CD34 IR, the presence of strong MAP2-positive neoplastic glial cells, as well as the usually higher proliferation activity (>1%) support the diagnosis of a pilocytic astrocytoma. Astrocytoma (WHO Grade II) As discussed above, the detection of dysplastic neurons and their differentiation from entrapped neurons within pre-existing brain parenchyma infiltrated by a diffuse astrocytoma represent a major diagnostic problem which may be addressed immunohistochemically. In this differential diagnosis, the presence of MAP2-positive neoplastic glial cells, the lack of CD34 staining, as well as the higher proliferation activity compared to GG support the diagnosis of astrocytoma. Pleomorphic Xanthoastrocytoma This tumor may have a prominent reticulin stroma, eosinophilic bodies and large pleomorphic cells that can resemble the dysplastic neurons of GG. Immunocytochemical analysis may be helpful to reveal the astroglial nature of these large cells, displaying strong GFAP IR; CD34 expression has been also observed (Blumcke and Wiestler, 2002). Thus, the lack of a neuronal dysplastic component together with the MAP2- positivity of neoplastic astrocytes and the higher proliferation activity compared to GG may help to differentiate these two entities. However, tumors with features of both pleomorphic xanthoastrocytoma and GG have been reported (Evans et al., 2000). Oligodendroglioma (WHO Grade II) A differential diagnosis with grade II oligodendroglioma could be considered in cases of GG with

259

prominent clear cell elements, resembling oligodendrocytes. The absence of dysplastic neuronal cells, as well as the distinct pattern of IR observed with MAP2 in oligodendroglioma and the higher proliferation activity may be helpful in the differential diagnosis. Dysembryoplastic Neuroepithelial Tumor (DNT) The differential diagnosis with DNT may represent a problem in case of small biopsies with few typical architectural features (such as multifocal nodular appearance and “floating” neurons in DNT) or in case of a DNT with a prominent glial component.

Coexistence with Cortical Dysplasia Dysplastic disorganization of the cortex near but separate from the tumor has been reported in patients with GNTs (Blümcke et al., 2009). According to the presently used classification of cortical dysplasia (Palmini et al., 2004) the large majority of GNT cases are combined with mild malformations of cortical development (mMCD) or with focal cortical dysplasia (FCD) type I (Fig. 29.2k). Whether cortical dysplasia combined with GNTs represents a distinct type of dysplasia (different from isolated FCDs) is a matter of ongoing debate. In addition, the potential contribution of this peritumoral cortical disorganization to the mechanisms of epileptogenesis will need further research and evaluation.

Prognostic Factors and Surgical Outcome Although GG are generally benign neoplasms, tumor recurrence, malignant progression and secondary glioblastoma multiforme (GBM) have been observed in some patients. Tumor mass effects and pharmacoresistent seizures are common manifestations of GG. Early surgical resection of GG reduces longterm morbidity and mortality from seizures, making surgery the treatment of choice for most patients when compared to medical management (Morris et al., 1998; Aronica et al., 2001a). Comparability of postoperative seizure outcome between different series is

260

complicated by differences in the follow up period and definition of seizure freedom, i.e. implication of patients suffering exclusively from auras or not. In large series the reported results of postoperative seizure freedom range from 63% to 90 (Morris et al., 1998; Aronica et al., 2001a; Im et al., 2002; Giulioni et al., 2006; Ogiwara et al., 2010). Some authors reported improved prognosis regarding seizure outcome in younger patients (Aronica et al., 2001a; Majores et al., 2008), whereas others found no correlation with age at the time of surgery (Giulioni et al., 2006; Ogiwara et al., 2010). Complete surgical resection indicates an optimal prognosis, even if significant reduction of symptoms including seizure freedom can be often achieved with partial resection. Particularly anaplastic features and malignant progression were associated with older age at surgery, subtotal resection of the tumor, extratemporal location and absence of epilepsy (Aronica et al., 2001a; Luyken et al., 2004; Majores et al., 2008). Additionally, tumor localisation and resection strategies affect surgical outcome. Whereas in series with tailored surgery using intraoperative electrocorticography (ECoG) best results are reported in patients with temporal GG (Ogiwara et al., 2010), resection of tumor alone (lesionectomy) has been associated with a less satisfactory outcome in temporo-mesial GG than in neocortical temporal or extratemporal GG (Giulioni et al., 2006). Those findings support an epileptogenic effect of the peritumoral zone especially in temporo-mesial GG. Furthermore, association with cortical dysplasia has been reported to determine a less effective control of seizures after surgery (Im et al., 2002). Concerning the histopathological parameters, anaplastic changes in the glial component, such as increased cellularity and mitotic activity, the presence of microvascular proliferation and necrosis, as well as high Ki-67 and TP53 indices, indicate a malignant progression and are associated with a less favorable outcome. A more recent analysis of the neuropathological features of GG with recurrence and malignant progression indicates the presence of a gemistocytic cell component and focal tumor cell associated CD34 immunolabeling as significant predictors of an adverse clinical course (Majores et al., 2008). However, malignant progression to a GBM usually is associated with loss of GG features, including loss of CD34 immunolabeling (Majores et al., 2008). Post-operative complications correspond

E. Aronica and P. Niehusmann

with tumor localisation. Most frequently neurological deficits like quadrantanopia, transitory aphasia, as well as memory decline are reported (Clusmann et al., 2002; Ogiwara et al., 2010).

Molecular Genetics and Pathogenesis In comparison to other tumor entities, little is known about the molecular pathogenesis of GG. Screening for genomic alterations by chromosomal and array-based comparative genomic hybridization (CGH) in a large series of 61 GG showed aberrations in 66% of GG (Hoischen et al., 2008). The most frequent aberration was gain of chromosome 7 (21%). Unsupervised cluster analysis of genomic profiles detected two major subgroups (group I: complete gain of 7 and additional gains of 5, 8 or 12; group II: no major recurring imbalances, mainly losses). A comparison of CGH data from GG and diffuse astrocytomas (WHO grade II) revealed gain of chromosome 5 significantly more frequent in GG. Interestingly, by unsupervised cluster analysis, all but one diffuse astrocytoma formed subclusters within group II of GG showing no concordant pattern of genomic imbalances. This finding suggests that the GG cluster defined by combined gains of chromosomes 5, 7, 8 and 12 (group I) represents a subgroup that can be genetically distinguished from diffuse astrocytomas, although cellular elements in the latter and the glial component of GG are generally not distinguishable with respect to their cytoand histopathological characteristics. With interphase FISH experiments the authors localized the imbalances on the cellular level to a subpopulation of glial cells, while no dysplastic neuronal cells carried those aberrations. In the same study, the analysis of two primary GG and their anaplastic recurrences identified genetic aberrations commonly associated with malignant gliomas (losses of CDKN2A/B and DMBT1 or a gain/amplification of CDK4), which were found in the anaplastic tumors and were already present in the respective GG (Hoischen et al., 2008). Interestingly, GG from patients with long-standing epilepsy had a significantly lower median number of imbalances per case than tumors from patients without epilepsy. There may be an association between these results and the observation that GG-patients with long-standing epilepsy have a lower recurrence rate and a better

29 Gangliogliomas: Molecular Pathogenesis and Epileptogenesis

clinical course than patients without epilepsy (Luyken et al., 2004). In the limited number of cytogenetically analyzed GG in further studies, numerical or structural aberrations of chromosome 7 were the only alterations detected in more than two cases (for review see Louis et al., 2007). Gain of chromosome 7 was also found recurrently by chromosomal CGH analysis (Squire et al., 2001; Yin et al., 2002; Pandita et al., 2007). The focal nature and differentiated glioneuronal phenotype as well as the expression of the stem cell epitope CD34 and coexistence with cortical dysplasia point toward an origin of GG from developmentally compromised or dysplastic precursor lesions (Blumcke and Wiestler, 2002; Fauser et al., 2004). In addition, histological similarities hint at an etiological relation with other primary low-grade brain tumors such as dysembryoplastic neuroepithelial tumours (DNETs) or diffuse and pilocytic astrocytomas. However, genetic studies have argued against the contribution of mutational events in genes with pathogenic relevance in other brain tumors (eg, TP53, EGFR, PTEN) to play a major role in GG (reviewed in Becker et al., 2006). Analysis of genes involved in the reelin pathway with a major role in neuronal development and cellular migration showed no evidence for mutations (e.g. CDK5, DCX, CDK5R1, DAB1) (Becker et al., 2002; Kam et al., 2004) On the other hand, GG did show lower mRNA expression of these genes compared with normal central nervous system tissue controls. Array analysis of epilepsy-associated GG revealed reduced expression, of the LIM-domain-binding 2 (LDB2) gene (Fassunke et al., 2008). Since LDB2 is known to play a critical role in brain development during embryogenesis, its reduced expression may represent an additional molecular mechanism contributing to the development of an aberrant neuronal network in GG (Fassunke et al., 2008). Further molecular genetic studies of GG suggest an activation of the PI3K-mTOR pathway, which is involved in cell size- and growth-control, cortical development and neuronal migration. Ligand binding to insulin receptors or growth factor receptors activates the cascade components phosphatidylinositol-3 kinase (Pi3K), Akt, TSC1 (hamartin), TSC2 (tuberin), mTOR (mammalian target of rapamycin), and the transcription factors p70S6kinase (S6K) and ribosomal S6 protein (S6) (Fig. 29.3a).

261

Hamartin and tuberin constitute a tumor suppressor complex with a central role in the PI3K-mTOR pathway (Potter et al., 2001). Several studies examined aberrant patterns of allelic variants in the TSC1 and TSC2 genes in GG and FCD IIa/IIb (reviewed in Becker et al., 2006). In contrast to FCD IIb, that showed an accumulation of partially coding polymorphisms in TSC1, FCD IIa and GG exhibit increased incidences of polymorphisms in TSC2. Polymorphisms in intron 4 and exon 41 of TSC2 were substantially increased in GG compared with controls. A somatic mutation in TSC2 intron 32 was identified in the glial cellular elements but not in the neurons of a single GG, compatible with a clonal expansion of the glial component within this tumor. The differential patterns of allelic variants in TSC1 and TSC2 argue against FCD IIb as a dysplastic precursor lesion for GG. The ERM proteins (ezrin, radixin and moesin) interact with hamartin and regulate cell adhesion and migration. Whereas sequence analysis of ezrin and radixin showed only occasional polymorphisms in epilepsy-associated tumors and FCDs, immunohistochemical data revealed aberrant labeling of ERM proteins in a high percentage of dysplastic elements in different glioneuronal lesions including GG (reviewed in Becker et al., 2006). Recent findings of further alterations in the Pi3K-mTOR cascade strongly support a pathogenetic relevance of this pathway for GNT (Boer et al., in press); (Fig. 29.3a). In addition, increased expression of components of this signaling pathway was not detected in patients with DNT, suggesting a different pathogenetic origin for this tumor. An intriguing question for the future is to further elucidate the mechanisms that regulate the Pi3K-mTOR signaling to explain the observed differences in activation of this pathway in the different developmental lesions.

Mechanisms of Epileptogenesis The association between epilepsy and brain tumors has been observed for over one century. In 1882 Hughlings Jackson made the important observation that often epilepsy represents the initial and only clinical manifestation of glial tumors; moreover he was the first to recognize the relationship between

262

E. Aronica and P. Niehusmann

Fig. 29.3 Pathogenesis and epileptogenesis. Panel (a): Schematic representation of the Pi3K-mTOR signaling pathway. Ligand binding to insulin receptors or growth factor receptors trigger phosphatidylinositol-3 kinase (Pi3K), which in turn activates the phosphoinositide-dependent protein kinase 1 (PDK1) by phosphorylation. Akt is phosphorylated and activated by phosphorylated (p)-PDK1 or by the adhesion molecule on glia (AMOG) independently of PDK1 and Pi3K. P-Akt inactivates the tumor suppressor protein tuberin (TSC2) by phosphorylation which results in the indirect activation of the mammalian target of rapamycin (mTOR). Downstream phosphorylation of the eukaryotic initiation factor 4E binding protein

1 (4E-BP1) releases the eukaryotic initiation factor 4E (eIF4E). EIF4E interacts with phosphorylated eIF4G to activate capdependent mRNA translation which enhances cell size and cell proliferation. Cell size and proliferation are also regulated by phosphorylation of the ribosomal protein S6 kinase (p70S6K) and its downstream effector ribosomal protein S6. The ERM proteins (Ezrin, radixin and moesin) interact with hamartin (TSC1) and regulate cell adhesion and migration. Components of the pathway studied in ganglioglioma are shaded in blue. Panel (b): alterations at cellular/molecular/circuit level in the lesion, in the perilesional region and in the global network potentially contributing to epileptogenesis in GG

tumor epileptogenicity and involvement of cortical gray matter in patients with brain tumors (Jackson, 1882). The advent of the neurochirurgical treatment of epilepsy associated brain lesions confirmed these initial observations and several clinical studies emphasize that pharmacologically intractable epilepsy critically affects the daily life of patients with brain tumors, even if the tumor is under control (for review see van Breemen et al., 2007; Shamji et al., 2009). The incidence of brain tumors in patients with epilepsy is about 4% and the frequency of epilepsy in patients with brain tumors is 30% or more depending on the type of the tumor (van Breemen et al., 2007). In principle, any tumor (extra-axial, intra-axial, benign or malignant, common or uncommon) can cause seizures. However, patients with supratentorial low-grade tumors are more likely to develop epilepsy (van Breemen et al., 2007). In particular GG represents the most frequent tumor in patients with focal pharmacologically intractable epilepsy (Blumcke and Wiestler, 2002; Luyken et al., 2004).

