M E T H O D S I N M O L E C U L A R M E D I C I N E TM
Pancreatic Cancer Methods and Protocols Edited by
Gloria H. Su
Pancreatic Cancer
M E T H O D S I N M O L E C U L A R M E D I C I N E™
John M. Walker, SERIES EDITOR 113. 113 Multiple Myeloma: Methods and Protocols, edited by Ross D. Brown and P. Joy Ho, 2005 112. Molecular Cardiology: Methods and Protocols, 112 edited by Zhongjie Sun, 2005 111. 111 Chemosensitivity: Volume 2, In Vivo Models, Imaging, and Molecular Regulators, edited by Rosalyn D. Blumethal, 2005 110. 110 Chemosensitivity: Volume 1, In Vitro Assays, edited by Rosalyn D. Blumethal, 2005 109. 109 Adoptive Immunotherapy, Methods and Protocols, edited by Burkhard Ludewig and Matthias W. Hoffman, 2005 108. 108 Hypertension, Methods and Protocols, edited by Jérôme P. Fennell and Andrew H. Baker, 2005 107. 107 Human Cell Culture Protocols, Second Edition, edited by Joanna Picot, 2005 106. 106 Antisense Therapeutics, Second Edition, edited by M. Ian Phillips, 2005 105. 105 Developmental Hematopoiesis: Methods and Protocols, edited by Margaret H. Baron, 2005 104. 104 Stroke Genomics: Methods and Reviews, edited by Simon J. Read and David Virley, 2004 103. 103 Pancreatic Cancer: Methods and Protocols, edited by Gloria H. Su, 2004 102 Autoimmunity: Methods and Protocols, edited 102. by Andras Perl, 2004 101 Cartilage and Osteoarthritis: Volume 2, 101. Structure and In Vivo Analysis, edited by Frédéric De Ceuninck, Massimo Sabatini, and Philippe Pastoureau, 2004 100. 100 Cartilage and Osteoarthritis: Volume 1, Cellular and Molecular Tools, edited by Massimo Sabatini, Philippe Pastoureau, and Frédéric De Ceuninck, 2004 99 Pain Research: Methods and Protocols, edited 99. by David Z. Luo, 2004 98 Tumor Necrosis Factor: Methods and Protocols, 98. edited by Angelo Corti and Pietro Ghezzi, 2004 97. 97 Molecular Diagnosis of Cancer: Methods and Protocols, Second Edition, edited by Joseph E. Roulston and John M. S. Bartlett, 2004
96. 96 Hepatitis B and D Protocols: Volume 2, Immunology, Model Systems, and Clinical Studies, edited by Robert K. Hamatake and Johnson Y. N. Lau, 2004 95. 95 Hepatitis B and D Protocols: Volume 1, Detection, Genotypes, and Characterization, edited by Robert K. Hamatake and Johnson Y. N. Lau, 2004 94 Molecular Diagnosis of Infectious Diseases, 94. Second Edition, edited by Jochen Decker and Udo Reischl, 2004 93. 93 Anticoagulants, Antiplatelets, and Thrombolytics, edited by Shaker A. Mousa, 2004 92. 92 Molecular Diagnosis of Genetic Diseases, Second Edition, edited by Rob Elles and Roger Mountford, 2004 91. 91 Pediatric Hematology: Methods and Protocols, edited by Nicholas J. Goulden and Colin G. Steward, 2003 90. 90 Suicide Gene Therapy: Methods and Reviews, edited by Caroline J. Springer, 2004 89. The Blood–Brain Barrier: Biology and 89 Research Protocols, edited by Sukriti Nag, 2003 88. 88 Cancer Cell Culture: Methods and Protocols, edited by Simon P. Langdon, 2003 87. 87 Vaccine Protocols, Second Edition, edited by Andrew Robinson, Michael J. Hudson, and Martin P. Cranage, 2003 86. 86 Renal Disease: Techniques and Protocols, edited by Michael S. Goligorsky, 2003 85. 85 Novel Anticancer Drug Protocols, edited by John K. Buolamwini and Alex A. Adjei, 2003 84. 84 Opioid Research: Methods and Protocols, edited by Zhizhong Z. Pan, 2003 83. 83 Diabetes Mellitus: Methods and Protocols, edited by Sabire Özcan, 2003 82 Hemoglobin Disorders: Molecular Methods 82. and Protocols, edited by Ronald L. Nagel, 2003 81. 81 Prostate Cancer Methods and Protocols, edited by Pamela J. Russell, Paul Jackson, and Elizabeth A. Kingsley, 2003
M E T H O D S I N M O L E C U L A R M E D I C I N E™
Pancreatic Cancer Methods and Protocols
Edited by
Gloria H. Su Departments of Otolaryngology/Head and Neck Surgery and Pathology Columbia University, New York, NY
Humana Press
Totowa, New Jersey
© 2005 Humana Press Inc. 999 Riverview Drive, Suite 208 Totowa, New Jersey 07512 www.humanapress.com All rights reserved. No part of this book 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. Methods in Molecular Medicine™ is a trademark of The Humana Press Inc. The content and opinions expressed in this book are the sole work of the authors and editors, who have warranted due diligence in the creation and issuance of their work. The publisher, editors, and authors are not responsible for errors or omissions or for any consequences arising from the information or opinions presented in this book and make no warranty, express or implied, with respect to its contents. This publication is printed on acid-free paper. ∞ ANSI Z39.48-1984 (American Standards Institute) Permanence of Paper for Printed Library Materials. Cover illustration: Foreground: Figure 1D, Chapter 18, Zebrafish as a Model for Pancreatic Cancer Research, N. S. Yee and M. Pack. Background: Figure 1D, Chapter 1, Identification and Analysis of Precursors to Invasive Pancreatic Cancer, R. H. Hruban, R. E. Wilentz, and A. Maitra. Cover design by Patricia F. Cleary. For additional copies, pricing for bulk purchases, and/or information about other Humana titles, contact Humana at the above address or at any of the following numbers: Tel.: 973-256-1699; Fax: 973-256-8341; Email:
[email protected]; or visit our Website: www.humanapress.com Photocopy Authorization Policy: Authorization to photocopy items for internal or personal use, or the internal or personal use of specific clients, is granted by Humana Press Inc., provided that the base fee of US $25.00 per copy is paid directly to the Copyright Clearance Center at 222 Rosewood Drive, Danvers, MA 01923. For those organizations that have been granted a photocopy license from the CCC, a separate system of payment has been arranged and is acceptable to Humana Press Inc. The fee code for users of the Transactional Reporting Service is: [1-58829107-3/05 $25.00]. Printed in the United States of America. 10 9 8 7 6 5 4 3 2 1 Library of Congress Cataloging in Publication Data Pancreatic cancer : methods and protocols / edited by Gloria H. Su. p. ; cm. — (Methods in molecular medicine ; 103) Includes bibliographical references and index. ISBN 1-58829-107-3 (alk. paper) e-ISBN 1-59259-780-7 1. Pancreas—Cancer—Research—Methodology. 2. Pancreas—Cancer—Molecular aspects. [DNLM: 1. Pancreatic Neoplasms—genetics. 2. Genetic Techniques. 3. Pancreatic Neoplasms—therapy. WI 810 P18825 2005] I. Su, Gloria H. II. Series. RC280.P25P3565 2005 616.99’437—dc22 2004002436
Preface Pancreatic ductal adenocarcinoma is the fourth leading cause of cancer death in the United States. Annually approximately 30,000 Americans are diagnosed with the disease and most will die from it within five years. Pancreatic ductal adenocarcinoma is unique because of its late onset in age, high mortality, small tumor samples infiltrated with normal cells, and a lack of both early detection and effective therapies. Some of these characteristics have made studying this disease a challenge. Pancreatic cancer develops as a result of the accumulation of genetic alterations in cancer-causing genes, such as the oncogenes and the tumor-suppressor genes. In the last decade, major progress has been made in identifying important oncogenes and tumor-suppressor genes for the disease. In Pancreatic Cancer: Methods and Protocols, we review the classical techniques that have contributed to the advances in pancreatic research and introduce new strategies that we hope will add to future breakthroughs in the field of cancer biology. Pancreatic Cancer: Methods and Protocols provides a broad range of protocols for molecular, cellular, pathological, and statistical analyses of sporadic and familial pancreatic cancer. It covers topics from in vitro cell cultures to in vivo mouse models, DNA to protein manipulation, and mutation analyses to treatment development. We believe that our book will prove an invaluable source of proven protocols for those who are interested in either basic or translational research in pancreatic cancer. Gloria H. Su
v
Contents Preface .............................................................................................................. v Contributors ..................................................................................................... ix 1 Identification and Analysis of Precursors to Invasive Pancreatic Cancer Ralph H. Hruban, Robb E. Wilentz, and Anirban Maitra ..................... 1 2 Optimal Molecular Profiling of Tissue and Tissue Components: Defining the Best Processing and Microdissection Methods for Biomedical Applications G. Steven Bova, Isam A. Eltoum, John A. Kiernan, Gene P. Siegal, Andra R. Frost, Carolyn J. M. Best, John W. Gillespie, and Michael R. Emmert-Buck .......................... 15 3 Immunohistochemistry and In Situ Hybridization in Pancreatic Neoplasia Robb E. Wilentz, Ayman Rahman, Pedram Argani, and Christine Iacobuzio-Donahue ................................................. 67 4 Practical Methods for Tissue Microarray Construction Helen L. Fedor and Angelo M. De Marzo ........................................... 89 5 Xenografting and Harvesting Human Ductal Pancreatic Adenocarcinomas for DNA Analysis Kimberly Walter, James Eshleman, and Michael Goggins ................ 103 6 Culture and Immortalization of Pancreatic Ductal Epithelial Cells Terence Lawson, Michel Ouellette, Carol Kolar, and Michael Hollingsworth .......................................................... 113 7 DNA Methylation Analysis in Human Cancer Carmelle D. Curtis and Michael Goggins ......................................... 123 8 Digital Single-Nucleotide Polymorphism Analysis for Allelic Imbalance Hsueh-Wei Chang and Ie-Ming Shih ................................................ 137 9 Representational Difference Analysis as a Tool in the Search for New Tumor Suppressor Genes Antoinette Hollestelle and Mieke Schutte ........................................ 143
vii
viii
Contents
10 Serial Analyses of Gene Expression (SAGE) Jia Le Dai ........................................................................................... 161 11 Oligonucleotide-Directed Microarray Gene Profiling of Pancreatic Adenocarcinoma David E. Misek, Rork Kuick, Samir M. Hanash, and Craig D. Logsdon ................................................................... 175 12 Identification of Differentially Expressed Proteins in Pancreatic Cancer Using a Global Proteomic Approach Christophe Rosty and Michael Goggins ............................................ 189 13 Detection of Telomerase Activity in Patients with Pancreatic Cancer Kazuhiro Mizumoto and Masao Tanaka ........................................... 199 14 Serological Analysis of Expression cDNA Libraries (SEREX): An Immunoscreening Technique for Identifying Immunogenic Tumor Antigens Yao-Tseng Chen, Ali O. Gure, and Matthew J. Scanlan ................... 207 15 Modeling Pancreatic Cancer in Animals to Address Specific Hypotheses Paul J. Grippo and Eric P. Sandgren ................................................. 217 16 Strategies for the Use of Site-Specific Recombinases in Genome Engineering Julie R. Jones, Kathy D. Shelton, and Mark A. Magnuson ................ 245 17 Primary Explant Cultures of Adult and Embryonic Pancreas Farzad Esni, Yoshiharu Miyamoto, Steven D. Leach, and Bidyut Ghosh ......................................................................... 259 18 Zebrafish as a Model for Pancreatic Cancer Research Nelson S. Yee and Michael Pack ...................................................... 273 19 Development of a Cytokine-Modified Allogeneic Whole Cell Pancreatic Cancer Vaccine Dan Laheru, Barbara Biedrzycki, Amy M. Thomas, and Elizabeth M. Jaffee ................................................................. 299 20 Overview of Linkage Analysis: Application to Pancreatic Cancer Alison P. Klein .................................................................................. 329 Index ............................................................................................................ 343
Contributors PEDRAM ARGANI • Department of Pathology, The Johns Hopkins University School of Medicine, Baltimore, MD CAROLYN J. M. BEST • Pathogenetics Unit, National Cancer Institute, National Institutes of Health, Bethesda, MD BARBARA BIEDRZYCKI • The Sidney Kimmel Cancer Center, The Johns Hopkins University School of Medicine, Baltimore, MD G. STEVEN BOVA • Departments of P of Genetic Medicine, The Johns Hopkins Medical Institutions, Baltimore, MD HSUEH-WEI CHANG • Department of Pathology, The Johns Hopkins University School of Medicine, Baltimore, MD YAO-TSENG CHEN • Weill Medical College of Cornell University and Ludwig Institute for Cancer Research, New York Branch, New York, NY CARMELLE D. CURTIS • Department of Pathology, The Johns Hopkins University School of Medicine, Baltimore, MD JIA LE DAI • Department of Molecular Pathology, M. D. Anderson Cancer Center, University of Texas, Houston, TX ANGELO M. DE MARZO • Department of Pathology, The Brady Urological Institute, and The Sidney Kimmel Cancer Center, The Johns Hopkins University School of Medicine, Baltimore, MD ISAM A. ELTOUM • Departments of Pathology, Cell Biology, and Surgery, and the UAB Comprehensive Cancer Center, University of Alabama at MICHAEL R. EMMERT-BUCK • Pathogenetics Unit, National Cancer Institute, National Institutes of Health, Bethesda, MD JAMES ESHLEMAN • Department of Pathology and The Oncology Center, The Johns Hopkins University School of Medicine, Baltimore, MD FARZAD ESNI • Department of Surgery, The Johns Hopkins University School of Medicine, Baltimore, MD HELEN L. FEDOR • Department of Pathology, The Johns Hopkins University School of Medicine, Baltimore, MD ANDRA R. FROST • Departments of Pathology, Cell Biology, and Surgery, and the UAB Comprehensive Cancer Center, University of Alabama at Birmingham, Birmingham, AL BIDYUT GHOSH • Department of Surgery, The Johns Hopkins University School of Medicine, Baltimore, MD JOHN W. GILLESPIE • Science Applications International Corporation, National Cancer Institute, Bethesda, MD MICHAEL GOGGINS • Departments of Pathology, Medicine, and The Oncology Center, The Johns Hopkins University School of Medicine, Baltimore, MD
x
Contributors
PAUL J. GRIPPO • Department of Surgery, Northwestern University Medical School, Chicago, IL ALI O. GURE • Ludwig Institute for Cancer Research, New York Branch, New York, NY SAMIR M. HANASH • Department of Pediatric Oncology, University of Michigan, Ann Arbor, MI ANTOINETTE HOLLESTELLE • Department of Medical Oncology, Josephine Nefkens Institute, Erasmus MC, Rotterdam, The Netherlands MICHAEL A. HOLLINGSWORTH • Eppley Institute for Research in Cancer and Allied Diseases, University of Nebraska Medical Center, Omaha, NE RALPH H. HRUBAN • Department of Pathology and The Oncology Center, The Johns Hopkins University School of Medicine, Baltimore, MD CHRISTINE IACOBUZIO-DONAHUE • Department of Pathology, The Johns Hopkins University School of Medicine, Baltimore, MD ELIZABETH M. JAFFEE • The Sidney Kimmel Cancer Center, The Johns Hopkins University School of Medicine, Baltimore, MD JULIE R. JONES • Department of Molecular Physiology and Biophysics, Vanderbilt University School of Medicine, Nashville, TN JOHN A. KIERNAN • Department of Anatomy and Cell Biology, University of Western Ontario, London, Canada ALISON P. KLEIN • Statistical Genetics Section, National Human Genome Research Institute, National Institutes of Health CAROL KOLAR • Eppley Institute for Research in Cancer and Allied Diseases, University of Nebraska Medical Center, Omaha, NE RORK KUICK • Department of Pediatric Oncology, University of Michigan, Ann Arbor, MI DAN LAHERU • The Sidney Kimmel Cancer Center, The Johns Hopkins University School of Medicine, Baltimore, MD TERENCE LAWSON • Department of Pharmaceutical Sciences, Eppley Institute for Research in Cancer and Allied Diseases, University of Nebraska Medical Center, Omaha, NE STEVEN D. LEACH • Department of Surgery and The Oncology Center, The Johns Hopkins University School of Medicine, Baltimore, MD CRAIG D. LOGSDON • Department of Physiology, University of Michigan, Ann Arbor, MI MARK A. MAGNUSON • Department of Molecular Physiology and Biophysics, Vanderbilt University School of Medicine, Nashville, TN ANIRBAN MAITRA • Department of Pathology, The Oncology Center, and Institute of Genetic Medicine, The Johns Hopkins University School of Medicine, Baltimore, MD DAVID E. MISEK • Department of Pediatric Oncology, University of Michigan, Ann Arbor, MI
Contributors
xi
YOSHIHARU MIYAMOTO • Department of Surgery, The Johns Hopkins University School of Medicine, Baltimore, MD KAZUHIRO MIZUMOTO • Department of Surgery and Oncology, Graduate School of Medical Sciences, Kyushu University, Fukuoka, Japan MICHEL M. OUELLETTE • Eppley Institute for Research in Cancer and Allied Diseases, University of Nebraska Medical Center, Omaha, NE MICHAEL PACK • Departments of Medicine and Cell and Developmental Biology, University of Pennsylvania School of Medicine, Philadelphia, PA AYMAN R AHMAN • Department of Pathology, The Johns Hopkins University School of Medicine, Baltimore, MD CHRISTOPHE ROSTY • Department de Pathologie, Institut Curie, Paris, France ERIC P. SANDGREN • Department of Pathobiological Sciences, School of Veterinary Medicine, University of Wisconsin–Madison, Madison, WI MATTHEW J. SCANLAN • Ludwig Institute for Cancer Research, New York Branch, New York, NY MIEKE SCHUTTE • Department of Medical Oncology, Josephine Nefkens Institute, Erasmus MC, Rotterdam, The Netherlands KATHY D. SHELTON • Department of Molecular Physiology and Biophysics, Vanderbilt University School of Medicine, Nashville, TN IE -MING SHIH • Departments of Pathology, Gynecology and Obstetrics, and The Oncology Center, The Johns Hopkins University School of Medicine, Baltimore, MD GENE P. SIEGAL • Departments of Pathology, Cell Biology, and Surgery, and the UAB Comprehensive Cancer Center, University of Alabama at Birmingham, Birmingham, AL MASAO TANAKA • Department of Surgery and Oncology, Graduate School of Medical Sciences, Kyushu University, Fukuoka, Japan AMY M. THOMAS • The Sidney Kimmel Cancer Center, The Johns Hopkins University School of Medicine, Baltimore, MD KIMBERLY WALTER • Department of Pathology, The Johns Hopkins University School of Medicine, Baltimore, MD ROBB E. WILENTZ • Department of Pathology, The Johns Hopkins University School of Medicine, Baltimore, MD NELSON S. Y EE • Department of Medicine, University of Pennsylvania School of Medicine, Philadelphia, PA
Precursors to Invasive Pancreatic Cancer
1
1 Identification and Analysis of Precursors to Invasive Pancreatic Cancer Ralph H. Hruban, Robb E. Wilentz, and Anirban Maitra Summary Histologically distinct noninvasive precursor lesions have been recognized in the pancreas for close to a century. The recent development of a consistent reproducible nomenclature and classification system for these lesions has been a major advance in the study of these noninvasive precursors. The “pancreatic intraepithelial neoplasia” or PanIN system was developed at a National Cancer Institutes sponsored think tank in Park City, Utah. Numerous studies have now demonstrated that genetic alterations in cancer-associated genes are more common in higher grade PanIN lesions then they are in lower grade PanIN lesions, and that higher grade PanIN lesions have many of the same genetic alterations that are found in invasive ductal adenocarcinomas of the pancreas. Thus, just as there is a progression in the colorectal of adenomas to invasive adenocarcinoma, so too is there a progression in the pancreas of histologically lowgrade PanIN, to high-grade PanIN to invasive ductal adenocarcinoma. Key Words: Precursor; panIN; neoplasia; intraepithelial; in situ.
1. Introduction It has been estimated that this year approx 30,000 Americans will be diagnosed with pancreatic cancer and 30,000 will die of it (1). Pancreatic cancer is one of the deadliest forms of cancer for two reasons. First, most patients do not come to clinical attention until after the cancer has spread to other organs (2,3). Second, the vast majority of pancreatic cancers do not respond to existing chemo- and radiation therapies (2). Early detection offers one of the best opportunities to reduce the human suffering caused by pancreatic cancer (4,5). The first step in developing a new screening test for pancreatic cancer is to improve our understanding of early pancreatic cancer and its precursors. It is hoped that a better understanding of the precursor lesions in the pancreas will lead to new ways to diagnose and From: Methods in Molecular Medicine, Vol. 103: Pancreatic Cancer: Methods and Protocols Edited by: G. Su © Humana Press Inc., Totowa, NJ
1
2
Hruban et al.
treat pancreatic cancer before it spreads to other organs (5). This chapter reviews our current understanding of the three most common precursor lesions in the pancreas and the methods for studying these lesions. 2. Materials 2.1. Precursor Lesions in the Pancreas Three histologically well-defined precursors to invasive adenocarcinoma of the pancreas have been identified. These include “pancreatic intraepithelial neoplasia,” “intraductal papillary mucinous neoplasms,” and “mucinous cystic neoplasms.” 2.2. Pancreatic Intraepithelial Neoplasia A growing body of evidence suggests that histologically well-defined lesions in the small ducts and ductules in the pancreas are the precursors to infiltrating ductal adenocarcinomas of the pancreas (6–10). For years, these lesions were known by a variety of different names including “hyperplasia,” “dysplasia,” “duct lesions,” “metaplasia,” and “carcinoma in situ,” and for decades there have been no uniform standards for classifying the lesions seen (11,12). The lack of a uniform nomenclature and standards to classify these lesions made it virtually impossible to compare one study to another, and it greatly impeded our understanding of precursor lesions in the pancreas. An international group of pathologists was therefore assembled at a Pancreatic Cancer Think Tank held in Park City, Utah in September 1999. Based on our current understanding of the genetic alterations present in these duct lesions, it was the consensus that the lesions represented early neoplasms. The nomenclature “Pancreatic Intraepithelial Neoplasia” (PanIN) was therefore adopted and uniform criteria were established for the grading of PanINs (see http://pathology.jhu.edu/pancreas_panin) (11). The criteria for the grading are reviewed in Subheading 3.1. The international acceptance of this new nomenclature and classification system will greatly facilitate the study of these important precursors to infiltrating ductal adenocarcinomas of the pancreas. 2.3. Intraductal Papillary Mucinous Neoplasm Intraductal papillary mucinous neoplasms (IPMNs) of the pancreas were first described in the 1980s by Oshashi. IPMNs are being recognized with greater frequency in the United States (13), and it is clear that they too can be a precursor to infiltrating carcinoma (14). The classification system for IPMNs is presented in detail in Subheading 3.2. In brief, IPMNs are large papillary tumors (they have fingerlike projections) that involve the main pancreatic ducts and that produce excess amounts of mucin (15–18). Because of this excess mucin,
Precursors to Invasive Pancreatic Cancer
3
IPMNs frequently distend the pancreatic ducts, and patients with this tumor are often found to have mucin oozing from the ampulla of Vater on endoscopy (15–18). IPMNs are distinguished from PanINs by their larger size. IPMNs are grossly and/or radiographically visible, whereas PanINs are microscopic lesions. Because of their large size, IPMNs are easier to study than are PanINs, and they have therefore served as a useful model tumor to study precursor lesions and the progression to infiltrating carcinoma in the pancreas (19). 2.4. Mucinous Cystic Neoplasm Mucinous cystic neoplasms (MCNs) are rare tumors of the pancreas that arise primarily in women (20,21). Like IPMNs, MCNs produce abundant mucin. Unlike IPMNs, mucinous cystic neoplasms do not involve the pancreatic duct system (20). In addition to the mucin-producing neoplastic epithelial cells, MCNs have a characteristic “ovarian” stroma (20) that is not seen in PanINs or IPMNs. MCNs can progress over time into an invasive pancreatic cancer (21). A detailed description of the classification of MCNs is provided in Subheading 3.3. (20). 3. Methods 3.1. Classification of PanINs The current system for classification of PanINs is based on a number of studies that have correlated microscopic findings with genetic alterations (6,7, 9,10,22). These studies have established that the small proliferative lesions in the pancreatic ducts are neoplasms—that is, the lesions harbor clonal mutations in cancer-associated genes. In addition, they have demonstrated progression of mutational events, such that few genetic alterations are found in PanINs without cytologic or architectural atypia, while the genetic alterations in the histologically higher grades of PanIN approach those found in infiltrating ductal adenocarcinomas (6,7,9,10,22). Based on these studies, the histological classification system for PanINs shown in Table 1 has been established (11). Examples of each grade of PanIN are available on the Web (http://pathology.jhu.edu/ pancreas_panin) and are shown in Fig. 1. This classification system, which implies a progression from normal duct epithelium, to PanIN-1, to PanIN-2, to PanIN-3, to invasive ductal adenocarcinoma, is supported by clinical cases in which patients with PanIN-3 later develop an infiltrating ductal adenocarcinoma (14,23,24) (see Note 1). 3.2. Classification of IPMNs IPMNs are histologically classified into four groups (Table 2) in the World Health Organization (WHO) classification scheme (25). IPMN-adenoma is the
4
Hruban et al.
Table 1 Proposed Pancreatic Intraepithelial Neoplasia Nomenclature a Normal: The normal ductal epithelium is a cuboidal epithelium without significant atypia. PanIN-1A (pancreatic intraepithelial neoplasia 1A): These are flat epithelial lesions composed of tall columnar cells with basally located nuclei and abundant supranuclear mucin. The nuclei are small and round to oval in shape. PanIN-1B (pancreatic intraepithelial neoplasia 1B): These epithelial lesions have a papillary, micropapillary, or basally pseudostratified architecture but are otherwise identical to PanIN-1A. PanIN-2 (pancreatic intraepithelial neoplasia 2): Architecturally these mucinous epithelial lesions may be flat but are mostly papillary. Cytologically, these lesions have moderate atypia. This atypia may include some loss of polarity, nuclear crowding, enlarged nuclei, pseudostratification, and hyperchromatism. These nuclear abnormalities fall short of those seen in PanIN-3. Mitoses are rare, but when present are nonluminal (not apical) and are not atypical. True cribriform structures with luminal necrosis and marked cytologic abnormalities are generally not seen and, when present, should suggest the diagnosis of PanIN-3. PanIN-3 (pancreatic intraepithelial neoplasia 3): Architecturally, these lesions are usually papillary or micropapillary. True cribriforming, the appearance of “budding off” of small clusters of epithelial cells into the lumen, and luminal necrosis should all suggest the diagnosis of PanIN-3. Cytologically, these lesions are characterized by a loss of nuclear polarity, dystrophic goblet cells (goblet cells with nuclei oriented toward the lumen and mucinous cytoplasm oriented toward the basement membrane), mitoses that may occasionally be abnormal, nuclear irregularities, and prominent (macro) nucleoli. The lesions resemble carcinoma at the cytonuclear level, but invasion through the basement membrane is absent. a Based
on (11).
lowest grade of IPMN and, like PanIN-1, IPMN-adenomas lack significant cytologic and architectural atypia (Fig. 2A). IPMN-borderline are noninvasive IPMNs with moderate nuclear and cytologic atypia. IPMN-carcinoma in situ are noninvasive IPMNs with significant architectural and cytologic atypia (Fig. 2B). Finally, IPMN invasive carcinomas are those IPMNs with an associated invasive carcinoma. IPMNs are associated with an invasive carcinoma in approximately 35% of the cases and this invasive component usually has a ductal (tubular) or colloid (mucinous) microscopic appearance (14). In the most recent WHO classification scheme, IPMNs with an invasive carcinoma are collectively designated as “papillary mucinous carcinomas” regardless of the histologic type of the invasive cancer (25) (see Note 2).
Precursors to Invasive Pancreatic Cancer
5
Fig. 1. Pancreatic intraepithelial neoplasia (PanIN). Increasing nuclear and cytologic atypia is seen as one progresses from PanIN-1A (A), to PanIN-1B (B), to PanIN-2 (C), to PanIN-3 (D). The lesion in panel (C) shows more atypia than the usual PanIN-2 and is approaching the level of PanIN-3. Table 2 Classification of Intraductal Papillary Mucinous Neoplasms IPMN-adenoma: These lesions have only minimal architectural and cytologic dysplasia. The papillae have well-defined fibrovascular cores and the epithelial cells are oriented perpendicular to the papillae. The epithelial cells contain abundant mucin, the nuclei are small and uniform and nucleoli are not prominent. Mitoses are absent. IPMN-borderline: These noninvasive neoplasms show moderate dysplasia. The papillae are not as well defined as they are for IPMN-adenoma, and the nuclei show moderate nuclear pleomorphism and hyperchromasia. Occasional nuclei may contain a conspicuous nucleolus and mitoses can be seen. IPMN-carcinoma in situ: These noninvasive neoplasms, also known as “intraductal carcinomas,” show significant nuclear dysplasia. True cribriforming of the papillae can be seen, as can focal necrosis. Nuclei are irregularly shaped, nucleoli are prominent and mitoses are frequent. IPMN-invasive carcinoma: These lesions are defined by the presence of an invasive carcinoma arising in association with an IPMN. The invasive carcinoma is usually either a tubular (ductal) or colloid carcinoma.
3.3. Classification of Mucinous Cystic Neoplasms As noted previously, MCNs can be distinguished from IPMNs because MCNs have a characteristic “ovarian” stroma and because, unlike IPMNs, MCNs do
6
Hruban et al.
Fig. 2. Intraductal papillary mucinous neoplasms (IPMNs). An IPMN-borderline showing moderate nuclear and architectural atypia is shown in (A), while an IPMNcarcinoma in situ showing significant dysplasia is illustrated in (B).
not involve the main pancreatic ducts. The classification system for MCNs and IPMNs is, however, similar (Table 3). MCN-adenomas contain a single layer of mucinous epithelium that lacks significant cytological and architectural atypia (Fig. 3A) (20,26). In borderline mucinous cystic neoplasms, the epithelium may form papillae, but only moderate cytologic and nuclear atypia are seen (20, 26). In MCN-carcinoma in situ, no invasive carcinoma is identified; however, the epithelium does show significant atypia including a high mitotic activity, cribriform or bridging structures, and marked nuclear pleomorphism (Fig. 3B) (20,26). MCN-invasive carcinoma should be diagnosed when tissue-invasive carcinoma arises in association with a MCN. Because a single MCN can show
Precursors to Invasive Pancreatic Cancer
7
Table 3 Classification of Mucinous Cystic Neoplasms MCN-adenoma: These noninvasive neoplasms are characterized by the formation of cystic spaces lined by a single row of columnar mucin-producing cells with uniform small nuclei. Nucleoli are inconspicuous and mitoses are absent. MCN-borderline: These non-invasive neoplasms show moderate dysplasia with nuclear crowding and pleomorphism. Mitoses may be seen as well as small nucleoli. MCN-in situ carcinoma: These non-invasive neoplasms show significant nuclear and architectural dysplasia. Architecturally, the papillae lack fibrovascular cores, and cribriforming and necrosis may be seen. Cytologically, significant nuclear pleomorphism, mitoses and prominent mucleoli are noted. MCN-invasive carcinoma: These are tissue invasive adenocarcinomas arising in association with a MCN. The invasive component usually resembles an invasive ductal carcinoma.
a range of architectural and cytological atypia, from adenoma to invasive carcinoma (21), MCNs need to be entirely histologically examined before they can be classified definitively (20,21) (see Note 3). 3.4. Microdissection of Precursors Because of their important role in the development of invasive neoplasms, much effort has been directed to the study of precursor lesions in the pancreas (see Note 4). The isolation of pure populations of neoplastic precursor cells not contaminated by adjacent normal non-neoplastic cells is an essential first step in the molecular analysis of these precursors. A variety of techniques have been described, including the use of a fine needle controlled by a hydraulic micromanipulator (27), laser capture microdissection (LCM) (28), and “epithelial aggregate separation and isolation” (EASI) (29). 3.4.1. Fine-Needle Microdissection C. A. Moskaluk and S. E. Kern described a simple technique for microdissecting precursor lesions that reliably produces polymerase chain reaction (PCR)amplifiable DNA from lesional tissue less than 0.1 mm in diameter (27). 1. Sections 7-µm thick of formalin-fixed, paraffin-embedded tissue sections are placed on glass slides. 2. From each tissue block, an additional 4-µm section, immediately adjacent to the previous section, is prepared to serve as a scout slide. This slide is stained with hemotoxylin and eosin. 3. The 7-µm sections are deparaffinized, stained with hemotoxylin and eosin, and incubated for 2 minutes in a 2.5% glycerol solution.
8
Hruban et al.
Fig. 3. Mucinous cystic neoplasms (MCNs). MCN-adenoma with only minimal atypia (A) contrasts with a MCN-carcinoma in situ (B) which shows significant nuclear and architectural atypia. Note the ovarian stroma.
4. The slides are air-dried and microdissected using an inverted microscope and hydraulic micromanipulator arm. 5. DNA is extracted in a series of buffers.
This technique will yield PCR products for 50 cells in the 150-basepair (bp) range in most cases, but approx 50–70% of the 50 cell samples will fail to yield PCR products larger than 400 bp (27). 3.4.2. Laser Capture Microdissection Laser capture microdissection (LCM) is the technique most commonly utilized to obtain relative pure populations of precursor cells (28):
Precursors to Invasive Pancreatic Cancer
9
1. Formalin-fixed paraffin-embedded or fresh-frozen tissue sections are placed on an untreated glass slide and stained with hemotoxylin and eosin. 2. A plastic cap coated with ethylene vinyl acetate transfer film is placed over the slides and the slide is placed in the LCM apparatus. 3. In the microscope (Pix Cell II LCM Microscope, Arcturus Engineering, Mountain View, CA), the operator identifies and selects the lesion of interest and then activates a laser within the microscope optics. 4. At the precise location selected, the film is melted and the cell sample selected is bonded onto the pleotic cap. The rest of the tissue is left behind. 5. The cap with the adherent tissue is then placed on an Eppendorf tube for nucleic acid (DNA, RNA) extraction and subsequent PCR.
Additional details and methods are available on the NIH LCM website (http:// dir.nichd.nih.gov/lcm/lcm.htm). 3.4.3. Epithelial Aggregate Separation and Isolation Maitra et al. have described an enhancement to LCM that can be used to enrich samples for neoplastic cells (29). This method, called “epithelial aggregate separation and isolation,” or EASI, is applicable to fresh tissues only. 1. The tissue is sectioned and gently scraped with the edge of a plain, uncharged, microscope glass slide. 2. The material adherent to this slide is then spread evenly onto the surface of a second uncharged slide. 3. Slides are immediately fixed in 95% methanol for 2 minutes and stained with hematoxylin and eosin. 4. Epithelial aggregates on these slides can then be microdissected using an LCM. Alternatively, manual methods can also be used.
The advantages of this technique are that the discreteness of the epithelial clusters helps reduce background inflammatory and stromal elements and that large areas can be sampled (28). 3.5. Immunohistochemical Labeling of Precursor Lesions Immunohistochemical labeling can also be used to examine precursor lesions in the pancreas. Immunohistochemical labeling has the advantage that tissue morphology is preserved. For example, Wilentz et al. have shown that immunohistochemical (IHC) labeling for the DPC4 (SMAD4) gene product accurately reflects DPC4/SMAD4 gene status (30) and IHC labeling has been used to evaluate PanINs, IPMNs, and MCNs (10,31,32). In brief, IHC entails: 1. Unstained 5-µm sections of the tissues are cut and placed on treated slides. 2. Sections are deparaffinized by routine techniques, treated with sodium citrate buffer (diluted to 1X from 10X heat-induced epitope retrieval buffer, Ventanta-Bio Tek Solutions, Tuscon, AZ), and then steamed for 20 minutes at 80°C.
10
Hruban et al.
3. After cooling, slides are labeled with a 1:1000 dilution of the monoclonal antibody to Dpc4/Smad4 (Clone B8, Santa Cruz). 4. The anti-Dpc4 antibodies can then be detected by adding biotinylated secondary antibodies and 3,3'-diaminobenzidine.
By substituting the primary antibody, this protocol can also be used to detect other antigens. 3.6. Tissue Array Analysis One important recent advance in immunohistochemical technique is the development of high-throughput tissue arrays (33–35). Tissue arrays can also be used to study precursor lesions in the pancreas. A tissue array consists of multiple tissue samples embedded in rows and columns in one paraffin block. Slides can be routinely cut from this arrayed paraffin block. Thus, instead of studying multiple slides, each of which contains one sample, one can perform experiments on a few slides, each of which contains multiple samples. In brief, tissue array experimentation consists of the following three steps: 1. Creation of the tissue array block: The tissue arrayer removes a focus of tissue from a donor paraffin block and transfers it to a specific coordinate on the array block. This is done multiple times to create a tissue array block. 2. Production and study of slides from the tissue array block: Up to 300 3- to 6-µm slides can be cut from the tissue array block. Verification of the appropriate tissue within the block is made by examination of a hematoxylin and eosin–stained slide. Immunohistochemical studies can then be performed. 3. Interpretation of data derived from the slides: Data can be interpreted manually at a traditional microscope or electronically with the aid of a computerized database system.
4. Notes 1. The new PanIN nomenclature and classification system should be used whenever studies of small duct lesions in the pancreas are reported. 2. IPMNs are being recognized with increasing frequency in the United States. Because of their large size, these neoplasms are a good model system with which to study the progression from a noninvasive precursor to an invasive cancer of the pancreas. 3. Because high-grade dysplasia and even invasive carcinoma can arise focally in MCNs, MCNs should be examined in their entirety at the light microscopic level. 4. A variety of microdissection techniques are available for enrichment of neoplastic cells from a heterogeneous background for molecular analyses.
Acknowledgments We thank Sandy Markowitz for her hard work and dedication in preparing this manuscript.
Precursors to Invasive Pancreatic Cancer
11
References 1. Greenlee, R. T., Hill-Harmon, M. B., Murray, T., and Thun, M. (2001) Cancer Statistics, 2001. CA Cancer J. Clin. 51, 16–36. 2. 2 Warshaw, A. L. and Castillo, C. F. D. (1992) Pancreatic carcinoma. N. Engl. J. Med. 326, 455–465. 3. 3 Niederhuber, J. E., Brennan, M. F., and Menck, H. R. (1995) The national cancer data base report on pancreatic cancer. Cancer 76, 1671–1677. 4. 4 Goggins, M., Canto, M., and Hruban, R. H. (2000) Can we screen high-risk individuals to detect early pancreatic carcinoma? J. Surg. Oncol. 74, 243–248. 5. 5 Hruban, R. H., Canto, M. I., and Yeo, C. J. (2001) Prevention of pancreatic cancer and strategies for management of familial pancreatic cancer. Dig. Dis. 19, 76–84. 6. 6 Hruban, R. H., Wilentz, R. E., and Kern, S. E. (2000) Genetic progression in the pancreatic ducts. Am. J. Pathol. 156, 1821–1825. 7. 7 Hruban, R. H., Wilentz, R. E., Goggins, M., Offerhaus, G. J. A., Yeo, C. J., and Kern, S. E. (1999) Pathology of incipient pancreatic cancer. Ann. Oncol. 10, S9–S11. 8. 8 McCarthy, D. M., Brat, D. J., Wilentz, R. E., et al. (2001) Pancreatic intraepithelial neoplasia and infiltrating adenocarcinoma: Analysis of progession and recurrence by DPC4 immunohistochemical labeling. Hum. Pathol. 32, 638–642. 9. 9 Wilentz, R. E., Geradts, J., Maynard, R., et al. (1998) Inactivation of the p16 (INK4A) tumor-suppressor gene in pancreatic duct lesions: Loss of intranuclear expression. Cancer Res. 58, 4740–4754. 10. 10 Wilentz, R. E., Iacobuzio-Donahue, C. A., Argani, P., et al. (2000) Loss of expression of Dpc4 in pancreatic intraepithelial neoplasia: Evidence that DPC4 inactivation occurs late in neoplastic progression. Cancer Res. 60, 2002–2006. 11. 11 Hruban, R. H., Adsay, N. V., Albores-Saavedra, J., et al. (2001) Pancreatic intraepithelial neoplasia (PanIN): A new nomenclature and classification system for pancreatic duct lesions. Am. J. Surg. Pathol. 25, 579–586. 12. 12 Cubilla, A. L. and Fitzgerald, P. J. (1976) Morphological lesions associated with human primary invasive nonendocrine pancreas cancer. Cancer Res. 36, 2690–2698. 13. 13 Sohn, T. A., Yeo, C. J., Cameron, J. L., Iacobuzio-Donahue, C. A., Hruban, R. H., and Lillemoe, K. D. (2001) Intraductal papillary mucinous neoplasms of the pancreas: An increasingly recognized clinicopathologic entity. Ann. Surg. 234, 313–321. 14. 14 Seidel, G., Zahurak, M., Iacobuzio-Donahue, C., et al. (2002) Almost all infiltrating colloid carcinomas of the pancreas and periampullary region arise from in situ papillary neoplasms: A study of 39 cases. Am. J. Surg. Pathol. 26, 56–63. 15. 15 Azar, C., Van de Stadt, J., Rickaert, F., et al. (1996) Intraductal papillary mucinous tumours of the pancreas. Clinical and therapeutic issues in 32 patients. Gut 39, 457–464. 16. 16 Nagai, E., Ueki, T., Chijiiwa, K., Tanaka, M., and Tsuneyoshi, M. (1995) Intraductal papillary mucinous neoplasms of the pancreas associated with so-called “mucinous ductal ectasia.” Histochemical and immunohistochemical analysis of 29 cases. Am. J. Surg. Pathol. 19, 576–589.
12
Hruban et al.
17. 17 Paal, E., Thompson, L. D., Przygodzki, R. M., Bratthauer, G. L., and Heffess, C. S. (1999) A clinicopathologic and immunohistochemical study of 22 intraductal papillary mucinous neoplasms of the pancreas, with a review of the literature. Mod. Pathol. 12, 518–528. 18. Z’graggen, K., Rivera, J. A., Compton, C. C., et al. (1997) Prevalence of activating K-ras mutations in the evolutionary stages of neoplasia in intraductal papillary mucinous tumors of the pancreas. Ann. Surg. 226, 491–498. 19. Fujii, H., Inagaki, M., Kasai, S., et al. (1997) Genetic progression and heterogene19 ity in intraductal papillary-mucinous neoplasms of the pancreas. Am. J. Pathol. 151, 1447–1454. 20. Wilentz, R. E., Albores-Saavedra, J., and Hruban, R. H. (2000) Mucinous cystic 20 neoplasms of the pancreas. Semin. Diagn. Pathol. 17, 31–42. 21. 21 Wilentz, R. E., Albores-Saavedra, J., Zahurak, M., et al. (1999) Pathologic examination accurately predicts prognosis in mucinous cystic neoplasms of the pancreas. Am. J. Surg. Pathol. 23, 1320–1327. 22. Moskaluk, C. A., Hruban, R. H., and Kern, S. E. (1997) p16 and K-ras gene muta22 tions in the intraductal precursors of human pancreatic adenocarcinoma. Cancer Res. 57, 2140–2143. 23. Brat, D. J., Lillemoe, K. D., Yeo, C. J., Warfield, P. B., and Hruban, R. H. (1998) 23 Progression of pancreatic intraductal neoplasias to infiltrating adenocarcinoma of the pancreas. Am. J. Surg. Pathol. 22, 163–169. 24. Brockie, E., Anand, A., and Albores-Saavedra, J. (1998) Progression of atypical 24 ductal hyperplasia/carcinoma in situ of the pancreas to invasive adenocarcinoma. Ann. Diagn. Pathol. 2, 286–292. 25. Longnecker, D. S., Adler, G., Hruban, R. H., et al. (2000) Intraductal papillarymucinous neoplasms of the pancreas, in (Hamilton S. R. and Aaltonen, L. A. eds.), Pathology and Genetics of Tumours of the Digestive System. Lyon: IARC Press, pp. 237–240. 26. Zamboni, G., Kloppel, G., Hruban, R. H., et al. (2000) Mucinous cystic neoplasms of the pancreas, in (Hamilton, S. R. and Aaltonen, L. A. eds.), Pathology and Genetics of Tumours of the Digestive System. Lyon: IARC Press, pp. 234–236. 27. Moskaluk, C. A. and Kern, S. E. (1997) Microdissection and polymerase chain 27 reaction amplification of genomic DNA from histological tissue sections. Am. J. Pathol. 150, 1547–1552. 28. Bonner, R. F., Emmert-Buck, M., Cole, K., et al. (1997) Laser capture microdis28 section: Molecular analysis of tissue. Science 278, 1481–1483. 29. 29 Maitra, A., Wistuba, I. I., Virmani, A. K., et al. (1999) Enrichment of epithelial cells for molecular studies. Nat. Med. 5, 459–463. 30. Wilentz, R. E., Su, G. H., Dai, J. L., et al. (2000) Immunohistochemical labeling 30 for Dpc4 mirrors genetic status in pancreatic: A new marker of DPC4 inactivation. Am. J. Pathol. 156, 37–43. 31. Iacobuzio-Donahue, C. A., Klimstra, D., Adsay, N. V., et al. (2000) DPC-4 pro31 tein is expressed in virtually all human intraductal papillary mucinous neoplasms
Precursors to Invasive Pancreatic Cancer
32.
33. 33 34. 34
35.
13
of the pancreas: Comparison with conventional ductal carcinomas. Am. J. Pathol. 24, 1544–1548. Iacobuzio-Donahue, C. A., Wilentz, R. E., Argani, P., et al. (2000) Dpc4 protein in mucinous cystic neoplasms of the pancreas: Frequent loss of expression in invasive carcinomas suggests a role in genetic progression. Am. J. Surg. Pathol. 157, 755–761. Kononen, J., Bubendorf, L., Kallioniemi, A., et al. (1998) Tissue microarrays for high-throughput molecular profiling of tumor specimens. Nat. Med. 4, 844–857. Mucci, N. R., Akdas, G., Manely, S., and Rubin, M. A. (2000) Neuroendocrine expression in metastatic prostate cancer: Evaluation of high throughput tissue microarrays to detect heterogeneous protein expression. Hum. Pathol. 31, 406–414. Kallioniemi, O. P., Wagner, U., Kononen, J., and Sauter, G. (2001) Tissue microarray technology for high-throughput molecular profiling of cancer. Hum. Mol. Genet. 10, 657–662.
14
Hruban et al.
Molecular Profiling of Tissue
15
2 Optimal Molecular Profiling of Tissue and Tissue Components Defining the Best Processing and Microdissection Methods for Biomedical Applications G. Steven Bova, Isam A. Eltoum, John A. Kiernan, Gene P. Siegal, Andra R. Frost, Carolyn J. M. Best, John W. Gillespie, and Michael R. Emmert-Buck Summary Isolation of well-preserved pure cell populations is a prerequisite for sound studies of the molecular basis of pancreatic malignancy and other biological phenomena. This chapter reviews current methods for obtaining anatomically specific signals from molecules isolated from tissues, a basic requirement for productive linking of phenotype and genotype. The quality of samples isolated from tissue and used for molecular analysis is often glossed-over or omitted from publications, making interpretation and replication of data difficult or impossible. Fortunately, recently developed techniques allow life scientists to better document and control the quality of samples used for a given assay, creating a foundation for improvement in this area. Tissue processing for molecular studies usually involves some or all of the following steps: tissue collection, gross dissection/identification, fixation, processing/embedding, storage/archiving, sectioning, staining, microdissection/annotation, and pure analyte labeling/identification. High-quality tissue microdissection does not necessarily mean high-quality samples to analyze. The quality of biomaterials obtained for analysis is highly dependent on steps upstream and downstream from tissue microdissection. We provide protocols for each of these steps, and encourage you to improve upon these. It is worth the effort of every laboratory to optimize and document its technique at each stage of the process, and we provide a starting point for those willing to spend the time to optimize. In our view, poor documentation of tissue and cell type of origin and the use of nonoptimized protocols is a source of inefficiency in current life science research. Even incremental improvement in this area will increase productivity significantly. Key Words: Molecular profiling; tissue processing; tissue staining; sample processing; laser microdissection; RNA; DNA; quality control; workflow management.
From: Methods in Molecular Medicine, Vol. 103: Pancreatic Cancer: Methods and Protocols Edited by: G. Su © Humana Press Inc., Totowa, NJ
15
16
Bova et al.
1. Introduction Isolation of well-preserved pure cell populations is a prerequisite for sound studies of the molecular basis of pancreatic malignancy and other biological phenomena. This chapter reviews current methods for obtaining anatomically specific signals from molecules isolated from tissues, a basic requirement for productive linking of phenotype and genotype. The quality of samples isolated from tissue and used for molecular analysis is often glossed over or omitted from publications, making interpretation and replication of data difficult or impossible. Fortunately, recently developed techniques allow life scientists to better document and control the quality of samples used for a given assay, creating a foundation for improvement in this area. Tissue processing for molecular studies usually involves some or all of the steps identified in Fig. 1. This diagram will serve as a guide for the remainder of the discussion in this chapter. Great tissue microdissection does not necessarily mean great samples to analyze. The quality of biomaterials obtained for analysis is highly dependent on steps upstream and downstream from tissue microdissection. It is worth the effort of every laboratory to optimize and document its technique at each stage of the process. Isolation of molecular materials from tissue components is a field in rapid evolution, and creativity in developing better ways to obtain pure cell populations and pure components is needed. In our view, poor documentation of tissue and cell type of origin and the use of nonoptimized protocols is a source of inefficiency in current life science research. Even incremental improvement in this area will increase productivity significantly. Most of the discussion in this chapter refers to cells in solid tissues; it applies equally to cells from body fluids or tissue aspirates when these cells are placed on glass slides or membrane-coated slides for microdissection. Flow cytometric cell purification is not discussed in detail here, but should be considered as an alternative to microdissection techniques whenever intact cells or cell components can be conveniently disaggregated and flow-separated based on reliable immunostaining or other features. Before starting a study requiring isolated cells or cell components, it is wise to consider the following: • What biomolecules (DNA, RNA, protein, carbohydrate, lipid) need to be recovered, how much is needed, and what level of purity is acceptable? Preliminary experiments may be needed to define how much starting material is needed and how pure the sam-ples need to be. • What is the required starting condition of the tissues to be dissected? If the tissue is frozen, how soon after loss of blood perfusion (loss of blood supply and/or nutrient supply) will it be frozen, and will the delay between perfusion loss and freezing/chem-
Molecular Profiling of Tissue
17
Fig. 1. Tissue processing for molecular analysis: flow diagram. ical fixation affect the biomolecules you seek to examine? Will drugs that the tissue donor has received affect the molecules to be analyzed within the tissue? Does the tissue have a high level of endogenous or exogenous bacterial or fungal DNase or
18
Bova et al.
RNase activity? Will frozen sections provide adequate histological detail to allow dissection of the cells of interest? If the tissue was chemically fixed (denoted “fixed” in the remainder of this chapter), are the target biomolecules in acceptable condition for the planned assay? For example, synthesis of full-length cDNA would be unlikely from formalin-fixed tissues, and target structures may not be adequately visualized from frozen sections, depending on how the structures are to be identified, and how the tissue sample was frozen. • What tissue dissection method will be most verifiable (i.e., providing evidence that the cells targeted were actually obtained without contamination), efficient, economical, and best documented?
Combining detailed answers to these questions with informed selection from the various options discussed in this chapter will optimize molecular profiling productivity. The following information is based on methods currently in use in our laboratories. However, it is not meant to be encyclopedic and the cited references provide a good starting point for additional reading. Also note that the history of tissue fixation and microdissection is not covered here. A brief review of this history is contained in a recent review by Eltoum et al. (1) and is touched on by Srinivasan et al. (2), and some of the key molecular methods discussed here (and additional topics) are also detailed in articles from the NCI Laboratory of Pathology Pathogenetics Unit (3–10). Some of the useful textbooks in the field are also listed in the bibliography (11–14). The books by Kiernan (13, 14) provide valuable information on chemical changes induced by fixation and staining that may be useful for those wishing to design and test new protocols. An e-mail listserver for broadcasting specific histotechnology-related questions (with a searchable archive of past questions) is also available (15). 1.1. Biosafety Issues Be sure to consider biosafety needs related to tissue handling in your laboratory. Tissues should always be handled using Universal Precautions. Fresh or processed tissues or their components should not come in direct contact with skin or mucous membranes, and in situations where tissue components could be released in the air, ventilatory isolation (by wearing masks and using biosafety hoods or other containment devices) should be used. Immunizations for preventable infections such as hepatitis B should be considered if the risk of exposure is considered significant. Higher levels of isolation are required if exposure to tuberculosis, prion disease, or other infectious disease is likely. If the tissues studied could contain particularly toxic substances (radioactive isotopes, chemotherapy drugs, etc.), appropriate steps should be taken to prevent significant exposure by laboratory personnel.
Molecular Profiling of Tissue
19
1.2. Tissue Collection and Processing The best tissue collection method will depend on the specifics of your situation. Interval between loss of blood perfusion and cooling or fixation, method of fixation, and uniformity of labeling and processing methods are some of the critical parameters for any study and are discussed in more detail below. “Fixation” of tissues can occur through freezing and/or through chemical fixation. We discuss only formaldehyde fixation (because it is still the standard) and alcohol-based tissue fixation (because it is a simple and inexpensive alternative that has worked in our hands), but the reader should be aware that a number of proprietary fixatives that claim to provide good histologic and molecular preservation are also available. 1.3. Staining of Tissue Sections Hematoxylin and eosin (H&E) staining has been the standard diagnostic tissue section staining method for more than a century. For molecular analyses, under the right conditions (see Subheading 3.), DNA and RNA can be obtained from H&E-stained material. Hematoxylin stains negatively charged molecules including nucleic acids and rough endoplasmic reticulum blue-violet, and eosin stains positively charged moieties including positively charged amino acids pinkred. Eosins are halogenated derivatives of fluorescein, and eosin Y is the form of eosin in most common use. Both RNA and DNA can be isolated from H&Estained sections, if the tissue is well preserved and stained properly. Because it fluoresces, eosin interferes with many protein analyses using fluorescent detection. It should also be noted that Mayer’s hematoxylin itself does not stain tissue. In solution, oxidizing agents such as alum (AlK[SO4]2.12H2O) convert hematoxylin to hematein. The correct terminology for this stain is Mayer’s hemalum, which is a concatenated product of hematein and alum. Other types of hematoxylin use other oxidizing agents. Methylene blue is a cationic dye. It stains DNA, RNA, and carbohydrate polyanions. Cytoplasm is strongly stained if a cell is rich in RNA (neurons with Nissl substance, secretory cells, etc.) or anionic mucosubstances (heparin in mast cells; many types of mucous). It is used prior to microdissection for DNA and protein isolation, but not for RNA isolation (16). Methyl green stains nuclei dark green, cytoplasm light green. According to a credible but nonpeer-reviewed study by Agilent, methyl green was best for RNA isolation when compared to the other stains mentioned here (16); methyl green is also reportedly compatible with DNA and protein isolation. Please note that what is currently sold as “methyl green” is actually “ethyl green” chemically. True “methyl green” has not been available for about 30 yr (17).
20
Bova et al.
Nuclear fast red stains nuclei dark red, cytoplasm lighter red. It is always used in conjunction with an aluminum salt, and its mechanism of action is not known. It is important to buy the right dye (Chemical Index or CI 60760) because the same name is sometimes used to label on other dyes that will not work (17). In the same Agilent study, nuclear fast red performed as well as H&E for RNA isolation, but better than methylene blue. 1.3.1. Immunostaining for Microdissection Tissue microdissection for molecular analysis is frequently limited by the difficulty in identifying cell types and structures by morphology combined with tinctorial (e.g., H&E) staining alone. The NCI Laboratory of Pathology Pathogenetics Unit and others have developed rapid immunostaining procedures for microdissection and RNA extraction from frozen sections (6), as summarized here. This method allows mRNA analysis of specific cell populations that have been isolated according to immunophenotype. Sections fixed in acetone, methanol, or ethanol/acetone give excellent immunostaining after only 12–25 min total processing time. Specificity, precision, and speed of microdissection are markedly increased due to improved identification of desired cell types. 1.4. Preparation of Cytologic Specimens for Microdissection Cells centrifuged from body fluids or fine-needle aspirates, or cells propagated in vitro can be prepared for microdissection by making direct smears or through a number of effective proprietary methods for creating thin layers of cells in designated areas of microscope slides. A subset of less adherent cells within fresh tissues can also be rapidly sampled by gentle scraping with a scalpel blade and then rapidly spreading the scraped sample onto a glass slide with the blade. The choice of strategy for preparing cell suspensions for microdissection depends on the anticipated cellularity of the sample. Highly cellular samples can be prepared as direct smears and effectively utilized for laser microdissection, less cellular samples can be concentrated using one of the proprietary cell concentration methodologies such as Cytospin® (Thermo Shandon Inc.), or more recent technologies such as ThinPrep (Cytyc Corp.) or AutocytePrep (TriPath Imaging). 1.5. Manual Microdissection of Blocks and Slides 1.5.1. Cryostat-Based Manual Dissection of Frozen Tissue Blocks It is often possible to obtain sufficient purity and relatively prodigious quantities of DNA, RNA, or protein from serial manual dissection of frozen tissue blocks directly on a suitable cryostat. Below we describe a method that can
Molecular Profiling of Tissue
21
increase purity from 10–50% to 75–95% for cell types that grow in macroscopic clusters. 1.5.2. Manual Microdissection of Tissue Sections on Slides Several manual microdissection methods can be performed on glass slides, and innovation in manual microdissection methods has continued despite the recent development of laser-based microdissection approaches. Techniques using hand-held tools (18), mechanical micromanipulators (19), manually cutting out areas of sections mounted on cellophane tape (20), ultrasonic oscillating needles (21), and methods specific to cytology specimens (22,23), have been described. Along these lines, Eppendorf has recently marketed a “Piezo Power Micro Dissection (PPMD) System” that is inexpensive relative to the laser dissection systems, and may work well when small quantities of dissected material are needed. The advantage of manual dissection is simpler equipment requirements, making it accessible to most laboratories. Its disadvantages are as follows: it is time consuming; it has a steep learning curve; the smallest dissectable region of interest (ROI) is generally significantly larger than that routinely obtainable with laser-based approaches; and documentation of manual dissection is usually not of as high a quality because it does not fit easily into the manual microdissection workflow. 1.6. Laser-Based Tissue Microdissection Systems Arcturus, Leica, and Zeiss/PALM laser-based Tissue Section Microdissection Systems are discussed here. All three systems are effective depending on specific needs of the user, and each instrument has its advantages and disadvantages. A comparison of the Arcturus, Leica, and Zeiss/PALM systems is contained in Table 1. Laser tissue microdissection systems have also recently been made available by Bio-Rad (Hercules, CA) and MMI AG (Glottbrugg, Switzerland), neither of which is discussed here. 1.6.1. Arcturus PixCell IIe LCM (Laser Capture Microdissection) System LCM utilizes an infrared laser integrated into a standard inverted microscope, and is based on patented “Laser Capture Microdissection”(LCM) technology originally described by Emmert-Buck et al. (24), and licensed to Arcturus Inc. (Santa Clara, CA, USA). Arcturus introduced its first PixCell system based on this technology in 1996. In LCM, a transparent plastic (CapSure™, Arcturus, Mountain View, CA, USA) cap with attached ethylene vinyl acetate (EVA) transparent thermoplastic membrane is placed on the surface of a non-coverslipped, stained tissue section mounted on a standard glass slide. The EVA film is in direct contact (CapSure Macro caps) or slightly above (CapSure HS caps) the tissue section (Fig. 2). CapSure HS caps are designed to reduce or eliminate the
22 Table 1 Comparison of Critical Features of Arcturus, Leica, and PALM Tissue Microdissection Systems Available in 2003
Tissue Isolation Method(s)
22
ROI: Region of interest
Selection of ROI
Leica AS LMD
PALM Micro Beam
(a) User melts EVA (ethylene vinyl acetate) plastic onto ROI using IR laser (980–1064 nm), plastic is lifted from slide, and ROI remains attached to EVA, while remainder of tissue remains on slide.
(a) UV-A Laser (337 nm) cutting of circumscribed ROI from inverted membrane-coated slides and cover slips. (b) “Cold” ablation of undesired areas with UV-A laser (337 nm)
Manual control of stage motion linked to video image or direct observation through microscope.
Automated optically-controlled laser beam targeting controlled by user selection of ROI on computer screen (neither stage nor laser moves). User can also view through eyepieces. User can select multiple ROIs within one field at a time.
(a) Noncontact transfer of ROI from slide surface by laser pressure catapulting (LPC) (b) “Cold” ablation of undesired areas with UV-A laser (337 nm) (c) UV-A laser cutting of circumscribed ROI from membranecoated slides and cover slips, followed by laser pressure catapulting of selected membrane region Automated stage motion (stepping size 128 nm) controlled by user selection of ROI on computer screen. User can select multiple coded ROIs within and outside initial image followed by automated collection process.
Bova et al.
Arcturus PixCell IIe
Lower size limit of a circular ROI
23 Disposables
Through eyepieces or on video or computer screen. 4 up to 40 objectives only, cannot use 100 objective (working distance limitation). Because of refractive index mismatch, visualization is often difficult or inadequate in regular use, depending on cells desired and tissue background. Theoretically, if plastic film adhesion is sufficient, can go to 5 µm or less. In practice, cells in tissues often will not be removed with only a 5-µm diameter circular spot, because adhesion to the slide is greater than adhesion to the plastic film. Dissection of organelles not feasible. Required use of original CapSure® LCM Macro Cap or CapSure® LCM HS Cap and associated materials
Through eyepieces or on computer screen. Leica recently improved manufacture of filmcoated slides, improving visualization of tissue.
Visualization on computer screen. In a demo, visualization decreased on membrane, and similar to Arcturus on regular slide, but massive improvement in visualization when dilute sterile mineral oil “liquid cover slip” placed on tissue. Selected ROIs color coded on computer screen.
At X66 high dry magnification, able to dissect 5-µm diameter area without difficulty. X100 not available on model tested, but reportedly can be done on coated cover slip.
Depending on magnification and and N.A. of objective. <1 µm diameter cutting can be routinely achieved with X100 1.3 N.A. objective. Dissection of organelles claimed by manufacturer, but not tested.
Required use of membranecoated slides. Can purchase from Leica, or can make your own with superglue and film roll purchased from Leica.
Can use device without proprietary disposables. Optional use of system with proprietary PALM membrane coated slides and coverslips.
23
(continued)
Molecular Profiling of Tissue
Visualization of ROI
24
Table 1 (Continued)
24
Leica AS LMD
PALM Micro Beam
Service history
Simpler device than the other two considered. IR laser is solid state, should never need service. History of excellent service in the United States.
No firsthand or anecdotal information available, but should be handled by Leica technicians. Pulsed nitrogen laser estimated good for 2 million pulses (approx 2 years).
Fluorescence option
Yes, does allow image light integration.
Yes. No information on integration of light.
Can fully dissect samples greater than 20 microns thick in a single pass? Software and documentation of dissection
No. Adhesive property of cap film not sufficient to pull off tissue.
Yes. Useful if tissue is homogeneous.
No firsthand or anecdotal experience available. Pulsed Nitrogen laser estimated good for 2 million pulses (approx 2 years), but is covered by service contract. Service will be handled by fully PALM trained Carl Zeiss USA technicians. Yes. Currently does not allow image light integration, but software modification to allow this is in pipeline. Yes. Useful if tissue is homogeneous.
Documentation of images as desired by user during dissection. Documentation of laser settings and images taken recorded in job file written to hard drive.
Standard software allows documentation of images only. Image database upgrade allows all electronic parameters to be recorded. Database can be linked to other software.
Documentation of images as desired by user during dissection. Key laser and slide position settings also stored automatically.
Bova et al.
Arcturus PixCell IIe
25
Special notes
Cost of instrument
Numerous publications illustrating isolation of various cellular components using LCM. Because dissection is contact-based, care must be taken to avoid contamination of cap with unwanted material; this can be avoided with special LCM caps with outer rails. Microsecond heating of tissue does occur during dissection, but does not appear significant. Must dissolve material off cap. Arcturus has developed an automated system for dissection based on same EVA melting principles with same issues and limitations. This device not reviewed here (cost is said to be $186K).
Approx $124K with fluorescence.
Newest instrument of group, but several publications citing its use. No clear indication that UV-A laser causes significant degradation of dissected tissue components, but this has not been rigorously tested.
Many publications illustrating isolation of various cell components. No clear indication that UV-A laser causes significant degradation of dissected tissue components, but this has not been rigorously tested.
Can be used to dissect living cells in culture using special culture dishes. Recently improved quality of membrane coated slides, vastly improving quality of visualization and laser cutting. Unclear if ablation by laser produces microaerosol that should be isolated from user. Approx $90K base, approx $120K with fluorescence and database software
Can be used to dissect living cells in special culture dishes. Can dissect tissues covered with light sheen of mineral oil allowing better visualization. Unclear if laser pressure catapulting causes microaerosol of tissue particles, and possible respiratory exposure, although this appears possible. Approx $160 K
25
(continued)
Molecular Profiling of Tissue
Quality of dissected materials for molecular analysis
26
Table 1 (Continued)
26
Arcturus PixCell IIe
Leica AS LMD
Cost of disposables
$3 or more per cap depending on cap type used
Bottom line: pros
Relatively simple to operate, good track record when: (a) visualization is not a problem, (b) area of interest is not too small, and (c) tissue “wants” to come off slide but does not come off too easily
Membrane-coated slides required, No disposables required, but and cost in range of $3–4 per slide membrane coated slides and coverslips cost in range of $4 each from PALM, but can be purchased elsewhere Dissections nearly always Middle of the road of the three work—you are consistently able devices in terms of device to get material off the slide. complexity Controls of focus, x/y/z, objective, and illumination convenient to user
PALM Micro Beam
Imaging sufficient for nearly any dissection can be obtained, by selecting from range of objectives, oil/no oil, “liquid cover slip.”
Excellent software interface Software is good, easy to use.
Company has partnered with Qiagen to create reagent systems.
Dissector has a large number of useful options in dissecting: combining ablation, laser catapulting, using membrane or not using membrane-coated slides. Excellent software interface, excellent documentation of dissection
Bova et al.
Company has excellent service record, excellent record of reagent manufacture.
Ability to store data in openly accessible database useful for some users.
Reasonably frequent problems getting targeted tissue to come off slide, wasting tissue and time
All dissections must occur on membrane, cannot use archival slides. Most labs opt to buy membrane slides, whose cost mounts quickly.
27
Difficult to get single cells when strong intercellular bonds present or tissue strongly adherent to slide
Dissected regions sometimes do not fall easily into cap. Need to “chase” membrane piece around.
Magnification not sufficient to do some single cell work. No X100 option available.
In dry environment, static electricity can interfere with gravity-based transfer to tube.
Visualization often not good, and few parameters to play with because “liquid cover slip” cannot be used
Track record relatively short relative to other two systems, but substantial installed base now present.
Special precautions/time needed to reduce potential for contamination by unwanted cells
More expensive More complex machine, learning curve higher, laser will require more maintenance. Larger footprint than other two devices No doubt that many investigators report good results obtaining various biomolecule types from dissections, but rigorous testing of laser pressure catapulting’s effects on nucleic acids and proteins has not been done to our knowledge.
Molecular Profiling of Tissue
Bottom line: cons
User must change objectives manually—can get out of sync with software.
Proprietary caps cost add up.
27
28
Bova et al.
Fig. 2. Arcturus laser capture microdissection (LCM) system.
risk of transfer of nontargeted cells. Viewing the tissue section under the microscope (directly or on a monitor), the investigator activates the infrared laser when the desired ROI is aligned with a targeting laser beam. The infrared laser (980– 1064 nm) pulse causes localized melting of the thermoplastic membrane, and expansion of the molten plastic causes it to contact the tissue adjacent to ROI. The EVA plastic then resolidifies, remaining adherent to the targeted ROI. As a result, areas where the EVA has come into contact with targeted tissue become optically clearer, allowing the user to identify easily tissues that have already been targeted. Other descriptions of the LCM process are contained in Eltoum et al. (1), and Curran (25), and others. A large number of experiments utilizing LCM have been published, illustrating that sufficient DNA, RNA, and protein can been obtained using LCM (24,26–30) for many types of experiments. Momentary heating of the tissue occurs with each contact by the molten EVA, but the effect of this heating generally appears to be minimal based on the success investigators have had with the device to date. Arcturus’ PixCell II software allows for visualization and documentation of images of ROI before and after dissection and the tissue transferred to the cap. After the LCM session is complete, the cap is removed from the slide. If the adherence of the EVA plastic to the targeted tissue is stronger than its adherence to the slide, the targeted tissue remains attached to the cap, leaving unwanted tissue regions on the slide, as illustrated in Fig. 2.
Molecular Profiling of Tissue
29
When dissection is complete for a given cap, Capsure® Macro caps with attached microdissected cells are placed onto microcentrifuge tubes containing an appropriate buffer for molecular analysis (caps were engineered to fit Eppendorf brand tubes). Capsure® HS caps must be attached to matching ExtracSure™ extraction devices, which allow extraction in small buffer volumes. The tube is then inverted and incubated as needed to allow microdissected cell components to go into solution. Positive aspects of the Arcturus LCM technology include: • • • • •
•
• •
•
Single cells and large areas can be dissected in many situations. Single cells can be “cherry-picked” from cell preparation slides. Equipment is relatively simple to manage and maintain. The laser is solid state and should rarely if ever need replacing. The Arcturus company provides generally excellent support including protocols and kits tailored to the LCM system. Challenges associated with the LCM technology include: Transfer of tissue from slide to LCM cap sometimes requires troubleshooting. Tissue that is targeted may remain on the slide after the CapSure® cap is removed, indicating that adhesion between the thermoplastic and the tissue is too weak, or the adhesion between the tissue and slide is too strong. Insufficient adhesion of the thermoplastic to the tissue can be caused by insufficient dehydration of the target tissue, which can sometimes be overcome by repeat immersion of the target slide in fresh xylene. When adhesion of the tissue to the glass slide is greater than tissue adhesion to the thermoplastic film, this can be difficult to overcome. Unevenness of tissue surfaces, caused by wrinkled sections or resulting from marked differences in tissue firmness (such as with plant tissues where stiff cellulose is next to soft internal cellular elements, or with tissues containing bone or cartilage), also can make dissection difficult. In some instances, internal bonding within tissue elements prevents selective tearing of individual cells from adjacent cells, preventing LCM altogether. It can be costly because of the need to use proprietary Arcturus CapSure® caps, which cost in the range of $2–3 apiece. Minimal dissectable area is variable. Single cells can easily be dissected in some tissues, while in others single-cell dissection is impossible because of excessive adhesion to the slide, or because of the above-mentioned resistance to tearing (internal bonding) of target tissue. Visualization in some tissues is not adequate. This is most often due to an index of refraction mismatch between the tissue section and air between the tissue and CapSure™ cap. With the PALM Microbeam and Leica LMD systems, with tissue mounted on membrane-coated slides, a “liquid cover slip” can be placed on the tissue section to provide visualization similar to ordinary cover-slipped sections, without interfering with dissection. This is not feasible with the Arcturus IIe LCM system, as liquid would interfere with adherence between the EVA plastic and the tissue.
30
Bova et al.
Fig. 3. Leica AS LMD laser microdissection system.
Arcturus recently marketed a new device called Autopix™, which is an automated, enclosed version of the technology contained in the PixCell IIe. Arcturus continues to market the PixCell IIe, which at the time of this writing is probably the most prevalent laser-based microdissection technology. Arcturus uses Olympus microscope in its systems, but has its own service and support team covering all aspects of its devices. 1.6.2. PALM Micro Beam Research System PALM Microlaser Technologies AG (Bernried, Germany), founded in 1993, originally focused on use of microscope-based ultraviolet (UV) lasers in assistedreproduction technology such as zona pellucida drilling, but moved into the area of tissue section microdissection when prodded by pathologists and researchers faced with difficulty in obtaining pure cell samples. PALM marketed its first system specifically for tissue microdissection around 1994, and its current device is called the Micro Beam microdissection system. PALM’s tissue section microdissection systems differ in a number of important respects from the Arcturus and Leica technologies, and these differences are illustrated in Figs. 2–4. The most important positives include the ability to obtain improved visualization by using a “liquid cover slip,” relative predictability of obtaining targeted ROIs, relative predictability in obtaining ROI in the 2- to 5 µm diameter range, rela-
Molecular Profiling of Tissue
31
Fig. 4. Zeiss/PALM Micro Beam laser microdissection system.
tive precision of the ablating UV beam (compared to Leica LMD), availability of PALM’s patented Laser Pressure Catapulting (LPC) technology for removal of targeted tissue from the slide (as compared to gravity used by Leica), and flexibility in the choice of mounting platforms (glass or special plastic membranes). In addition, the PALM system software was intuitive and adequately sophisticated both for dissections and documentation in our testing. The most important potential challenges with the current PALM technology are as follows: their technology is somewhat more expensive; UV lasers require more maintenance and are more expensive; and in the United States at least, PALM relies on Zeiss for service and support, an arrangement that may be difficult but for which we have no direct experience. PALM recently created a similar arrangement with Olympus in Europe, and it is unclear whether the PALM technology will continue to be supported on both Zeiss and Olympus platforms, and how these platforms might differ. Similar to the Arcturus system, the best current evidence of its effectiveness for a range of molecular studies is the large number of publications using material isolated using the PALM system (31–34). 1.6.3. Leica LMD Microdissection System Leica was the third major entrant into the laser microdissection device field, and its current Leica LMD system has features that in some ways are hybrids between the Arcturus and PALM systems, as summarized in Table 1 and Figs. 2–4. Similar to the PALM system, its laser is an ablating 337-nm laser in the UV-A range. Similar to the Arcturus, its hardware is compact and its software
32
Bova et al.
is relatively bare bones compared to the PALM software (the version tested in 2003 did not provide more advanced documentation tools, such as allowing direct annotation of images within the Leica software). Positive attributes include several convenient features including a single manual control box for all major dissection functions. These include integration of objective selection and software adjustment to a new magnification (not supported in the PALM version tested), convenient ability to send dissected ROIs from different regions of a slide to separately identified microcentrifuge tubes, and the ability to use ordinary intact microcentrifuge tubes for collection of dissected materials (not possible on either the Arcturus or PALM systems tested). Significant challenges with the Leica system include required use of membrane coated slides, occasional failure of the targeted ROI to drop into the collection tube after dissection, relative lack of fine focus of the UV laser beam compared to the PALM system, and poor performance when using UV illumination. The Leica LMD system is relatively new, but publications citing its use in molecular studies have already appeared (35,36). Used carefully, all three of the laser-based tissue microdissection systems discussed provide technology sufficient for isolation of most ROI, obtaining material of sufficient quality for many if not most downstream molecular analyses. Selection of the best system for an application should be based on one’s specific needs and resources available, taking into consideration the various positive attributes and challenges listed. Technologies are evolving rapidly, so consulting with company websites and representatives is recommended prior to making purchase decision. More recently marketed tissue microdissection systems from Bio-Rad (Hercules, CA) and MMI AG (Glottbrugg, Switzerland), and other new entrants have not been reviewed by us but should also be considered. 2. Materials 2.1. Rapid Tissue Freezing 2.1.1. Isopentane Method 1. 2. 3. 4. 5.
Isopentane. Dry ice. Ice bucket. Metal bowl, 500-mL or larger volume. Metal or Plexiglas basket corresponding to metal bowl to hold tissue while freezing.
2.1.2. Gentle-Jane® Method 1. Gentle-Jane® (Instrumedics) snap-freezing device. 2. Liquid nitrogen. 3. OCT (“optimal cutting temperature”) compound (Tissue-Tek, others), Cryogel (Instrumedics), or other embedding medium.
Molecular Profiling of Tissue
33
2.2. Preparation of Cytologic Specimens for Microdissection or Direct Molecular Analysis 2.2.1. Tissue Disaggregation Methods Solid tissues can be disaggregated to obtain pure cell mixtures ready for further purification and/or live cell experiments or for direct molecular separation and phenotyping. 2.2.1.2. ENZYMATIC
AND/OR
MECHANICAL DISAGGREGATION
1. Fresh tissue to be disaggregated. (This methods may also work on frozen tissues [37], and success may be increased if the tissues are frozen in media containing dimethyl sulfoxide (DMSO), that may protect cell membranes.) 2. Sterile solution compatible with cells of interest such as RPMI-1640 or Hank’s balanced salt solution (HBSS), with addition of serum, defined additives (growth factors, insulin, cytokines, etc.), and antibiotics as desired. 3. Enzyme(s) such as collagenase 1A (Sigma, St. Louis, MO, USA) to be used for disaggregation (see example) if needed. 4. Sterile scalpels and/or automated mechanical disaggregation system (such as Medimachine™ marketed by BD (formerly Becton Dickenson, Franklin Lakes, NJ).
2.2.2. Direct Smears of Liquids Containing Suspended Cells 1. Standard (uncoated) glass microscope slides. 2. 50–250 µL of fluid to be smeared. 3. Hemocytometer cover slip or other device narrower than glass slide to use as a cell spreader. 4. 70% ethanol.
2.2.3. Tissue Scraping Fresh tissues containing cells that can be detached intact with mild shearing force can be rapidly sampled by scraping with a sterile scalpel blade and then rapidly spreading the scraped liquid sample onto a glass slide with the blade. A basic requirement for use of this technique together with microdissection is that the intended target cells can be readily identified cytologically or by cytology together with specific staining characteristics. 2.2.4. Cell Concentration Technologies 2.2.4.1. DENSITY
AND
SIZE GRADIENT SEPARATIONS
Many types of cells in solution can be separated by density and/or size, without significant damage. For example, from disaggregated living liver cells, hepatocytes, Kupffer cells, liver endothelial cells, and bile duct epithelial cells can be separated using density gradients (38):
34
Bova et al.
1. 2. 3. 4. 5.
Percoll (Amersham), RediGrad (Amersham), Ficoll-Paque Plus (Amersham), others. Sterile 50-mL conical tubes. Diluent (usually phosphate-buffered saline or similar). Red cell lysis buffer (ACK by Bio-Whittaker, others). Centrifuge.
2.2.4.2. MAGNETIC-BEAD
OR OTHER
BEAD-BASED SEPARATIONS
Immunobead-based separation has been used effectively for concentrating multiple myeloma cells from fresh bone marrow aspirates (39) for molecular analysis. With a well-optimized combination of disaggregation methods and specific antibodies, other types of fresh or frozen tissue can be separated using a combination of density gradient and magnetic-bead separations. Methods will vary depending on tissues separated and target cells desired. Listed here is a general pattern. 1. Aqueous sample containing intact cells of interest admixed with other cells. 2. Erythrocyte lysis buffer (if needed). Two such buffers are: 0.32 M sucrose, 10 mM Tris-HCl, 5 mM MgCl2, 1% Triton X-100, or alternatively hypotonic solutions (0.15 M NH4Cl in 0.1 mM Tris-HCl). 3. Appropriate nonspecific (blocking) and specific antibodies (multiple suppliers) for selection of wanted or unwanted cells. 4. Paramagnetic beads coated with secondary antibody specific to the primary antibodies used (Dynal, Oslo, Norway, and others). 5. Magnet (Dynal, others) for separation of beads, leaving target cells in supernatant, or on beads as designed.
2.2.4.3. SEMI-AUTOMATED CELL CONCENTRATION METHODS
These include Cytospin® (Thermo Shandon Inc.) preparations, or by more advanced thin-layer technology called ThinPrep® (Cytyc Corp.) or AutocytePrep (TriPath Imaging). The Cytospin® technique uses proprietary centrifuge cartridges to separate cells in suspension and place them on specific areas of a microscope slide ready for staining. ThinPrep® also uses proprietary solutions (separate protocols for fine-needle aspirates, body fluids, and mucoid samples) and instrumentation to produce cells on slides. Tripath’s solution is targeted at Papanicolau cervical smears but could be modified for other purposes. Consult the manufacturer’s manual for details. 2.3. Chemical Fixatives Alcohol-based and formaldehyde-based tissue fixation methods are presented, as they currently come closest to providing both sufficient histologic detail for light microscopy and sufficiently preserved material for molecular analysis of varying types. Metal-based fixatives (Zenker’s, Harris hematoxylin, etc.) can
Molecular Profiling of Tissue
35
provide excellent histology but can interfere with many molecular analyses and are not discussed. 2.3.1. 10% Neutral Buffered Formalin (10% NBF) (see Notes 1 and 2) Many laboratories purchase 10% NBF ready-made from a wide variety of suppliers. It is relatively easy to make, however, and in some instances (see Notes) making your own may provide significant advantages: 1. 2. 3. 4. 5. 6.
Fume hood. 2-L graduated cylinder with stir bar. Magnetic stirring plate. 1 L of 10% formalin (100 mL of 37% formaldehyde diluted 1:10). 4 g of sodium phosphate monobasic (Na2H2PO4 . H2O). 6.5 g of sodium phosphate dibasic (anhydrous) (Na2 H2 PO4).
In a fume hood, add stir bar, sodium phosphate monobasic, and sodium phosphate dibasic to empty graduated cylinder on stir plate (see Note 3). Add 900 mL of water followed by 100 mL of 37% formaldehyde (see Note 4). Stir until phosphate dissolves (see Note 5). 2.3.2. Alcohol-Based Solutions (see Note 6) 1. 70% ethanol (see Note 7). To make 1 L, add 700 mL of 100% ethanol to a graduated cylinder, and add water to 1 L. The amount of water needed to reach 1000 mL will be roughly 330 mL because of volume contraction caused by hydrogen bond formation. 2. 4.2:2.0:1.8 (v/v/) Ethanol–methanol–water fixative (see Notes 8–10). To make 8 L, to a 10-L graduated cylinder, add 1800 mL of H2O, 2000 mL of 100% methanol, and 4200 mL of 100% ethanol. Mix well with a stir bar and bring the volume up to 8 L with water) to make up for volume shrinkage due to hydrogen bonding.
2.4. Combined Formalin Fixation/Sucrose Infusion for Cryostat Sectioning In situations in which standard frozen sections may not provide adequate histology for downstream microdissection, and where standard formalin fixation and paraffin embedding will not provide adequate DNA, RNA, or protein quality, an alternate approach is to perform short-term fixation with freshly prepared 2% or 4% buffered formaldehyde (as opposed to using commercially prepared 10% formalin, which is 4% formaldehyde but also contains a small amount of methanol, which coagulates proteins), then rinsing the sample with glycine (to remove excess formaldehyde and inactivate formaldehyde) followed by sucrose infusion.
36
Bova et al.
1. Fresh 2% or 4% neutral buffered formaldehyde, or modified Millonig formalin (see Subheading 4.3.1.) if isoosmotic fixation is desired. Note that this freshly made “formaldehyde” is the same as what many call “paraformaldehyde.” 2. Sterile phosphate-buffered saline (PBS). 3. 0.5 M Sucrose in PBS.
2.5. Standard Automated Clearing and Paraffin Embedding of Fixed Tissues Most molecular laboratories will work with histology laboratories for tissue clearing (usually with xylenes) and embedding (usually with paraffins). Modern paraffin tissue embedding occurs almost entirely in automated processors under vacuum. Molecular researchers must recognize, however, that standard surgical pathology tissue processors may start with a formalin step, and thus if formalin exposure is to be avoided, a separate no-formalin processor run must be created. In addition, processors on which formalin is routinely used may allow formalin exposure even when not selected. Higher temperatures and longer exposure times than desired may also be commonplace. Examination of these parameters with one’s support laboratory is recommended. Paraffins for embedding vary in melting temperature and quantity of added polymer plastic as stiffening agent. Increased polymer allows greater stiffness and thinner sectioning, but also increases infiltration time. 2.6. Sectioning of Tissue Blocks for Microdissection 2.6.1. Frozen Sections 1. 2. 3. 4.
Cryostat. OCT or other embedding compound. Mounting chucks for cryostat. Dry ice in styrofoam container.
2.6.2. Paraffin Sections 1. 2. 3. 4.
Microtome. Water bath filled with deionized/distilled ddH2O for floating sections. Clean, disposable blade. Oven for baking slides.
2.7. Staining of Tissue Sections for Microdissection 2.7.1. Mayer’s Hemalum and Eosin Staining for DNA, RNA, and Protein Recovery 1. Fresh xylenes (Sigma, others). 2. Fresh 100%, 95%, 70% ethanol (use diethyl pyrocarbonate [DEPC]-treated water for dilution of 100% ethanol).
Molecular Profiling of Tissue
37
3. Mayer’s hematoxylin (Sigma, Richard-Allan, others). 4. Eosin Y (Sigma, Richard-Allan, others). 5. Mini-protease inhibitor tablets (Roche) (if protein recovery is desired).
2.7.2. Immunostaining Prior to Laser Microdissection 1. 2. 3. 4. 5.
6. 7. 8. 9. 10.
DEPC-treated H2O (Invitrogen/Research Genetics) (for preparation of PBS, alcohols). Superfrost Plus glass slides (Fisher Scientific). Cold acetone. 1X PBS, pH 7.4. DAKO Quick Staining kit (DAKO Corp.), a three-step streptavidin–biotin technique with prediluted mono- or polyclonal (rabbit) primary antibodies optimized for very short staining times. Diaminobenzidine (DAB). Hematoxylin solution, Mayer’s (Sigma). 70%, 95%, 100% ethanol. Xylenes, mixed, ACS grade (Sigma). Placental RNase inhibitor (Perkin Elmer, Branchburg).
2.7.3. Methylene Blue Staining 1. Methylene blue (0.05% in water, Sigma 31911-2). 2. DEPC-treated water (Invitrogen/Research Genetics, others).
2.7.4. Methyl Green Staining 1. Methyl green solution (actually “ethyl green;” see Subheading 1.3.) (Dako, S1962). 2. DEPC-treated water (Invitrogen/Research Genetics, others).
2.7.5. Nuclear Fast Red 1. Nuclear fast red solution (Dako, S1963). 2. DEPC-treated water (Invitrogen/Research Genetics, others).
2.8. Manual Microdissection 2.8.1. Cryostat-Based Microdissection of Tissue Blocks 1. 2. 3. 4. 5. 6. 7. 8. 9. 10.
Cryostat with blade guards. OCT or other embedding compound. Brightfield microscope. Glass slides. 50-mL conical tubes. 16-in. long smooth forceps. Insulated bucket containing liquid nitrogen. Sterile one-sided straight edge blades. Mayer’s hemalum and eosin slide staining setup. Logbook or computerized logging system.
38
Bova et al.
2.8.2. Manual Microdissection of Tissue Sections on Glass Slides The method described is a combination of the one in use in the NCI Pathogenetics Unit and the one described by Moskaluk and Kern (40): 1. Stained glass-slide-mounted sections without coverslips and adjacent stained, cover-slipped sections. 2. Dissecting microscope or standard inverted microscope with hydraulic micromanipulator arm, if available. 3. Sterile 30-gage needle on syringe. 4. 2.5% Glycerol solution. 5. Agarose and buffers as needed for specific dissections.
2.9. Laser Microdissection of Tissue Sections on Slides 2.9.1. Arcturus LCM System 1. 2. 3. 4.
LCM device. CapSure dissection caps. Post-It® Notes or Arcturus PrepStrips™. Perkin Elmer/Applied Biosystems GeneAmp 500-µL thin-walled polymerase chain reaction (PCR) reaction tubes, cat. no. N8010611.
2.9.2. PALM Microbeam System 2.9.3. Leica LMD System 2.10. Isolation of Analyte from Microdissected Materials 2.10.1. Isolation of DNA DNA isolation kits are available from a number of manufacturers. For small samples, kits may be more cost-effective than nonkit methods. For large samples (if >10 µg of DNA will be isolated), we believe nonkit methods are most cost effective. 2.10.1.1. KIT-BASED METHODS A number of kits are available for DNA isolation. For LCM-derived materials, we tested Trizol, DNAzol, Trireagent, Easy-DNA (Invitrogen), and the DNeasy kit (Qiagen), against sodium dodecyl sulfate (SDS)–phenol–chloroform extraction, and Tween-20–phenol–chloroform extraction, and of the kits, the DNeasy kit performed best in our hands (GSB Laboratory), although SDS– phenol–chloroform provided a better yield, in our hands, DNA size is generally larger with DNeasy than with phenol–chloroform extraction. DNeasy is also easier to use because of its lower toxic chemical content. We describe a method using the DNeasy kit and LCM caps, although this could be applied to microdissected material from any source. Several more recently marketed DNA isolation kits are also available, but have not been tested by us.
Molecular Profiling of Tissue
39
1. DNeasy kit (Qiagen) or similar kit. The DNeasy kit contains 2-mL collection tubes, a series of proprietary (“black box”) buffers: ATL (lysis buffer), AL (lysis buffer contains guanidine hydrochloride), AW1 wash buffer (contains guanidine hydrochloride), AW2 wash buffer, AE elution buffer; also proteinase K (activity = 600 mAU/mL solution or 40 mAU/mg of protein). 2. LCM cap removal device (if LCM caps are source of DNA). 3. 100% (Absolute) ethanol.
2.10.1.2. SDS/PHENOL–CHLOROFORM EXTRACTION
We compared Tween-20–phenol–chloroform extraction to SDS–phenol– chloroform extraction, and in our hands (GSB laboratory), SDS–phenol–chloroform extraction provided double or greater yields, and better quality DNA. Our SDS–phenol–chloroform extraction method is described here and is based on the method of Goelz et al. (41). The procedure given is for large quantities of DNA (10–1000 µg), and can be scaled down for smaller samples. 1. 1 L of DNA digestion buffer. a. Sterile bottle. b. 750 mL of sterile, deionized/distilled H2O (ddH2O). c. 50 mL of 1 M Tris-HCl (pH 8). d. 100 mL of 0.5 M EDTA, pH 8. e. 100 mL of 20% SDS. Combine the ingredients in the bottle. Gently add 20% SDS to avoid foaming. Gently mix the solution, and label the bottle: DNA extraction buffer: 50 mM TrisHCl, 50 mM EDTA, 2% SDS. This can be stored at room temperature. Place 6 mL in a sterile 50-mL conical tube to isolate DNA from 100 to 300 6-µm frozen tissue sections. Scale down for smaller amounts of tissue. 2. DNA digestion materials. a. 5 mL of proteinase K (15.6 mg/mL) (Roche Molecular Biochemicals, cat. no. 1-373-196). b. Incubator. c. Rocking device, such as Belly-Button Shaker/Rocker (Stovall Life Science, Inc.). Set the incubator (or water bath) to 48°C and allow to equilibrate. To each sample (6 mL), add 38 µL of proteinase K. Incubate on undulating Belly Button (Stovall) or other gentle agitation device at 48°C for 12–18 h (typically overnight) (see Notes 11 and 12). 3. 50 mL of LoTE DNA suspension buffer. a. Sterile 50-mL conical tube. b. 44.5 mL of ddH2O. c. 450 µL of 1 M Tris-HCl, pH 8.0. d. 9 µL of 0.5 M EDTA, pH 8.0. Add ddH2O, Tris-HCl, and EDTA to a conical tube; mix well. Filter sterilize if desired. Label tube. Final concentrations are: 10 mM Tris-HCl, 1 mM EDTA. Keep refrigerated.
40
Bova et al.
4. DNA purification. a. One box of serum separation tubes (SST) (Becton Dickinson-Vacutainer Systems, cat. no. 366512). (Also consider using Phase Lock Gel™ tubes of various sizes, marketed by Eppendorf.) b. Phenol–chloroform–isoamyl alcohol (25:24:1 v/v/v) (UltraPure Gibco BRL). c. Sterile 50-mL conical tubes. d. Centrifuge and rotor capable of handling SST and 50-mL conical tubes.
2.10.2. Isolation of RNA 2.10.2.1. RNASE AWAY™ (MOLECULAR BIOPRODUCTS) OR OTHER RNASE INACTIVATING SOLUTION 2.10.2.2. KIT-BASED METHOD (PICOPURE™ RNA ISOLATION KIT
BY
ARCTURUS)
This kit contains several proprietary (“black box”) buffers, including conditioning buffer (CB), GITC-based extraction buffer (XB), 70% ethanol (EtOH), wash buffer 1 (W1), wash buffer 2 (W2), elution buffer (EB), RNA purification columns, collection tubes, and microcentrifuge tubes. Although the description is focused on material derived from LCM, it is equally applicable to tissue isolated using any type of microdissection. Small-quantity RNA isolation kits are also available from Ambion, Qiagen, and others. 2.10.2.3. PHENOL–CHLOROFORM BASED RNA EXTRACTION METHOD 1. Rnase-free microcentrifuge tubes. 2. DEPC-treated water purchased ready for use (Invitrogen, Sigma, others) or made by adding 0.2 mL of DEPC (Sigma and others) per 100 mL of ddH2O, shaking vigorously to get DEPC into solution, and autoclaving to inactivate remaining DEPC. Caution: Handle DEPC only in a fume hood, as it may be a carcinogen. 3. 30 mL of GITC denaturing solution made with 29.3 mL of ddH20, 1.76 mL of 0.75 M Na citrate, pH 7.0, 2.64 mL of 10% (w/v) N-lauroylsarcosine, 25.0 g of guanidine thiocyanate (dissolves with stirring at 65°C), and 35 µL of 2-mercaptoethanol. Final concentrations are 4 M guanidinium isothiocyanate, 0.5% N-lauroylsarcosine, 25 mM sodium citrate, and 0.1 M 2-mercaptoethanol. 4. 2 M Sodium acetate pH 4 (add 1.64 g of anhydrous sodium acetate to 4 mL of water and 3.5 mL of glacial acetic acid, and bring the volume to 10 mL with ddH2O). 5. Water-saturated buffered phenol. Dissolve 10 g of phenol crystals in ddH2O at 65°C. Mix dissolved phenol with 200 mM Tris-base until the aqueous solution reaches pH 8. Remove the upper water phase and store at 4°C for up to 1 mo. 6. 70% Ethanol (prepared with DEPC-treated water). 7. 49:1 (v/v) chloroform–isoamyl alcohol. 8. 100% isopropanol. 9. Refrigerated microcentrifuge or other refrigerated centrifuge capable of handling microcentrifuge tubes. 10. Tissue disaggregating device, as needed.
Molecular Profiling of Tissue
41
11. Formamide purchased ready for use (Sigma, Molecular Research Center, others) or prepare in the laboratory by mixing with 1 g of AG 501-X8 ion-exchange resin (BioRad, others) per 10 mL of formamide for 45 min and filter at room temperature.
2.10.2.4. DNASE TREATMENT
OF
TOTAL RNA
1. DEPC-treated ddH2O (see Subheading 2.10.2.3. for recipe). 2. RNase inhibitor (Perkin Elmer). 3. DNase I (GenHunter Inc) and DNase I buffer (GenHunter, Inc.).
2.10.3. Isolation of Protein for 2-Dimensional Gel Electrophoresis Protein extraction methods vary widely depending on the intended analysis method and are not detailed here. Recent papers by Ornstein (9), Simone (10), and Emmert-Buck et al. (42) contain useful protocols for protein extraction from microdissected frozen tissue material, and a 1998 paper by Ikeda et al. (43) suggests that paraffin-embedded tissue can also be used for proteomic analysis. 3. Methods 3.1. Rapid Tissue Freezing (see Notes 13 and 14) 3.1.1. Isopentane Method 1. Place 400 mL of isopentane (highly flammable) in metal container large enough to hold a corresponding Plexiglas or metal basket (see Note 15). 2. Place bed of dry ice in ice bucket. 3. Place the metal container containing isopentane onto dry ice. Allow 15–30 min for a cool-down period. Put the tissue to be frozen into the metal or Plexiglas basket long enough for complete freezing, Excess isopentane (highly flammable) can be removed with a paper towel (see Note 16).
3.1.2. Gentle-Jane® Method In our hands, because of the rapidity of freezing, this method provides superior frozen section histology, and in some cases provides as good as or better histology than from formalin-fixed, paraffin-embedded samples (see Note 17). Rapid freezing reduces ice crystal size, preserving cellular architecture. Significant drawbacks are covered in the Notes (see Note 18). 1. Place sufficient liquid nitrogen in the Gentle-Jane (Instrumedics) liquid nitrogen container to cover the metal piston (see Note 19). 2. Wait until liquid nitrogen stops boiling (metal and liquid nitrogen at equal temperatures). 3. Place tissue on freezing pedestal. 4. Move metal piston to piston-holder, and gently lower onto tissue located on the freezing pedestal. 5. Remove the frozen tissue from the pedestal.
42
Bova et al.
3.2. Preparation of Cytologic Specimens for Microdissection 3.2.1. Tissue Disaggregation Methods A generic method for tissue disaggregation is defined below. Specific methods will need to be optimized depending on the specific input tissues and output needs of the method. The method described is based on reports by Novelli et al. (44) of the use of collagenase and the Medimachine™ for isolation and downstream analysis of cutaneous lymphocytes (44) and by Brockhoff et al. for isolation of colon carcinoma cells for downstream flow cytometry (37). 3.2.1.1. ENZYMATIC
AND/OR
MECHANICAL DISAGGREGATION (SEE NOTE 20)
1. Mechanically disaggregate tissue using sterile scalpel blades or the Medimachine™ as recommended by the manufacturer (http://www.bdbiosciences.com/immunocytometry_systems/brochures/pdf/mmach_download.pdf). The Medimachine system requires selection of the cutting device (Filcon™) mesh pore size desired (35 or 50 µm; 30 µm is for isolation of nuclei, 50 µm for whole cells), and the specific filter pore shape (syringe type for larger volumes, cup type for smaller volumes) and pore size desired (10–500 µm). Alternatively, if manual mechanical disaggregation is performed with sterile scalpel blades, 70-µm (or other) pore size tissue strainers (Falcon, BD Inc., Franklin Lakes, NJ) can be used. 2. If desired, either prior to (if the tissue sample is very small) or after mechanical disaggregation, the suspended tissue clumps can be exposed to disaggregating enzymes such as collagenase 1A (Sigma). Overexposure can lead to decreased cell integrity and viability, so optimization of exposure duration and temperature is critical for success.
3.2.2. Direct Smears of Liquids Containing Suspended Cells (see Note 21) 1. Place one drop of liquid sample on a glass slide. 2. Drag the hemocytometer cover slip over the liquid sample. 3. Drop immediately (before drying) into fresh 70% ethanol for 10 dips. Then airdry or proceed to staining.
3.2.3. Tissue Scraping (see Notes 22 and 23) 1. Gently scrape a sterile blade against wet tissue surface to collect fluid on the edge of the blade. 2. Drag the blade gently across the dry glass slide. 3. Immediately plunge into 70% ethanol for 20 dips, then proceed to staining or airdry.
3.2.4. Cell Concentration Technologies 3.2.4.1. DENSITY
AND
SIZE GRADIENT SEPARATIONS (SEE NOTE 24)
1. Dilute cells, and prepare gradient in 50-mL conical tubes as directed by the manufacturer.
Molecular Profiling of Tissue
43
2. Centrifuge the cells as directed by manufacturer (usually approx 10,000g if cells are in saline), wash if desired, and recentrifuge. 3. Lyse red cells if desired, wash, recentrifuge, resuspend, and count cells before downstream molecular analysis or further separation by other methods.
3.2.4.2. MAGNETIC-BEAD BASED SEPARATIONS (SEE NOTE 25)
Magnetic-bead (or other bead-based) separations can be used separately or together with microdissection-based isolation. For particularly bloody specimens, red blood cell lysis agents should be added to bead-based protocols. Density-based separation methods (such as Ficoll gradients) can also add power to bead-based techniques. 1. Obtain sterile aqueous samples containing intact cells of interest admixed with other cells. 2. If necessary, use density gradient techniques to remove erythrocytes or use erythrocyte lysis buffer such as those listed. 3. Wash cells and incubate with appropriate nonspecific (blocking) and specific antibodies for selection of wanted or unwanted cells. 4. Wash cells and incubate with paramagnetic beads coated with secondary antibody directed against the primary antibodies used (Dynal, Oslo, Norway, and others). 5. Magnetic (Dynal, others) separation of beads, leaving target cells in supernatant, or on beads as desired.
3.2.4.3. SEMI-AUTOMATED CELL CONCENTRATION METHODS (SEE NOTE 26).
For Cytospin® (Thermo Shandon, Inc.), ThinPrep® (Cytyc Corp.), or AutocytePrep (TriPath Imaging) contact the manufacturer for specific materials and methods. 3.3. Combined Formalin Fixation/Sucrose Infusion for Cryostat Sectioning (see Note 27) 1. Fix samples (dimensions as small as possible to allow rapid penetration of fixative) in 2% or 4% buffered formaldehyde at 4°C for 4–24 h (duration needs to be optimized for your tissue/application). When possible, perfuse the entire organ with saline followed by formaldehyde prior to sectioning and placement in fixative solution. Usually this can be accomplished only via volume replacement through the left ventricle. 2. Wash samples with sterile 4°C PBS. 3. Place samples in 0.5 M sucrose in PBS at 4°C. When tissue sinks, infusion is complete. Do not allow the infusion to go longer than 1 d (reduces quality of tissue morphology). 4. Place tissue samples in OCT or other embedding compound, and freeze rapidly. Section on a cryostat.
44
Bova et al.
3.4. Standard Automated Clearing and Paraffin Embedding of Fixed Tissues (see Note 31) See Subheading 2.5. for comments. In terms of standard embedding materials, we have had good results with type VI paraffin (Richard Allan Scientific, Kalamazoo, MI), but many other products are available that are suitable. 3.5. Sectioning of Tissue Blocks for Microdissection 3.5.1. Frozen Sections (see Notes 29 and 30) 1. 2. 3. 4.
Place a small amount of OCT onto a chuck in the cryostat. Immediately place the frozen tissue block onto the chuck. After the block is adherent to the chuck, place it in position onto the microtome. Proceed to cut tissue sections onto glass slides. Reference H&E slides are cut at 6µm thickness, slides for microdissection are typically cut at 8-µm, or thicker if ROI is still distinguishable at this thickness. 5. Immediately place the slides onto dry ice. Slides can be stained and dissected or stored at 80°C.
3.5.2. Paraffin Sections (see Notes 31 and 32) 1. Place tissue block onto a microtome. 2. Proceed to cut tissue sections. 3. Sections are floated on the surface of a bath containing distilled deionized water (ddH2O) at 39°C. 4. Floating sections are lifted onto glass slides. 5. Slides are baked at 60°C for 5–10 min, until the wax surrounding the tissue goes from whitish to translucent.
3.6. Staining of Tissue Sections for Microdissection 3.6.1. Mayer’s Hemalum and Eosin Staining for DNA, RNA, and Protein Recovery (see Notes 33 and 34) 1. Replace reagents frequently for optimal results, and to reduce the risk of cross contamination. 2. If protein recovery is desired from sections, add one complete miniprotease inhibitor tablet (Roche) per 10 mL of reagent, except xylene. 3. Stain each slide-mounted section as follows (see Notes 35 and 36): Reagents in order of treatment
Time
Fresh xylenes 1 (to remove paraffin) Fresh xylenes 2 (to remove paraffin) 100% ethanol 95% ethanol 70% ethanol DEPC-treated water
3 min 2 min 15 s 15 s 15 s 10 s
Molecular Profiling of Tissue Reagents in order of treatment Mayer’s hematoxylin DEPC-treated water 70% ethanol Eosin Y 95% ethanol 95% ethanol 100% ethanol 100% ethanol Fresh xylenes (for dehydration) Air dry (allow xylenes to evaporate) Sections are now ready for microdissection.
45 Time 6–30 s (optimize to tissue) 10 s 15 s 1–5 s (optimize to tissue) 15 s 15 s 15 s 15 s 60 s 2 min
3.6.2. Immunostaining Prior to Laser Microdissection (see Notes 37–39) Frozen sections are immunostained under RNase-free conditions using a rapid three-step streptavidin–biotin technique followed by dehydration. 1. Cut 8-µm thick serial sections of snap-frozen tissue blocks on a standard cryostat with a new disposable cryostat blade. 2. Mount the tissue sections on Superfrost Plus glass slides and store immediately at 80°C. 3. Thaw the frozen sections at room temperature for 30–60 s without drying. 4. Fix by immersing immediately in cold acetone for 5 min. 5. Rinse the slides briefly in 1X phosphate-buffered saline (PBS, see recipe in Notes), pH 7.4. (see Note 40). 6. Using the DAKO Quick Staining kit, immunostain the slides by incubating the slides at room temperature with the primary and avidin-linked secondary antibodies and the horseradish peroxidase for 90–120 s each, rinsing briefly with 1X PBS between each step. 7. Develop the color with diaminobenzidine (DAB) for 3–5 min in the presence of dilute H2O2 and counterstain with Mayer’s hemalum for 15–30 s. 8. Dehydrate the sections sequentially in 70%, 95%, 100% ethanol (15 s each), and xylenes (twice for 2 min each). 9. Air-dry. 10. The immunostained sections are then ready for LCM.
3.6.3. Methylene Blue Staining (see Note 41) Reagents in order of treatment
Time
Fresh xylenes1 (to remove paraffin) Fresh xylenes 2 (to remove paraffin) 100% ethanol 95% ethanol 70% ethanol
3 min 2 min 15 s 15 s 15 s
46
Bova et al. Reagents in order of treatment
Time
DEPC-treated water Methylene blue DEPC-treated water Fresh xylenes (for dehydration if using Arcturus PixCell) Air-dry (allow xylenes to evaporate)
10 s 5–10 min 10 s 60 s 2 min
3.6.4. Methyl Green Staining (see Note 42) Reagents in order of treatment
Time
Fresh xylenes 1 (to remove paraffin) Fresh xylenes 2 (to remove paraffin) 100% ethanol 95% ethanol 70% ethanol DEPC-treated water Methyl green solution DEPC-treated water Fresh xylenes (for dehydration if using Arcturus PixCell) Air-dry (allow xylenes to evaporate)
3 min 2 min 15 s 15 s 15 s 10 s 5 min 10 s 60 s 2 min
3.6.5. Nuclear Fast Red (see Note 43) Reagents in order of treatment
Time
Fresh xylenes 1 (to remove paraffin) Fresh xylenes 2 (to remove paraffin) 100% ethanol 95% ethanol 70% ethanol DEPC-treated water Nuclear fast red solution DEPC-treated water Fresh xylenes (for dehydration if using Arcturus PixCell) Air-dry (allow xylenes to evaporate)
3 min 2 min 15 s 15 s 15 s 10 s 5–10 min 10 s 60 s 2 min
3.7. Manual Microdissection 3.7.1. Cryostat-Based Microdissection of Tissue Blocks The advantages of this cryostat-based dissection technique are that a large amount of material can be obtained for analysis in a relatively short time, with minimal required equipment and expertise needed, and easy to manage documentation. The disadvantage is that even the highest possible purity obtained using this technique is not sufficient for some experiments.
Molecular Profiling of Tissue
47
1. Mount the block to be dissected in the cryostat using OCT or other mounting medium. Be sure to surround the block with mounting media so it has a strong base. 2. Cut face section of tissue to be dissected, mount on an ordinary glass slide, stain with H&E, and cover slip. 3. Examine section microscopically with 4, 10 objectives and clearly mark unwanted areas with a lab marker. Label slide as section 1 (etc.) and record in the logbook. 4. Remove the cryostat blade, or cover with a knife guard. Lock the reciprocating block holder in place using the cryostat-locking mechanism. Place the marked glass slide against the block to identify area to be cut. Draw dots or lines directly on frozen block with the marker to indicate where cuts should be made. 5. Use new heavy-duty single-edge blade to carefully cut away unwanted tissue, and trim OCT from edge of block, leaving a wider base of OCT close to the cryostat chuck (reduces risk that block will pop off chuck after dissection) Caution: Take great care to avoid cutting yourself (see Note 44). 6. Cut another face section of the block and mark this slide as section 2. Examine to be sure that all unwanted tissue has been dissected. Repeat dissection and staining until only desired tissue remains. Estimate the percentage purity of dissected material and record in logbook. 7. Cut fifty 6-µm sections serially and allow them to stack on the cryostat cutting plate (which is kept clean). Place collected sections in a cold sterile prelabeled 50-mL conical tube using long forceps prechilled by dipping in liquid nitrogen just prior to collection. Keep everything that contacts the sections frozen to avoid tissue sticking to the forceps and to the sides of the 50-mL conical tube (see Note 45). 8. To monitor dissection progress, interval sections are mounted on glass slides and H&E stained (usually before and after each episode of dissection at 300- to 600-µm intervals). In some cases undesired areas of the frozen block are easy to identify grossly based on their appearance in the initial face section, allowing continuous visual monitoring and removal of undesired areas without the need to repeat slide mounting and staining. 9. A record of estimated purity of the resulting sample, the number of sections, and cross-sectional area dissected should be kept for reference during analysis of molecular data and for monitoring of the percentage yield in downstream DNA/RNA/protein isolation protocols (see Note 46).
3.7.2. Manual Microdissection of Tissue Sections on Glass Slides 1. Prepare stained sections 7–10 µm thick, on glass slides without cover slips. For each section to be dissected, prepare an adjacent stained cover slipped section on a glass slide as a “scout section.” Prepare a dissection station using a standard inverted microscope using a 30-gage needle on a syringe as the microdissecting tool (see Note 47). 2. If desired (test necessity of this step for tissues to be dissected), place sections for dissection in 2.5% glycerol solution. 3. While viewing the tissue through the microscope, gently scrape the cell population of interest with the needle. The dissected cells will become detached from the
48
Bova et al.
slide and form small dark clumps of tissue that can be collected on the needle by electrostatic attraction. Several small tissue fragments can be procured simultaneously. Collection of an initial fragment on the tip of the needle will assist in procuring subsequent tissue. 4. The tip of the needle with the procured tissue fragments should be carefully placed into a 1.5-mL tube containing the appropriate buffer. Gentle shaking of the tube will ensure the tissue detaches from the tip of the needle (see Notes 48–51).
3.8. Laser Microdissection of Tissue Sections on Slides 3.8.1. Arcturus LCM System The described procedure is for use with the Pixcell I or II Laser Capture Microdissection System manufactured by Arcturus Corporation (Santa Clara, CA), and assumes rudimentary knowledge of the function of the components of the instrument and the software that accompanies the instrument. The procedure can be divided into three basic steps: slide positioning, microdissecting with the laser, and collecting the microdissected cells. Additional methods information can be found at the Arcturus Engineering website (www.arctur.com). 3.8.1.1. POSITIONING
THE
SLIDE
FOR
MICRODISSECTION
Wear gloves when microdissecting to avoid contamination of LCM specimens. Clean the microscope stage and capping station with 95% ethanol before beginning the microdissection to reduce the possibility of contamination. With the stage vacuum off, place the glass slide with the section to be microdissected on the microscope stage and identify the target of interest. When viewing uncoverslipped sections for dissection, put the light diffuser piece (white membrane) in place to improve visualization. If necessary, refer to an adjacent cover slipped section for orientation. When the target of interest is identified, place the stagemotion joystick so that it is perpendicular to the tabletop and then slide the glass slide so that the target of interest is in the center of the field of view. Turn on the stage vacuum. If the slide does not cover the vacuum chuck holes when centered, move the slide as little as necessary until it just covers the holes and continue to the next step. Identify targets of interest and take “roadmap” images of targets of interest if desired using the Arcturus software. 3.8.1.2. MICRODISSECTION
Pick up a CapSure™ LCM cap from the loaded cassette module with the placement arm, moving the placement arm toward the caps to ensure that the cassette module is engaged so that the first available cap is aligned, and lifting the transport arm until the cap detaches from the base slide in the cassette module. The arm must then be positioned over the tissue, ensuring that the area to be
Molecular Profiling of Tissue
49
microdissected is still in the microscopic field of view. Finally, the arm is gently lowered so that the cap contacts the tissue section. If there are folds in the tissue, the cap may not make direct contact with the entire surface at that area and transfer efficiency is compromised. Therefore, it is advisable to inspect the tissue before placing the cap. If any tissue is mounded or folded, it is best not to place the cap over that area. Alternatively, the folded area of the tissue can be scraped off the slide using a sterile razor blade, leaving only flat portions of the tissue section. The tissue section must be dry and uncover-slipped during dissection. Next, enable the diode laser and visualize the tracking beam on the monitor as well as through the eyepieces. Once the laser is focused using the laser-tracking beam and then the focus of the tissue is adjusted using 20 objective, one is ready for dissection. Refocus the laser beam for each tissue section and slide at the 7.5-µm beam setting by turning the laser-focusing wheel until the tracking beam shows a bright spot with a well-defined edge. There should be no bright rings surrounding the central spot. There is no need to refocus the 15 µm or 30 µm beams. They are automatically calibrated once the 7.5-µm beam is focused. Adjust the laser power and pulse duration settings for the particular spot size following the manufacturer’s recommendation and as needed to obtain good tissue “wetting” by molten plastic, as indicated by clear areas formed with each laser pulse. If the edges of the area of clearing are not well delineated, check to make sure the tissue section (where the cap is placed) is flat. If this fails to correct the problem, we recommend increasing the power and/or duration gradually, testing for wetting with higher energy/duration. Move the stage with the joystick, targeting selected cells by firing the button-actuated laser. When targeting is complete, lift the placement arm and inspect the area in which the laser was fired for removal of cells. Dissected areas should show near total removal of tissue. Quality of dissection should be checked occasionally by releasing the vacuum slide holder, moving the slide so that a clean area without tissue is in the microscopic field of view, lowering the cap to the slide, and scanning the surface of the cap. The microdissected tissue should be visible on the cap surface. If this is not the case, there are several explanations and potential remedies (see Notes 52–61). Avoid lifting and lowering the cap after firing the laser and capturing tissue. Already-captured tissue attached to the LCM cap will be placed on the histologic section away from its point of origin, resulting in a layering effect, which can limit contact of the cap with the tissue and compromise the effectiveness of LCM. However, before proceeding with a largescale microdissection, we recommend that the user check for quality of dissection after several pulses. This should not interfere with subsequent dissection and the cap can be placed back in the same position. In addition, dense, dark, or thick samples may occlude the tracking beam. If this occurs, the intensity of
50
Bova et al.
the tracking beam should be increased. Once LCM is achieved successfully with initial test pulses, the remainder of the desired cells can be efficiently microdissected. 3.8.1.3. COLLECTION OF THE MICRODISSECTED CELLS After the microdissection is completed, the placement arm with the cap is moved to the “unload platform” and, using the cap insertion tool, the cap is removed. Cells that were not selected for capture may stick to the surface of the cap at random, and it is important to remove this unwanted tissue. This can be accomplished by using CapSure Pads®, purchased from the manufacturer, which have a sticky surface. A less costly alternative to the CapSure Pad is to use the sticky surface of Post-it® Notes, which can be used after the cap has been removed from the unload platform (gently touch the cap three separate times to clean areas of the Post-it Note, then view the cap microscopically to ensure that all loose material is removed). To obtain successful removal of unwanted cells, one may have to repeat this two or three times. When removal of unwanted material is complete, use the cap insertion tool to place the LCM cap onto a Perkin Elmer/Applied Biosystems GeneAmp 500-µL thin-walled PCR reaction tube (cat. no. N8010611) containing the appropriate amount of lysis buffer. The buffer used will depend upon the analyte, for example, RNA, DNA, or protein, and the method of analysis. Do not seat the LCM cap fully on the microcentrifuge tube, as this will cause undue stretching and possible reagent leakage. Leave a 2-mm gap between the cap top and the rim of the microcentrifuge tube to avoid this problem. For best results, use only Eppendorf tubes, as they are well-matched to the size of the lower portion of the CapSure™ caps. Invert the tube so that lysis buffer contacts the cap surface, and place on ice or refrigerate until the microdissection session is over, to preserve the analyte. If regular CapSure™ caps are used (not HS caps described later), an exposed blank LCM cap control is recommended for each experiment to ensure that nonspecific transfer is not occurring during microdissection. This is best performed by placing an LCM cap on the tissue section being dissected and aiming and firing the laser at regions where there are no cells or structures present, for example, lumens of large vessels, cystic structures, and so forth (alternatively one can place a portion of the LCM cap “off” the tissue and target this region). The cap should be processed through the buffer and analysis methodology applied. This serves as a negative control.
3.8.2. LCM Method for Epithelial Cell Enrichment Maitra et al. have described an enhancement to LCM that can be used to enrich samples for neoplastic cells (45). This method, called “epithelial aggregate separation and isolation,” or EASI, is applicable to fresh tissues only.
Molecular Profiling of Tissue
51
1. The tissue is sectioned and gently scraped with the edge of a plain, uncharged, microscope glass slide. 2. The material adherent to this slide is then spread evenly onto the surface of a second uncharged slide. 3. Slides are immediately fixed in 95% methanol for 2 min and stained with H&E. 4. Epithelial aggregates on these slides can then be microdissected using an LCM. Alternatively, manual methods can also be used. 5. The advantages of this technique are that the discreteness of the epithelial clusters helps reduce background inflammatory and stromal elements and that large areas can be sampled (46).
3.8.3. LCM Dissection of Single or Small Numbers of Cells Arcturus Engineering has developed a line of related products specially designed for high sensitivity capture and extraction of a single cell or a small number of cells. There are three key components of the system: (1) a preparation strip that flattens the tissue section and removes loose debris, (2) a highsensitivity transfer cap (CapSure HS™ cap) that keeps the cap surface out of contact with the untargeted section surface, and (3) a low-volume reaction chamber that fits onto the high sensitivity transfer caps and accepts a low volume of lysis or digestion buffer while sealing out any nonselected material from the captured cells. The surface coated with polymer contacts only the tissue in the area that the laser is fired, thus reducing nonspecific transfer of tissue. Using the HS caps, it is preferable to capture cells as close to the center of the cap as possible. Unlike basic LCM using standard caps, the HS caps can be repositioned as often as needed to keep the targets toward the center of the cap, the cap surface does not contact the tissue except at the area that the laser is fired. It is important to stay within the capture ring because areas outside the ring will be excluded from the low-volume reaction tube. After the intended microdissection is completed, the HS cap is placed on the unload platform, picked with the cap insertion tool, and placed into the alignment tray. The specialized low-volume reaction chamber is then positioned over the cap. The chamber has a port for insertion of the lysis/digestion buffer. Ten microliters of the desired buffer is delivered into the fill port, which is covered securely with a thin-walled Gene-Amp PCR tube. 3.8.4. PALM Microbeam System Follow the manufacturer’s instructions for general use of the instrument. 3.8.5. Leica LMD System As for the PALM system, follow the manufacturer’s instructions for general use of the instrument.
52
Bova et al.
3.9. Isolation of Analyte from Microdissected Material 3.9.1. Isolation of DNA 3.9.1.1. KIT-BASED METHOD (BASED ON DNEASY KIT BY QIAGEN) (This protocol applies to all types of tissue material microdissected from glass slides. Ignore references to LCM caps if material was manually dissected or dissected using other devices). 1. Remove the LCM Cap using the cap-removal tool, add 45 µL of proprietary Buffer ATL and 5 µL of proteinase K mixture (45:5) per cap, and replace the cap, taking care not to fully seat the cap (causes stretching and later leakage). 2. Mix the solution by vortexing briefly, flick down the tube in an upside down position so that the cells on the cap make contact with the solution, vortex five times at 1-s intervals, and incubate at 55°C for 3 h in a Hybaid oven or other oven that provides gentle rocking motion. Vortex again three times. 3. Centrifuge the tubes at 5000 rpm (2700g) for 5 min at room temperature in a microcentrifuge. 4. Heat tubes at 95°C at 600 rpm using a Thermomixer R (Eppendorf) or other heating device for 10 min to inactivate proteinase K (if sample is to be used directly for PCR). 5. Add 2 µL of 0.5 mg/mL of RNase A, vortex briefly, and incubate at 37°C for 30 min. 6. Combine all samples from each case (if multiple caps per unique case) into a fresh microcentrifuge tube, add 150 µL of proprietary buffer AL, immediately mix by vortexing, and incubate at 70°C for 10 min. (White precipitate may form after addition of AL buffer, but will dissolve while incubating at 70°C.) 7. Add 150 µL of room temperature 100% ethanol and mix thoroughly by vortexing. 8. Transfer the sample into a DNeasy mini-column sitting in the provided 2-mL collection tube and centrifuge at 8000 rpm (6800 g) for 1 min at room temperature. 9. Transfer the column into a fresh collection tube, add 500 µL of buffer AW1, and centrifuge for 1 min at the same speed. 10. Transfer the column into a fresh collection tube, add 500 µL of buffer AW2, and centrifuge for 3 min at 17,900g (13,000 rpm with Eppendorf 5417C microcentrifuge, using standard fixed angle rotor) to dry the column. 11. To elute DNA from the column, transfer the column into a 1.5-mL microcentrifuge tube, add 200 µL of buffer AE, and centrifuge for 1 min at 10,600g (10,000 rpm with Eppendorf 5417C microcentrifuge, using standard fixed angle rotor). To increase recovery of the DNA, use the buffer containing the DNA after the first elution to elute the column one more time. 12. Add 1 µL of 20 mg/mL glycogen, 20 µL of 3 M sodium acetate, pH 5.2, 440 µL of 100% ethanol, mix by vortexing and incubate at 20°C for at least 1 h (we are unable to find a reference supporting an optimal time for precipitation). 13. Centrifuge the sample at 13,000 rpm at +4°C for 30 min, aspirate the supernatant, and resuspend the pellet in 1 mL of 70% ethanol. Centrifuge the tube for 10 min, aspirate ethanol, and air-dry the pellet for 5–10 min.
Molecular Profiling of Tissue
53
14. Resuspend each DNA pellet in 5 µL of loTE buffer, pH 7.4 (see recipe below), combine all the sample in one tube and use 0.5 µL of the sample to quantitate using pico-green (or other) assay.
3.9.1.2. SDS/PHENOL–CHLOROFORM EXTRACTION 1. Transfer samples to a first SST tube. (This can be performed in 1.5-mL microcentrifuge tubes if volume requirements are smaller.) 2. Add 6 mL of phenol–chloroform–isoamyl alcohol (25:24:1) to each tube (see Note 62). 3. Vortex five times, setting 8. 4. Centrifuge at 2000g (STH-750 rotor), 20 min, room temperature. 5. Transfer the top (aqueous) layer to a second SST tube. 6. Add 6 mL of phenol–chloroform–isoamyl alcohol (25:24:1) to each tube. 7. Vortex five times (VWR vortex Genie, setting 8). 8. Centrifuge at 2000g, 20 min, room temperature. 9. Transfer the aqueous layer to a third SST tube. 10. Add 6 mL of phenol–chloroform–isoamyl alcohol (25:24:1) to each tube. 11. Vortex five times, setting 8. 12. Centrifuge at 2000g, 20 min, room temperature. 13. Transfer the aqueous layer to a 50-mL first sterile blue top conical tube. 14. Add 5.75 mL of chloroform–0.25 mL of isoamyl alcohol to each tube. 15. Vortex five times, setting 8. 16. Centrifuge at 2000g (3100 rpm in Sorvall Super T21 Refrigerated Centrifuge STH750 rotor, swing buckets), 20 min, at room temperature. 17. Transfer the aqueous phase to a second 50-mL conical tube. 18. Add 2.5 mL of ammonium acetate, mix gently. 19. Add 15 mL of “frozen” 100% ethanol, mix gently. 20. Store in 20°C, overnight. 21. Centrifuge at 3800g (4300 rpm in Sorvall Super T21 Refrigerated Centrifuge STH750 rotor, swing buckets), 30 min, 4°C. 22. Discard the supernatant gently into a waste flask, preserving pellet. 23. Wash pellet with 10 mL of 70% EtOH (room temperature), set for 10 min. 24. Centrifuge at 3800g, 10 min, 4°C. 25. Discard the supernatant into the waste flask. 26. Wash with 10 mL of 70% EtOH (room temperature), set for 10 min. 27. Centrifuge at 4300 rpm (STH-750 rotor), 10 min, 4°C. 28. Gently discard the supernatant into the waste flask. 29. Inverted at a 30° angle, air dry for 25 ± min (do not overdry or DNA will not go into solution). 30. Resuspend in 200–1000 µL or more of LoTE (volume dependent on predicted yield and desired storage concentration), and store at 4°C, overnight. 31. Gently mix DNA and transfer DNA to a microcentrifuge tube or other storage device.
54
Bova et al.
3.9.2. Isolation of RNA (see Notes 63 and 64) We have obtained high-quality total RNA from microdissected samples using a phenol–chloroform-based approach, and also using the PicoPure RNA isolation kit (Arcturus), a column-based approach that is optimized for use with microdissected cells. Several other kit-based methods for isolation of total RNA from small tissue samples are currently on market from companies such as Ambion, Qiagen, and others (47), but we have not had the opportunity to test them. The phenol–chloroform and Arcturus kit-based methods are covered here. Prior to performing total RNA extraction of any kind, it is wise to clean all pipettors with a RNase removal product such as RNase AWAY™ (Molecular BioProducts) or RNase ZAP™ (Ambion, Austin, TX). 3.9.2.1. KIT-BASED METHOD (PICOPURE RNA ISOLATION KIT
BY
ARCTURUS)
1. Dispense extraction buffer (XB) and incubate as follows: a. Pipet 50 µL of well-mixed XB into a Perkin Elmer/Applied Biosystems GeneAmp 500-µL thin-walled PCR reaction tube (cat. no. N8010611) 0.5-mL microcentrifuge tube. This is the only type of tube that Arcturus currently recommends using with the original non-LCM caps currently, as other tubes vary in diameter and may be more prone to leakage. b. Place CapSure LCM cap onto the microcentrifuge tube using an LCM cap insertion tool. c. Invert the CapSure Cap-microcentrifuge tube assembly. Tap the microcentrifuge tube to ensure all XB is covering the CapSure LCM cap. d. Incubate assembly for 30 min at 42°C. 2. Centrifuge assembly at 800g (approx 3500 rpm) for 2 min to collect cell extract. 3. Remove the LCM cap and save the tube with the extract in it. 4. Proceed with the RNA isolation or freeze the extract at 80°C. 5. Precondition the RNA purification column: a. Pipet 250 µL conditioning buffer (CB) onto the purification column filter membrane. b. Incubate the column with conditioning buffer at RT for 5 min. c. Centrifuge the purification column in the provided collection tube at 16,000g (approx 14,000 rpm) for 1 min. 6. Pipet 50 µL of 70% ethanol (EtOH) into the cell extract from step 4. Mix well by pipetting. Do not centrifuge. 7. Pipet the cell extract and EtOH mixture into the preconditioned purification column. The combined volume will be approx 100 µL. 8. Centrifuge for 2 min at 100g (approx 1200 rpm), immediately followed by 16,000g (14,000 rpm) for 30 s. 9. Pipet 100 µL of wash buffer 1 (W1) into the purification column and centrifuge for 1 min at 16,000g (approx 14,000 rpm). A DNase step may be performed at this point: a. DNase treatment (Qiagen RNase-free DNase Set, cat. no. 79254).
Molecular Profiling of Tissue
10. 11.
12. 13. 14.
55
b. Pipet 5 µL of DNase I stock solution into 35 µL of buffer RDD. Mix by inverting. c. Pipet the 40 µL of DNase incubation mix directly onto the purification column membrane. d. Incubate at RT for 15 min. e. Pipet 40 µL of PicoPure RNA Kit wash buffer 1 (W1) into the purification column. Centrifuge at 8000g (approx 10,000 rpm) for 15 s. Pipet 100 µL wash buffer 2 (W2) into the purification column and centrifuge for 1 min at 16,000g (approx 14,000 rpm). Pipet another 100 µL of wash buffer 2 into the purification column and spin for 2 min at 16,000g (14,000 rpm). Check the column for any residual wash buffer. If present, centrifuge for an additional minute. Transfer the purification column to a new microcentrifuge tube provided in the kit. Pipet 11 µL (maximum is 30 µL) elution buffer (EB) directly onto the membrane of the purification column. Incubate the purification column for 1 min at RT and then centrifuge at 1000g (approx 3800 rpm) for 1 min and then maximum speed for 1 min to elute RNA. Use RNA immediately or store at 80°C until use.
3.9.2.2. PHENOL–CHLOROFORM-BASED RNA EXTRACTION METHOD
This method is modified from Chomczynski and Sacchi (48): 1. Ensure an RNase-free environment, including reducing likelihood that pippetors contain RNases as described in the preceding. 2. Dissolve tissue obtained by microdissection by placing in 200 µL of GITC denaturing solution. Invert several times over the course of 2 min to digest the tissue off the cap. If tissue does not dissolve completely, use an appropriate disaggregating device to homogenize. 3. For specimens that originated from paraffin-embedded tissue, it is helpful to include an incubation step (with the tube inverted) of 20 min at 60°C to further liberate the RNA from the tissue. 4. Remove the solution from the reagent tube and replace it in a sturdy RNase-free 1.5-mL microcentrifuge tube (cap must seal tightly, and must be able to withstand centrifugation at 10,000g; test this if necessary). 5. Add 20 µL (0.1X volume) of 2 M sodium acetate, pH 4.0. 6. Add 220 µL (1X volume) of water-saturated phenol. 7. Add 60 µL (0.3X volume) of chloroform–isoamyl alcohol. 8. Shake the tube vigorously for 15 s. 9. Place on wet ice for 15 min. 10. Centrifuge at 10,000g for 30 min at 4°C to separate the aqueous and organic phases. 11. Transfer the upper aqueous layer to a fresh tube (see Note 65). 12. Add to aqueous layer, 1–2 µL of glycogen (10 mg/mL) and 200–300 µL of cold isopropanol (i.e., equal volume). Glycogen facilitates visualization of the pellet, which can be problematic when using small amounts of RNA.
56
Bova et al.
13. Place samples at 80°C (some use 20°C) for at least 30 min. It may be left overnight. 14. Before centrifuging, the tubes may need to be thawed slightly if they have solidified during the isopropanol precipitation. 15. Centrifuge for 30 min at 4°C with cap hinges pointing outward so that the location of the pellet can be better predicted. 16. Remove the supernatant and wash with 300 µL of cold 70% ethanol. Add the alcohol and centrifuge for 5 min at 4°C. 17. Remove the supernatant. 18. Let the pellet air dry on ice to remove any residual ethanol. Overdrying prevents the pellet from resuspending easily (drying is not necessary however if the RNA is to be resuspended in formamide). 19. The pellet may be stored at 80°C until use or proceed to DNase treatment (see below). 20. Dissolve the RNA pellet in 10–20 µL DEPC-treated water and store at 80°C, or dissolve in a similar amount of deionized formamide by passing the solution a few times through a pipett tip and store at 20°C or 80°C.
3.9.2.3. DNASE TREATMENT OF TOTAL RNA (ALTERNATE TO METHOD IN SUBHEADING 3.9.2.1., STEPS 9A–9E) 1. DNase treatment is highly recommended for microdissected cells. Genomic DNA contamination is often problematic with these samples, possibly due to the small DNA fragments that are created during tissue processing and are difficult to purify from RNA. 2. To an RNA pellet, add 15 µL of DEPC-treated water and 1 µL (20 U/µL) RNase inhibitor (Perkin Elmer). 3. Gently mix by flicking until the pellet is dissolved. 4. Pulse spin on microcentrifuge. 5. Add 2 µL of 10X DNase buffer (GenHunter) and 2 µL (10 U/µL) DNase I (GenHunter; 20 U total). 6. Incubate at 37°C for 2 h. 7. Reextract RNA by adding: a. 2 µL 2 M sodium acetate, pH 4.0. b. 22 µL of water-saturated phenol. c. 6 µL of chloroform-isoamyl alcohol. 8. Shake vigorously for 15 s. 9. Place on wet ice for 5 min. 10. Centrifuge at 10,000g for 10 min at 4°C. 11. Transfer upper layer to a fresh tube. 12. Continue with RNA extraction from step 12 in Subheading 3.9.2.2., adjusting the volume of isopropanol accordingly.
4. Notes 1. Commercial preparations of 10% neutral buffered formalin contain 1–10% methanol (13). This is added to inhibit polymerization of formaldehyde, which can grad-
Molecular Profiling of Tissue
2. 3.
4.
5.
6.
7. 8.
9. 10.
11.
12.
57
ually precipitate out of solution. Commercially prepared NBF is not used for electron microscopy because the added methanol can cause coagulation of proteins. Similarly, molecular laboratories wishing to avoid protein coagulation should consider making formaldehyde solutions from scratch for this reason. Formaldehyde is a carcinogen in rodents and can induce allergies in humans. Use appropriate means to avoid inhalation and contact exposure under all circumstances. A 37% formaldehyde solution can be prepared by dissolving 37 g of paraformaldehyde powder (e.g., Sigma) in 100 mL of water in a fume hood. Heat to 70°C (no higher—to avoid decomposition) to allow paraformaldehyde to dissolve and dissociate into a mixture of polymeric and nonpolymeric formaldehyde, and cool to room temperature. Note that many laboratories call any fixative made directly from paraformaldehyde powder “paraformaldehyde,” which is something of a misnomer because all formaldehyde solutions are a mixture of formaldehyde and paraformaldehyde. The term “formalin” should be abandoned because of the confusion it causes, but nonetheless is in common usage. “10% formalin” by tradition is what is obtained when 37% formaldehyde is diluted 1:10. The correct term is 3.7% formaldehyde. The resulting pH of this 10% neutral buffered formalin solution is approx 6.8, and is hypotonic at approx 165 mosM. Isotonic 3.7% formaldehyde can be prepared by increasing sodium phosphate monobasic to 18.6 g, eliminating sodium phosphate dibasic, and adding 4.2 g of NaOH. This is known as Modified Millonig Formalin, and isotonic (310 mosM), with a pH of 7.2–7.4. It can be used both for histologic preparation of tissues for light and electron microscopy (12). Note that formaldehyde itself is reported not to be osmotically active (49). Molecular studies performed in our laboratories have shown that that both 70% ethanol and a 4.2:1.8:2.0 (v:v:v) of ethanol–methanol–water provide excellent histologic detail, and markedly improved DNA (7), RNA (7), and protein quality (4) compared to standard 10% NBF fixation. 70% Ethanol is easy to prepare, but may be open to abuse (human consumption) in some laboratory environments. The final concentration of ethanol in this mixture is 52.5%, and the final concentration of methanol is 25%. The mixture needs to be marked as toxic, as it can cause blindness if ingested. Methanol penetrates tissues measurably faster than ethanol in our studies (G. S. B., unpublished data). In a blinded survey of practicing pathologists, kidney and prostate histology were better with this ethanol–methanol mixture than with 70% ethanol, although both were rated acceptable (7). Complete proteinase K digestion is critical to downstream DNA isolation efficiency. If proteins are inaccessible to proteinase K because of residual paraffin, or other hydrophobic compounds, DNA recovery will be affected. Even when proteins are accessible to the enzyme, incomplete digestion will also markedly reduce yield. Well-digested samples are clearer, without clumps, and flow freely. Add more proteinase K and incubate longer if digestion appears incomplete.
58
13.
14.
15. 16.
17.
18.
19.
20.
21.
Bova et al. Proteinase K functions at least between 37°C and 55°C, and may function at room temperature, although we have not tested this. We use 48°C, and this works. Direct immersion of tissue in liquid nitrogen often leads to poor histological preservation because freezing is not rapid enough to prevent formation of ice crystals large enough to damage cell morphology, and should be avoided. A jacket of insulating nitrogen gas forms around the tissue after immersion, slowing heat transfer. Two methods that provide more rapid freezing and better histology are discussed. Tissues thicker than roughly 2 mm tend to crack, probably because of stresses caused by differential expansion of the tissue as the wave of freezing passes through the tissue from outside to inside. Multiple blocks can be frozen simultaneously if the basket is large enough If freezing in OCT is desired, plastic cryomolds can be purchased for this purpose. Tissues frozen at 80°C or below for long periods with no covering oil or other material become severely dessicated (“freezer burn”), which, however, may also occur in samples stored in OCT over longer periods. Instrumedics (Hackensack, NJ) sells oil for covering tissues to prevent desiccation, but we do not have independent confirmation that this technique works. This is our preferred method for rapid freezing, because it does not require use of flammable liquid, and in our hands (G. S. B.) provides histological quality as good as standard H&E-stained formalin fixed paraffin-embedded sections. Only one block can be frozen at a time (taking 1 min or so per block) using this device and if many blocks need to be frozen simultaneously, this method can be prohibitively slow. The device can be used without an embedding compound, but may cause the tissue to stick to the piston or to the base. We always use an embedding compound, and cut clear plastic from a report cover to approx 3 cm 3 cm, cover this with 4 mm of embedding compound, freeze enough to make firm (but not crack) and use this as a base on which to place tissue to be frozen. A layer of embedding media is put over the entire top and sides of the tissue, and the cold metal piston is then lowered onto the tissue. The two greatest advantages of disaggregation of tissues are that living cells can be obtained and that whole cells are recoverable. The disadvantages are that tissue morphology is lost, and a reliable molecular method (usually antibody-specific staining) is required to identify the cells of interest after disaggregation. The same basic caveats apply to cytologic specimens as to histologic sections: alcohol fixation is preferred, especially for RNA analysis. a. Smeared cells should not be allowed to dry on the slide prior to fixation, particularly for those using LCM (difficult to remove cells from slide surface). Fixed and stained cells should be adequately dehydrated prior to attempting LCM. b. We prefer to prepare cytologic smears with a hemocytometer cover (rather than using a separate glass slide) because the width is slightly less than that of the standard glass microscopic slide and the resulting smear (and cells) are not spread to the edge (or off) of the slide, where they are difficult or impossible to stain and microdissect.
Molecular Profiling of Tissue
59
22. As for direct smears of liquids containing suspended cells, smeared cells should not be allowed to dry on the slide prior to fixation, particularly for those using LCM (difficult to remove cells from slide surface). Fixed and stained cells should be adequately dehydrated prior to attempting LCM. 23. This technique works best when target cells are less adherent to tissue than nontarget cells. Most epithelia are less adherent than underlying stroma for example. 24. Density and size-based separation techniques traditionally has been used to separate blood mononuclear cells, cell organelles, and microorganisms, but has not been used routinely for disaggregated solid tissue cell separation, because of the time the cells spend unfixed and because of the loss of tissue morphology. Theoretically, these problems are surmountable through isosmotic fixation of cells, and if the cells are phenotypically distinct based on a specific antibody or other marker, this method may become more useful as specific phenotypic immunomarkers and other biomarkers are identified. An advantage of this method when compared to tissue sectioning is that whole cells are obtained. 25. A large amount of pertinent information about magnetic bead-based separations is contained in the Dynal website (www.dynalbiotech.com). Other manufacturers also offer bead-based separation systems. 26. Cytospin® preparations can be used for any cytologic sample but are preferred for samples of low cellularity. Another alternative to handling samples of low cellularity is to centrifuge the sample, pour off the supernatant, and make a direct smear from the sediment concentrated in a low volume of liquid. Particularly bloody specimens may benefit from Ficoll separation. If such a separation is used, to avoid RNA, DNA, or protein degradation, the cytologic samples should be processed and fixed in 95% ethanol before processing. 27. This method was originally developed for electron microscopy, and can provide improved histology over standard frozen section techniques, and theoretically will provide biomolecular integrity intermediate between frozen and formalin-fixed, paraffin-embedded tissue. We are not aware of its use in molecular studies to date, but it warrants testing for molecular profiling studies requiring excellent histology, moderate biomolecule quality, and avoidance of the relatively high temperature exposures of standard tissue processing. It is also notable that since proteins should not be denatured using this technique, and because heating (antigen retrieval) can reverse formaldehyde crosslinks, native proteins may be obtainable using this technique. 28. It is notable that an effort is currently underway by Sakura Finetek (Torrance, CA) and the University of Miami School of Medicine to replace current standard surgical pathology tissue processing methods with the goal of reducing time and to increase the molecular preservation and analyzability of samples (50). 29. Frozen sections allowed to come to room temperature and dry prior to staining will usually adhere too strongly to the slide, preventing effective LCM. Other types of microdissection are not affected by drying. 30. If frozen tissue sections are to be used for RNA analysis, it is essential that they be stored at 80°C and used within 2–3 mo of sectioning. This conclusion is based
60
31.
32.
33.
34.
35.
36.
37.
38.
Bova et al. an anecdotal experience in which tissue microdissected from frozen sections on slides stored for 2 yr did not yield high quality RNA (further study may be needed). If slides are to be stored in a slide file, allowed to cool first. Storage in slide boxes with space between slides is probably optimal to reduce risk of damage to sections and contamination, but this has not been studied to our knowledge. The need for DEPC-treated water in the section-floating bath to eliminate RNase exposure and/or use of protease inhibitors in the water bath to eliminate protease exposure has not been tested to our knowledge. Immunostaining of paraffin sections floated in ordinary tap water has been the rule for several decades, so it is likely that paraffin inhibits significant protein degradation, but depending on your applications, testing under controlled conditions may be warranted for potential effects of proteases, especially on low-abundance proteins. Use the least amount of Mayer’s hemalum (MH) and eosin necessary for visualization of the cells of interest. Recovery of DNA, RNA, and protein is improved as MH and eosin content decreases. Decreased stain content can be achieved by decreasing time of exposure or by decreasing concentration of stain used. Decreased concentration (we use 10% of normal) provides greater control if staining is done manually. Minimized MH intensity in sections also allows better visualization during microdissection, because MH-stained areas appear much darker than normal when no coverslip is in place. MH-stained areas in “normally” MH&E stained tissue sections appear black when no cover slip is present because of light scattering at the tissue–air interface (scattering occurs because of refractive index mismatch between air and tissue). Poor LCM (Arcturus PixCell device) transfers will occur if sections are not fully dehydrated and xylene treated. Be sure that the 100% ethanol used are fresh to ensure dehydration. Xylenes should be changed when cloudy. The final xylene step is necessary for successful Arcturus LCM, but may not be necessary for other types of microdissection. Xylene treatment appears to remove the alcohols more effectively, allowing the LCM EVA plastic to adhere better to the tissue, although this is not proven. When using membrane-coated slides (with PALM, Leica devices) with xylene, test blank membrane-coated slides in xylenes prior to staining sections. Membrane formulations are changing, and some membranes may dissolve in xylenes. The mRNA recovered from tissue with a short blood perfusion-free interval and that is rapidly frozen and immunostained is generally of high quality. Single-step PCR allows amplification of fragments of more than 600 bp from both housekeeping genes, for example, -actin, as well as cell-specific messages, for example, CD4 or CD19, using cDNA derived from less than 500 immunostained, microdissected cells (NCI Laboratory results). For primary antibodies other than those included in a commercial kit such as in DAKO Quick Staining kit, the dilutions should be determined individually. Add placental RNase inhibitor to the primary antibody and the DAB solution in a concentration of 200–400 U/mL. All solutions are prepared with DEPC-treated water.