The cellular mechanism(s) underlying the epileptogenicity of brain tumors are still not clearly understood. A number of hypotheses have been put forward during the past decades that could explain increased excitability in patients with brain tumors (van Breemen et al., 2007; Shamji et al., 2009). It is likely that multiple mechanisms are involved, including both tumor related factors (tumor type, tumor location), as well as genetic and peritumoral and changes (van Breemen et al., 2007; Shamji et al., 2009). Figure 29.3b shows the mechanisms potentially contributing to epileptogenesis in GG.

Tumor Related Factors The usual localization of GG in the temporal lobe, often with cortical involvement, may only partially explain the raised likelihood of epileptic activity in this entity. The strong association of GG with chronic

29 Gangliogliomas: Molecular Pathogenesis and Epileptogenesis

seizures could also reflect the peculiar cellular composition and neurochemical profile of these tumors. Several studies have demonstrated the intrinsic epileptogenicity of GG, indicating the presence of a hyperexcitable neuronal component (for review see Blumcke and Wiestler, 2002). Analysis of electrocorticographic patterns in GG reveals that a relatively high neuronal density in the lesion is associated with highly epileptiform discharge patterns, such as continuous spiking or recruiting discharges (Ferrier et al., 2006). The immunocytochemical profile of the dysplastic neuronal component of GG indicates high expression of specific glutamate receptors (GluR) subtypes including both ionotropic and metabotropic glutamate receptors (mGluR). These findings support a central role for glutamatergic transmission in the mechanisms underlying the intrinsic, high epileptogenicity of GG (Aronica et al., 2001b). More recent studies analyzing the gene expression profile of GG support the existence of developmental alterations of the balance between excitation and inhibition within the lesion (Samadani et al., 2007; Aronica et al., 2008; Fassunke et al., 2008). Single cell analysis demonstrates a prominent expression of the mGluR5 in the neuronal component of GG (Samadani et al., 2007). In contrast, a downregulation of several GABAa receptor (GABAA R) subunits (including α1, α5, ß1, ß3 and δ subunit) was detected in GG, suggesting an impairment of GABAergic inhibition (Samadani et al., 2007; Aronica et al., 2008). Increase of the sodium-potassium chloride co-transporter (NKCC1) expression and a reduced expression of the potassium-chloride co-transporter KCC2 has been also reported (Aronica et al., 2008). The expression patterns of NKCC1 and KCC2 resemble the expression patterns observed in immature brain and support the hypothesis of a failure of developmental maturation in the pathogenesis of GG. Moreover, the reported deregulation of these transporters could actively contribute to the epileptogenicity of GG, via modulation of GABA receptors (Yamada et al., 2004). Both gene expression and immunocytochemical studies provide evidence for a prominent activation of the inflammatory response in GG. In particular, upregulation of inflammatory interleukins (such IL1β), activation of the complement cascade and the Toll-like receptor pathway are particularly consistent in GG (Aronica et al., 2008). Interestingly, several observations strongly support the proconvulsant and ictogenic properties of inflammatory molecules and

263

suggest a pathogenic role for inflammation in epilepsy (reviewed in Vezzani and Granata, 2005). The prominent increased expression of the immune response with activation of the complement cascade and the production of pro-inflammatory cytokines observed in GG may induce alterations of the blood brain barrier (BBB), which could play an additional role in the tumor epileptogenicity (Shamji et al., 2009). Decreased expression of glial glutamate transporters, in GG in combination with an enhanced amount of astrocytes, may also contribute to tumor epileptogenesis, increasing the extracellular glutamate concentration (Samadani et al., 2007). An aditional potential mechanism underlying tumorassociated epilepsy is represented by an altered potassium homeostasis. The gene expression profile of GG with low expression of several potassium channel genes suggests a disturbed ion homeostasis and transport that could lead to increased excitability (Aronica et al., 2008). Alterations in Pi3K-mTOR pathway components in GG have been recently reported (Samadani et al., 2007; Schick et al., 2007; Boer et al., in press). Interestingly, recent studies have demonstrated the critical role of this pathway in epileptogenesis and the epileptogenic potential of mTOR inhibitors (Zeng et al., 2008). However, additional experimental work is needed to clarify the molecular mechanisms by which mTOR pathway activation may influence the epileptogenesis in GG. Alterations in antiepileptic drug targets (such as ion channels and neurotransmitter receptors) and overexpression of multidrug transporters, such as P-glycoprotein (P-gp;(Aronica et al., 2003)) reported in GG, may likely underlie the drug refractoriness observed in patients with GG.

Peritumoral Changes The peritumoral region may also be relevant for the generation and propagation of seizure activity (van Breemen et al., 2007). The epileptogenicity of the peritumoral zone is supported by both functional and immunocytochemical studies, showing network alterations and revealing cytoarchitectural and neurochemical changes in the cortex resected from patients with intractable epilepsy associated with different types of focal brain lesions, including glial tumors (van Breemen et al., 2007; Shamji et al., 2009).

264

Abrupt tissue damage with isolation of cortical area (deafferentation) and hemosiderin deposition have been suggested to be implicated as cause of seizure activity in rapidly progressive high grade tumors. Experimental studies have shown that glioma invasion may alter the discharge properties of neighbouring neurons, converting them to bursting cells and hence providing “pacemaker” cells driving the neuronal networks surrounding the tumour (Kohling et al., 2006). Hypoxia and acidosis, ionic changes, enhanced intercellular communication through increased expression of gap junction channels and increased level in the peritumoral region have also been suggested as potential mechanisms affecting epileptogenesis in gliomas (van Breemen et al., 2007; Shamji et al., 2009). Perilesional changes, involving both excitatory and inhibitory pathways have been reported in GG (Aronica et al., 2007; for review see Blumcke, 2009). The reported perilesional changes suggest, in particular, a complex alteration of the GABAergic system in patients with GG (Aronica et al., 2007). Finally, the potential contribution of peritumoral cortical disorganization (coexistence with cortical dysplasia; as discussed above) has to be considered in the evaluation of the epileptogenicity of GG. Acknowledgments We thank Dr. Albert Becker (Dept. of Neuropathology, University of Bonn Medical Center) for his input in the final version of this chapter.

References Aronica E, Gorter JA, Jansen GH, van Veelen CW, van Rijen PC, Leenstra S, Ramkema M, Scheffer GL, Scheper RJ, Troost D (2003) Expression and cellular distribution of multidrug transporter proteins in two major causes of medically intractable epilepsy: focal cortical dysplasia and glioneuronal tumors. Neuroscience 118:417–429 Aronica E, Boer K, Becker A, Redeker S, Spliet WG, van Rijen PC, Wittink F, Breit T, Wadman WJ, Lopes da Silva FH, Troost D, Gorter JA (2008) Gene expression profile analysis of epilepsy-associated gangliogliomas. Neuroscience 151:272–292 Aronica E, Leenstra S, van Veelen CW, van Rijen PC, Hulsebos TJ, Tersmette AC, Yankaya B, Troost D (2001a) Glioneuronal tumors and medically intractable epilepsy: a clinical study with long-term follow-up of seizure outcome after surgery. Epilepsy Res 43:179–191 Aronica E, Yankaya B, Jansen GH, Leenstra S, van Veelen CW, Gorter JA, Troost D (2001b) Ionotropic and metabotropic glutamate receptor protein expression in glioneuronal tumors

E. Aronica and P. Niehusmann from patients with intractable epilepsy. Neuropathol Appl Neurobiol 27:1–16 Aronica E, Redeker S, Boer K, Spliet WG, van Rijen PC, Gorter JA, Troost D (2007) Inhibitory networks in epilepsyassociated gangliogliomas and in the perilesional epileptic cortex. Epilepsy Res 74:33–44 Becker AJ, Klein H, Baden T, Aigner L, Normann S, Elger CE, Schramm J, Wiestler OD, Blumcke I (2002) Mutational and expression analysis of the reelin pathway components CDK5 and doublecortin in gangliogliomas. Acta Neuropathol 104:403–408 Becker AJ, Blumcke I, Urbach H, Hans V, Majores M (2006) Molecular neuropathology of epilepsy-associated glioneuronal malformations. J Neuropathol Exp Neurol 65:99–108 Blumcke I (2009) Neuropathology of focal epilepsies: a critical review. Epilepsy Behav 15:34–39 Blumcke I, Wiestler OD (2002) Gangliogliomas: an intriguing tumor entity associated with focal epilepsies. J Neuropathol Exp Neurol 61:575–584 Blümcke I, Vinters HV, Armstrong D, Aronica E, Thom M, Spreafico R (2009) Malformations of Cortical Development and Epilepsies: Neuropathological findings with emphasis on Focal Cortical Dysplasia. Epileptic Disord 11:181–193 Boer K, Troost D, Timmerman W, Spliet WGM, van Rijen PC, Aronica E (2009) Pi3K-mTOR signaling and AMOG expression in epilepsy-associated glioneuronal tumors. Brain Pathol Apr 7. [Epub ahead of print] Clusmann H, Schramm J, Kral T, Helmstaedter C, Ostertun B, Fimmers R, Haun D, Elger CE (2002) Prognostic factors and outcome after different types of resection for temporal lobe epilepsy. J Neurosurg 97:1131–1141 Evans AJ, Fayaz I, Cusimano MD, Laperriere N, Bilbao. JM (2000) Combined pleomorphic xanthoastrocytomaganglioglioma of the cerebellum. Arch Pathol Lab Med 124:1707–1709 Fassunke J, Majores M, Tresch A, Niehusmann P, Grote A, Schoch S, Becker AJ (2008) Array analysis of epilepsyassociated gangliogliomas reveals expression patterns related to aberrant development of neuronal precursors. Brain 131:3034–3050 Fauser S, Becker A, Schulze-Bonhage A, Hildebrandt M, Tuxhorn I, Pannek HW, Lahl R, Schramm J, Blumcke I (2004) CD34-immunoreactive balloon cells in cortical malformations. Acta Neuropathol 108:272–278 Ferrier CH, Aronica E, Leijten FS, Spliet WGM, van Huffelen AC, van Rijen PC, Binnie CD (2006) Electrocorticographic discharge patterns in glioneuronal tumors and focal cortical dysplasia. Epilepsia 47:1477–1486 Giulioni M, Gardella E, Rubboli G, Roncaroli F, Zucchelli M, Bernardi B, Tassinari CA, Calbucci F (2006) Lesionectomy in epileptogenic gangliogliomas: seizure outcome and surgical results. J Clin Neurosci 13:529–535 Hoischen A, Ehrler M, Fassunke J, Simon M, Baudis M, Landwehr C, Radlwimmer B, Lichter P, Schramm J, Becker AJ, Weber RG (2008) Comprehensive characterization of genomic aberrations in gangliogliomas by CGH, array-based CGH and interphase FISH. Brain Pathol 18:326–337 Im SH, Chung CK, Cho BK, Wang KC, Yu IK, Song IC, Cheon GJ, Lee DS, Kim NR, Chi JG (2002) Intracranial ganglioglioma: preoperative characteristics and oncologic outcome after surgery. J Neuro Oncol 59:173–183

29 Gangliogliomas: Molecular Pathogenesis and Epileptogenesis Jackson JH (1882) Localized convulsions from tumor of the brain. Brain 5:364–374 Kam R, Chen J, Blumcke I, Normann S, Fassunke J, Elger CE, Schramm J, Wiestler OD, Becker AJ (2004) The reelin pathway components disabled-1 and p35 in gangliogliomas– a mutation and expression analysis. Neuropatho Appl Neurobiol 30:225–232 Kohling R, Senner V, Paulus W, Speckmann EJ (2006) Epileptiform activity preferentially arises outside tumor invasion zone in glioma xenotransplants. Neurobiol Dis 22:64–75 Louis, DN, Ohgaki, H, Wiestler, OD and Cavanee, WK (eds) (2007) WHO classification of tumours of the central nervous system. IARC, Lyon Luyken C, Blumcke I, Fimmers R, Urbach H, Wiestler ODand, Schramm J (2004) Supratentorial gangliogliomas: histopathologic grading and tumor recurrence in 184 patients with a median follow-up of 8 years. Cancer 101:146–155 Majores M, von Lehe M, Fassunke J, Schramm J, Becker AJ, Simon M (2008) Tumor recurrence and malignant progression of gangliogliomas. Cancer 113:3355–3363 Morris HH, Matkovic Z, Estes ML, Prayson RA, Comair YG, Turnbull J, Najm I, Kotagal Pand, Wyllie E (1998) Ganglioglioma and intractable epilepsy: clinical and neurophysiologic features and predictors of outcome after surgery. Epilepsia 39:307–313 Ogiwara H, Nordli DR, DiPatri AJ, Alden TD, Bowman RM, Tomita T (2010) Pediatric epileptogenic gangliogliomas: seizure outcome and surgical results. J Neurosur Pediatr 5:271–276 Palmini A, Najm I, Avanzini G, Babb T, Guerrini R, FoldvarySchaefer N, Jackson G, Luders HO, Prayson R, Spreafico R, Vinters HV (2004) Terminology and classification of the cortical dysplasias. Neurology 62:S2–8 Pandita A, Balasubramaniam A, Perrin R, Shannon P, Guha A (2007) Malignant and benign ganglioglioma: a pathological and molecular study. Neuro Oncol 9:124–134

265

Potter CJ, Huang Hand, Xu T (2001) Drosophila Tsc1 functions with Tsc2 to antagonize insulin signaling in regulating cell growth, cell proliferation, and organ size. Cell 105:357–368 Quinn B (1998) Synaptophysin staining in normal brain: importance for diagnosis of ganglioglioma. Am J Surg Pathol 22:550–556 Samadani U, Judkins AR, Akpalu A, Aronica E, Crino PB (2007) Differential cellular gene expression in ganglioglioma. Epilepsia 48:646–653 Schick V, Majores M, Koch A, Elger CE, Schramm J, Urbach H, Becker AJ (2007) Alterations of phosphatidylinositol 3kinase pathway components in epilepsy-associated glioneuronal lesions. Epilepsia 48(Suppl 5):65–73 Shamji MF, Fric-Shamji ECand, Benoit BG (2009) Brain tumors and epilepsy: pathophysiology of peritumoral changes. Neurosurg Rev 32:275–284 Squire JA, Arab S, Marrano P, Bayani J, Karaskova J, Taylor M, Becker L, Rutka J, Zielenska M (2001) Molecular cytogenetic analysis of glial tumors using spectral karyotyping and comparative genomic hybridization. Mol Diagn 6:93–108 van Breemen MS, Wilms EB, Vecht CJ (2007) Epilepsy in patients with brain tumours: epidemiology, mechanisms, and management. Lancet Neurol 6:421–430 Vezzani A, Granata T (2005) Brain inflammation in epilepsy: experimental and clinical evidence. Epilepsia 46:1724–1743 Yamada J, Okabe A, Toyoda H, Kilb W, Luhmann HJ, Fukuda A (2004) Cl- uptake promoting depolarizing GABA actions in immature rat neocortical neurones is mediated by NKCC1. J Physiol 557:829–841 Yin XL, Hui AB, Pang JC, Poon WS, Ng HK (2002) Genomewide survey for chromosomal imbalances in ganglioglioma using comparative genomic hybridization. Cancer Genet Cytogenet 134:71–76 Zeng LH, Xu L, Gutmann DH, Wong M (2008) Rapamycin prevents epilepsy in a mouse model of tuberous sclerosis complex. Ann Neurol 63:444–453

Chapter 30

Epilepsy-Associated Gangliogliomas: Identification of Genes with Altered Expression Albert J. Becker

Abstract Gangliogliomas are the most frequent neoplasms in patients with temporal lobe epilepsy (TLE). The characteristic histopathological composition of glial and neuronal elements, the focal nature and their differentiated phenotype and benign biological behavior suggest an origin from a developmentally compromised or dysplastic precursor lesion e.g. by neoplastic transformation of the glial component. The complex cellular admixture of these glioneuronal tumors represents a particular challenge for functional genomics. Considering the large number of expressed genes already in normal neuronal and glial cell types, unraveling of comprehensive, transcriptional patterns in glioneuronal neoplasms as well as cell-specific expression profiles are critical to further understand the molecular pathology of gangliogliomas. New developments in microarray and gene expression profiling strategies allow transcriptional studies for (a) many mRNAs in parallel or even the entire number of human genes as well as (b) starting from small amounts of tissue or individual cells. In concert with mRNA amplification strategies from single cells as well as laser-microdissection approaches and quantitative reverse transcription-polymerase chain reaction (RT-PCR) technologies, comprehensive expression analyses can be carried out in bioptic tumor tissue or even cellular subpopulations. The integrated interpretation of gene expression data with comprehensive data from genomic as well as promoter analyses represents an intriguing perspective to improve the

A.J. Becker () Department of Neuropathology, University of Bonn Medical Center, D-53105, Bonn, Germany e-mail: [email protected]

understanding and develop new treatment regimens for gangliogliomas. Keywords Gangliogliomas · Temporal lobe epilepsy · Microdissection · Oligonucleotide · Reelin

Introduction A major clinical feature of glioneuronal tumors is given by their manifestation with partial seizures. Within a number of different entities, gangliogliomas represent the most frequent tumors in surgical specimens from epileptic patients. They represent approximately 5% of brain tumors in childhood, but are rare in adults (Blümcke et al., 1999; Luyken et al., 2004). Gangliogliomas WHO grade I are most frequently present within the temporal lobe (>70%) (Blümcke and Wiestler, 2002). In gangliogliomas with anaplastic features (WHO grade III) the temporal lobe appears to be less frequently affected (Blümcke and Wiestler, 2002). The dual composition of dysmorphic neuronal cell combined with glial cells represents a histopathological hallmark of gangliogliomas (Fig. 30.1). The neoplastic nature of the tumor is reflected by the proliferative but rarely mitotic activity of the glial cell component, whereas the neuronal tumor element is generally considered to be non-neoplastic. Nuclear labeling for the proliferating cell nuclear antigen Ki-67 can observed in astrocytic tumor elements (Wolf et al., 1994). The focal nature of gangliogliomas, the differentiated glioneuronal phenotype and the benign clinical character have substantiated the hypothesis that these tumors derive from developmentally compromized or dysplastic precursor lesions

M.A. Hayat (ed.), Tumors of the Central Nervous System, Volume 5, DOI 10.1007/978-94-007-2019-0_30, © Springer Science+Business Media B.V. 2012

267

268

A.J. Becker

Fig. 30.1 Neuropathological hallmarks of gangliogliomas. (a) Gangliogliomas show a composition of glial and dysmorphic neuronal elements. The arrow points to a large, irregularly oriented dysmorphic neuron with aberrantly clustered Nissl-substance (hematoxylin-eosin (HE), scale bar: 50 μm). (b) Immunohistochemically, neoplastic astroglial cells are marked by strong expression of glial fibrillary acid protein

(GFAP) in delicate processes, that form a fibrillary matrix. (c) Immunohistochemistry with an antibody directed against synaptophysin underlines the irregular orientation of neurons interspersed within the glial matrix of the tumor. Perisomatic accumulation of synaptophysin is a typical finding in gangliogliomas (arrow)

(Blümcke et al., 1999). Correspondingly, the stem cell epitope CD34 has been observed abundantly expressed in gangliogliomas (Blümcke et al., 1999). These characteristics of gangliogliomas, i.e. an admixture of strikingly heterogeneous cellular populations suggest a complex etiology and pathogenesis. In order to learn more about the pathological basis of these intriguing neoplasm, refined gene expression analysis tools provide excellent technical options, which are complementary applied.

microarrays as well as to resolute gene expression on a cellular level (Fassunke et al., 2008, 2004). An important prerequisite for expression profiling experiments is given by the use of suitable controls or matched pairs of samples such as histopathologically or immunohistochemically characterized cellular elements, identified cell populations, brain regions or groups of neuropathologically characterized individuals (Crino et al., 2001; Fassunke et al., 2004). With respect to the composition of gangliogliomas the aspect of appropriate control tissue is particularly important. One approach to overcome this problem is to use ‘control’ brain tissue adjacent to ganglioglioma portions matched for grey and white matter composition of identical patients, which excludes differential expression due to differences in genetic background of different tumor and control individuals (Fassunke et al., 2008). With respect to expression array technologies largescale oligonucleotide (Lipshutz et al., 1999) and glass microscope slide DNA arrays (Brown and Botstein,

Expression Analysis: Methodology The development of expression microarray technology, i.e. cDNA and oligonucleotide arrays, allows highly parallel analysis of human tissue gene expression profiles (Bowtell, 1999; Brown and Botstein, 1999; Lipshutz et al., 1999). In addition to genome-wide studies, particularly real-time RT-PCR is a powerful approach to validate expression profiles obtained with

30 Epilepsy-Associated Gangliogliomas: Identification of Genes with Altered Expression

269

Fig. 30.2 Gene expression analysis – technical approaches. (a) Oligonucleotide arrays (e.g. Affymetrix system) represent a standard approach for highly parallel gene expression analyses based on non-competitive hybridization. Initial double strand cDNA synthesis is followed by an RNA-polymerase step, i.e. moderate amplification of templates biotin-labeling resulting in cRNA, which is hybridized. Each template transcript is represented by several short oligonucleotide sequences on the array. The specificity of hybridization for perfect sequences (PM) is monitored by a mismatch (MM) sequence with minute alteration in the DNA-sequence composition. This procedure allows large-scale parallel expression analyses with high specificity. (b) In order to learn about specific gene expression patterns in individual, highly defined cellular elements, complementary

approaches are useful such as in situ-RT and immunolaser microdissection. This procedure was used to unravel the cellular nature of CD34-expressing ganglioglioma components. As initial step we applied in situ-reverse transcription on fresh frozen sections from gangliogliomas. After in situ-RT and an immunohistochemical reaction with an antibody directed against CD34 (arrow marks a CD34 expressing cellular element, scale bar: 200 μm) laser microdissection of an individual cell was carried out. PCR analysis with lineage specific primers reveals represented the final step of this experiment (lane 1: GFAP, lane 2: NFM, lane 3: MBP, lane 4: HLA-DQ, lane 5: CD34). The result suggests a neuronal nature of CD-34 expressing cellular elements of gangliogliomas

1999) hybridized with fluorescent probes and analyzed by specific detection systems are widely applied (Fig. 30.2). Large-scale expression arrays have proven to constitute useful tools for monitoring gene expression and open new avenues for the identification of pathogenesis-related molecules and mechanisms of central nervous tissue disorders including low-grade brain tumors (Aronica et al., 2008; Fassunke et al., 2008). A certain limitation of microarray systems with respect to complex tissue samples reflected by summative expression profiles is the need of sufficient amounts of region/cell type specific mRNA (Becker et al., 2003; Duggan et al., 1999). The analysis of cellular pathology particularly in complex glioneuronal lesions and tumors represents a critical issue since mRNA isolates from gangliogliomas contain combinations of different cell types (Becker et al., 2007). Therefore, expression differences may be

contaminated due to heterogeneous cellular composition within the sample of mRNA origin for expression analysis rather than to prove true disease associated alterations in gene expression. In order to analyze specific cellular populations, separation may be needed. As a particular powerful technique used in glioneuronal lesions, aRNA amplification has been applied to generate sufficient amounts of aRNA for array hybridization starting from individual cells (Crino et al., 2001). A particular advantage of this method is the option to determine differential expression of large numbers of genes starting from minute amounts of mRNA isolated of even individual cells. (Laser)-microdissection of individual cells or cellular groups represents a highly useful technology to be combined with aRNA or RT-PCR protocols to determine changes of gene expression in ganglioglioma pathogenesis (Becker et al., 2006).

270

Differential Gene Expression in Gangliogliomas In addition to their biphasic composition by dysplastic neuronal cells as well as neoplastic glial components, the latter showing proliferative activity (Wolf et al., 1994), particularly the expression of the stem cell epitope CD34 in gangliogliomas has substantiated the hypothesis of an origin from developmentally compromised or dysplastic precursor malformations (Blümcke et al., 1999; Fassunke et al., 2004). However, the cellular nature of CD34 expressing elements in gangliogliomas had remained enigmatic. In order to characterize CD34 expressing ganglioglioma components, we have performed single-cell mRNA analysis in gangliogliomas (Fig. 30.2) (Fassunke et al., 2004). After in situ-RT and immunohistochemistry by an antibody against CD34 we used laser microdissection of individual ganglioglioma components. Subsequently, we used PCR systems with different lineage markers, and observed co-expression of CD34 and neurofilament (NFM) protein but not other lineage markers (Fig. 30.2). This finding is in line with neuronal elements of compromised differentiation indicated by expression of CD34 as inherent components of gangliogliomas. Complementary to this qualitative expression analysis approach starting from distinct laser microdissected cellular elements, we have addressed alterations in components of a major signal pathway for neuronal development and migration, i.e. the signaling cascade (Gilmore and Herrup, 2000). Lower mRNA transcript numbers of CDK5, DCX, p35 and dab1, coding for proteins that operate in the reelin signaling cascade, were found in gangliogliomas compared to normal brain tissue controls (Becker et al., 2001; Kam et al., 2004). Low expression of several components of the reelin signaling cascade may point towards impaired reeling cascade signaling in gangliogliomas as factor related to its dysmorphic appearance. Epigenetic modification of factors of the reelin signaling cascade (Kobow et al., 2009) should be taken into consideration in this context in the future. In order to learn about differential expression patterns with emphasis on developmental mechanisms more comprehensively, several large scale expression array studies have been carried out in gangliogliomas.