Molecular Profiling of Tissue
61
39. If RNA is to be isolated from dissected material, it is essential that microdissection proceed immediately after slide preparation since significant RNA degradation may occur in fully dehydrated tissue sections after just one hour at room temperature. In addition, captured cells should be extracted with GITC (guanidinium isothiocyanate) buffer as soon as possible. 40. PBS recipe: a. Dissolve 8 g of NaCl, 0.2 g of KCl, 1.44 g of Na2PO4, and 0.24 g of KH2PO4 in 800 mL of distilled H2O. b. Adjust pH to 7.4 with HCl. c. Adjust volume to 1 L with additional distilled H2O. d. Filter sterilize, or sterilize by autoclaving. 41. Stains nuclei dark blue, stains cytoplasm where there are concentrations of RNA. As stated in the introduction, methylene blue staining was associated with poor RNA recovery in one non-peer-reviewed but credible study (16). 42. Stains DNA and RNA greenish blue. Provided best RNA recovery in the same study mentioned earlier (16). 43. Stains nuclei dark red, cytoplasm lighter red. Performed as well as H&E for RNA recovery in the same study (16). 44. Be sure to trim away OCT or other mounting medium as much as possible as it can interfere with downstream protocols. 45. Dissected sections can be retained in 50-mL conical tubes at 80°C for long periods prior to further processing. 46. If sections need to be placed directly in digestion buffer for DNA, RNA, or protein isolation, carefully place frozen sections in the bottom of the tube (50-mL conical tube for large volume of sections, or microcentrifuge tube for small volume of sections), and use gentle mechanical rocking to get all the sections to go into solution. 47. If no micromanipulator arm is available, the dissector should prop his or her elbow on a solid surface adjacent to and at the same height as the stage of the microscope to stabilize the dissecting hand. It is helpful to rest the ulnar aspect of the dissecting hand on the stage of the microscope and move the needle into the microscopic field, a few millimeters above the tissue. In this way, the dissecting arm and hand can be rested on solid support surfaces. 48. Pressing down on the shaft of the syringe to inject an air bubble into the extraction solution helps to detach the tissue from the needle and prevents any fragments from remaining lodged in the barrel of the needle. 49. Placement of frozen tissue sections directly on agarose coated slides can be helpful in maintaining enzyme stability if the investigator wishes to recover proteins in native form from a frozen section. In addition, agarose gels can be prepared or soaked in custom buffers that will bathe the frozen section prior to and during the microdissection, for example, pH, salt concentration, proteinase inhibitors, and so forth, can be varied specifically for a given enzyme. Some members of the NCI group also prefer to use the agarose-coated slide microdissection approach for recovery of mRNA.
62
Bova et al.
50. Slides for microdissection are prepared by placing 200 µL of warm agarose on standard uncoated glass slides, covering with a glass slip, and allowing the gel to polymerize. The glass slip is removed from the slide and the frozen tissue section is immediately placed onto the agarose gel. For best results, the freshly cut section should be transferred directly from the cryostat to the agarose-coated slide. 51. The dissector may find it easier to “tease” the tissue apart since the tissue remains bathed in the fluid from the gel and can be gently pulled apart. The tissue will also separate along tissue planes, for example, stroma and epithelium will easily separate from each other. The dissected tissue can be gently picked up from the slide or, alternatively, the dissector can use a needle to physically cut the agarose and procure both the agarose and the tissue fragment together. 52. Ensure that the slide is completely free of xylene. Use a Cleantex Microduster (MG Chemicals, others) or other drying mechanism if necessary to dry the slide thoroughly. 53. Ensure that tissue sections are flat. Wrinkles can be shaved off with sterile razor blades. The section should be dipped in xylene after shaving the wrinkles to ensure that no contaminating debris remains on the section. 54. Ensure that silver metal weight on cap support arm is resting freely on slide. Add additional weight temporarily. 55. Refocus the laser beam. Increase laser power and/or pulse duration. 56. Change the cap. Not all caps perform equally. Cap shelf-life may be important (but no good data on this is readily available). We recommend buying relatively small numbers of caps so that one’s stock remains relatively new. 57. Repeat dehydration of the specimen with fresh xylenes. Submerge for 1 min or more and allow drying in hood for 1–5 min. If LCM is still not successful, pass the slides through 95% alcohol for 30 s twice, absolute alcohol for 30 s twice, and then xylenes for 1–5 min. 58. Cut a new tissue section onto a new glass slide, ensuring that the frozen sections or cytologic specimens have not been allowed to dry on the slide prior to fixation. For formalin-fixed sections, either do not bake or decrease the baking time. 59. Try a different brand or type of glass slide. 60. Check lab humidity. Under highly humid conditions, water may be transferred to the tissue rapidly, reducing EVA adhesion to tissue. Attempt to lower humidity if this could be a problem. 61. If still not successful, talk with other researchers to continue to troubleshoot the problem. In addition, Arcturus Engineering technical support (easily available on the WWW) has been very helpful in many situations. 62. Phenol–chloroform–isoamyl alcohol (25:24:1) can be prepared in the laboratory, but take great care to handle phenol only in a fume hood and with appropriate protective gear, as it can cause severe burns. Phenol metls at 41°C and boils at 61°C. Phenol must be appropriately pH-buffered prior to mixing. 63. It is estimated that a typical mammalian cell contains 2 pg of total RNA per cell, therefore, to achieve 5 µg of total RNA, the lower limit for some expression arrays,
Molecular Profiling of Tissue
63
will require the microdissection of 2.5 million cells, a daunting task. Therefore, some authors have advocated amplification of RNA or resultant cDNA prior to hybridization with these larger arrays, even though this may introduce some degree of amplification bias. 64. The duration of the actual microdissection session on each frozen section should be limited to 15–30 min for optimal RNA preservation. Samples for protein analysis are also best processed as for RNA analysis, but reagents should include protease inhibitors. 65. If any of the lower organic phase was accidentally transferred during RNA isolation and may be contaminating the aqueous phase, this will interfere with the subsequent isopropanol precipitation. To remove any residual organics from the aqueous layer, add one volume of 100% chloroform, mix well, and centrifuge for 10 min at 4°C to again separate the aqueous and organic phases. Transfer the upper layer to a new tube.
Acknowledgments Special thanks to Drs. Eun-Chung Park, Douglas P. Clark, and Anirban Maitra for critical reading and suggestions for improvement of this manuscript, and to Steven H. Chen for technical assistance in preparation of the manuscript. Portions of the text are adapted from Best and Emmert-Buck (3) with permission of the publisher Future Drugs Ltd. (London, England). Figures 2–5 are adapted from Eltoum, Siegal and Frost (1) with permission of Lippincott Williams & Wilkins (Philadelphia). References 1. 1 Eltoum, I. A., Siegal, G. P., and Frost, A. R. (2002) Microdissection of histologic sections: Past, present, and future. Adv. Anat. Pathol. 9, 316–322. 2. 2 Srinivasan, M., Sedmak, D., and Jewell, S. (2002) Effect of fixatives and tissue processing on the content and integrity of nucleic acids. Am. J. Pathol. 161, 1961–1971. 3. Best, C. J. and Emmert-Buck, M. R. (2001) Molecular profiling of tissue samples using laser capture microdissection. Expert. Rev. Mol. Diagn. 1, 53–60. 4. 4 Ahram, M., Flaig, M. J., Gillespie, J. W., et al. (2003) Evaluation of ethanol-fixed, paraffin-embedded tissues for proteomic applications. Proteomics 3, 413–421. 5. 5 Englert, C. R., Baibakov, G. V., and Emmert-Buck, M. R. (2000) Layered expression scanning: Rapid molecular profiling of tumor samples. Cancer Res. 60, 1526–1530. 6. 6 Fend, F., Emmert-Buck, M. R., Chuaqui, R., et al. (1999) Immuno-LCM: Laser capture microdissection of immunostained frozen sections for mRNA analysis. Am. J. Pathol. 154, 61–66. 7. 7 Gillespie, J. W., Best, C. J., Bichsel, V. E., et al. (2002) Evaluation of non-formalin tissue fixation for molecular profiling studies. Am. J. Pathol. 160, 449–457. 8. 8 Gillespie, J. W., Ahram, M., Best, C. J., et al. (2001) The role of tissue microdissection in cancer research. Cancer J. 7, 32–39.
64
Bova et al.
9. 9 Ornstein, D. K., Gillespie, J. W., Paweletz, C. P., et al. (2000) Proteomic analysis of laser capture microdissected human prostate cancer and in vitro prostate cell lines. Electrophoresis 21, 2235–2242. 10. Simone, N. L., Remaley, A. T., Charboneau, L., et al. (2000) Sensitive immuno10 assay of tissue cell proteins procured by laser capture microdissection. Am. J. Pathol. 156, 445–452. 11. Bancroft, J. D. and Gamble, M. (2002) Theory and Practice of Histological Techniques, 5th ed. Edinburgh: Churchill Livingstone. 12. Carson, F. L. (2003) Histotechnology: A Self Instructional Text, 2nd ed. Chicago: ASCP Press. 13. Kiernan, J. A. (2001) Histological and Histochemical Methods, 3rd ed. Oxford: Oxford University Press. 14. Kiernan, J. A. and Mason, I. (eds.) (2002) Microscopy and Histology for Molecular Biologists: A User’s Guide. London: Portland Press. 15. Histonet Listserver Information http://www.histonet.org/. 16. Gassmann, M. (2003) Quality Assurance of RNA derived from laser microdissected tissue samples obtained by the PALM(R) MicroBeam System using the RNA 6000 Pico LabChip(R) kit. 1-8. 2003. Agilent Technologies. 17. Horobin, R. W. and Kiernan, J. A. (eds.) (2002) Conn’s Biological Stains: A Handbook of Dyes, Stains and Fluorochromes for Use in Biology and Medicine, 10th ed. Published for the Biological Stain Commission by BIOS Scientific Publishers, Distributed in US by Springer-Verlag (U.S.), Oxford, UK. 18. 18 Zhuang, Z., Bertheau, P., Emmert-Buck, M. R., et al. (1995) A microdissection technique for archival DNA analysis of specific cell populations in lesions <1 mm in size. Am. J. Pathol. 146, 620–625. 19. 19 Lee, J. Y., Dong, S. M., Kim, S. Y., Yoo, N. J., Lee, S. H., and Park, W. S. (1998) A simple, precise and economical microdissection technique for analysis of genomic DNA from archival tissue sections. Virchows Arch. 433, 305–309. 20. 20 Gupta, S. K., Douglas-Jones, A. G., and Morgan, J. M. (1997) Microdissection of stained archival tissue. Mol. Pathol. 50, 218–220. 21. 21 Harsch, M., Bendrat, K., Hofmeier, G., Branscheid, D., and Niendorf, A. (2001) A new method for histological microdissection utilizing an ultrasonically oscillating needle: Demonstrated by differential mRNA expression in human lung carcinoma tissue. Am. J. Pathol. 158, 1985–1990. 22. 22 Beltinger, C. P. and Debatin, K. M. (1998) A simple combined microdissection and aspiration device for the rapid procurement of single cells from clinical peripheral blood smears. Mol. Pathol. 51, 233–236. 23. 23 Zhang, Z., Nakamura, M., Taniguchi, E., Shan, L., Yokoi, T., and Kakudo, K. (1997) A simple approach to single-cell microdissection and molecular analysis. Anal. Quant. Cytol. Histol. 19, 514–518. 24. 24 Emmert-Buck, M. R., Bonner, R. F., Smith, P. D., et al. (1996) Laser capture microdissection [see Comments]. Science 274, 998–1001. 25. 25 Curran, S., McKay, J. A., McLeod, H. L., and Murray, G. I. (2000) Laser capture microscopy. Mol. Pathol. 53, 64–68.
Molecular Profiling of Tissue
65
26. 26 Goldsworthy, S. M., Stockton, P. S., Trempus, C. S., Foley, J. F., and Maronpot, R. R. (1999) Effects of fixation on RNA extraction and amplification from laser capture microdissected tissue. Mol. Carcinog. 25, 86–91. 27. 27 Lawrie, L. C., Curran, S., McLeod, H. L., Fothergill, J. E., and Murray, G. I. (2001) Application of laser capture microdissection and proteomics in colon cancer. Mol. Pathol. 54, 253–258. 28. 28 Craven, R. A. and Banks, R. E. (2002) Use of laser capture microdissection to selectively obtain distinct populations of cells for proteomic analysis. Methods Enzymol. 356, 33–49. 29. Nakazono, M., Qiu, F., Borsuk, L. A., and Schnable, P. S. (2003) Laser-capture 29 microdissection, a tool for the global analysis of gene expression in specific plant cell types: Identification of genes expressed differentially in epidermal cells or vascular tissues of maize. Plant Cell 15, 583–596. 30. Luzzi, V., Mahadevappa, M., Raja, R., Warrington, J. A., and Watson, M. A. 30 (2003) Accurate and reproducible gene expression profiles from laser capture microdissection, transcript amplification, and high density oligonucleotide microarray analysis. J. Mol. Diagn. 5, 9–14. 31. Burgemeister, R., Gangnus, R., Haar, B., Schutze, K., and Sauer, U. (2003) High 31 quality RNA retrieved from samples obtained by using LMPC (laser microdissection and pressure catapulting) technology. Pathol. Res. Pract. 199, 431–436. 32. Fink, L., Kohlhoff, S., Stein, M. M., Hanze, J., et al. (2002) cDNA array hybridi32 zation after laser-assisted microdissection from nonneoplastic tissue. Am. J. Pathol. 160, 81–90. 33. Cohen, C. D., Grone, H. J., Grone, E. F., Nelson, P. J., Schlondorff, D., and 33 Kretzler, M. (2002) Laser microdissection and gene expression analysis on formaldehyde-fixed archival tissue. Kidney Int. 61, 125–132. 34. Kleeberger, W., Rothamel, T., Glockner, S., Lehmann, U., and Kreipe, H. (2000) 34 Laser-assisted microdissection and short tandem repeat PCR for the investigation of graft chimerism after solid organ transplantation. Pathobiology 68, 196–201. 35. 35 Inoue, K., Sakurada, Y., Murakami, M., Shirota, M., and Shirota, K. (2003) Detection of gene expression of vascular endothelial growth factor and flk-1 in the renal glomeruli of the normal rat kidney using the laser microdissection system. Virchows Arch. 442, 159–162. 36. Mori, M., Mimori, K., Yoshikawa, Y., et al. (2002) Analysis of the gene-expres36 sion profile regarding the progression of human gastric carcinoma. Surgery 131, S39–S47. 37. Brockhoff, G., Fleischmann, S., Meier, A., Wachs, F. P., Hofstaedter, F., and 37 Knuechel, R. (1999) Use of a mechanical dissociation device to improve standardization of flow cytometric cytokeratin DNA measurements of colon carcinomas. Cytometry 38, 184–191. 38. Pertoft, H. (2000) Fractionation of cells and subcellular particles with Percoll. J. 38 Biochem. Biophys. Methods 44, 1–30. 39. 39 Tai, Y. T., Teoh, G., Shima, Y., et al. (2000) Isolation and characterization of human multiple myeloma cell enriched populations. J. Immunol. Methods 235, 11–19.
66
Bova et al.
40. 40 Moskaluk, C. A. and Kern, S. E. (1997) Microdissection and polymerase chain reaction amplification of genomic DNA from histological tissue sections. Am. J. Pathol. 150, 1547–1552. 41. Goelz, S. E., Hamilton, S. R., and Vogelstein, B. (1985) Purification of DNA from 41 formaldehyde fixed and paraffin embedded human tissue. Biochem. Biophys. Res. Commun. 130, 118–126. 42. Emmert-Buck, M. R., Gillespie, J. W., Paweletz, C. P., et al. (2000) An approach 42 to proteomic analysis of human tumors. Mol. Carcinog. 27, 158–165. 43. 43 Ikeda, K., Monden, T., Kanoh, T., et al. (1998) Extraction and analysis of diagnostically useful proteins from formalin-fixed, paraffin-embedded tissue sections. J. Histochem. Cytochem. 46, 397–403. 44. 44 Novelli, M., Savoia, P., Cambieri, I., et al. (2000) Collagenase digestion and mechanical disaggregation as a method to extract and immunophenotype tumour lymphocytes in cutaneous T-cell lymphomas. Clin. Exp. Dermatol. 25, 423–431. 45. Maitra, A., Wistuba, I. I., Virmani, A. K., et al. (1999) Enrichment of epithelial 45 cells for molecular studies. Nat. Med. 5, 459–463. 46. 46 Guerrero, R. B., Batts, K. P., Brandhagen, D. J., Germer, J. J., Perez, R. G., and Persing, D. H. (1997) Effects of formalin fixation and prolonged block storage on detection of hepatitis C virus RNA in liver tissue. Diagn. Mol. Pathol. 6, 277–281. 47. 47 Ohyama, H., Zhang, X., Kohno, Y., and Alevizos, M. (2000) Laser capture microdissection-generated target sample for high-density oligonucleotide array hybridization. Biotechniques 29, 530–536. 48. Chomczynski, P. and Sacchi, N. (1987) Single-step method of RNA isolation by 48 acid guanidinium thiocyanate-phenol-chloroform extraction. Anal. Biochem. 162, 156–159. 49. Maunsbach, A. B. (1966) The influence of different fixatives and fixation meth49 ods on the ultrastructure of rat kidney proximal tubule cells. II. Effects of varying osmolality, ionic strength, buffer system and fixative concentration of glutaraldehyde solutions. J. Ultrastruct. Res. 15, 283–309. 50. Morales, A. R., Essenfeld, H., Essenfeld, E., Duboue, M. C., Vincek, V., and Nadji, M. (2002) Continuous-specimen-flow, high-throughput, 1-hour tissue processing. A system for rapid diagnostic tissue preparation. Arch. Pathol. Lab. Med. 126, 583–590.
Immunohistochemistry and In Situ Hybridization
67
3 Immunohistochemistry and In Situ Hybridization in Pancreatic Neoplasia Robb E. Wilentz, Ayman Rahman, Pedram Argani, and Christine Iacobuzio-Donahue Summary There are many types of pancreatic neoplasms. Pathologic examination, which includes both routine (e.g., hematoxylin-and-eosin staining) and ancillary (e.g., immunohistochemistry and in situ hybridization) techniques, is essential in correctly typing a pancreatic neoplasm. This chapter focuses on the use of immunohistochemistry and in situ hybridization in the differentiation of pancreatic neoplasms. The materials and methods of these two techniques are described in detail. Key Words: Pancreatic cancer; pancreas; ductal adenocarcinoma; immunohistochemistry; in situ hybridization.
1. Introduction There are many different types of pancreatic neoplasms. Therefore, one must always consider a number of benign and malignant neoplasms when making the diagnosis of a pancreatic tumor. Pathologic examination of a tumor, which includes both routine (e.g., hematoxylin and eosin staining) and ancillary (e.g., immunohistochemistry and in situ hybridization) techniques, is essential to its accurate diagnosis. Needless to say, the correct diagnosis by means of these studies is extremely important, as each pathological entity is also clinically distinct, with its own clinical associations and outcome. This chapter thus focuses on the use of immunohistochemistry and in situ hybridization in the differentiation of primary pancreatic neoplasms. Each of the neoplasms is discussed in reference to its immunohistochemical/in situ hybridization staining pattern. In general, neoplasms in the pancreas can be primary to the gland, metastatic, or systemic. Primary neoplasms arise in the pancreas. From: Methods in Molecular Medicine, Vol. 103: Pancreatic Cancer: Methods and Protocols Edited by: G. Su © Humana Press Inc., Totowa, NJ
67
68
Wilentz et al.
Primary pancreatic tumors can show differentiation toward exocrine (ducts and acini), endocrine (islets), or mesenchymal (vessels, nerves, muscle, and fat) structures. Exocrine neoplasms can be solid or cystic. Metastatic cancers originate elsewhere and spread by blood or lymphatics to the pancreas. Systemic malignancies such as lymphomas derive from the blood or lymph nodes themselves; by definition, they simultaneously involve multiple sites, one of which may be the pancreas. This chapter focuses on neoplasms that are primary to the pancreas. 1.1. Primary Solid Exocrine Neoplasms 1.1.1. Ductal Adenocarcinoma Ductal adenocarcinoma is the most common primary malignancy of the pancreas. It accounts for almost three fourths of all cancers that involve the pancreas (1,2). Ductal adenocarcinomas are composed of infiltrating glands surrounded by dense, reactive fibrous tissue. The nuclei of the cells are irregular, they lack polarity, and they have prominent nucleoli. Immunohistochemically, ductal adenocarcinomas usually express cytokeratin, epithelial membrane antigen (EMA), carcinoembryonic antigen (CEA), and carbohydrate antigen (CA) 19-9. Unfortunately, reactive cells can also produce these antigens; therefore, these studies do not distinguish between reactive and neoplastic glands in the pancreas (3–5). A recently developed immunohistochemical stain for the Dpc4 protein may solve this problem. DPC4 is a tumor-suppressor gene that is inactivated in more than 55% of pancreatic adenocarcinomas (6–9). The DPC4 gene is not inactivated in reactive or normal epithelium. We recently showed that immunohistochemistry for the Dpc4 protein is an extremely sensitive and specific marker of DPC4 gene inactivation (10). Therefore, the loss of Dpc4 expression may help establish the diagnosis of cancer in the pancreas (Fig. 1). The presence of the Dpc4 protein, however, does not rule out cancer, as some (approx 45%) pancreatic adenocarcinomas do not inactivate the DPC4 gene. Also, inactivation of DPC4 does not prove that a lesion is invasive carcinoma, as it is seen in approximately 30% of high-grade pancreatic intraepithelial neoplasias (PanINs) (carcinomas in situ; see Chapter 1) (11). 1.1.2. Medullary Carcinoma Recently Goggins et al. described a subset of five poorly differentiated adenocarcinomas with a syncytial growth pattern, pushing borders, and extensive necrosis. They showed that these cancers are associated with a specific genetic abnormality called “DNA replication errors” (“microsatellite instability”) and that most had wild-type K-ras genes (12). We have subsequently confirmed these results by studying an additional 13 medullary carcinomas (13).
Immunohistochemistry and In Situ Hybridization
69
Fig. 1. (A) Hematoxylin and eosin–stained section of an infiltrating duct adenocarcinoma of the pancreas. Note the invasive glands within a fibrotic stroma. (B) Immunohistochemical stain for Dpc4. Expression is lost in the malignant gland in the left side of the image, while the normal pancreas on the right shows intact expression.
Medullary carcinoma is an important variant of pancreatic cancer to recognize for a number of reasons. First, the initial data suggest that medullary carcinomas may have a better outcome than conventional ductal adenocarcinomas,
70
Wilentz et al.
even though the number of patients studied with medullary carcinoma is still too small to produce a statistically significant survival difference (12). Second, our studies show that a small minority (6%) of medullary carcinomas harbor Epstein–Barr virus RNA by in situ hybridization (Fig. 2) (13). Finally, medullary carcinoma may represent the first instance in which pancreatic cancer histology can identify an inherited susceptibility to developing pancreatic carcinoma. This can occur in two ways. First, we have shown that patients with pancreatic medullary carcinomas are more likely to have first-degree relatives with pancreatic and extrapancreatic cancers (13). Second, medullary carcinoma of the pancreas may be a manifestation of an inherited cancer syndrome. One patient from our studies had a pancreatic medullary carcinoma and a synchronous cecal colonic carcinoma. Both carcinomas had microsatellite instability, a finding that suggests that medullary carcinomas of the pancreas may be a part of the hereditary nonpolyposis colorectal carcinoma (HNPCC) syndrome (13). Such syndromes are caused by germline mutations in one of the DNA mismatch repair genes (MSH2, MLH1, etc.), and these defects can be identified through immunohistochemical absence of the protein derived from one of these mismatch repair genes (Fig. 2). 1.1.3. Acinar Cell Carcinoma Acinar cell carcinomas account for only 5% of pancreatic cancers, and they have a distinct histology, immunohistochemical profile, and clinical presentation (14–24). Acinar cell carcinomas show acinar differentiation by light microscopy. That is, these tumors form clusters of cells around small, central lumina. Stromal desmoplasia is decidedly less prominent than it is in ductal adenocarcinoma. Immunohistochemistry often reveals the expression of trypsin, lipase, chymotrypsin, and/or amylase. Abraham et al. recently reported that a minority (25%) of acinar cell carcinomas have alterations in the APC/b-catenin pathway; these alterations can be detected by abnormal nuclear accumulation on a b-catenin immunostain. Also, acinar cell carcinomas, in contrast to ductal adenocarcinomas, do not show loss of Dpc4 expression (25). As many as 20% of patients with this carcinoma develop subcutaneous fat necrosis, an erythema
Fig. 2. (Opposite page) (A) Medullary carcinoma of the pancreas. The panel on the left shows the histology of medullary carcinoma, with a syncytial growth pattern and intratumoral lymphocytes (hematoxylin and eosin stain). The panel on the right demonstrates the positivity for Epstein–Barr virus RNA. (B) Immunohistochemical stains for DNA mismatch repair genes MLH1 and MSH2. This medullary carcinoma has lost nuclear expression for Mlh1 (left), thus explaining its “mutator phenotype” (“microsatellite instability”) status. The immunohistochemical stain for Msh2 (right) is positive.
Immunohistochemistry and In Situ Hybridization
71
72
Wilentz et al.
nodosum-like rash, peripheral eosinophila, and/or polyarthralgias. These symptoms result from the release of lipase by the tumor (19,21,26–30). 1.1.4. Pancreatoblastoma Pancreatoblastoma is a neoplasm that occurs primarily in children under 15 yr of age (31–36). Microscopically, pancreatoblastomas contain undifferentiated and differentiated components. The undifferentiated part consists of small cells with a syncytial growth pattern and round nuclei. The differentiated segment can show a spectrum of appearances, including squamous, acinar, and endocrine differentiation. Typically, however, these tumors contain nests of squamoid cells in a sea of uniform, undifferentiated cells. Immunohistochemically, pancreatoblastomas sometimes label for somatostatin and CEA (31). In addition, Abraham et al. recently showed that the majority of pancreatoblastomas contain alterations in the APC/b-catenin pathway; these alterations can be detected by abnormal nuclear accumulation on a b-catenin immunostain (Fig. 3) (37). 1.2. Primary Cystic Exocrine Neoplasms 1.2.1. Serous Cystadenomas Serous cystadenomas are benign neoplasms that are spongy, well circumscribed, and multilocular. They sometimes contain a central, calcified scar. The cysts harbor watery, clear or brown fluid, and a layer of simple cuboidal cells usually lines the cysts. These cells have clear cytoplasm and round, uniform nuclei. Immunohistochemistry reveals the expression of cytokeratin and EMA but, in contrast to mucinous cystic neoplasms, they do not express CEA. 1.2.2. Mucinous Cystic Neoplasms The discussion of these neoplasms was covered primarily in “Precursors to Pancreatic Cancer” (see Chapter 1). Immunohistochemically, these tumors demonstrate the expression of CEA, CA 19-9, cytokeratin, and EMA by the epithelial cells (38–40). Scattered cells expressing chromogranin, neuron-specific enolase (NSE), and/or serotonin can also be seen within the epithelium (38). The ovarian-like stroma may express inhibin, estrogen receptor, and progesterone receptor. In addition, like ductal adenocarcinoma, the majority of infiltrating adenocarcinomas associated with mucinous cystic neoplasms show loss of Dpc4 expression. The noninvasive portions of mucinous cystic neoplasms, however, do not lose expression of Dpc4 (41). 1.2.3. Intraductal Papillary Mucinous Neoplasms The discussion of these neoplasms was covered primarily in “Precursors to Pancreatic Cancer” (see Chapter 1). Immunohistochemically, these tumors, in
Immunohistochemistry and In Situ Hybridization
73
Fig. 3. (A) Hematoxylin and eosin–stained section of a pancreatoblastoma. Note the squamous island (center) in a “sea” of small blue cells. (B) b-Catenin immunostain of a pancreatoblastoma. There is nuclear accumulation of the b-catenin protein in some of the cells. The majority of pancreatoblastomas demonstrate nuclear accumulation of the b-catenin protein, reflecting abnormalities in the APC/b-catenin signaling pathway. (Photographs courtesy of Susan C. Abraham, M.D., Department of Pathology, Mayo Clinic, Rochester, MN.)
74
Wilentz et al.
contrast to ductal adenocarcinomas and mucinous cystic neoplasms and whether invasive or not, virtually always express Dpc4 (42). 1.2.4. Solid-Pseudopapillary Neoplasm Solid-pseudopapillary neoplasms of the pancreas occur primarily in women in their 20s (43–51). These tumors contain solid areas, cysts, pseudopapillae, hemorrhage, and necrosis (43–50). The solid areas comprise nests of small, uniform cells with pink cytoplasm. Cysts are filled with blood, and the pseudopapillae usually have fibrovascular cores. Immunohistochemically, some of these tumors express a1-antitrypsin and/or lipase (43,45–47). Recently, Abraham et al. showed that these tumors virtually always demonstrate nuclear b-catenin immunoreactivity and are associated with corresponding b-catenin gene mutations (52). 1.3. Primary Endocrine Neoplasms There are two types of endocrine neoplasms of the pancreas: (1) well- and moderately differentiated neuroendocrine neoplasms and (2) poorly differentiated (high-grade) neuroendocrine carcinoma (small cell carcinoma). As their name implies, well- and moderately differentiated endocrine neoplasms show clear signs of endocrine differentiation, both morphologically (nested and trabecular growth pattern, fine chromatin) and immunohistochemically. Unlike welland moderately differentiated endocrine neoplasms, endocrine differentiation in poorly differentiated neuroendocrine carcinomas is not obvious, either histologically or immunohistochemically. Like mucinous cystic neoplasms of the pancreas, well- and moderately differentiated endocrine neoplasms can be benign, borderline (“of uncertain malignant potential”), or malignant. Unfortunately, unlike with mucinous cystic neoplasms, it is often very difficult to place an endocrine tumor into one of these three categories based on histology alone. Instead, the best way to determine if a well- or moderately differentiated endocrine neoplasm is benign or malignant is to consider its clinical behavior. Another approach to assessing the malignant potential of a well- or moderately differentiated endocrine neoplasm is measuring the tumor’s proliferation rate immunohistochemically (53). Unfortunately, this technique is not entirely reliable. Some low-grade carcinomas have a low (1% positive cells) Ki-67 proliferation index, while some benign neoplasms have high indices (53). Nevertheless, a proliferation rate above 10% usually indicates malignancy (53). Well- and moderately differentiated endocrine neoplasms demonstrate a nested or tabecular growth pattern and contain fine, uniformly distributed chromatin. They label immunohistochemically for chromogranin, synaptophysin,
Immunohistochemistry and In Situ Hybridization
75
and NSE. In addition, these neoplasms sometimes express cytokeratin and neurofilament. Furthermore, these neoplasms may express specific pancreatic and/or extrapancreatic hormones, for example, insulin, glucagon, or serotonin (54–62). High-grade (poorly differentiated) neuroendocrine carcinomas are infiltrative, hemorrhagic, necrotic, and firm. They have extremely high nuclear-to-cytoplasmic ratios, nonapparent nucleoli, and obvious necrosis. Nuclear molding, where nuclei wrap around one another, is prominent. Both the mitotic rate (at least 10 per 10 high-power fields) and the Ki-67 proliferation index (at least 10%) are high (2,63–67). High-grade neuroendocrine carcinomas sometimes show faint or focal expression of endocrine markers, including NSE, synaptophysin, and chromogranin (1,2,2,67,68). In addition, immunolabeling for cytokeratin is sometimes positive in these tumors. 1.4. Primary Mesenchymal Neoplasms Benign and malignant mesenchymal tumors of the pancreas are extremely rare. Schwannoma, leiomyosarcoma, liposarcoma, primitive neuroectodermal tumor (PNET), and malignant fibrous histiocytoma have all been reported in the pancreas (69–76). Generally, survival rates for patients with sarcomas are low, as they are for sarcomas primary in other locations. Immunohisochemically, these tumors stain as they do in the locations in which they typically arise. 2. Materials 2.1. Immunohistochemistry 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18.
Polylysine-coated or charged slides of paraffin-embedded tissue of choice. Xylene. 100% Ethanol. 70% Ethanol. 50% Ethanol. Deionized water. Deionized water with 0.1% Tween-20. 10 mM citrate buffer, pH 6.5. Phosphate-buffered saline (PBS)/Tween-20 (PBST). 3% Hydrogen peroxide. PBS. Primary antibody. Secondary antibody. Avidin–biotin–horseradish peroxidase complex. 3',3-Diaminobenzidine (DAB). Hematoxylin. Mounting medium. Cover slip.
76
Wilentz et al.
2.2. In situ Hybridization 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17.
18. 19. 20. 21. 22. 23. 24. 25. 26.
27. 28. 29. 30. 31. 32. 33.
Sense and antisense primers to gene of choice. Phenol–chloroform, pH 8.0. Ethanol. LoTE: 3 mM Tris-HCl, pH 8.0, 0.2 mM EDTA, pH 8.0. Ultrapure deionized water. 10X DIG (digoxigenin) RNA labeling mix (Roche). 10X transcription buffer. RNase inhibitor. T7 polymerase. DNase I, RNase-free. 0.2 M EDTA. Polylysine-coated or charged paraffin-embedded tissue of choice. Diethyl pyrocarbonate (DEPC)-treated water. Xylene. 50%, 70%, and 100% ethanol. 1% Hydrogen peroxide. Proteinase K: The stock solution is made by dissolving one 100-mg bottle of proteinase K powder in 5 mL of RNase-free dH2O. Freeze small aliquots (e.g., 500 µL) at -20°C. When ready to use, thaw aliquot and dilute to optimal concentration in 50-mL tubes with 37°C prewarmed TBS. TBS, pH 7.5: 100 mM Tris-HCl, pH 7.5, and 150 mM NaCl. TBST: Add 111 mL of 10X Tris-buffered saline with Tween-20 directly to 1 L of RNase-free water. Hybridization solution. Incubation chamber gasket. RNase A/T1 cocktail (Ambion). Formamide (deionized). 20X Saline sodium citrate (SSC). 50X SSC stringent wash concentrate (DAKO). Blocking buffer: First make 100 mM Tris-HCl, pH 7.5, 150 mM NaCl, and 1% blocking reagent (highly purified casein). With pipetting and vortexing, blocking reagent will dissolve at 55°C in <1 h. Store at 4°C and use within a month (longer storage may lead to higher background). Immediately before adding to sections, centrifuge the rabbit immunoglobulin fraction in microfuges, and add a 1:20 dilution to the blocking buffer. To remove any precipitated blocking reagent, centrifuge in a tabletop centrifuge for 5 min. Store at room temperature during the assay. Rabbit horseradish peroxidase (HRP)-anti-DIG. Biotinyl-tyramide. Streptavidin. DAB. Hematoxylin. Mounting medium. Cover slip.
Immunohistochemistry and In Situ Hybridization
77
3. Methods 3.1. Immunohistochemistry Immunohistochemistry essentially consists of three parts: (1) slide preparation, (2) primary antibody incubation, and (3) reaction detection. Because immunohistochemistry is most commonly performed on formalin-fixed, paraffinembedded tissues, the generalized procedure for tissues prepared in this manner is discussed. Immunohistochemistry for frozen tissues can be performed with slight modifications of the following procedures, for example, by deleting the need for slide deparaffination and antigen retrieval. 3.1.1. Slide Preparation First, slides must be prepared for subsequent incubation with antibody. To do this, we must not only “uncover” the antigen (“antigen retrieval”) but also eliminate any factors that will detract from the subsequent incubation reaction. The two most important procedures to accomplish this include “endogenous peroxidase quenching” and “serum blocking.” These procedures are performed in the following order: 1. “Endogenous peroxidase quenching” is performed by treatment with hydrogen peroxide. Endogenous peroxidases must be eliminated because peroxidase activity develops the commonly used chromagen at the end of the immunohistochemical procedures. 2. “Antigen retrieval” is necessary because formalin fixation crosslinks and denatures proteins such that many epitopes are inaccessible to antibodies. Methods of antigen retrieval include treatment with sodium citrate buffer or incubation with a protease such as pronase or pepsin. Although the mechanism of antigen retrieval is unknown, this technique has revolutionized immunohistochemistry over the past decade. 3. “Serum blocking” is performed with serum from the species in which the secondary antibody was made. If the species-specific serum is added to a slide, it will bind excess secondary antibody and decrease nonspecific binding.
3.1.2. Primary Antibody Incubation “Prepared” slides are treated with the primary antibody. Of course, the conditions under which the primary antibody is added must be optimized. Important variables include the antibody dilution, exposure time, and temperature. A slide to which normal saline is added serves as a negative control. An appropriate positive control (a tissue that bears the antigen of interest) is also used. 3.1.3. Reaction Detection The reaction of antibody to its target can now be assayed. Detection is performed by adding three reagents in a set order: biotinylated secondary antibodies,
78
Wilentz et al.
avidin–biotin–peroxidase complex, and DAB. Biotinylated secondary antibodies bind the Fc portion of the primary antibody. Avidin–biotin–peroxidase complex, which is not yet bound to capacity with biotin, binds the biotin on the secondary antibody. Other avidin–biotin–peroxidase complexes then bind to the biotin on the avidin–biotin–peroxidase complex that is bound to the secondary antibody. This creates a “chain reaction” of avidin–biotin–peroxidase binding. DAB is oxidized by the peroxidase within the avidin–biotin–peroxidase complexes, resulting in a brown color change. The DAB reaction is amplified because of the avidin–biotin–peroxidase “chain reaction,” and thus the sensitivity of the reaction is dramatically increased. Sections are counterstained with hematoxylin so that tissue compartments can be visualized. The following protocols are derived from routine protocols (77): 3.1.4. Deparaffination 1. Heat slides to 68°C for 10 min. 2. Rinse in xylene twice for 5 min each. 3. Hydrate slides through graded alcohols (dip slides in 100%, 70%, and 50% ethanol in that order). 4. Dip slides in deionized water. 5. Wash slides in deionized water with 0.1% Tween-20.
3.1.5. Antigen Retrieval/Peroxidase Quenching 1. 2. 3. 4. 5. 6.
Place slides in 10 mM citrate buffer, pH 6.5. Steam slides for 14 min. Cool slides for 5 min. Wash slides with PBST for 5 min. Incubate slides for 7 min in 3% hydrogen peroxide. Rinse slides twice in PBS.
3.1.6. Primary Antibody Incubation 1. Incubate slides with primary antibody for 1 h (or optimized length of time) at room temperature or at 4°C in a humid chamber. 2. Rinse slides in PBS.
3.1.7. Reaction Detection 1. Incubate slides with secondary antibody (e.g., at a 1:1000 dilution in PBST) for 30 min at room temperature. 2. Rinse slides in PBS. 3. Add avidin–biotin–horseradish peroxidase complex for 30 min at room temperature. 4. Rinse slides in PBS. 5. Add DAB for 5 min at room temperature.
Immunohistochemistry and In Situ Hybridization
79
6. 7. 8. 9.
Rinse slides in PBS. Counterstain slides with hematoxylin for 30 s. Wash slides with deionized water. Dehydrate slides in graded alcohols and xylene (dip in 50%, 70%, and 100% ethanol in that order, then dip in xylene). 10. Mount slides with a cover slip.
3.2. In Situ Hybridization The process of in situ hybridization consists of two steps: first, making probes, and second, detecting gene expression with the probes. For reasons described below, we routinely use digoxigenin-labeled riboprobes in our system. Digoxigenin is a derivative of the steroid digoxin, and it is useful because antibodies can be directed to it. To make these digoxigenin-labeled riboprobes, DNA from an IMAGE clone is amplified by polymerase chain reaction (PCR) (see Note 1). The amplification is performed twice, each time with either the reverse or forward primer containing a T7 promoter sequence at its 5'end (see Note 2). Each of the PCR products is then subjected to in vitro transcription with T7 polymerase. The uracil used during in vitro transcription contains a digoxigenin moiety within it, and thus a transcribed riboprobe is labeled with digoxigenin. The DNA templates made with the T7 promoter sequence on the reverse primer will generate antisense probes, while those made with the T7 promoter sequence on the forward primer will create sense probes. The riboprobes are stored at -80°C until use (see Note 3). Once the probes have been made, multiple samples can be assayed for gene expression. Expression is detected by adding the riboprobe to formalin-fixed, paraffin-embedded tissue sections under set conditions. These hybridization conditions are determined based on the GC content and the length of a given probe. The sections are subsequently treated with RNase to degrade singlestranded RNA so that only double-stranded RNA (probe bound to the message) remains. The sections are incubated with mouse monoclonal anti-digoxigenin antibodies, which bind to the digoxigenin-containing double-stranded RNA complex. Secondary biotinylated antibodies attach to the primary antibodies, and biotin–avidin–peroxidase complex binds the biotin within the secondary antibodies. DAB reacts with peroxidase of the biotin–avidin–peroxidase complex, creating a brown color at sites of hybridization. Counterstaining with hematoxylin allows slides to be studied under a microscope for gene expression within tissue compartments. Sense probes are substituted for antisense probes as controls. An alternative to this digoxigenin system relies on radioactive probes. We have chosen the digoxigenin in situ hybridization system over a radioactive one for four reasons. First, and most obviously, digoxigenin-based hybridization experiments eliminate the potential biohazards associated with radioactivity
80
Wilentz et al.
use. Second, digoxigenin probes have long half-lives; once made, each probe can last months to years if stored at -80°C. Third, it is easy to interpret tissue morphology under the microscope with the digoxigenin system. Fourth, the digoxigenin system produces permanent color changes, whereas radioactive probes fade over time. Hence, although a disadvantage of the digoxigenin system is that it is less sensitive than a radioactive one, these four advantages tend to outweigh this problem. Alternatives to single-stranded RNA probes are double-stranded cDNA or oligonucleotide probes. We choose to use the single-stranded riboprobes for four reasons. First, RNA probes are more specific than cDNA and oligonucleotide probes. This is because RNA:RNA hybrids are more stable than RNA:DNA hybrids, and therefore more stringent washes can be performed during hybridizations. Second, riboprobes are easy to make by in vitro transcription, as described above. Third, unlike double-stranded DNA probes, single-stranded riboprobes do not self-anneal, and the entire probe mixture (not just half of it, i.e., the antisense DNA strand) is available to bind to the target. Fourth, unlike oligonucleotide probes, RNA probes can be as long as 400 basepairs (vs up to 40), thereby allowing greater label incorporation. The protocol presented in the next section, which takes advantage of the points explained above, has already been used to demonstrate specific gene expression patterns within the neoplastic glands and the non-neoplastic stroma that characterize pancreatic cancer (78). This technique allows one to validate gene expression within cancers and also to localize this gene expression to specific histologic compartments. This technique will continue to be useful in the evaluation of the expression of new genes, many of which will have prognostic and therapeutic applications. 3.2.1. Nonradioactive mRNA In Situ Hybridization This protocol was adapted from C. Iacobuzio-Donahue, MD, and S. E. Kern, MD (The Johns Hopkins Medical Institutions). 3.2.1.1. MAKING DIGOXIGENIN-LABELED RIBOPROBES 1. Use PCR to generate 400 to 500-bp-sized DNA templates for antisense or sense riboprobes by incorporating the following T7 promoter into the 5' end of the antisense or sense primer (see Notes 1–3): 5'CTAATACGACTCACTATAGGG 3'. 2. Collect the PCR product into one tube, extract twice with phenol–chloroform, pH 8.0, and ethanol precipitate. 3. Resuspend pellet in LoTE, measure DNA concentration using spectrophotometer, and dilute to 50 ng/µL.
Immunohistochemistry and In Situ Hybridization
81
4. (In vitro transcription) Mix at room temperature the following reagents in the order shown: dH2O water 9 µL 200 ng of DNA template (50 ng/µL) 4 µL 10X DIG RNA labeling mix 2 µL 10X Transcription buffer 2 µL RNase inhibitor 1 µL T7 polymerase 2 µL 5. Incubate reaction 2 h at 37°C. 6. Add 2 µL of DNase I, RNase-free and incubate at 37°C for 15 min. 7. Add 2 µL of 0.2 M EDTA and place on ice. 8. Measure RNA concentration and raise volume to 200 ng/µL with DEPC-treated water. 9. Determine the amount of digoxigenin incorporation into the newly made riboprobe with a nucleic acid detection kit (optional step).
3.2.1.2. IN SITU DETECTION
Day 1 1. Have your paraffin-embedded sections cut onto microscope slides that are polylysine coated or charged. Once cut, these sections can be stored in any clean, dry container until ready for use. 2. Incubate slides in xylene for 5 min (see Note 4). 3. Incubate in 100% ethanol for 5 min. 4. Incubate in 70% ethanol for 5 min. 5. Incubate in 50% ethanol for 5 min. 6. Rinse in dH2O twice, then incubate in 1% hydrogen peroxide for 10 min at room temperature. Rinse again in dH2O twice. 7. Incubate 5 min in TBS. 8. Immediately immerse sections in proteinase K diluted in fresh prewarmed TBS (see Note 5). Incubate 30 min at 37°C in a water bath. 9. Rinse with 2X TBST for 3 min. 10. Add in a 1.5-mL RNase-free microfuge tube digoxigenin-labeled riboprobe to fresh hybridization solution (final concentration 100–200 ng/mL). Ensure a dilution of at least 1:50. Vortex-mix, then denature the probe at 65°C in a heating block for 2 min and place immediately on ice. Add 300 µL of the diluted probe solution to a large incubation chamber gasket. Invert slide, and press firmly onto the gasket containing the riboprobe. Incubate overnight in a covered water bath (see Note 6). The incubation temperature is 15–25°C below the hybridization temperature (HT) predetermined for the individual probe (this calculated temperature is the temperature at which 50% of the probe is hybridized): HT = 41+ 0.41(%GC) - 500/n, where %GC is the percentage of GC in the probe (e.g., 60) and n represents the length of the probe in bases.
82
Wilentz et al.
Day 2 1. Rinse excess probe from slides by incubating in 50-mL tubes containing 30 mL of 2X SSC for at least 5 min in a 45°C water bath (see Note 7). 2. Rinse with 2X SSC, then incubate with 250 µL of RNase A/T1 cocktail, diluted 1:35 in 2X SSC, at 37°C for 30 min in a covered water bath. Use an incubation chamber to prevent evaporation of solution. 3. Stringently wash slides with prewarmed 2X SSC, 50% formamide DI, in a 50-mL tube twice at 5–8°C below the HT for 20 min each. 4. Stringently wash slides with prewarmed 0.08X SSC once for 20 min at 5–8°C below the HT. Make this final wash solution using 50X SSC Stringent Wash Concentrate, which also contains detergent and blocking agent. 5. Rinse twice with 1X TBST for 3 min at room temperature. 6. Quickly dry around the sections with a Kimwipe. Next, circumscribe the outside edges of the slide with a heat-resistant pap pen. Before the sections dry, add 150 µL of blocking buffer (see Notes 8 and 9). Incubate for 30 min at room temperature. 7. Incubate sections with rabbit HRP-anti-DIG diluted 1:100 in blocking buffer for 30 min at room temperature. 8. Wash three times for 5 min with 1X TBST with agitation. (You can use a platform shaker set to 2–3.) 9. Add one drop of ready-to-use biotinyl-tyramide directly to slides. Incubate in dark 15 min at room temperature. (Use an inverted ice bucket to accomplish this.) 10. Wash three times for 5 min with 1X TBST with agitation (see Note 10). 11. Incubate sections with secondary streptavidin for 15 min at room temperature. 12. Wash three times for 5 min with 1X TBST with agitation. 13. Develop with DAB, made from diluting DAB concentrate 1:50 in dilution buffer. Add approx 150 µL to each section (see Note 11). 14. Counterstain in a Coplin jar for 15–30 s with Mayer’s Hematoxylin. Rinse twice with dH2O. 15. Dehydrate sections in graded alcohols, 30 s each: 50% ethanol, then 70% ethanol, followed by 100% ethanol. 16. Place in xylene 1 min. 17. Finally, place glass cover slip on a Kimwipe, add one drop of mounting medium to cover slip, invert slide, and press section onto coverslip (see Notes 12 and 13).
3.3. Tissue Array Analysis One important recent advance in immunohistochemical and in situ technique is the development of high-throughput tissue arrays (79–81). This is covered in “Precursors to Pancreatic Cancer” (see Chapter 1). 4. Notes 1. In a typical reaction using an amplified DNA template (e.g., purified PCR product, IMAGE clone, etc.), sufficient product can be derived from as little as three 15-µL PCR reactions of 28 cycles. Run a small amount of PCR product on an agarose gel.
Immunohistochemistry and In Situ Hybridization
2.
3.
4. 5.
6. 7. 8.
9. 10. 11.
12.
13.
83
If a robust PCR product is present and background bands are absent, gel purification is unnecessary. DNA templates made with the T7 promoter on the original reverse primer will generate antisense probes, while those made with the T7 promoter on the forward primer will generate sense probes. Riboprobes are very stable, and, once made, they can be stored for months to years at -80°C. However, if RNA integrity is ever in question, verify RNA integrity and concentration by running 5 µL (1 µg) of RNA on a 6% TBE-urea gel alongside known concentrations of marker. Use 50-mL tubes to incubate slides. Each 50-mL tube can fit two outward-facing slides. Thirty milliliters of solution is sufficient to immerse sections. The optimal concentration of proteinase K to be used for a run of in situ experiments using a particular probe must be determined, but a good starting point is to try 10–20 µg/mL of proteinase K in TBS. Overdigestion of tissues can result in high background, while underdigestion results in no or very weakly detectable signal at the end of the experiment. Also, the addition of glass slides to prewarmed TBS usually lowers the temperature several degrees. Therefore, incubations must be timed for 30 min after the temperature reaches 37°C inside the conical tube. (Check the temperature inside the conical tube with a thermometer periodically.) You can probe slides overnight in the water bath by placing slides on a glass plate that is resting on a tube holder within the water bath. Remember to allow the SSC/formamide solution to come up to the desired stringent wash temperature once slides are immersed before timing the incubation. Before using blocking buffer, be sure to add a 1:20 dilution of rabbit immunoglobulin fraction. Before dilution, all antibodies should be spun in a microfuge (e.g., 5 min at maximum speed) to prevent precipitate-related background. Do not add sodium azide to blocking solution as this can inactivate HRP over long periods of time! Once biotin is deposited, sterile conditions are no longer necessary. Coplin jars for all washes are suitable from this point on. Color develops quickly with DAB. Monitor color development under a microscope, and be prepared to quickly immerse slides in a Coplin jar filled with water to stop the reaction. If the ribroprobe is not well labeled with digoxigenin, it may help to add more probe for a better signal. The best amount to add has to be determined on a set of sections hybridized with a range of probe concentrations diluted in hybridization buffer. However, if increased probe concentration does not help, the final signal can be increased with no noticeable increase in background by designing additional probes against nonoverlapping regions of the transcript and by adding 100– 200 ng/mL concentrations of these additional probes in the same hybridization reaction. In contrast, for certain transcripts (particularly abundant ones), nonspecific background caused by the binding of riboprobe can be eliminated completely (without reducing the signal) by titrating down the concentration of probe.
84
Wilentz et al.
References 1. Cubilla, A. L. and Fitzgerald, P. J. (1984) Tumors of the Exocrine Pancreas, 2nd Series ed.Washington, D. C.: Armed Forces Institute of Pathology. 2. O’Connor, T. P., Wade, T. P., Sunwoo, Y. C., et al. (1992) Small cell undifferentiated carcinoma of the pancreas. Report of a patient with tumor marker studies. Cancer 70, 1514–1519. 3. 3 Kim, J., Ho, S. B., Montgomery, C. K., and Kim, Y. S. (1990) Cell lineage markers in human pancreatic cancer. Cancer 66, 2134–2143. 4. 4 Loy, T. S., Springer, D., Chapman, R. K., Diaz-Arias, A. A., Bulatao, I. S., and Bickel, J. T. (1991) Lack of specificity of monoclonal antibody B72.3 in distinguishing chronic pancreatitis from pancreatic adenocarcinoma. Am. J. Clin. Pathol. 96, 684–688. 5. 5 Shimizu, M., Saitoh, Y., Ohyanagi, H., and Itoh, H. (1990) Immunohistochemical staining of pancreatic cancer with CA19-9, KM01, unabsorbed CEA, and absorbed CEA. Arch. Pathol. Lab. Med. 114, 195–200. 6. 6 Hahn, S. A., Schutte, M., Hoque, A. T. M. S., et al. (1996) DPC4, a candidate tumor suppressor gene at human chromosome 18q21.1. Science 271, 350–353. 7. 7 Hahn, S. A., Hoque, A. T. M. S., Moskaluk, C. A., et al. (1996) Homozygous deletion map at 18q21.1 in pancreatic cancer. Cancer Res. 56, 490–494. 8. 8 Schutte, M., Hruban, R. H., Hedrick, L., et al. (1996) DPC4 gene in various tumor types. Cancer Res. 56, 2527–2530. 9. 9 Rozenblum, E., Schutte, M., Goggins, M., et al. (1997) Tumor-suppressive pathways in pancreatic carcinoma. Cancer Res. 57, 1731–1734. 10. 10 Wilentz, R. E., Su, G. H., Dai, J. L., et al. (2000) Immunohistochemical labeling for Dpc4 mirrors genetic status in pancreatic: A new marker of DPC4 inactivation. Am. J. Pathol. 156, 37–43. 11. 11 Wilentz, R. E., Iacobuzio-Donahue, C. A., Argani, P., et al. (2000) Loss of expression of Dpc4 in pancreatic intraepithelial neoplasia: Evidence that DPC4 inactivation occurs late in neoplastic progression. Cancer Res. 60, 2002–2006. 12. 12 Goggins, M., Offerhaus, G. J. A., Hilgers, W., et al. (1998) Pancreatic adenocarcinomas with DNA replication errors (RER+) are associated with wild-type K-ras and characteristic histopathology: Poor differentiation, a syncytial growth pattern, and pushing borders suggest RER+. Am. J. Pathol. 152, 1501–1507. 13. 13 Wilentz, R. E., Goggins, M., Redston, M., et al. (2000) Genetic, immunohistochemical, and clinical features of medullary carcinomas of the pancreas: A newly described and characterized entity. Am. J. Pathol. 156, 1641–1651. 14. 14 Miller, J. R., Baggenstoss, A. H., and Comfort, M. W. (1951) Carcinoma of the pancreas. Effect of histological type and grade of malignancy on its behavior. Cancer 4, 233–241. 15. 15 Chen, J. and Baithun, S. I. (1985) Morphological study of 391 cases of exocrine pancreatic tumours with special reference to the classification of exocrine pancreatic carcinoma. J. Pathol. 146, 17–29.
Immunohistochemistry and In Situ Hybridization
85
16. 16 Morohoshi, T., Held, G., and Klöppel, G. (1983) Exocrine pancreatic tumours and their histological classification. A study based on 167 autopsy and 97 surgical cases. Histopathology 7, 645–661. 17. 17 Hoorens, A., Lemoine, N. R., McLellan, E., et al. (1993) Pancreatic acinar cell carcinoma: An analysis of cell lineage markers, p53 expression, and Ki-ras mutation. Am. J. Pathol. 143, 685–698. 18. 18 Klimstra, D. S., Heffess, C. S., Oertel, J. E., and Rosai, J. (1992) Acinar cell carcinoma of the pancreas: A clinicopathologic study of 28 cases. Am. J. Surg. Pathol. 16, 815–837. 19. 19 MacMahon, H. E., Brown, P. A., and Shen, E. M. (1965) Acinar cell carcinoma of the pancreas with subcutaneous fat necrosis. Gastroenterology 49, 555–559. 20. 20 Mah, P., Loo, D. C., and Tock, E. P. C. (1974) Pancreatic acinar cell carcinoma in childhood. Am. J. Dis. Child. 128, 101–104. 21. 21 Robertson, J. C. and Eeles, G. H. (1970) Syndrome associated with pancreatic acinar cell carcinoma. Br. Med. J. 2, 708–709. 22. 22 Stamm, B., Burger, H., and Hollinger, A. (1987) Acinar cell cystadenocarcinoma of the pancreas. Cancer 60, 2542–2547. 23. 23 Ulich, T., Cheng, L., and Lewin, K. J. (1982) Acinar-endocrine cell tumor of the pancreas. Report of a pancreatic tumor containing both zymogen and neuroendocrine granules. Cancer 50, 2099–2105. 24. 24 Webb, J. N. (1977) Acinar cell neoplasms of the exocrine pancreas. J. Clin. Path. 30, 103–112. 25. 25 Abraham, S. C., Wu, T. T., Hruban, R. H., et al. (2002) Genetic and immunohistochemical analysis of pancreatic acinar cell carcinoma: Frequent allelic loss on chromosome 11p and alterations in the APC/beta-catenin pathway. Am. J. Pathol. 160, 953–962. 26. Auger, C. (1947) Acinous cell carcinoma of the pancreas with extensive fat necrosis. Arch. Pathol. 43, 400–405. 27. 27 Burns, W. A., Matthews, M. J., Hamosh, M., Weide, G. V., Blum, R., and Johnson, F. B. (1974) Lipase-secreting acinar cell carcinoma of the pancreas with polyarthropathy. A light and electron microscopic, histochemical, and biochemical study. Cancer 33, 1002–1009. 28. Osborne, R. R. (1950) Functioning acinous cell carcinoma of the pancreas accompanied with widespread focal fat necrosis. Arch. Intern. Med. 85, 933–943. 29. 29 Hruban, R. H., Molina, J. M., Reddy, M. N., and Boitnott, J. K. (1987) A neoplasm with pancreatic and hepatocellular differentiation presenting with subcutaneous fat necrosis. Am. J. Clin. Pathol. 88, 639–645. 30. 30 Belsky, H. and Cornell, N. W. (1955) Disseminated focal fat necrosis following radical pancreatico-duodenectomy for acinous carcinoma of head of pancreas. Ann. Surg. 141, 556–562. 31. 31 Buchino, J. J., Castello, F. M., and Nagaraj, H. S. (1984) Pancreatoblastoma. A histochemical and ultrastructural analysis. Cancer 53, 963–969.
86
Wilentz et al.
32. 32 Benjamin, E. and Wright, D. H. (1980) Adenocarcinoma of the pancreas of childhood: A report of two cases. Histopathology 4, 87–104. 33. 33 Frable, W. J., Still, W. J. S., and Kay, S. (1971) Carcinoma of the pancreas, infantile type. A light and electron microscopic study. Cancer 27, 667–673. 34. 34 Horie, A., Yano, Y., Kotoo, Y., and Miwa, A. (1977) Morphogenesis of pancreatoblastoma, infantile carcinoma of the pancreas. Report of two cases. Cancer 39, 247–254. 35. 35 Taxy, J. B. (1976) Adenocarcinoma of the pancreas in childhood. Report of a case and a review of the English language literature. Cancer 37, 1508–1518. 36. 36 Tsukimoto, I., Watanabe, K., Lin, J., and Nakajima, T. (1973) Pancreatic carcinoma in children in Japan. Cancer 31, 1203–1207. 37. 37 Abraham, S. C., Wu, T. T., Klimstra, D. S., et al. (2001) Distinctive molecular genetic alterations in sporadic and familial adenomatous polyposis-associated pancreatoblastomas: Frequent alterations in the APC/beta-catenin pathway and chromosome 11p. Am. J. Pathol. 159, 1619–1627. 38. 38 Albores-Saavedra, J., Angeles-Angeles, A., Nadji, M., Henson, D. E., and Alvarez, L. (1987) Mucinous cystadenocarcinoma of the pancreas. Morphologic and immunocytochemical observations. Am. J. Surg. Pathol. 11, 11–20. 39. 39 Yu, H. C. and Shetty, J. (1985) Mucinous cystic neoplasm of the pancreas with high carcinoembryonic antigen. Arch. Pathol. Lab. Med. 109, 375–377. 40. 40 Yamaguchi, K. and Enjoji, M. (1987) Cystic neoplasms of the pancreas. Gastroenterology 92, 1934–1943. 41. 41 Iacobuzio-Donahue, C. A., Wilentz, R. E., Argani, P., et al. (2000) Dpc4 protein in mucinous cystic neoplasms of the pancreas: Frequent loss of expression in invasive carcinomas suggests a role in genetic progression. Am. J. Surg. Pathol. 24, 1544–1548. 42. 42 Iacobuzio-Donahue, C. A., Klimstra, D., Adsay, N. V., et al. (2000) DPC-4 protein is expressed in virtually all human intraductal papillary mucinous neoplasms of the pancreas: Comparison with conventional ductal carcinomas. Am. J. Pathol. 157, 755–761. 43. 43 Learmonth, G. M., Price, S. K., Visser, A. E., and Emms, M. (1985) Papillary and cystic neoplasm of the pancreas—an acinar cell tumour? Histopathology 9, 63–79. 44. 44 Boor, P. J. and Swanson, M. R. (1979) Papillary-cystic neoplasm of the pancreas. Am. J. Surg. Pathol. 3, 69–75. 45. 45 Wilentz, R. E., Chung, C. H., Sturm, P. D. J., et al. (1998) K-ras mutations in the duodenal fluid of patients with pancreatic carcinoma. Cancer 82, 96–103. 46. Klöppel, G., Morohoshi, T., John, H. D., et al. (1981) Solid and cystic acinar cell tumour of the pancreas. A tumor in young women with favorable prognosis. Virchows Arch. [A] 392, 171–183. 47. 47 Schlosnagle, D. C. and Campbell, W. G. (1981) The papillary and solid neoplasm of the pancreas: A report of two cases with electron microscopy, one containing neurosecretory granules. Cancer 47, 2603–2610. 48. Riela, A., Zinsmeister, A. R., Melton, L. J. I., Weiland, L. H., and DiMagno, E. 48 (1992) Increasing incidence of pancreatic cancer among women in Olmsted County, Minnesota, 1940 through 1988. Mayo Clin. Proc. 67, 839–845.
Immunohistochemistry and In Situ Hybridization
87
49. 49 Morrison, D. M., Jewell, L. D., McCaughey, W. T. E., Danyluk, J., Shnitka, T. K., and Manickavel, V. (1984) Papillary cystic tumor of the pancreas. Arch. Pathol. Lab. Med. 108, 723–727. 50. 50 Kuo, T.-T., Su, I.-J., and Chien, C.-H. (1984) Solid and papillary neoplasm of the pancreas. Report of three cases from Taiwan. Cancer 54, 1469–1474. 51. 51 Hamoudi, A. B., Misugi, K., Grosfeld, J. L., and Reiner, C. B. (1970) Papillary epithelial neoplasm of pancreas in a child. Report of a case with electron microscopy. Cancer 26, 1126–1134. 52. 52 Abraham, S. C., Klimstra, D. S., Wilentz, R. E., Wu, T. T., and Hruban, R. H. (2002) Solid-pseudopapillary tumors of the pancreas frequently harbor alterations in the APC/β-catenin pathway. Am. J. Pathol. 160, 1361–1369. 53. 53 La Rosa, S., Sessa, F., Capella, C., et al. (1996) Prognostic criteria in nonfunctioning pancreatic endocrine tumors. Virchows Arch. 429, 323–333. 54. 54 Wilson, B. S. and Lloyd, R. V. (1984) Detection of chromogranin in neuroendocrine cells with a monoclonal antibody. Am. J. Pathol. 115, 458–468. 55. 55 Chejfec, G., Falkmer, S., Grimelius, L., et al. (1987) Synaptophysin. A new marker for pancreatic neuroendocrine tumors. Am. J. Surg. Pathol. 11, 241–247. 56. Hoefler, H., Denk, H., Lackinger, E., Helleis, G., Polak, J. M., and Heitz, P. U. (1986) Immunocytochemical demonstration of intermediate filament cytoskeleton proteins in human endocrine tissues and (neuro-) endocrine tumors. Virchows Arch. [A] 409, 609–626. 57. 57 Perez, M. A., Saul, S. H., and Trojanowski, J. Q. (1990) Neurofilament and chromogranin expression in normal and neoplastic neuroendocrine cells of the human gastrointestinal tract and pancreas. Cancer 65, 1219–1227. 58. Klöppel, G. and Heitz, P. U. (1988) Pancreatic endocrine tumors. Pathol. Res. Pract. 183, 155–168. 59. 59 Heitz, P. U., Kasper, M., Polak, J. M., and Klöppel, G. (1982) Pancreatic endocrine tumors: Immunocytochemical analysis of 125 tumors. Hum. Pathol. 13, 263–271. 60. 60 Kimura, W., Kuroda, A., and Morioka, Y. (1991) Clinical pathology of endocrine tumors of the pancreas. Analysis of autopsy cases. Dig. Dis. Sci. 36, 933–942. 61. 61 Capella, C., Heitz, P. U., Höfler, H., Solcia, E., and Klöppel, G. (1995) Revised classification of neuroendocrine tumors of the lung, pancreas and gut. Virchows Arch. 425, 547–560. 62. 62 Kruseman, A. C., Knijnenburg, G., de la Riviere, G. B., and Bosman, F. T. (1978) Morphology and immunohistochemically-defined endocrine function of pancreatic islet cell tumours. Histopathology 2, 389–399. 63. 63 Reyes, C. V. and Wang, T. (1981) Undifferentiated small cell carcinoma of the pancreas: A report of five cases. Cancer 47, 2500–2502. 64. 64 Hobbs, R. D., Stewart, A. F., Ravin, N. D., and Carter, D. (1984) Hypercalcemia in small cell carcinoma of the pancreas. Cancer 53, 1552–1554. 65. Ibrahim, N. B. N., Briggs, J. C., and Corbishley, C. M. (1984) Extrapulmonary 65 oat cell carcinoma. Cancer 54, 1645–1661. 66. 66 Morant, R. and Bruckner, H. W. (1989) Complete remission of refractory small cell carcinoma of the pancreas with cisplatin and etoposide. Cancer 64, 2007–2009.
88
Wilentz et al.
67. Solcia, E., Capella, C., and Klöppel, G. (1997) Atlas of tumor pathology: Tumors of the pancreas, 3rd Series ed.Washington, D. C.: Armed Forces Institute of Pathology. 68. Sessa, F., Bonato, M., Frigerio, B., et al. (1990) Ductal cancers of the pancreas 68 frequently express markers of gastrointestinal epithelial cells. Gastroenterology 98, 1655–1665. 69. de Alava, E., Torramadé, J., and Vázquez, J. J. (1993) Leiomyosarcoma of the pancreas. Virchows Arch. [A] 422, 419–422. 70. Elliott, T. E., Albertazzi, V. J., and Danto, L. A. (1980) Pancreatic liposarcoma. 70 Case report with review of retroperitoneal liposarcomas. Cancer 45, 1720–1723. 71. 71 Ishikawa, O., Matsui, Y., Aoki, Y., Iwanaga, T., Terasawa, T., and Wada, A. (1981) Leiomyosarcoma of the pancreas. Report of a case and review of the literature. Am. J. Surg. Pathol. 5, 597–602. 72. Pascal, R. R., Sullivan, L., Hauser, L., and Ferzli, G. (1989) Primary malignant 72 fibrous histiocytoma of the pancreas. Hum. Pathol. 20, 1215–1217. 73. 73 Ross, C. F. (1951) Leiomyosarcoma of the pancreas. Br. J. Surg. 39, 53–56. 74. 74 Suster, S., Phillips, M., and Robinson, M. J. (1989) Malignant fibrous histiocytoma (giant cell type) of the pancreas. A distinctive variant of osteoclast-type giant cell tumor of the pancreas. Cancer 64, 2303–2308. 75. Danner, D. B., Hruban, R. H., Pitt, H. A., Hayashi, R., Griffin, C. A., and Perlman, 75 E. J. (1994) Primitive neuroectodermal tumor arising in the pancreas. Mod. Pathol. 7, 200–204. 76. 76 Urban, B. A., Fishman, E. K., Hruban, R. H., and Cameron, J. L. (1992) CT findings in cystic schwannoma of the pancreas. J. Comput. Assist. Tomogr. 16, 492–496. 77. Brown, D., Lydon, J., McLaughlin, M., Stuart-Tilley, A., Tyszkowski, R., and 77 Alper, S. (1996) Antigen retrieval in cryostat tissue sections and cultured cells by treatment with sodium dodecyl sulfate (SDS). Histochem. Cell Biol. 105, 261–267. 78. 78 Iacobuzio-Donahue, C. A., Ryu, B., Hruban, R. H., and Kern, S. E. (2002) Exploring the host desmoplastic response to pancreatic carcinoma: Gene expression of stromal and neoplastic cells at the site of primary invasion. Am. J. Pathol. 160, 91–99. 79. Kononen, J., Bubendorf, L., Kallioniemi, A., et al. (1998) Tissue microarrays for 79 high-throughput molecular profiling of tumor specimens. Nat. Med. 4, 844–847. 80. 80 Mucci, N. R., Akdas, G., Manely, S., and Rubin, M. A. (2000) Neuroendocrine expression in metastatic prostate cancer: Evaluation of high throughput tissue microarrays to detect heterogeneous protein expression. Hum. Pathol. 31, 406–414. 81. Kallioniemi, O. P., Wagner, U., Kononen, J., and Sauter, G. (2001) Tissue microarray technology for high-throughput molecular profiling of cancer. Hum. Mol. Genet. 10, 657–662.
Tissue Microarray Construction
89
4 Practical Methods for Tissue Microarray Construction Helen L. Fedor and Angelo M. De Marzo Summary The tissue microarray (TMA) of Kononen et al. is an extension of an idea originally developed by Battifora and consists of an array of cylindrical cores of paraffin-embedded tissue that are removed from preexisting “donor” paraffin blocks. The donor block is a standard tissue block that may be from surgical pathology, autopsy, or research material. A morphologically representative area of interest within the donor block is identified under the microscope using a stained section (usually hematoxylin and eosin stained) on a glass slide as a guide. The tissue cores are removed from the donor and inserted into a “recipient” paraffin block usually using a custom patented instrument from Beecher Instruments. Using a precise spacing pattern, tissues are inserted at high density, with up to 1000 tissue cores in a single paraffin block. Sections from this block that are cut with a microtome are placed onto standard slides that can then be used for in situ analysis. Depending on the overall depth of tissue remaining in the donor blocks, tissue arrays can generate between 100 and 500 sections. Once constructed tissue microarrays can be used with a wide range of techniques including histochemical staining, immunohistochemical/immunofluorescent staining, or in situ hybridization for either DNA or mRNA. In this chapter we present methods of TMA construction with emphasis on providing detailed tips and techniques. Key Words: Tissue microarrays; image analysis; bioinformatics; immunohistochemistry.