A.J. Becker

Recently, we have carried out an expression array study starting from discrete microdissected ganglioglioma and adjacent control brain tissue obtained from the neurosurgical access area to the tumor of identical patients that were carefully matched for equivalent glial and neuronal elements (Fassunke et al., 2008). A cluster analysis with all present genes based on euclidean distance as a samplewise distance measure and average linkage as a setwise distance measure showed tumors and controls to be fairly well separated (Fassunke et al., 2008). Intriguingly, even in highly differentiated tumors such as gangliogliomas, the parameter “tumor” apparently was more distinctive than “identical genetic background” with respect to comprehensive gene expression. Although gene expression in a considerable range of molecular cascades was affected in gangliogliomas, differential expression of more than individual genes related to only a certain circumscribed number of elementary cellular functions and molecular pathways, i.e., regulation of chromatin state and transcription factors, intracellular signal transduction, transduction of extracellular signals and cell adhesion, control of cell cycle and proliferation, development and differentiation (Fig. 30.3). As particularly interesting molecules, we observed two LIM domain-interacting transcripts, i.e. LIM domain only 4 (LMO4) and LIM domain binding 2 (LDB2), as strikingly lower in expression in gangliogliomas compared to controls (Fassunke et al., 2008). LIM domain-containing proteins are critical regulators determining cellular fate and differentiation in embryonal development (Dawid et al., 1998). Furthermore, LIM-HD transcription factors were observed as substantially present in neurons and essential for their development (Benveniste et al., 1998; Pfaff et al., 1996; Way and Chalfie, 1988). In order to follow further on a potential functional role of LDB2 in neuronal and ganglioglioma development, we used LDB2 silencing by specific shRNAs in cultured primary neurons. These experiments demonstrated a striking deficit of arborization in neuronal development (Fassunke et al., 2008) reflecting features of dysmorphic neuronal components of gangliogliomas. Functionally, neuronal elements with impaired connectivity by significantly contribute to impaired neuronal network function and resulting hyperexcitability.

30 Epilepsy-Associated Gangliogliomas: Identification of Genes with Altered Expression

271

Fig. 30.3 Overview of differential gene expression in gangliogliomas. Intriguingly, differential expression of more than individual genes in gangliogliomas related to a rather defined number of elementary cellular functions and molecular pathways. Only a minor fraction of differentially expressed transcripts was not sorted in one of the functional groups. A detailed list of differentially expressed transcript has been presented before (Fassunke et al., 2008)

Correspondingly, alterations of the balance between excitation and inhibition were recently reported by important comprehensive expression array studies in gangliogliomas (Aronica et al., 2008; Samadani et al., 2007). A highly refined single cell expression analysis showed abundance of the metabotropic glutamate receptor (mGluR) 5 in neuronal ganglioglioma elements (Samadani et al., 2007). Conversely, reduced expression of several GABAa receptor (GABAAR) subunits was present in gangliogliomas, which may reflect insufficient GABAergic inhibition (Aronica et al., 2008; Samadani et al., 2007). Aronica and colleagues reported on aberrant expression of a variety of ion channels and ion transporters (Aronica et al., 2008) that may point to impaired ion homeostasis in gangliogliomas, which represents a general factor for epileptogenicity (Beck and Yaari, 2008; Becker et al., 2008). A recent factor that gains considerable importance in the pathogenesis of gangliogliomas and seizure development is given by the strong activation of particularly innate inflammatory reactions (Aronica et al., 2008). Increased expression of inflammatory interleukins (e.g., IL-1β), activation of the complement cascade and the Toll-like receptor pathway have been detected partially by expression array studies (Aronica

et al., 2005, 2008). As these data and considerations may have demonstrated, expression array studies have provided substantially improved insights in molecules and pathogenetic cascades operating in gangliogliomas. How are we going to optimally use and exploit this data treasure in the future?

Discussion Large-scale expression data on gangliogliomas has come up with a substantial number of differentially expressed transcripts. Which of these findings are pathogenetically important, which represent side effects? Can we learn from differential expression patterns about transcriptional command structures active in gangliogliomas? Considering these questions, expression array data can gain an overadditive power by integration with complementary technical approaches that aim to understand the functional impact of differential gene expression as well as to unravel genomic aberrations that interfere with gene expression in gangliogliomas. Complex analyses starting from expression array data to characterize concerted promoter control

272

modules taking into account differential promoter activation due to slight allelic variants of gene promoter and resulting variability in transcription factor binding affinity in humans will provide completely novel insights into the pathogenesis of complex disorders such as glioneuronal neoplasms (Kasowski et al., 2010; Marsh et al., 2009). Further important insights into the pathogenesis of gangliogliomas may be expected by the comprehensive analysis of expression array and comparative genomic hybridization (CGH) data. A recent, somewhat unexpected finding was given by a high incidence of genomic aberrations observed in more than 50% of gangliogliomas under study (Hoischen et al., 2008). Considering the glial ganglioglioma component as neoplastic and cytologically resembling many features of low-grade astrocytomas, we compared the CGH patterns of gangliogliomas and diffuse astrocytomas (WHO grade II). In an unsupervised cluster analysis, diffuse astrocytomas formed subclusters within a group of gangliogliomas that demonstrated no strong similarities in the pattern of genomic imbalances. In contrast a genetically clearly distinct ganglioglioma cluster characterized by combined gains of chromosomes 5, 7, 8 and 12 (group I) was clustered distinct from diffuse astrocytomas, albeit strong histological similarities of the constituting astroglial elements. The analysis of these CGH data with comprehensive gene expression array results of respective gangliogliomas will provide novel information on putative functional consequences of genetic aberrations with respect to gene expression. This approach will comprehensively provide insight in amplification of oncogenes/loss of heterozygosity of tumor suppressor genes and corresponding alterations of gene expression. Considering the recently published fact that array-CGH detected genetic alterations such as amplification of the oncogene CDK4 in a fraction of tumor cells in gangliogliomas (WHO grade I) can predict a malignant recurrence as glioblastoma multiforme, in which tumor cells derived from this fraction represent the majority of cellular elements (Hoischen et al., 2008), expression-/CGH-array based biomarkers for also low grade brain tumors may be expected in the future. Such strategies will be a starting point for personalized therapy also of generally rather benign brain neoplasms such as gangliogliomas.

A.J. Becker

References Aronica E, Gorter JA, Redeker S, Ramkema M, Spliet WG, van Rijen PC, Leenstra S, Troost D (2005) Distribution, characterization and clinical significance of microglia in glioneuronal tumours from patients with chronic intractable epilepsy. Neuropathol Appl Neurobiol 31:280–291 Aronica E, Boer K, Becker A, Redeker S, Spliet WG, van Rijen PC, Wittink F, Breit T, Wadman WJ, Lopes, da Silva FH, Troost D, Gorter JA (2008) Gene expression profile analysis of epilepsy-associated gangliogliomas. Neuroscience 151:272–292 Beck H, Yaari Y (2008) Plasticity of intrinsic neuronal properties in CNS disorders. Nat Rev Neurosci 9:357–369 Becker AJ, Löbach M, Klein H, Normann S, Nöthen MM, von Deimling A, Mizuguchi M, Elger CE, Schramm J, Wiestler OD, Blümcke I (2001) Mutational analysis of TSC1 and TSC2 genes in gangliogliomas. Neuropathol Appl Neurobiol 27:105–114 Becker AJ, Chen J, Zien A, Sochivko D, Normann S, Schramm J, Elger CE, Wiestler OD, Blümcke I (2003) Correlated stageand subfield-associated hippocampal gene expression patterns in experimental and human temporal lobe epilepsy. Eur J Neurosci 18:2792–2802 Becker AJ, Blümcke I, Urbach H, Hans V, Majores M (2006) Molecular neuropathology of epilepsy-associated glioneuronal malformations. J Neuropathol Exp Neurol 65:99–108 Becker AJ, Figarella-Branger D, Wiestler OD, Blümcke I (2007) Ganglioglioma and gangliocytoma. In: Louis, DN, Ohgaki, H, Wiestler, OD, Cavenee, WK (eds) WHO classification of tumours of the central nervous system. IARC, Lyon:103–105 Becker AJ, Pitsch J, Sochivko D, Opitz T, Staniek M, Chen CC, Campbell KP, Schoch S, Yaari Y, Beck H (2008) Transcriptional upregulation of Cav 3.2 mediates epileptogenesis in the pilocarpine model of epilepsy. J Neuroscience 28:13341–13353 Benveniste RJ, Thor S, Thomas JB, Taghert PH (1998) Cell typespecific regulation of the drosophila FMRF-NH2 neuropeptide gene by Apterous, a LIM homeodomain transcription factor. Development 125:4757–4765 Blümcke I, Wiestler OD (2002) Gangliogliomas: an intriguing tumor entity associated with focal epilepsies. J Neuropathol Exp Neurol 61:575–584 Blümcke I, Giencke K, Wardelmann E, Beyenburg S, Kral T, Sarioglu N, Pietsch T, Wolf HK, Schramm J, Elger CE, Wiestler OD (1999) The CD34 epitope is expressed in neoplastic and malformative lesions associated with chronic, focal epilepsies. Acta Neuropathol 97:481–490 Bowtell DD (1999) Options available-from start to finish-for obtaining expression data by microarray. Nat Genet 21: 25–32 Brown PO, Botstein D (1999) Exploring the new world of the genome with DNA microarrays. Nat Genet 21:33–37 Crino PB, Duhaime AC, Baltuch G, White R (2001) Differential expression of glutamate and GABA-A receptor subunit mRNA in cortical dysplasia. Neurology 56:906–913 Dawid IB, Breen JJ, Toyama R (1998) LIM domains: multiple roles as adapters and functional modifiers in protein interactions. Trends Genet 14:156–162

30 Epilepsy-Associated Gangliogliomas: Identification of Genes with Altered Expression Duggan DJ, Bittner M, Chen Y, Meltzer P, Trent JM (1999) Expression profiling using cDNA microarrays. Nat Genet 21:10–14 Fassunke J, Majores M, Ullmann C, Elger CE, Schramm J, Wiestler OD, Becker AJ (2004) In situ-RT and immunolaser microdissection for mRNA analysis of individual cells isolated from epilepsy-associated glioneuronal tumors. Lab Invest 84:1520–1525 Fassunke J, Majores M, Tresch A, Niehusmann P, Grote A, Schoch S, Becker AJ (2008) Array analysis of epilepsyassociated gangliogliomas reveals expression patterns related to aberrant development of neuronal precursors. Brain 131:3034–3050 Gilmore EC, Herrup K (2000) Cortical development: receiving reelin. Curr Biol 10:R162–166 Hoischen A, Ehrler M, Fassunke J, Simon M, Baudis M, Landwehr C, Raldwimmer B, Lichter B, Schramm J, Becker AJ, Weber RG (2008) Comprehensive characterization of genomic aberrations in gangliogliomas by comparative genomic hybridization (CGH), array-based CGH and interphase-FISH. Brain Pathol 18:326–337 Kam R, Chen J, Blümcke I, Normann S, Fassunke J, Elger CE, Schramm J, Wiestler OD, Becker AJ (2004) The reelin pathway components disabled-1 and p35 in gangliogliomas – a mutation and expression analysis. Neuropathol Appl Neurobiol 30:225–232 Kasowski M, Grubert F, Heffelfinger C, Hariharan M, Asabere A, Waszak SM, Habegger L, Rozowsky J, Shi M, Urban AE, Hong MY, Karczewski KJ, Huber W, Weissman SM, Gerstein MB, Korbel JO, Snyder M (2010) Variation in transcription factor binding among humans. Science 328: 232–235

273

Kobow K, Jeske I, Hildebrandt M, Hauke J, Hahnen E, Buslei R, Buchfelder M, Weigel D, Stefan H, Kasper B, Pauli E, Blümcke I (2009) Increased reelin promoter methylation is associated with granule cell dispersion in human temporal lobe epilepsy. J Neuropathol Exp Neurol 68:356–364 Lipshutz RJ, Fodor SP, Gingeras TR, Lockhart DJ (1999) High density synthetic oligonucleotide arrays. Nat Genet 21: 20–24 Luyken C, Blumcke I, Fimmers R, Urbach H, Wiestler OD, Schramm J (2004) Supratentorial gangliogliomas: histopathologic grading and tumor recurrence in 184 patients with a median follow-up of 8 years. Cancer 101:146–155 Marsh B, Stevens SL, Packard AE, Gopalan B, Hunter B, Leung PY, Harrington CA, Stenzel-Poore MP (2009) Systemic lipopolysaccharide protects the brain from ischemic injury by reprogramming the response of the brain to stroke: a critical role for IRF3. J Neurosci 29:9839–9849 Pfaff SL, Mendelsohn M, Stewart CL, Edlund T, Jessell TM (1996) Requirement for LIM homeobox gene Isl1 in motor neuron generation reveals a motor neuron-dependent step in interneuron differentiation. Cell 84:309–320 Samadani U, Judkins AR, Akpalu A, Aronica E, Crino PB (2007) Differential cellular gene expression in ganglioglioma. Epilepsia 48:646–653 Way JC, Chalfie M (1988) mec-3, a homeobox-containing gene that specifies differentiation of the touch receptor neurons in C. elegans. Cell 54:5–16 Wolf HK, Müller MB, Spänle M, Zentner J, Schramm J, Wiestler OD (1994) Ganglioglioma: a detailed histopathological and immunohistochemical analysis of 61 cases. Acta Neuropathol 88:166–173

Index

A Abdulrauf, S.I., 139 Acetazolamide, CA inhibitor, 68 Actin-associated proteins, 84 Adenomatous polyposis coli gene (APC), 41 Ageing, and cancer, 5 Agonistic Fas receptors, 25 AGT, see O6 -methylguanine-DNA methyltransferase (MGMT) Alkylating agents, 89 Allograft implantation models, 188 Alpha-carbonic anhydrase family, 65 Altinoz, M.A., 60 5-Aminolevulinic acid (5-ALA), 239 Anaplastic astrocytomas, 36–37, 135, 213 CBV measurements, 216 MGMT IHC of, 92 p16/INKa4 protein, relation with p53, 27–28 Angiogenesis, 135–136 Angiogenesis-related proteins, 137–138 Antiapoptotic proteins, see Astrocytoma(s) Antisera, 108 ANXA1 (annexin 1), 49 Aphasia, 224 Apolipoprotein Apia-I, 191 Apoptosis, 136, 146, 153 activation, and signaling pathways, 23–24 Bcl-2 controls, 29–30 biological significance, 23 caspases role in, 30 defined, 23, 121–122 FasL and TRAIL mediated, 25 in gliomas, study using MRS apoptotic cell density, 123 ca 2.8 ppm Lip/MM peak from PUFAs, 125–126

methodology, 122 taurine concentration in glioma biopsies, 124–126 ligands and effector caspases suppress, 25 pathways of, 121 proteins, inhibitors of, 30–31 regulatory function of PTEN, 28 repression of Bcl-2 and survivin, 26, 32 role in gliomas, 30, 33 TNF-induced, 25 ubiquitination role in, 31 Aquaporin-1 (AQP1), 240 Argon lasers, 174 Aronica, E., 271 Assimakopoulou, M., 61 Asthagiri, A.R., 247 Astrocytic tumors, see Astrocytoma(s) Astrocytoma(s), 57–58, 70, 135, 213–214, 259 anaplastic astrocytoma, 36 antagonist RU486 role in, 60 antiapoptotic proteins role in Bcl-2 proteins family, 29–30 death ligands, 23–26 IAPs, activation in gliomas, 30–32 p53, and cell cycle progression E2F role, 26–27 PTEN relationship with p53, 28–29 receptors and messengers role, 23–26 survivin and cell cycle progression, 32–33 TNF-induced NF-κB activation in, 25 biopsy, analysis of, 122–126 carbonic anhydrase IX diagnostic tool in grading astrocytomas, studies, 69–70 evaluation in tumors, 68 prognostic significance of, 68–69 role, 68–70

M.A. Hayat (ed.), Tumors of the Central Nervous System, Volume 5, DOI 10.1007/978-94-007-2019-0, © Springer Science+Business Media B.V. 2012

275

276

Astrocytoma(s) (cont.) CBV/rCBV measurements predictive role, in survival and recurrence, see Cerebral blood volume (CBV) cell motility, IF proteins role GFAP, 85–86 nestin, 81, 85–86 synemin, 82–85 vimentin, 81, 85–86 cerebellar, spontaneous regression of, see Cerebellar astrocytomas (CA) diffuse astrocytoma, 36 doppel protein functionality biochemical features, and cellular localization, 18–20 distinguishing subtypes of, 16 functional pathways analysis, 19–20 gene expression analysis, tumor marker, 13, 15–17 glial tumors, gene expression, 16, 18 interaction analysis, 19–20 over-expression, 16–17 glioblastoma multiforme, 35 higher MVD in p53 mutated low-grade, 140 high recurrence rate of, factors, 81 hormonal therapy for, 58, 62 Human Angiogenesis Array, 137–138 IF proteins role in cell motility, 85–86 IHC for MVD, 136–137 Laser-Induced Fluorescence Spectroscopy (LIFS) for, 162 low grade, p53 involvement, 135 pilocytic astrocytoma, 36 P53 mutation, 138 association with MVD, 138–139 characterization, 137 higher number of vessels, 138 PR isoforms expression regulation, and malignancy grades, 59–60, 62 function in growth of, 57, 60–61 progesterone (P) genomic mechanism of action, 58 role in cell growth and proliferation of, 57, 60 regression, 143 SEGA, see Subependymal giant cell astrocytomas (SEGAs) statistical analyses, 138 taurine role, in apoptosis, see Apoptosis

Index

therapy, synemin and other IF proteins prospects, 86–87 treatments, depending factors, 58 See also Gliomas Autism, 149–150, 154 Axin, tumor suppressor, 40–41 B Bannykh, S.I., 199 Bcl-2 protein family, 23 controls apoptosis, action mechanism, 29–30 groups, 29 Bigner, S.H., 111 Biomarker discovery, methods cytogenetic method, 111–112 digital karyotyping technique, 114–115 EGFR amplification technique, 111–112 EGFRvIII expression, 112 genomic characterization technique, 116–117 immunologic methods, 108–109 integrated genomic analysis, 115–116 karyotypic analysis, 111 large-scale array analysis, 113 using double-minute (DM) chromosomes, 111 using gangliosides, 109–110 using gene expression arrays, 114 using GLI transcript, 112–113 using tenascin, 110 Bisdas, S., 216–218 Bisulfite DNA treatment, 6–7 Bjerkvig, R., 109 Bleier, A.R., 180 Blood brain barrier (BBB), 188 Bonéy-Montoya, J., 60 Borit, A., 143 Bourdon, M.A., 110 Bouvier, C., 199 Bovenzi, V., 60 Bovine spongiform encephalopathy (BSE), 13 Bowers, D.C., 201 Bown, S.G., 173–174 Brain Cancer, nanotechnology based therapy methods to increase targeting specificity, 191–192 nanomaterials delivery, 190–191 convection-enhanced delivery (CED), 192–193 systemic delivery, 190–191 nanoparticle formulations, 189–190 strategies to overcome BBB, 192 See also Cancer

Index

Brain thermal lesions, evolution on MRI, 181–182 Brain tumors, 57, 60, 135, 195 cell cultures, 107 computer-based QOL monitoring, ePROM, 226–227 fluorescence signature and, 170 Laser-Induced Fluorescence Spectroscopy (LIFS) for, 162 malignant, nanotechnology-based therapy for, 187–189 PROM, see electronic Patient-Reported Outcome Monitoring (ePROM) specific Pros, 225 symptoms, 223–224 TR-LIFS, tool for intra-operative diagnosis, see Time-resolved laser induced fluorescence spectroscopy (TR-LIFS) variety of, and treatment possibilities, 223 Breast cancer, CA IX as tumor biomarker, 67 Brell, M., 92 Britz, G.W., 235 Brunetaud, J.M., 177 C Cabrera-Munoz, E., 61 Cachexia, 196 CA9 gene, 65, 67 Calcifications, 196, 199 Camacho-Arroyo, I., 58–59 Camptothecin, 146 Cancer ageing and, 5 and DNA methylation failure, 5–6 ERK/MAPK pathway deregulation, 102 importance, of imaging of HIF-1-active tumors, 129 NF-κB in genesis and progression of, 25 p16/INKa4 protein role, 27–28 role of synemin in, 83 -targeted PROs, 225 types, 223 Cantharidinare, 147 Cao, V.T., 93 Capper, D., 92–93 Carbonic anhydrase (CAs), 65 Carbonic Anhydrase IX (CA IX), 65–66 diagnostic evaluation monoclonal M75 anti-human CA IX antibody use, 68

277

diagnostic tool, in grading astrocytomas, 70 expression in Barrett’s-associated adenocarcinomas, 66 and disease prognosis, 67 ectopic, 66 and hypoxia responsive element (HRE), 67 increased, 66–67 malignant gliomas, 67 in normal tissue, 66 predominant CA isozyme in tumors, 66 prognostic significance (studies), as valuable marker, 67–70 role in astrocytomas, 68–70 role in carcinogenesis, function mechanism, 67–68 Carboplatin, in treatment of PMAs, 207 Carpentier, A., 175–176, 180, 183 Carroll, R.S., 60 β-Catenin, 35, 37–39, 42–43 See also Wnt signaling pathway Cathepsin D, lysosomal proteinase, 74 CBV, see Cerebral blood volume (CBV) CD95 ligand (CD95L), 23 CD34 protein, 256 Cellular differentiation, 136 Central nervous system (CNS), 187–188 and doppel protein expression, 13, 15, 19 germ cell tumors, 95 arise from pleuropotential embryonic cells, 95 in Asian population, higher incidence, 97 benign and malignant, 95 clinical presentation, tumors in, 96–97 germinoma and non-germinoma, histological examination, 95–96 locations, mainly midline, 95–97 hemangioblastomas, see Hemangioblastomas low-grade neoplasms PAs, pediatric tumors of, see Pilocytic astrocytomas (PAs) malignant tumors, nanotechnology-based therapy for, 187–189 tumors and EGBs presence diagnosis, 78 Ceppa, E.P., 198, 207 Cerebellar astrocytomas (CA), 143 regression, 144, 147 criteria for, 144 CT scans, spontaneous regression, 144 hormonal changes and chances of, 145 imaging surveillance in, 145 incidence and time course of, 144–145 mechanism and multiple factors, 145–146

278

Cerebellar astrocytomas (CA) (cont.) residual tumours, smaller volume and higher chances of, 146 therapy for, surgical resection, 144 Cerebellar lesions, 195 Cerebral blood volume (CBV) estimation, by MRI arterial spin labeling techniques, 215 DSC-MRI, 214–215 gradient echo pulse sequence use, 215 preload approach, problem, 214 T1-based methods, 214 measurements, predictive value, 217–219 biases in prediction, 217 correlation between histopathological grade and, 218 cut-off values, 218 data collection from longitudinal studies, 219 EORTC criteria, 218 rCBV value, after combined radiation and temozolamide therapy, 219 methodological considerations, and survival rates, 215–216 CBV mapping, 215 histogram analysis, of rCBV values, 216 hot-spot ROI method, 216 limitations, 216 in predicting astrocytoma histopathologic grade, 214 Cerebrospinal fluid (CSF), 149, 203 Chang, S.D., 248 Chemotherapy, for PMA, 203–204, 206–208 Chen, Y.Y., 43 Chikai, K., 207 Chinese herbs, 146–147 ChIP-on-chip method, 9 Choi, Y.J., 47, 51 Choroid plexus, 65 Cisplatin (CDDP), in treatment of PMAs, 203, 207 Coagulation factor III (CF III), 138, 140 Coagulation factor VIIa, 140 CO2 lasers, 174 Colchicinamide, 147 Colchicine, 146 Coley’s vaccine, 145 α(II) collagen prolyl-4-hydroxylase, 139 Comincini, S., 18 Computer-based Health Evaluation System (CHES), 227