1. Introduction Tissue microarrays (TMAs) are emerging as a breakthrough in our ability to analyze rapidly the expression of existing and new biomarkers using archival pathology specimens. Multitissue blocks were first introduced by Battifora et al. in the so-called “sausage” or Multi-Tissue Tumor Block (MTTB), where up to 100 separate tissues were processed together into a single paraffin block (1). Recently, Kononen et al. introduced a new method of combining multiple tissues into a single paraffin block that uses a novel sampling approach with regular size and shaped tissues. This allows for many more specimens to be precisely From: Methods in Molecular Medicine, Vol. 103: Pancreatic Cancer: Methods and Protocols Edited by: G. Su © Humana Press Inc., Totowa, NJ
89
90
Fedor and De Marzo
arrayed into a single paraffin block (2). Several sources of information are available for tissue microarray protocols, tips, techniques, and troubleshooting. These include recent reviews (3–7), a detailed web site with protocols (http:// www.yalepath.org/DEPT/research/YCCTMA/YTMA_protocol.pdf), and sites developed by the NIH (http://resresources.nci.nih.gov/tarp/; http://www.nhgri. nih.gov/DIR/CGB/TMA). Tissue microarray construction entails several key aspects that will be presented as an overview here, with details found in Subheading 3. These aspects include the following: (1) purpose of the TMA; (2) TMA design; (3) selection of appropriate tissue blocks; (4) identifying regions of interest within donor tissue blocks; (5) data handling; (6) array construction; (7) TMA block sectioning; (8) TMA slide staining; and (9) image handling (optional). In this chapter we present detailed methods of TMA construction, using our experience constructing more than 320 TMAs containing more than 45,000 tissue cores as a guide. In depth information regarding array construction and troubleshooting is also available form the Beecher Instruments website (http://www.beecher instruments.com/index.html) and instruction manual. 1.1. Purpose of the TMA There are unlimited types of TMAs that may be of use depending on the project at hand. We have found that very useful arrays can be constructed from an assortment of normal human tissues obtained as discarded material from surgical specimens. These arrays are ideal for working up new antibodies or new probes for in situ hybridization. A second general type of array is one with a small number of diseased tissues from several patients. For example, when evaluating a new antibody for prostate cancer we often use an array with 20 cases of tumor and normal to quickly assess whether the marker appears of interest. A more sophisticated TMA can be constructed containing samples from many more patients (hundreds or more), and this may consist of a set of several individual TMA blocks. The true power of this type of approach is to correlate staining results with clinical outcome, if available. If possible, it is always advisable to include some matched normal tissue from which the tumor or other diseased tissue is derived. For our prostate cancer arrays, we provide a matched normal appearing epithelial sample from each sample of cancer. 1.2. TMA Design Depending on the purpose of the TMA, the design may vary greatly. There are no established guidelines for TMA design. For pancreatic cancer, in which the prediction of clinical outcome is often not an important issue, the main purpose in constructing a TMA is to simply assemble a convenient series of
Tissue Microarray Construction
91
Table 1 Typical Core Spacing and Number of Cores Using Various Needle Sizes Needle size 0.6 mm 1.0 mm 1.5 mm
Spacing between samples
Array format
Total number of cores
0.2 mm 0.3 mm 0.4 mm
20 20 cores 16 13 cores 11 9 cores
400 208 99
patients into one or a few TMA blocks. For example, we have produced arrays of invasive pancreatic adenocarcinoma, pancreatic intraepithelial neoplasia, intraductal pancreatic mucinous neoplasms, and so forth. These arrays typically contain between 50 and 100 patient samples. One key consideration is the size of the punch. Four sizes are available: 0.6 mm, 1.0 mm, 1.5 mm, and 2.0 mm. Although some investigators prefer the 2.0mm size, these large cores may damage the donor and recipient blocks (see Note 1). If larger cores are desired, the 1.5-mm core may be acceptable, as it seems to cause less damage but still covers a large area. The 1.5-mm core is the preferred core size of most of the pancreatic cancer arrays that have been constructed at our institution, as pancreatic cancer can be so sclerotic and tumor cells are often widely dispersed. For the majority of TMAs in our laboratory, however, the 0.6-mm are used, as many studies have shown that one to four cores of this size from a given tissue yield as much information as standard tissue sections (8–12). The number of cores that can be put into a single TMA block is another important consideration. The reported record number of cores is 1200. But it is usual not to place more than 600 (see Note 2). The spacing between samples is the next consideration. Spaces from 0.1 mm and higher have been used successfully (see Note 3). Table 1 shows a convenient spacing pattern between cores and the number of cores that can be put into the array block without rotating the block during construction. The array is usually laid out in an electronic spreadsheet. Figure 1 shows a sample “map” used to design and construct an example array utilizing the 0.6mm punch. The resulting array will have 10 cores across (x-axis) and 6 cores down (y-axis). This TMA is designed to have four replicate cores from each of the sample areas (individual tissue diagnoses), with the tumor and normal tissues from each case in the same row. All pertinent information that identifies a sample should be included in the map design. In most of our TMAs we place control normal tissues or cell lines in strategic regions throughout the blocks. Any kind of control tissue may be used: normal tissues, cell lines, or animal tissue (see Note 4). Often, we place entire columns of various control tissues between the tumor and normal tissue, or
92
Fedor and De Marzo
Fig. 1. Example of spreadsheet containing TMA core data.
asymmetrically at one end of the block. It is quite important to never construct an array block with complete symmetry, as one can lose orientation. For cell lines we fix cells in 10% neutral buffered formalin at room temperature overnight and embed them in 1% agarose prepared in phosphate-buffered saline (13). These cell plugs are then processed into paraffin blocks just as if they were pieces of tissue, making them available as donor blocks for TMAs. Navigating a stained TMA slide can be cumbersome in that it is easy to lose one’s place regarding the x- and y-coordinates. Building in an orientation marker that indicates where the array begins can be very helpful. As first shown to our group by the M. A. Rubin laboratory (now at Harvard University), we use five tissue cores arranged in a plus sign (+), three cores placed vertically intersecting three cores placed horizontally. This is positioned in the upper left hand corner of the block. Although many TMA laboratories prefer to group array cores in separate regions of the blocks, we do not do this, as this can hamper automated image acquisition, at least on some systems. 1.3. Tissue Selection The selection and collection of tissue blocks to be included in TMAs is the most time-consuming aspect of the entire project. Appropriate samples from clinical specimens are identified from the pathology archives of the given insti-
Tissue Microarray Construction
93
tution. Glass slides corresponding to the entire cases for surgical pathology specimens are retrieved and reviewed to select appropriate candidate blocks (see Notes 5 and 6). Corresponding blocks are then obtained from the tissue archives. The area to be sampled, which usually represents a region corresponding to a specific pathological diagnosis (“tissue diagnosis”) (14) or relevant normal tissue, is circled with a Pilot Pen (or similar xylene-free pen) by a pathologist or highly trained technician. Because the block may contain more than one tissue diagnosis of interest for obtaining tissue cores for a TMA, each circled tissue diagnosis on the slide is also assigned a letter or number. 1.4. TMA Construction Details of array construction and TMA block sectioning are presented below. We recommend that you familiarize yourself with the equipment by constructing a practice array. The first couple of array blocks made will be far from perfect. By the third, expert status should be achieved (see Notes 7 and 8). 1.5. TMA Block Sectioning Sectioning an array can be accomplished using standard sectioning procedures, although it is highly recommended to use a dedicated microtome. Prior to use the recipient blocks are faced off on this same microtome. In this fashion the amount of realignment of the microtome block holder is minimized. All the array blocks produced in the laboratory will have the same orientation, maximizing the number of slides that an array will yield (see Note 9). Changing blades frequently is advisable (see Note 10). Several investigators use the “Paraffin Sectioning Aid System” from Instrumedics Inc. This is called the Tape Transfer Method. Because we are still in the process of perfecting this method (see Notes 11 and 12), we refer the reader to the Instrumedics website (http://www.instrumedics.com). 1.6. TMA Slide Staining TMA slides can be stained with any stain that can be used for standard slides, with out any modifications (see Notes 11 and 12). 1.7. TMA Slide Data Handling In terms of TMA data handling, many investigators simply view the slides under a microscope and record their observations for each spot on paper or directly into a spreadsheet. Although presentation of detailed methods of TMA data handling are beyond the scope of this chapter, it should be pointed out that several highly specialized systems are in place and under development to store TMA data, as well as to scan TMA slides using automated imaging acquisition and to either score the images on the computer screen, or to have the system
94
Fedor and De Marzo
perform automated image analysis (see refs. 14–16 and http://www.pathology. pitt.edu/apiii02/Sci-DeMarzo.htm). The method of data storage and retrieval differs somewhat in the different systems. For example, data can either be entered into a spreadsheet (15), where they can be reformatted suitably for hierarchical clustering analysis or other statistical analysis, or be directly entered into a relational database (14; http://www.pathology.pitt.edu/apiii02/Sci-DeMarzo.htm), in which it too can be exported in a form suitable for statistical or other analysis. Demonstrations of some of these systems, in which the noncommercial software components can be obtained, are available online (http://tmaj.pathology. jhmi.edu; http://genome-www.stanford.edu/TMA/explore.shtml). 2. Materials 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12.
Manual tissue arrayer (Beecher Instruments, Sun Prarie, WI). Tissue Array punches, sizes 0.6 mm, 1.0 mm, 1.5 mm (Sun Prarie, WI). Paraplast X-tra (Fisher Scientific, Suwanee, GA, cat. no. 23-021-401). Cap gap slides (Fisher Scientific, Suwanee, GA, cat. no. 12-548-6A). Oven (Fisher Scientific, Suwanee, GA, cat. no. 11-695-1). Magnifier on stand with attached light (Fisher Scientific, Suwanee, GA, cat. no. 8-882). Stainless steel molds, 6 mm (Allegiance, Columbia, MD, cat. no. M7300-4). Tissue cassettes (Allegiance, Columbia MD, cat. no. TN045). Floatation water bath (Allegiance, Columbia MD, cat. no. M7654-1). Accu Edge blades (Allegiance, Columbia, MD, cat. no. M7321-41). Pilot Pen ultrafine point (Register Office Supply, Baltimore MD, cat. nos. 44104, 44103, and 44102 red, blue, and black). Automated rotary microtome (Leica, Deerfield, IL, cat. no. RM2155).
3. Methods Although a semiautomated tissue microarrayer is now available from Beecher Instruments, all the methods presented here involve the manual array system. 3.1. Tissue Selection and Donor Block Preparation 1. Identify the tissue of choice using a standard tissue section on a glass slide stained with hematoxylin and eosin (H&E). An immunostained slide may also be employed (see Notes 5 and 6). 2. Circle the area to be sampled directly on the glass slide, which usually represents a region corresponding to a specific pathological diagnosis (“tissue diagnosis”) (14) or relevant normal tissue, with a Pilot Pen (or similar xylene-free pen). Assign each separate region (tissue diagnosis) its own number or letter, such that it can be uniquely identified given the case number, block designation, and tissue diagnosis designation.
Tissue Microarray Construction
95
3. Overlay the circled H&E glass slide and the area of interest and identify the corresponding region on the block and circle with a laboratory marker (see Note 13). 4. The wax is melted at 56–59°C and then poured into a deep mold. A standard tissue cassette is placed on top (see Note 14). 5. Allow the block to chill completely at room temperature. After cooling, separate the cassette and mold.
3.2. Equipment Setup for TMA Construction 1. Each block and slide pair should be arranged in the order that they will be used following the map design. 2. Place the recipient block into the base plate and position the base plate on the arrayer (see Note 9). 3. Position a pair of punches in the arrayer (see Note 15). The larger one is used for extracting tissue from the donor block and is placed in the right punch holder (for right-handed persons). The smaller one is for making a hole in the recipient block where the donor tissue will be positioned, and is placed in the left punch holder (use the opposite configuration for left handed persons; see Note 16). 4. To ensure the alignment of the punches, first move the recipient punch into position and make a mark in the paraffin. Then do the same for the donor punch. These marks should coincide precisely (see Notes 17 and 18). 5. Move the needles to the position of the first punch with the x- and y-axis micrometer adjustment knobs. The position of the punches over the block can be assessed by gently pushing down on them until a mark is made in the paraffin, continuing to make adjustments with the micrometer knobs until the desired position is attained. 6. Zero the micrometers.
3.3. Array Construction (Assuming 0.6-mm Punches) At this point it may be necessary to adjust the depth stop, which dictates how deep the punch will be in the recipient block. Tightening or loosening the nut at the top left of the turret accomplishes this. The nut stops the downward movement of the turret to achieve the desired depth (see Note 19). 1. Apply pressure to the top of the turret to bring down the needle. A squeezing motion is used to push the tissue core into the paraffin. 2. Rotate the arm of the punch to the left and then back to the right, while maintaining pressure on the turret top. This rotating motion helps free the core from the recipient paraffin block (see Notes 20 and 21). 3. Release the pressure on the turret, allowing the springs to raise the turret to its resting position. 4. Press down on the stylus to eject the paraffin core from the punch and examine its length (see Note 22). 5. Make any necessary adjustment with the depth stop nut. 6. Swing the turret to the right to allow the donor needle to be brought into proper position.
96
Fedor and De Marzo
Fig. 2. Correct and incorrect core placement. The placement of the core into the recipient block is the most challenging part of array construction. (A) Incorrect placement of tissue cores in a recipient block. (B) Correct placement. Tissue protruding from the surface of the block can be gently pressed down with a clean glass slide, until it becomes flush with the surface.
7. Place the bridge over the recipient block. 8. The donor block with the area of interest circled is placed on the bridge under the needle and a punch is taken from the inscribed sample area by repeating the procedure for removing the paraffin core from the recipient block (see Notes 1, 19, 21, 23, and 24). 9. The bridge and donor block are removed. 10. The core is inserted into the previously made hole by bringing the turret down until the lower punch surface is directly over the hole that was just created. 11. Keep the pressure on the turret and slowly push down on the stylus, guiding the core into the hole while expelling the tissue from the punch. This sample will have the coordinates 01, 01 from the map. See Fig. 2 for examples of proper and improper core placement (see Notes 20, 25–28). 12. Move the x-axis micrometer knob to 0.8 mm and repeat (see Notes 22, 29, and 30). 13. Continue across the row until the last core has been placed. 14. Now use the x-axis micrometer knob to go back to position zero and move the yaxis knob to 0.8 mm, and repeat until the block is complete (see Notes 31 and 32).
3.4. TMA Block Sectioning 1. On completion of the array, remove it from the base plate and place into a 37°C oven for 15 min face down on a clean glass slide. This facilitates the adherence of the cores to the walls of the punches in the recipient block. 2. Remove the slide/block combination from the oven and apply gentle and even pressure. This evens out any irregularities in the block surface and removes the bulging of the block center that occurs during array construction (see Note 33). 3. Immediately place the TMA block onto a block of ice and wait until the complex is completely cooled before disassembly. 4. Set the temperature in the water bath to 30°C. Section blocks at 4–5 µm, always placing sections on the slide in the same orientation (see Notes 34 and 35).
Tissue Microarray Construction
97
5. Dry the slides overnight in a vertical position. 6. After drying slides place slides front to back and put into a stack, wrap tightly with Parafilm, label, and then store at 20°C (see Note 36).
3.5. Preparation of Cell Culture Blocks as TMA Block Donors 1. Grow cells using your preferred method, depending on the cell line. 2. When cell cultures are 50–75% confluent, detach cells and resuspend in 10% (v/v) phosphate-buffered formalin at room temperature. The time of fixation can vary depending on the application; we typically fix over night. 3. Pellet fixed cells by centrifugation at 500g for 10 min, wash once in 1X PBS, and pellet again. 4. Resuspend cell pellets in an equal volume of 0.8% agarose (prepared in 1X PBS) at 42°C. 5. Prefill the tapered end of a 0.6-mL microfuge tube with agarose and let solidify. 6. Transect a 1000-µL plastic pipet tip approximately 3 mm from the end of the tip using a razor blade. 7. Using the transected pipet tip, transfer the agarose/cell mixture to the microfuge tube that was previously filled with agaraose. 8. While the agarose is still melted, add a wooden toothpick to the tube. This facilitates removal of the cells when the agarose is solidified. 9. After the agarose has solidified, remove the cell plug using the embedded toothpick and section the plug in half, generating two cell blocks. 10. Process the agarose plugs into paraffin blocks using standard tissue processing. Several plugs can be placed into a single paraffin block.
4. Notes 1. Donor blocks placed under a low wattage bulb will warm up slightly, making them a little softer, less likely to crack, and easier to punch. 2. Because of the limitation in travel along the x-axis, to put more than 600 cores into a single TMA block, it is necessary to remove the block from the base plate and rotate the block 180°. If the block is removed, it is difficult to realign the rows and columns. 3. During construction the block will begin to bulge up. Incorporating space between cores minimizes this. Greater spacing will cause less bulging. 4. The use of control tissue is optional, but often quite helpful when examining new and existing immunohistochemical or other markers. In addition, it can serve as a quality control during block construction. For example, once a number of tumor tissues with similar morphologies are placed into a TMA, there is no simple way to check if the tissues were correctly placed. If the array was designed with liver tissue to be placed in position X5,Y3, then the expectation is that the final block will contain such tissue in the proper location. Finding a number of such tissues in place is quite reassuring that the block was indeed correctly constructed. 5. Cut a new H+E before evaluating the blocks to be used. This will ensure that the tissue has not been cut through and that the tissue diagnosis is correct.
98
Fedor and De Marzo
6. Performing immunocytochemistry stains prior to selection of tissue samples will help evaluate the quality of fixation. 7. Practice with the 1.0- or 1.5-mm punch before beginning your array. It is easier to visualize what you are doing with the larger size punches. 8. It may be necessary to use a magnifier during array construction. We use a magnifier with an attached light. 9. It is important to face off the recipient block using a dedicated rotary microtome prior to use. This helps ensure that the block face is smooth and that all the arrays that are made will be in the identical plane. This minimizes the amount of block realignment that will be necessary during sectioning and helps to optimize the number of complete sections that an array block will yield. The backside of the block should also be checked to to ensure that it is level as well. 10. Use the practice block to test out the proper water bath temperature before cutting the first array block. 11. Immunocytochemistry using the capillary gap method will not work with the tape transfer slides, unless paired with a blank capillary gap slide. 12. The Ventana autostainer is not compatible with tape transfer slides; the oil cover slip seems to interfere with the staining process. 13. If the area within the block that is selected contains tissue that is quite thin, it may be advantageous to select an alternate sample. Thin donor blocks will not yield many Tissue Microarray slides. In general, the tissue in the donor block should be at least 1.5 mm thick, although at times certain types of samples may not contain this much tissue. In this case, one must decide on the relative merits of using a TMA vs simply cutting standard blocks and staining standard slides. For very precious specimens such as brain biopsies we have successfully prepared TMAs using quite thin specimens. 14. Although other types of paraffin can be employed, in our recipient blocks we use Paraplast X-tra, which is softer than standard paraffin. We also use deep molds. Care should be exercised to avoid trapping air bubbles under the cassette. 15. Check all needles on receipt; some arrive bent. 16. The most agility is required while expressing the tissue into the recipient block. A margin of 2.5 mm of paraffin is recommended around the entire array. This prevents the recipient block from cracking during construction. 17. Overtightening of the screws on the arrayer can damage components. It is necessary to turn them only until they are tight. 18. Adjustments can be made to the positions of the punch in the holder. There are four screws holding the V-block; they all need to be loosened before turning the front to back setscrew. Clockwise turning will bring the punch forward. Counterclockwise turning will bring it backward. Left-to-right alignment is achieved using the setscrews on the left and right sides of the turret. 19. The depth of the recipient core can be set by the depth stop, which is located in the upper left hand corner of the turret. But there is no depth stop incorporated into the design of the Tissue Arrayer that will dictate the depth of the donor core. This depth cannot be the same when you are going from one block to another because
Tissue Microarray Construction
20. 21.
22.
23. 24. 25.
26.
27. 28.
29. 30. 31. 32.
99
tissue blocks are of varying thickness, which will determine the limit of the depth of the cores that can be taken. Beecher has recently engineered a Depth Stop Kit, which will aid in extracting reproducible lengths of donor cores. Although an experienced person may not need these, a novice will find them useful. Always return the arms of the needles to their original positions. This ensures that the orientation of all the cores will be reproducible. If the core sticks in a block, rotate the stylus back and forth, this will help free the core from the block. In addition, tamping the top of the stylus will help to remove it. After doing several punches the stylus/punch complex may retain some paraffin. Moving the stylus up and down and wiping it with a Kimwipe will remove the paraffin. Using xylene is not necessary and may damage the punch. Occasionally it is necessary to remove the stylus from the punch and wipe it off. You can make your own depth stop by cutting a small section of a thin transfer pipet to the desired length and placing it at the top of the arraying needle. There is no depth stop to aid in removal of tissue samples from the donor block. This is refined with some practice. Do not push the donor punch all the way into its respective hole; let it protrude slightly. It can then be gently pressed in further with a clean glass slide. This ensures that its position is level with the recipient block surface, not in too deep or protruding out. It is possible to cut down the size of a donor core that is too long. Eject the core from the stylus and place on a clean flat surface. Use a clean razor blade to cut the core to the desired length. The core can than be placed into the recipient block with a pair of forceps. This is easier with the larger size punches. It is possible to remove a wrongly placed core with the donor punch. It may leave some material behind. Make a note of it on the map. Cores that are placed too deeply may be removed with the smaller size needle (recipient needle). An additional core can then be inserted. This process should not be considered routine for several reasons. The hole in the recipient block may become enlarged and some residual tissue will be left behind. Cores placed erroneously may also be removed in this fashion. Be sure to make a note whenever this is necessary because the residual tissue may yield incorrect data. The array block should be wiped, brushed, or dusted after the placement of every punch. It is easiest to move horizontally across the map because the x-axis adjustment knob is more accessible. The punches can be sharpened with a fine sharpening stone if they become dull (after thousands of punches). Several accessories have been added to the Beecher product line that are useful and worth mentioning. The Four-Block indexer allows four replicate blocks to be made simultaneously. Extended hours of making tissue microarrays can lead to wrist fatigue and aggravate carpal tunnel syndrome. The Motorized Positioner for the Manual Arrayer can alleviate this.
100
Fedor and De Marzo
33. To avoid damaging the array block when you take it from the 37°C oven, practice with a couple of empty recipient blocks to familiarize yourself on how the block/ slide complex feels after warming. 34. Also, make sure the section is placed parallel with the slide edge. Having the sections in the proper orientation on the slide greatly facilitates scanning of the slide with automated slide scanners. 35. Recently developed specialized tools are available to help align the block in the microtome to minimize tissue loss, although they are not compatible with all microtomes (http://www.microm.de/products-accessories-Histo.htm). 36. When sectioning a TMA block, we typically cut 20 sections at a time. No published reports on the best methods of storing TMA slides are available. While some are stored under nitrogen gas, we store unstained sections without baking the slides and at –20°C. We have had good success with this method with several antibodies over several months’ time. However, the long-term effects of this storage method have not been examined.
Acknowledgments We would like to thank Cellie Southerland for her diligence and expertise in TMA construction and for providing us with a large number of useful insights for this manuscript. We also thank Don Vindivich for expertise and help in preparing Fig. 2. This work was funded by Public Health Services Specialized Program in Research Excellence (SPORE) in Prostate Cancer grant no. P50CA58236, The CaPCURE Foundation, and The Lee Family Foundation. References 1. 1 Battifora, H. (1986) The multitumor (sausage) tissue block: Novel method for immunohistochemical antibody testing. Lab. Invest. 55, 244–248. 2. 2 Kononen, J., Bubendorf, L., Kallioniemi, A., et al. (1998) Tissue microarrays for high-throughput molecular profiling of tumor specimens. Nat. Med. 4, 844–847. 3. 3 Moch, H., Kononen, T., Kallioniemi, O. P., and Sauter, G. (2001) Tissue microarrays: What will they bring to molecular and anatomic pathology? Adv. Anat. Pathol. 8, 14–20. 4. 4 Bubendorf, L., Nocito, A., Moch, H., and Sauter, G. (2001) Tissue microarray (TMA) technology: Miniaturized pathology archives for high-throughput in situ studies. J. Pathol. 195, 72–79. 5. 5 Rimm, D. L., Camp, R. L., Charette, L. A., Olsen, D. A., and Provost, E. (2001) Amplification of tissue by construction of tissue microarrays. Exp. Mol. Pathol. 70, 255–264. 6. 6 Rimm, D. L., Camp, R. L., Charette, L. A., Costa, J., Olsen, D. A., and Reiss, M. (2001) Tissue microarray: A new technology for amplification of tissue resources. Cancer J. 7, 24–31.
Tissue Microarray Construction
101
7. Jensen, T. A. and Hammand, M. E. H. (2001) The tissue microarray-a technical guide for histologists. J. Histotechnol. 24, 283–287. 8. Camp, R. L., Charette, L. A., and Rimm, D. L. (2000) Validation of tissue micro8 array technology in breast carcinoma. Lab. Invest. 80, 1943–1949. 9. 9 Torhorst, J., Bucher, C., Kononen, J., et al. (2001) Tissue microarrays for rapid linking of molecular changes to clinical endpoints. Am. J. Pathol. 159, 2249–2256. 10. 10 Rubin, M., Dunn, R., Strawderman, M., and Pienta, K. J. (2002) Tissue microarray sampling strategy for prostate cancer biomarker analysis. Am. J. Surg. Pathol. 26, 312–319. 11. Nocito, A., Bubendorf, L., Maria Tinner, E., et al. (2001) Microarrays of bladder 11 cancer tissue are highly representative of proliferation index and histological grade. J. Pathol. 194, 349–357. 12. Hoos, A., Urist, M. J., Stojadinovic, A., et al. (2001) Validation of tissue micro12 arrays for immunohistochemical profiling of cancer specimens using the example of human fibroblastic tumors. Am. J. Pathol. 158, 1245–1251. 13. Meeker, A. K., Gage, W. R., Hicks, J. L., et al. (2002) Telomere length assess13 ment in human archival tissues: Combined telomere fluorescence in situ hybridization and immunostaining. Am. J. Pathol. 160, 1259–1268. 14. Manley, S., Mucci, N. R., De Marzo, A. M., and Rubin, M. A. (2001) Relational 14 database structure to manage high-density tissue microarray data and images for pathology studies focusing on clinical outcome: The prostate specialized program of research excellence model. Am. J. Pathol. 159, 837–843. 15. Liu, C. L., Prapong, W., Natkunam, Y., et al. (2002) Software tools for high15 throughput analysis and archiving of immunohistochemistry staining data obtained with tissue microarrays. Am. J. Pathol. 161, 1557–1565. 16. Camp, R. L., Chung, G. G., and Rimm, D. L. (2002) Automated subcellular localization and quantification of protein expression in tissue microarrays. Nat. Med. 8, 1323–1327.
102
Fedor and De Marzo
Xenografting Ductal Pancreatic Adenocarcinomas
103
5 Xenografting and Harvesting Human Ductal Pancreatic Adenocarcinomas for DNA Analysis Kimberly Walter, James Eshleman, and Michael Goggins
Summary Xenotransplantation (xenografting) of primary cancers or cancer cell lines into immunodeficient mice is a commonly used technique to assess tumor growth in response to a variety of experimental agents. When primary pancreatic cancers are xenografted, cancer cells proliferate in the mouse, but the human stroma does not. This growth pattern enables facile genetic analysis of cancer genetics profiles without the contamination of admixed stromal cells typical of primary cancers. Key Words: DNA; pancreatic cancer; xenograft.
1. Introduction Most pancreatic adenocarcinomas are composed of a complex mixture of normal and neoplastic cells. Non-neoplastic cells present in the tumor include stromal, vascular endothelial, and inflammatory cells (1). This cellular heterogeneity frequently confounds studies of DNA isolated from such tumors, which may contain more normal cell DNA than tumor DNA. Therefore, methods are needed to obtain pure tumor cell populations for DNA analysis. Human tumor xenografting allows the simultaneous isolation and expansion of tumor cells. Tumors are subcutaneously implanted into immunodeficient athymic mice, which lack T-cells and are therefore the ideal system for growth of cancer cells (2,3). This is a fertile environment for tumor cells to proliferate, and produces greater neoplastic cellularity in the harvested tumor. This chapter presents a summary of the methods used to xenograft and harvest primary pancreatic adenocarcinomas and cell lines.
From: Methods in Molecular Medicine, Vol. 103: Pancreatic Cancer: Methods and Protocols Edited by: G. Su © Humana Press Inc., Totowa, NJ
103
104
Walter et al.
2. Materials 2.1. Primary Tumor Xenografting 1. Mice: We use 3- to 4-wk-old male CD1 NuNu (nude) mice. 2. Human ductal pancreatic adenocarcinoma fresh section, at least 3 mm3 per site of implant (we implant two mice, two sites per mouse for each tumor), stored in minimal essential media (MEM) containing 1% penicillin–streptomycin, kept on ice (see Note 1). 3. Sterile surgery pack: One 250-mL beaker containing a piece of aluminum foil (large enough to cover the beaker) and dissecting instruments, which include two pairs of fine dissecting scissors, two pairs of fine forceps, and one pair of gripping forceps. Wrap the beaker in two autoclave-proof cloth towels, tape tightly with autoclave tape, and autoclave the pack to sterilize. 4. Betadine solution. 5. Ether anesthetic, diluted to 1.9% (see Note 2). 6. Sterile absorbable surgical sutures, 4-0 chromic gut. 7. Sterile Petri dishes, 10 cm diameter. 8. Single-edge razor blades. 9. Gauze pads, 7.6 cm ´ 7.6 cm. 10. Matrigel basement membrane matrix, store in 0.2-mL aliquots at -20°C; thaw at time of xenografting (see Note 3). 11. Laminar flow hood.
2.2. Cell Line Xenografting (Additional Materials Needed) 1. 2. 3. 4. 5. 6.
Pancreatic adenocarcinoma cell lines in culture. Versene 1:5000. Trypsin-EDTA, 0.05%. 1-mL syringes, one per sample. 15-mL conical tubes, one per sample. Hemacytometer.
2.3. Xenograft Tumor Harvesting (Additional Materials Needed) 1. Calipers. 2. Cryovials. 3. Liquid nitrogen.
3. Methods 3.1. Primary Tumor Xenografting 1. In a laminar flow hood, open the surgery pack and place the beaker and dissecting tools on the cloth towel. The towel will be used as the sterile surface for surgery. 2. Allow the aliquot of Matrigel to thaw and pour it into a sterile Petri dish. 3. Place the tumor in the Matrigel and cut it into four pieces approx 0.3–0.5 cm in size. Coat the tumor in Matrigel before implanting it.
Xenografting Ductal Pancreatic Adenocarcinomas
105
Fig. 1. Tumor implantation. (A) Subcutaneous pocket.
4. Apply a small amount of ether to a gauze pad and place it inside the beaker. The beaker will be used as an anesthesia chamber. Transfer one mouse from its cage to the beaker and cover the beaker with the aluminum foil. Keep the mouse inside the anesthesia chamber until the mouse is unconscious (see Note 4). 5. Place the mouse on the sterile cloth. Using the gauze pads, apply betadine thoroughly to the right and left scapular regions of the mouse (to sterilize the skin at the site of implantation). 6. Place the mouse on its side. Lift the skin with the forceps and use the scissors to make an incision 3–4 mm in length in the scapular region. Insert the forceps into the interscapular space and make a pocket for the tumor by opening and sliding the forceps toward the head of the mouse (see Note 5) (Fig. 1A). 7. Implant a piece of tumor at least 3 mm3 in size, using the forceps to slide it into the subcutaneous pocket (see Note 6) (Fig. 1B). Be sure that the tumor is secured in the pocket. 8. Using the gripping forceps, make one or two sutures to close the incision. Hold the skin of the mouse with the fine forceps and the suture needle with the gripping forceps to pull the suture through the skin (Fig. 1C). Tie the suture twice and cut the remaining ends close to the knot. The sutures will be absorbed by the mouse. 9. Repeat the procedure on the opposite scapula of the mouse. If the mouse is regaining consciousness, place it back in the anesthesia chamber just until the breathing rate slows again. 10. After completing the surgery, return the mouse to its cage and place it on its side. Check after several minutes that the mouse has regained consciousness (see Note 7).
106
Walter et al.
Fig. 1. (B) Tumor insertion. (C) Suture.
3.2. Cell Line Xenografting by Injection 1. In a tissue culture flask, grow the number of cells desired for xenograft injection (1–5 million cells per injection site). Remove the media from the flask and wash the cells with 5–10 mL of Versene. Harvest the cells from the flask by adding 5–10 mL of trypsin. Transfer the cells to a 15-mL conical tube and centrifuge to obtain a cell pellet. 2. Resuspend the cells in 1 mL of MEM and count. 3. Aliquot the desired number of cells into 1.5-mL tubes and centrifuge. We inject 1 or 5 million cells in 200 µL of serum-free MEM.
Xenografting Ductal Pancreatic Adenocarcinomas
107
Fig. 2. Measurement of tumor size. 4. Resuspend the cells in the appropriate volume of MEM. Plan for extra cell volume. If you are injecting 1 million cells/200 µL at two sites, you will need 400 µL total volume; therefore, prepare about 2.5 million cells in 500 µL cell volume. Keep the cells on ice until injection. 5. Inside the laminar flow hood, set up the sterile working surface. Anesthetize the mouse as described for primary tumor xenografting. 6. Sterilize the skin at the site of injection (scapula) with Betadine. 7. Using sterile technique, remove the needle from the syringe and aspirate the appropriate volume of cell suspension into the syringe, being careful not to aspirate air into the syringe. If air is aspirated, release it from the syringe before injecting the cells, to avoid injecting air subcutaneously. Reattach the needle to the syringe. 8. Lift the skin with the forceps and inject the cells subcutaneously (see Note 8). 9. Repeat the injection on the opposite scapula of the mouse as described. Return the mouse to the cage, placing it on its side. Check the mouse after 15–20 min to confirm that it has regained consciousness.
3.3. Xenograft Tumor Harvesting 1. Inside the laminar flow hood, set up the sterile working surface and anesthesia chamber as described. 2. Anesthetize the mouse, place it on the working surface, and use the calipers to measure the tumor size (Fig. 2). 3. Using forceps and scissors, make an incision at the base of the tumor and cut the skin around the circumference of the tumor (Fig. 3A). The tumor will be encapsulated apart from the abdominal wall. Once the skin is cut around the circumference of the tumor, the tumor should begin to fall away from the abdominal wall and can easily be excised from the wall (Fig. 3B).
108
Walter et al.
Fig. 3. Tumor harvest. (A) Cutting around circumference of tumor. (B) Removal of encapsulated tumor. 4. Transfer the tumor to a Petri dish and use the forceps and scissors to remove the skin. 5. With a razor blade, cut the tumor into sections of desired size, place in cryovials, and snap-freeze in liquid nitrogen. Store the harvested tumor at -80°C (see Notes 9 and 10).
4. Notes 1. If the tumor cannot be xenografted immediately after resection from the patient, store the tumor on ice in MEM; do not freeze. We have successfully xenografted
Xenografting Ductal Pancreatic Adenocarcinomas
2.
3. 4.
5.
6. 7.
8.
9.
10.
11.
109
tumors up to 4 h after resection; however, it is ideal to xenograft tumors immediately after resection. We have used methoxyflurane (Metofane) as an anesthetic; however, Metofane is no longer available for purchase in the United States. Alternatives include halothane and isoflurane, which volatize to a higher percentage than methoxyflurane, producing faster induction and awakening. These agents can easily overanesthetize and kill. We have successfully used a third alternative, ether, which is also more volatile than methoxyflurane and must be used carefully to avoid overanesthesia. The effective concentration of ether to produce anesthesia is 1.9%, which can be produced with 0.8 mL/L of volume. The use of Matrigel as a vehicle for subcutaneous engraftment of human carcinomas greatly enhances the growth of the implanted carcinomas (4,5). Keep the mouse in the anesthesia chamber until the breathing rate slows to approximately one breath per second. Begin surgery as soon as the breathing has slowed to this rate, to avoid overdosing the mouse. The younger the mouse, the more sensitive it will be to the anesthetic. It is critical to make a pocket under the skin before implanting the tumor, to keep the tumor under the skin while suturing the incision. It may be necessary to slide the forceps forcefully under the skin to make a pocket. Tumor sections less than 3 mm3 in size do not grow well when implanted. We implant sections between 3 and 5 mm3 in size. The amount of time required for the mouse to regain consciousness is proportional to the amount of anesthetic to which it was exposed. It may take 15–20 min for the mouse to regain consciousness. On injection of the cells, a small bump will form. If the cells have been injected too superficially (intradermally), a bleb will form and the cells may leak out of the mouse. To avoid an intradermal injection, track the needle in slightly before injecting. Tumor tissue can also be fixed in formalin for histological examination (6). Alternatively, the tumor can be reimplanted and expanded further in another mouse. We have reimplanted tumors by first coating them in Matrigel directly after excision and then implanting them as described. Tumors can also be reimplanted after cryopreservation (7). We harvest tumors when they have grown to a size of 1.5–2.0 cm. Allowing the tumor to grow larger than 2.0 cm usually restricts the blood supply to the tumor and results in necrosis of cells in the tumor bed. Most primary tumors will take within 3–4 mo of implantation. Usually, a tumor nodule will remain visible under the skin from the time of implantation to the time that tumor growth begins. If there is no tumor growth or visible nodule after 3 mo, it is unlikely that the tumor will take at all. However, we have found that in rare cases a tumor will begin growing up to 1 yr after the date of implantation. For unknown reasons, some primary cancers do not grow at all when xenografted. For cell lines, the time from injection to the time that tumor growth begins will vary, depending on the doubling rate of the cell line. Most cell lines begin growing
110
Walter et al.
Fig. 4. Common tumor growth patterns. (A) Ulcerated tumor. (B) Tumor with hemorrhagic cyst.
2–3 wk after injection. Xenografts that are harvested and reimplanted usually grow immediately. As expected, metastatic adenocarcinomas also grow faster than nonmetastatic tumors. 12. With subcutaneous xenografting, the tumor is encapsulated and fails to metastasize either regionally or distally (2,8,9). 13. The mechanical stress of the tumor on the epidermal layer may cause the tumor to ulcerate, or grow through the epidermal layer (Fig. 4A). If this occurs, harvest the tumor to prevent infection of the skin and necrosis of the tumor cells.
Xenografting Ductal Pancreatic Adenocarcinomas
111
14. Fluid-filled cysts and hemorrhagic cysts may form adjacent to the tumor. As shown in Fig. 4B, a hemorrhagic cyst has formed above the tumor. The cyst can usually be distinguished from the tumor before harvesting. If a minimum tumor size is desired, measure the tumor (fibrous mass) only and do not include the cyst in the size measurement. 15. Mouse stromal cells grow in xenografted tumors. When using methods such as PCR to study these tumors, make sure that primers are not complementary to mouse DNA.
References 1. 1 Van Weerden, W. M. and Romijn, J. C. (2000) Use of nude mouse xenograft models in prostate cancer research. The Prostate 43, 263–271. 2. Mueller, B. M. and Reisfeld, R. A. (1991) Potential of the scid mouse as a host for human tumors. Cancer Metastas. Rev. 10, 193–200. 3. Pandelouris, E. M. (1968) Absence of thymus in a mouse mutant. Nature 217, 370–371. 4. 4 Pretlow, T. G., Delmoro, C. M., Dilley, G. G., Spadafora, C. G., and Pretlow, T. P. (1991) Transplantation of human prostatic carcinoma into nude mice in Matrigel. Cancer Res. 51, 3814–3817. 5. 5 White, L., Sterling-Lewis, K., Kees, U. R., and Tobias, V. (2001) Medulloblastoma/ primitive neuroectodermal tumour studied as a Matrigel enhanced subcutaneous xeno-graft model. J. Clin. Neurosci. 8, 151–156. 6. 6 Mohammad, R. M., Dugan, M. C., Mohamed, A. N., et al. (1998) Establishment of a human pancreatic tumor xenograft model: Potential application for preclinical evaluation of novel therapeutic agents. Pancreas 16, 19–25. 7. 7 Sorio, C., Bonora, A., Orlandini, S., et al. (2001) Successful xenografting of cryopreserved primary pancreatic cancers. Virchows Arch. 438, 154–158. 8. Xinfu, F., Besterman, J. M., Monosov, A., and Hoffman, R. M. (1991) Models of human metastatic colon cancer in nude mice orthotopically constructed by using histologically intact patient specimens. Proc. Natl. Acad. Sci. USA 88, 9345–9349. 9. Manzotti, C., Audisio, R. A., and Pratesi, G. (1993) Importance of orthotopic implantation for human tumors as model systems: Relevance to metastasis and invasion. Clin. Exp. Metastas. 11, 5–14.
112
Walter et al.
Culture and Immortalization of DECs
113
6 Culture and Immortalization of Pancreatic Ductal Epithelial Cells Terence Lawson, Michel Ouellette, Carol Kolar, and Michael Hollingsworth Summary Some populations of the epithelial cells from the duct and ductular network of the mammalian pancreas have been isolated and maintained in vitro for up to 3 mo. These cells express many of the surface factors that are unique to them in vivo. They also retain significant drugand carcinogen-metabolizing capacity in vitro. In this chapter we review the progression of the methods for the isolation, culture and maintenance in vitro for these cells from the earliest when only duct/ductular fragments were obtainable to the current ones which provide epithelial cells. The critical steps in the isolation process are identified and strategies are provided to facilitate these steps. These include the selection of tissue digestive enzymes, the importance of extensive mincing before culture and the importance of roles of some co-factors used in the culture medium. Key Words: Immortalization; hTERT; senescence; telomerase; ductal epithelial cells.
1. Introduction Pancreatic ductal epithelial cells (DECs) are involved in the pathogenesis of a number of disease conditions, including cancer, pancreatitis, and cystic fibrosis; however, extensive studies of normal DECs and alterations in them that contribute to different disease conditions have not been reported. One significant problem is our inability to produce reasonable quantities of normal DECs for study. Long-term primary culture of functional pancreatic ductal epithelial cells and their precursors has remained an important goal of a few laboratories. In this chapter, we summarize previous efforts at culturing DECs, and present a summary of new methods that have been moderately successful in establishing immortalized DEC lines.
From: Methods in Molecular Medicine, Vol. 103: Pancreatic Cancer: Methods and Protocols Edited by: G. Su © Humana Press Inc., Totowa, NJ
113
114
Lawson et al.
1.1. History of DEC Culture Development of cell culture methods to produce large quantities of human pancreatic DECs, particularly from adult tissue, has been slow. Human pancreas in large volume is seldom available for laboratory use, and the pancreas itself is a complex organ with endocrine, exocrine, and ductal functions in several different cell types. Ducts comprise only 5–10% of the cells of the human pancreas. Evolution of techniques to separate cell types and of preferential media to nourish the ductal epithelium over other cell types proceeded at first on animal, usually rodent or bovine, tissues. Before 1980, these animal tissues were usually cultured by explant of small dissected ductal fragments from autopsied organs. Human tissues from fresh autopsy could be similarly derived and held in culture for up to 4 mo (1) with normal epithelial morphology. Exploration of techniques for culture of isolated ductal cells (singly or in clusters) followed. Advances in this area enabled researchers to separate ductal epithelium for culture by removing or discouraging growth of acinar cells, islet cells, and fibroblasts. Most of the developmental work was performed in animal models (rat, hamster, guinea pig). Digestion of minced pancreatic tissue followed by sieving and microscopic examination to select for ductal fragments improved the quality of starting preparations. Initial efforts to differentiate cell types included exploration of expression of selected enzymatic activities, but these were problematic. Alkaline phosphatase activity is very low in acinar cells compared with levels in ductal epithelium, but is elevated in blood vessel endothelium. Carbonic anhydrase is detectable only in large (common) ducts. Most enzyme activities declined markedly with time in culture, suggesting that suboptimal culture conditions were being employed (2). Culture was attempted on several extracellular matrices. Agarose (3) and collagen gel (4) encouraged maintenance of a typical epithelial cobblestone morphology. Matrigel (BD Biosciences) failed to encourage growth while porous cellulose ester membranes (culture inserts) produced cuboidal polar cells with numerous mitochondria and microvilli on the apical surface (5). More recently human DECs have successfully been cultured (6) on matrix secreted by the 804G cell line (7), which is stimulatory to duct cell proliferation, but not islets. Ductal tissue from explant culture when treated with methylnitrosamine produces cells that form carcinomas in mice (8). Human fetal pancreas explants grow acinar cells and ductal cells. After methyl nitrosourea treatment, ductal cells proliferate and form tumorigenic cells (9). Early attempts to use postmortem adult tissue from autopsy resulted in explant cultures that grew very slowly (10). Fetal pancreas was more easily cultured by dissecting and mincing ductal tissue, followed by explant culture (11,12). Fetal pancreatic DECs produced bicarbonate, confirming a measure of ductal function, and expressed cytokeratins 18 and 19 in a manner consistent with intact human fetal ductal tissue.
Culture and Immortalization of DECs
115
A systematic exploration of trophic growth factors in guinea pig DECs revealed differences between in vivo and in vitro growth requirements. Epidermal growth factor (EGF) is necessary for DEC proliferation (13) while a variety of hormones that affect pancreas in vivo (cerulein, carbachol, bombesin, secretin, RPA) have little effect in culture (14). Other strong mitogens for DECs include hepatocyte growth factor and insulin-like growth factor-1, with platelet-derived growth factor and gastrin active to a lesser degree. The same study showed secretin, somatostatin, and interleukin-6 (IL-6) had no effect on proliferation (15). Hepatocyte growth factor was mitogenic only for ductal cells, not for islets (6). A consensus of these studies suggests that EGF, dexamethasone or hydrocortisone, cholera toxin, and bovine pituitary extract enhance and are probably required for human DEC culture, while insulin, transferrin, selenium, and trypsin inhibitors are often useful additions to the media base. Basal media can be any of a number of common mixtures: RPMI-1640, CMRL 1066, Dulbecco’s modified Eagle’s medium (DMEM), DME/F12. Markers for ductal cells include cytokeratins 7, 8, 18, 19, and CA 19-9, which are positive for ductal cells but negative for acinar and islet cells. CK7 and CK19 are found in centroacinar and less differentiated ductal cells, CK8 and CK18 are characteristic of differentiated forms. Cystic fibrosis transmembrane conductance regulator (CFTR) and the mucin MUC1 have been demonstrated in adult human DECs (16,17). Keratin-1 can be used to confirm the epithelial nature of ductal cells or tissues (18). Carbonic anhydrase II is useful both as a marker for ductal cells and as a confirmation of their ductal function (19). Recent SAGE analysis of primary ductal cell cultures has provided a baseline of transcripts in primary epithelial cells (20). Acinar cells are positive for amylase, chymotrypsin, and g-glutamyl transferase; fibroblasts for vimentin. Common culture protocols tend to be of two types, explant or enzymatic digest. Explant cultures require large amounts of tissue from which the ductal tree is dissected, then minced and placed into plastic dishes (sometimes with a collagen matrix). In these cultures, the epithelial cells begin to grow out from the tissue explants and can form cell monolayers. Some authors suggest moving explants to new culture plates once epithelial outgrowth is established, where they may start another cellular outgrowth. Use of enzymatic digestion (generally collagenase) for tissue disruption may be applied to minced whole tissue or to dissected ductal tissue. This tissue digest may be used without modification, or ductal tissue can be selected by sieving the digest or by use of a Ficoll gradient to separate the lighter ductal clumps from lighter islets and heavier acinar tissue (13). A protocol adapted for using very small tissue samples (postsurgical) or the remains of pancreas digestion after removal of islets for transplantation (17) selects for ductal tissue by culturing the primary preparation for several days in a thick collagen gel until ductal tissue forms cysts that are removed and
116
Lawson et al.
redigested to a duct-enriched product (2). Contaminating fibroblasts may be removed from culture by scraping, use of cloning rings to select pure DEC colonies, selective trypsinization, treatment with Geneticin (15) or cholera toxin (21) in the culture medium, or by treatment with a monoclonal antibody to fibroblast surface membrane molecules followed by complement lysis (22). An early metabolic study reported human ductal explants capable of activating polycyclic aromatic hydrocarbons (benzo[a]pyrene and 7,12-dimethylbenz [a]anthracene) to metabolic intermediates that bind to DNA suggesting CYP1A1 or 2B1 activity (23). Explant-derived human ductal epithelial cells exposed to MNNG or 4-nitroquinoline-1-oxide produce morphological changes consistent with human pancreatic adenocarcinoma in vivo (24). Cultured human DECs are capable of activating heterocyclic amines to mutagenic species, which suggests CYPIA2 activity (25). Cyclic AMP can be detected after secretin stimulation (15). Human DECs from three donors metabolize nitrosamines (BOP and the tobaccospecific NNK) and a heterocyclic amine (PhIP) to mutagenic species demonstrating CYP1A2 and 2E1 activity. The same cells show a response to ethanol treatment consistent with CYP2E1 induction (17). Human fetal DEC in culture express an AE2 chloride–bicarbonate exchanger (26). Ductal epithelial cells have been observed to be phenotypically unstable in culture. Adult acinar cells produce both acinar and ductal antigens, with the ductal phenotype characteristic of conditions involving rapid growth (27). Human acinar cells have been reported to transdifferentiate to ductal phenotype in only 2 wk of culture (28). Other investigators suggest that this phenomenon is attributable to growth from stem cells in the primary tissue isolate (29). Morphology is also unstable in some mediums or culture conditions. Adult human DECs assume a fibroblastic morphology in minimal media on uncoated tissue culture plastic while retaining their epithelial cytokeratins and negative vimentin status (19). Similar changes in phenotype between islet and ductal forms seem to be related to the matrix used for culture. Rat tail collagen gels allowed islets to lose insulin production and to form cystic structures that expressed CK19 and keratin-1, which are characteristics of ductal epithelium (18,30). Further culture of the cystic forms for 7–10 d gave rise to single endocrine islet cells (18). Use of laminin-1/nidogen matrix with a tumor line sharing many components of normal DEC-induced permanent phenotypic changes in the cells (31). A ductal mono-layer isolated from nonendocrine adult human pancreas digest when overlaid with Matrigel gave rise to cystic forms that exhibited budding of islet-like insulin-producing structures (32). Ductal forms isolated from a Ficoll cell pellet after islet harvest produced less insulin than lighter forms from higher in the Ficoll gradient (32). A comparison of primary tissue from main duct (by both dissection and enzymatic digestion), from islet/duct-enriched tissue in threedimensional culture, and from acinar cells cultured in several media and on or
Culture and Immortalization of DECs
117
in varied matrices shows that expansion of all initial cell types yields cells of ductal phenotype. Tissue explants from the main duct in this study failed to maintain viability and developed fibroblast overgrowth. Further growth of the islet/duct-enriched and acinar-based cells resulted in cystic forms of ductal phenotype (33). These derived ductal cells in further culture produced insulin promoter factor 1, which is required for endocrine cell neogenesis (34). Use of similarly derived cells for reversal of insulin-dependent diabetes in mice has recently been reported (35). These observations of phenotypic flexibility suggest that careful attention should be paid to culture conditions when maintaining human DEC cultures. In summary, culture conditions and protocols have now been explored and nutrient requirements defined enough to allow investigators to use human DECs as a research tool, as opposed to an end in themselves. The next area of research emphasis is the immortalization of human DEC. 1.2. Immortalization of Pancreatic Ductal Epithelial Cells A major obstacle to the establishment of human cell lines is cellular senescence. Telomere-controlled senescence is caused by the shortening of telomeres that occurs each time normal human cells divide (36). The transfer of an exogenous hTERT cDNA, encoding the catalytic subunit of human telomerase, can prevent telomere-controlled senescence and can be used to immortalize normal human cells without causing cancer-associated changes or altering phenotypic properties (37–39). However, under current culture conditions, the growth of human pancreatic ductal epithelial cells appears to be limited by premature forms of senescence, independent of telomere shortening, that hTERT alone fail to overcome (our unpublished observations). Overcoming premature senescence is most often achieved with the use of viral oncogenes, such as the HPV16 E6/E7 or the SV40 large T antigen (40). Thus, the few lines of human pancreatic ductal epithelial cells available thus far have all been established with the use of these oncogenes (41–43). Described below is our adaptation of the procedure utilized by M. S. Tsao and collaborators (Ontario Cancer Center, Ontario, Canada) for extending the lifespan of these cells with HPV16 E6/E7 (41,42). 2. Materials 1. 2. 3. 4. 5. 6.
Cell line PA317 LXSN 16E6E7 (from ATCC CRL-2203) (44). Dulbecco’s modified Eagle’s medium (DMEM) with glutaMAX™. Cosmic calf serum (HyClone). Gentamycin. Trypsin-EDTA: 0.05% trypsin, 0.53 mM EDTA. Polysulfone filter, 0.45 µm (ALADN™).
118 7. 8. 9. 10. 11. 12. 13. 14. 15.
Lawson et al.
Polybrene. Keratinocyte serum-free medium (or KSF). Bovine pituitary extract (Gibco BRL). Epidermal growth factor (or EGF) (Gibco BRL). HPDE medium: KSF supplemented with 50 mg/mL of bovine pituitary extract and 5 ng/mL of EGF. Mg2+- and Ca2+-free Hank’s balanced salt solution. Trypsin inhibitor, soybean. Bovine serum albumin (BSA). G418.
3. Methods 3.1. Preparing Viral Supernatants Carrying HPV16 E6/E7 1. Thaw and grow PA317 LXSN 16E6E7 cells in DMEM supplemented with 10% Cosmic calf serum and 50 µg/mL of gentamycin. 2. Pass cells at a split ratio of 1:6 to 1:12 with the use of trypsin-EDTA. 3. Freeze half the cells in Cosmic calf serum containing 5% dimethyl sulfoxide (DMSO). 4. Once the remaining cells have reached 80% confluence, wash once with KSF. 5. Collect viruses overnight in a minimal volume of HPDE medium (6 mL/75 cm2). 6. Harvest supernatant, force through a polysulfone filter and add polybrene to 4 µg/mL. 7. Use viral supernatants immediately or freeze for later use.
3.2. Extending Lifespan with HPV16 E6/E7 1. In ice-cold Mg2+- and Ca2+-free Hank’s balanced salt solution, isolate by dissection fragments of large pancreatic ducts (see Note 1). 2. Cut the dissected fragments into 1-mm blocks. 3. Seed the tissues blocks in HPDE medium. 4. Once sheets of epithelial cells have emerged surrounding the seeded blocks (after 3–6 d), replace medium with a viral supernatant carrying E6/E7 (2 mL/10 cm2). 5. Incubate overnight at 37°C. 6. Remove viral supernatant and replenish cells with fresh HPDE medium. 7. Let cells divide one or twice (2–3 d). 8. Pass the transduced cells: Dissociate cells in trypsin-EDTA, neutralize the trypsin with one volume of KSF containing 0.1% trypsin inhibitor and 0.1% BSA, replate cells at a split ratio of 1:4 to 1:8 in fresh HPDE medium. 9. Select for viral integration with the addition of 400 µg/mL of G418. Apply selection for a period of 2–3 wk. Change medium every 2–3 d. 10. In medium HPDE, maintain cells in log-phase growth until crisis is induced by terminal telomere shortening. Change medium every 2–3 d. 11. Expand individual clones that overcame crisis (see Note 2). 12. Test each clone for the expression of markers of pancreatic ductal epithelial cells (see Note 3).
Culture and Immortalization of DECs
119
4. Notes 1. Other approaches may be used to purify human pancreatic ductal epithelial cells, which typically use the enzymes collagenase (see previous subheadings of this chapter) or liberase (43) to dissociate the tissues samples. 2. To allow a more efficient bypass of crisis, the E6/E7-expressing cells should be transduced with retroviral vectors carrying an hTERT cDNA. Viruses carrying hTERT can be produced by the transient transfection of vector pBabePuro-hTERT (45,46) into amphotropic packaging cells phoenix-A (Garry P. Nolan, Stanford University, CA, USA) or PT67 (ATCC CRL-12284). 3. Markers of human pancreatic ductal epithelial cells express by the immortalized cells should include carbonic anhydrase II, MUC-1, and the cytokeratins 7, 8, 18, and 19 (42).
References 1. 1 Jones, R. T., Hudson, E. A., and Resau, J. H. (1981) A review of in vitro and in vivo culture techniques for the study of pancreatic carcinogenesis. Cancer 47, 1490–1496. 2. 2 Githens, S., Holmquist, D. R., Whelan, J. F., and Ruby, J. R. (1981) Morphologic and biochemical characteristics of isolated and cultured pancreatic ducts. Cancer 47, 1505–1512. 3. Githens, S. and Whelan, J. F. (1983) Isolation and culture of hamster pancreatic ducts. J. Tissue Cult. Methods 8, 97–103. 4. 4 Richards, J., Pasco, D., Yang, J., Guzman, R., and Nandi, S. (1983) Comparison of the growth of normal and neoplastic mouse mammary cells on plastic, on collagen gels and in collagen gels. Exp. Cell Res. 146, 1–14. 5. Hootman, S. R. and Logsdon, C. D. (1998) Isolation and monolayer culture of guinea pig pancreatic duct epithelial cells. In Vitro 24, 566–574. 6. 6 Lefebvre, V. H., Otonkoski, T., Ustinov, J., Huotari, M. A., Pipeleers, D. G., and Bouwens, L. (1998) Culture of adult human islet preparations with hepatocyte growth factor and 804G matrix is mitogenic for duct cells but not for beta cells. Diabetes 47, 134–137. 7. 7 Langhofer, M., Hopkinson, S. B., and Jones, J. C. R. (1993) The matrix secreted by 804G cells contains laminin-related components that participate in hemidesmosome assembly in vitro. J. Cell Sci. 105, 753–764. 8. 8 Parsa, I., Marsh, W. H., and Sutton, A. L. (1980) An in vitro model of pancreas carcinogenesis. Am. J. Pathol. 98, 649–662. 9. 9 Parsa, I., Marsh, W. H., Sutton, A. L., and Butt, K. M. H. (1981) Effects of dimethylnitrosamine on organ-cultured adult human pancreas. Am. J. Pathol. 102, 403–411. 10. 10 Resau, J. H., Cottrell, J. R., Elligett, K. A., and Hudson, E. A. (1987) Cell injury and regeneration of human epithelium in organ culture. Cell Biol. Toxicol. 3, 441–458. 11. 11 Harris, A. and Coleman, L. (1988) Cultured epithelial cells derived from foetal pancreas as a model for the study of cystic fibrosis: Further analysis on the origins and nature of the cell types. J. Cell Sci. 90, 73–77.
120
Lawson et al.
12. 12 Harris, A., Chalkley, G., Goodman, S., and Coleman, L. (1991) Expression of the cystic fibrosis gene in human development. Development 113, 305–310. 13. Githens, S., Patke, C. L., and Schexnayder, J. A. (1994) Isolation and culture of 13 rhesus monkey pancreatic ductules ductule-like epithelium. Pancreas 9, 20–31. 14. 14 Verme, T. B. and Hootman, S. R. (1990) Regulation of pancreatic duct epithelial growth in vitro. Am. J. Physiol. 258, G833–G840. 15. Vila, M. R., Lloreta, J., Schussler, M. H., Berrozpe, G., Welt, S., and Real, F. X. 15 (1995) New pancreas cancers cell lines that represent distinct stages of ductal differentiation. Lab. Invest. 72, 395–404. 16. Chambers, J. A. and Harris, A. (1993) Expression of the cystic fibrosis gene and 16 the major pancreatic mucin gene, MUC1, in human ductal epithelial cells. J. Cell Sci. 105, 417–422. 17. Kolar, C., Caffrey, T., Hollingsworth, M., et al. (1997) Duct epithelial cells cul17 tured from human pancreas processed for transplantation retain differentiated ductal characteristics. Pancreas 15, 265–271. 18. Kerr-Conte, J., Pattou, F., Lecomte-Houcke, M., et al. (1996) Ductal cyst forma18 tion in collagen-embedded adult human islet preparations. A means to the reproduction of nesidioblastosis in vitro. Diabetes 45, 1108–1114. 19. Trautmann, B., Schlitt, H.-J., Hahn, E. G., and Lohr, M. (1993) Isolation, culture, 19 and characterization of human pancreatic duct cells. Pancreas 8, 248–254. 20. 20 Ryu, B., Jones, J., Hollingsworth, M. A., Hruban, R. H., and Kern, S. E. (2001) Invasion-specific genes in malignancy: Serial analysis of gene expression comparisons of primary and passaged cancers. Cancer Res. 61, 1833–1838. 21. 21 Harris, A. and Coleman, L. (1987) Establishment of a tissue culture system for epithelial cells derived from human pancreas: A model for the study of cystic fibrosis. J. Cell Sci. 87, 695–703. 22. Singer, K. H., Scearce, R. M., Tuck, D. T., Whichard, L. P., Denning, S. M., and 22 Haynes, B. F. (1989) Removal of fibroblasts form human epithelial cell cultures with use of a complement fixing monoclonal antibody reactive with human fibroblasts and monocytes/macrophages. J. Invest. Dermatol. 92, 166–170. 23. Harris, C. C., Autrup, H., Stoner, G., et al. (1977) Metabolism of benzo(a)pyrene 23 and 7,12-dimethylbenz(a)anthracene in cultured human bronchus and pancreatic duct. Cancer Res. 37, 3349–3355. 24. Jones, R. T., Barrett, L. A., van Haaften, C., Harris, C. C., and Trump, B. F. (1977) 24 Carcinogenesis in the pancreas. I. Long-term explant culture of human and bovine pancreatic ducts. J. Natl. Cancer Inst. 58, 557–565. 25. Lawson, T. and Kolar, C. (1994) Mutagenicity of heterocyclic amines when acti25 vated by pancreas tissue. Mutation Res. 325, 125–128. 26. 26 Hyde, K., Harrison, D., Hollingsworth, M. A., and Harris, A. (1999) Chloride-bicarbonate exchangers in human fetal pancreas. Biochem. Biophys. Res. Commun. 263, 315–321. 27. 27 de Lisle, R. C. and Logsdon, C. D. (1990) Pancreatic acinar cells in culture: Expression of acinar and ductal antigens in a growth-related manner. Eur. J. Cell Biol. 51, 64–75.
Culture and Immortalization of DECs
121
28. 28 Hall, P. A. and Lemoine, N. R. (1993) Rapid acinar to ductal transdifferentiation in cultured human exocrine pancreas. J. Pathol. 166, 97–103. 29. Vilá, M. R., Lloreta, J., and Real, F. X. (1994) Normal human pancreas cultures display functional ductal characteristics. Lab. Invest. 71, 423–431. 30. Yuan, S., Rosenberg, L., Paraskevas, S., Agapitos, D., and Duguid, W. P. (1996) 30 Transdifferentiation of human islets to pancreatic ductal cells in collagen matrix culture. Differentiation 61, 67–75. 31. Paddenberg, R., Flocke, K., Elsasser, H. P., Lesch, G., Heidtmann, H. H., and Mannherz, H. G. (1998) Phenotypical changes of a human pancreatic adenocarcinoma cell line after selection on laminin-1/nidogen (LM/Ng) substratum. Eur. J. Cell Biol. 76, 51–64. 32. 32 Bonner-Weir, S., Taneja, M., Weir, G. C., et al. (2000) In vitro cultivation of human islets from expanded ductal tissue. Proc. Natl. Acad. Sci. USA 97, 7999– 8004. 33. 33 Gmyr, V., Kerr-Conte, J., Vanderwalle, B., Proye, C., Lefebvre, J., and Pattou, F. (2001) Human pancreatic ductal cells: Large scale isolation and expansion. Cell Transplant. 10, 109–121. 34. 34 Gmyr, V., Kerr-Conte, J., Belaich, S., et al. (2000) Adult human cytokeratin 19positive cells reexpress insulin promoter factor 1 in vitro: Further evidence for pluripotent pancreatic stem cells in humans. Diabetes 49, 1671–1680. 35. 35 Ramiya, V. K., Maraist, M., Arfors, K. E., Schatz, D. A., Peck, S. B., and Cornelius, J. G. (2000) Reversal of insulin-dependent diabetes using islets generated in vitro from pancreatic stem cells. Nat. Med. 6, 278–282. 36. Ouellette, M. M. and Lee, K. (2001) Telomerase: Diagnostics, cancer therapeu36 tics and tissue engineering. Drug Discov. Today 6, 1231–1237. 37. 37 Bodnar, A. G., Ouellette, M., Frolkis, M., et al. (1998) Extension of life-span by introduction of telomerase into normal human cells. Science 279, 349–352. 38. Morales, C. P., Holt, S. E., Ouellette, M., et al. (1999) Absence of cancer-associ38 ated changes in human fibroblasts immortalized with telomerase. Nat. Genet. 21, 115–118. 39. 39 Jiang, X. R., Jimenez, G., Chang, E., et al. (1999) Telomerase expression in human somatic cells does not induce changes associated with a transformed phenotype. Nat. Genet. 21, 111–114. 40. Shay, J. W. and Wright, W. E. (2001) Aging. When do telomeres matter? Science 40 291, 839–840. 41. 41 Furukawa, T., Duguid, W. P., Rosenberg, L., Viallet, J., Galloway, D. A., and Tsao, M. S. (1996) Long-term culture and immortalization of epithelial cells from normal adult human pancreatic ducts transfected by the E6E7 gene of human papilloma virus 16. Am. J. Pathol. 148, 1763–1770. 42. Ouyang, H., Mou, L. J., Luk, C., et al. (2000) Immortal human pancreatic duct 42 epithelial cell lines with near normal genotype and phenotype. Am. J. Pathol. 157, 1623–1631. 43. Jesnowski, R., Muller, P., Schareck, W., Liebe, S., and Lohr, M. (1999) Immor43 talized pancreatic duct cells in vitro and in vivo. Ann. NY Acad. Sci. 880, 50–65.
122
Lawson et al.
44. 44 Halbert, C. L., Demers, G. W., and Galloway, D. A. (1992) The E6 and E7 genes of human papillomavirus type 6 have weak immortalizing activity in human epithelial cells. J. Virol. 66, 2125–2134. 45. Ouellette, M. M., Aisner, D. L., Savre-Train, I., Wright, W. E., and Shay, J. W. 45 (1999) Telomerase activity does not always imply telomere maintenance. Biochem. Biophys. Res. Commun. 254, 795–803. 46. Counter, C. M., Hahn, W. C., Wei, W., et al. (1998) Dissociation among in vitro telomerase activity, telomere maintenance, and cellular immortalization. Proc. Natl. Acad. Sci. USA 95, 14723–14728.
DNA Methylation Analysis in Human Cancer
123
7 DNA Methylation Analysis in Human Cancer Carmelle D. Curtis and Michael Goggins
Summary Many tumor suppressor genes (such as p16, Rb, VHL, E-cadherin, and hMLH1) that are silenced by mutation are also inactivated by gene silencing through DNA methylation. Characterization of genes hypermethylated in human cancers but not in normal tissues not only provides insights into cancer biology but also permits the use of methylation-specific polymerase chain reaction–based assays that could serve as diagnostic tests for the early detection and early diagnosis of this disease. To this end, research aimed at the identification and characterization of the methylation status of known and candidate tumor suppressor genes is one strategy for finding putative diagnostic markers. This chapter describes several methods of methylation analysis. Key Words: Methylation analysis; polymerase chain reaction; CpG island; DNA methyltransferase; restriction enzyme digestion; bisulfite modification; immunohistochemistry; high performance liquid chromatography; methylation-specific oligonucleotide microarray; representation differential analysis; gene expression profiling; 5-Aza-2'-deoxycytidine; restriction landmark genome scanning.
1. Introduction In the human genome, DNA methylation is restricted to cytosines of CpG dinucleotides, which are often clustered into CpG-rich regions known as CpG islands. CpG (cytosine-phosphoguanine) islands are defined as regions of the genome that are at least 200 basepairs (bp) with a GC content of >60% and have a CpG doublet of >10/100 bp. CpG islands are usually found at the 5' end of genes near promoters (1). Outside of CpG islands evolutionary pressures have reduced the prevalence of CpGs as a defense against the deamination of methylated cytosines that leads to thymidine mutation. Methylation of the C5 position of cytosines’ DNA requires S-adenosylmethionine and is catalyzed by the
From: Methods in Molecular Medicine, Vol. 103: Pancreatic Cancer: Methods and Protocols Edited by: G. Su © Humana Press Inc., Totowa, NJ
123
124
Curtis and Goggins
action of DNA methyltransferases (DNMT). In eukaryotic cells DNA methylation is often associated with changes in the transcriptional activity of adjacent promoters. DNA methylation is also associated with genomic imprinting (2), X-chromosome inactivation (3), embryonic development (4,5), and protection against foreign DNA (6). Many tumor suppressor genes (such as p16, Rb, VHL, E-cadherin, and hMLH1) (7–11) that are silenced by mutation are also inactivated by gene silencing through DNA methylation. Characterization of genes hypermethylated in human cancers but not in normal tissues not only provides insights into cancer biology but also permits the use of methylation specific polymerase chain reaction (PCR)–based assays that could serve as diagnostic tests for the early detection and early diagnosis of this disease. To this end, research aimed at the identification and characterization of the methylation status of known and candidate tumor suppressor genes is one strategy for finding putative diagnostic markers. This chapter will describe several methods of methylation analysis including detection of methylated CpGs by: 1. 2. 3. 4. 5. 6. 7.
Restriction enzyme digestion. Bisulfite modification and methylation specific PCR (MSP). Immunohistochemistry and high-performance liquid chromatography (HPLC). Methylation-sensitive nucleotide primer extension (Ms-SnPE). Combined bisulfite restriction analysis (COBRA). Methylation-specific oligonucleotide (MSO) microarray. Methylated CpG island amplification/representation differential analysis (MCA/ RDA). 8. Gene expression profiling before and after 5-aza-2'-deoxycytidine (5-aza-dC). 9. Restriction landmark genome scanning (RLGS).
In addition, Table 1 lists these methods. 1.1. Restriction Enzyme Digestion The use of restriction enzymes is often the first step in determining the methylation status of DNA. In this method, genomic DNA is cut with methylationsensitive and -insensitive restriction enzymes such as HpaII and MspI, respectively. This is an enzyme set that recognizes CCGG sites. Both will cut at the internal cytosine if the DNA is unmethylated but only MspI will cut if this site is methylated. A comparison of the restriction patterns will show which cytosine residues are methylated. Other methylation-sensitive and -insensitive isoschizomers may be used. This method can be used with Southern analysis and PCR. 1.2. Bisulfite Modification and Methylation-Specific PCR (MSP) Bisulfite modification of DNA is the other common means besides restriction digestion of directly characterizing the methylation status of DNA. It is usually combined with other techniques such as PCR and/or sequencing. Bisulfite treat-
DNA Methylation Analysis in Human Cancer
125
Table 1 Methylation Analysis Methods Methylation analysis methods Type of detection
Source of isolated DNA
Restriction digestion
Cells
Bisulfite sequencing MSP IHC HPLC Ms-SNuPE COBRA MSO microarray MCA/RDA 5-Aza-dC treatment followed by RT-PCR RLGS
Global, specific (at the site of digestion) Specific Regional Global Global, regional Regional, specific Regional, specific (at the site of digestion) Regional Regional Regional
Frozen, paraffin, cells Frozen, paraffin, cells Cells
Global, regional
Frozen, paraffin, cells
Frozen, Frozen, Frozen, Cells Frozen, Cells
paraffin, cells paraffin paraffin, cells paraffin, cells
Various methylation analysis methods are shown. Three different types of detection are listed. Global refers to genome-wide methylation. Regional detection is within a certain gene or region of the genome and specific detection is at a specific base. Sources of isolated DNA include DNA isolated from cell lines, frozen, or paraffinized tissues. Abbreviations are the same as those in the text.
ment of genomic DNA changes unmethylated cytosines to uracils, resulting in thymidines incorporation during PCR. Primers are designed to target bisulfite modified DNA—that is, all cytosines are converted to thymidines unless the cytosines are part of a CpG dinucleotide. To sequence bisulfite modified DNA, primers are usually designed so that they avoid CpGs and thus PCR amplification leads to the generation of PCR products that correspond to bisulfite modified DNA that may have either methylated cytosines or converted thymidines at CpGs internal to the sequencing primer. Methylation-specific PCR is a methylation analysis technique first described by Herman et al. (9). In this method, bisulfite-treated genomic DNA is used as the template for PCR reactions. Primers specific for methylated or unmethylated DNA are designed to amplify regions of 80–250 bp. Methylase treatment followed by bisulfite treatment of DNA is used to generate a positive control. Isolated DNA from cell lines treated with the demethylating agent 5-aza-2'-deoxycytidine (5-aza-dC) can serve as a negative control. Low-level methylation of only a small percentage of templates such as agerelated methylation in normal tissues would usually not be detectable using bisulfite sequencing because it will be masked by the abundance of unmethylated
126
Curtis and Goggins
templates. Such low-level methylation can be detected using the more sensitive MSP. 1.3 Global Measurement of DNA Methylation by Immunohistochemical Staining (IHC) and High-Performance Liquid Chromatography (HPLC) Examining global methylation will give an idea of the overall level of 5-methylcytosine (5mC) within a cell or tissue. Most cancers show hypermethylation of tumor suppressor genes when compared to normal tissues, but globally cancers have a lower level of 5mC present (12,13). IHC with a monoclonal anti-5mC antibody (14), and HPLC are two methods for analyzing global methylation. A detailed description of HPLC is beyond the scope of this chapter. 1.4. Methylation-Sensitive Nucleotide Primer Extension (Ms-SNuPE) Methylation-sensitive nucleotide primer extension is another method that uses bisulfite modified DNA to access the methylation status of specific CpG dinucleotides. After bisulfite modification and PCR, samples are subjected to primer extension. One advantage of this technique is that small amounts of DNA such as DNA from microdissected sections can be used. 1.5. Combined Bisulfite Restriction Analysis (COBRA) COBRA is ideal when working with a small sample size such as DNA isolated from paraffin. As its name suggests, COBRA is a combination of several methods including bisulfite treatment and PCR, followed by restriction enzyme digestion. This technique is a sensitive and reliable method to quantitate the methylation status of specific CpGs of a gene. 1.6. Methylation-Specific Oligonucleotide (MSO) Microarray MSO microarray uses bisulfite-modified DNA as a template for PCR (15). DNA is hybridized to an array of thousands of oligonucleotides that are 19–20 nucleotides in length. These oligonucleotides can distinguish between methylated and unmethylated cytosine residues. This technique has been used to identify methylation differences between normal and cancer tissues. 1.7. Methylated CpG Island Amplification/ Representation Differential Analysis (MCA/RDA) MCA/RDA is a subtractive and kinetic enrichment technique that relies on the use of methylation-specific restriction enzymes to differentiate between methylated and unmethylated CpGs in different DNA samples. After restric-
DNA Methylation Analysis in Human Cancer
127
tion digestion, differentially methylated CpGs are identified through a series of steps. First, restriction-digested DNA is subject to PCR amplification after ligation of linkers to the digested DNA. The amplicons generated after PCR can then be identified using RDA. The RDA step involves two or three rounds of subtractive hybridization followed by PCR amplification of subtracted DNA products. This technique involves characterization of a large number of clones, therefore it is relatively labor intensive compared to other strategies for identifying novel changes in DNA methylation between tissues such as gene expression profiling before and after 5-aza-2'-deoxycitidin treatment. 1.8. Gene Expression Profiling Before and After 5-Aza-2'-Deoxycytidine (5-Aza-dC) Several groups have recently demonstrated that this is a highly effective strategy for identifying novel aberrantly methylated gene candidates in cancers (16–18). 1.9. Restriction Landmark Genome Scanning (RLGS) Restriction landmark genome scanning has successfully been used to profile a large number of methylation alterations in multiple tissues (19,20), but it has limitations because it cannot discriminate between methylation and deletion events, the latter of which is common in carcinoma. 2. Materials 2.1. Restriction Enzyme Digestion 1. 2. 3. 4. 5. 6.