Index

Cottingham, S.L., 206–207 CpG islands, 4–5, 90 Curzerenone, 147 CyberKnife SRS, 249 Cyst, 239–240 Cystic tumors, see Pilocytic astrocytomas (PAs) Cytogenetic method, 111–112 Cytoskeletal protein, 81 D Death domain (DD), 24 Death effector domain (DED), 24 Death inducing signaling complex (DISC), 24 Death ligands, 23 Death receptors, 23–24 De Porter, J., 175, 180 Desmuslin, 82 Devaux, B.C., 176 Dhermain, F., 218 Di, C., 115 Dickkopf family proteins, 35, 39 Diencephalic syndrome, 204 Dietary supplements, anti-cancer effects, 146–147 Diffuse astrocytoma(s), 10, 36 main tumor entities, 89–90 MGMT IHC expression, 89–90 compared to MGMT promoter methylation, 91 in glioma and non-neoplastic cells, 89, 91–92 marker of patient outcome, 92–93 technical considerations, 90–91 MGMT protein and resistance alkylating agents, 89–90 Digital karyotyping technique, 114–115 Digital oscilloscope, 164 3,3-Diindolylmethane (DIM), 146 Diode laser, 176 Dirven, C.M., 200–201 Discriminant function analysis (DFA), 165 Dishevelled (Dvl), 40 D-54 MG glioma cells, 109 DNA, apoptotic DNA fragmentation, 121 DNA methylation, 3–4 analysis, methods based on DNA chemical modification, 6 ChIP-on-chip method, 9 MALDI-TOF mass spectrometry, 8 methylation-sensitive restriction enzymes usage, 6 methylation-specific PCR (MSP), 7

Index

microarray expression profiling, 8 primers usage, 8 restriction landmark genomic scanning (RLGS) method, 8–9 techniques comparison, 7 in astrocytic tumors, for diagnosis and prognosis, 9–10 failure, as cause of disease or cancer, 5–6 hypomethylation, 5–6 relevance in normal cells, 4 role in mammalian CNS development and function, 9 DNA methyltransferases, 4 DNA methyltransferases (DNMT) genes, 4 DNA repair, 136 DNMT, see DNA methyltransferases Doppel gene, discovery, 13–15 Doppel protein biochemical features, 15, 18–19 cell migration process, contribution in, 13, 20 functional pathways, 19–20 and interaction analysis, 19–20 post-translational modifications, 18 and prion proteins, comparison and similarities, 13–15, 18, 20 similar exon–intron architecture, with prion gene, 14 and structural characterization, 15 Double-minute (DM) chromosomes, 111 Doxorubicin-loaded, PEG coated PHDCA nanoparticles, 191 Dysembryoplastic neuroepithelial tumours (DNETs), 259, 261 E E2F family protein, 27 EGBs, see Eosinophilic granular bodies (EGBs) EGFR, see Epidermal growth factor receptor gene (EGFR) EGFRvIII expression, 112 electronic Patient-Reported Outcome Monitoring (ePROM) implementation of, 226–227 software, 227 tele-monitoring, 227–228 usage of proxy-rating, 228–229 Ellmann, S., 60 Endostatin, 139 EORTC Brain Cancer Module (BN20), 226

279

EORTC QLQ-BN20, 226, 228 Eosinophilic granular bodies (EGBs), 73–74 cathepsin D role, in cell apoptosis, 76 contents and morphologies, in tumor, 74 H&E (hematoxylin and eosin) staining and anti-GFAP and anti-CSE1L antibodies, 75 Mayer’s hematoxylin for staining, 75–76 observation protocol using immunohistochemistry, 77 with antibodies against LAMP-1 and LAMP-2, 73–74 origin, 75–77 PAS (periodic acid-Schiff) staining and anti-GFAP and anti-CSE1L antibodies, 75 relation to lysosomal system, 73 role in cyst development in pilocytic astrocytomas, 73, 75–77 Epidermal growth factor receptor gene (EGFR), 111–112, 192 Epigenetics defined, 3 states, modification, 4–5 Epigenetic therapy, 11 Epileptogenicity, of brain tumors cellular composition and neurochemical profile, 262–263 of peritumoral zone, 263–264 ERM proteins, 261 European Organization for Research and Treatment of Cancer core questionnaire (EORTC QLQ-C30), 225, 229 Everolimus, 45, 50–51 F FACT-Br Symptom Index (FBrSI), 225–226 FasL (Fas ligand), 25 Fernandez, C., 207 Ferroli, P., 235 Fiber optic probe, 164 Fisher, B.J., 201 Fisher, P.G., 195 FLAIR image, 218 Fluorescence measurements types, 162 Fluorescence spectroscopy, 162 Fluorophores, 162, 169 Foltz, G., 39 Franz, D.N., 52 FRAT1 (frequently arranged in advanced T-cell lymphomas-1), 40

280

Freilinger, A., 47 Functional Assessment of Cancer Therapy-Brain Module (FACT-Br), 225–226 Functional Assessment of Cancer Therapy general version (FACT-G), 225 Functional Assessment of Chronic Illness Therapy (FACIT) measurement system, 225 Furnari, F.B., 187 Fusion genes BRAF and KIAA1549, in PAs, 100–101 MAPK activation via RAF1, 101 in pilocytic astrocytoma, KIAA1549, 100–101 Fuss, M., 218 FVIII-immunostaining, 139 G Gajjar, A., 195 Gamma-glutamyl-Semethylselenocysteine (GGMSC), 146 Gamma Knife SRS, 247 Gangliocytoma, 258–259 Gangliogliomas (GG) alterations in Pi3K-mTOR pathway components, 263 association with temporal lobe epilepsy (TLE), 267 cellular pathology analysis, 269 coexistence with cortical dysplasia, 259 CT for imaging, 254 differential diagnosis, 258–259 differential gene expression in, 270–271 LIM domain-interacting transcripts, 270 reelin signaling cascade, factors modification, 270 epileptogenicity, 263 expression of specific glutamate receptors (GluR) subtypes, 263 gene expression analysis, approaches, 272 aRNA amplification method, 269 glass microscope slide DNA arrays, 268–269 microarray systems, limitations, 269 oligonucleotide arrays, 268–269 in situ-RT and immunolaser microdissection, 268–269 histopathological features, and hallmark of, 257–258, 268–269 IHC characterization of, 258 imaging, 254 incidence and age/sex distribution, 254 intraoperative diagnosis, 256

Index

localization, 254 macroscopy of, 256 mechanisms of epileptogenesis, 261–264 molecular pathogenesis of, 260–261 neurophysiological features, 254–256 origin from, 268–269 pathogenesis, 263 perilesional changes, 264 prognostic factors and surgical outcome, 259–260 symptoms, 254 Gangliosides antigens, 109–110 patterns, 109–110 GBM, see Glioblastoma multiforme (GBM) Genetically engineered mouse models (GEMMs), 188 Genomic characterization technique, 116–117 Germ cell tumors arise from pleuropotential embryonic cells, 95 in Asian population, higher incidence, 97 benign and malignant, 95 clinical presentation, tumors in basal ganglia region, 96 pineal region, symptoms, 96–97 predilection for males, during puberty time, 97 suprasellar region, 96 germinona, 95 histological examination, cell types, 96 magnetic resonance imaging of, 96 locations, mainly midline, 95 and non-germinoma, histological examination, 95 choriocarcinoma, and teratomas, 96 yolk-sac tumors and embryonal carcinoma, 96 SEER-17 registry data for study, 97 See also Central nervous system (CNS) GFAP (Glial fibrillary acidic protein), 75, 82–83, 85–87 Giannini, C., 200 Glial tumors, malignancy grades, 16 Glioblastoma multiforme (GBM), 187, 214 develop de novo, 36 doppel expression, 16 epigenetic alterations in, 9 genetic changes associated, 37 GPNMB gene role, 153 gross total surgical resection (GTR) in, 233 nanotechnology-based delivery of therapeutics, features, 188–189

Index

primary and secondary, chromosomal aberrations in, 36 TP53 mutation, 37 Glioblastomas, 36, 135–136, 161 CA IX expression in, 69–70 cell types, 37 EGFRvIII mutation, 112 GBM, see Glioblastoma multiforme (GBM) p16/INKa4 protein, relation with p53, 27–28 poor prognosis, 58 Gliomas, 57, 135, 161 animal model systems, for preclinical trials, 188 astrocytomas, 35–36, 57 cell culture studies, limitation, 188 diagnosis, in vivo measurements fluorescence signature of diffuse astrocytoma, 170 spectroscopic parameters analysis, 170 TR-LIFS characteristics, 170 ependymomas, 35–36 of grades III-IV, 187 histological characteristics for grading, 35 IAPs activation in, 31–32 malignancy grades, and survival rates, 57–58 malignant, 187–188 mixed oligoastrocytomas, 35 nuclear polymorphism, 36 oligodendrogliomas, 35–36 progress, microvascular hyperplasia, 135 spectroscopic classification, 164 surgical resection of, 161–162 survivin and cell cycle progression, 32–33 Wnt/β-catenin/Tcf signaling pathway components in, 35, 38 Axin and APC proteins, 38 Axin-APC-GSK3β, 41 β-catenin, central player, 38, 41–42 Dishevelled (Dvl), 40 DKK family, 39–40 FRAT1, 40 Frizzled family, 37–38 frizzled receptors, 40 Lef/Tcf family transcription factors, 42 LRP5 and LRP6, 37, 39 neural stem cells regulation, 38 overview, 37 pygopus 2, 43 sFRPs and Dickkopf family, extracellular inhibitors, 39

281

Wnt genes associated with, 37, 39–40 Wnt inhibitory factor-1 (WIF-1), antagonist, 39 Wnt proteins, 39–40 Wnt receptor complex, 37 in women, 146 Glioneuronal lesions, 253, 261 Glioneuronal tumors, see Gangliogliomas (GG) GLI transcript, 112–113 Glutamate decarboxylase (GAD), 169 Glycogen synthase kinase 3 (GSK 3), 41 Glycosylphosphatidyl-inositol (GPI), 14–15, 18 Gonzalez-Aguero, G., 60–61 Goodwin, T., 97 Gottfried, O.N., 207 Gotze, S., 39 GPI, see Glycosylphosphatidyl-inositol (GPI) GPNMB (glycoprotein nmb), 49 Grasbon-Frodl, E.M., 91 Groucho (Grg/TLE) family, transcriptional co-repressors, 42 GTPase-activating protein, 47 Guba, M., 49 Guerra-Araiza, C., 59 Gunny, R.S., 146 Guo, G., 41 H Haapasalo, H., 201 Harada, H., 130 Harringtonine, 146 Health-related quality of life (HRQOL) assessment, 223 Helin, K., 58 Hemangioblastomas, 233, 239–240, 242, 245, 250 associated with von Hippel-Lindau syndrome, 233 conventional radiotherapy for, 246 cysts, formation mechanisms in aquaporin-1 (AQP1) role, 240 histologic appearance of, 246 intraoperative ICG videography use in, 233–235 resection of intracranial and spinal, 235–237 photodynamic diagnosis (PDD), with 5-ALA, 239–240 radiosurgery for contraindications, 246 current indications, 246 radiosurgical complications, 249–250 rationale for use, 246 stereotactic, 246–249

282

Hemangioblastomas (cont.) radiosurgical treatment of, key studies, 247–249 rare vascular tumors, 233 recurrence rate, 239, 245 residual tumors detection, by ALA-derived PpIX fluorescence method, 241–242 treatment modalities, 250 types, as per MRI studies, 240 Heme pathway regulation, ALAS feedback inhibition role, 240–241 synthesis steps, 240 Hengstschlaeger, M., 50 Henske, E.P., 48 Hernandez-Hernandez, T., 60 HIF-1-active tumors imaging, using 123 I-IPOS based on pretargeting approach, concept, 131–132 hypoxia imaging concept, 130–131 size-exclusion analysis, 131 High mobility group (HMG), 42 High-resolution magic angle spinning (HRMAS) 1 H MRS, 121–122 lipid peaks, 126 of non-necrotic (top spectrum) and necrotic (middle spectrum) biopsy, 123, 125 taurine concentrations, 126 Higuchi, N., 178 Hilvo, M., 68 Hirai, T., 218 Histology, of LITT-induced lesions, see Laser interstitial thermotherapy (LITT) Histone modifications, 3 HMG, see High mobility group (HMG) Homoharringtonine, 146 Horbinski, C., 199 Howng, S.L., 40 HRMAS; 1H MRS, see High-resolution magic angle spinning (HRMAS) Hulleman, E., 58 Human angiogenesis array, 137–138 Human ICF syndrome, 5 Humphrey, P.A., 112 Hydrocephalus, 149–150 Hydroxycamptothecin, 146 Hyperplasia, 135 Hyperthermia, 177 Hypoxia-inducible factors (HIFs), 47, 67 Hypoxic regions, 129