20 µg of DNA. HpaII (New England Biolabs). MspI (New England Biolabs). Loading buffer. 10X enzyme buffer (provided with enzyme). ddH2O.
2.2. Bisulfite Modification (21) 1. 2. 3. 4. 5. 6. 7. 8. 9.
2 M NaOH. 10 mM hydroquinone (Sigma). 3 M sodium bisulfite (Sigma). 80% isopropanol. ddH2O. Wizard Purification DNA resin (Promega). Genomic DNA from cells or tissues. Mineral oil. Low-concentration Tris-EDTA: 10 mM Tris-HCl, pH 8.0, and 1 mM EDTA.
128
Curtis and Goggins
2.3. Methylation-Specific PCR (MSP) 1. 10X PCR buffer: 166 mM ammonium sulfate, 670 mM Tris-HCl, pH 8.8, 67 mM MgCl2, and 100 mM 2-mercaptoethanol. 2. Bisulfite-treated sample DNA. 3. 1.25 mM of each dNTP/50 µL reaction. 4. 300 ng/50 µL reaction each sense and antisense primer. 5. 5-Aza-2'-deoxycytidine. 6. Cell culture.
2.4. Immunohistochemistry (IHC) (14) 1. 2. 3. 4. 5. 6. 7. 8.
Formalin-fix paraffin-embedded section. Monoclonal anti-5mC antibody (22). Wash buffer. 10 mM citric acid, pH 6. 3.5 N HCl. 3.0% H2O2. 1% preimmune goat serum. Biotin–streptavidin detection system (Signet).
2.5. High-Performance Liquid Chromatography (HPLC) 1. 2. 3. 4. 5. 6.
DNA sample. 3 mM Tris-HCl. 0.2 mM EDTA. LC-18 reverse-phase column (Supelcosil, Inc.). Nuclear P1 (Boehringer-Mannheim). Bacterial alkaline phosphatase (Sigma).
2.6. Methylation-Sensitive Nucleotide Primer Extension (Ms-SNuPE) 1. 2. 3. 4. 5. 6. 7. 8. 9. 10.
Bisulfite-modified DNA. Buffer: 10 mM Tris-HCl, pH 8.3, 1.5 mM MgCl2, 50 mM KCl, 0.1% gelatin/mL. 100 µM of each dNTP/ 25 µL reaction. 0.5 µM final concentration of each sense and antisense primer. 2% agarose gel. Qiaquick gel extraction kit (Qiagen). 10X PCR buffer. Taq DNA polymerase (Boehringer Mannheim). TaqStart antibody (Clontech). [32P]dCTP or [32P]dTTP.
2.7. Combined Bisulfite Restriction Analysis (COBRA) (23) 1. Bisulfite-modified DNA. 2. 10 K MWCO nanospin plus filters (Gelman Sciences). 3. HpaII (New England Biolabs).
DNA Methylation Analysis in Human Cancer
129
2.8. Methylation-Specific Oligonucleotide (MSO) Microarray (15) 1. 2. 3. 4. 5.
DNA isolated for cell lines or tissue. Bisulfite-modified DNA samples. 1X microspotting solution (Telechem). Terminal transferase (New England Biolabs). Superaldehyde-coated glass slides (Telechem).
2.9. Methylated CpG Island Amplification/ Representation Differential Analysis (MCA/RDA) 1. 2. 3. 4. 5.
5 µg of DNA. SmaI (New England Biolabs). XmaI (New England Biolabs). RMCA adapter (27). RDA buffer: 10 mM 670 Tris-HCl, pH 8.8, 1.5 mM MgCl2, 50 mM KCl, 0.5 M betaine, 2% DMSO. 6. 200 µM of each dNTP/100-µL reaction. 7. 100 pmol of RMCA. 8. Taq polymerase (Life Technologies, Inc.).
2.10. Gene Expression Profiling Before and after 5-Aza-2'-Deoxycytidine 1. Cell lines. 2. Cell media. 3. 5-Aza-2'-Deoxycytidine (Sigma).
2.11. Restriction Landmark Genome Scanning (RLGS) The materials are listed in detail by Costello et al. (19,20). 3. Methods 3.1. Restriction Enzyme Digestion 1. Digest 10 µg of DNA with an enzyme that produces the fragment of interest, and with HpaII. 2. In a separate reaction, digest 10 µg of DNA with the same enzyme of choice and with MspI. Make sure the digest is carried out in a buffer that is compatible for each enzyme. 3. Use a total volume of 50 µL and overlay with oil to prevent evaporation. 4. Incubate overnight at the optimal temperature. 5. Heat the reactions in order to inactivate the enzymes (refer to the manufacturer’s protocols). 6. Concentrate the samples to a volume of 20–30 µL by using a speed vacuum. 7. Add loading buffer and electrophorese samples. 8. Southern blot and hybridize samples using standard procedures (see Note 1).
130
Curtis and Goggins
3.2. Bisulfite Modification 1. Prepare 3 M sodium bisulfite by adding 1.88 g of sodium to a total volume of 5 mL with dH2O (makes enough for nine samples). 2. Adjust to pH 5.0 with NaOH and set reagents on ice. 3. Dilute 1 µg of DNA into a total volume of 50 µL with dH2O and add 5.5 µL of 2 M NaOH. 4. Incubate 10 min in a 37°C water bath to denature dsDNA. 5. Add 30 µL of 10 mM hydroquinone followed by 520 µL of 3 M sodium bisulfite. 6. Mix solution well then layer with mineral oil. 7. Incubate at 50°C overnight for 16 h in a heating block. 8. Cover samples with foil to keep out light. 9. Purify the DNA by following the Wizard Purification kit protocol. 10. Precipitate DNA by adding 1 µL of glycogen with 10 M ammonium acetate and ethanol.
3.3. Methylation-specific PCR 1. Prepare PCR mixes for the methylated and unmethylated reactions. Each reaction contains 1X PCR buffer, 50 ng of bisulfite-modified DNA, 1.25 mM of each dNTP, and 300 ng of each primer. 2. Add water to a final volume of 50 µL. 3. Hot-start at 95°C for 5 min before adding 1.25 U of Taq polymerase (BRL). 4. Begin thermocycling for 35 cycles of 30 s at 95°C, 30 s at the annealing temperature, and 30 s at 72°C, followed by 4 min at 72°C. 5. Load 10 µL of each sample onto a nondenaturing 6–8% polyacrylamide gel. 6. Stain with ethidium bromide and visualize with UV light (see Note 2).
3.4. Immunohistochemistry 1. Use formalin-fixed, paraffin-embedded sections of approximately 4 µm in thickness. (Use a cancer-free section for the control.) 2. Retrieve antigen by placing slides in 10 mM citric acid, pH 6.0. Microwave at full power for 10 min. 3. Place slides in 3.5 N HCl for 15 min at room temperature to expose CpGs. 4. Place slides in 3.0% H2O2 for 4 min to inhibit endogenous peroxidase activity. 5. Block slides by incubating with 1% preimmune goat serum for 20 min at room temperature. 6. Rinse with wash buffer and add the primary anti-5mC antibody at a concentration of 5 µg/mL for 1 h at room temperature. 7. Use a serially prepared slide without anti-5mC as a control (see Note 3).
3.5. High-Performance Liquid Chromatography 1. Resuspend DNA at a concentration of 0.5 µg/mL into 3 mM Tris-HCl, and 0.2 mM EDTA, pH 7. 2. Denature by placing sample into a boiling water bath for 2 min.
DNA Methylation Analysis in Human Cancer
131
3. Digest with nuclease P1 and bacteria alkaline phosphatase (24). 4. Load 10 µg of DNA onto a Supelcosil LC-18 DB reverse-phase column.
3.6. Methylation-Sensitive Nucleotide Primer Extension 1. Add approximately 50 ng bisulfite-modified DNA to a total volume of 25 µL containing 10 mM Tris-HCl, pH 8.3, 1.5 mM MgCl2, 50 mM KCl, 0.1% gelatin/mL, 100 µM of each dNTP, and 0.5 µM of each primer. 1. Hot-start with 1:1 Taq/Taqstart antibody (Clontech). 3. Thermocycle 94°C for 3 min followed by 35 cycles of 94°C for 30 s, 68°C for 30 s, and 72°C for 30 s. 4. Electrophorese products on a 2% agarose gel. 5. Isolate bands with Qiaquick gel extraction kit (Qiagen). 6. For Ms-SNuPE, incubate approx 10–50 ng of PCR product as template in 1X PCR buffer, 1 µM forward primer, 1 µM reverse primer, 1 µCi of [32P]dCTP ([32p]dTTP may also be used), and 1 U of Taq polymerase. 7. Hot-start and add Taq/Taqstart antibody. 8. For primer extension, incubate at 95°C for 1 min, 50°C for 2 min, and 72°C for 1 min.
3.7. Combined Bisulfite Restriction Analysis 1. Design PCR primers that complement the original DNA sequence and do not contain CpG sites. 2. Purify PCR with 10 K MWCO nanospin plus filters (Gelman Sciences). 3. Rinse twice with 200 µL of dH2O. 4. Digest the purified DNA with a restriction enzyme containing a site with a CpG in the unmodified DNA to check efficiency of bisulfite modification. (Also perform a control digest. Xiong et al. recommend HspII, which will recognize unmethylated DNA [23].) 5. Electrophorese on an 8% denaturing polyacrylamide gel. 6. Southern blot and hybridize samples according to standard protocols (see Note 4).
3.8. Methylation-Specific Oligonucleotide Microarray 1. Design several sets of oligonucleotide pairs. Each should have two to four CpG sites of interest. 2. Attach amino-linked C6 linker to the 5' end (Integrated DNA Technologies). 3. Dilute linker attached oligonucleotides to 50 pmol/µL in 1X microspotting solution (Telechem). 4. Print 0.05–0.1 pmol of each oligonucleotide in quadruplicate onto superaldehydecoated glass slides (Telechem) with a GMS 417 microarrayer (Affymetrix). 5. Wash slides to remove any unbound oligonucleotides (Telechem). 6. Bisulfite treat DNA and PCR amplify as described above. 7. Use terminal transferase (New England Biolabs) to 3' label PCR products with Cy5dCTP (Amersham Pharmacia). 8. Remove unincorporated dCTP with the micro-Biospin column (Bio-Rad).
132
Curtis and Goggins
9. Resuspend labeled product into Unihybridization solution (1:4 v/v Telechem). 10. Denature at 95°C for 5 min and apply to glass slide. 11. Cover samples with a cover slip and hybridize samples at 45°C for 4 h in a moist chamber. 12. Rinse slide and wash twice at room temperature with 2X SSC–0.2% SDS for a total of 15 min. 13. Wash twice with 2X SSC at room temperature for 5 min. 14. Centrifuge sample for 5 min at 58g to dry slides. 15. For data analysis, scan the microarray slide with a GenePix 4000A scanner (Axon Instruments) at 600 PMT (see Note 5).
3.9. Methylated CpG Island Amplification/ Representation Differential Analysis The RDA step is performed in similar fashion to the original method described by Lisitsyn and Wigler (25) and performed by Schutte et al. (26). The full RDA/ MCA protocol is available at www.mdanderson.org/leukemia/methylation. It is described here briefly. 1. Digested 5 µg of DNA with SmaI and XmaI (New England Biolabs). Other methylation-sensitive restriction enzymes can be used but this combination is helpful for identifying CpG islands. 2. Ligate the restriction fragments to RMCA adapter and amplify by PCR in RDA, 200 µM each dNTP, 100 pmol of RMCA 24mer primer, and 15 U Taq polymerase (Gibco-BRL) in a final reaction volume of 100 µL. 3. Incubated the reaction mixture at 72°C for 5 min and at 95°C for 3 min. 4. Thermocycle 25 cycles of 1 min at 95°C and 3 min at 77°C followed by a final extension of 10 min at 77°C. If one is trying to identify aberrantly methylated DNA from cancer DNA compared to non-neoplastic DNA, then the MCA amplicons generated during the MCA step can be used as the tester for RDA and MCA amplicon generated from a mixture of DNA from the normal tissues can be used as the driver. 5. Perform RDA on these MCA amplicons using different adapters, JMCA and NMCA. Sequences of adapters used for MCA/RDA are available from the authors on request. 6. After the third round competitive hybridization and selective amplification, the RDA difference products of second and third round amplifications are cloned into pBluescript II plasmid vector (Stratagene). 7. Characterize the sequence of clones recovered after MCA/RDA DNA from each clone by amplifying with T3 and T7 primers and then sequenced using KS primer. 8. To determine the methylation status of MCA/RDA clones in cancer and normal tissues, clones can be screened by hybridizing them to a dot blot of MCA products of DNA samples under analysis. 9. Digest plasmid DNA containing each independent clone with SmaI. 10. Recover DNA fragments from an agarose gel and used as a probe for dot blot hybridization.
DNA Methylation Analysis in Human Cancer
133
11. Blot 1-µL aliquots of the mixture of 10X SSC and MCA products from the driver and from the tester both before and after each of the three rounds of RDA amplification/hybridization onto nylon membranes in duplicate. 12. Hybridize the membranes overnight with 32P-labeled probes. 13. Wash the membranes and expose to Kodak X-ray film (see Note 6).
3.10. Gene Expression Profiling Before and after 5-Aza-2'-Deoxycytidine 1. Treat cultured cells with 1 µM 5-aza-dC continuously for 4 d. Each day add new media and fresh 5-aza-dC. Use PBS instead of 5-aza-dC for mock-treated controls. One common gene expression profiling method employed is the Affymetrix platform. 2. Isolate total RNA from cultured cells using TRIZOL reagent (Invitrogen, Carlsbad, CA). 3. Purify with the RNeasy Mini Kit (QIAGEN, Valencia, CA). 4. Synthesize first- and second-stranded cDNA from 10 µg of total RNA using T7(dT)24 primer (Genset Corp., South La Jolla, CA) and the SuperScript Choice system (Invitrogen). 5. Labeled cRNA is synthesized from the purified cDNA by in vitro transcription (IVT) reaction using the BioArray HighYield RNA Transcript Labeling Kit (Enzo Diagnostics, Inc., Farmingdale, NY) at 37°C for 6 h. 6. The cRNA is fragmented at 94°C for 35 min in a fragmentation buffer (40 mmol/ L of Tris-acetate, pH 8.1, 100 mmol/L of potassium acetate, and 30 mmol/L magnesium acetate). 7. Hybridize the fragmented cRNA to the Human Genome U133A chips (Affymetrix, Santa Clara, CA) with 18,462 unique gene/EST transcripts at 45°C for 16 h. 8. Wash and stain according to the manufacturer’s instructions for the Affymetrix Fluidics Station. 9. Scan the probes using the Affymetrix GeneChip scanner. Calculate signal intensity for each transcript (background-subtracted and adjusted for noise) by using Microarray Suite Software 5.0 (Affymetrix) (see Note 7).
3.11. Restriction Landmark Genome Scanning Restriction landmark genome scanning (RLGS) involvess six steps. These steps include isolation, enzymatic processing, first dimension electrophoresis, in-gel digestion, second-dimension electrophoresis, and analysis of DNA. These procedures are described in detail by Costello et al. (19,20). 4. Notes 1. Bands within the MspI lane correspond to completely digested DNA that does not contain methylated sites. Bands within the HpaII lane can be compared to those in the MspI lane to reveal regions of methylation. Restriction digestion results may not be easily interpreted due to many different bands that correspond to methylation
134
2.
3. 4.
5.
6.
7.
Curtis and Goggins
of DNA at several sites. Restriction digestion and Southern hybridization are not adequate methods for analysis of paraffin extracted DNA. Also prepare positive, negative, and water controls. For demethylated control, add 1 µM 5-aza-dC to media containing cultured cells. Treat cells daily by changing media and adding fresh 5-aza-dC. Harvest cells and isolate DNA. Bisulfite sequencing is a particularly good method for studying individual CpGs. For best results place hydroquinone and bisulfite on ice and make sure they are dissolved. Another consideration is the care of DNA after bisulfite treatment. Grunau et al. showed that during the process of bisulfite treatment, DNA is substantially degradation. In a study using HPLC or quantitative PCR to measure intact DNA, Grunau et al. found that between 84% and 96% of DNA was degraded during the process of bisulfite treatment (28). They also state that resuspending treated DNA in Tris-EDTA instead of water helps to prevent rapid degradation (28). Store DNA at −20°C, but for best results DNA may be stored at −70°C. When using limited amounts or low quality DNA, a protocol with a PCR amplification step is helpful. For visualization, Piyathilake et al. (14) suggest the biotin–streptavidin detection system (Signet). Lightly counterstain with hemotoxylin. Comparing bands of DNA from cut versus uncut lanes will reveal the presence or absence of methylation. Make sure bisulfite treatment is complete by comparing samples to a control digest. Images may be further analyzed with Microsoft Excel. Statistical analysis may be achieved with SigmaStat 2.0 software (Jandel Scientific). Gitan et al. suggest scanning at approx 600 PMT to minimize background and to maintain linearity. Clones should represent sequences containing methylated SmaI sites at each end and hybridization is expected to occur in MCA amplicons generated from DNA samples that also have methylation at the SmaI restriction sites in the clone sequence. As the SmaI restriction site tends to occur in CpG islands, methylated clones identified using MCA/RDA usually arise in CpG islands. In addition to analysis of clones by blot hybridization, the cloned sequences can be tested in cancer and noncancerous DNA samples to determine methylation patterns, using bisulfite sequencing and MSP. The preceding MCA/RDA protocol includes a few modifications to improve the efficiency of the MCA/RDA technique. We included betaine in the PCR reaction and amplified the methylated templates under a higher annealing temperature (77°C). This combination can uniformly amplify a mixture of DNA with different GC content (28). This modification might have enhanced the amplification of distinct clones instead of Alu repetitive sequences that accounted for 60% of the recovered clones using the original protocol (29). A fivefold increase in a gene’s level of expression often indicates that the gene in the untreated cells is suppressed by methylation. Expression changes can be confirmed by reverse transcriptase-PCR.
References 1. Bird, A. (1992) The essentials of DNA methylation. Cell 70, 5–8. 2. Bird, A. (1999) DNA methylation de novo. Science 286, 2287–2278.
DNA Methylation Analysis in Human Cancer
135
3. Mohandas, T., Sparkes, R. S., and Shapiro, L. J. (1981) Reactivation of an inactive human X chromosome: Evidence for X inactivation by DNA methylation. Science 211, 393–396. 4. Okano, M., et al. (1999) DNA methyltransferases Dnmt3a and Dnmt3b are essential for de novo methylation and mammalian development. Cell 99, 247–257. 5. Okano, M., Xie, S., and Li, E. (1998) Cloning and characterization of a family of novel mammalian DNA (cytosine-5) methyltransferases. Nat. Genet. 19, 219–220. 6. Yoder, J. A., Walsh, C. P., and Bestor, T. H. (1997) Cytosine methylation and the ecology of intragenomic parasites. Trends Genet. 13, 335–340. 7. Baylin, S. B., et al. (1998) Alterations in DNA methylation: A fundamental aspect of neoplasia. Adv. Cancer Res. 72, 141–196. 8. Herman, J. G., et al. (1996) Methylation-specific PCR: A novel PCR assay for methylation status of CpG islands. Proc. Natl. Acad. Sci. USA 93, 9821–9826. 9. Herman, J. G., et al. (1998) Incidence and functional consequences of hMLH1 promoter hypermethylation in colorectal carcinoma. Proc. Natl. Acad. Sci. USA 95, 6870–6875. 10. Rountree, M. R., et al. (2001) DNA methylation, chromatin inheritance, and cancer. Oncogene 20, 3156–3165. 11. Ueki, T., et al. (2000) Hypermethylation of multiple genes in pancreatic adenocarcinoma. Cancer Res. 60, 1835–1839. 12. Gama-Sosa, M. A., et al. (1983) The 5-methylcytosine content of DNA from human tumors. Nucl. Acids Res. 11, 6883–6894. 13. Feinberg, A. P., et al. (1988) Reduced genomic 5-methylcytosine content in human colonic neoplasia. Cancer Res. 48, 1159–1161. 14. Piyathilake, C. J., et al. (2001) Altered global methylation of DNA: An epigenetic difference in susceptibility for lung cancer is associated with its progression. Hum. Pathol. 32, 856–862. 15. Gitan, R. S., et al. (2002) Methylation-specific oligonucleotide microarray: A new potential for high-throughput methylation analysis. Genome Res. 12, 158–164. 16. Suzuki, H., et al. (2002) A genomic screen for genes upregulated by demethylation and histone deacetylase inhibition in human colorectal cancer. Nat. Genet. 31, 141–149. 17. Yamashita, K., et al. (2002) Pharmacologic unmasking of epigenetically silenced tumor suppressor genes in esophageal squamous cell carcinoma. Cancer Cell. 2, 485–495. 18. Sato, N., et al. (2003) Discovery of novel targets for aberrant methylation in pancreatic carcinoma using high-throughput microarrays. Cancer Res. 63, 3735–2742. 19. Costello, J. F., Plass, C., and Cavenee, W. K. (2002) Restriction landmark genome scanning. Methods Mol. Biol. 200, 53–70. 20. Costello, J. F., Smiraglia, D. J., and Plass, C. (2002) Restriction landmark genome scanning. Methods 27, 144–149. 21. Frommer, M., et al. (1992) A genomic sequencing protocol that yields a positive display of 5-methylcytosine residues in individual DNA strands. Proc. Natl. Acad. Sci. USA 89, 1827–1831.
136
Curtis and Goggins
22. Reynaud, C., et al. (1992) Monitoring of urinary excretion of modified nucleosides in cancer patients using a set of six monoclonal antibodies. Cancer Lett. 61, 255–262. 23. Xiong, Z. and Laird, P. W. (1997) COBRA: A sensitive and quantitative DNA methy-lation assay. Nucl. Acids Res. 25, 2532–2534. 24. Gehrke, C. W., et al. (1984) Quantitative reversed-phase high-performance liquid chromatography of major and modified nucleosides in DNA. J. Chromatogr. 301, 199–219. 25. Lisitsyn, N. and Wigler, M. (1993) Cloning the differences between two complex genomes. Science 259, 946–951. 26. Schutte, M., et al. (1995) Identification by representational difference analysis of a homozygous deletion in pancreatic carcinoma that lies within the BRCA2 region. Proc. Natl. Acad. Sci. USA 92, 5950–5954. 27. Grunau, C., Clark, S. J., and Rosenthal, A. (2001) Bisulfite genomic sequencing: systematic investigation of critical experimental parameters. Nucl. Acids Res. 29, E65–E65. 28. Baskaran, N., et al. (1996) Uniform amplification of a mixture of deoxyribonucleic acids with varying GC content. Genome Res. 6, 633–638. 29. Toyota, M., et al. (1999) CpG island methylator phenotype in colorectal cancer. Proc. Natl. Acad. Sci. USA 96, 8681–8686.
Allelic Imbalance Detection by Digital SNP
137
8 Digital Single-Nucleotide Polymorphism Analysis for Allelic Imbalance Hsueh-Wei Chang and Ie-Ming Shih
Summary Digital single-nucleotide polymorphism (SNP) analysis is developed to amplify a single template from a pool of DNA samples, thereby generating the amplicons that are homogeneous in sequence. Different fluorophores are then applied as probes to detect and discriminate different alleles (paternal vs maternal alleles or wild-type vs mutant), which can be readily counted. In this way, digital SNP analysis transforms the exponential and analog signals from conventional polymerase chain reaction (PCR) to linear and digital ones. Digital SNP analysis has the following advantages. First, statistical analysis of the PCR products becomes available as the alleles can be directly counted. Second, this technology is designed to generate PCR products of the same size; therefore, DNA degradation would not be a problem as it commonly occurs when microsatellite markers are used to assess allelic status in clinical samples. Last, digital SNP analysis is designed to amplify a relatively small amount of DNA samples, which is available in some clinical samples. Digital SNP analysis has been applied in quantification of mutant alleles and detection of allelic imbalance in clinical specimens and it represents another example of the power of PCR and provides unprecedented opportunities for molecular genetic analysis. Key Words: Digital; molecular genetics; mutation; allelic imbalance; polymerase chain reaction; single-nucleotide polymorphism.
1. Introduction Genetic instability is a defining molecular signature of most human cancers (1,2), and at the molecular level it is characterized by allelic imbalance (AI), representing losses or gains of defined chromosomal regions. Analysis of AI is useful in elucidating the molecular basis of cancer and also provides a molecular basis for cancer detection. There are, however, at least two major problems associated with the current methods for assessing AI using microsatellite markers.
From: Methods in Molecular Medicine, Vol. 103: Pancreatic Cancer: Methods and Protocols Edited by: G. Su © Humana Press Inc., Totowa, NJ
137
138
H-W. Chang and I-M. Shih
First, DNA purified from microdissected tissues or body fluids is a mixture of neoplastic and non-neoplastic DNA and the latter, released from non-neoplastic cells, can mask AI because it is difficult to quantify the allelic ratio using microsatellite markers. Second, such DNA is often degraded to a variable extent, producing artifactual enrichment of smaller alleles when microsatellite markers are used for analysis (3). To overcome these obstacles associated with the molecular genetic analysis of AI, we employed a recently developed polymerase chain reaction (PCR)-based approach called digital single-nucleotide polymorphism (SNP) analysis in which the paternal or maternal alleles within a plasma DNA sample are individually counted, thus allowing a quantitative measure of such imbalance in the presence of normal DNA. Digital SNP analysis is based on the concept of digital PCR (4). Therefore, digital SNP analysis is a powerful tool to assess allele status of tumor cells when the presence of contamination from normal DNA is inevitable or only a minimal amount of DNA is available for assay. 2. Materials 1. Molecular beacons (MB) and oligonucleotide primers (see Subheading 3.1.). 2. Regular PCR reagents (10X PCR buffer provided by the company [DMSO], 50 mM MgCl2, and 25 mM dNTP) and Taq enzyme. 3. TE Buffer: 10 mM Tris and 1 mM EDTA. 4. Biomek2000 Working Station (Beckman) or equivalent. 5. 96- and 384-well plates. 6. Salad spinner or centrifuge (spin the 384-well plate). 7. PCR machine with block for 384 wells. 8. Fluorometer with appropriate adaptor to read 384-well plates.
3. Methods 3.1. Molecular Beacon and Primer Design 1. Several SNP websites are available to identify appropriate SNPs, such as http:// www.genome.wi.mit.edu/SNP/human/index.html or http://www.ncbi.nlm.nih. gov/SNP. 2. If the NCI website is used, http://lpg.nci.nih.gov/html-snp/imagemaps.html ® select one’s chromosomal regions by directly clicking the SNP map under the chromosome ® click the left column (distribution of SNP) of the SNPs one prefers based on gene name, adjacent microsatellite marker, genetic, physical maps or banding segment ® SNP viewer ® self-explanatory and you will see the flanking sequence data around the SNPs. 3. Choose the SNPs with allelic frequency close to each other [e.g., 12 A and 12 G from 24 different expressed sequence tags (ESTs)]. Go to Tool ® Design Primer ® copy the sequence to new file.
Allelic Imbalance Detection by Digital SNP
139
4. Design the primers that amplify approx 100 bp products containing the SNPs using the same guidelines as usual (approx 18–20 bp length, Tm: approx 58–62°C, end with 5'-CG-3' if possible). Design the molecular beacon with the following features: length: 18–21 bp with the SNP around the middle of the sequence, then add the stem loop structure as underlined (5'-CACG-nnnnnnnnnnnnnn-CGTG) so the total length of the molecular beacon is 26–29 bp. Check the Tm of the molecular beacon around 51.5 ± 2°C (without the stem sequence) or 72 ± 3.4°C (the whole molecular beacon). Try to avoid secondary structure within the beacon other than the stem loop on both ends (only 4 bp). Then design the molecular beacon for the other allele with the same sequence except for the SNP. A website is available to assist the design (see Note 1). 5. Repeat the above procedures to find more SNPs. 6. Order molecular beacons. Beacons should be 5' labeled with either fluorescein or HEX (or other fluorophores) and the 3' should be labeled with Dabcyl (the quencher). A 200 µM scale should be sufficient for at least 60 assays. All beacons should be gel purified.
3.2. Molecular Beacon Testing 1. Use the panel of control DNA from several healthy individuals. 2. For each molecular beacon set, sequence the PCR products amplified from control DNA (usually four or five samples are enough to obtain homozygous alleles). 3. Test the molecular beacon on seven control DNA samples without sequencing them.
3.3. Set Up the Reactions in a 96-Well Plate 1. Make up the PCR premix for one 384-well plate (1253 µL of water, 145 µL of 10X PCR buffer, 88 µL of DMSO, 45 µL of 50 mM MgCl2, 12 µL of 25 mM dNTP) without Taq, primers, beacons, and template. The detailed protocol has been described (5–9). They are good at –20°C for at least 3 mo. 2. Before the experiment, for one 384-well plate one need to add 14 µL of Taq, 3 µL of primers-F (1 µg/µL), 12 µL of primers-R (1 µg/µL), and 20 µL of MB mixture (10 µM) into the PCR premix and aliquot to a 96-well plate. 3. Add DNA templates (see Note 2), two allele-specific, homozygous control DNA samples and negative control (TE buffer).
3.4. Transfer from a 96-Well Plate to a 384-Well Plate 1. Apply 5 µL of mineral oil in each well of 384-well plates using automatic pipetting system. Open the BioWorks software for Biomek2000 Working Station ® set up the plate configuration to make 5 µL per well ® and run the program. 2. Apply the PCR mixture from 16 wells or 32 wells in a 96-well plate to a 384-well plate by Biomek2000 ® set program to make 3 µL per well for the purpose of DNA dilution with 0.5 genomic equivalent per well in 384-well plate ® make sure the pipet tip used is correct (20 µL) ® run ® accept all.
140
H-W. Chang and I-M. Shih
3.5. PCR Protocol 1. Place the adhesive plastic plate cover on the plates. Rub the cover evenly and gently with the plastic device. This rubbing is critical to keep samples at edge from evaporation. Salad spinning is highly recommended before PCR. 2. PCR was performed in a single step with the following protocol: 94oC (1 min); four cycles of 94°C (15 s), 64°C (15 s), 70°C (15 s); four cycles of 94°C (15 s), 61°C (15 s), 70°C (15 s); four cycles of 94°C (15 s), 58°C (15 s), 70°C (15 s); 60 cycles of 94°C for (15 s), 55°C (15 s), 70°C (15 s); 94°C (1 min), and 60°C (5 min). Use hot lid, simulated plate, and volume set to 8 µL.
3.6. Reading Using Cytofluor Galaxy (BMG) 1. Open the BMG cytofluor in the program ® control ® plate out ® insert the plate ® plate in ® measure. If the reading is near 65,000 in raw data, decrease the gain of the specific fluorophore and repeat the above until the optimal intensity range is achieved (the highest reading approx 45,000–55,000) (see Note 3).
3.7. Analysis of Digital SNP 1. In the analysis format, go to Summary in Excel sheet ® adjust the ratio of one control allele to be 1 and the other control above 1. 2. To obtain digitalized results, the number of positive wells (green, red, or yellow) should not exceed 250 (optimal: 150–220 positive wells per plate). 3. SPRT analysis. To determine whether there is statistical significance for AI, we employed the sequential probability ratio test (SPRT) (10). This method allows two probabilistic hypotheses to be compared as data accumulate, and facilitates decisions about the presence or absence of allelic imbalance after study of a minimum number of samples. The details and application of the SPRT in allelic counting have been previously reported (6–12). If the ratio is above lower curve, then interpretation is allelic imbalance; if below higher curve, then interpretation is allelic balance; if between higher and lower curve, then interpretation is not informative and more wells are required to conclude anything. Alternatively, a receiver-operating characteristic (ROC) curve can be constructed to determine the sensitivity and specificity using a series of allelic ratio cutoffs (12).
4. Notes 1. Alternatively, a software is available for design of molecular beacon and primer. Please refer to “Beacon Designer 2” for Molecular beacon, TaqMan® probe and primer pair design software for Windows. http://www.premierbiosoft.com/ 2. The DNA samples are purified form Qiagen PCR kit and the final volume in elution buffer is approx 150 µL. To determine the amount of DNA for digital SNP analysis, the DNA concentration was measured using the PicoGreen® dsDNA quantitation kit (Molecular Probes, Inc.) following the manufacturer’s instruction. The fluorescence intensity was measured by a FLUOstar Galaxy fluorescence microplate reader with an excitation at 480 nm and an emission at 520 nm.
Allelic Imbalance Detection by Digital SNP
141
3. One can adjust the gain by selecting one positive control well → select one fluorophore → gain adjustment (well) → repeat for the other fluorophore. Alternatively, one can select the whole plate for gain adjustment by selecting all the wells. The instrument will then determine the highest reading of any well and adjust the gain so that this reading is no higher than 90% of 65,000.
References 1. 1 Lengauer, C., Kinzler, K. W., and Vogelstein, B. (1998) Genetic instabilities in human cancers. Nature 396, 643–649. 2. 2 Cahill, D. P., Kinzler, K. W., Vogelstein, B., and Lengauer, C. (1999) Genetic instability and Darwinian selection in tumours. Trends Cell Biol. 9, M57–M60. 3. 3 Liu, J., Zabarovska, V. I., Braga, E., Alimov, A., Klien, G., and Zabarovsky, E. R. (1999) Loss of heterozygosity in tumor cells requires re-evaluation: The data are biased by the size-dependent differential sensitivity of allele detection. FEBS Lett. 462, 121–128. 4. 4 Vogelstein, B. and Kinzler, K. W. (1999) Digital PCR. Proc. Natl. Acad. Sci. USA 96, 9236–9241. 5. 5 Shih, I. M., Zhou, W., Goodman, S. N., Lengauer, C., Kinzler, K. W., and Vogelstein, B. (2001) Evidence that genetic instability occurs at an early stage of colorectal tumorigenesis. Cancer Res. 61, 818–822. 6. 6 Zhou, W., Galizia, G., Goodman, S. N., et al. (2001) Counting alleles reveals a connection between chromosome 18q loss and vascular invasion. Nat. Biotechnol. 19, 78–81. 7. 7 Shih, I. M., Wang, T. L., Traverso, G., et al. (2001) Top-down morphogenesis of colorectal tumors. Proc. Natl. Acad. Sci. USA 98, 2640–2645. 8. 8 Singer, G., Kurman, R. J., Chang, H.-W., Cho, S. K. R., and Shih, I.-M. (2002) Diverse tumorigenic pathways in ovarian serous carcinoma. Am. J. Pathol. 160, 1223–1228. 9. 9 Shih, I. M., Yan, H., Speyrer, D., Shmookler, B. M., Sugarbaker, P. H., and Ronnett, B. M. (2001) Molecular genetic analysis of appendiceal mucinous adenomas in identical twins, including one with pseudomyxoma peritonei. Am. J. Surg. Pathol. 25, 1095–1099. 10. Royall, R. (1997) Statistical Evidence: A Likelihood Primer. London: Chapman and Hall. 11. 11 Chang, H.-W., Ali, S. Z., Cho, S. R., Kurman, R. J., and Shih, I. M. (2002) Detection of allelic imbalance in ascitic supernatant by digital SNP analysis. Clin. Cancer Res. 8, 2580–2585. 12. Chang, H.-W., Lee, S. M., Goodman, S. N., et al. (2002) Assessment of plasma DNA levels, allelic imbalance and CA-125 as diagnostic tests for cancer. J. Natl. Cancer Inst. 94, 1697–1703.
142
H-W. Chang and I-M. Shih
Representational Difference Analysis
143
9 Representational Difference Analysis as a Tool in the Search for New Tumor Suppressor Genes Antoinette Hollestelle and Mieke Schutte
Summary The recognition of a homozygous deletion of genetic material in a tumor genome has been instrumental in several tumor suppressor gene searches. The representational difference analysis (RDA) allows one to identify homozygous deletions even from among the high background of allelic losses that is typical for most cancers. RDA is a polymerase chain reaction (PCR)-based subtractive hybridization method. Two major obstacles to successful enrichment of target sequences from complex genomes were circumvented by RDA. Incomplete reassociation of complex DNA populations is overcome by using representative subpopulations of the tester and driver genomes. In addition, reiterated hybridization, selection, and amplification of the difference products introduces a kinetic component in the enrichment of target sequences. RDA thus enables the identification of homozygous deletions as small as 100 kilobases. Here, we provide a detailed protocol of the RDA procedure, including reflections on frequently encountered technical problems and on the particulars of its application in cancer. Key Words: Allelic loss; gene identification; homozygous deletion; kinetic enrichment; protocol; representational difference analysis; subtractive hybridization; tumor suppressor gene.
1. Introduction Cancer is characterized by the dysregulation of cellular processes that govern the proliferation of cells, their differentiation, their integrity, and their death. This loss of regulation has a genetic basis (1). Cancer differs from most other genetic diseases in that multiple genes are involved and that most of the mutations are somatically introduced. The genes involved in cancer are classified as oncogenes (with dominant “gain of function” mutations), tumor suppressor genes (recessive “loss of function” mutations), and DNA repair genes (“guardians of the genome”).
From: Methods in Molecular Medicine, Vol. 103: Pancreatic Cancer: Methods and Protocols Edited by: G. Su © Humana Press Inc., Totowa, NJ
143
144
Hollestelle and Schutte
Although most currently known cancer genes have been identified through linkage analysis, the recognition of chromosomal aberrations involving the putative gene was pivotal in most gene searches. Representational difference analysis (RDA) can be used to identify a variety of genetic aberrations, by molecular comparison of a tumor genome with the genome of non-neoplastic cells from the same donor (2,3). RDA has proven to be particularly powerful in the identification of homozygous deletions from among a high background of heterozygous deletions in tumor genomes. Homozygous deletions result in the total loss of genetic information, and the cellular effect of most large homozygous deletions is therefore assumed to be deleterious. Indeed, homozygous deletions are rare and, when present, they are relatively small. The a priori premise that a homozygous deletion in a cancer represents the genetic locus of a tumor suppressor gene is therefore considered to be strong. The potential of the RDA is illustrated by the identification of the BRCA2 and PTEN tumor suppressor genes, both of which were identified through homozygous deletions that had been found by RDA (4,5). Several modifications of the RDA procedure have also proven their worth in cancer, as well as in other fields of research. The genomic RDA has been coupled to chromosome-specific YAC clone arrays to provide information on the position of the target genes (6). The use of cDNA as starting material instead of DNA results in isolation of differentially expressed sequences (7). Coupling of cDNA RDA to microarray hybridization allowed high-throughput analysis of multiple representations (8). Finally, methylated CpG island amplification (MCA) RDA is designed to isolate sequences that are differentially methylated in normal and tumor cells (9). 1.1. The Evolution of Subtractive Hybridization Techniques RDA is a subtractive hybridization technique (2,10). In subtractive hybridization techniques, one DNA population (the “driver”) is hybridized in excess with a second DNA population (the “tester”), which is similar but not identical to the first DNA population. The differences between these two DNA populations that are present in the tester but not in the driver are the “target” sequences, which are the objectives for isolation. Typically, the tester DNA is digested with a restriction endonuclease and mixed with an n-fold excess of sheared driver DNA. The DNA mixture is then denatured and allowed to reassociate to recover three types of hybrids: tester–tester and driver–driver homohybrids and tester– driver heterohybrids. The chance that a nontarget tester sequence hybridizes with its tester complement is substantially smaller than the chance that it hybridizes with its complement from the driver DNA population (i.e., once vs n times). The target sequence is present only in the tester and will therefore always hybridize with its complement from the tester DNA population. Because only tester DNA had been digested with a restriction endonuclease, the tester–tester homo-
Representational Difference Analysis
145
Fig. 1. The reassociation reaction of eukaryotic genomes spans up to eight Cot values and can be divided into three phases corresponding the three components of an eukaryotic DNA population: highly repetitive sequences (the fast component), moderately repetitive sequences (the intermediate component), and unique sequences (the slow component). Co, the DNA concentration at the beginning of the reassociation reaction; t, reaction time (13).
hybrids can now be cloned selectively, generating a library in which the target sequences are enriched n times in proportion to the nontargets. The excess of driver DNA is thus used to literally drive the nontarget tester sequences from the pool of tester DNA. Lamar and Palmer first used subtractive hybridization successfully in 1984 to generate a library enriched for murine Y chromosome DNA, by hybridizing sheared DNA from a female in 100-fold excess with MboI-digested DNA from a male (11). A variation on this method was published by Kunkel et al., who used a phenol-enhanced reassociation technique (PERT) to increase reassociation rates and thus isolated a probe from an X chromosome interstitial deletion in a male patient (12). However, this approach was still not powerful enough to isolate small genetic aberrations from complex mammalian genomes, mainly because of two problems. The first problem was incomplete reassociation. Reassociation of a complex DNA population takes place in three phases: the fast, intermediate, and slow components (Fig. 1; [13]). Highly repetitive sequences, such as microsatellites or ALU sequences, reassociate relatively fast with their abundant (near-identical) complements. In the intermediate component, reassociation of moderately repetitive sequences takes place, such as the immunoglobulin superfamily of genes. The unique sequences in the genome will reassociate in the slow component of the reassociation curve, if they reassociate at all. Complete reassocia-
146
Hollestelle and Schutte
tion of highly complex genomes, such as the human genome, is practically impossible. Target sequences usually are unique sequences and thus reassociate during the last phase of the reaction. Incomplete reassociation therefore preferentially prevents the isolation of target sequences. The second problem was insufficient enrichment of target sequences. Lamar and Palmer and Kunkel et al. had been successful because their target regions were the Y chromosome and a 5-Mb hemizygous deletion, respectively (11, 12). The abundance of restriction fragments from these target regions was such that they were still able to isolate target sequences, even with a limited enrichment of the target sequences and consequently a rather poor representation of the target regions in the difference products. Insufficient enrichment will, however, severely hamper the isolation of sequences from smaller target regions, as will usually be the case. Strauss et al. improved the subtractive hybridization technique by applying multiple rounds of hybridization, hence increasing the enrichment through the second-order kinetics of self-association of target sequences that had been enriched in the previous round of hybridization (14). Biotinylation of the driver allowed the removal of driver–driver and tester–driver hybrids after hybridization by using avidin-coated polystyrene beads. Tester–tester homohybrids were subsequently denatured and again hybridized with new biotinylated driver DNA. The difference products from the last round were ligated to adapters, amplified, and cloned. Although this approach was a substantial improvement of the subtractive hybridization technique, the isolated probe originated from a 5-kb deletion representing only 1/4000th of the yeast genome. Also, the complexity of the yeast genome is about one order of magnitude less than that of a typical mammalian genome and incomplete reassociation thus should not have been a problem. Wieland et al. used multiple rounds of hybridization to isolate sequences from hemizygous target regions of 30-kb and 1-Mb in a human genome (15). They opted for biotinylation of the tester DNA population that had been digested with Sau3A restriction endonuclease, upon which adapter primers were ligated to the tester. Thus generated tester was mixed with sheared driver DNA, denatured, and allowed to reassociate to 90% completion. The ssDNA fraction was then isolated by using hydroxylapatite. As the target sequences are unique sequences and thus presumably had not yet associated with their complements, the isolated ssDNA fraction would be enriched for target sequences. The ssDNA was subsequently mixed with new driver and the hybridization and selection cycle was reiterated twice. The difference products from the last round were purified by avidin/biotin affinity chromatography, amplified, and cloned. An 100- to 1000-fold enrichment was thus obtained in the complex human genome. The isolation of sequences from target regions smaller than 1 kb became feasible only with the introduction of the RDA by Lisitsyn et al. in 1993 (2).
Representational Difference Analysis
147
RDA differed from earlier subtractive hybridization techniques in that the genomic complexity was reduced by the use of “representations” of the tester and driver DNA populations. Both the tester and the driver DNA populations were digested with a restriction endonuclease, ligated to adapter primers, and amplified by the polymerase chain reaction (PCR). Such “whole genome” PCR will preferentially amplify DNA fragments smaller than 1000 to 1500 bp (16), thus generating amplicons that represent a subpopulation of the original tester and driver DNA populations. The choice of restriction endonuclease used to digest the original DNA populations determines the complexity of the generated representations. BamHI, for example, generates a mean fragment length of about 5500 bp in the human genome (17) and results in an estimated 55-fold reduction of genomic complexity in the representations (2), whereas HindIII generates an average fragment length of about 2000 bp and results in an estimated eightfold complexity reduction (recognition sequences are GGATCC and AAGCTT, respectively). Although the use of representations implies that only a proportion of the target sequences can possibly be isolated, it does allow complete reassociation of human sequences. When necessary, several representations generated by using different restriction endonucleases may be analyzed to isolate more of the target sequences present in the original tester DNA population. In the RDA, the use of adapter primers also enabled the efficient amplification of DNA populations at various steps in the procedure. Adapter primers were used to generate tester and driver representations in sufficient amounts, thereby allowing the use of small quantities of starting material. Adapter primers were removed from both representations prior to the hybridization reaction. New adapter primers were then ligated to only the tester representation, thus allowing the selective amplification of only the tester–tester homohybrids after the hybridization reaction. After this PCR, the tester–driver and driver– driver hybrid molecules had become single-stranded and were removed by digestion with mung bean nuclease (with specificity for ssDNA). The remaining dsDNA population would thus be enriched for target sequences. After reiterated hybridization, selection, and amplification, the difference products were cloned and analyzed. The highly selective and efficient PCR-based isolation of tester–tester homohybrids in the RDA took full advantage of the kinetic enrichment resulting from the multiple rounds of subtractive hybridization. It is this kinetic component of the RDA that allows an enrichment of target sequences of about 105-fold after two rounds and more than 1010-fold after three rounds in a genome as complex as the human genome (2). 1.2. Representational Difference Analysis: The Procedure The RDA consists of two phases: the generation of the representations and the subtractive hybridization, which is reiterated until the target sequences are
148
Hollestelle and Schutte
Table 1 Fragment Lengths for Restriction Endonucleases Commonly Used in RDA (17)
Enzyme
Sequence
Mean fragment lengths
BamHI BglII EcoRI HindIII Sau3A/MboI
GGATCC AGATCT GAATTC AAGCTT GATC
5534 2699 3013 1873 318
Percent of fragments <100 bp
Percent of fragments <1000 bp
1.7 3.5 3.1 4.9 26.4
16.5 30.9 28.2 41.3 95.8
Fig. 2. Ligation of RBam24 adapters to BamHI-digested DNA fragments and subsequent PCR amplification (see Subheading 3.1.2.).
sufficiently enriched to be cloned and analyzed. In our protocol (see Subheading 3.), the representations are generated by using BamHI restriction endonuclease (see Subheading 3.1.1.), generating fragments smaller than 1 kb for about 15% of the original DNA populations (Table 1). BglII, EcoRI, and HindIII restriction endonucleases are also commonly used in RDA (see Note 1). RBam24 adapter primers are then ligated to the 5' ends of the DNA fragments (see Subheading 3.1.2.). The RBam12 adapter is used to generate a double-stranded ligation template with the RBam24 adapter, but RBam12 will not be ligated to the DNA fragments because this synthetic molecule lacks the 5' phosphate group. The ligation mixture is then incubated at 72°C, during which the RBam12 adapters dissociate from the DNA fragments. The 3' ends of the DNA fragments are subsequently extended by Taq polymerase, thereby generating primer sites for the RBam24 adapters (Fig. 2). The two populations of DNA fragments are then amplified in a whole genome PCR using RBam24 as primers (see Sub-
Representational Difference Analysis
149
heading 3.1.3.). It is critical to generate these amplicons in a highly reproducible manner, to prevent the introduction of artifactual differences between the representations (see Note 2). The tester amplicons are selected for DNA fragments between 150- and 1000-bp of size using a preparative low-melting-point agarose gel (see Subheading 3.1.4. and Note 3). If the amplification of the tester representation had unintentionally been somewhat more efficient with regard to the larger fragments than the amplification of the driver representation, the size selection ensures that these artifactual differences between the two representations are removed. Finally, both representations are amplified to produce sufficient tester and driver amplicons for reiterated subtractive hybridization (see Subheading 3.1.5.). The RBam24 adapters are removed from both the tester and driver amplicons and JBam24 adapters are ligated to the tester amplicons (see Subheadings 3.1.6. and 3.1.7.). For the subtractive hybridization (Fig. 3), tester and driver amplicons are mixed in a 1:80 ratio, denatured, and allowed to reassociate at high DNA and salt concentrations (10 µg of DNA/µL and 1 M NaCl; see Subheading 3.2.1. and Note 4). The tester–tester homohybrids are selectively amplified using JBam24 primers (see Subheading 3.2.2.). After this PCR, the tester–driver and driver– driver hybrids will be single-stranded and are degraded by mung bean nuclease (see Subheading 3.2.3.). Another PCR is subsequently performed to generate sufficient difference products (see Subheading 3.2.4.). For the second round of subtractive hybridization, the JBam24 adapters of the difference products from the first round are exchanged for NBam24 adapters, whereafter these DNA fragments are mixed with new driver amplicons (see Table 2, Subheading 3.2.5.). JBam24 and NBam24 adapters alternate in all following rounds of subtractive hybridization. Enrichment of target sequences can be visualized by electrophoresis of the difference products in standard agarose gels, where a smear of DNA fragments from the representations gradually is replaced in successive rounds by a banding pattern of DNA fragments of discrete length. Thus enriched target sequences should now be analyzed for their presence in the tester and driver representations, to discriminate target sequences from nontarget sequences that have escaped the subtractive hybridization. The ratio of tester:driver DNA during the hybridization reaction determines the subtractive enrichment. Although a 1:80 ratio of tester:driver DNA theoretically could result in an enrichment of target sequences of about 80 times, the real subtractive enrichment is estimated to be 50 (2). Subsequent rounds of hybridization also involve kinetic enrichment, owing to the second-order kinetics of self-association of target sequences that had been enriched in a previous round. The kinetic enrichment is determined by the total enrichment achieved in the previous round (squared). For example, if the subtractive enrichment e is 50, the total enrichment after two rounds of hybridization is 50 ´ (50)2 = 503
150
Hollestelle and Schutte
Fig. 3. Schematic representation of the subtractive hybridization as described in Subheading 1.2.
(or 105), and the total enrichment after three rounds is 50 ´ (503)2 = 507 (or more than 1010). 1.3. Representational Difference Analysis of Tumor Genomes The RDA has been pivotal in the identification of the BRCA2 and PTEN tumor suppressor genes (4,5). For both genes, the identification of a homozygous deletion in a tumor genome narrowed the gene search to only a few hundreds of kilobases, thus allowing the identification of the target gene by a position cloning approach. Although Lisitsyn et al. have shown that RDA can be applied to identify deletions as well as amplifications in tumor genomes (3), we believe that RDA’s potential to identify homozygous deletions is most valuable in the
Representational Difference Analysis
151
identification of cancer genes. Homozygous deletions can be identified by RDA even from among the typically high background of heterozygous deletions present in most tumor genomes (4). RDA identifies heterozygous deletions only when the deletions involve restriction fragment length polymorphisms of which the length of the lost DNA fragment is within the size range of the representations and the retained DNA fragment is outside this size range. In other words, the 2:1 allele ratio within the original genomes is converted to a 1:0 allele ratio in the representations, and the enrichment of the smaller “target” allele is thus no longer prevented by the larger allele that had been retained in the original driver population (i.e., the tumor). For homozygous deletions, both alleles are absent from the original driver and thus all target sequences (whether or not polymorphic) will be identified by RDA provided their length is within the size range of the representations. Assuming a polymorphism frequency in the human genome of about 1 in 300 bp, RDA will identify target sequences from a homozygously deleted region 50 times more efficient than from a heterozygous deletion when using a restriction endonuclease with a 6-bp recognition sequence. Indeed, we isolated eight times more frequently target sequences from heterozygous deletions than from homozygous deletions after two rounds of subtractive hybridization (4), whereas target sequences from heterozygous deletions were only twice as frequent after three rounds (3). The RDA will isolate one DNA fragment for each 50–150 kb of homozygous deletion (10,18). The use of a single restriction endonuclease to generate the representations therefore generally suffices to identify homozygous deletions in a tumor genome, if present. The application of RDA in cancer requires pure tumor cell populations. With an estimated subtractive enrichment e of 50 (2), as few as 2% contaminating non-neoplastic cells in a tumor sample will render the sample essentially useless as a driver in the RDA. Cancer cell lines, cancer xenografts, or highly purified cancer cells, for example, by microdissection or FACS, should therefore be used as starting material in the RDA. We are, however, hesitant to use primary tumor specimens because DNA isolated from such samples frequently is of poor quality, which will introduce artifactual target sequences. When using cancer xenografts, we recommend doubling the reassociation time of the first hybridization reaction, to compensate for the increased genomic complexity in the murine background (4). Although any normal, non-neoplastic human DNA sample can be used as tester, the polymorphic differences between two unrelated DNA samples (approx 0.3% between two human genomes) will introduce unwanted target sequences. Constitutional DNA from the donor of the tumor is therefore preferred. We recommend using samples from female patients, as half of all male tumors have deleted Y chromosome sequences. These hemizygous deletions will be isolated by RDA as effective as homozygous deletions. When
152
Hollestelle and Schutte
using male tumor samples, the isolation of primarily Y sequences can be prevented by admixture of DNA from an Y monochromosomal somatic cell hybrid to the tumor DNA (Schutte, unpublished results). 2. Materials (see Note 5) 1. LoTE: 3 mM Tris-HCl, pH 8.0, 0.2 mM EDTA, pH 8.0. 2. Adapter primers: HPLC purified and reconstituted in LoTE at 50 µM. RBam12: 5'-GATCCTCGGTGA-3' RBam24: 5'-AGCACTCTCCAGCCTCTCACCGAG-3' JBam12: 5'-GATCCGTTCATG-3' JBam24: 5'-ACCGACGTCGACTATCCATGAACG-3' NBam12: 5'-GATCCTCCCTCG-3' NBam24: 5'-AGGCAACTGTGCTATCCGAGGGAG-3' For sequences of adapters to other endonucleases, see ref. 10. 3. Agarose (molecular biology grade). 4. BamHI restriction endonuclease, supplied with 10X restriction buffer. 5. EE 3X buffer: 30 mM EPPS (N-[2-hydroxyethyl]piperazine-N'-[3-propanesulfonic acid], 3 mM EDTA, pH 8.0. 6. Ept (DNA precipitation using ethanol): Mix one volume of aqueous solution with one-third volume of 10 M NH4OAc and 2 µL of glycogen carrier, then mix with three volumes of 100% ethanol. The DNA is pelleted by microcentrifugation at maximum speed for 5 min at room temperature. 7. Glycogen (Molecular biology-grade). 8. Ipt (DNA precipitation using isopropanol; used for removing adapter primers): Mix one volume of aqueous solution with one-tenth volume of 3 M NaAc, pH 5.2, and 2 µL of glycogen carrier, then mix with one volume of isopropanol. The DNA is pelleted by microcentrifugation at maximum speed for 5 min at room temperature. 9. T4 Ligase (400 New England Biolabs-U/µL), supplied with 10X T4 ligase buffer. 10. LMP: Low-melting-point agarose (molecular biology grade). 11. MBN: Mung bean nuclease (10 New England Biolabs-U/µL), supplied with 10X MBN buffer. 12. 25 mM Magnesium chloride. Preferably purchased in solution (PCR-grade) as problems may be encountered with “home-made” MgCl2 solutions owing to adherence of the salt molecules to certain types of plastic. 13. Mineral oil (molecular biology grade). 14. dNTPs: An equal molarity mixture (10 mM each) of dATP, dCTP, dGTP, and dTTP diluted in LoTE. 15. PC8 (extraction with phenol–chloroform–isoamylalcohol [25:24:1], pH 8.0): Mix one volume of aqueous solution with one volume of PC8. Separate the phases by microcentrifugation at maximum speed for 1 min at room temperature. 16. PCR: Performed in thin-walled tubes, with two drops of mineral oil, in an Omnigene thermocycler (Hybaid) using the tube control mode.
Representational Difference Analysis 17. 18. 19. 20. 21.
22. 23. 24. 25.
153
Q20 tip: Column for anion-exchange chromatography (Qiagen 10023). QBT: Equilibration buffer QBT (Qiagen 19054). QC: Wash buffer QC (Qiagen 19055). QF: Elution buffer QF (Qiagen 19056). RDA 10X buffer: 670 mM Tris, pH 8.8, 160 mM (NH4)2SO4, 100 mM b-mercaptoethanol, 1 µg/µL of bovine serum albumin (acetylated and molecular biology grade). 5 M Sodium chloride. Taq polymerase (5 Promega-U/µL). 50 mM Tris buffer, pH 8.9. W: Wash with 70% ethanol; includes vortexing and 1 min microcentrifugation at maximum speed. Routinely, a DNA precipitation is followed by two washes. When the next procedure is a ligation, three washes are performed.
3. Methods 3.1. Generation of the Representations 3.1.1. Digestion of Genomic DNA 1. Incubate genomic DNA in a restriction reaction mixture for 1 h at 37°C: 1 µg of genomic DNA (driver or tester) + 30 µL of BamHI 10X restriction buffer + x µL of LoTE + 1 µL of BamHI enzyme (10 U/µL) = 300 µL. 2. Extract the DNA with PC8. 3. Ethanol precipitate the DNA (Ept), wash three times with 70% ethanol (3W), and resuspend the DNA in 6 µL of LoTE. 4. Repeat the digestion and purification until (near) complete digestion (see Note 1).
3.1.2. Ligation to RBam24 Adapters 1. Mix 6 µL of digested DNA (driver or tester from Subheading 3.1.1.) + 10 µL RBam12 primers (50 µM) + 10 µL of RBam24 primers (50 µM) + 3 µL of 10X T4 ligase buffer = 29 µL. Bring the reaction mixture to 50°C, and then gradually cool the sample over a period of 1 h down to 10°C. 2. Add 1 µL of T4 ligase to the reaction mixture and incubate overnight at 16°C (see Note 6).
3.1.3. PCR-1 Amplicons 1. Set up the PCR mix: 3 µL of template DNA (approx 100 ng driver or tester from Subheading 3.1.2.) + 18 µL of 10X RDA buffer + 32 µL of MgCl2 (25 mM) + 4 µL of dNTPs (10 mM) + 119 µL of LoTE = 176 µL. 2. First incubate the PCR mix for 3 min at 72°C, then add 20 µL of Taq polymerase (20 U; diluted 1:5 in 1X RDA buffer). 3. Incubate the PCR–Taq mix for 3 min at 72°C, then add 4 µL of RBam24 primers (50 µM).
154
Hollestelle and Schutte
Fig. 4. PCR amplification curve of tester (T) and driver (D) amplicons for a range of 14–24 cycles. Smearing of ssDNA toward the wells (20, 22, and 24 cycles) and inefficient amplification of the larger sequences (24 cycles) are both indicative of exhaustion of PCR reagents.
4. Perform an amplification curve with 14, 16, 18, and 20 cycles of 1 min at 94°C and 3 min at 72°C, with a final extension of 5 min at 72°C (see Note 2). 5. Analyze 5–10 µL of the products on a 1% agarose gel. Run the gel exactly 3 min at 100 V, and quantify by comparison to a DNA standard. About 10 µg of amplicons should be generated per 200 µL of PCR reaction. Run the gel for another hour, and select the sample with the number of cycles that gives maximal recovery, but does not show smearing of the products toward the well (Fig. 4). 6. Add 6 µL of RBam24 primers (50 µM), 1 µL of 10X RDA buffer, 2 µL of LoTE, and 1 µL of Taq polymerase (5 U) to the selected PCR product with the optimal number of cycles. Then perform an additional cycle of 1 min at 94°C and 5 min at 72°C. 7. Extract the DNA with PC8. Ethanol precipitate (Ept), wash (2W), and resuspend the DNA in 10 µL of LoTE.
3.1.4. Size Selection of Tester Amplicons 1. Load 5–10 µg of tester amplicons (about half of the products generated in Subheading 3.1.3.) on a 1.5% LMP agarose gel, separate, and excise the amplicons from the gel between 150 and 1000 bp (see Note 3). 2. Purify the amplicons by Qiagen chromatography: Melt the agarose block at 70°C, add five volumes of QBT, and mix. Keep the gel mix at 70°C and load in 500-µL aliquots on a Q20 tip that has been equilibrated with 1 mL of QBT. Then wash with 6 mL of QC and elute with 800 µL of QF. 3. Precipitate by adding 2 µL of glycogen and 650 µL of isopropanol, then microcentrifuge at maximum speed for 10 min at room temperature, wash (2W), and resuspend in LoTE at approx 100 ng/µL (assume 30% recovery). 4. Gel-quantify and adjust the DNA to a final concentration of 100 ng/µL.
Representational Difference Analysis
155
3.1.5. PCR-2 Amplicons 1. Perform a PCR on approx 1.5 ng of the gel-purified tester amplicons from Subheading 3.1.4., in 200 µL PCR reactions, as described in Subheading 3.1.3. Generate an amplification curve with 14, 16, 18, and 20 cycles. 2. Analyze 5–10 µL on a gel, and quantify and select the sample with the optimal number of cycles, as described in Subheading 3.1.3. 3. Generate approx 5 µg of tester amplicons and approx 120 µg of driver amplicons, according the selected optimal number of PCR cycles (including an additional cycle). 4. Extract the DNA samples with PC8. Isopropanol precipitate (Ipt), wash (2W), and resuspend the DNA in LoTE. 5. Gel-quantify and adjust the amplicons to a final concentration of 1 µg/µL.
3.1.6. Digestion of Amplicons 1. Scale up the following reaction according to the amount of amplicons to be digested: 1 µg amplicons (driver or tester from Subheading 3.1.5.) + 1 µL BamHI 10X restriction buffer + 0.4 µL of BamH1 enzyme (10 U/µL) + 7.6 µL of LoTE = 10 µL. 2. Incubate for 1 h at 37°C. 3. Extract the DNA sample with PC8. Isopropanol precipitate (Ipt without glycogen), wash (3W), and resuspend the DNA in LoTE. 4. Gel-quantify and adjust the amplicons to a final concentration of 0.5 µg/µL.
3.1.7. Ligation of Tester Amplicons to JBam24 Adapters 1. Set up the annealing reaction: 4 µL of tester amplicons (2 µg from Subheading 3.1.6.) + 20 µL of JBam12 primers (50 µM) + 20 µL of JBam24 primers (50 µM) + 6 µL 10X T4 ligase buffer + 8 µL of LoTE = 58 µL. Bring the reaction mixture at 50°C, and then gradually cool the sample over a period of 1 h down to 10°C. 2. Add 2 µL of ligase and incubate overnight at 16°C. 3. Test the ligation by PCR using either the previous primers (RBam24) or the new primers (JBam24), with approx 35 ng of the ligated template in a 50 µL of PCR reaction, for 25 cycles. The amplification with the previous primers generally gives some product, but the amount of product should not be more than half the amount generated with the new primers (see Note 6).
3.2. Subtractive Hybridization 3.2.1. Hybridization-1 1. In a microcentrifuge tube, mix 500 ng of tester-JB amplicons (from Subheading 3.1.7.) and 40 µg of driver amplicons (from Subheading 3.1.6.) in a final volume of 200 µL. 2. Extract the DNA with PC8. Isopropanol precipitate (Ipt without glycogen), wash (2W) and carefully resuspend the DNA in approx 3.5 µL 3X EE buffer to a final volume of 4 µL (see Note 4). 3. Overlay with one drop of mineral oil. 4. Heat for 5 min at 98°C.
156
Hollestelle and Schutte
5. Add 1 µL of 5 M NaCl at 98°C, carefully mix by pipetting (see Note 4). 6. Incubate for 20 h at 67°C (see Note 4). 7. Dilute to 500 µL final volume with LoTE (see Note 4).
3.2.2. PCR-1 of Hybridization-1 1. Perform two PCR reactions with each 50 µL of diluted Hyb-1 product as template (from Subheading 3.2.1.). Perform PCR as described in 3.1.3., except that JBam24 primers are used, 10 cycles are performed, and Taq polymerase is added after 3 min at 85°C in stead of 72°C (see Note 7).
3.2.3. Mung Bean Nuclease Treatment of Hybridization-1 1. Both PCR products of 3.2.2 are purified by extracting the DNA with PC8, precipitating with ethanol (Ept), and washing (2W). Each PCR product is resuspended in 5 µL of LoTE. 2. Add to each reaction: 29 µL of LoTE + 4 µL of 10X MBN buffer + 2 µL of MBN enzyme (10 U/µL) = 40 µL. 3. Incubate for 30 min at 30°C (see Note 8). 4. Add 160 µL of 50 mM Tris, pH 8.9, and inactivate for 5 min at 98°C. 5. Pool both MBN-treated Hyb-1 samples.
3.2.4. PCR-2 of Hybridization-1 1. Perform an amplification curve with 15, 17, and 19 cycles in 200 µL of PCR reactions with 30 µL of MBN-treated Hyb-1 sample as template (from Subheading 3.2.3.). 2. Analyze 5–10 µL of the products on gel and select the sample with the optimal number of cycles. 3. Perform an additional cycle of amplification on the selected PCR sample with the optimal number of cycles, as described in Subheading 3.1.3. 4. Extract the DNA with PC8. Ethanol precipitate (Ept), wash (2W), and resuspend the DNA in 10 µL of LoTE. 5. Analyze 5–10 µL of the difference products on a 1% agarose gel.
3.2.5. Successive Rounds of RDA 1. For the second round of hybridization and selection, go to Subheadings 3.1.6. and 3.1.7. for switching the primers of the generated tester difference product. JBam24 and NBam24 primers alternate with each round. 2. Proceed with hybridization-2 as for hybridization-1 (Subheadings 3.2.1.–3.2.4.). 3. Adjust the quantities of tester difference product according to the indications in Table 2. These amounts serve as a guideline; the amounts depend on the fraction of target sequences in the tester sample and on the enrichment factor of the previous round(s).
4. Notes 1. It is important to obtain near complete digestions, as partial digestions result in different cleavage patterns among tester and driver and will thus introduce arti-
Representational Difference Analysis
157
Table 2 Quantities of Tester or Tester Difference Product (TDP) For Each Hybridization No. Hyb 1 2 3 4
Amount of DNA used in hybridization reaction 500 ng of tester-JB/40 µg of driver 50 ng of TDP-NB/40 µg of driver 100 pg of TDP-JB/40 µg of driver 5 pg of TDP-NB/40 µg of driver
Primers JBam24 NBam24 JBam24 NBam24
factual difference products. Upon electrophoresis of BamHI-digested genomic DNA fragments, the peak of ethidiumbromide intensity should run at 8–8.5 kb. When restriction endonucleases are used that produce mean fragment lengths larger than BamHI, it will be more difficult to generate reproducible representations. In addition, the representations may then contain a too small fraction of the original sample to allow isolation of sufficient target sequences. When using endonucleases that produce mean fragment lengths smaller than HindIII, the complexity of the human DNA samples will not be reduced enough to allow complete reassociation. It is recommended to use restriction endonucleases that produce mean fragment lengths in the range of those generated by HindIII and BamHI (10,17; Table 1). 2. The PCR in RDA resembles a “whole genome PCR” (16) in that an entire DNA population of sequences is amplified instead of an unique sequence, as is typical for a standard PCR. A whole genome PCR generates more products and therefore requires more PCR reagents and enzyme. Exhaustion of reagents affects primarily the amplification of longer DNA fragments. To obtain reproducible representations, the PCR should be optimized for the number of cycles, without exhausting any of the PCR reagents. This can be visualized on a standard agarose gel. Smearing of the amplicons toward the well is indicative for ssDNA and thus exhaustion of the PCR reagents (Fig. 4). After choosing the optimal number of cycles, an additional cycle with new primers and polymerase is performed to ensure the production of dsDNA. 3. Fragments smaller than 150-bp and larger than 1 kb are removed from the representation. Larger fragments are removed, as competition in the PCR affects primarily the amplification of larger fragments, which may result in artifactual differences between the tester and driver representations. Fragments smaller than 150 bp will be largely made up of adapter sequences, resulting in inappropiate reassociation of adapter sequences rather than DNA sequences from the test samples. 4. The efficiency of the hybridization reaction is determined by the DNA concentration (C0t: C0 times t, where C0 is the DNA concentration at the beginning of the hybridization reaction and t is the time of reannealing). The amount of driver amplicons should therefore be quantified accurately to ensure that the DNA concentration is near its limit of solubility (approx 10 µg/µL). It is also critical to quantify the tester amplicons accurately because the ratio of driver to tester DNA determines
158
5.
6.
7. 8.
Hollestelle and Schutte
the enrichment of the hybridization reaction. After resuspension of the 40-plus µg driver and tester, the DNA solution should be viscous and look cloudy. When insecure, it is strongly recommended to experiment with salmon sperm DNA or similar. Care should be taken that the 5 M NaCl is indeed mixed with the DNA solution, as this high-salt solution tends to “jump” into the oil layer (use a P2 pipet). Dilution of the hybridization mixture after the hybridization reaction is to be performed stepwise, as dilution of a high-salt solution may present problems. From the second round onwards, hybridization times should not exceed 20 h in order to exploit the achieved enrichment of the target sequences compared to the nontarget tester sequences (i.e., their reassociation kinetics have shifted to the left in the Cot curve; see Fig. 1). Prepare all solutions using milliQ purified and autoclaved water. Frequently take fresh aliquots of solutions and use filter tips throughout, to prevent tester to driver cross-contaminations. RDA is a complicated technique, and it is essential that the experimentator masters basic molecular biology methods. We strongly recommend reading a textbook on hybridization kinetics and the major reviews on the RDA method (2,3,10,13). The most common reasons for failing RDA are deviations of the protocol or taking short cuts. We strongly advise to follow the protocol as it is. Our protocol describes RDA of genomic DNA. Although the essence of the RDA protocol is the same for other applications, such as cDNA RDA (7) or MCA RDA (9), we refer to the original reports for the particulars of these modifications. Note that the ligation units are given in NEB units (67 NEB units equal 1 cohesive ends unit). When switching adapters make sure that the amount of PCR product generated with the old adapters does not exceed the amount of PCR product generated with the new adapters, as this will affect the enrichment of the hybridization reaction. The high-stringency incubation of 3 min at 85°C is intended to reduce priming mediated by hybrids of near-identical repetitive elements (19). For degradation of ssDNA after the hybridization and amplification, mung bean nuclease should be used and not S1 nuclease. Both nucleases are ssDNA specific, but mung bean nuclease is unsensitive for nicks and nucleotide mismatches whereas S1 nuclease is not. The use of S1 nuclease would result in degradation of hybrids that are not exactly complementary, such as those that contain mutations introduced during the various PCR amplification steps.