Index

I ICG, see Indocyanine green (ICG) Ichikawa, T., 48 Ichimura, K., 58 IHC markers, 90 ([123 I]iodobenzoyl)norbiotinamide (123 I-IBB), 129 123 I-IPOS, probe, 129 Immunohistochemistry, 131 Indirubin, 147 Indocyanine green (ICG), 234 intracranial practical applications, 234–235 intraoperative use, in hemangioblastomas, 233–237 resection of intracranial and spinal, 235 Inhibitor of apoptosis (IAP) proteins family activation in gliomas, 31–32 classification, 30–31 eight human members, 30 survivin, and cell cycle progression, 32–33 Inoki, K., 49 Inorganic nanoparticles, 189 Integrated genomic analysis, 115–116 Intermediate filament (IF) proteins functions, 81–82 GFAP (Glial fibrillary acidic protein), 75, 82–83, 85–87 nestin, 81–82, 85–87 prospects for astrocytoma therapy, 86–87 synemin, 81–83 vimentin, 81, 83, 85–87 See also Astrocytoma(s) Interstitial hyperthermia concept, 173–174 Intracranial germ cell tumors, malignant, 95 Intracranial xenografts, 191 Intraoperative ICG videography, applications, 234 in identifying regions of tumor, 235 in the resection of 100 craniotomies of mixed pathology, 235 in the resection of hemangioblastoma, 235 in resection of spinal hemangioblastomas, 235–236 to stain glioma margins in animal models, 234–235 in vascular neurosurgery, 234 Intratumoral, 240–241 accumulation, 191 administration, 191 cysts, 240–242 hemorrhages, 150 infusion, 192 necrosis, 240 Isbert, C., 177

Index

Ischemic tumour necrosis, 145 3 6 -isoLD1, expression of, 109 Ivanov, S., 69 Ivarsson, K., 177 J Jagannathan, J., 242 Jawahar, A., 248 Johnson, M.W., 48 Jolesz, F.A., 179 Jó´zwiak, S., 48, 51–52 K Kageji, T., 207 Kahn, T., 178, 180–181, 183 Kangasniemi, M., 176, 178, 181–182 Kano, H., 247 Karabagli, H., 247, 249 Karayan-Tapon, L., 92 Karnofsky performance score, 204 Karyotypic analysis, 111 Kastner, P., 58–59 Kazuno, M., 140 Keene, D., 97 Kerr, J.F., 23 Khalid, H., 60 KIAA1549 protein, 101 Kickhefel, A., 180 Klein, R., 200 Knudson’s two-hit model, tumor development, 48 Kohler, G., 108 Komakula, S.T., 207 Komotar, R.J., 207 Kondo, I., 111 Korkolopoulou, P., 69 Korur, S., 41 Kou, L., 176 Kraus, W.L., 59 Krueger, D.A., 52 Kudo, T., 130 Kuratsu, J., 97 L Lam, C., 45, 52 LAMP-1 and 2 (Lysosomal membrane proteins), 74–78 Lange, C.A., 58–59 Large-scale array analysis, 113 Laser fibers, 176

283

Laser-induced fluorescence spectroscopy (LIFS), 162 Laser interstitial thermotherapy (LITT), 173 applied to brain tumors, clinical studies CT-guided stereotactic procedures, 182 devices, manufactured for, 184–185 Gd-DTPA-enhancing rim, 183 heating process and temperature elevation, 183 histological analysis, 183 hyperthermia treatment with Nd-YAG laser, 182 real-time magnetic resonance-guided LITT system, 183 for deepseated tumors, 174 limitations, 175 MRI imaging and, 179 brain thermal lesions on MRI, evolution, 181–182 MRI thermal imaging sequences, 179–180 real time computation, 180–181 procedures on patients, with brain metastases, 174 with real-time MRI, 174 treatments consist of, 173 Laser (Light Amplification by Stimulated Emission of Radiation), 173–174 control delivery software with real time with MRI thermometry analysis, 181 emission in active medium, 175 functioning principles, 175–176 history, 174–175 interactions with biological tissues, mechanisms, 177 histology of LITT-induced lesions in brain tissue, 178–179 immediate and secondary, 177–178 thermal dosimetry, 179 main elements, 175 physics, 175 produce high energy light, properties, 175 technology, fundamental principles, 175 transmission of beam, 176 used in LITT, 176–177 use in neurosurgery, 173 Law, M., 218 LDL receptors (LDLR), 191 Lee, N., 52 Lef/Tcf family transcription factors, 42 Leonhardt, S.A., 58 Leon, S.P., 139 Lev, M.H., 218

284

Libermann, T.A., 111–112 LIM-domain-binding 2 (LDB2) gene, 261 LINAC-based radiosurgery, 247 Lipid NPs, categories, 190 Liposomes, 190 LITT, see Laser interstitial thermotherapy (LITT) Liu, X., 42 Liu, Z.J., 60 Lockshin, R.A., 23 Loncaster, J.A., 67 LTF (lactotransferrin), 49 Lycobetaine, 147 Lysosomal protease, 73–74 Lysosomes, 73–74 M Magnetic resonance imaging (MRI), 145, 254 astrocytomas characterization, imaging method, 214 brain, case report, 204–206 contrast T1 MRI follow-up after LITT treatment, 182 coronal MRI images, 255 diffusion-weighted, 214 DSC-MRI, 214–215 for evaluating intracranial, neoplastic disease, 196 with gadolinium enhancement at T1/T2/FLAIR weighting, 246 of gangliogliomas (GG), 253 imaging and LITT, 179 perfusion-weighted, 214 for pilocytic astrocytomas, 197 thermal imaging sequences, 179–180 tumor determined by, 183 T2-weighted (FLAIR) MR image, 218 to visualize spinal hemangioblastomas, 236 Mahaley, S.M., 108 MALDI-TOF mass spectrometry, 8 Male fertility, and doppel, 15, 21 Mamelak, A.N., 207 Mammal development, and DNA methylation role, 4–5 MAPK, see Mitogen activated protein kinase (MAPK) pathway Maser (Microwaves Amplification by Stimulated Emission of Radiation), 174 Massimino, M.L., 19 Mast cells, 245 Matrix metalloproteinase-9 (MMP-9), 140

Index

Matsunaga, S., 248 Maxwell, J.A., 91 McCowage, G., 207 McKenna, N.J., 58 Menovsky, T., 179 Mental retardation, 150 Methylation, importance in clinic, 10–11 Methylation-specific PCR (MSP) method, 7–8 O6 -Methylguanine-DNA methyltransferase (MGMT), 9, 89 IHC expression in diffuse gilomas, see Diffuse astrocytoma(s) immunohistochemistry with clone MT3.1 and MT23.3, 92 double, 91–92 expression, 9, 89–90 marker of patient outcome, 93 and non-neoplastic cells, 91–92 technical considerations, 90–91 tumors identification with loss of MGMT expression, 92–93 promoter methylation assays application and limitations of, 91 MGMT, see O6 -Methylguanine-DNA methyltransferase MGMT methylation, 9–11 MIB-1 labeling index, 200 Microarray expression profiling, 8 Microdissection approach, 267, 269–270 Microvascular hyperplasia, 135 Microvessel density (MVD) astrocytoma evaluation, 136 degree of vascularization, 136 function estimatimation, of p53 mutation status, 137 IHC detected P53 protein, 138 Milstein, C., 108 Mineura, K., 91 Mi, R., 51 Missense mutant p53 proteins, 139 Mitogen activated protein kinase (MAPK) pathway activation via BRAF fusion gene, in PAs, 103 alternative activation mechanisms, 103–104 targeted therapy against AZD6244, 103 PLX4032 inhibitor, 103 Sorafenib, target, 103 Mizobuchi, Y., 40 Momparler, R.L., 60

Index

Monoclonal antibodies (MAbs), 108–110 Monocrotaline, 147 Moore R.C., 15 Mordon, S., 177 Moss, J.M., 247, 249 Motor deficit, 224 MRI imaging, 145 See also Magnetic resonance imaging (MRI) MRI thermal imaging sequences, 179–180 MSP derived methods, 8 mTOR (mammalian Target Of Rapamycin) pathway, 46 inhibitors, anti-angiogenic effects effects on cell cultures and animal models of TSC, 51–52 everolimus (RAD001), selective nature, 52 rapamycin action process, 49–50 sirolimus, mechanism involved, 51 temsirolimus (CCI-779), action mode, 49–51 in Subependymal Giant Cell Astrocytoma, 48 genes regulating activity, 48 targeting, 52–53 in Tuberous Sclerosis Complex, 46–47 autophagy inhibition, 47 genes regulated by HIFs, 46 inactivating mutations TSC1/TSC2 gene, 46 Ras-homoloque-enriched in brain (Rheb) target, 46 mTOR protein, 46 Mulac-Jericevic, B., 59, 61 Muller, W., 40 Murai, Y., 235 N Nakasu, S., 93 Nanooncology, 187 Nanotechnology, 187–189 based delivery of therapeutics to GBM, 188 for brain cancer, 189 nanomaterials delivery, 190 convection-enhanced delivery (CED), 192–193 methods to increase targeting specificity, 191–192 strategies to overcome BBB, 192 systemic delivery, 190–191 nanoparticles (NPs), 188 designed to delivery, 188–189 formulations, 189

285

inorganic, 189 lipid, 190 liposomes, 190 polymeric, 189 polymeric micelles, 190 xenograft systems models, tumor biology, 188 Natural remedies and herbs, anti-cancer effects, 146–147 Necrosis prediction, 180 Necrotic biopsies, analysis of, 122–126 Neodimium(Nd)-YAG lasers, 173–177, 179–180, 182–183 Neovascularization, 135–136, 139 Nestin, 81–82, 85–87 Neural stem cells-gliomas, 39 Neurooncology, 107 Neurosurgery interstitial hyperthermia in, 174 ultrasound-based monitoring techniques not feasible for, 179 use of intraoperative ICG, 234 use of lasers in, 173 NF1 (Neurofibromatosis 1) gene, 100, 102 Niemela, M., 249 Nikuseva Martic, T., 41 NPTX1 (neuronal pentraxin I), 49 Nuclear factor-κB, 25 Nuclear medicine imaging, 130 Nuclear polymorphism, in tumor cells, 36 O Oligodendroglioma, 259 Oligonucleotide arrays, 268–269 Olson, J.J., 60 Oridonin, 147 Oxygen-dependent degradation domain (ODD) fusion proteins, 130 P PAI-1 gene, 140 Paixão Becker, A., 199 Paralogue compensation process, 18 Park, Y.S., 248 Parsons, D.W., 115 Pastorek, J., 65 Pastoreková, S., 68 Patient-reported outcomes (PROs), 224–225 commonly used PRO instruments, 229 computer-based, 226–228

286

Patient-reported outcomes (PROs) (cont.) tele-monitoring, 227–228 See also electronic Patient-Reported Outcome Monitoring (ePROM) Pecina-Slaus, N., 41 Pediatric gliomas, 195 high grade, and for MGMT IHC, 93 See also Pilocytic astrocytomas (PAs) Petronio, J., 207 P53 gene expression, 140 immunohistochemistry, 137 -mediated regulation on angiogenesis, in low grade astrocytomas, 135, 139–140 mutations, 138 in astrocytomas, 135, 137–138 characterization of, 137 protein expressions, 138–139 tumor suppressor, 136 P-glycoprotein (P-gp), 263 Phosphatidylethanolamine (PE), 190 Phospholipid, 190–191 Photoablative effect, 177 Photochemical effect, 177 Photodynamic diagnosis (PDD), for residual tumors, 240 Photomechanical effect, 177 Photosensitizing agent, 177 Photothermal effect, 177 PI3K-mTOR pathway, 261 Pilocytic astrocytomas (PAs), 36, 73–74, 195–196, 203, 259 biphasic tumors, 76, 99 cerebellum, most frequent site, 99 classic form, Rosenthal fibers presence, 196–198 common pediatric tumors of CNS, 73, 99 EGBs presence in microcysts, 76–77 histological analyses, 198–199 histopathology, 73 imaging characteristics, 196–197 immunohistochemical analyses, 199–201 and increased intracranial pressure in, 76, 78 LAMP-1, LAMP-2, and cathepsin D involvement in EGBs formation, 76 leptomeningeal infiltration in, 199 lysosomal proteinases, in EGBs role, 76–77 management complexity, 196

Index

MAPK activation via RAF fusion genes, 99–101 BRAF:KIAA1549 fusion genes, 100–101 BRAF V600E mutation, 103 fusion variants, 100–101 mutation of KRAS and BRAF, 101–102 NF1 gene mutation, 102 targeted therapy against MAPK pathway, 102–103 MIB-1 labeling indices of, 200–201 molecular genetic changes in, 100 oligodendroglioma-like features, 199 of optic pathways, 195 role of EGBs in cyst development in, 73, 75–77 slow-growing tumors, 76 treatment strategies, 196 tumorigenesis, importance of MAPK signaling, 99–100 variants, 99, 197–198 See also Eosinophilic granular bodies (EGBs) Pilomyxoid astrocytoma (PMA), 198, 203 case reports, 204–206 chemotherapeutic regimens, 207–208 clinical characteristics, 203 different features from PA, 203, 206 drug combination cisplatin (CDDP)/carboplatin (CBDCA) and etoposide, 203 higher rate of recurrence, 203, 207 management of, 203 MIB-1 labeling index, 204 with monomorphous pilomyxoid features, 206 patients with CSF dissemination, 204 temozolomide as first line adjuvant chemotherapy, 207–208 therapeutic strategies for PMA patients, 207 WHO classification, 207 See also Pilocytic astrocytomas (PAs) Pinski, J., 60 Piroli, G., 60 Plasminogen activator (PA), 140 Pleomorphic xanthoastrocytoma, 259 P53, nuclear phosphoprotein activity regulation, 26 cell cycle progression, control mechanism, 26–27 as genome guardian, 26 p16/INKa4 protein, relation with, 27–28 relationship between PTEN and, 28–29 Pollack, I.F., 93, 195 Polymeric micelles, 190