References 1. 1 Vogelstein, B. and Kinzler, K. W. (1993) The multistep nature of cancer. Trends Genet. 9, 138–141. 2. 2 Lisitsyn, N., Lisitsyn, N., and Wigler, M. (1993) Cloning the differences between two complex genomes. Science 259, 946–951. 3. 3 Lisitsyn, N. A., Lisitsina, N. M., Dalbagni, G., et al. (1995) Comparative genomic analysis of tumors: Detection of DNA losses and amplification. Proc. Natl. Acad. Sci. USA 92, 151–155.
Representational Difference Analysis
159
4. 4 Schutte, M., da Costa, L. T., Hahn, S. A., et al. (1995) Identification by representational difference analysis of a homozygous deletion in pancreatic carcinoma that lies within the BRCA2 region. Proc. Natl. Acad. Sci. USA 92, 5950–5954. 5. 5 Li, J., Yen, C., Liaw, D., et al. (1997) PTEN, a putative protein tyrosine phosphatase gene mutated in human brain, breast, and prostate cancer. Science 275, 1943–1947. 6. 6 Zeschnigk, M., Horsthemke, B., and Lohmann, D. (1999) Detection of homozygous deletions in tumors by hybridization of representational difference analysis (RDA) products to chromosome-specific YAC clone arrays. Nucl. Acids Res. 27, e30. 7. Hubank, M. and Schatz, D. G. (1994) Identifying differences in mRNA expression 7 by representational difference analysis of cDNA. Nucl. Acids Res. 22, 5640–5648. 8. 8 Welford, S. M., Gregg, J., Chen, E., et al. (1998) Detection of differentially expressed genes in primary tumor tissues using representational difference analysis coupled to microarray hybridization. Nucl. Acids Res. 26, 3059–3065. 9. Toyota, M., Ho, C., Ahuja, N., et al. (1999) Identification of differentially methy9 lated sequences in colorectal cancer by methylated CpG island amplification. Cancer Res. 59, 2307–2312. 10. 10 Lisitsyn, N. and Wigler, M. (1995) Representational difference analysis in the detection of genetic lesions in cancer. Methods Enzymol. 254, 291–304. 11. Lamar, E. E. and Palmer, E. (1984) Y-encoded, species-specific DNA in mice: 11 Evidence that the Y chromosome exists in two polymorphic forms in inbred strains. Cell 37, 171–177. 12. Kunkel, L. M., Monaco, A. P., Middlesworth, W., Ochs, H. D., and Latt, S. A. 12 (1985) Specific cloning of DNA fragments absent from the DNA of a male patient with an X chromosome deletion. Proc. Natl. Acad. Sci. USA 82, 4778–4782. 13. Lewin, B. (1994) Genome size and genetic content, in Genes V. Oxford: Oxford University Press, pp. 657–676. 14. Strauss, D. and Ausubel, F. M. (1990) Genomic subtraction for cloning DNA corresponding to deletion mutations. Proc. Natl. Acad. Sci. USA 87, 1889–1893. 15. Wieland, I., Bolger, G., Asouline, G., and Wigler, M. (1990) A method for differ15 ence cloning: Gene amplification following subtractive hybridization. Proc. Natl. Acad. Sci. USA 87, 2720–2724. 16. Kinzler, K. W. and Vogelstein, B. (1989) Whole genome PCR: Application to the 16 identification of sequences bound by gene regulatory proteins. Nucl. Acids Res. 17, 3645–3653. 17. Bishop, D. T., Williamson, J. A., and Skolnick, M. H. (1983) A model for restric17 tion fragment length distributions. Am. J. Hum. Genet. 35, 795–815. 18. 18 Schutte, M., Rozenblum, E., Moskaluk, C. A., et al. (1995) An integrated highresolution physical map of the DPC/BRCA2 region at chromosome 13q12. Cancer Res. 55, 4570–4574. 19. Schutte, M., da Costa, L. T., Moskaluk, C. A., et al. (1995) Isolation of YAC insert sequences by representational difference analysis. Nucl. Acids Res. 23, 4127–4133.
160
Hollestelle and Schutte
Serial Analyses of Gene Expression
161
10 Serial Analyses of Gene Expression (SAGE) Jia Le Dai Summary Serial analysis of gene expression (SAGE) is a molecular biology technique that was developed to measure the global gene expression levels. It has been applied successfully to characterize transcriptomes, compare the transcript levels between normal and diseased tissues, and uncover novel molecules within defined signal transduction pathways. A detailed description is presented in this chapter of the procedures involved to prepare the SAGE libraries. Protocols for automated sequencing and other standard molecular biology techniques can be found elsewhere, and thus are not included herein. Key Words: Serial analysis of gene expression; tags; ditags; anchoring enzyme; tagging enzyme; cDNA.
1. Introduction With the advent of completion of the human genome project (1), more focus has been shifted toward the elucidation of the roles these genes play in normal and diseased tissues. To measure gene expression level in a global and comprehensive manner becomes both the opportunity and challenge of biomedical research. A number of technologies have been developed during the past two decades, including subtractive hybridization, differential display, representative display analysis, and cDNA or oligonucleotide arrays. These methods, however, have their own limitations. For example, subtractive hybridization and differential display are unable to provide comprehensive expression analyses and may miss transcripts that express at low levels. The cDNA or oligonucleotide arrays can be used only to evaluate the expression of previously identified genes. A new technique known as serial analysis of gene expression (SAGE) was developed in 1995 by Vogelstein and Kinzler’s group that allows one to study the gene expression in an unbiased and comprehensive way (2). In contrast to array technology, SAGE does not require prior knowledge of the gene, and the From: Methods in Molecular Medicine, Vol. 103: Pancreatic Cancer: Methods and Protocols Edited by: G. Su © Humana Press Inc., Totowa, NJ
161
162
Dai
data from SAGE are readily comparable between experiments. SAGE represents a powerful tool that has been used successfully to characterize transcriptomes, compare the gene transcription difference between normal and diseased tissues, and uncover novel signal pathway components (3–7). SAGE is based on two major principles. First, a short sequence (10 bp) provides sufficient information to be assigned to a unique transcript. A 10-bp sequence has a complexity of 410 (approx 1 106). Statistical calculation predicts that this complexity exceeds the number of genes in the human genome and the subset of genes expressed in given tissues. In fact, it has been shown more than 95% of the tags can be uniquely assigned to specific genes (2). Second, concatemerized forms can be generated from these short sequences that are suitable for high-throughput DNA sequencing. A single sequencing reaction is able to screen up to 30 tags. However, a meaningful SAGE library requires at least 50,000 tags. Both the throughput speed and cost have so far restricted its widespread application. Nevertheless, owing to the tremendous advantages this technique can offer, SAGE remains the method of choice for gene expression analyses. The SAGE technique involves multistep procedures that are diagrammed in Fig. 1. Like every gene expression profiling method, SAGE starts with conversion of mRNA to double-stranded cDNA using oligo-dT and reverse transcriptase. In SAGE, a 5'-biotinylated oligo-dT primer is used to immobilize the prepared cDNAs onto streptavidin-coated magnetic beads. This allows one to recover the 3' cDNA sequences following the anchoring enzyme digestion and to separate the SAGE tags from the rest of the 3' cDNA sequences (discussed below). The double-stranded cDNAs are first digested by a 4-bp cutter such as NlaIII (anchoring enzyme). The cDNA fragments that are 5' to the 3'most NlaIII site can be washed off, and only the 3' cDNAs that follow the NlaIII site can be captured. The isolated cDNAs are then split into two parts, each of which is ligated with a linker adaptor (linker A or B) that has a cohesive end with the NlaIII overhang. The linker adaptors serve at least two purposes. First, it provides the recognition sequences for the tagging enzyme. Second, it provides the primer sequence for subsequent polymerase chain reaction (PCR) amplification. The cDNAs are then digested with BsmFI (tagging enzyme) that is a type IIS restriction enzyme. BsmFI cleaves 14–15 bp 3' to its recognition sequence that is located on the linker primers. As a result, a 14- to 15-bp small fragment
Fig. 1. (Opposite page) A diagram of the multistep SAGE procedures. Doublestranded cDNAs are prepared using poly(A) RNA purified from varied tissue sources. SAGE ditags are generated from cDNAs through a series of restriction enzyme digestions, DNA ligations, and PCR amplifications. The ditags are then concatemerized, subcloned into a vector, and sequenced (see text for details.)
Serial Analyses of Gene Expression
163
164
Dai
(SAGE tag) together with the primer sequence is released. Every tag creates a small but unique sequence representing a specific transcript. Two pools of linker-adapted SAGE tags are first subject to Klenow fill-in. Then the two pools of tags are mixed and tail-to-tail blunt-end ligated in the presence of T4 DNA ligase. The ligation products contain the ligated tags (ditags) and the linker primer sequences on both ends. To amplify ditags, the ligation products are PCR amplified using primers 1 and 2, which are homologous to linker sequences. The ditags within the PCR products are then released from the linker adaptors by NlaIII digestion. The ditags are separated from the linker sequences by electrophoresis. The individual ditags are then contatermized by DNA ligation. The contatemerized product is then subcloned into a plasmid (digested with SphI) that bears an NlaIII overhang (CATG). After E. coli transformation, individual colony is PCR amplified using primers 1 and 2. The PCR products are then subject to direct sequencing using automated DNA sequencer. The tag sequences in the concatemer can be precisely identified, as each ditag is discretely bounded by the recognition sequences for the anchoring enzyme. The sequencing data will then be processed by SAGE software. Only the tags that meet the set criteria are accepted onto the tag list. For example, the ditags that fall out of the normal length range will be eliminated. A report is generated profiling the sequence and the occurrence of every unique tag, and tag abundances between projects may be compared. 2. Materials (see Note 1) 2.1. mRNA Preparation and cDNA Synthesis 1. 2. 3. 4. 5. 6.
Trizol reagent (Life Technologies Inc., cat. no. 10296-028). Messagemaker kit (Life Technologies Inc., cat. no. 10298-016). Superscript Choice System (Life Technologies Inc., cat. no. 18090-019). 10 U/µL of T4 polynucleotide kinase. 10 mM ATP. Biotinylated oligo-dT18 (polyacrylamide gel electrophoresis [PAGE]-purified) (see Note 2). 7. Linkers (PAGE-purified) (see Note 2): 1A: 5'-TTTGGATTTGCTGGTGCAGTACAACTAGGCTTAATAGGGACATG-3' 1B: 5'-TCCCTATTAAGCCTAGTTGTACTGCACCAGCAAATCC-3' (amino-modified C7) 2A: 5'-TTTCTGCTCGAATTCAAGCTTCTAACGATGTACGGGGACATG-3' 2B: 5'-TCCCCGTACATCGTTAGAAGCTTGAATTCGAGCAG-3' (amino-modified C7)
2.2. Preparation of Ditags 1. NlaIII (10 U/µL) (see Note 3). 2. BsmFI (4 U/µL).
Serial Analyses of Gene Expression 3. 4. 5. 6. 7. 8.
165
Klenow (1 U/µL). Dyna-bead M-280 Streptavidin beads (10 mg/mL, Dynal, cat. no. 112.05). Magnet stand (Promega, cat. no. Z5342). Primer 1: 5'-GGATTTGCTGGTGCAGTACA-3'. Primer 2: 5'-CTGCTCGAATTCAAGCTTCT-3'. 2X B+W buffer (10 mM Tris-HCl, pH 7.5, 1 mM EDTA, 2.0 M NaCl).
2.3. Contatemerization of Ditags and Automated Sequencing 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12.
T4 DNA ligase (5 U/µL). T4 DNA ligase (1 U/µL). SphI. pZErO-1 (Invitrogen, cat. no. K2500-01). Zeocin (Invitrogen, cat. no. R250-01). ElectroMAX DH10B competent E. coli (Life Technologies, cat. no. 18290-015). MicroPulser Electroporator (Bio-Rad, cat. no. 165-2100). Qiagen gel extraction kit (Qiagen, cat. no. 28704). ABI-21M13 Dye Primer FS sequencing kit (Perkin Elmer, cat. no. 402111). ABI-373 automatic sequencer. M13F primer: 5'-GTAAAACGACGGCCAGT-3'. M13R primer: 5'-GGAAACAGCTATGACCATG-3'.
2.4. Other Reagents and Equipments 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21.
Diethyl pyrocarbonate (DEPC)-treated H2O. Distilled H2O. PC8: Phenol–chloroform equilibrated with Tris-HCl, at pH 8.0. 10X PCR buffer: 166 mM (NH4)2SO4, 670 mM Tris-HCl, pH 8.8, 67 mM MgCl2, 100 mM -mercaptoethanol. 5 U/µL of Taq DNA polymerase (Life Technologies Inc., cat. no. 10342-020). 20 mg/mL of glycogen. 10 mM dNTP. Dimethyl sulfoxide (DMSO). 10 M NH4OAc. 2 M NaClO4. 100% Ethanol. 70% Ethanol. 100% Isopropanol. TE: 3 mM Tris-HCl, pH 7.5, 0.2 mM EDTA. 40% Acrylamide (19:1 acrylamide:bis). 1 kb plus DNA ladder (Life Technologies Inc., cat. no. 10787-018). 20-bp DNA ladder (GenSura, cat. no. SLL-101). SYBR green I (Molecular Products, cat. no. S-7567). 10X TAE. 10% ammonium persulfate. N,N,N',N'-Tetramethylethylenediamine (TEMED).
166
Dai
22. Spin-X (Costar, cat. no. 8161). 23. PCR thermal cycler.
3. Methods 3.1. mRNA Preparation and cDNA Synthesis 1. Prepare total RNA from culture cells or tissues using Trizol Reagent or guanidium isothiocynate/CsCl gradient method. It is recommended to obtain approx 1 mg total RNA for poly(A) RNA extraction (see Notes 4 and 5). 2. Determine RNA quality by agarose/formaldehyde gel electrophoresis. The general formaldehyde gel procedures are provided elsewhere (8). Stain the gel with ethidium bromide and examine the gel on a UV transilluminator. Intact RNA should show distinct 28S and 18S ribosomal RNA bands at 4.7 and 1.9 kb, respectively (see Note 6). 3. Prepare poly(A) RNA using the Messagemaker kit according to the manufacturer’s protocol. 4. Use 5 µg of poly(A) RNA to make the double-stranded cDNA. A cDNA synthesis kit such as Superscript Choice System (from Life Technologies) should be used (see Note 7). PAGE-purified 5'-biotinylated oligo-dT18 is used to substitute the regular oligo-dT. Follow the manufacturer’s instruction to make the first and second strand of cDNA. 5. Following the cDNA synthesis, increase the sample volume to 200 µL by adding H2O. Extract samples twice with 200 µL of PC8 by vortex and centrifuging. Collect the aqueous phase to a fresh tube, and add 100 µL of 10 M NH4OAc, 3 µL of glycogen (20 mg/mL), and 700 µL of 100% ethanol. Mix by vortex and centrifuge in a microfuge at 14,000g for 15 min. Carefully decant the supernatant, and wash the DNA pellet twice with 70% ethanol. Air-dry the pellet. Resuspend the purified cDNA in 20 µL of TE. 6. To check the quality of cDNA, run 1 µL of the above cDNA on a 1.0% agarose gel using 1X TAE. A high-quality cDNA should yield a smear ranging from 500 bp to 8–10 kb (see Note 6).
3.2. Preparation of cDNA Tags and Ditags 1. Prior to proceeding to this step, the linkers 1B and 2B need to be phosphorylated and annealed to linkers 1A and 2A, respectively. Set up the kinase reaction as follows, mix, and incubate at 37°C for 30 min (see Note 8): 350 ng/µL of linker 1B or 2B 9 µL H2O 6 µL 1X T4 kinase buffer 2 µL 10 mM ATP 2 µL 10 U/µL of T4 polynucleotide kinase 1 µL Total 20 µL
Serial Analyses of Gene Expression
167
Inactivate the T4 kinase by incubating the reactions at 65oC for 10 min. 2. To anneal linkers, add 9 µL of linker 1A or 2A (both 350 ng/µL) to the above 20 µL of linker 1B or 2B kinasing reaction, respectively. Incubate at 95°C for 2 min, and slowly cool down to room temperature over the period of at least 1 h. The easiest way is to use a large-beaker water bath with constant stirring. Prior to use, the quality of annealed linkers needs to be analyzed for ligation efficiency. To test this, set up the following test and control reactions and incubate at 15°C for 1 h (see Note 8): 1A/1B 1A/1B 2A/2B 2A/2B test control test control 200 ng/µL of annealed linker 1A/1B 200 ng/µL of annealed linker 2A/2B H2O 5X T4 DNA ligase buffer 1 U/µL of T4 DNA ligase Total
1 µL
1 µL
6 µL 2 µL 1 µL 10 µL
7 µL 2 µL 0 µL 10 µL
1 µL 6 µL 2 µL 1 µL 10 µL
1 µL 7 µL 2 µL 0 µL 10 µL
After the incubation, run the ligation control and test reactions on a 12% polyacrylamide gel. Use a 20-bp DNA ladder as a marker. Set up the 12% polyacrylamide gel (40 mL) by adding the following components: H2O 23.5 mL 10X TAE 4.0 mL 40% acrylamide 12 mL 10% ammonium persulfate 350 µL TEMED 30 µL The gel is run at 150 V using 1X TAE until the bromophenol blue dye reaches the bottom. The gel is stained in SYBR green I (1:10,000 dilution) for 30 min and visualized by UV. Expected results: In the test reaction, the majority (>80%) of the linkers should form dimers (approx 80 bp). In the control reaction, only the linker monomers are present at the approx 40-bp region. The annealed linker can be used for the following ligation reaction only if the linker dimers exceed 70% of the total linker input. Store the annealed linkers at –20°C until use. 3. Cut the cDNA from step 5 of Subheading 3.1. with an anchoring enzyme (NlaIII) (see Note 3). Set up a 200 µL of NlaIII digestion by mixing the following components in order, and incubate at 37°C for 1 h (see Note 8): cDNA 10 µL H2O 163 µL 100 X Bovine serum albumin (BSA) 2 µL Buffer 4 (10X) 20 µL 10 U/µL of NlaIII 5 µL Total 200 µL After the digestion, perform PC8 extraction and DNA precipitation as described in step 5 of Subheading 3.1. Resuspend the digested cDNA in 20 µL of TE.
168
Dai
4. Aliquot 100 µL of Dynabead M-280 Streptavidin Beads each into two 1.5-mL microfuge tubes. Use a magnet stand to separate the bead from buffer. Wash once with 200 µL of 1X B+W buffer. Then add 100 µL of 2X B+W buffer and 90 µL of H2O to each tube. Split the NlaIII-digested cDNA into the two tubes (10 µL each). Incubate at room temperature for 15 min with gentle mixing every 5 min. Capture the bead by the magnet stand, and wash the beads three times with 200 µL of 1X B+W buffer and once with 200 µL of TE. 5. For the two tubes, set up linkers 1A/1B and 2A/2B ligation reactions as follows:
Washed cDNA-coupled beads H2O Annealed linker 1A/1B Annealed linker 2A/2B 5X ligase buffer Total
Tube 1
Tube 2
Yes 25 µL 5 µL
Yes 25 µL
8 µL 38 µL
5 µL 8 µL 38 µL
Incubate the tubes at 50°C for 2 min, and then at room temperature for 15 min. Add 2 µL of T4 DNA ligase (5 U/µL) to each reaction, and incubate at 15°C for 2 h. After the ligation, wash the beads three times with 200 µL of 1X B+W buffer. Transfer the beads to a clean tube. Then wash once with 200 µL of 1X B+W buffer and then twice with 200 µL of 1X buffer 4 (see Note 9). 6. To release the cDNA tags, the above cDNA-bound beads will be subject to tagging enzyme (BsmFI) digestion. For each tube, set up the digestion by mixing the following components in sequence (see Note 8), and incubate at 65°C for 1 h with gentle tapping every 20 min. H2O 88 µL 10X buffer 4 10 µL 100X BSA 1 µL 4 U/µL of BsmFI 1 µL Total 100 µL Following the digestion, use the magnet stand to capture the beads. Save the supernatant in another clean microfuge tube. Increase the volume to 200 µL by adding H2O, and extract DNA with PC8. In the 200-µL sample, add 3 µL of glycogen, 100 µL of 10 M NH4OAc, and 900 µL of 100% ethanol. Incubate at 80°C for 30 min. Centrifuge at 14,000g at 4°C for 30 min. Wash twice with 70% ethanol. Resuspend DNA pellet with 10 µL of TE. 7. Blunt-end released cDNA tags by Klenow filling-in. For each of the above tubes, set up the fill-in reaction by adding the following components in sequence (see Note 8). Incubate at room temperature for 20 min. cDNA tags 10 µL H2O 28.5 µL 10X buffer 2 5 µL 100X BSA 1 µL
Serial Analyses of Gene Expression
169
10 mM dNTP 2.5 µL 1 U/µL of Klenow 3 µL Total 50 µL Carry out the PC8 extraction and ethanol precipitation as described in step 6 of Subheading 3.2. Resuspend DNA in 6 µL of TE. 8. To generate ditags, ligate the blunt-ended products from above by mixing the following components in 0.2-mL PCR tubes (see Note 8), and incubate at 15°C for 16 h. Also set up a control reaction (one without ligase) to examine the PCR specificity in step 9. Test Control DNA from tube 1 2 µL 2 µL DNA from tube 2 2 µL 2 µL 5X T4 DNA ligase buffer 1.2 µL 1.2 µL 5 U/µL of T4 ligase 0.8 µL H2O 0.8 µL Total 6 µL 6 µL Following the ligation, take out 1 µL of the ligation reaction to make 1:20, 1:50, 1:200, and 1:400 dilutions. 9. Set up PCR reactions to amplify the ditags by mixing the following components. Use the above serial dilutions of ligation products to optimize the DNA template input. Test Control H2O 22.5 µL 22.5 µL 10X PCR buffer 4 µL 4 µL 10 mM dNTP 7.5 µL 7.5 µL 350 ng/µL of primer 1 1 µL 1 µL 350 ng/µL of primer 2 1 µL 1 µL DMSO 3 µL 3 µL Diluted ditag with ligase 1 µL 0 µL Diluted ditag without ligase 0 µL 1 µL Total 40 µL 40 µL Heat the reaction at 95°C for 2 min, and hold at 78°C. Dilute Taq polymerse in 1X PCR buffer by mixing (see Note 8): H2O 8 µL 10X PCR buffer 1 µL 5 U/µL of Taq DNA polymerase 1 µL Total 10 µL When the thermal cycler block is on hold at 78°C, quickly add 10 µL of diluted Taq into PCR tubes, and mix by pipetting. Proceed PCR cycles: 26 cycles of 94°C 30 s, 55°C 30 s, 72°C 30 s, and 1 cycle of 72°C 10 min. The PCR condition varies from one machine from the other. You may adjust the parameters to yield the optimal result. Run 10 µL of PCR products on a 12% polyacrylamide gel and stain the
170
10.
11.
12.
13.
Dai
gel as described in step 2 of Subheading 3.2. A major band is expected at approx 102-bp that represents the amplification of linker-ditag. In addition, a band (approx 80 bp) that reflects linker amplification is usually of equal intensity as the 102bp band. No 102-bp product is expected for the ligase-negative control reaction. Use the DNA dilution that gives the highest 102-bp signal for large-scale PCR amplification. Set up large-scale PCR reactions (200 of 50 µL) for each library, using the optimized condition. Pool the PCR products into 30 1.5-mL microfuge tubes (approx 300 µL each). Extract with PC8. DNA is then ethanol-precipitated by adding 150 µL of 10 M NH4OAc, 3 µL of glycogen, and 900 µL of 100% ethanol to each tube. Centrifuge at 14,000g for 15 min. Wash twice with 70% ethanol. Resuspend PCR products in 10 µL of TE for each tube. A total of 300 µL of DNA will be collected. Run PCR products on a 12% polyacrylamide gel. Add 75 µL of DNA loading buffer to the pooled PCR products. Load 12.5 µL of sample on each lane (total of 10 lanes). Run gel at 150 V until dye reaches the bottom of the gel. After staining, excise the bands at approx 102 bp. Put three slices of gel into every 0.5-mL microfuge tube that has been pierced at the bottom. Put each 0.5-mL tube into a 1.5-mL tube. Centrifuge for 2 min or until all gel slices are meshed and deposited in the 1.5-mL tubes. To each tube, add 250 µL of H2O and 40 µL of 10 M NH4OAc, and incubate at 65°C for 1 h. Separate the supernatant from gel pieces by spinning SpinX columns for 2 min at 14,000g. Transfer the elutes to another clean tube, and add 3 µL of glycogen, 100 µL of NH4OAc, and 900 µL of 100% ethanol. Centrifuge at 14,000g for 15 min. Wash the DNA pellet twice with 70% ethanol. Air-dry and resuspend the ditag DNA in 10 µL of TE. A total of 100 µL of DNA will be obtained. Purified PCR products will be digested with NlaIII to release the ditags (see Notes 3 and 8). Set up the following digestion, and incubate at 37°C for 1 h. PCR products 100 µL H2O 216 µL 100X BSA 4 µL 10X Buffer 4 40 µL 10 U/µL NlaIII 40 µL Total 400 µL Following the incubation, the digestion mixture will be subject to PC8 extraction and ethanol precipitation as described in step 6 of Subheading 3.2. Resuspend DNA in 40 µL of TE. Run the digested products on a 12% polyacrylamide gel at 150 V for 2 h (see Note 10). Add DNA loading buffer, and load 12.5 µL of DNA in each lane (four lanes total). Stain the gel as described in step 2 of Subheading 3.2. Cut out the bands at the 24- to 26-bp area. Add two gel slices to each 0.5-mL microfuge tube that have been pierced at the bottom (two tubes total). Add each 0.5-mL tube to a 1.5-mL tube. Centrifuge until gel slices enter the 1.5-mL tubes. Add 250 µL of H2O and 40 µL of 10 M NH4OAc, and incubate at 37°C for 2 h. Separate the supernatant from gel pieces by Spin-X columns. For every tube, add 80 µL of 10 M NH4OAc, 3 µL of glycogen, and 1 mL of 100% ethanol. Mix and chill in 80°C freezer for
Serial Analyses of Gene Expression
171
30 min, and centrifuge at 14,000g at 4°C for 30 min. Wash the DNA pellet twice with 70% ethanol. Resuspend the ditag DNA in a total of 7 µL of TE. 14. Calculate the ditag concentration by ethidium bromide dot quantitation. Prepare DNA standard by serial diluting to 0, 1, 2.5, 5, 7.5, 10, and 20 ng/µL. Make sample dilution in 1:5, 1:25, and 1:125. On a piece of Saran wrap, dot and mix 4 µL of 1 µg/mL of ethidium bromide and 4 µL of DNA standards or 4 µL of diluted DNA sample. Photograph under UV and estimate DNA concentration by intensity comparison. A total of a few hundred nanograms of ditags is expected.
3.3. Concatemerization of Ditags and Automated Sequencing 1. In a microfuge tube, mix the following and incubate at 16oC for 1–3 h (see Notes 8 and 10). Pooled purified ditags 7 µL 5X ligase buffer 2 µL 5 U/µL of T4 DNA ligase 1 µL Total 10 µL Add 2.5 µL of DNA loading buffer to the above reaction, heat at 65°C for 5 min, and place on ice for 5 min. 2. Run a 1.0% agarose gel (approx 15 cm in length) using 1X TAE buffer. Load the entire 12.5-µL sample in one lane in parallel with a 1-kb DNA ladder. Run the gel at 100 V for 2 h. Stain the gel with SYBR Green I for 30 min. 3. Excise two regions of 0.6–1.2 kb and 1.2–2.5 kb. Extract DNA from agarose using the Qiagen gel extraction kit following the manufacturer’s protocol. Use 200 µL of TE as a final elute. Collect elutes in a microfuge tube, add 3 µL of glycogen, 100 µL of 7.5 M NH4OAc, and 700 µL of 100% ethanol. Centrifuge in a microfuge at 14,000g for 15 min. Wash the pellet twice with 70% ethanol. Resuspend the purified ditag contatemer in 6 µL of TE. 4. To clone the concatemer to pZErO-1, pZErO-1 will first be digested with SphI and purified by PC8 extraction and ethanol precipitation. Set up the following reaction and incubate at 16°C for 16 h (see Note 8). 25 ng/µL of SphI-digested pZErO-1 1 µL Purified contatemer 6 µL 5X ligase buffer 2 µL 5 U/µL of T4 DNA ligase 1 µL Total 10 µL Add 190 µL of H2O to the reaction, and extract with 200 µL of PC8. Precipitate and wash DNA as described in step 5 of Subheading 3.1. Resupend the ligation sample with 10 µL of TE. 5. Use 1 µL of purified ligation reaction to perform E. coli transformation of ElectroMAX DH10B cells by electroporation following the manufacturer’s instruction. Plate the bacterial in ten 100-mm agar plates that contain Zeocin (50 µg/mL). After 16 h of incubation at 37°C, examine the colonies on plates. The control plates should have no colony while the test plates should contain hundreds.
172
Dai
6. Perform PCR to check the insert sizes. Add the following components in a 0.2mL PCR tube. For large screening, 96-well plates may be used. H2O 14.5 µL 10X PCR buffer 2 µL 10 mM dNTP 1.25 µL 350 ng/µL of M13F 0.5 µL 350 ng/µL of M13R 0.5 µL DMSO 1.25 µL Total 20 µL Use a sterile tip to gently touch colony and dip the tip into PCR mixture. Mix by pipetting. Heat the reaction at 95°C for 2 min, and hold at 78°C. Dilute Taq polymerase in 1X PCR buffer by mixing (see Note 8): H2O 4.3 µL 10X PCR buffer 0.5 µL 5 U/µL of Taq DNA polymerase 0.2 µL Total 5 µL Add 5 µL of diluted Taq into each PCR tube when the block is on hold at 78°C. Mix by pipetting and proceed to 25 cycles of 94°C for 30 s, 55°C for 30 s, and 72°C for 60 s, followed by one cycle of 72°C for 10 min. Run 5 µL of the reaction on a 1.0% agarose gel. 7. Pick the clones that run above 700 bp. This will ensure that at least 16 ditags is contained in the concatemer. For the positive clones, transfer approx 20 µL of the remaining PCR reactions to a fresh 96-well plates. To each well, add a cocktail of 28 µL of H2O, 15 µL 2 M NaClO4, and 33 µL of 100% isopropanol using multichannel pipetman. Centrifuge at 6000g at 4°C for 15 min using a plate adaptor. Carefully discard the supernatant. Gently tap the plate up-side-down on paper towels. Add 100 µL of 70% ethanol to each well. Centrifuge for 5 min. Tap to eliminate the residual ethanol. Air-dry the DNA pellet. Resuspend pellets in 25 µL H2O, and store in –20°C until sequencing. 8. Use 5 µL of the above PCR products and an ABI-21M13 Dye Primer FS sequencing kit to set up automated sequencing reactions, following the manufacturer’s protocol. Load the samples on ABI373 automated sequencer. The SAGE data will be extracted and analyzed by the SAGE software. The SAGE software can be downloaded from http://www.sagenet.org (see Notes 11 and 12).
4. Notes 1. Use only high-quality molecular biology grade reagents. Reagents used in the protocol are either purchased from manufacturers or are prepared with DEPCtreated H2O (for total RNA harvest, poly(A) RNA preparation, and cDNA synthesis) or distilled H2O (for all enzymatic reactions). 2. Linkers that are of full-length and high quality are crucial to SAGE. Biotinylated oligo-dT18 and linkers 1A, 1B, 2A, and 2B need to be ordered as PAGE-purified primers.
Serial Analyses of Gene Expression
173
3. NlaIII is a very labile enzyme. Store at −20°C for only short time. For longer storage, keep it at −80°C. 4. Use of silicon-treated microfuge tubes is highly recommended. 5. Trizol works well for cultured cells, but guanidium isothiocynate is required for tissue samples to obtain clean RNA preparation. 6. The RNA quality directly determines the quality of SAGE library. Before preparing for poly(A) RNA, the quality of total RNA needs to be determined by agarose/ formaldehyde gel electrophoresis. The quality of double-stranded cDNA also needs to be determined by agarose gel electrophoresis. 7. Five micrograms of polyA RNA is generally required to make a high-quality SAGE library. For reasons that are still not clear, lower amount of RNA leads to poor yield of the 102-bp PCR product. For experiments in which only smaller amount of tissue is available, refer to modifications reported earlier (9–11) and the MicroSAGE protocol (through http://www.sagenet.org). 8. The enzyme activity from manufacturers may vary from lot to lot. Calculate the units required for every reaction and adjust the volume with H2O accordingly. 9. To prevent excessive linker amplification (80-bp bands) in step 9 of Subheading 3.2., the repeated washing procedures in step 5 of Subheading 3.2. is critical. Transferring the beads to another clean tube seems to significantly decrease residual linker molecules retained on tubes. 10. The success of ditag concatemerization depends largely on the purity and quantity of ditag. The procedures described in steps 12 and 13 of Subheading 3.2. normally produce a total of a few hundred nanograms of ditag that is generally needed for an efficient ligation. Also, the linker contamination in the ditag DNA prevents the concatemerization from proceeding. Thus the complete separation of linkers and ditags by electrophoresis is critical for the subsequent ligation process. Two hours of ligation at 16°C is generally sufficient for formation of long concatemers. The optimal incubation time may vary for different experiments. 11. High-quality sequencing is crucial for the SAGE data. Sequencing errors create contaminated tag counts. SAGE software has been designed to eliminate some of these errors. However, sequencing error is still a major source of nonconfirmable tags. 12. A few useful websites are available for useful SAGE information: http://www. sagenet.org, http://www.ncbi.nlm.nih.gov/SAGE, and http://www.labonweb.com.
References 1. 1 Venter, J. C., Adams, M. D., Myers, E. W., et al. (2001) The sequence of the human genome. Science 291, 1304–1351. 2. 2 Velculescu, V. E., Zhang, L., Vogelstein, B., and Kinzler, K. W. (1995) Serial analysis of gene expression. Science 270, 484–487. 3. 3 Polyak, K., Xia, Y., Zweier, J. L., Kinzler, K. W., and Vogelstein, B. (1997) A model of p53-induced apoptosis. Nature 389, 300–305. 4. 4 Zhang, L., Zhou, W., Velculescu, V. E., et al. (1997) Gene expression profiles in normal and cancer cells. Science 276, 1268–1272.
174
Dai
5. 5 Velculescu, V. E., Madden, S. L., Zhang, L., et al. (1999) Analysis of human transcriptomes. Nat. Genet. 23, 387–388. 6. He, T.-C., Sparks, A. B., Rago, C., et al. (1998) Identification of c-MYC as a 6 target of the APC pathway. Science 281, 1509–1512. 7. 7 Yu, J., Zhang, L., Hwang, P. M., Kinzler, K. W., and Vogelstein, B. (2001) PUMA induces the rapid apoptosis of colorectal cancer cells. Mol. Cell 7, 673–682. 8. Ausubel, F. M., Brent, R., Kingston, R. E., et al. (2000) Current Protocols in Molecular Biology. New York: John Wiley & Sons. 9. Datson, N. A., van der Perk-de Jong, J., van den Berg, M. P., de Kloet, E. R., and 9 Vreugdenhil, E. (1999) MicroSAGE: a modified procedure for serial analysis of gene expression in limited amounts of tissue. Nucl. Acids Res. 27, 1300–1307. 10. Virlon, B., Cheval, L., Buhler, J. M., Billon, E., Doucet, A., and Elalouf, J. M. 10 (1999) Serial microanalysis of renal transcriptomes. Proc. Natl. Acad. Sci. USA 96, 15286–15291. 11. St. Croix, B., Rago, C., Velculescu, V., et al. (2000) Genes expressed in human tumor endothelium. Science 289, 1197–1202.
Gene Profiling of Pancreatic Adenocarcinoma
175
11 Oligonucleotide-Directed Microarray Gene Profiling of Pancreatic Adenocarcinoma David E. Misek, Rork Kuick, Samir M. Hanash, and Craig D. Logsdon Summary Successful gene profiling studies involve careful experimental design, use of sensitive and accurate technologies, and statistically valid analysis of experimental results. In this chapter we describe our approach to the profiling of pancreatic adenocarcinoma to illustrate the various steps and methods involved in this type of study. Pancreatic adenocarcinoma is a particularly challenging subject for gene profiling, as these tumors have a profound desmoplastic response such that neoplastic epithelium makes up only a small proportion of the tissue mass. We have utilized statistical comparisons of gene expression between adenocarcinoma, normal pancreas, samples of chronic pancreatitis, and pancreatic cancer cell lines that provides a means to deduct the influence of the stromal elements. We utilized oligonucleotide-directed gene chips (Affymetrix), as they allow the simultaneous interrogation of thousands of genes in an efficient, reproducible, sensitive, and highly quantitative manner. The details of the approach we utilized are reported here, including information on experimental design, sample collection, expression level measurements, and data analysis for gene profiling. Key Words: Gene profiling; microassay; desmoplasia; stroma; biomarker; gene chip; chronic pancreatitis.
1. Introduction Successful gene profiling studies involve careful experimental design, use of sensitive and accurate technologies, and statistically valid analysis of experimental results. In this chapter we describe our approach to the profiling of pancreatic adenocarcinoma to illustrate the various steps and methods involved in this type of study. Pancreatic adenocarcinoma is a particularly challenging subject for gene profiling, as these tumors have a profound desmoplastic response such that neoplastic epithelium makes up only a small proportion of the tissue mass. Thus, while there have been a number of gene profiling studies in this From: Methods in Molecular Medicine, Vol. 103: Pancreatic Cancer: Methods and Protocols Edited by: G. Su © Humana Press Inc., Totowa, NJ
175
176
Misek et al.
disease (1–5), the genes thus far identified have not all been localized to the neoplastic epithelium. We have recently described studies in which we have utilized statistical comparisons of gene expression between adenocarcinoma, normal pancreas, samples of chronic pancreatitis, and pancreatic cancer cell lines that provides a means to deduct the influence of the stromal elements (6). We utilized oligonucleotide-directed gene chips (Affymetrix), as they allow the simultaneous interrogation of thousands of genes in an efficient, reproducible, sensitive, and highly quantitative manner. The details of the approach we utilized are reported here. Thus, information is provided on experimental design, sample collection, expression level measurements, and data analysis for gene profiling. 2. Materials 2.1. Surgical Specimens and Pancreatic Tumor Cell Lines 1. Patient-derived pancreatic tumors, uninvolved pancreas, and chronic pancreatitis specimens that were surgically removed, flash frozen in liquid nitrogen, then stored at -80°C until use. 2. Pancreatic tumor-derived cell lines were obtained from the American Type Culture Collection (ATCC; Manassas, VA) and propagated according to instructions from the ATCC.
2.2. Isolation of Total RNA 1. 2. 3. 4. 5. 6. 7.
OCT freezing media (Miles Scientific). TRIzol reagent (Invitrogen). Chloroform (Sigma). Isopropanol (Sigma). Glycogen (Ambion). Acid phenol–chloroform (5:1) (Ambion). RNeasy Mini Kit (Qiagen).
2.3. Preparation of cDNA 1. T7-(dT)24 Primer HPLC purified DNA (GENSET Corp.) 5'-GGCCAGTGAATTG TAATACGACTCACTATAGGGAGGCGG-(dT)24-3'. 2. Superscript II with DTT and 5X first-strand buffer (Invitrogen). 3. 10 mM dNTP mix (Invitrogen). 4. Diethyl pyrocarbonate (DEPC)-treated water (Ambion). 5. 5X second-strand buffer (Invitrogen). 6. E. coli DNA ligase (Invitrogen). 7. E. coli DNA polymerase I (Invitrogen). 8. E. coli ribonuclease H (Invitrogen). 9. T4 DNA polymerase (Invitrogen). 10. 0.5 M EDTA (Sigma).
Gene Profiling of Pancreatic Adenocarcinoma 11. 12. 13. 14. 15.
177
2-mL Phase Lock Gel Heavy (Eppendorf). Buffered phenol–chloroform–isoamyl alcohol (Ambion). 7.5 M ammonium acetate (Sigma). Absolute ethanol (stored at -20°C). 80% Ethanol (stored at -20°C).
2.4. Synthesis of Biotin-Labeled cRNA, Cleanup, and Quantitation 1. Enzo BioArray HighYield RNA Transcript Labeling Kit (Affymetrix). 2. RNeasy Mini Kit (Qiagen). 3. DEPC-treated water (Ambion).
2.5. cRNA Fragmentation 1. 2. 3. 4. 5.
Trizma base (Sigma). Magnesium acetate (Sigma). Potassium acetate (Sigma). Glacial acetic acid (Sigma). DEPC-treated water (Ambion).
2.6. Preparation of Hybridization Cocktail and Hybridization of Probe Array 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12.
Nuclease-free water (Ambion). Acetylated bovine serum albumin (BSA; 50 mg/mL; Invitrogen). Herring sperm DNA (Promega Corp.). 20X eukaryotic hybridization controls (Affymetrix). Control oligo B2 (Affymetrix), high-performance liquid chromatography (HPLC)purified, 5'-bioGTCGTCAAGATGCTACCGTTCAGGA-3'. 5 M NaCl, RNase-free, DNase-free (Ambion). 2-Morpholino ethanesulfonic acid (MES) free acid monohydrate (Sigma). MES sodium salt (Sigma). 0.5 M EDTA (Sigma). Tough Spots Label Dots (USA Scientific). Surfact-Amps 20 (10% Tween-20) (Pierce Chemical). Thin wall orifice pipet tips, 1–250 µL (Dot Scientific).
2.7. Washing, Staining, and Scanning of Probe Arrays 1. 2. 3. 4. 5. 6. 7. 8. 9.
Nuclease-free water (Ambion). Acetylated BSA (50 mg/mL; Invitrogen). Streptavidin, R-phycoerythrin (Molecular Probes). 5 M NaCl, RNase-free, DNase-free (Ambion). Phosphate-buffered saline, Ca2+- and Mg2+-free (Invitrogen). 20X SSPE (Ambion). Goat IgG, reagent grade (Sigma). Anti-streptavidin antibody (goat), biotinylated (Vector Laboratories). Antifoam O-30 (Sigma).
178
Misek et al.
10. Surfact-Amps 20 (10% Tween-20) (Pierce Chemical). 11. Thin wall orifice pipet tips, 1–250 µL (Dot Scientific). 12. Amber-colored microfuge tubes (Life Science Products).
3. Methods 3.1. Experimental Design Successful interpretation of microarray data requires a careful consideration of the data being analyzed. Differences in the levels of specific mRNAs in different tissue samples may be caused by variations in cellular composition as well as changes in gene expression levels. This is an especially critical consideration in the case of pancreatic cancer. Pancreatic tumors consist of neoplastic epithelial cell islands embedded in a profound stroma consisting of connective tissue, vascular elements, and inflammatory cells. In contrast, normal pancreas consists primarily of pancreatic acinar cells. Thus, direct comparisons between normal pancreas and pancreatic adenocarcinoma are likely to identify large numbers of genes that do not originate within neoplastic epithelium. To overcome this inherent obstacle, we utilized samples of chronic pancreatitis. Chronic pancreatitis is an inflammatory disease that results in the replacement of normal pancreatic parenchyma with stromal elements (7). Morphologically the stroma of pancreatic tumors and of focal regions of chronic pancreatitis is identical. Thus, comparison of samples from pancreatic adenocarcinomas with samples of chronic pancreatitis is useful to deduct the influence of the stromal elements. To further increase the specificity of the selected genes to those expressed in neoplastic epithelium, we used comparisons between pancreatic cancer cell lines and normal pancreas. Direct comparisons of pancreatic cancer cell lines with normal pancreas are not very useful as the cell lines have numerous alterations in gene expression caused by adaptation to in vitro growth. However, expression data from the cell lines are useful as genes more highly expressed in both pancreatic cancer cell lines and pancreatic tumors than in normal pancreas are unlikely to originate from stroma. 3.2. Preparation of Total RNA 1. All tumors, uninvolved pancreas, and chronic pancreatitis samples were processed in a similar fashion. Frozen tumor samples were embedded in OCT freezing media, cryotome sectioned (5 µM), and evaluated by routine hematoxylin and eosin (H&E) stains by a surgical pathologist. Areas of relatively pure tumor, chronic pancreatitis, or normal pancreas were microdissected (see Note 1) and processed for RNA isolation. Cells grown in culture were lysed by direct addition of TRIzol to the culture dish, then scraped and homogenized further by repetitive pipetting, as per the manufacturer’s instructions.
Gene Profiling of Pancreatic Adenocarcinoma
179
2. Single isolates of frozen microdissected tumor samples were homogenized in the presence of TRIzol and allowed to incubate at room temperature for 5 min to promote the complete dissociation of ribonucleoprotein complexes. 3. The TRIzol homogenates were centrifuged at 13,000g for 10 min at 4°C to pellet out all insoluble material (see Note 2). The cleared homogenates were transferred to a clean microfuge tube. To each sample 0.2 mL of chloroform per milliliter of TRIzol was added. The samples were vortex-mixed, then centrifuged at 13,000g for 15 min at 4°C to promote phase separation. The RNA-containing clear upper aqueous phase was carefully transferred (while avoiding the interphase) to a clean microfuge tube (usually about 0.6 mL). 4. The total RNA was precipitated by addition of 1 µL of glycogen (5 mg/mL) and 0.5 mL of isopropanol per 1 milliliter of TRIzol in the initial homogenization. The samples were centrifuged at 13,000g for 10 min at 4°C. The supernatant was discarded and the remaining pellet was washed once in 80% ice-cold ethanol, and then air-dried. The pellet was resuspended in 100 µL of DEPC-treated H2O. 5. One hundred microliters of acid phenol-chloroform (5:1) was added to each sample (see Note 3). The samples were vortexed, then centrifuged at 13,000g to promote phase separation. The RNA-containing upper aqueous phase was transferred to a clean microfuge tube and then further cleaned by passage through a RNeasy spin column, using all optional steps, as per the manufacturer’s instructions. 6. The RNA should be quantitated using A260/A280 ratios and its quality visually assessed by running an aliquot in an ethidium bromide stained 1% agarose gel (denaturing gels are usually not necessary). Samples that did not reveal intactness of the RNA and approximately equal 18S and 28S ribosomal bands were excluded from further study.
3.3. Preparation of cDNA Briefly stated, 5 µg of total RNA was converted into double-stranded cDNA by reverse transcription using an oligo (dT)24 primer containing a T7 RNA polymerase promoter site added 3' of the poly T. 1. Five micrograms of total RNA is suspended in 11 µL of DEPC-treated H2O, then mixed with 1 µL of T7- (dT)24 primer (100 pmol/µL). Place microfuge tube at 70°C (we find that a PCR machine set manually works best for this and all subsequent incubations) for 10 min, followed by a quick centrifugation and immediate incubation on ice. 2. Add 4 µL of 5X first-strand cDNA buffer, 2 µL of 0.1 M dithiohreitol (DTT) (final concentration of 10 mM), 1 µL of 10 mM dNTP mix (final concentration of 500 µM each), and 1 µL of SuperScript II RT (200 U/µL). Mix well, centrifuge briefly, then incubate the microfuge tube at 42°C in a dry-block heater for 1 h. 3. Place first-strand reactions on ice, and when cooled, centrifuge briefly to lower the condensation. Add to the first-strand synthesis 91 µL of DEPC-treated H2O, 30 µL of 5X second-strand reaction buffer, 3 µL of 10 mM dNTP mix (final concentration of 200 µM each), 1 µL of E. coli DNA Ligase (10 U/µL), 4 µL of E. coli
180
Misek et al.
DNA polymerase I (10 U/µL), and 1 µL of E. coli RNase H (2 U/µL). Gently tap the tube to mix the contents, briefly centrifuge, and incubate at 16°C in a dryblock heater for 2 h. 4. Add 2 µL of T4 DNA polymerase (10 U) to the tube, gently tap the tube to mix the contents, and incubate the tube at 16°C for 5 min. 5. Add 10 µL of 0.5 M EDTA to stop the reaction. Add 162 µL of buffered phenol– chloroform–isoamyl alcohol (25:24:1) to the reaction mixture and vortex-mix to form an emulsion. Pellet the Phase Lock Gel (1.5-mL tube) by centrifugation at 16,000g in a microfuge for 30 s. Transfer the entire cDNA–phenol–chloroform– isoamyl alcohol, then centrifuge at 16,000g for 2 min. Transfer the aqueous upper phase to a clean 1.5-mL microfuge tube. 6. The double-stranded cDNA was precipitated by addition of 1 µL of glycogen (5 mg/ mL), 82 µL of 7.5 M ammonium acetate, and 610 µL of ice-cold absolute ethanol, vortex-mixed, then pellet by centrifugation at 16,000g for 20 min at room temperature. The supernatant was discarded and the remaining pellet was washed once in 80% ice-cold ethanol and then air-dried. The pellet was resuspended in 12 µL of DEPC-treated H2O and then stored at -80°C until use (usually the following day).
3.4. Preparation of Biotin-Labeled cRNA Essentially the protocol followed is as supplied by the manufacturer with the Enzo BioArray HighYield RNA Transcript Labeling Kit. 1. The 12 µL of cDNA (prepared as described in Subheading 2.3.) was thawed on ice and mixed with the individual components of the labeling kit (4 µL of 10X HY reaction buffer, 4 µL of 10X biotin-labeled ribonucleotides, 4 µL of 10X DTT, 4 µL of 10X RNase inhibitor mix, and 2 µL of T7 RNA polymerase). The reagents were mixed by gently tapping on the tube, collected at the bottom of the tube by a brief centrifugation, then incubated at 37°C in a dry-block heater for 5 h. The reaction was gently mixed every 45 min by gently tapping on the tube. Usually 30–80 µg of transcribed RNA product is produced during the 5-h incubation, which is sufficient for many hybridization reactions (see below). 2. To remove unincorporated NTP’s from the transcribed RNA so that the transcribed product can be properly quantitated, the biotinylated RNA was cleaned by passage through a RNeasy spin column (the biotinylated RNA should not be phenol–chloroform extracted, as biotinylated RNA does not cleanly partition into the aqueous phase and substantial loss of RNA may result), using all optional steps, as per the manufacturer’s instructions. Subsequently, the cRNA should be quantitated using A260/A280 ratios.
3.5. Fragmentation of the cRNA for Preparation of the Target 1. Fifteen micrograms of biotinylated cRNA was placed in a clean microfuge tube, and the volume was adjusted to 32 µL with DEPC-treated H2O. Eight microliters of 5X fragmentation buffer (200 mM Tris-acetate, pH 8.1, 500 mM potassium acetate, 150 mM magnesium acetate) was added. The reagents were mixed by gently
Gene Profiling of Pancreatic Adenocarcinoma
181
tapping on the tube, collected at the bottom of the tube by a brief centrifugation, then incubated at 94°C in a dry-block heater for 35 min. After cooling the mixture on ice, the reagents were again collected at the bottom of the tube by brief centrifugation. The fragmented cRNA was stored at -20°C until further use.
3.6. Hybridization, Washing, Staining, and Scanning of Probe Arrays It is necessary to prepare a hybridization cocktail prior to hybridization of the Test3 array. We generally prepare sufficient cocktail (300 µL) to hybridize up to five arrays sequentially (The Human Genome U95 series, for example, were designed by Affymetrix as five arrays/complete set [comprising 60,000 known genes and EST’s total, 12,000 per array]). 1. 12X MES stock (0.33 M MES, free acid monohydrate and 0.89 M MES, sodium salt, pH 6.7) was prepared, passed through a 0.22-µm filter and stored at 4°C. 2X MES hybridization buffer (consisting of 8.3 mL of 12X MES stock, 17.7 mL of 5 M NaCl, 4.0 mL of 0.5 M EDTA, 0.1 mL of 10% Tween-20, and 19.9 mL of DEPCtreated H2O) was also prepared and stored at room temperature. 2. Wet the Test3 array by adding 100 mL of 1X MES hybridization buffer through the lower septa (insert a pipet tip through the upper septa so that it will function as an air vent), using a micropipet and thin wall oriface pipet tips. Incubate the Test3 array at 45°C for up to 1 h, with rotation at 60 rpm in a rotisserie hybridization oven. 3. Thaw the 15 µg of fragmented cRNA (in 40 µL prepared as above). Add 150 µL of 2X hybridization buffer, 84 µL of DEPC-treated H2O, 3 µL of acetylated BSA (50 mg/mL), 3 µL of herring sperm DNA (10 mg/mL), 5 µL of control oligo B2 (3 nM), and 15 µL of 20X eukaryotic hybridization controls (bioB, bioC, bioD, and cre [final concentration of 1.5, 5, 25, and 100 pM respectively]). 4. Incubate the hybridization cocktail at 94°C for 5 min in a dry-block heater (we find that a PCR machine set manually works best), followed by incubation at 45°C for 5 min. Vortex-mix the tubes briefly, then centrifuge the hybridization cocktails at 16,000g for 5 min at room temperature to pellet insoluble material from the mixture. 5. Remove the 1X MES hybridization buffer through the lower septa (insert a pipet tip through the upper septa so that it will function as an air vent), using a micropipet and thin wall orifice pipet tips. Bang the array firmly on the laboratory bench to drive the remaining 1X MES hybridization buffer to the bottom of the array, then remove the buffer with a clean pipet tip. Introduce 80 µL of the hybridization cocktail into the array through the lower septa (this volume will leave an air bubble inside the array chamber that will serve as a “stir bar” during the hybridization). Seal the two septa with Tough-spots so that there is no evaporation of buffer. Incubate the Test3 array at 45°C for approx 16 h, with rotation at 60 rpm in a rotisserie hybridization oven. 6. Stringent wash buffer (83.3 mL of 12X MES stock, 5.2 mL of 5 M NaCl, 1 mL of 10% Tween-20, and 910.5 mL of molecular biology grade water) was prepared, passed through a 0.22-µm filter, and stored at room temperature. Nonstringent wash buffer (300 mL of 20X SSPE, 1 mL of 10% Tween-20, and 698 mL of molecular
182
Misek et al.
biology grade water) was prepared and passed through a 0.22-µm filter. After filtration, 1 mL of 5% Antifoam O-30 was added (this will turn the solution slightly turbid) and the solution was stored at room temperature. 2X stain buffer (41.7 mL of 12X MES stock, 92.5 mL of 5 M NaCl, 2.5 mL of 10% Tween-20, and 113 mL of molecular biology grade water) was also prepared and passed through a 0.22µm filter. Following filtration, 0.5 mL of 5% Antifoam O-30 was added (this will turn the solution slightly turbid) and the solution was stored at room temperature. 7. It is necessary to prepare the stain reagents immediately prior to removal of the arrays from the hybridization oven. The proper staining of the probe arrays requires a series of three stains, with the first and third being identical. The first (and the third) stain was prepared by mixing 600 µL of 2X stain buffer, 540 µL of DEPCtreated water, and 12 µL of streptavidin, R-phycoerythrin (this is light labile, thus care must be exercised to prepare and keep this solution in the dark), in duplicate, for each array to be stained. We typically prepare several additional aliquots each time we stain arrays in case that they would be needed. This solution is passed through a 0.22-µm syringe filter and 1152-µL aliquots were distributed into ambercolored microfuge tubes. Forty-eight microliters of acetylated BSA is added to each aliquot. The tubes are vortexed, then centrifuged to remove insoluble material (precipitated streptavidin, R-phycoerythrin, and albumin) from the solution. The top 600 µL from each aliquot was transferred to a fresh amber-colored microfuge tube), and was used in the staining reaction (the remainder was discarded). The second stain was prepared by mixing 600 µL of 2X stain buffer, 538 µL of DEPC-treated water, 48 µL of acetylated BSA, 12 µL of reagent grade goat IgG (10 mg/mL), and 7.2 µL of anti-streptavidin IgG (0.5 mg/mL) for each array to be stained. We typically prepare several additional aliquots in case they would be needed. The tubes were vortexed, then centrifuged to remove insoluble material (precipitated IgG and albumin) from the solution. The top 600 µL from each aliquot was transferred to a fresh microfuge tube (this solution is kept in clear microfuge tubes), and was used in the staining reaction (the remainder was discarded). 8. Following removal of the probe arrays from the hybridization oven it is necessary to retrieve the hybridization cocktail from the probe array (as it can be reused for subsequent hybridization reactions). Remove the hybridization cocktail through the lower septa (insert a pipet tip through the upper septa so that it will function as an air vent), using a micropipet and thin wall orifice pipet tips. Make sure that you place the cocktail back into the microfuge tube that contains the remainder of the aliquot. Bang the array firmly on the laboratory bench to drive the remaining hybridization cocktail to the bottom of the array, and then remove the cocktail with a clean pipet tip. Place this small quantity of cocktail back into the microfuge tube that contains the remainder of the aliquot as well. Introduce 100 µL of the nonstringent wash buffer into the array through the lower septa. The tube containing the hybridization cocktail should be stored at -20°C (or for longer term storage at -80°C). 9. Instructions on the proper use and care of the Affymetrix Fluidics Station (used for the washing and staining of the hybridized probe arrays) can be found in the Affymetrix GeneChip Expression Analysis Technical Manual (available from Affymetrix).
Gene Profiling of Pancreatic Adenocarcinoma
183
10. Following staining, the Test3 probe arrays are scanned using the GeneArray scanner from Affymetrix. The scanned image is provided as a .dat file. Once generated, the .dat file is analyzed by the Affymetrix (version 4.0 Microarray Suite) software, which generates a .chp file. Typically, we check information generated in the .chp file to determine sample quality assurance issues such as background, 5'-3' concordance ratios of b-actin and GAPDH (a ratio close to 1 suggests that all cRNA transcripts are close to full length; a ratio of more than 3 suggests that only a third of the transcripts are full length [the cutoff of a successful sample according to the Affymetrix Technical Manual]). 11. Once the quality of the sample has been determined to be good by hybridization on the Test3 array, the same sample that was hybridized to the Test3 array should be used for hybridization to the GeneChip Array. Wet the GeneChip Array by adding 250 mL of 1X MES hybridization buffer through the lower septa (insert a pipet tip through the upper septa so that it will function as an air vent), using a micropipet and thin wall orifice pipet tips. Incubate the GeneChip Array at 45°C for up to 1 h, with rotation at 60 rpm in a rotisserie hybridization oven. 12. Incubate the hybridization cocktail at 94°C for 5 min in a dry-block heater (we find that a PCR machine set manually works best), followed by incubation at 45°C for 5 min. Vortex-mix the tubes briefly, then centrifuge the hybridization cocktails at 16,000g for 5 min at room temperature to pellet insoluble material from the mixture. 13. Remove the 1X MES hybridization buffer through the lower septa (insert a pipet tip through the upper septa so that it will function as an air vent), using a micropipet and thin wall orifice pipet tips. Bang the array firmly on the laboratory bench to drive the remaining 1X MES hybridization buffer to the bottom of the array, then remove the buffer with a clean pipet tip. Introduce 200 µL of the hybridization cocktail into the array through the lower septa (this volume will leave an air bubble inside the array chamber that will serve as a “stir bar” during the hybridization). Seal the two septa with Tough-spots so that there is no evaporation of buffer. Incubate the GeneChip Array at 45°C for approx 16 h, with rotation at 60 rpm in a rotisserie hybridization oven. It is necessary to prepare the stain reagents immediately prior to removal of the arrays from the hybridization oven. This step is identical to that described in Subheading 3.6.7. 14. Following removal of the probe arrays from the hybridization oven it is necessary to retrieve the hybridization cocktail from the probe array (as it can be reused for subsequent hybridizations). Remove the hybridization cocktail through the lower septa (insert a pipet tip through the upper septa so that it will function as an air vent), using a micropipet and thin wall orifice pipet tips. Make sure that you place the cocktail back into the microfuge tube that contains the remainder of the aliquot. Bang the array firmly on the lab bench to drive the remaining hybridization cocktail to the bottom of the array, and then remove the cocktail with a clean pipet tip. Place this small quantity of cocktail back into the microfuge tube that contains the remainder of the aliquot as well. Introduce 250 µL of the nonstringent wash buffer into the array through the lower septa. The tube containing the hybridization cocktail should be stored at -20°C (or for longer term storage at -80°C).
184
Misek et al.
15. Instructions on the proper use and care of the Affymetrix Fluidics Station (used for the washing and staining of the hybridized probe arrays) can be found in the Affymetrix GeneChip Expression Analysis Technical Manual (available from Affymetrix). It should be noted that there are differences in the Fluidics protocols utilized between the washing and staining of Test3 arrays, and those utilized when washing and staining of actual GeneChip Arrays. 16. Following staining, the GeneChip probe arrays are scanned using the GeneArray scanner from Affymetrix. The scanned image (Fig. 1) is shown as a .dat file. Once generated, the .dat file is analyzed by the Affymetrix (version 4.0 Microarray Suite) software, which generates a .chp file. Typically, we do check information generated in the .chp file to determine sample quality assurance issues such as background, 5'-3' concordance ratios of b-actin and GAPDH (although it would be rare that these values would differ substantially from those observed on the Test3 array).
3.7. Reading and Statistical Analysis 1. After hybridization steps the chips are scanned at 3 µm per pixel resolution using the GeneArray scanner from Affymetrix. For each 24 ´ 24 µm feature, the Affymetrix (version 4.0 Microarray Suite) software ignores a one-pixel border, and the 75th percentile of the remaining pixels is stored in a file (.CEL files). There are typically 20 pairs of features (probe-pairs) on HuGeneFL chips for each transcript (probe-set), 20 of which are designed to be complementary to a specific sequence (perfect match = PM features), and the other 20 being identical except that the central base has been altered (mismatch = MM features). On the more recent chips produced by Affymetrix (human U95 series and U133 series, Mouse U74 series) there are more features, these occupying 20- or 18-µm squares, and there are fewer probe-pairs per probe-set on average (16 or 11 typically). We have developed software to read the .CEL files and their descriptions and perform some processing of the data (written in C). The code performs most of the calculations described below except for the quantile-normalization. Downloads and more documentation are available (http://dot.ped.med.umich.edu:2000/pub/index.html). The idea of a standard chip is used, which is usually selected to be a chip with high signals and low background. Probe-pairs for which either the PM or MM feature are saturated (pure white = large pixel values) in the image of the standard, or for which PM - MM < C on the standard are excluded from further consideration. These probe-pairs usually give highly negative PM - MM values on all chips. (C was usually chosen to be -1000 prior to a reduction of the voltage across the PMT of the scanner. For subsequent scans, -100 is a typical choice.) Saturated features (measures at least 98% of the maximum pixel value for the chip) on other chips are imputed separately for PM or MM values. For a saturated PM value the ratios of nonsaturating PM values for the chip divided by the standard are averaged for a probe-set by taking the antilogarithm of the mean of the log ratios. This factor is multiplied by the PM values of the standard to obtain imputed values for the chip under consideration. (The original values are replaced by the imputed values only if the imputed values are larger.) The MM values are imputed similarly.
Gene Profiling of Pancreatic Adenocarcinoma
185
Fig. 1. Example of a scanned image of an Affymetrix human FL gene chip hybridized with a sample prepared from human pancreatic adenocarcinoma. Each feature is approx 25 µm2. The Affymetrix reader quantitates the intensity of the fluorescence signal for each feature and uses the difference in intensities between perfect-match and mismatched oligonucleotide features to measure mRNA levels. 2. A one-sided Wilcoxon signed-rank test is performed on the PM - MM differences to help judge if the transcript represented by the probe-set is being detected. More recent Affymetrix software (Microarray Suite 5.0) also performs a signed-rank test for this purpose. 3. The average intensity for each probe-set is computed as the mean of the PM - MM differences, after trimming away the 25% highest and lowest differences. This is sometimes referred to as the trimmed-mean to distinguish it from the analogous “average difference” computed by the Affymetrix software. Normalization is usually by one of two methods. For method 1, we select a set of reference probe-sets
186
Misek et al.
that are used to normalize each chip by adjusting by a single scale factor. The reference set can be chosen by asking that a probe-set is infrequently among the most or least intense probe-sets among in the study, or by asking that a certain minimum percentage of the chips give small p-values for signed-rank test for “detected.” However the set is chosen, a normalization factor is obtained using the reference probe-sets by computing the antilogarithm of the mean log ratios of the trimmed means for the selected chip divided by the standard. We usually prefer a second method, for which detailed mathematics and code (written by Kerby Sheddon) is also available at the website above, and that is referred to as quantilenormalization. Using quantile-normalization the distribution of trimmed-means is adjusted to more nearly match that of a standard chip by making 100 (or 20 or 50, etc.) individual quantiles (or percentiles) have the same values, using a piecewise linear function. (The “standard chip” being normalized to can be a real chip’s trimmed-means, or some artificially computed standard such as the median value for each quantile over a set of chips.) The first and last intervals are normalized by fitting a regression line through these largest (or smallest) values for the two chips. This method is sometimes used after removing a set of 65 (varies with the chip) probe-sets that serve quality control purposes or give highly negative average intensity values. Alternatively, you can specify a bound such that genes whose rank range exceeds the bound will not be used to calculate the quantile adjustment. The procedure is performed by a separate C++ program, and thus can also be used for data obtained from other kinds of assays. 4. For comparisons of two chips, two-sided Wilcoxon signed-rank tests can be performed on the differences of the PM - MM values for each probe-set, after normalization with method 1. Probe-pairs for which either chip has saturated values for PM or MM are excluded, excepting cases where a PM value is saturated on the chip that gave a larger PM - MM value (even before imputing it), in which case the imputed PM - MM value is used. These tests are somewhat analogous to certain calls obtainable from the Affymetrix software. For larger experiments that include some replication of the samples in the design we usually fit models to the quantileadjusted (and log-transformed) trimmed-mean data rather than considering the probepair level data, since signed-rank tests at the probe-pair level treat the PM - MM values for probe-pairs on the same chip as independent assays, which is not true. We suggest using these signed-rank tests only on preliminary experiments to help decide if it is worth running additional samples.
3.8. Selection Strategy 1. Genes expressed specifically in neoplastic epithelium of pancreatic adenocarcinoma were identified as those whose expression in adenocarcinoma samples that was more than twofold the value of normal (to eliminate normal pancreatic genes) and chronic pancreatitis (to eliminate stromal genes) and also more than twofold higher in pancreatic cancer cell lines compared to normal pancreas (as a further criterion for neoplastic cell expression).
Gene Profiling of Pancreatic Adenocarcinoma
187
2. Genes expressed specifically in chronic pancreatitis were identified as those whose expression in chronic pancreatitis samples was more than twofold the value of either normal (to eliminate normal pancreatic genes) and adenocarcinoma (to eliminate stromal genes). 3. Genes expressed specifically in stromal elements were identified as those whose expression was more than twofold that of normal pancreas in both chronic pancreatitis and adenocarcinoma samples and whose expression in pancreatic cancer cell lines was less than twofold increased over normal pancreas.
4. Notes 1. Microdissection consisted of using a fresh razor blade to trim the frozen sample to the area identified by the pathologist as being highly enriched for either cancer or focal pancreatitis. These trimmed pieces were then reembedded in OCT to facilitate sectioning. After sampling a final section was cut and stained with H&E to verify that the sectioning did not pass out of the tissue area of interest. 2. We have found that some samples contain significant amounts of insoluble material, especially in source tissues such as pancreas and liver. Therefore, we have found it beneficial to incorporate this centrifugation step. Although it is not imperative that it be performed, it will reduce the amount of interphase found after phase separation and will result in a cleaner harvesting of the aqueous phase. We have also found that inclusion of this step will greatly reduce the background in lower quality samples. 3. We have found that inclusion of an acid phenol extraction will greatly reduce the background in lower quality samples. With the inclusion of the centrifugation step in Note 1 and the acid phenol extraction, we have been able to reprocess a number of samples successfully that have failed on the Test3 array.
References 1. 1 Argani, P., Rosty, C., Reiter, R. E., et al. (2001) Discovery of new markers of cancer through serial analysis of gene expression: Prostate stem cell antigen is overexpressed in pancreatic adenocarcinoma. Cancer Res. 61, 4320–4324. 2. 2 Iacobuzio-Donahue, C. A., Ryu, B., Hruban, R. H., and Kern, S. E. (2002) Exploring the host desmoplastic response to pancreatic carcinoma: Gene expression of stromal and neoplastic cells at the site of primary invasion. Am. J. Pathol. 160, 91–99. 3. 3 Zhang, L., Zhou, W., Velculescu, V. E., et al. (1997) Gene expression profiles in normal and cancer cells. Science 276, 1268–1272. 4. Zhou, W., Sokoll, L. J., Bruzek, D. J., et al. (1998) Identifying markers for pancreatic cancer by gene expression analysis. Cancer Epidemiol. Biomark. Prev. 7, 109–112. 5. 5 Gress, T. M., Muller-Pillasch, F., Geng, M., et al. (1996) A pancreatic cancer-specific expression profile. Oncogene 13, 1819–1830. 6. Logsdon, C. D., Simeone, D., Binkley, C., et al. (2003) Molecular profiling of pancreatic adenocarcinoma and chronic pancreatitis identifies multiple genes differentially regulated in pancreatic cancer. Cancer Res. 63, 2649–2657 [erratum in: Cancer Res. 63, 3445]. 7. Bruno, M. J. (2001) Current insights into the pathogenesis of acute and chronic pancreatitis. Scand. J. Gastroenterol. 234(Suppl), 103–108.
188
Misek et al.
Identification of Differentially Expressed Proteins
189
12 Identification of Differentially Expressed Proteins in Pancreatic Cancer Using a Global Proteomic Approach Christophe Rosty and Michael Goggins Summary Proteomics is the term used for the large-scale analysis of proteins in biological fluids or cells by biochemical methods. Two approaches are used for proteomics analysis: two-dimensional polyacrylamide gel electrophoresis (2D-PAGE) and a mass-spectrometry-based approach, such as surface-enhanced laser desorption ionization (SELDI). SELDI can be used for large protein profiling or peptide identification after enzymatic digestion. In pancreatic cancer, proteomics analysis can be performed with the aim to identify all differentially expressed proteins in cancer cells vs normal pancreatic cells. Protein profiling of pancreatic juice or serum may also identify biomarkers for pancreatic cancer that could be used as diagnostic markers or therapeutic targets. This chapter outlines the use of 2D-PAGE and SELDI for profiling the protein content of pancreatic juice samples and for identifying proteins differentially expressed in pancreatic cancer patient samples compared to control patient samples. Key Words: Proteomics; two-dimensional polyacrylamide gel electrophoresis; mass spectrometry; surface-enhanced laser desorption ionization; biomarkers; serum; pancreatic juice.
1. Introduction Proteomics is the term now used for the large-scale analysis of proteins in biological fluids or cells by biochemical methods (see ref. 1 for review). Proteomics focuses on multiprotein complex analysis rather than on individual proteins by studying the interplay of multiple proteins in a complex protein network. New gene expression methodologies, such as DNA microarrays, are very powerful techniques that allow the simultaneous quantification of thousands of mRNAs in a population of cells at a given time. However, levels of mRNA expression do not necessarily correlate with levels of the corresponding protein in a cell (2). Anderson and Seilhamer have reported a 48% correlation coefficient between mRNA and the corresponding protein levels in human liver (3). From: Methods in Molecular Medicine, Vol. 103: Pancreatic Cancer: Methods and Protocols Edited by: G. Su © Humana Press Inc., Totowa, NJ
189
190
Rosty and Goggins
This can be attributable to several parameters: stability of mRNAs, efficiency of protein translation, stability of proteins and their turnover rate. Moreover, each gene may give rise to more than one protein because of endogenous posttranslational modifications (glycosylation, phosphorylation, and so forth) or environmental modifications. For all these reasons, the analytical characterization of the proteome is the most accurate way to give a comprehensive cellular activity snapshot of a given population of cells at a given time. Applications of proteomics techniques are of three main types: (1) identifying each single protein to create a catalog of the protein content of a sample; (2) protein-expression profiling to identify proteins in a sample as a function of a particular state of the organism or the cell (e.g., disease state vs normal state)— differential display proteomics will allow one to identify proteins that are upor down-regulated in a disease-specific manner; and (3) protein-interaction mapping to determine protein functional networks, such as signal transduction cascades. Early proteomic approaches used two-dimensional polyacrylamide gel electrophoresis (2D-PAGE) to separate proteins according to their charge and molecular mass (4,5). 2D-PAGE has been extensively used to study changes in protein expression in cell lines and bulk tissue specimens (6–8) but it has several limitations, such as reproducibility, poor resolution of hydrophobic proteins, and the difficulty of visualizing the less abundant proteins. Recently, the development of new mass-spectrometry-based technologies provides investigators with sensitive and high-throughput approaches for studying the proteome. Matrix-assisted laser desorption ionization time-of-flight (MALDI-TOF) is one such approach in which the protein mixture is analyzed after addition of a light-absorbing chemical matrix (9). After the mixture is spotted to a slide, a laser beam hits the surface and ionizes the protein that will fly through the analyzer to the detector. The speed for the ionized protein to fly is proportional to its mass-to-charge ratio. Surface-enhanced laser desorption ionization (SELDI) is a variant of MALDITOF in which small amounts of proteins are applied to a biochip coated with specific chemical matrices (10). These technologies can be used for large protein profiling or peptide identification after enzymatic digestion. Biomarkers for cancer can be identified using a 2D-PAGE approach or a massspectrometry-based approach. A 2D-PAGE approach has been used to identify tumor-specific altered proteins in esophageal cancer (11), in squamous cell carcinoma of the bladder (12), or in colorectal carcinoma (13). Protein profiling using SELDI is now becoming a first choice method to screen biological fluids. Differentially expressed peaks can be easily and rapidly identified by comparing multiple protein profiles using bioinformatical tools such as the SELDI Biomarker wizard program. A SELDI approach has demonstrated its utility in biomarker discovery for prostate carcinoma and bladder carcinoma (14,15). Newer
Identification of Differentially Expressed Proteins
191
bioinformatics protocols based on cluster analysis are being utilized to identify subtle differences in protein spectra based on relative differences in peak size as opposed to detecting absolute qualitative differences in peaks between samples. Using this approach Liotta and colleagues were recently able to identify peak spectra that segregated cancer samples from noncancer samples and that were able to correctly classify all serum from patients with ovarian carcinoma, with a specificity of 95% (16). Interestingly, none of the ovarian cancer specific peaks were proteins, as the peaks were in the approx 1000 daltons range suggesting they were peptides or other small molecules. In pancreatic cancer, proteomics analysis can be performed with the aim to identify all differentially expressed proteins in cancer cells vs normal pancreatic cells. Protein profiling of pancreatic juice or serum may also identify biomarkers for pancreatic cancer that could be used as diagnostic markers or therapeutic targets (17,18). This chapter outlines the use of 2D-PAGE and SELDI for profiling the protein content of pancreatic juice samples and for identifying proteins differentially expressed in pancreatic cancer patient samples compared to control patient samples (Fig. 1). 2. Materials 1. Pancreatic juice (PJ) samples, collected at the time of surgery for pancreatectomy or during endoscopic retrograde cholangiopancreatography, are stored at -80°C in 2 mM 4-(2-aminoethyl)-benzolsulfonyfluorid (AEBSF) protease inhibitor cocktail. 2. pH 3–10 nonlinear gel strips, 180 mm long and 3 mm wide. 3. Immobilized pH gradient (IPG) strip tray. 4. 8 M urea, 4% 3-(3-cholamidopropyl)dimethylammonio)-1-propanesulfonic acid (CHAPS), 50 mM dithiothreitol (DTT), and a trace of bromophenol blue. 5. 8 M urea, 2% CHAPS, 10 mM DTT, 2% pH 3.5–10 carrier ampholytes (Merck), and a trace of bromophenol blue (see Note 1). 6. 50 mM Tris-HCl, pH 8.4, 6 M urea, 30% glycerol, 2% sodium dodecyl sulfate (SDS), and 10 mM DTT. 7. 50 mM Tris-HCl, pH 8.4, 6 M urea, 30% glycerol, 2% SDS, 2.5% iodoacetamide. 8. 1.875 M Tris-HCl, pH 8.8. 9. 30% acrylamide, 0.8% bisacrylamide. 10. 10% SDS. 11. Ammonium persulfate, a 10% solution diluted in dH2O freshly prepared and stored at 4°C. 12. 50 mM Tris, 198 mM glycine, 0.1% SDS, pH 8.3. 13. 0.5% agarose in 50 mM Tris, 198 mM glycine, 0.1% SDS, pH 8.3. 14. 40% methanol, 7% acetic acid. 15. 50% methanol. 16. 20% ethanol. 17. 25% ammonium. 18. 10 N sodium hydroxyde.
192
Rosty and Goggins
Identification of Differentially Expressed Proteins 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37.
193
0.01% citric acid and 0.1% formaldehyde. 5% Tris, 2% acetic acid. Phosphate-buffered saline (PBS), solution pH 7.4. Freshly prepared 8 M urea and 1% CHAPS in PBS solution, pH 7.4. Freshly prepared 1 M urea diluted in PBS, pH 7.4. 100 mM CuSO4. Immobilized affinity chromatography 3 (IMAC3) ProteinChip array (Ciphergen). ProteinChip bioprocessor (Ciphergen). Saturated sinapinic acid in 50% acetonitrile–0.5% trifluoroacetic acid. SELDI ProteinChip reader (Ciphergen), with the ProteinChip software. 0.1 M ammonium bicarbonate, pH 8.0. 50% acetonitrile, 0.1 M ammonium bicarbonate, pH 8.0. 100% acetonitrile. 25 mM ammonium bicarbonate, pH 8.0, 1 µg/µL of trypsin (from lyophilized bovine pancreatic trypsin) in 10 mM HCl (see Note 2). Normal Phase 1 (NP1) ProteinChip array. 10% saturated a-cyano-4-hydroxy-cinnamic acid (CHCA) in 50% acetonitrile– 0.25% trifluoroacetic acid. N,N,N',N'-Tetramethylethylenediamine (TEMED), butanol, silver nitrate. SpeedVac. Electrophoresis power supply.
3. Methods 3.1. Sample Preparation 1. Mix 50 µL of PJ with 50 µL of a freshly prepared solution containing 8 M urea, 4% CHAPS, 50 mM DTT, and a trace of bromophenol blue.
3.2. Protein Separation by 2D-SDS-PAGE 3.2.1. First-Dimension Immobilized pH Gradient (IPG) 1. Rehydration of the IPG gel strips overnight with 25 mL of a solution containing 8 M urea, 2% CHAPS, 10 mM DTT, 2% pH 3.5–10 carrier ampholytes, and a trace of bromophenol blue (see Note 1). 2. Place the IPG strips, humid electrode wicks, electrodes, and sample cups in the strip tray. 3. Apply 100 µL of the diluted sample at the cathodic end of the IPG strip. 4. Running conditions: Constant 1 mA/gel with linear increase from 300 to 3500 V over 3 h, followed by 3500 V for 3 h. Fig. 1. (Opposite page) Strategy for detection and identification of differentially expressed proteins in pancreatic juice from patients with pancreatic cancer using a first 2D-PAGE protein separation and SELDI for protein identification. After 2D gel electrophoresis, protein mass can be determined using mass spectrometry prior to protease digestion. Individual peptides are indicated by the arrows in (C).
194
Rosty and Goggins
5. Equilibrate the strips with 100 mL of a solution containing 50 mM Tris-HCl, pH 8.4, 6 M urea, 30% glycerol, 2% SDS, and 10 mM DTT for 12 min. Block the SH groups with 100 mL of a solution containing 50 mM Tris-HCl, pH 8.4, 6 M urea, 30% glycerol, 2% SDS, 2.5% iodoacetamide, and a trace of bromophenol blue for 5 min.
3.2.2. Second Dimension SDS-PAGE 1. Prepare a 10% resolving acrylamide SDS-PAGE gel by mixing 10 mL of 30% acrylamide–0.8% bisacrylamide with 6 mL of 1.875 M Tris-HCl, pH 8.8, 300 µL of 10% SDS, 13.6 mL of dH2O, 20 µL of TEMED, and 200 µL of 10% ammonium persulfate (see Note 3). 2. Pour resolving gel in gel rig until 7 mm of the top of the plate and overlay with water-saturated butanol for 1 h. 3. Discard the butanol and replace with a solution containing 50 mM Tris–198 mM glycine–0.1% SDS, pH 8.3, and leave overnight. 4. After IPG gel strips equilibration, cut the strips to size. 5. Overlay the second dimension gel with a solution containing 0.5% agarose heated at 70°C in 50 mM Tris–198 mM glycine–0.1% SDS, pH 8.3, and immediately load the IPG strips through it. 6. Running conditions: constant 40 mA/gel, 100 V, 5 h.
3.2.3. Protein Detection 1. 2. 3. 4. 5.
6. 7. 8.
9. 10.
Remove the gel from the glass plates and soak in dH2O for 5 min. Fix the gel in 40% methanol, 7% acetic acid. Wash the gel in 50% methanol, three changes in 1 h total time. Wash the gel in 20% ethanol, two changes in 20 min total time. Prepare ammoniacal silver nitrate solution: Dissolve 6 g of silver nitrate with 30 mL of dH2O, and mix into a solution containing 160 mL of dH2O, 10 mL of concentrated ammonia (25%), and 1.5 mL of 10 N sodium hydroxide. Stain the gel in the ammoniacal silver nitrate solution for 30 min. Wash four times in dH2O for 4 min. Develop the image in a solution containing 0.01% citric acid and 0.1% formaldehyde for 5–10 min. When the background stain appears, stop with a solution containing 5% Tris and 2% acetic acid. Scan the gel and save the image as a TIFF file. Use the 2D image analysis software Melanie 3 (Geneva Bioinformatics) to identify and quantify protein spots, and to compare results from different samples.
3.3. Protein Profiling Using SELDI 1. Dilute 20 µL of PJ sample in 30 µL of PBS solution, pH 7.4, containing 8 M urea and 1% CHAPS. 2. Vortex the diluted sample on ice for 5 min. 3. Add 100 µL of 1 M urea diluted in PBS, pH 7.4, add dilute 1:9 with PBS, pH 7.4. 4. Prepare an eight-spot IMAC3 array by loading each spot with 10 µL of 100 mM CuSO4 for 15 min. Repeat once. Rinse each spot with dH2O.
Identification of Differentially Expressed Proteins
195
5. Assembly the array on the bioprocessor. 6. Add 350 µL of diluted pancreatic juice sample onto each spot and incubate for 20 min on a shaker. 7. Wash each spot with the same buffer used for PJ sample dilution. 8. Apply on each spot 0.5 mL of saturated sinapinic acid diluted in 50% acetonitrile–0.25% trifluoroacetic acid. 9. Run the ProteinChip in the SELDI reader with the following conditions: sensitivity 10, laser intensity between 250 and 280. Collect an average 80 nitrogen laser shots. 10. Identify cluster of peaks differentially present in cancer samples compared to normal samples using the Biomarker wizard function of the ProteinChip software.
3.4. Protein Identification (see Note 2) 3.4.1. In-Gel Protease Digestion 1. Run another 2D-SDS-PAGE and cut the gel piece containing the protein of interest, found to be differentially expressed in cancer samples compared to control samples. 2. Incubate the gel piece in 0.4 mL of 0.1 M ammonium bicarbonate, pH 8.0, on a shaker for 10 min. Repeat once with fresh solution. 3. Wash the gel piece with 0.5 mL of 50% acetonitrile–0.1 M ammonium bicarbonate, pH 8.0, on a shaker for 1 h. 4. Replace the buffer with 50 µL of 100% acetonitrile. Incubate for 15 min. Air-dry in a SpeedVac for 15 min. 5. Add to 50 µL of 25 mM ammonium bicarbonate, pH 8.0, containing 0.1 µg/µL of trypsin and incubate in a dry air incubator at 37°C for 16 h (see Note 4).
3.4.2. Peptide Mapping by SELDI Analysis 1. Apply 2 µL of the ammonium bicarbonate solution surrounding the gel piece on a NP1 ProteinChip array. Air dry for 10 min (see Note 5). 2. Wash the ProteinChip spot with 5 µL of dH2O. Air-dry for 5 min. 3. Apply 0.5 µL of 10% saturated CHCA diluted in 50% acetonitrile–0.25% trifluoroacetic acid: add 200 µL of 100% acetonitrile and 200 µL of 0.5% trifluoroacetic acid to CHCA, vortex-mix, and let stand for 5 min. Centrifuge at 10,000g for 30 s before diluting 1:10 in 50% acetonitrile–0.25% trifluoroacetic acid. Air-dry for 5 min. 4. Run the ProteinChip in the SELDI reader with the following conditions: time lag focusing 2 kDa, sensitivity 8, laser intensity between 150 and 200. Fire continuously until approx 100 transients have been averaged (see Note 6).
3.4.3. Protein Database Search 1. List all peaks from the spectrum. Choose a minimum of seven different peaks with the highest intensity (see Note 7). 2. Open one of the following Web protein databases: http://prowl.rockefeller.edu/ profound_bin/WebProFound.exe, http://us.expasy.org/tools/peptident.html, http:/
196
Rosty and Goggins
/www.mann.embl-heidelberg.de/GroupPages/PageLink/peptidesearchpage.html, http://prospector.ucsf.edu/ucsfhtml4.0/msfit.htm. 3. Set database search parameters: Specify the mass range as close to the protein of interest as possible, specify cysteine residues have been modified by acrylamide, peptide mass tolerance to 1.0 Da, maximum missed cleavages to 1–4.
4. Notes 1. First dissolve urea and CHAPS in dH2O in a 37°C water bath, then add DTT and ampholytes. Always use a fresh buffer for sample dilution. 2. The protocol for protein identification described in this chapter follows the 2DPAGE approach in which candidate proteins have been identified in the gels. Protein identification following SELDI profiling can be achieved after purification of the protein of interest according its mass and biochemical properties (type of surface it binds). 3. The second dimension gel can be made without adding SDS to obtain a more homogeneous gel and to reduce the concentration of unpolymerized monomers in the polyacrylamide. The SDS present in the running buffer is sufficient to maintain the proteins with the necessary negative charges. Piperazine diacrylyl (PDA) and sodium thiosulfate can be added to the gel to reduce the background in silver staining. 4. Make 5-µL trypsin aliquots stored in −80°C no longer than 6 mo by dissolving 25 µg of trypsin in 25 µL of 10 mM HCl to produce a 1 µg/µL trypsin solution. Just before use, thaw a trypsin aliquot and add 45 µL of 25 mM ammonium bicarbonate, pH 8.0, to obtain a 0.1 µg/µL trypsin final concentration. 5. If the buffer has evaporated after overnight digestion, resuspend the gel piece with 20 µL of 25 mM ammonium bicarbonate, pH 8.0, vortex-mix, and allow to equilibrate for 30 min before use on the ProteinChip. 6. When running the ProteinChip in the SELDI reader, it is better to calibrate the reader within the appropriate mass range, to ensure the best mass accuracy. Calibration can be performed before running the chip by using a combination of three to five peptide standards (Ciphergen) on a NP1 chip or along with the chip reading by adding the same peptide standards to the mixture to analyze. Check on the protein spectrum if the digested trypsin peak appears at 2163.3 Da. If it is not present, it may be related to low trypsin activity. If it is present with a slightly different mass, it may be used to internally calibrate the reader. 7. To determine which peptides are derived from the protein digestion itself, it is recommended to process separetely a piece of protein-free SDS-PAGE gel. All peaks common in the sample of interest and the no-protein control mass spectrometry profiles will be ignored. SELDI can determine protein mass with a 0.01–0.1% mass accuracy. This level of resolution may not always be sufficient to identify peptide fragments. Thus, peptide mass fingerprinting may require the use of more sensitive MALDI mass spectrometers or electrospray mass spectrometry (ESI).
References 1. 1 Pandey, A. and Mann, M. (2000) Proteomics to study genes and genomes. Nature 405, 837–846.
Identification of Differentially Expressed Proteins
197
2. 2 Gygi, S. P., Rochon, Y., Franza, B. R., and Aebersold, R. (1999) Correlation between protein and mRNA abundance in yeast. Mol. Cell Biol. 19, 1720–1730. 3. 3 Anderson, L. and Seilhamer, J. (1997) A comparison of selected mRNA and protein abundances in human liver. Electrophoresis 18, 533–537. 4. Wilkins, M. R., Pasquali, C., Appel, R. D., et al. (1996) From proteins to proteomes: Large scale protein identification by two-dimensional electrophoresis and amino acid analysis. Biotechnology (NY) 14, 61–65. 5. 5 Anderson, N. G. and Anderson, N. L. (1996) Twenty years of two-dimensional electrophoresis: Past, present and future. Electrophoresis 17, 443–453. 6. 6 Young, D. S. and Tracy, R. P. (1995) Clinical applications of two-dimensional electrophoresis. J. Chromatogr. A 698, 163–179. 7. 7 Okuzawa, K., Franzen, B., Lindholm, J., et al. (1994) Characterization of gene expression in clinical lung cancer materials by two-dimensional polyacrylamide gel electrophoresis. Electrophoresis 15, 382–390. 8. 8 Franzen, B., Hirano, T., Okuzawa, K., et al. (1995) Sample preparation of human tumors prior to two-dimensional electrophoresis of proteins. Electrophoresis 16, 1087–1089. 9. 9 Shevchenko, A., Loboda, A., Ens, W., and Standing, K. G. (2000) MALDI quadrupole time-of-flight mass spectrometry: A powerful tool for proteomic research. Anal. Chem. 72, 2132–2141. 10. Hutchens, T. W. and Yip, T. T. (1993) New desorption strategies for the mass spectrometric analysis of macromolecules. Rapid Commun. Mass Spectrom. 7, 576–580. 11. 11 Emmert-Buck, M. R., Gillespie, J. W., Paweletz, C. P., et al. (2000) An approach to proteomic analysis of human tumors. Mol. Carcinog. 27, 158–165. 12. 12 Ostergaard, M., Wolf, H., Orntoft, T. F., and Celis, J. E. (1999) Psoriasin (S100A7): A putative urinary marker for the follow-up of patients with bladder squamous cell carcinomas. Electrophoresis 20, 349–354. 13. 13 Jungblut, P. R., Zimny-Arndt, U., Zeindl-Eberhart, E., et al. (1999) Proteomics in human disease: Cancer, heart and infectious diseases. Electrophoresis 20, 2100–2110. 14. 14 Vlahou, A., Schellhammer, P. F., Mendrinos, S., et al. (2001) Development of a novel proteomic approach for the detection of transitional cell carcinoma of the bladder in urine. Am. J. Pathol. 158, 1491–1502. 15. Wright, G. L. Jr., Cazares, L. H., Leung, S. M., et al. (1999) ProteinChip surface enhanced desorption/ionization (SELDI) mass spectrometry: A novel protein biochip technology for detection of prostate cancer biomarkers in complex protein mixtures. Prostate Cancer Prostat. Dis. 2, 264–276. 16. 16 Petricoin, E. F., Ardekani, A. M., Hitt, B. A., et al. (2002) Use of proteomic patterns in serum to identify ovarian cancer. Lancet 359, 572–577. 17. 17 Rosty, C., Christa, L., Kuzdzal, S., et al. (2002) Identification of hepatocarcinomaintestine-pancreas/pancreatitis-associated protein I as a biomarker for pancreatic ductal adenocarcinoma by protein biochip technology. Cancer Res. 62, 1868–1875. 18. Koopman, J., Zhang, Z., White, N., et al. (2004) Serum diagnosis of pancreatic adenocarcinoma using surface-enhanced laser desorption and ionization mass spectrometry. Clin. Cancer Res. 10, 860–868.
198
Rosty and Goggins
Telomerase Activity in Pancreatic Cancer
199
13 Detection of Telomerase Activity in Patients with Pancreatic Cancer Kazuhiro Mizumoto and Masao Tanaka Summary Telomerase, which ensures the unlimited proliferation by adding TTAGGG repeat at the end of the chromosome, is strongly activated at a very high incidence in a variety of malignant neoplasms including pancreatic cancer. In addition to the acquisition of the immortality, telomerase plays an important role in the aggressive behavior of pancreatic cancer. Invasiveness of human pancreatic cancer cells correlates well with telomerase activity. Exposure of pancreatic cancer to anticancer drugs up-regulates telomerase activity, and the increase in telomerase activity correlates with resistance to the drug-induced apoptosis. More important, diagnositic values of telomerase activity are highly focused because of the lack of other specific genetic markers for pancreatic cancer. Samples of pancreatic juice are obtained at endoscopic retrograde pancreatography using a balloon catheter after intraveneous injection of secretin. Because the pancreatic juice has strong protease and RNase activity, addition of protease inhibitors and RNase inhibitors in the telomerase extraction buffer is necessary for the detection of telomerase activity in the pancreatic juice. A telomeric ladder was detected in 80% patients with carcinoma, whereas only 4.3% patients with adenoma and none with chronic pancreatitis showed positive telomerase activity. Key Words: Telomerase; diagnosis; molecular marker; pancreatic cancer; pancreatic juice; endoscopic retrograde pancreatography.
1. Introduction Chromosome ends are composed of telomere that protects chromosomes from various stimuli that damage DNA. Telomere is shortened at each cell division, and the limit of the cell division based on shortening of the telomere length regulates the life span of somatic cells (1,2). In contrast to the somatic cells, carcinoma cells can acquire unlimited proliferation, that is, immortality, by adding TTAGGG repeat at the end of the chromosome through activation of telomerase (3,4). Telomerase is strongly activated at very high incidence in a variety of From: Methods in Molecular Medicine, Vol. 103: Pancreatic Cancer: Methods and Protocols Edited by: G. Su © Humana Press Inc., Totowa, NJ
199
200
Mizumoto and Tanaka
malignant neoplasms including pancreatic cancer (5,6). Research on activation of telomerase is fundamentally important to understand cancer biology. In addition to the immortal character derived from telomerase activation, various roles of telomerase in pancreatic cancer has been documented in relation to the aggressive biological properties of this tumor, for example, strong invasive ability and potent drug resistance. Invasiveness of human pancreatic cancer cells correlates well with telomerase activity and inhibition of telomerase by transfection with antisense oligonucleotides of telomerase RNA decreases their motility and invasion rates (7). Exposure of pancreatic cancer to anticancer drugs up-regulates telomerase activity, and the increase in telomerase activity correlates with resistance to the drug-induced apoptosis in pancreatic cancer, indicating the anti-apoptotic effects of telomerase (8). Mutation of the p53 gene is frequently observed in pancreatic cancer and correlates with poor prognosis of the tumor. Wild-type p53 gene transduction to the pancreatic cancer harboring p53 mutation inhibits telomerase activity through down-regulation of hTERT mRNA, a catalytic subunit of telomerase (9). Telomerase plays an important role in the aggressive biological behavior of pancreatic cancer as mentioned above, and the diagnostic values of telomerase is more highly focused because of the lack of other specific genetic markers in pancreatic cancer. Using the telomerase amplification protocol (TRAP) developed by Kim et al. (10), we reported that 80% of pancreatic cancer tissues presented high telomerase activity, whereas adenoma, chornic pancreatitis, and normal tissues displayed very weak activity (11). K-ras point mutation is also frequently observed in pancreatic cancer tissues, but it is also present in chronic pancreatitis tissues (12). Specificity of p53 gene mutation in pancreatic cancer is quite high, but its positivity rate is not sufficient for diagnostic use (13). Therefore, telomerase could be the only specific and sensitive marker for distinguishing pancreatic cancer from pancreatitis and adenoma. In this chapter, we describe how to collect a sufficient amount of the pancreatic juice for telomerase assay during endoscopic retrograde pancreatography (ERP). We could obtain appropriate samples of pancreatic juice, which was rich in duct epithelial cells, by the use of a balloon catheter, and a telomeric ladder was detected in 20 out of 24 (80%) patients with carcinoma (Table 1). Meanwhile, only 1 out of 23 (4.3%) patients with adenoma and none of 23 chronic pancreatitis showed positive telomerase activity. 2. Materials 1. Telomerase extraction buffer (CHAPS lysis buffer): 10 mM Tris-HCl, pH 7.5, 1 mM MgCl2, 1 mM EGTA, 0.1 mM 4-(2-aminoethyl)-benzenesulfonylfluoride hydrochroride, 5 mM b-mercaptoethanol, 0.5% 3-(3-cholamidopropyl) dimethylammonio)1-propanesulfonic acid (CHAPS), 10% glycerol, 1 µg/mL each of protease inhib-
Telomerase Activity in Pancreatic Cancer
201
Table 1 Positivity Rates of Telomerase Activity in Pancreatic Juice Obtained at Endoscopic Retrograde Pancreatography Diagnosis Carcinoma Adenoma Pancreatitis
2. 3. 4. 5. 6. 7. 8. 9.
Number of patients
Number of patients with positive telomerase activity (%)
24 23 23
20 (83.3%) 12 (4.3%) 0 (0%)
itors (antipain, leupeptin, phophoramidon, elastatinal, pepstatin A, chymostatin, Peptide Institute, Osaka, Japan). TRAP reaction buffer: 20 mM Tris-HCl, pH 8.3, 1 mM MgCl2, 63 mM KCl, 0.05% Tween-20, 1 mM EGTA. Deoxynucleotide triphophates (dATP, dTTP, dGTP, dCTP). [a-32P]dCTP. TS primer (5'-AATCCGTCGAGCAGAGTT-3'). CX primer (5'-CCCTTACCCTTACCCTTACCCTAA-3'). Taq polymerase (Promega, Madison, WI). Control pancreatic cancer cell. 12% polyacrylamide nondenaturing gel.
3. Methods 3.1. Collection of Pancreatic Juice Samples of the pancreatic juice for the measurement of telomerase activity are obtained at ERP using a balloon catheter (see Note 1). A 350-cm no. 6 French balloon catheter (wedge pressure catheter JX-283; Arrow International Inc., Reading, PA) is inserted through a duodenoscope into the pancreatic duct and pancreatography is performed. For the precise examination of the pancreatic duct, contrast medium (sodium diatrizoate) is injected after removal of the duodenoscope while leaving the catheter in the pancreatic duct with the aid of the inflated balloon (Fig. 1). Spot films obtained with the patient in the supine position are useful to take clear pancreatograms. To collect a sufficient amount of the pancreatic juice containing pancreatic duct epithelial cells, 1 U/kg body weight of secretin (Eisai, Tokyo, Japan) is injected intraveneously. The pancreatic juice is collected for 15–20 min through the balloon catheter. A sample obtained for the first 5 min is discarded because of contamination of contrast medium. Even if a large number of epithelial cells exist in the first sample, the cells float in the pancreatic juice mixed with contrast medium after centrifugation, and thus sufficient cell pellets could not be collected from the pancreatic
202
Mizumoto and Tanaka
Fig. 1. Collection of pancreatic juice samples for telomerase assay. The balloon catheter is necessary to obtain sufficient samples.
juice for the first 5 min (see Note 2). The pancreatic juice subsequently obtained is used for cytology and telomerase activity determination. For the cytological examination, a part of the pancreatic juice is centrifuged at 400–700g for 5 min, and smears are submitted to Papanicolaou staining (see Note 3). The samples are considered as appropriate for telomerase activity assay if the presence of epithelial cells is confirmed. For patients with a stenotic or an obstructed pancreatic duct, the brushing technique may be necessary to yield a sufficiently large number of cells for the analysis. A brush catheter (BC24Q, Olympus, Tokyo, Japan) is inserted into the stenotic duct, and the duct is scrubbed several times. The tip with brush is immediately washed in saline, which serves as samples for cytological examination and telomerase activity assay. Biopsy samples from the stenotic or obstructed duct are also adequate for telomerase activity. The samples for telomerase assay are immediately centrifuged at 1000g for 5 min at 4°C. Supernatant is completely discarded and the pellet or the biopsy sample is washed once in ice-cold phosphate-buffered saline (PBS), centrifuged again at 1000g for 5 min at 4°C, and then stored at -80°C until use for telomerase assay. Prompt and complete washing is extremely important because
Telomerase Activity in Pancreatic Cancer
203
of the pancreatic juice has strong activity of protease and RNase that disturb the telomerase assay (see Note 4). 3.2. TRAP 1. Resuspend the cell pellets in 20 µL of telomerase extraction buffer containing protease inhibitors. 2. Incubate the suspension for 30 min on ice. 3. Centrifuge the lysates at at 16,000g for 20 min at 4°C. 4. Transfer the supernatant into a fresh tube and the concentration of protein in the supernatant is measured by Bradford assay. The supernatant can be stored at -80°C until use. 5. Prepare extracts from pancreatic cancer cell lines equivalent to 10, 100, and 1000 cells as positive controls. The extracted samples from pancreatic cancer cells are incubated at 85°C for 10 min to use it as a negative control. 6. Incubate the extracts (0.06–0.6 µg) with TRAP reaction buffer containing 50 µM dNTPs, 0.3 µCi of [a-32P]dCTP, 2 U of Taq DNA polymerase, 0.1 µg of TS primer at 20°C for 30 min, and heat to 90°C for 3 min. 7. Add 0.1 µg of CX primer to the reaction mixture. 8. Subject the reaction mixture to 31 cycles of polymerase chain reaction (PCR) at 94°C for 30 s, 50°C for 30 s and 72°C for 90 s. 9. Analyze the PCR products by electrophoresis in 0.5X Tris-borate EDTA buffer on 12% polyacrylamide gels. 10. Gels are processed for autoradiography and exposed for 8–12 h. Telomerase activity in the pancreatic juice from patients with pancreatic cancer is detected as sixnucleotide repeat ladder (Fig. 2). 11. Measure signal intensity by NIH image. 12. Express relative telomerase activity as the equivalent telomerase intensity of the number of pancreatic cancer cell lines.
4. Notes 1. Use of a balloon catheter is essential to obtain enough cell pellets for telomerase assay. Many cells will be lost during the sampling procedure if the conventional catheter for ERP is used. The balloon catheter is also important to delineate possible minimal changes in the duct by ductal neoplasms. 2. Contrast medium is seen in syringe while sampling with careful observation. If a large amount of contrast medium are present in the pancreatic juice, the samples must be discarded even after 5 min following the secretin injection. 3. If a large number of lymphocytes are stained in cytology, false-positive results may be obtained. 4. The pancreatic juice is very rich in protease and RNase. A telomerase detection kit does not contain sufficient protease inhibitors to detect telomerase activity in the pancreatic juice. Thus, protease inhibitors must be added to the telomerase extraction buffer. Addition of RNase may also be useful for the detection of telomerase activity in the pancreatic juice (14).
204
Mizumoto and Tanaka
Fig 2. Telomerase amplification protocol. Telomeric ladder could be seen in the pancreatic juice samples from pancreatic cancer patients but not from noncancerous patients. Positive controls of pancreatic cancer cells lines are used for semiquantitative determination of telomerase activity.
References 1. 1 Harley, C. B., Futcher, A. B., and Greider, C. W. (1990) Telomeres shorten during ageing of human fibroblasts. Nature 345, 458–460. 2. Shay, J. W., Werbin, H., and Wright, W E. (1994) Telomere shortening may contribute to aging and cancer. A perspective. Mol. Cell Different. 2, 1–21. 3. 3 Chadeneau, C., Hay, K., Hirte, H. W., et al. (1995) Telomerase activity associated with acquisition of malignancy in human colorectal cancer. Cancer Res. 55, 2533–2536. 4. 4 Morin, G. B. (1989) The human telomere terminal transferase enzyme is a ribonucleoprotein that synthesizes TTAGGG repeats. Cell 88, 521–529. 5. 5 Mizumoto, K., Suehara, N., Muta, T., et al. (1996) Semi-quantitative analysis of telomerase in pancreatic ductal adenocarcinoma. J. Gastroenterol. 31, 894–897l. 6. 6 Niiyama, H., Mizumoto, K., Sato, N., et al. (2001) Quantitative analysis of hTERT m RNA expression in colorectal cancer. Am. J. Gastroenterol. 96, 1895–1900. 7. 7 Sato, N., Maehara, N., Mizumoto, K., et al. (2000) Telomerase activity of cultured human pancreatic carcinoma cell lines correlates with their potential for migration and invasion. Cancer 91, 496–504.
Telomerase Activity in Pancreatic Cancer
205
8. 8 Sato, N., Mizumoto, K., Kusumoto, M., et al. (2000) Up-regulation of telomerase activity in human pancreatic cancer cells after exposure to etoposide. Br. J. Cancer 82, 1819–1826. 9. Kusumoto, M., Ogawa, T., Mizumoto, K., et al. (1999) Adenovirus-mediated p53 9 gene transduction inhibits telomerase activity independent of its effects on cell cycle arrest and apoptosis in human pancreatic cancer cells. (1999) Clin. Cancer Res. 5, 2140–2147. 10. 10 Kim, N. W., Piatyszek, M. A., Prowse, K. R., et al. (1994) Specific association of human telomerase activity with immortal cells and cancer. Science 266, 2011–2015. 11. 11 Suehara, N., Mizumoto, K., Muta, T., et al. (1997) Telomerase elevation in pancreatic ductal carcinoma compared to nonmalignant pathological states. Clin. Cancer Res. 3, 993–998. 12. Furuya, N., Kawa, S., Akamatsu, T., et al. (1997) Long-term follow-up of patients with chronic pancreatitis and k-ras gene mutation detected in pancreatic juice. Gastroenterology 113, 593–598. 13. Casey, G., Yamanaka, Y., Friess, H., et al. (1993) p53 mutations are common in 13 pancreatic cancer and are absent in chronic pancreatitis. Cancer Lett. 69, 151–160. 14. Nakamura, Y., Tahara, E., Tahara, H., et al. (1999) Quantitative reevaluation of telomerase activity in cancerous and noncancerous gastrointestinal tissues. Mol. Carcinogen. 26, 312–320.
206
Mizumoto and Tanaka
SEREX Technique
207
14 Serological Analysis of Expression cDNA Libraries (SEREX) An Immunoscreening Technique for Identifying Immunogenic Tumor Antigens Yao-Tseng Chen, Ali O. Gure, and Matthew J. Scanlan Summary SEREX (serological analysis of recombinant tumor cDNA expression libraries) is a technique designed to isolate tumor antigens that have elicited high-titer IgG responses in human hosts. This is an immunoscreening method for gene cloning, with two key features that distinguish it from earlier immunoscreenings used to identify targets in autoimmune diseases. First, the assay was designed, at last originally, to analyze autologous immunological responses to cancer, that is, the tumor cDNA library and the screening serum were obtained from the same patient. Second, the screening is performed with serum samples at high dilution (1:1000–1:100), and the secondary antibody used was specific for human IgG. This ensures that only antigens eliciting high-titer IgG responses will be isolated. This latter feature is important in that a key purpose of SEREX is to identify tumor antigens that have elicited both humoral and cell-mediated immune responses in cancer patients. Several tumor antigens identified by SEREX on various types of cancer indeed showed such characteristics and these antigens are being tested as targets for therapeutic cancer vaccines. Key Words: Tumor immunology; immunotherapy; cancer vaccine; tumor antigen, SEREX.
1. Introduction In the early 1990s, Boon and van der Bruggen (1) reported the cloning of MAGE-1, the first human tumor antigen shown to have elicited a spontaneous cytotoxic T-lymphocyte (CTL) response in a cancer patient. This led to work on antigen-specific cancer vaccines, and the identification of additional tumor antigens, particularly the ones that are immunogenic in human hosts, became a major focus in the field of tumor immunology. However, the cDNA transfection-based
From: Methods in Molecular Medicine, Vol. 103: Pancreatic Cancer: Methods and Protocols Edited by: G. Su © Humana Press Inc., Totowa, NJ
207
208
Chen et al.
Fig. 1. The SEREX technique.
genetic approach that has led to the identification of MAGE is technically difficult and requires the establishment of autologous CTL lines and tumor cell lines from the same patient, a task not achievable for most epithelial tumor types. To circumvent this limitation, a serologically based approach, SEREX (serological analysis of recombinant tumor cDNA expression libraries), was designed to isolate tumor antigens that have elicited high-titer IgG responses in human hosts (2) (Fig. 1). In their original paper describing this technique (2), Pfreundschuh and his colleagues identified a panel of antigens that included MAGE-1 and tyrosinase, two antigens known to elicit CD8+ T-cell responses. This indicates that SEREX is capable of isolating tumor antigens that have elicited CTL-medicated immune responses, and many articles based on SEREX have since appeared in the literature (see [3] for a review). A SEREX collaborative group, established by the Ludwig Institute for Cancer Research, has applied this technique to many common tumor types, including melanoma, colon cancer, renal cancer, breast cancer, lung cancer, esophageal cancer, and so forth, and more than 2000 clones have been deposited in the SEREX database (http://www2.licr.org/CancerImmunome DB), representing hundreds, if not thousands, of different gene products. Many genes of interests are within this pool of SEREX-defined antigens, including mutational antigens (e.g., p53), cancer/testis (CT) antigens (e.g., NY-ESO-1, SSX2, CT7), and differentiation antigens (e.g., NY-BR-1) (2,4–7). Although the concept behind SEREX is straightforward, two key technical challenges needed to be resolved. One involved the elimination of antibodies
SEREX Technique
209
in human sera that react with bacterial or phage components. This is absolutely essential because such contaminating antibodies would completely obscure the detection of other classes of antibodies. The other was related to the presence of variable numbers of B-cells in tumors; these give rise to immunoglobulin G (IgG) mRNA, which is expressed and detected in SEREX, and a strategy to eliminate these clones needed to be developed. Once a positive clone is identified and purified, the following steps are usually taken to analyze the clones and identify genes of interest: (1) DNA sequencing to establish identity, similarity, or uniqueness with regard to genes in the existing data banks and the search for possible structural (e.g., mutational) abnormalities; (2) analysis of the mRNA expression pattern in normal tissues and in tumors; and (3) immunogenicity as measured by the frequency of antibodies in sera from normal individuals and patients with the same tumor type. Subsequent analysis of interesting clones showing cancer-relatedness in terms of sequence abnormalities, expression patterns, or seroreactivity includes chromosomal mapping, generating monoclonal antibodies for biochemical and immunohistochemical study, and serological surveys of antibody reactivity in patients with various types of human cancer. 2. Materials 2.1. RNA Extraction and cDNA Library Construction 1. Micro-FAST Track mRNA Isolation Kit (Invitrogen). 2. ZAP Express XR Library Construction Kit (Stratagene). Common reagents needed for bacteria and phage propagation and storage can be found in the manual included with this kit and in common molecular biological protocols, and are not repeated here. These include the preparation of LB and NZYCM bacterial plates, NZY medium, top agar, maltose solution, MgSO4 solution, SM buffer for phage storage, antibiotic stocks (e.g., tetracycline and kanamycin).
2.2. Immunoscreening 1. Nitrocellulose membranes (Schleicher & Schuell, PROTRAN BA85, 135 mm and 82 mm). 2. Binding buffer: 0.1 M NaHCO3, pH 8.3. 3. Tris-buffered saline (TBS), 20 mM Tris-HCl, pH 7.5, 150 mM NaCl. 4. TBST: TBS with 0.05% Tween-20. 5. Blocking solution: TBS with 5% nonfat dry milk. 6. TBS–1% BSA (bovine serum albumin). 7. 10% sodium azide. 8. 0.5 M isopropylthio-b-D-galactoside (IPTG) (Invitrogen). 9. Goat anti-human IgG, Fc-g specific, alkaline phosphatase-conjugated (Jackson ImmunoResearch).
210
Chen et al.
10. Alkaline phosphatase reaction buffer: 100 mM Tris-HCl, pH 9.5, 100 mM NaCl, 50 mM MgCl2. 11. Developing solution for alkaline phosphatase: 300 mL of alkaline phosphatase reaction buffer, 82.5 µg/mL of 5-bromo-4-chloro-3-indolyl-phosphate (BCIP), and 100 µg/mL of nitroblue tetrazolium chloride (NBT).
3. Methods 3.1. RNA Extraction and Construction of cDNA Expression Library 1. Prepare total RNA from fresh-frozen tumor or normal tissue, or from cultured cell lines by guanidinium thiocyanate–CsCl gradient method (8). Use enough material to generate at least 150–300 µg of total RNA (see Note 1). 2. Use at least 150 µg of total RNA to isolate poly(A)+ RNA. Several commercial kits are available for this purpose, for example, Micro-FAST Track mRNA Isolation Kit (Invitrogen). Determine RNA yield by measuring optical density (OD) absorbance at 260 nm. 3. Use 5 µg of poly(A)+ RNA for one cDNA library construction. lZAP-Express vector (Stratagene) is most commonly used for SEREX libraries, and the manufacturer’s protocol is followed (see Note 2). The size of the library varies, but a library size of at least 1 ´ 106 should be obtained for a meaningful SEREX screening. 4. The library can be amplified following the manufacturer’s protocol if desired. Screening of an unamplified library, however, is often preferred.
3.2. Absorption of Human Serum The serum absorption step is crucial to remove antibodies reactive to E. coli and to the phage. Unabsorbed serum invariably gives a high background in the subsequent primary screening, and will totally obscure even the strong positive clones. For screening of 20 nitrocellulose filters in each set of experiment (see below), 300–400 mL of diluted serum is needed. As most SEREX analysis uses serum at a final dilution of 1:100 to 1:400, this translates to a need of absorbing 1–5 mL of undiluted serum. 1. Preparation of E. coli/l bacteriophage affinity columns: a. Amplify two to five plates of wild-type lZAP Express bacteriophage in E. coli XL1 Blue MRF at a concentration of 104 pfu per 15-cm plate of NZYCM agar, for 15 h at 37°C. b. Add 10 mL of binding buffer per agar plate, and incubate at 4°C overnight with gentle shaking. c. Collect the resultant supernatant from the agar plate and sonicate to lyse residual E. coli. d. Couple the lysate to CNBr-Sepharose 4B (Amersham Pharmacia Biotech) as per the manufacturer’s instructions. 2. Pack resin in a 15-mL disposable column to a 5-mL bed volume. Wash with 10 volumes of TBS.
SEREX Technique
211
3. Dilute serum 1:10 with TBS–1% BSA. 4. Pass the diluted serum through the column five times, wash the column with TBS– 1% BSA, and dilute the absorbed serum to the desired dilution (1:100–1:400) with TBS–1% BSA. Add sodium azide to 0.02% and store at 4°C.
3.3. Primary Immunoscreening Primary screening of at least 5 ´ 105 to 1 ´ 106 is needed for a complete library screening. This translates to approx 100–200 plates at a density of 5000 pfu per plate. Screening at higher density may risk decreasing the signal intensity of the positive clones and is not recommended. At least 20 plates can be comfortably processed at one time. All steps are at room temperature unless specified. 1. Plate out recombinant phage on NZYCM plates at a density of approx 5000 per 15-cm plate, using top agar containing IPTG (30 µL of 0.5 M IPTG per 6 mL of top agar). Incubate overnight at 37°C. 2. Overlay with premarked nitrocellulose membranes (S&S). Puncture three or four holes with a needle on the filters and agar for future orientation. Incubate at 37°C for an additional 2–4 h. 3. Transfer the membranes to a tray containing approx 200 mL of TBST. Scrub the membrane to physically remove bacterial debris. Wash with 300 mL of TBST for 10 min while shaking. 4. Repeat washing with TBST three times, 10 min each. 5. Transfer to 300 mL of blocking solution (TBS with 5% nonfat dry milk). Incubate for 1 h while shaking. 6. Wash twice with TBST, 10 min each, followed by one washing in TBS for 10 min. 7. Prepare 20 Petri dishes, pipet 15 mL of diluted absorbed serum to each dish. 8. Transfer one filter to each Petri dish, with phage side facing up. Incubate at room temperature overnight with continuous shaking. 9. Pool all filters in a tray, wash twice with 300 mL of TBS-T, followed by one washing with TBS, 10 min each with shaking. 10. Incubate with goat anti-human IgG, Fc-g fragment specific, alkaline phosphataseconjugated (Jackson Immunolabs), 1:3000 dilution in 300 mL of TBS with 1% BSA. Incubate for 1 h at room temperature with shaking. 11. Wash 3 times with TBS while shaking. 12. Prepare developing solution for alkaline phosphatase and pipet 15 mL into each of the 20 plates. 13. Transfer filters into individual Petri dishes. Shake while developing. Monitor closely the developing step and stop as soon as the signals are clearly visible or when the background phage plaques are appearing (Fig. 2). This usually takes 3 to <10 min. Stop by transferring the filters to a tray of water (see Note 3). 14. Locate the positive phage plaques by aligning the needle marks on the filter and the plate. Recover the positive phage with a Pasteur pipet and dispense each agar core with positive phage into 0.5 mL of SM buffer. Add 20 µL of chloroform (see Note 4).
212
Chen et al.
Fig. 2. An example of a primary screening filter, showing two positive clones. The background phage plaques gave no background in this screening and were not visible.
15. Incubate overnight to allow phage diffusion into buffer. Centrifuge the agar core and cell debris. Transfer the phage suspension to a fresh tube. Titer the phage if desired.
3.4. Secondary Immunoscreening Secondary immunoscreening serves the purpose of purifying the positive clones, at the same time allowing the detection and elimination of cloning encoding IgG, which would appear as positive clones in the primary screening. 1. Plate out each primary positive clone at a density of 100–300 pfu per 8-cm plate. This may be achieved by empirically plating out 0.5–1 µL from the 0.5-mL phage suspension from above; however, titering individual phage suspension will allow a better estimation. 2. Follow the procedure above for primary screening to step 6, before the serum incubation step. The only difference is that while marking the nitrocellulose filters, mark clone designations on both top and bottom halves of the filters. 3. Cut each membrane into the top and bottom halves. Process the top halves through serum incubation and washing steps, while leaving the bottom halves of the filters in TBS overnight at room temperature. Instead of incubating in individual Petri dish, the serum incubation step for secondary screening can be carried out as a pool in a tray, or in individual Petri dish if desired.
SEREX Technique
213
Fig. 3. Example of a positive secondary screening filter. Negative phage plaques were present on this filter; some of the negative plaques overlapped with positive ones, resulting in positive signals in the shape of incomplete circles.
4. Pool all filters and process for secondary antibody incubation and developing as above for primary screening, again in a tray with shaking. A phage clone encoding IgG would show positive signals on both top and bottom halves, whereas a true SEREX-positive clone would be positive on the top half of the filter, but not the bottom half (Fig. 3, see also Note 5). 5. Isolate the positive clones and dispense into 0.5 mL of SM buffer with chloroform as above. Centrifuge and transfer the phage suspension to a fresh tube after overnight incubation. This is the purified phage stock. If the purity of the clone is in doubt, a tertiary screening should be performed to ensure that the positive clone is purified.
3.5. In Vivo Excision of Plasmid DNA The positive clones can now be converted to pBK-CMV plasmid form by in vivo excision. This process infects the phagemid-containing XL1-blue cells with an M13 filamentous helper phage (ExAssist, Strategene). This ultimately leads to release of the plasmid into the culture media, which is then used to infect XLOLR cells. The XLOLR cells are then plated out onto LB/Kanamycin plates. 1. Grow up XL-1 blue in NZY with 0.2% maltose, and XLOLR in NZY. Estimate 400 µL of culture per clone to be excised. Grow at 37°C with vigorous shaking for 3–4 h, till OD600 of approx 0.6–0.7. 2. Centrifuge and resuspend both bacteria in 10 mM MgSO4 to OD600 of 1.0. 3. Combine in a 15-mL polypropylene tube 200 µL of XL-1 blue, 250 µL of phage suspension (out of 500 µL), and 1 µL of ExAssist helper phage. Incubate for 15 min at 37°C.
214
Chen et al.
4. Add 3 mL of NZY media; incubate for 3 h at 37°C with shaking. 5. Heat at 65–70°C for 20 min. Centrifuge the bacterial debris at 1000g for 15 min. 6. Transfer the supernatant to a fresh tube. This is the excised phagemid stock that can be stored at 4°C. 7. Combine in a 1.5-mL tube 200 µL of XLOLR and 100 µL of the phagemid stock; incubate for 15 min at 37°C. Add 3 mL of NZY media and incubate at 37°C for 45 min with shaking. 8. Plate out 100 µL on LB plates containing kanamycin at 50 µg/mL. Incubate overnight. This should normally yield dozens to hundreds of colonies. The plasmidcontaining bacteria is now ready for small-scale plasmid DNA preparation and subsequent analysis, for example, DNA sequencing.
4. Notes 1. Good RNA quality is crucial to the success of the library construction. Although this is usually not a problem for cell lines, preparing high-quality RNA from tumor samples is more challenging. Once the tumor is surgically resected, a portion of the specimen taken for RNA preparation should be snap-frozen in liquid nitrogen, and areas of necrosis should be avoided. We have found the traditional guanidinium thiocyanate–CsCl gradient method to be most reliable in generating good quality RNA for this purpose. 2. Although the manufacturer’s protocol usually yields a good library, we did encounter occasional failures. The steps that we have found crucial are the cDNA blunting step, EcoRI adapter ligation step, and the final step of recovering cDNA inserts for ligating to vector. We have made the following modifications that might be helpful should one encounter difficulties in cDNA library construction: a. For the cDNA blunting step with Pfu polymerase, we have added a phenol–chloroform extraction and ethanol precipitation step after second-strand cDNA synthesis. The precipitated DNA is reconstituted in 150 µL of H2O, 5 µL first-strand synthesis buffer, and 50 µL of second-strand buffer, and the blunting reaction was then performed with the addition of dNTP mix and Pfu polymerase. b. For the EcoRI adapter ligation step, the aliquot of EcoRI adapter to be used was heated at 70°C for 2 min and allowed to return to room temperature gradually for optimal annealing. This was then followed by the low-temperature ligation step. c. For recovering the cDNA inserts after XhoI digestion, we found it a simple alternative to pass the cDNA through a 1% agarose gel, cut out cDNA species above 300 bp, electroelute, followed by phenol–chloroform extraction and ethanol precipitation. The quantity of the cDNA inserts in the final preparation should be estimated by ethidium bromide agarose gel method to allow setting up the ligation at optimal vector/insert ratio. This procedure can be found in the manual of the cDNA library construction kit (Stratagene). 3. The background may still be high at this point, which happens with occasional serum samples and usually cannot be removed by additional column absorption. This background usually improves significantly after one or two rounds of screening, or one may incubate the diluted serum overnight at room temperature with
SEREX Technique
215
mock experimental nitrocellulose membranes prepared from 15-cm NZYCM agar plates containing 5000–10,000 pfu wild-type λZAP Express bacteriophage. 4. Although it may be possible to isolate a pure positive phage clone at this stage, it is advisable to mix the positive clone with at least one negative phage clone. The negative clone will serve as an internal control for background staining on the same filter at secondary screening. Without such internal negative control on the same filter, it is often difficult to distinguish a weak positive clone from background, and one will likely miss weak-positive clones, or isolate false-positives. The latter would lead to an artificially high number of SEREX-defined antigens that are not truly immunogenic. 4. This step of identifying IgG-coding clones is obviously not necessary if the library is derived from a cell line unrelated to B-cells. For tissue-derived libraries, the frequency of IgG clones varies significantly, reflecting the numbers of B-cells in the specimen. If IgG clones predominate the positive clones at the primary screening, it would be useful to carry out the prescreening step with the secondary antibody first, so as to eliminate the IgG clones and incubate only the non-Ig-positive clones with serum. Occasionally the number of IgG clones is so high (>200 per nitrocellulose filter in our experience) that this protocol would be impossible to follow. For such cases, one can perform a prescreening of the primary filters with secondary antibodies linked to a different chromogen reaction (e.g., hydrogen peroxidase–DAB combination). After this step, the IgG clones on the filters can be marked, and the filters can be washed and one can then proceed to the serum incubation step. Alternatively, one may also try to subtract the Ig mRNA from the total RNA before cDNA library construction (4). This, however, often leads to substantial loss of the RNA, and is not advisable if the total RNA available is limited in quantity.
References 1. 1 Boon, T. and van der Bruggen, P. (1996) Human tumor antigens recognized by T lymphocytes. J. Exp. Med. 183, 725–729. 2. 2 Sahin, U., Tureci, O., Schmitt, H., et al. (1995) Human neoplasms elicit multiple specific immune responses in the autologous host. Proc. Natl. Acad. Sci. USA 92, 11810–11813. 3. Chen, Y.-T., Scanlan, M. J., Obata, Y., and Old, L. J. (2000) Identification of human tumor antigens by serological expression cloning (SEREX), in Biologic Therapy of Cancer (Rosenberg, S. A., eds.), Philadelphia: Lippincott Williams & Wilkins, pp. 557–570. 4. 4 Scanlan, M. J., Chen, Y. T., Williamson, B., et al. (1998) Characterization of human colon cancer antigens recognized by autologous antibodies. Int. J. Cancer 76, 652–658. 5. 5 Chen, Y. T., Gure, A. O., Tsang, S., et al. (1998) Identification of multiple cancer/ testis antigens by allogeneic antibody screening of a melanoma cell line library. Proc. Natl. Acad. Sci. USA 95, 6919–6923.
216
Chen et al.
6. 6 Chen, Y. T., Scanlan, M. J., Sahin, U., et al. (1997) A testicular antigen aberrantly expressed in human cancers detected by autologous antibody screening. Proc. Natl. Acad. Sci. USA 94, 1914–1918. 7. 7 Jager, D., Stockert, E., Gure, A. O., et al. (2001) Identification of a tissue-specific putative transcription factor in breast tissue by serological screening of a breast cancer library. Cancer Res. 61, 2055–2061. 8. Frederick, M., Ausubel, R. B., and Kingston, R. E. (1997) Preparation and analysis of RNA, in Current Protocols in Molecular Biology, vol. 1 (Chanda, V. B., ed.), New York: John Wiley & Sons, pp. 4.2.1–4.2.9.
Modeling Pancreatic Cancer in Animals
217
15 Modeling Pancreatic Cancer in Animals to Address Specific Hypotheses Paul J. Grippo and Eric P. Sandgren Summary Multiple experimental approaches have been employed to study exocrine pancreatic cancer, including the use of animals as surrogates for the human disease. Animals have the advantage that they can be manipulated to address specific hypotheses regarding mechanisms underlying this disease. Implicit in this opportunity is the necessity to match the question being asked with an appropriate animal model. Several approaches to modeling pancreatic cancer have been established that involve animals. First, xenogeneic cell transplantation, generally into immunocompromised rodent subcutis or pancreas, allows examination of (1) the effect of host environment on human or rodent pancreatic cancer cells, (2) whether specific genetic changes in donor cells correlate with certain cancer cell behaviors, and (3) novel approaches to cancer therapy or imaging of tumor growth. Second, carcinogen administration, typically to hamster or rat, allows examination of whether specific genetic, biochemical, cellular, and tissue phenotypic changes, including progression to neoplasia, accompany exposure to a particular chemical. Third, genetically engineered animals, usually transgenic or gene targeted mice, allow examination of (1) whether genetic changes, including oncogene overexpression/mutation or tumor suppressor gene loss, can increase the risk for neoplastic progression, (2) whether specific genetic changes can cooperate during pancreatic carcinogenesis, and (3) how the genetic signature of a neoplasm correlates with particular biological aspects of tumor initiation and progression. Collectively, these experimental approaches permit detailed exploration of pancreatic cancer genetics and biology in the whole animal context, thereby mimicking the environment in which human disease occurs. Key Words: Animal model; chemical carcinogen; embryonic stem cells; gene targeted mice; hamster; inducible transgene; Kras; mouse; oncogene; pancreatic cancer; rat; transgenic mice; tumor suppressor gene; xenogeneic cell transplantation.
1. Introduction Research in the fight against pancreatic cancer attempts both to understand the etiology of this disease and to identify specific molecular targets for therapeutic intervention. There are several ways to address these issues using animal models. From: Methods in Molecular Medicine, Vol. 103: Pancreatic Cancer: Methods and Protocols Edited by: G. Su © Humana Press Inc., Totowa, NJ
217
218
Grippo and Sandgren
Transplanting human pancreatic tumor tissue into an immunodeficient animal can provide an in vivo system for evaluating the activity of certain therapies against human cancer cells. Rodents exposed to carcinogens can be used to identify the molecular and cellular effects of carcinogenic chemicals. Transgenic and gene knockout models can be used to engineer in vivo the molecular events believed to contribute to pancreatic carcinogenesis. Each approach allows us to answer selected questions in pancreatic cancer research and to improve our understanding of the pathogenesis of this disease. This chapter correlates hypothesis testing to specific types of animal models. First, because the object of inducing pancreatic cancer in animals is to model the disease in humans, we briefly review phenotypic and molecular features of this disease. 1.1. Pancreatic Adenocarcinoma: A Molecular Model of the Disease In the United States, pancreatic adenocarcinoma is the fifth leading cause of cancer death (1), with a 3% 5-yr survival rate following diagnosis (2). Ductal adenocarcinoma accounts for 85% of all neoplasms in the pancreas (3,4). Metastases are common, particularly to the lymph nodes, liver, lung, adrenal glands, kidney, and stomach (5–7). Acinar cell carcinomas are present in fewer than 10% of patients (3). Although many genetic alterations have been identified in pancreatic adenocarcinoma, how these changes collaborate during initiation of the disease and its progression to metastasis remains incompletely understood. Some alterations may contribute causally to one or several aspects of disease progression, whereas others may not be relevant. Identifying causal alterations has become a major objective in this field. Based on data correlating genetic mutations with cellular alterations in human pancreatic cancer, a general model of molecular pathogenesis for this disease has been proposed (3,8). Early genetic changes in pancreatic adenocarcinoma cells include mutation of the K-ras gene (90–95%) and increased expression of Her-2/neu (9,10). Both alterations can be identified in early lesions and tumors and are implicated in the initiation of pancreatic carcinogenesis (8). Subsequent genetic changes include mutations in chromosome 9p (75%) (8–11), leading to inactivation of the p16 gene, and up-regulation of transforming growth factor-a (TGF-a) expression (12,13). Late events implicated in progression include alterations in chromosome 17p and 18q, leading to dysfunction or deletion of p53 and DPC (deleted in pancreatic cancer) (8–11,13). Other genetic, cellular, and tissue alterations almost certainly contribute causally to the development of pancreatic adenocarcinoma. 1.2. Animal Models of Pancreatic Cancer Animal model systems have been used to study specific molecular or cellular alterations thought to contribute to the development of human pancreatic
Modeling Pancreatic Cancer in Animals
219
adenocarcinoma. Transplantation of primary or cultured human pancreatic cancer cells into immunodeficient mice and rats has allowed investigators to examine how human tumor cells respond to mesenchymal tissue and to certain therapeutic regimens. Carcinogen-treated rodents provide models of how pancreatic epithelium responds to chemical exposure and can be used to identify genetic mutations associated with a specific carcinogen. Transgenic and gene-targeted mice provide models of pancreatic carcinogenesis in which the phenotypic effects of selected genetic changes can be evaluated. 2. Xenogeneic Cell Transplantation 2.1. Methodology Xenogeneic transplantation of human pancreatic cancer into immunocompromised mice presents a simple yet powerful approach to study human cancer in an in vivo mammalian system. Orthotopic transplants specifically involve isolation of either established pancreatic cancer cell lines or cells from primary neoplasms and injection of these cells into the appropriate site in immunocompromised animals, usually nude or severe combined immunodeficient (SCID) mice. Nude mice display defective development of the thymus and hair follicles, and thus lack T-lymphocytes and hair. SCID mice are defective in immunoglobulin gene and T-cell receptor gene rearrangements, and thus lack mature B- and T-lymphocytes. There are more than 25 published lines of cultured pancreas cells derived from human pancreatic adenocarcinomas (22 are reviewed in ref. 14). These cells have been derived from primary human tumors (including AcPC1 and PANC-1) (15) or derived from cells isolated at biopsy from tumors, adjacent tissue, or ascites fluid of pancreatic cancer patients (HuP-T1, T3, and T4 were started from cells present in ascites fluid of pancreatic cancer patients with peritonitis; 16). Cell lines typically display molecular alterations observed in primary pancreatic tumors, although adaptation to culture undoubtedly causes selection for additional changes. Tumor cells can be injected or transplanted subcutaneously, underneath a membrane (kidney capsule or abdominal wall), within the peritoneum, or as an orthotopic transplant directly into the pancreas (including primary pancreatic duct injections). Following transplantation, characteristics of tumor cell growth, invasion, and metastasis are identified. Orthotopic transplantation typically results in tumor growth in the pancreas, with only occasional tumor formation in the peritoneum, and has become the most popular approach because tumor phenotype and behavior most closely recapitulate the human disease. Orthotopic transplants are more difficult, as they require rodent surgery and hence greater risk to the animal. The simplicity of subcutaneous transplants is an advantage when seeking a rough evaluation of tumor growth capacity.
220
Grippo and Sandgren
Transplanted cells are allowed to grow for several weeks, and their ability to produce tumors and/or metastases is noted. This allows for the study of cellular interactions between tumor cells, stromal cells, and normal pancreatic cells in the appropriate tissue context (17–19). Furthermore, various therapies can be evaluated to identify their ability to target and destroy the cancer cells. 2.2. Questions Addressed 1. Does pancreatic cancer cell growth differ in distinct host environments? 2. Is there a correlation between specific genetic changes in tumor cells and the aggressiveness of tumor growth, specifically invasion and metastases? 3. Can tumor growth be monitored in vivo using imaging methods? 4. Will specific therapeutic regimens kill or inhibit human pancreatic cancer cells in vivo?
2.3. Representative Findings (see Table 1) For imaging studies, mice typically are implanted with tumor cells under the skin. Advanced imaging systems such as positron emission tomography (PET) have been used to monitor growth of transplanted tumors, especially in hairless rodents such as nudes, without euthanizing the animal. Other approaches include introducing the green flourescent protein (GFP) coding region into the pancreatic cancer cell line prior to transplant. To evaluate the effect of a specific genetic alteration, human pancreatic cancer cell lines can be genetically manipulated, then transplanted into immunocompormised rodents. Tumor phenotype is compared to that of the original unmodified cell line. Finally, gemcitabine, wortmannin, and several other drugs have been evaluated using orthotopically transplanted human tumors in rodents. 3. Carcinogen-Induced Cancer in Animal Models Animals exposed to carcinogenic chemicals provide an in vivo system with which to identify the molecular and cellular changes induced by chemicals that may cause pancreatic cancer. Hamsters and rats respond to certain carcinogens by developing exocrine pancreatic lesions, including neoplasms. Mouse pancreas appears resistant to most carcinogens. 3.1. Methodology Carcinogens can be administered systemically via injection or gavage or targeted by infusion into the primary pancreatic duct or by surgical implantation of a crystal or solid pellet into pancreatic parenchyma. Hamsters and rats are used most frequently with these techniques. 3.2. Questions Addressed 1. Can a specific chemical induce neoplasia in the pancreas?
Modeling Pancreatic Cancer in Animals
221
2. What genetic, biochemical, cellular, and tissue phenotypic changes are associated with exposure to a particular chemical, and how do these changes correlate with subsequent development of neoplasia?
3.3. Representative Findings 3.3.1. Hamsters An excellent model of pancreatic ductal adenocarcinoma is produced by administration of either N-nitroso-bis-(2-oxypropyl) amine (BOP) or N-nitrosobis-(2-hydroxypropyl) amine (BHP) to hamsters (20,21). BOP and BHP are strong alkylating agents that promote O6 methylation of guanine in DNA (22,23). DNA adducts induce nucleotide mutations, and the K-ras gene is one preferential target (24). Mutations in the p53 allele are not seen in primary tumors, although they have been observed in transplantable tumor tissue and cell cultures derived from primary tumors (25,26). Hamsters are administered 250 mg/kg of BHP by subcutaneous injection once a week for up to 44 wk. After 9 wk of treatment, a few abnormal acini are evident. Six weeks later, pseudoductules and small cystic foci are evident. The two primary lesion types are (1) cystic ductal complexes that contain luminal spaces lined by cuboidal epithelial cells surrounded by fibrous tissue, and (2) tubular ductal complexes that contain ductular structures composed of cuboidal and columnar epithelia (25–29). Finally, cystadenomas and ductal adenocarcinomas develop (23). BHP-treated hamsters present both the K-ras mutation and a tumor/lesion morphology that is similar to human pancreatic cancer. BHP-treated hamsters also develop primary liver and lung tumors (30). This model has been used to address the cell type of origin of pancreatic adenocarcinoma, and both ductal and islet epithelium have been implicated (31,32). 3.3.2. Rats The most common model in rats is asazerine-induced pancreatic cancer. Asazerine is an alkylating carcinogen that damages DNA. Unrepaired DNA adducts induce mutations, and neoplasia develops almost exclusively in acinar cells. K-ras mutations have not been identified in these neoplasms (29,33). A second rat model involves surgically implanting 2–3 mg of crystalline 7,12dimethylbenz[a]anthracene (DMBA) into the head of the pancreas. DMBA has been used extensively as a tumor promoter in mouse skin and liver, and functions by forming adenosine/guanine DNA adducts (30,34,35). Local DMBA treatment induced the development of ductular to glandular adenocarcinomas of the pancreas in 80% of DMBA-treated rats with a mean latency of 6.5 mo and without development of primary tumors in other tissues. These pancreatic tumors are accompanied by ductal proliferation and atypia and are locally invasive and metastatic to liver and intestine. Another report described similar lesions
222
Table 1 Applications of Orthotopic Transplantation Related to Pancreatic Research
222
Tumor type Treatment
Host
Results
Pancreatic cancer cell lines
Tumor fragments into pancreas
Nude mouse
Resection and adjuvant gemcitabine
Nude mouse
Directly into proximal part of pancreas
SCID mouse
Implanted into pancreas or abdominal wall
SCID mice
Overexpression of TIMP-1
Nude mouse
Intratumoral injection of p202/SN2 complex
Nude mouse
Immunize with T-cells from donor mice Injected into pancreas or spleen
C57BLl/6 MUC1.Tg Nude mouse
Daily oral doses of PKI 166
Nude mouse
Implanted into head of the pancreas
Nude mouse
PKC-a antisense oligonucleotide
SyrianGolden hamster SyrianGolden hamster Nude rat
Realtime optical imagining of primary and metastatic tumor events Reduced recurrent and metastastic events versus control mice Recapitulation of the metastatic process evident in humans Identify metastasis-associated genes Orthotopic tumors grew faster than the ectopic tumors Higher levels of tumor-dervied VEGF protein Attenuate tumor growth, decrease metastasis, and inhibit angiogenesis Reduced expression of angiogenic markers following p202 treatment Detection of immune responses to muc1 and moderate protection against tumor challenge Increased cell motility and expression of MMP9 Reduced expression of E-cadherin Demonstrate decreased microvessel density; increased tumor cell and endothelial cell apoptosis; correlated with therapeutic success Up-regulated PKC-a activity Doubling of the mortality rate Down-regulation of PKC-a activity Increased survival 75% of surgically resected hamsters cured Nearly 50% tumor free after over a year Increased infiltration of NK cells and macrophages Increased tumor uptake of CEA antibody Reduced volume of pancreatic tumor and angiogenesis Increased apoptotic rate
Mab against CA19-9-CVF Oral administration of MMI-166
SyrianGolden hamster
(78) (79) (80) (81) (82) (83) (84) (77) (85) (86) (87) (87) (88) (89) (90)
Grippo and Sandgren
Surgical resection of pancreatic tumor
Reference
Primary human pancreatic tumors
IP injection of BB94 (MMP inhibitor) or BB94 + A.4.6.1 (VEGF inhibitor)
IP injection of AG3340 (MMP inhibitor)
Pancreas implant
223
Subcutaneous or orthotopic implantation in the pancreas Implanted into body of pancreas Primary rodent tumors
Pancreas implant
Pancreas implant Orthotopic implantation auristatin-PE, gemcitabine, or both
Inhibition of PKB/Akt phosphorylation by wortmannin (91) fivefold increase in apoptosis with both versus each one or negative control Increased survival and reduced metastasis; (92) Nude mouse lower MMP activity (93) Additive effects in reduced tumor size, metastasis, and angiogenesis Reduced primary tumor volume, invasion, and metastasis; (94) SCID mouse induced necrosis, differentiation, and fibrosis in the tumors and inhibited angiogenesis Nude mouse Extensive local tumor growth with metastasis to the (95) stomach, duodenum, liver, adrenal gland, diaphragm, and lymph nodes Site of administration and source of cancer tissue affects (96) Nude mouse tumor growth and apoptosis Athymic mouse VEGF protein levels were elevated in malignant ascites by (97) 15-fold compared to control mice with equivalent tumors SyrianGolden Tumor tissue transplants had direct invasion to (98) hamster adjacent organs with a higher rate of liver metastasis versus cell implantation Immuncomp Production of moderately to well-differentiated ductal (99) Lewis rat tumors surrounded by dense stroma (100) SCID mice Reduced tumor growth following combination therapy as analyzed by MRI visualization of tumors SCID mouse
Modeling Pancreatic Cancer in Animals
Pancreatic cancer cell lines (cont’d) IV wortmannin and gemcitabine
223
224
Grippo and Sandgren
in rats in which the pancreatic duct was infused with a DMBA solution, leading to ductal adenocarcinoma in 50% of treated rats (30). These lesions also have not been associated with a mutation in K-ras. 4. Genetically Engineered Animal Models 4.1. Transgenic Mice Mice are resistant to the carcinogen treatments used in hamsters and rats. However, transgenic approaches have proven effective at inducing exocrine pancreatic neoplasia in mice. 4.1.1. Methodology Generating transgenic mice requires introduction of a DNA transgene (including promoter, coding region, polyadenylation signal, and possibly other regulatory elements) into a fertilized mouse embryo (Fig. 1). The transgene will integrate into genomic DNA in a fraction of the injected eggs. Potential transgene-positive founders and their offspring can be screened for possession of the transgene by evaluating genomic DNA. These techniques are labor intensive but have become standard in many research institutions and have been extended to rats. The critical issue in generating transgenic animals is selection of both coding sequence and gene regulatory element(s). The coding sequence is selected to encode the gene whose product is to be evaluated as a potential causative agent in pancreatic cancer. The gene regulatory element is selected to target expression of the coding region to the desired cell type. The resulting transgenic mice are used to demonstrate molecular, cellular, and whole animal effects of the transgeneexpressed protein. Furthermore, mice carrying a single transgene can be mated with other transgenic mice to generate bi- and tritransgenic mice as a means to evaluate transgene interactions. Inducible transgenic systems permit temporal regulation of transgene expression (Fig. 2) (36–38). Recombination transgenic systems can employ two separate gene regulatory elements to drive expression of a transgene in a subset of cells that express both of the above genes (Fig. 3) (39–41). Finally, transgenic mice can be engineered to inactivate specific gene products by targeting expression of dominant negative proteins. 4.1.2. The First Generation: EL-TAg, EL-H-ras, EL-myc, EL-TGF-a 4.1.2.1. DESIGN
The first series of four transgenic models of pancreatic neoplasia that were developed includes EL-mutant H-ras, EL-SV40 TAg, EL-myc, and EL-TGF-a. These models took advantage of cloned DNA that contained gene regulatory elements from the rat pancreatic elastase (EL) gene (42,43). A brief description of these transgenic mice will illustrate how genetically engineered mouse
Modeling Pancreatic Cancer in Animals
225
Fig. 1. Transgene construction. (A) The elastase promoter (EL) flanked by metallothionein (MT) gene loci that contain hypersensitive sites (hs) plus the human growth hormone (hGH) polyadenylation signal [poly(A)]. The hs act as locus control regions and enhance transgene expression. Coding regions of choice can be inserted between the promoter and the poly(A) site. Those listed have been generated. (B) Regulatory elements from genes expressed in ductal epithelium such as cytokeratin 19 (CK19) or Mucin 1 (Muc1) have been engineered into a transgene cassette for expression of various coding regions. Human placental alkaline phosphatase (hPAP) has been used as a marker to identify cell types that activate expression of these gene regulatory elements. However, without pancreas specificity, multiple organs are targeted. (C) Regulatory elements from genes expressed in pancreatic progenitor cells (such as Pdx-1) can be used to target transgene expression.
models can address specific questions. In these studies, the primary goal was to investigate the cellular and tissue effects of overexpression of oncogenes or growth factors in acinar cells. Each coding sequence had some relevance to the human disease. TGF-a is overexpressed by many pancreatic neoplasms, H-ras shares extensive homology with K-ras, the viral TAg inhibits cellular p53 and
226
Grippo and Sandgren
Fig. 2. Tetracycline-inducible transgenes. An example of the tetracycline inducible system for targeting coding regions to pancreatic acinar cells. Inducible systems could be used to target pancreatic ductal and progenitor cells in a similar manner, as long as appropriate gene regulatory elements are available. Two transgene constructs are needed for this system to function, including: (A) Cell-specific expression of the effector transgene. In this example, the tetracycline transactivator (tTA) is expressed in acinar cells using the elastase promoter. (B) To identify which cell type is targeted and how effectively the system works, the target transgene expresses hPAP in the presence of the transactivator protein. (C) Coding regions such as mutant K-ras can be used in place of hPAP to target potentially oncogenic genes to pancreatic acinar cells.