Index

Polymeric nanoparticles, 189 POS and 123 I-IBB method for imaging HIF-1 active regions in tumors, 130–131 in vivo molecular imaging, 130 pretargeting approach, 131–133 oxygen-dependent degradable probe, development, 130 P response elements (PRE), 58 Pretargeting approach, 129, 131–133 advantages, 132 defined, 129, 131 for imaging of HIF-1-active tumors, principle, 132–133 Preusser, M., 92 Prion diseases, see transmissible spongiform encephalopathies (TSEs) Prion–doppel interaction, 13, 19–20 Prion-like protein, see Doppel protein PR isoforms, 57 expression, regulation of, 59–60, 62 estrogens effects, and interaction with ERs, 59 PR-B and PR-A, 57 regulation and function in astrocytomas, 57, 60–61 PR-A transfection effects, on U373 human astrocytomas growth, 61 regulation by phosphorylation, 58–59, 62 transcriptional activity of, 58–59 U373 and D54 cell lines, expressed in, 57 up-regulation by E, 59 See also Progesterone (P) Probe, 129–132 Proescholdt, M.A., 70 Progesterone (P), 57 genomic mechanism of action, PR interaction with, 58–59 induced cell proliferation, in cell lines, 60 interaction with intracellular receptor (PR), 57 and PR isoforms role in astrocytomas cell growth, 57, 60 Programmed cell death, see Apoptosis Proliferating cell nuclear antigen (PCNA), 27 Proptosis, 195 Proton-resonance frequency (PRF), 175 Protoporphyrin IX (PpIX), 239 PTEN, tumor suppressor gene, 28–29 P53, tumor suppressor gene, 136 Pu, P., 40–41 Pygopus 2, 42

287

Q Qin, K., 19 QOL-instrument, 225 R Radionuclides, 132 Radiosurgery, 173, 182, 245–246 contraindications, 246 current indications, 246 frame-based linear accelerator, 249 indications, 246 key studies, 247–249 radiosurgical complications, 250 rationale for use, 246 stereotactic, 245–249 See also Hemangioblastomas Radiotherapy, 145 Rapamycin, 45, 49–50 rCBV, see Relative cerebral blood volume (rCBV) Real time computation, 180–181 Real time MRI thermal imaging, 180 Reelin, 270 Relative cerebral blood volume (rCBV), 213, 215–218 Relaxivity, 214–215 Reoxygenation, 131 Restriction landmark genomic scanning (RLGS) method, 8–9 disadvantages, 9 Reticuloendothelial system (RES), 190 Rett’s syndrome, 5 Richardson, E.P. Jr, 143 Richer, J.K., 58–59, 61 RND3 (Rho family GTPase 3), 49 Rodriguez, F.J., 197 Roggendorf, W., 200 Roninson, I.B., 113 Rosenfeld, M.G., 58 Rosenthal fibers, 203, 206 Rosner, M., 50 Roth, W., 39 Rousseau, A., 57 Roux, F.X., 176–177, 179, 182 Ruby-based laser, 174–176 Rüegg, S., 49 S Saarnio, J., 66 Sager, G., 60 Sandlund, J., 67

288

Sapareto, S.A., 179 Saraswathy, S., 218 Sarcomas, 145 Sasai, K., 91 S100A11 (S100 calcium binding protein a11), 49 Sathornsumetee, S., 69 Saunders, D.E., 145 Schatz, S.W., 178 Schlosser, S., 93 Schober, R., 178 Schubert, G.A., 236 Schulze, C.P., 178–180 Schulze, P.C., 183 Schwabe, B., 181, 183 Secondary glioblastomas, 139 Seizures, 150 Serpin E1, 138, 140 sFRP (secreted Frizzled Related Protein), 39 SFRP4 (secreted frizzled-related protein4), 49 Shadoo proteins, 19 Shimizu, N., 111 Shou, J., 40 Silica nanocarriers, 189 Sirolimus, antiangiogenic effects, 49 Smith, M., 97 Smoots, D.W., 143, 145 Sorafenib, 103 Span, P.N., 67 Spontaneous regression, of tumour, 143–144 See also Cerebellar astrocytomas (CA) Spriggs, A.I., 111 Stafford, R.J., 176 Steady state fluorescence spectroscopy, 162 Stellar, S., 174 Stening, K., 60 Stereotactic radiosurgery, 246–249 Steroid hormones, 146 Steroid receptor coactivator (SRC) family, 58 Streptavidin, 137 Strong, J.A., 196 Students’t-test, 4 Stupp, R., 207 Subependymal giant cell astrocytomas (SEGAs), 47–49, 51–52 analysis of gene expression profiling in, 151 defined, 149–150 everolimus impact on, 45 gene expression profiling analysis Affymetrix microarrays, 151

Index

gene regulation, by mTOR kinase at transcriptional level, 155–156 genes with highest up-or down-regulation scores in, 152–153 mixed-lineage phenotype, 155 mTOR effector genes in, 49 mTOR pathway in, 48, 150–151 genes regulating activity, 49 targeting, everolimus and rapamycin treatment, 52–53 TSC1/TSC2 disruption, 48 restricted ability to differentiate into glial cells or neurons, 154 with Tuberous sclerosis complex (TSC), 45, 149–150 dual neuronal and glial origin, 48 epilepsy, symptom, 150 gene expression profiling, 151 mental retardation, 150 mixed cells glial/giant, 47 molecular pathophysiology of, 150–151 neurologic dysfunctions, genes down-regulation association with, 154 positive for GFAP, 48 tumorigenesis, genes up-regulation link with, 151–154 Subependymal giant cell tumor (SEGT), 48 Subependymal nodules (SENs), 149 Sugiyama, K., 174, 178–179, 182 Surgery cellular necrosis following vascular damage during, 145 treatment options for glioma, 161 treatment possibilities of brain tumours, 223 See also Neurosurgery Survivin, and cell cycle progression, 32–33 Sutton, C., 173–174 Synemin, 82–83 α- and β-synemin, expressed in astrocytoma cells, 83 characteristic, 82 contribution to astrocytoma cells, malignant behavior Boyden chamber assays, 83–84 down-regulation impact, 83–84 interaction with α-actinin, 83–84 invasion, 83 present in leading edges and ruffled membranes, 83, 85

Index

RNAi experiments, 83 siRNAs and scrape wound assays, 83–84 exhibit alternative splice variants, 82 intermediate filament (IF) protein, 82–83 positive regulator, of cell motility and proliferative capacity, 81, 83–84 prospects for astrocytoma therapy, 86–87 regulation in pathologies of CNS, 83 U-373 MG human glioblastoma cells staining, 85 See also Astrocytoma(s); Intermediate filament (IF) proteins Synthetic low-density lipoproteins (LDL), 191 T Tago, M., 248 Takei, H., 201 TAT-ODD-procaspase-3 (TOP3), 130 Temozolomide, chemotherapy for PMA, 207–208 Temporal lobe epilepsy (TLE), 267 Temsirolimus (CCI-779), 49–51 Tenascin, 110 Teratomas, 95 Tetrapyrrole, 240 Thermal dosimetry, 179 Thrombospondin-1, 138–139 Thrombospondin-2, 140 Tibbetts, K.M., 199, 201 Tihan, T., 198, 203, 206–207 Time-resolved fluorescence measurements, 162 Time-resolved laser induced fluorescence spectroscopy (TR-LIFS) advantages, 170 classification and prediction, fluorescence signal classification algorithm elements, 165–166 linear discriminant function analysis (DFA), 165 method limitation, 169 training and test phase, 165–166 TR-fluorescence characteristics, 166–168 classification results using, 168 clinical methods, 163 data analysis, 164 reduction and statistical analysis, 165 delivery catheter, 163–164 penetration depth for astrocytoma, 164

289

differentiate clearly LGG from normal tissues, 168–169 fluorescent data collection, 164 goal, 163, 170 histopathological analysis, of tumors, 164 intra-operative tool, instrumental setup, 163–164 in vivo measurements, for glioma diagnosis, see Gliomas NC measurements, 169 parameters selection, 164–165 statistical analysis, and classification, 168 time-resolved fluorescence characteristics, 166–168 TNFR associated factor (TRAF) family, 25 TNF superfamily, 23, 25, 32 TP53 mutation, 138 Tracz, R.A., 175, 178, 180 TRAIL, 24–25, 28–30 Transcriptional regulation, 149 Transfection, 136 Transmissible spongiform encephalopathies (TSEs), 13 TR-LIFS, see Time-resolved laser induced fluorescence spectroscopy (TR-LIFS) Trypsin-Giemsa banding technique, 111 TSC, see Tuberous sclerosis complex (TSC) Tsc1, 150–151, 154 TSC2-Rheb-mTOR pathway, 47 TSP-1 expression, 140 Tuberous sclerosis complex (TSC), 45–53, 149–150 Akt activation, 49 -associated lesions, 48 defined, 45 glial dysfunction and, 150 global gene expression profiling, 151 and analysis, 151 GTPase-activating protein (GAP), 46 loss of heterozygosity (LOH) in, 48 molecular pathophysiology, 150–151 role of mutations in TSC1 and TSC2, 151 mTORC1 functions, 151 mTOR inhibitors, cell cultures and animal models studies, 51–52 mTOR pathway in, 46–47 autophagy inhibition, 47 genes regulated by HIFs, 46 inactivating mutations TSC1/TSC2 gene, 46 Ras-homoloque-enriched in brain (Rheb) target, 46

290

Tuberous sclerosis complex (TSC) (cont.) SEGAs in, 51 dual neuronal and glial origin, 48 mixed cells glial/giant, 48 positive for GFAP, 48 TSC1/TSC2-mTOR signaling pathway, 151, 155–156 tumorigenesis hamatrin-tuberin complex inactivation, 49 Knudson’s two-hit model, 48 See also Subependymal giant cell astrocytomas (SEGAs) Tumorigenesis apoptosis relation with, 23 EGFRvIII role in, 112 genes up-regulated in SEGA, link to, 151, 153–154 hypomethylation role, 5 MAPK signaling to PA, 102 OTX2 gene, medulloblastoma, 115 and survivin overexpression, 32 in TSC, see Tuberous sclerosis complex (TSC) Tumorigenesis apoptosis relation with, 23 EGFRvIII role in, 112 genes up-regulated in SEGA, link to, 151, 153–154 hypomethylation role, 5 MAPK signaling to PA, 102 OTX2 gene, medulloblastoma, 115 and survivin overexpression, 32 in TSC, see Tuberous sclerosis complex (TSC) Tumours animal models for preclinical trials, 188 associated with NF1, 143 autoimmunity to, 145 of central nervous system, 203 hemangioblastoma, see Hemangioblastomas hypoxia, importance, 129–130 imaging HIF-1 active regions by 123 I-IBB and 123 IPOS method, 129, 131–133 in vivo molecular imaging, 130 pretargeting approach, 131–133 nuclear medicine imaging, 130 oxygen-dependent degradable probe, development, 130 regression, 146 slow-growing, 213 Tumstatin, 139 Turner, J.R., 66

Index

U Ubiquitin–proteasome system, 58 Ueda, M., 132 Ushio, Y., 97 V Vaporization, 177 Ventana Benchmark IHC system, 90 Videography, intraoperative ICG, 233–237 Vimentin, 81–83, 85–87 Vincristine, 207 Visual field defects, 224 von Hippel-Lindau syndrome, 233, 235 W Wang, E.M., 248 Wang, Z.X., 42 Weidner, N., 136 Wesseling, P., 139 Williams, C.M., 23 Wnt antagonists, sFRP and Dickkopf class, 39 Wnt genes, 37 Wnt signaling pathway, 35, 37–39 Axin-APC-GSK3β, 41–42 β-catenin role, 40–41 destruction complex, 38 Dishevelled (Dvl), and FRAT1 action, 40 inhibition by extracellular Wnt antagonists DKK glycoproteins family, 39 sFRPs and Dickkopf family, 39 Wnt inhibitory factor-1 (WIF-1), 39 Lef-1 and Tcf-4 factors, 42 neural stem cells-gliomas, 38 pygopus 2, 42 See also Gliomas Wong, A.J., 111 X Xenograft models, 188 Y YAG (Yttrium, Aluminum and Garnet) laser, 176 Yang, Z., 39 Yolk-sac tumors, 96 Yu, J.M., 40 Z Zhang, L.Y., 42 Zhang, Z., 40