Modeling Pancreatic Cancer in Animals
227
227
Fig. 3. Cre-Lox recombinant transgenes. Examples of the cre/lox recombinant system for targeting knockout mutations to pancreatic acinar cells are shown at the top. This and similar recombinant systems could be used to target pancreatic duct and progenitor cells. (A) The effector transgene is designed to produce cell-specific expression of cre. In this example, cre is expressed in pancreatic acinar cells using the elastase promoter. (B) The target is designed by flanking an endogenous target gene locus with Lox sites for subsequent cre-mediated recombination and loss of the gene locus. This requires use of gene targeting vectors for subsequent generation of genetically modified mice by ES cell blastocyst injection (see Fig. 4). (C) For this example, the promoter used in the effector transgene is from the cytokeratin 19 gene (CK19). (D) The target transgene employs a gene promoter distinct from that of the effector transgene. Recombination will occur in cells that express the gene whose promoter regulates the effector transgene (CK19). Recombinant transgene expression will be limited to cells that express the gene whose promoter regulates the target transgene (Muc1). In this way, the desired coding sequence (K-ras) will be expressed exclusively in the subset of cells that express both genes. In the example above, expression of K-ras will be limited to a subset of cells that express both CK19 and Mucin 1.
228
Grippo and Sandgren
Rb function, and c-myc may be up-regulated by altered b-catenin signaling and also is overexpressed by some pancreatic adenocarcinomas. 4.1.2.2. QUESTIONS ADDRESSED 1. Will overexpression of specific oncogenes and/or growth factors in exocrine pancreas cells increase the incidence of pancreatic cancer? 2. Will animal models of pancreatic cancer generated via transgene technology provide information about the biological basis of pancreatic cancer intiation and progression associated with specific genetic changes?
4.1.2.3. REPRESENTATIVE FINDINGS
Most human pancreatic cancers have a ductal histotype. However, ductal cells may not be the only cell of origin. A growing body of research suggests that a stem, islet, and/or acinar cell may be involved in pancreatic ductal carcinogenesis (23,44–49). Furthermore, acinar cell neoplasms develop in about 10% of patients with pancreatic cancer. Hence, use of EL gene regulatory elements served two purposes: (1) to generate models of acinar neoplasia and (2) to test whether altered gene expression in acinar cells in vivo can lead to ductal neoplasia. Phenotypic analysis of these transgenic mice demonstrated the following. First, mutant H-ras induced diffuse perinatal exocrine pancreatic hyperplasia and dysplasia, resulting in death of 90% of founder mice within several days of birth (50). The surviving EL-H-ras mice demonstrated acinar cell atrophy and occasional acinar cell carcinomas in very old mice (E. P. S., unpublished). Thus, mutant H-ras provides a potent growth stimulatory signal in developing pancreas but may have a less potent affect in adult pancreas. Second, EL-TAg mice developed acinar cell hyperplasia and acinar cell carcinomas by 3–7 mo of age, indicating that TAg-associated loss of tumor suppressor function (together with other TAg effects) is potently oncogenic (51). Third, EL-TGF-a mice developed severe pancreatic fibrosis, acinar tubular complexes, and acinar to ductal metaplasia (52). A small fraction of old mice developed neoplasms with acinar and ductal components. The conversion from acinar to ductal cells was accompanied by expression of ductal cell markers and led to the formation of tubular structures, and then malignant pancreatic tumors (49). Thus, although TGF-a increases slightly the risk of pancreatic cancer, its effects on tissue architecture are far more prominent. Finally, EL-myc transgenic mice developed acinar cell carcinomas and mixed carcinomas that displayed neoplastic acinar and neoplastic ductal cell components with accompanying fibrosis (53). The ductal components contained cells expressing the duct-specific antigen cytokeratin 19 (CK19) (53). The quiescent nature of normal ductal cells in these mice (<1% of cells undergoing DNA synthesis) suggested that primary ductal cells were not the source of duct-like cells in mixed neoplasms. Furthermore, focal collec-
Modeling Pancreatic Cancer in Animals
229
tions of amylase-positive, acinar-like cells in some neoplasms display induction of CK19 gene expression (Grippo and Sandgren, submitted). These data support suggestions that pancreatic carcinogenesis can be accompanied by acinarto-ductal metaplasia in vivo. Perhaps the most striking finding associated with studies of these four transgenic models is the diversity of lesion phenotype and behavior following expression of the different transgenes, illustrating the variable response of the pancreas to different genetic insults. These mice also provide tools that can be used to test experimental therapeutic regimens. 4.1.3. The Next Generation of Transgenic Mice There remains a need to generate additional models of pancreatic cancer, in particular by employing the most commonly identified genetic changes identified in human pancreatic adenocarcinoma and by targeting ductal epithelium. 4.1.3.1. MUTANT K-RAS TRANSGENICS 4.1.3.1.1. Design. Several recently developed transgenic mouse models have focused on the most prevalent genetic event associated with human pancreatic adenocarcinoma—mutation of the K-ras gene. Most K-ras mutations occur at codon 12, with a glycine to aspartate substitution. Several promoter elements have been employed in these studies, including EL and Mist-1, a transcription factor identified early in developing acinar cells. In view of the acinar-to-ductal metaplasia evident in EL-myc transgenic mouse pancreas, acinar cell targeting also was employed to attempt to induce ductal lesions. 4.1.3.1.2. Question Addressed. Will mutant K-ras expression in acinar cells lead to ductal carcinoma? 4.1.3.1.3. Representative Findings. Although most founder mice expressing the EL-K-ras transgene displayed perinatal pancreatic growth defects, two survived (54). Among offspring of one of those founders, nearly half developed pancreatic ductal carcinoma in situ at a year or more of age, often in association with increased stroma and pancreatic atrophy. Interestingly, the lesions were neither invasive nor metastastic, suggesting that K-ras mutation must collaborate with one or more additional genetic alterations to bring about preneoplastic progression. As in EL-myc transgenic mice, lesion pathogenesis appears to involve acinar to ductal metaplasia. However, unlike EL-myc mice, ductal neoplastic cells do not maintain a prominent association with neoplastic acinarlike cells. These findings strengthen the association between K-ras mutation and early preneoplastic development. Development of acinar metaplasia and pancreatic cancer of mixed histotype also is observed in the Mist1-K-ras knockin mice, in which Mist1 directs mutant K-ras expression to mature pancreatic acinar cells and to pancreatic precursor cells during development. Furthermore, tumor invasion and metastasis were observed in the Mist1-K-ras mice (Tuveson
230
Grippo and Sandgren
et al, personal communication). These models indicate that, although other genetic events probably contribute to the development of carcinoma in situ or cancer in these models, mutant K-ras can initiate mouse pancreatic carcinogenesis. 4.1.3.2. MUCIN
AND
CCKR TRANSGENICS
4.1.3.2.1. Design. Additional genes have been targeted to pancreatic epithelium, including Mucin 1 and CCKR. The intact human Mucin 1 gene was inserted into the mouse genome to generate expression of human Mucin 1 in multiple epithelial cells, including pancreatic epithelium (55). Overexpression of the gastrin/ CCK receptor (EL-CCKR and Elas-CCK2) was targeted to acinar cells. CCKR expression is increased locally in human pancreatic cancer (56), but it is unclear if a concomitant increase in CCK ligand exists (56–58). However, CCKR is increased in several rodent models of pancreatic cancer relative to normal pancreas, and there have been suggestions that increased CCKR signaling in both rodents and humans can enhance pancreatic tumor cell growth. 4.1.3.2.2. Questions Addressed 1. Can Mucin 1 protein serve as a target for immunotherapy in pancreatic neoplasms? 2. Does overexpression of CCKR in pancreatic epithelium cause abnormal acinar cell growth and lesion development?
4.1.3.2.3. Representative Findings. The goal of the Mucin 1 model was to assess mucin as an immunotherapeutic target. Human Mucin 1 expression did not induce pancreatic abnormalities. However, when Muc-1 transgenic mice were crossed with EL-TAg mice, which develop multiple pancreatic acinar carcinomas within several months of birth, bitransgenic mouse tumors expressing Muc-1 could be recognized by Muc-1-specific immunobased therapies, causing a delay in lesion progression (59). The goal of the CCKR models was to determine if increased gastrin receptor signaling would increase pancreatic growth and/or influence the development of neoplasia. Mice overexpressing CCKR displayed increased pancreas weight. In one model, no lesions were evident (61), whereas in the other model, homozygous transgenic mice developed occasional focal, small ductal lesions and, in a few mice, developed tumors from an acinar–ductal carcinoma sequence (61). These findings demonstrate that CCKR signaling enhances pancreatic growth, and support suggestions that it may influence the development of cancer. These mice also provide a model for therapies designed to inhibit signaling via this pathway. 4.1.3.3. DUCTAL TARGETING
4.1.3.3.1. Design. Acinar cells have been targeted frequently by oncogenic transgenes. Attempts at targeting pancreatic duct cells have been limited by the
Modeling Pancreatic Cancer in Animals
231
lack of a specific pancreatic duct cell gene regulatory element. Several studies have targeted ductal epithelium throughout the body using more widely expressed gene regulatory elements, such as cytokeratin 19 (CK19). 4.1.3.3.2. Question Addressed. Will targeting of oncogenes to pancreatic duct cells induce ductal adenocarcinoma reminiscent of the human disease? 4.1.3.3.3. Representative Findings. Human CK19 gene regulatory elements (hCK19) fused to the coding region of a marker gene (human placental alkaline phosphatase [hPAP]) targets expression to simple epithelium (including ducts) in the pancreas, kidney, liver, and mammary gland and to urothelial cells in the bladder (62). Mouse cytokeratin 19 gene regulatory elements (mCK19) were fused to the b-galactosidase gene, and this transgene was introduced into transgenic mice. Transgene expression was observed in pancreatic ductal cells but not acinar or islet cells (Rustgi et al., personal communication). hCK19-TAg transgenic mice provided a model of highly invasive transitional cell carcinoma in urinary bladder (62). This finding illustrates the limitation of using gene regulatory elements that target multiple cell types: lesions develop in the most susceptible cell type, not necessarily the desired cell type. The targeting of mutant K-ras with the mCK19 in transgenic mice leads to pancreatic ductal hyperplasia in some mice that is accompanied by a periductal lymphocytic infiltrate (63). This mouse model may serve as a platform for investigating the effects of additional genetic alterations. Similar studies using the Mucin 1 gene promoter are in progress. 4.1.3.4. FUTURE TRANSGENICS
Several cell targeting and technological advances have been added to the repertoire of pancreatic researchers, and these promise to improve our understanding of pancreatic carcinogenesis and provide models of the disease that more closely resemble the human disease. First, new targeting approaches direct expression to progenitor or stem cells in the pancreas. Pdx-1, a transcription factor expressed early during pancreatic development, is being used to target stem cells that give rise to both endocrine and exocrine tissue (64,65). These studies hold great promise for identifying signaling pathways that affect pancreatic differentiation and, by extension, may be active during metaplasia that occurs in injured or cancerous pancreas in the adult (66). Second, inducible transgene systems are being employed that allow manipulation of the timing of transgene expression (Fig. 2) (67). These systems can provide temporal regulation of mutant K-ras expression in mouse pancreas to address the age-related susceptibility to transformation and to identify whether continuous oncogene expression is required for lesion maintenance. Third, recombinant systems employing cre/lox or frt/flp allow targeted deletion of endogenous DNA strictly in pancreas cells (see Fig. 3 and tumor suppressor section below). Any cell type to which a transgene can be targeted can be manipulated in these ways.
232
Grippo and Sandgren
4.2. Tumor Suppressor Gene Knockouts 4.2.1. Design Loss of tumor suppressor gene function can be modeled by targeted deletion of a portion of a gene locus using embryonic stem (ES) cells (Fig. 4). Recent advances in this approach include use of the cre/lox recombination system, in which DNA flanked by lox sites are deleted by a cre-mediated recombination event (Fig. 3). These approaches are used to mimic, at the whole animal level or in specific tissues, the heterozygous or homozygous loss of specific tumor suppressor genes that typically occur in cancer cells. 4.2.2. Methodology The approach used for knockout technology involves recombination between a DNA vector typically containing a marker gene (such as b-gal) and an antibiotic resistance gene (such as neomycin) with a genomic target. The antibiotic resistance gene will serve to select for ES cell clones that carry the introduced DNA, and the marker allows these cells to be identified. The herpes simplex virus thymidine kinase gene usually is used as a second selectable marker, which kills cells (following exposure to the substrate gancyclovir) that have integrated the introduced DNA without homologous recombination, in which case the TK gene remains with the insert. Only cells that have gone through homologous recombination, which should delete the TK gene, will survive. Thus, ES cells resistant to both neomycin and gancyclovir are candidates for successful vector insertion. Further screening must be done to verify that the correct allele has been disrupted. ES cells that demonstrate loss of the targeted gene are injected into mouse blastocysts, which then are implanted into pseudopregnant mice for the generation of chimeric mice that contain the gene-targeted ES cells (Fig. 4). 4.2.3. Questions Addressed 1. Will homozygous or heterozygous loss of a specific tumor suppressor gene increase the incidence of pancreatic cancer? 2. Will tumor suppressor gene loss cooperate with oncogene expression to enhance pancreatic carcinogenesis?
4.2.4. Tumor Suppressor Gene-Targeted Mice Three tumor suppressor genes (TSG) often are altered in pancreatic cancer: p53, p16, and DPC. Currently, there are mice that model loss of each of these genes. p53 -/+ mice develop normally, but show a higher frequency of certain tissue abnormalities and tumors. p53 -/- mice develop a variety of tissue abnormalities and tumors, but none specific to the pancreas (68). Loss of one or both p16 allele(s) leads to greater susceptibility to tumor induction as demonstrated
Modeling Pancreatic Cancer in Animals
233
Fig. 4. Creating chimeric mice. Embryonic stem (ES) cells are isolated from a mouse with an agouti coat color (A) and manipulated to carry a targeted mutation in one chromosomal gene. The ES cells then are injected into blastocysts from a mouse with a black coat color (B). The blastocysts are implanted into a pseudopregnant recipient (C). Litters born to these mice should contain a few chimeric mice (D) that have some cells carrying the mutated ES cell gene and some cells carrying the wild-type host gene. Chimeric mice have a conspicuous, mixed coat color, which indicates that the ES cells have survived in the embryo. Chimeric mice are mated (E), and progeny are screened for evidence of the targeted mutation in the gene of interest. Heterozygous males and females (F) can be crossed to generate homozygous offspring that lack a functional gene (G).
234
Grippo and Sandgren
in the skin following DMBA and UVB treatment, with a more rapid onset in homozygous p16 null mice. Spontaneous soft tissue sarcomas or lymphomas did develop in homozygous p16 null mice at 5–9 mo of age, but no neoplasia was reported in the pancreas (69). Homozygous loss of DPC (Smad4 -/-) leads to death of the mice at E7.5 with lack of gastrulation and endoderm development. With loss of only one allele, Smad4 +/- mice develop intestinal polps within one year of age (70). None of these mice develop pancreatic abnormalities. Furthermore, combined heterozygous deficiency of p53 and p16 also did not induce pancreatic lesions, although it increased the incidence of neoplasia in other cell types (71). 4.2.5. Combinatorial Tumor Suppressor Gene-Targeted Mice Since pancreatic cancer cells have a high frequency of loss in p53, p16, and DPC (Smad4), the complete loss of two or all three might increase the likelihood of lesion development. When p16 homozygous null mice were crossed with p53 homozygous nulls, double null/null mutants displayed a higher frequency and broader spectrum of tumors compared to mice carrying any loss of p53 (+/- or -/-) in either p16 wild-type or p16 heterozygous null background. Even so, there were no reported abnormalities in the pancreas (71). These models demonstrate a limitation associated with knockout technology: gene deletion is present in every cell of the mouse. This is not true for most cancers, in which only cancer cells lose TSG function. However, cell-specific loss can be induced using cre/lox conditional knockout technology (Smad4 conditional knockouts have been generated; see 72). 5. An Integrative Approach to Animal Modeling 5.1. Transgenic Gene-Targeted Mice Crossing transgenic mice onto tumor suppressor null backgrounds can be used to generate improved mouse models of pancreatic cancer. Mice that overexpress TGF-a in a p53-/+ background develop aggressive, metastatic pancreatic neoplasia, despite the inefficient induction of pancreatic cancer by either genetic alteration alone, and maintain the acinar-to ductal metaplasia originally observed in EL-TGF-a mice (73). TGF-a transgenic mice deficient in both p16 and p53 develop serous cystadenomas (SCA) that recapitulate both the phenotypic and molecular characteristics observed in human SCA (74). With the addition of inducible transgene methodology, it becomes possible to generate mice that simultaneously gain expression of a mutated oncogene and lose functional activity of tumor suppressor gene products in the same cell type. For example, inducible targeting of mutant K-ras to specific pancreatic cells together with deletion of p16 and p53 should allow establishment of a
Modeling Pancreatic Cancer in Animals
235
mouse model with molecular characteristics that closely recapitulate the human disease. Furthermore, multiple inducible systems (tetracycline and estradiol) and conditional recombination systems (cre/lox and flp/frt) can be used together to model complex genomic changes in pancreatic cells. The ultimate strength of models of this type is the opportunity they provide to simultaneously or sequentially regulate specific oncogene or growth factor expression and loss of tumor suppressor gene function exclusively in the pancreas (see Fig. 5). (These issues are reviewed in ref. 37). 5.2. Carcinogen/Cancer Cell Administration to Genetically Modified Mice Investigators also can combine chemical treatment or cell transplantation with genetic manipulation of the target mouse. The primary phenotype of Smad4 -/+/ APCMin/+ mice is the development of intestinal polyps and colorectal carcinoma (75,76). When administered N-nitroso-N-methyl urea (NMU), these mice displayed increased frequency of acinar dysplasia and cystic structures in the pancreas with greater b-catenin expression relative to either manipulation alone. Thus, Smad4 loss can predispose for pancreatic lesion development, perhaps through interactions of the TGF-b and wnt signaling pathways (75). In a second model, Panc02 pancreatic tumor cells (syngenic to the C57BL/6 mouse) were transfected with a mucin 1-expressing transgene. These cancer cells were transplanted orthotopically into the pancreatic gastric lobe of Muc1 transgenic and wild-type C57BL/6 mice. Adoptive transfer of T-lymphocytes (generated from the spleen and lymph nodes of wild-type C57BL/6 mice challenged with Panc02/Muc1) protected wild type mice and prolonged the survival of Muc1 mice against orthotopic Panc02/Muc1 tumor cell challenge (77). 6. Summary Animal models of pancreatic cancer have created many opportunities to address a variety of scientific and medical questions. With better technology and more detailed genomic information, the integrative approach to modeling pancreatic cancer in animals allows us to study causal and sequential genetic events involved in pancreatic carcinogenesis in a stepwise fashion. Thus, mouse models provide highly specific data regarding the combined effects of environmental, cellular, and genetic factors on pancreatic carcinogenesis. However, as we develop these increasingly sophisticated methods to model human pancreatic cancer, it remains important to choose a model that is the most appropriate to address our specific scientific question. In view of the many unresolved issues that remain in this field, it is not likely that any single model will be appropriate or adequate to address all of them. Hence, the development of new animal models will remain an important component of future pancreatic cancer research.
236
236 Grippo and Sandgren
Fig. 5. Summary of animal models of pancreatic cancer.
Modeling Pancreatic Cancer in Animals
237
References 1. 1 Sohn, T. A. (2002) The molecular genetics of pancreatic ductal carcinoma. Minerva Chir. 57, 561–574. 2. Parker, S., Tong, T., Bolden, S., et. al. (1996) Cancer Statistics. CA J. Clin. 65, 5–27. 3. 3 Sakorafas, G. H. and Tsiotou, A. G. (1999) Multi-step pancreatic carcinogenesis and its clinical implications. Eur. J. Surg. Oncol. 25, 562–565. 4. 4 Hilgers, W. and Kern, S. E. (1999) Molecular genetic basis of pancreatic adenocarcinoma. Genes Chromosom. Cancer 26, 1–12. 5. 5 Caldas, C. and Kern, S. E. (1995) K-ras mutation and pancreatic adenocarcinoma. Int. J. Pancreatol. 18, 1–6. 6. 6 Gunji, N., Oda, T., Todoroki, T., et al. (1998) Pancreatic carcinoma: Correlation between E-cadherin and alpha-catenin expression status and liver metastasis. Cancer 82, 1649–1656. 7. 7 Visser, C. J., Bruggink, A. H., Korc, M., et al. (1996) Overexpression of transforming growth factor-alpha and epidermal growth factor receptor, but not epidermal growth factor, in exocrine pancreatic tumours in hamsters. Carcinogenesis 17, 779–785. 8. 8 Hruban, R. H. W. and Kern, S. E. (2000) Genetic progression in the pancreatic ducts. Am. J. Pathol. 156, 1821–1825. 9. 9 Mangray, S. and King, T. C. (1998) Molecular pathobiology of pancreatic adenocarcinoma. Front. Biosci. 3, D1148–1160. 10. Perugini, R. A., McDade, T. P., Vittimberga, F. J. Jr., and Callery, M. P. (1998) The molecular and cellular biology of pancreatic cancer. Crit. Rev. Eukaryot. Gene Express. 8, 377–393. 11. 11 Goggins, M., Kern, S. E., Offerhaus, J. A., and Hruban, R. H. (1999) Progress in cancer genetics: Lessons from pancreatic cancer. Ann. Oncol. 10, 4–8. 12. 12 Kumar, V., Bustin, S. A., and McKay, I. A. (1995) Transforming growth factor alpha. Cell Biol. Int. 19, 373–388. 13. 13 Sakorafas, G. H., Tsiotou, A. G., and Tsiotos, G. G. (2000) Molecular biology of pancreatic cancer; oncogenes, tumour suppressor genes, growth factors, and their receptors from a clinical perspective. Cancer Treat. Rev. 26, 29–52. 14. 14 Moore, P. S., Sipos, B., Orlandini, S., et al. (2001) Genetic profile of 22 pancreatic carcinoma cell lines. Analysis of K- ras, p53, p16 and DPC4/Smad4. Virchows Arch. 439, 798–802. 15. 15 Aoki, K., Yoshida, T., Matsumoto, N., Ide, H., Sugimura, T., and Terada, M. (1997) Suppression of Ki-ras p21 levels leading to growth inhibition of pancreatic cancer cell lines with Ki-ras mutation but not those without Ki-ras mutation. Mol. Carcinog. 20, 251–258. 16. 16 Kita, K., Saito, S., Morioka, C. Y., and Watanabe, A. (1999) Growth inhibition of human pancreatic cancer cell lines by anti-sense oligonucleotides specific to mutated K-ras genes. Int. J. Cancer 80, 553–558. 17. 17 Hahn, S. A., Seymour, A. B., Hoque, A. T., et al. (1995) Allelotype of pancreatic adenocarcinoma using xenograft enrichment. Cancer Res. 55, 4670–4675.
238
Grippo and Sandgren
18. 18 Capella, G., Farre, L., Villanueva, A., et al. (1999) Orthotopic models of human pancreatic cancer. Ann. NY Acad. Sci. 880, 103–109. 19. Schwarz, R. E., McCarty, T. M., Peralta, E. A., Diamond, D. J., and Ellenhorn, 19 J. D. (1999) An orthotopic in vivo model of human pancreatic cancer. Surgery 126, 562–567. 20. Gingell, R., Wallcave, L., Nagel, D., Kupper, R., and Pour, P. (1976) Metabolism 20 of the pancreatic carcinogens N-nitroso-bis(2-oxopropyl)amine and N-nitroso-bis (2-hydroxypropyl)amine in the Syrian hamster. J. Natl. Cancer Inst. 57, 1175–1178. 21. 21 Pour, P. and Althoff, J. (1977) The effect of N-nitrosobis(2-oxopropyl)amine after oral administration to hamsters. Cancer Lett. 2, 323–326. 22. Konishi, Y., Tsutsumi, M., and Tsujiuchi, T. (1998) Mechanistic analysis of pan22 creatic ductal carcinogenesis in hamsters. Pancreas 16, 300–306. 23. 23 Flaks, B., Moore, M. A., and Flaks, A. (1982) Ultrastructural analysis of pancreatic carcinogenesis. V. Changes in differentiation of acinar cells during chronic treatment with N-nitrosobis(2-hydroxypropyl)amine. Carcinogenesis 3, 485–498. 24. Feng, Z., Hu, W., Chen, J. X., et al. (2002) Preferential DNA damage and poor 24 repair determine ras gene mutational hotspot in human cancer. J. Natl. Cancer Inst. 94, 1527–1536. 25. Okita, S., Tsutsumi, M., Onji, M., and Konishi, Y. (1995) p53 mutation without 25 allelic loss and absence of mdm-2 amplification in a transplantable hamster pancreatic ductal adenocarcinoma and derived cell lines but not primary ductal adenocarcinomas in hamsters. Mol. Carcinog. 13, 266–271. 26. Erill, N., Cuatrecasas, M., Sancho, F. J., et al. (1996) K-ras and p53 mutations in 26 hamster pancreatic ductal adenocarcinomas and cell lines. Am. J. Pathol. 149, 1333–1339. 27. Flaks, B., Moore, M. A., and Flaks, A. (1981) Ultrastructural analysis of pancre27 atic carcinogenesis. IV. Pseudoductular transformation of acini in the hamster pancreas during N-nitroso-bis(2-hydroxypropyl)amine carcinogenesis. Carcinogenesis 2, 1241–1253. 28. 28 Flaks, B., Moore, M. A., and Flaks, A. (1982) Ultrastructural analysis of pancreatic carcinogenesis. VI. Early changes in hamster acinar cells induced by N-nitrosobis(2-hydroxypropyl)amine. Carcinogenesis 3, 1063–1070. 29. 29 Hall, P. A. and Lemoine, N. R. (1993) Models of pancreatic cancer. Cancer Surv. 16, 135–155. 30. Rivera, J. A., Graeme-Cook, F., Werner, J., et al. (1997) A rat model of pancrea30 tic ductal adenocarcinoma: Targeting chemical carcinogens. Surgery 122, 82–90. 31. Rao, M. S. (1987) Animal models of exocrine pancreatic carcinogenesis. Cancer Metastas. Rev. 6, 665–676. 32. Pour, P. M., Weide, L., Liu, G., et al. (1997) Experimental evidence for the origin 32 of ductal-type adenocarcinoma from the islets of Langerhans. Am. J. Pathol. 150, 2167–2180. 33. Schaeffer, B. K., Zurlo, J., and Longnecker, D. S. (1990) Activation of c-Ki-ras 33 not detectable in adenomas or adenocarcinomas arising in rat pancreas. Mol. Carcinog. 3, 165–170.
Modeling Pancreatic Cancer in Animals
239
34. 34 Dissin, J., Mills, L. R., Mains, D. L., Black, O. Jr., and Webster, P. D. (1975) Experimental induction of pancreatic adenocarcinoma in rats. J. Natl. Cancer Inst. 55, 857–864. 35. 35 Bockman, D. E., Black, O. Jr., Mills, L. R., Mainz, D. L., and Webster, P. D. (1976) Fine structure of pancreatic adenocarcinoma induced in rats by 7,12-dimethylbenz(a)anthracene. J. Natl. Cancer Inst. 57, 931–936. 36. 36 Albanese, C., Hulit, J., Sakamaki, T., and Pestell, R. G. (2002) Recent advances in inducible expression in transgenic mice. Semin. Cell Dev. Biol. 13, 129–141. 37. 37 Tuveson, D. A. and Jacks, T. (2002) Technologically advanced cancer modeling in mice. Curr. Opin. Genet. Dev. 12, 105–110. 38. Yamamoto, A., Hen, R., and Dauer, W. T. (2001) The ons and offs of inducible 38 transgenic technology: A review. Neurobiol. Dis. 8, 923–932. 39. 39 Grippo, P. J., Nowlin, P. S., Cassaday, R. D., and Sandgren, E. P. (2002) Cellspecific transgene expression from a widely transcribed promoter using Cre/lox in mice. Genesis 32, 277–286. 40. Lewandoski, M. (2001) Conditional control of gene expression in the mouse. 40 Nat. Rev. Genet. 2, 743–755. 41. 41 Sauer, B. (1998) Inducible gene targeting in mice using the Cre/lox system. Methods 14, 381–392. 42. Hammer, R. E., Swift, G. H., Ornitz, D. M., et al. (1987) The rat elastase I regula42 tory element is an enhancer that directs correct cell specificity and developmental onset of expression in transgenic mice. Mol. Cell Biol. 7, 2956–2967. 43. Ornitz, D. M., Palmiter, R. D., Hammer, R. E., et al. (1985) Specific expression 43 of an elastase-human growth hormone fusion gene in pancreatic acinar cells of transgenic mice. Nature 313, 600–602. 44. De Lisle, R. C. and Logsdon, C. D. (1990) Pancreatic acinar cells in culture: 44 Expression of acinar and ductal antigens in a growth-related manner. Eur. J. Cell Biol. 51, 64–75. 45. 45 Yuan, S., Duguid, W. P., Agapitos, D., Wyllie, B., and Rosenberg, L. (1997) Phenotypic modulation of hamster acinar cells by culture in collagen matrix. Exp. Cell Res. 237, 247–258. 46. Arias, A. E. and Bendayan, M. (1993) Differentiation of pancreatic acinar cells 46 into duct-like cells in vitro. Lab. Invest. 69, 518–530. 47. 47 Hall, P. A. and Lemoine, N. R. (1992) Rapid acinar to ductal transdifferentiation in cultured human exocrine pancreas. J. Pathol. 166, 97–103. 48. Vila, M. R., Lloreta, J., and Real, F. X. (1994) Normal human pancreas cultures 48 display functional ductal characteristics. Lab. Invest. 71, 423–431. 49. 49 Wagner, M., Luhrs, H., Kloppel, G., Adler, G., and Schmid, R. M. (1998) Malignant transformation of duct-like cells originating from acini in transforming growth factor transgenic mice. Gastroenterology 115, 1254–1262. 50. Quaife, C. J., Pinkert, C. A., Ornitz, D. M., Palmiter, R. D., and Brinster, R. L. 50 (1987) Pancreatic neoplasia induced by ras expression in acinar cells of transgenic mice. Cell 48, 1023–1034.
240
Grippo and Sandgren
51. 51 Ornitz, D. M., Hammer, R. E., Messing, A., Palmiter, R. D., and Brinster, R. L. (1987) Pancreatic neoplasia induced by SV40 T-antigen expression in acinar cells of transgenic mice. Science 238, 188–193. 52. 52 Sandgren, E. P., Luetteke, N. C., Palmiter, R. D., Brinster, R. L., and Lee, D. C. (1990) Overexpression of TGF alpha in transgenic mice: Induction of epithelial hyperplasia, pancreatic metaplasia, and carcinoma of the breast. Cell 61, 1121–1135. 53. 53 Sandgren, E. P., Quaife, C. J., Paulovich, A. G., Palmiter, R. D., and Brinster, R. L. (1991) Pancreatic tumor pathogenesis reflects the causative genetic lesion. Proc. Natl. Acad. Sci. USA 88, 93–97. 54. Grippo, P. J., Nowlin, P. S., Demeure, M. J., Longnecker, D. S., and Sandgren, 54 E. P. (2003) Preinvasive pancreatic neoplasia of ductal phenotype induced by acinar cell targeting of mutant kras in transgenic mice. Cancer Res. 63, 2016–2019. 55. Peat, N., Gendler, S. J., Lalani, N., Duhig, T., and Taylor-Papadimitriou, J. (1992) 55 Tissue-specific expression of a human polymorphic epithelial mucin (MUC1) in transgenic mice. Cancer Res. 52, 1954–1960. 56. Goetze, J. P., Nielsen, F. C., Burcharth, F., and Rehfeld, J. F. (2000) Closing the 56 gastrin loop in pancreatic carcinoma: Coexpression of gastrin and its receptor in solid human pancreatic adenocarcinoma. Cancer 88, 2487–2494. 57. Watson, S. A. and Caplin, M. (2002) Correspondence re: Weinberg et al., Cholecystokinin and gastrin levels are not elevated in pancreatic carcinoma. Cancer Epidemiol. Biomark. Prev. 10, 721–722. Cancer Epidemiol. Biomark. Prev. 11, 219. 58. Weinberg, D. S., Heyt, G. J., Cavanagh, M., Pitchon, D., McGlynn, K. A., and London, W. T. (2001) Cholecystokinin and gastrin levels are not elevated in human pancreatic adenocarcinoma. Cancer Epidemiol. Biomark. Prev. 10, 721–722. 59. 59 Mukherjee, P., Ginardi, A. R., Madsen, C. S., et al. (2000) Mice with spontaneous pancreatic cancer naturally develop MUC-1-specific CTLs that eradicate tumors when adoptively transferred. J. Immunol. 165, 3451–3460. 60. Yen, T. W., Sandgren, E. P., Liggitt, H. D., et al. (2002) The gastrin receptor 60 promotes pancreatic growth in transgenic mice. Pancreas 24, 121–129. 61. 61 Clerc, P., Leung-Theung-Long, S., Wang, T. C., et al. (2002) Expression of CCK2 receptors in the murine pancreas: Proliferation, transdifferentiation of acinar cells, and neoplasia. Gastroenterology 122, 428–437. 62. Grippo, P. J. and Sandgren, E. P. (2000) Highly invasive transitional cell carci62 noma of the bladder in a simian virus 40 T-antigen transgenic mouse model. Am. J. Pathol. 157, 805–813. 63. Brembeck, F. H., Schreiber, F. S., Deramaudt, T. B., et al. (2003) The mutant K63 ras oncogene causes pancreatic periductal lymphocytic infiltration and gastric mucous neck cell hyperplasia in transgenic mice. Cancer Res. 63, 2005–2009. 64. Gu, G., Dubauskaite, J., and Melton, D. A. (2002) Direct evidence for the pan64 creatic lineage: NGN3+ cells are islet progenitors and are distinct from duct progenitors. Development 129, 2447–2457.
Modeling Pancreatic Cancer in Animals
241
65. 65 Herrera, P. L., Nepote, V., and Delacour, A. (2002) Pancreatic cell lineage analyses in mice. Endocrine 19, 267–278. 66. Song, S. Y., Gannon, M., Washington, M. K., et al. (1999) Expansion of Pdx166 expressing pancreatic epithelium and islet neogenesis in transgenic mice overexpressing transforming growth factor alpha. Gastroenterology 117, 1416–1426. 67. Holland, A. M., Hale, M. A., Kagami, H., Hammer, R. E., and MacDonald, R. J. 67 (2002) Experimental control of pancreatic development and maintenance. Proc. Natl. Acad. Sci. USA 99, 12236–12241. 68. Donehower, L. A., Harvey, M., Slagle, B. L., et al. (1992) Mice deficient for p53 68 are developmentally normal but susceptible to spontaneous tumours. Nature 356, 215–221. 69. Serrano, M., Lee, H., Chin, L., Cordon-Cardo, C., Beach, D., and DePinho, R. A. 69 (1996) Role of the INK4a locus in tumor suppression and cell mortality. Cell 85, 27–37. 70. Sirard, C., de la Pompa, J. L., Elia, A., et al. (1998) The tumor suppressor gene 70 Smad4/Dpc4 is required for gastrulation and later for anterior development of the mouse embryo. Genes Dev. 12, 107–119. 71. Sharpless, N. E., Alson, S., Chan, S., Silver, D. P., Castrillon, D. H., and DePinho, 71 R. A. (2002) p16(INK4a) and p53 deficiency cooperate in tumorigenesis. Cancer Res. 62, 2761–2765. 72. Yang, X., Li, C., Herrera, P. L., and Deng, C. X. (2002) Generation of Smad4/ 72 Dpc4 conditional knockout mice. Genesis 32, 80–81. 73. 73 Wagner, M., Greten, F. R., Weber, C. K., et al. (2001) A murine tumor progression model for pancreatic cancer recapitulating the genetic alterations of the human disease. Genes Dev. 15, 286–293. 74. Bardeesy, N., Morgan, J., Sinha, M., et al. (2002) Obligate roles for p16(Ink4a) 74 and p19(Arf)-p53 in the suppression of murine pancreatic neoplasia. Mol. Cell Biol. 22, 635–643. 75. Cullingworth, J., Hooper, M. L., Harrison, D. J., et al. (2002) Carcinogen-induced 75 pancreatic lesions in the mouse: Effect of Smad4 and Apc genotypes. Oncogene 21, 4696–4701. 76. Takaku, K., Miyoshi, H., Matsunaga, A., Oshima, M., Sasaki, N., and Taketo, M. M. 76 (1999) Gastric and duodenal polyps in Smad4 (Dpc4) knockout mice. Cancer Res. 59, 6113–6117. 77. Morikane, K., Tempero, R., Sivinski, C. L., Kitajima, S., Gendler, S. J., and 77 Hollingsworth, M. A. (2001) Influence of organ site and tumor cell type on MUC1specific tumor immunity. Int. Immunol. 13, 233–240. 78. Bouvet, M., Wang, J., Nardin, S. R., et al. (2002) Real-time optical imaging of 78 primary tumor growth and multiple metastatic events in a pancreatic cancer orthotopic model. Cancer Res. 62, 1534–1540. 79. Lee, N. C., Bouvet, M., Nardin, S., et al. (2000) Antimetastatic efficacy of adju79 vant gemcitabine in a pancreatic cancer orthotopic model. Clin. Exp. Metastas. 18, 379–384.
242
Grippo and Sandgren
80. 80 Alves, F., Contag, S., Missbach, M., et al. (2001) An orthotopic model of ductal adenocarcinoma of the pancreas in severe combined immunodeficient mice representing all steps of the metastatic cascade. Pancreas 23, 227–235. 81. Tarbe, N., Evtimova, V., Burtscher, H., Jarsch, M., Alves, F., and Weidle, U. H. 81 (2001) Transcriptional profiling of cell lines derived from an orthotopic pancreatic tumor model reveals metastasis-associated genes. Anticancer Res. 21, 3221–3228. 82. 82 Tsuzuki, Y., Mouta Carreira, C., Bockhorn, M., Xu, L., Jain, R. K., and Fukumura, D. (2001) Pancreas microenvironment promotes VEGF expression and tumor growth: Novel window models for pancreatic tumor angiogenesis and microcirculation. Lab. Invest. 81, 1439–1451. 83. Bloomston, M., Shafii, A., Zervos, E. E., and Rosemurgy, A. S. (2002) TIMP-1 83 overexpression in pancreatic cancer attenuates tumor growth, decreases implantation and metastasis, and inhibits angiogenesis. J. Surg. Res. 102, 39–44. 84. Wen, Y., Yan, D. H., Wang, B., et al. (2001) p202, an interferon-inducible pro84 tein, mediates multiple antitumor activities in human pancreatic cancer xenograft models. Cancer Res. 61, 7142–7147. 85. Bruns, C. J., Harbison, M. T., Kuniyasu, H., Eue, I., and Fidler, I. J. (1999) In 85 vivo selection and characterization of metastatic variants from human pancreatic adenocarcinoma by using orthotopic implantation in nude mice. Neoplasia 1, 50–62. 86. Solorzano, C. C., Baker, C. H., Tsan, R., et al. (2001) Optimization for the block86 ade of epidermal growth factor receptor signaling for therapy of human pancreatic carcinoma. Clin. Cancer Res. 7, 2563–2572. 87. Denham, D. W., Franz, M. G., Denham, W., et al. (1998) Directed antisense 87 therapy confirms the role of protein kinase C-alpha in the tumorigenicity of pancreatic cancer. Surgery 124, 218–223; discussion 223–214. 88. Morioka, C. Y., Saito, S., Kita, K., and Watanabe, A. (2000) Curative resection 88 of orthotopically implanted pancreatic cancer in Syrian golden hamsters. Int. J. Pancreatol. 28, 207–213. 89. Juhl, H., Sievers, M., Baltzer, K., et al. (1995) A monoclonal antibody-cobra 89 venom factor conjugate increases the tumor-specific uptake of a 99mTc-labeled anti-carcinoembryonic antigen antibody by a two-step approach. Cancer Res. 55, 5749s–5755s. 90. 90 Matsushita, A., Onda, M., Uchida, E., Maekawa, R., and Yoshioka, T. (2001) Antitumor effect of a new selective matrix metalloproteinase inhibitor, MMI-166, on experimental pancreatic cancer. Int. J. Cancer 92, 434–440. 91. Ng, S. S., Tsao, M. S., Nicklee, T., and Hedley, D. W. (2001) Wortmannin inhib91 its pkb/akt phosphorylation and promotes gemcitabine antitumor activity in orthotopic human pancreatic cancer xenografts in immunodeficient mice. Clin. Cancer Res. 7, 3269–3275. 92. Zervos, E. E., Shafii, A. E., and Rosemurgy, A. S. (1999) Matrix metallopro92 teinase (MMP) inhibition selectively decreases type II MMP activity in a murine model of pancreatic cancer. J. Surg. Res. 81, 65–68.
Modeling Pancreatic Cancer in Animals
243
93. 93 Hotz, H. G., Hines, O. J., Hotz, B., Foitzik, T., Buhr, H. J., and Reber, H. A. (2003) Evaluation of vascular endothelial growth factor blockade and matrix metalloproteinase inhibition as a combination therapy for experimental human pancreatic cancer. J. Gastrointest. Surg. 7, 220–227; discussion 227–228. 94. Alves, F., Borchers, U., Padge, B., et al. (2001) Inhibitory effect of a matrix 94 metalloproteinase inhibitor on growth and spread of human pancreatic ductal adenocarcinoma evaluated in an orthotopic severe combined immunodeficient (SCID) mouse model. Cancer Lett. 165, 161–170. 95. 95 Fu, X., Guadagni, F., and Hoffman, R. M. (1992) A metastatic nude-mouse model of human pancreatic cancer constructed orthotopically with histologically intact patient specimens. Proc. Natl. Acad. Sci. USA 89, 5645–5649. 96. Farre, L., Casanova, I., Guerrero, S., Trias, M., Capella, G., and Mangues, R. 96 (2002) Heterotopic implantation alters the regulation of apoptosis and the cell cycle and generates a new metastatic site in a human pancreatic tumor xenograft model. FASEB J. 16, 975–982. 97. 97 Liu, C. D., Tilch, L., Kwan, D., and McFadden, D. W. (2002) Vascular endothelial growth factor is increased in ascites from metastatic pancreatic cancer. J. Surg. Res. 102, 31–34. 98. Morioka, C. Y., Saito, S., Ohzawa, K., and Watanabe, A. (2000) Homologous 98 orthotopic implantation models of pancreatic ductal cancer in Syrian golden hamsters: Which is better for metastasis research—cell implantation or tissue implantation? Pancreas 20, 152–157. 99. Hotz, H. G., Reber, H. A., Hotz, B., et al. (2001) An improved clinical model of 99 orthotopic pancreatic cancer in immunocompetent Lewis rats. Pancreas 22, 113– 121. 100. He, Z., Evelhoch, J. L., Mohammad, R. M., et al. (2000) Magnetic resonance imaging to measure therapeutic response using an orthotopic model of human pancreatic cancer. Pancreas 21, 69–76.
244
Grippo and Sandgren
Site-Specific Recombinases
245
16 Strategies for the Use of Site-Specific Recombinases in Genome Engineering Julie R. Jones, Kathy D. Shelton, and Mark A. Magnuson Summary Conventional gene targeting has been very useful in the study of gene function and regulation in mice. However, the methodologies involved have several limitations. First, mutations that cause embryonic lethality largely preclude studies of gene function at a later stage in development. Second, conditional and/or tissue-specific alterations of gene expression cannot be achieved using these methods. In addition, classical gene targeting can be difficult and time consuming. Strategies that make use of site-specific recombinases such as Cre and/or Flp have been developed in recent years to overcome these limitations. These new techniques include global and conditional knockouts, recombinase-mediated DNA insertion (RMDI), and recombinase-mediated cassette exchange (RMCE). Together, they have tremendously increased the number and variety of genetic manipulations that can be achieved. Key Words: Site-specific recombinases; gene targeting; Cre; loxP sites; Flp; FRT sites; embryonic stem cells.
1. Introduction The ability to manipulate the mouse genome by gene targeting in embryonic stem (ES) cells has greatly enhanced our ability to determine both the function and regulation of specific genes within a mammalian species. However, despite the advances in our knowledge that gene targeting has facilitated, the methods involved are often frustratingly slow and have intrinsic limitations, particularly with respect to the inability to perform conditional and tissue-specific gene manipulations. Thus, during the past decade strategies that make use of site-specific recombinases have attracted considerable attention. These strategies, when coupled with gene targeting in ES cells, allow for the selective manipulation of genes in a manner that does not depend on repetitive cycles of homologous recombination. Indeed, the combined use of gene targeting and site-specific recombinases has opened the door to a large number of new strategies whereby the function of specific genes can be explored (1). From: Methods in Molecular Medicine, Vol. 103: Pancreatic Cancer: Methods and Protocols Edited by: G. Su © Humana Press Inc., Totowa, NJ
245
246
Jones et al.
Site-specific recombinases function by binding to target DNA sequences that can be placed, via gene targeting, at single or multiple locations in the genome. Depending on the orientation and number of these recognition sites, it has become possible to generate DNA deletions, insertions, inversions, and translocations as a result of the catalytic power of these recombinases (2). Here, we describe strategies that make use of two site-specific recombinase systems: Cre-loxP and Flp-FRT. Both Cre and Flp are members of the integrase superfamily of sitespecific recombinases. Cre is a 38-kDa protein encoded by bacteriophage P1 whereas Flp, a 45-kDa protein, is found in the budding yeast Saccharomyces cerevisiae (2,3). Both Cre and Flp have minimal DNA recognition sequences of 34 basepairs (bp). Moreover, both enzymes perform DNA strand cleavage and religation as homotetramers. Interestingly, despite their origins in different phyla, both Cre and Flp contain four conserved amino acid residues in their recognition sites that are crucial to the DNA cleavage/ligation reaction (4). The 34-bp recognition sequence for Cre has been termed loxP (for locus of crossover [x] in P1) while the recognition sequence for Flp is termed an FRT site (for Flp recognition target). Both the loxP and minimal FRT site consist of two 13-bp inverted repeats that flank an 8-bp asymmetric core sequence. The relative orientations of the 8-bp core sequences give directionality to the target site and govern the strand crossover reaction catalyzed by the recombinase. For the loxP site, the 13-bp inverted sequence repeats are identical, whereas the minimal FRT site contains a single base difference in the symmetry elements (Fig. 1A). While the wild-type (wt) FRT site contains a third copy of the 13-bp repeat sequence, it is dispensable because efficient strand crossover can occur in its absence (4). Thus, the minimal FRT site shown in Fig. 1A is all that is necessary in the strategies that are described here. The Materials and Methods subheadings that follow outline the general steps necessary to utilize site-specific recombinases to design genetic loci at which DNA deletions, insertions, and exchanges can be induced to occur. Our goal is to present a general approach to the researcher for choosing an overall strategy that allows for the genetic manipulations of choice to be achieved. Additional information regarding the mechanisms of the reactions involved can be found in the literature. 2. Materials 2.1. Targeting Replacement Vector 1. DNA that is isogenic to the ES cells being used is recommended for the construction of a gene targeting vector (see Note 1). DNA probes and polymerase chain reaction (PCR) primer pairs are required to identify and validate recombination events by Southern blot and PCR analysis, respectively.
Site-Specific Recombinases
247
2. A detailed restriction map of the gene of interest is required for (1) the selection of DNA fragments for targeting vector arms, (2) the precise placement of recombinase recognition sites, (3) the design of primers for sequence verification of the final targeting vector construct, and (4) the synthesis of PCR primers for validation of gene targeting and recombination events. 3. Positive/negative selection cassette(s), for example, neomycin resistance, hygromycin resistance, herpes simplex virus thymidine kinase (HSVtk), diphtheria toxin, and so forth. 4. Plasmid backbone vectors for constructing gene targeting, insertion, and/or exchange vectors. Because the placement and orientation of recombinant recognition sites vary immensely by the gene being targeted and the strategy being applied, generic vectors are often not available. For this reason, gene targeting vectors that incorporate the use of site-specific recombinases often need to be assembled in a piecemeal process. 5. Mouse ES cell lines and culture media. 6. Southern blotting materials including both large and standard sized electrophoresis gel equipment, Zeta-Probe GT Genomic Tested Blotting Membrane (Bio-Rad), Stratagene Prime-It II random primer labeling kit, and various reagents to prepare electrophoresis buffers, gels, gel fixing and blotting solutions, and hybridization and wash solutions.
2.2. Recombinase Recognition Sites In addition to the FRT and loxP sites mentioned above, mutational studies have led to the identification of variant loxP sites that have functional characteristics that may make them useful in specific settings. For instance, in addition to catalyzing recombination between two loxP sites, Cre can also perform the same reaction between two lox511 sites. Lox511 is identical to loxP except for a single base mutation in the 8-bp spacer region (Fig. 1A). However, although lox511 can recombine with another lox511 site, it does so poorly with a loxP site. Thus, the combined use of both a loxP and a lox511 may simplify some strategies, such as recombinase-mediated cassette exchange. However, whether this is the case in practice remains to be determined. Similarly, lox71 and lox66 are variant sites that contain mutations in either the left end (LE) or right end (RE) symmetry elements. Recombination between a lox71 and lox66 generates a double-mutant lox (DM) site that contains both LE and RE mutations, and a wild-type loxP. The utility of this approach results from the fact that the DM lox site is poorly recognized by Cre, which may make it useful in the situation where Cre is used to catalyze gene insertion events as shown in Fig. 1B. 2.3. Cre and/or Flp Recombinase Both the Cre-loxP and Flp-FRT systems have been demonstrated to be powerful tools for gene manipulation; however, the marked differences in the origin
248
Jones et al.
Fig. 1. Cre and Flp recognition sites. (A) FRT, loxP, and variant lox sites all consist of two 13-bp inverted repeats flanking an 8-bp asymmetric spacer. (B) Left end (lox71) and right end (lox66) mutant lox sites can be used to achieve recombinase-mediated DNA insertion.
of the two proteins, bacterial virus vs budding yeast, have led to differences in the optimal temperature for enzymatic activity. The fact that yeast grow well at ambient temperatures (approx 30°C) may explain why Flp is less thermostable than Cre, whose temperature optimum appears more suited to normal growth conditions of its bacterial host, E. coli (37°C). Owing to this character-
Site-Specific Recombinases
249
Table 1 Characteristics of Cre and FLP Recombinases Cre
FLP
FLPe
Optimal temperature at or above 37°C Denatures at 54°C Recommended for mammalian systems
Optimal temperature near 30°C Denatures at 38°C Useful for applications that depend on tight regulation, not efficiency or recombination
Optimal temperature near 37°C Denatures at 40°C Recombination efficiency at 37°C similar to Cre
Can be used in organisms with temperatures < 35°C (i.e., flies, plants, fish)
Three to fivefold increase in recombination efficiency in mammalian cells compared to FLP
istic, Cre may be more naturally suited for use in mammals and ES cells. Fortunately, the thermolability of Flp has been overcome by the identification of amino acid substitutions that increase the thermostability of the enzyme (P2S, L33S, Y108N, and S294P) (5). The thermostable or enhanced version of Flp has been termed Flpe. The modified enzyme has an activity equivalent to that of Cre at body temperature (4). Some of the properties of Cre, Flp, and Flpe are listed in Table 1, along with recommendations for their usage (6). 2.4. Conditional and Global Gene Knockouts 1. Cre and Flpe expression plasmids and/or mice are available for use in obtaining the desired recombination events both in vitro and in vivo (see Notes 2 and 3). 2. PCR primer pairs and DNA probes to screen for recombination events.
2.5. Recombinase-Mediated DNA Insertion 1. An insertion vector containing a lox66 site and DNA sequence of interest. 2. Cre and Flpe expression plasmids or mice expressing Cre or Flpe from globally expressed promoters (see Notes 2 and 3). 3. PCR primers and DNA probes to screen for recombination.
2.6. Recombinase-Mediated Cassette Exchange 1. A plasmid containing the exchange DNA cassette flanked by loxP and an inverted loxP site (loxP inv) or lox511 and loxP inv. 2. Sources of Cre and Flpe expression plasmids and mice available for use in obtaining the desired recombination events both in vitro and in vivo (see Notes 2 and 3). 3. PCR primer pairs and DNA probes to screen for recombination events.
250
Jones et al.
3. Methods 3.1. The Targeting Vector Whether performing recombinase-mediated deletions, insertions, or exchange reactions, specific recombinase recognition sites must first be introduced into the gene of interest, usually in precise orientations and locations. This is done via homologous recombination in ES cells through the use of a replacement-type of gene targeting vector. This type of gene targeting contains both long and short arms of homologous DNA, where DNA crossover occurs (see Note 4). Gene targeting vectors of this design generally contain both positive and negative selection markers so as facilitate identification of clones in which the sought after homologous recombination event has occurred. The positive selectable marker allows for the identification of clones that have taken up and stably integrated the targeting vector DNA. The negative selectable marker, when placed in a position outside either the long or short arm of DNA homology, helps to limit growth of clones containing random integration events (7). It is important to note that HSVtk cannot be used simultaneously to facilitate the identification of correctly targeted ES cells and the negative selection for RMCE (as described in Subheading 3.4.). Thus, if it is deemed preferable to retain a negative selectable marker in the initial gene targeting step required for this particular strategy, then diphtheria toxin can be used in lieu of HSVtk. In almost any gene targeting strategy it is useful for the positive selection cassette (usually encoding a neomycin resistance gene) to be floxed (flanked by loxP) or flrted (flanked by FRT) to allow removal of these sequences once gene targeting has been achieved. A unique restriction enzyme site located outside the homologous arm sequences of the targeting vector is also necessary to enable linearization of the targeting vector prior to electroporation of ES cells. It is very helpful, if not essential, to have the complete DNA sequence of the gene of interest when designing a targeting vector that requires various arrangements of loxP and FRT sites. This information quickly identifies restriction sites for the placement of recombinase recognition sequences, aids in the design of oligonucleotide primers for DNA sequence confirmation of the assembled vector and for PCR analysis of the targeted allele, and facilitates the selection of DNA hybridization probes that do not contain repetitive DNA elements that are often encountered when a DNA probe is randomly chosen. 3.2. Global and Conditional Knockouts The most common use, to date, of site-specific recombinases has been to generate conditional gene knockouts in mice. This has generally been achieved through insertion of three loxP sites into a region of a gene, two of which flank
Site-Specific Recombinases
251
Fig. 2. Conditional and global knockouts generated by recombinase-mediated DNA excision. (A) Conditional gene knockout strategy using three loxP sites. The exon that is flanked by loxP sites should contain key sequences or a nonunit number of codons. Cre-mediated recombination is used to generate both conditional and null alleles. Partial recombination is required for generation of the conditional allele which contains two loxP sites.
a positive selectable marker and the third inserted 2–3 kb away in a manner by which a key region of the gene can later be deleted (Fig. 2A). However, the limitation of this approach has been the need for a partial recombination event (e.g., 3 2 loxP) as identification of clones and/or mice containing this intermediate recombinant can sometimes prove difficult. With the development of Flpe, it has been possible to use a combined approach, for example, by introducing both loxP and FRT sites in the targeted locus, so that different portions of the targeted allele can be selectively removed via independent steps using either Cre or Flpe (Fig. 2B). This approach is more efficient because a partial recombination step can be completely avoided, or put to some other use. A third strategy that involves three loxP and two FRT sites has also been utilized to generate conditional and null alleles (Fig. 2C). Such a design enables
252
Jones et al.
Site-Specific Recombinases
253
the removal of a positive selection cassette, or other DNA sequence, via either Cre or Flpe. The arrangement of sites shown is only one of several, and serves largely to illustrate the point that many different possibilities exist whereby these two recombinases can be used to generate multiple different alleles from a single targeting vector. In designing targeting vectors to disrupt gene function, caution must be taken in the placement of the lox and FRT sites to ensure a knockout of gene function and to prevent alternative splicing or generation of a protein with partial or novel function. For small genes, such problems can be avoided by eliminating the entire gene. For larger genes, eliminating one or more of the more proximal exons may be sufficient. Our general approach has been to delete exons (whose length in base pairs is not divisible by three) downstream of the initiator ATG so as to allow protein translation to begin. This approach forces a frameshift after removal of the floxed DNA segment, thereby ensuring that a partially functional protein is not expressed (4). 3.3. Recombinase-Mediated DNA Insertion Cre recombinases can be used for purposes other than gene deletions. For instance, to allow for the insertion of new DNA sequences into a gene locus, a strategy we term “recombinase-mediated DNA insertion” (RMDI) can be used. DNA insertion is achieved by Cre-mediated recombination between a loxP site contained within a plasmid and a loxP site that has been introduced into the genome by gene targeting (Fig. 1B, left panel). Normally, the efficiency of RMDI is quite low because the newly integrated DNA is flanked by two tandemly oriented loxP sites and thus will be easily excised again if any Cre is present. However, this obstacle can be overcome by the use of heteromeric lox sites that favor DNA insertion over excision (8,9). One approach is to use left end (lox71) and right end (lox66) mutant lox sites. Recombination between these mutant lox sites results in the generation of a double mutant (DM) lox site and a wildtype loxP site (Fig. 1B, right panel). Recombination between the DM lox and loxP is highly inefficient because Cre does not recognize the DM lox site very well. This enables the inserted DNA to be resistant to excision, thereby increasing the likelihood of obtaining stable integration.
Fig. 2. (Opposite page) (B) Strategy implementing two loxP sites and two FRT sites. In this example, Flpe-mediated recombination removes the selection cassette, and Cre recombinase is subsequently used to remove coding sequences and generate a null allele. (C) Use of a three loxP/two FRT strategy allows deletion of intervening DNA sequences via one or another recombinase.
254
Jones et al.
3.4. Recombinase-Mediated Cassette Exchange Another strategy for the insertion of DNA into a tagged chromosomal location is recombinase-mediated cassette exchange (RMCE) (10–13). One advantage of this strategy over that of recombinase-mediated DNA insertion is that it avoids the obligatory introduction of plasmid DNA sequences into the genome, as occurs in RMDI. In the RMCE strategy based on Cre, a chromosomal cassette is flanked by two inversely oriented loxP sites (Fig. 3A). In the presence of Cre, the chromosomal cassette can be exchanged for a second cassette carried by a plasmid and also flanked by two inversely oriented loxP sites. The impetus for the use of inversely oriented sites is that DNA excision does not occur. Rather, Cre-mediated intramolecular recombination of inverted lox sites results in inversion of the intervening DNA sequences. This allows use of the chemically induced negative selection provided by HSVtk to facilitate identification of DNA exchange events because intrachromosomal recombination does not remove the negative selection cassette. RMCE resulting from intermolecular recombination results in the replacement of the positive/negative selection cassettes with a second DNA sequence of interest. Because these clones eliminate the selection cassettes, identification of clones in which the exchange of DNA has occurred can be facilitated. However, if one chooses to make use of HSVtk in this manner, then it is important to note that the sustained presence of HSVtk in a targeted allele will markedly impair germline transmission of any given ES cell clone. Thus, RMCE must be performed while the ES cells are in culture. A caveat to the RMCE approach shown in Fig. 3A is that the exchange cassette can integrate in two different orientations. This means that half of the exchange products will be in the opposite orientation than may be desired. An alternate strategy that makes use of both lox511 and loxP sites, instead of inversely oriented loxP sites (Fig. 3B), may overcome this issue. This strategy makes use of the fact that recombination between lox511 and loxP is inefficient, so that directionality might be maintained. However, this approach rests on the assumption that both lox511 and loxP sites have equal recombination efficiencies. Because the kinetics, and possibly efficiency, of recombination between two loxP sites may differ from that of lox511 sites, RMCE using two loxP sites could prove to be a better strategy. At present, RMCE based on the dual loxP sites has been shown to work well whereas the combined site strategy remains hypo-thetical. Indeed, both of these strategies lie at the leading edge of what has been demonstrated to be possible when using site-specific recombinases. Thus, variations on these themes, as well as other improved strategies, are likely to evolve within the next few years. Lastly, while we have only explained RMCE strategies based on use of Cre, in principle these strategies could also use Flpe/FRT in which two inversely oriented FRT sites flank a DNA region of interest.
Site-Specific Recombinases
255
Fig. 3. Recombinase-mediated cassette exchange. (A) RMCE using loxP and loxP inv. Any region of DNA (indicated by box with number 1) can be replaced with any other DNA sequence (indicated by the box with number 2). Intermolecular recombination of loxP sites and loxPinv sites results in the exchange of a second DNA sequence for the positive and negative selection cassettes that were first introduced by gene targeting. Intramolecular recombination of loxP and loxPinv results in the new DNA being present in either of two orientations. (B) RMCE using lox511 and inverted loxP (loxPinv). This strategy may have the advantage of maintaining directionality of the new DNA sequence after recombination.
256
Jones et al.
4. Notes 1. For more detailed information on constructing gene targeting replacement vectors and other information pertinent to the culture of mouse ES cells, please see ref. (14). 2. Cre and Flp expression plasmids have previously been described (15–19). 3. Information regarding Cre and Flp transgenic mice can be found on the following web pages: www.mshri.on.ca/nagy/default.htm and http://jaxmice.jax.org/info/index. html. 4. In general, we suggest using a short arm greater than 1 kb in length and a long arm of at least 4 kb.
Acknowledgments We thank Qiaoming Long for critical reading of the manuscript. J. R. J. was supported by NIH institutional training grant DK07061-27. References 1. 1 Metzger, D. and Feil, R. (1999) Engineering the mouse genome by site-specific recombination. Curr. Opin. Biotech. 10, 470–476. 2. 2 Nagy, A. (2000) Cre recombinase: The universal reagent for genome tailoring. Genesis 26, 99–109. 3. 3 Babineau, D., Vetter, D., Andrews, B. J., et al. (1985) The FLP protein of the 2micron plasmid of yeast. J. Biol. Chem. 260, 12313–12319. 4. Dymecki, S. M. (2000) Site-specific recombination in cells and mice, in Gene Targeting: A Practical Approach (Joyner, A. L., ed.), New York: Oxford University Press, pp. 37–99. 5. 5 Buchholz, F., Angrand, P. O., and Stewart, A. F. (1998) Improved properties of FLP recombinase evolved by cycling mutagenesis. Nat. Biotech. 16, 657–662. 6. Buchholz, F., Ringrose, L., Angrand, P., Rossi, F., and Stewart, A. F. (1996) Different thermostabilities of FLP and Cre recombinases: Implications for applied site-specific recombination. Nucl. Acids. Res. 24, 4256–4262. 7. Torres, R. M. and Kuhn, R. (1997) in Laboratory Protocols for Conditional Gene Targeting (Torres, R. M. and Kuhn, R., eds.), Oxford: Oxford University Press, p. 167. 8. 8 Albert, H., Dala, E. C., Lee, E., and Ow, D. W. (1995) Site-specific integration of DNA into wild-type and mutant lox sites placed in the plant genome. Plant J. 7, 649–659. 9. 9 Araki, K., Araki, M., and Yamamura, K. (1997) Targeted integration of DNA using mutant lox sites in embryonic stem cells. Nucl. Acids Res. 25, 868–872. 10. 10 Bouhassira, E. E., Westerman, K., and Leboulch, P. (1997) Transcriptional behavior of LCR enhancer elements integrated at the same chromosomal locus by recombinase-mediated cassette exchange. Blood 90, 3332–3344. 11. 11 Seibler, J. and Bode, J. (1997) Double-reciprocal crossover mediated by Flprecombinase: A concept and an assay. Biochemistry 36, 1740–1747.
Site-Specific Recombinases
257
12. 12 Seibler, J., Schubeler, D., Fiering, S., Groudine, M., and Bode, J. (1998) DNA cassette exchange in ES cells mediated by Flp recombinase: An efficient strategy for repeated modification of tagged loci by marker-free constructs. Biochemistry 37, 6229–6234. 13. Bethke, B. and Sauer, B. (1997) Segmental genomic replacement by Cre-medi13 ated recombination: Genotoxic stress activation of the p53 promoter in singlecopy transformants. Nucl. Acids Res. 25, 2828–2834. 14. Joyner, A. L. (ed.) (1993) Gene Targeting: A Practical Approach. Oxford: Oxford University Press. 15. Xu, X., Li, C., Garrett-Beal, L., Larson, D., Wynshaw-Boris, A., and Deng, C. X. 15 (2001) Direct removal in the mouse of a floxed neo gene from a three-loxP conditional knockout allele by two novel approaches. Genesis 30, 1–6. 16. Schaft, J., Ashery-Padan, R., van der Hoeven, F., Gruss, P., and Stewart, A. F. 16 (2001) Efficient FLP recombination in mouse ES cells and oocytes. Genesis 31, 6–10. 17. Lauth, M., Moerl, K., Barski, J. J., and Meyer, M. (2000) Characterization of 17 Cre-mediated cassette exchange after plasmid microinjection in fertilized mouse oocytes. Genesis 27, 153–158. 18. Feng, Y., Seibler, J., Alami, R., et al. (1999) Site-specific chromosomal integration 18 in mammalian cells: Highly efficient Cre recombinase-mediated cassette exchange. J. Mol. Biol. 292, 779–785. 19. Soukharev, S., Miller, J. L., and Sauer, B. (1999) Segmental genomic replacement in embryonic stem cells by double lox targeting. Nucl. Acids Res. 27, e21.
258
Jones et al.
Primary Explant Cultures of Pancreas
259
17 Primary Explant Cultures of Adult and Embryonic Pancreas Farzad Esni, Yoshiharu Miyamoto, Steven D. Leach, and Bidyut Ghosh Summary The developmental plasticity of adult pancreas is evidenced by the ability to undergo conversion between different epithelial cell types. Specific examples of such conversions include acinar to ductal metaplasia, ductal to islet metaplasia, and generation of ductal structures within islets. Although 90% of human pancreatic cancers are classified as ductal adenocarcinoma, markers of all pancreatic epithelial cell types (acini, ductal, and endocrine) as well as markers of gastric and intestinal lineages can be detected in these tumors. In recent years considerable knowledge has been gained regarding regulation of cellular differentiation and various signaling pathways involved in normal and neoplastic pancreas through studies of pancreatic cancer and immortalized ductal cell lines. However, these studies provide little insight into the context of normal developmental cues, the disruption of which leads to organ pathology. Here we have described a detailed method for preparation, maintenance, and manipulation of adult and embryonic mouse pancreas. These methods may be utilized in studies involving normal epithelial differentiation, contributing to improved understanding of pancreatic development and disease. Key Words: Acini; adenovirus; cancer; collagen; development; dorsal bud; embryo; explant; metaplasia; mouse; pancreas; transdifferentiation.
1. Introduction Although significant progress has been made in identifying genetic changes underlying pancreatic cancer, little is known regarding how these changes affect normal pancreatic epithelium, or how changes in epithelial differentiation may contribute to this disease. In addition to influencing the phenotype and biologic behavior of fully developed pancreatic cancer, it is intriguing to consider that early changes in epithelial differentiation may also contribute to initiation of pancreatic cancer precursor lesions. In this regard, acinar-to-ductal metaplasia has frequently been proposed as an initiating mechanism for pancreatic cancer (reviewed in 1). Patients with acinar-to-ductal metaplasia arising in the From: Methods in Molecular Medicine, Vol. 103: Pancreatic Cancer: Methods and Protocols Edited by: G. Su © Humana Press Inc., Totowa, NJ
259
260
Esni et al.
setting of chronic pancreatitis carry a 16-fold increase in relative risk for pancreatic ductal adenocarcinoma (2), increasing to 50-fold in patients with familial chronic pancreatitis (3). Evidence that this alteration in epithelial differentiation may indeed be involved in pancreatic cancer initiation is provided by a mouse model of pancreatic ductal metaplasia initiated by chronic overexpression of transforming growth factor-a (TGF-a), in which metaplastic ductal epithelium progresses to pancreatic intraepithelial neoplasia (PanIN) and invasive pancreatic cancer (4). The ability of adult pancreas to undergo metaplastic conversion between epithelial cell types emphasizes the high level of developmental plasticity displayed by this tissue. Specific examples of developmental plasticity in adult pancreas include ductal-to-islet metaplasia (islet neogenesis; 5–10), duct generation within islets (11–13), and acinar-to-ductal metaplasia (14–16). Presumably, these profound changes in pancreatic epithelial differentiation reflect the common origin of islet, acinar, and ductal cell types from a shared precursor pool within embryonic pancreatic epithelium. The multipotent differentiation capacity of pancreatic epithelium is further exemplified by examination of lineage markers in human pancreatic cancer. Although 90% of all pancreatic neoplasms may be classified as ductal adenocarcinoma, these tumors typically express combinations of ductal, islet, acinar, and/or markers of gastric and intestinal epithelium (17,18). Although some insight regarding regulation of differentiation in normal and neoplastic pancreas has been gained from investigations utilizing pancreatic cancer cell lines and/or immortalized pancreatic ductal epithelial cells, it might be argued that changes in differentiation are best studied within the context of an intact epithelium. To this end, a number of primary explant culture systems have been developed for the in vitro study of pancreatic tissue (11,14,15,19,20). These systems have previously been valuable in addressing a variety of scientific questions, including studies of epithelial–mesenchymal signaling, transdifferentiation, islet neogenesis, and in vitro carcinogenesis. In this chapter, we provide detailed methodologies for the preparation, maintenance, and genetic manipulation of explant cultures from both adult and embryonic pancreas. Although these techniques may be applicable to establishing explant cultures of pancreatic epithelium from a variety of species, we primarily employ mouse pancreas in order to utilize tissue from targeted mouse strains. 2. Materials 2.1. Preparation and Culture of Adult Pancreas 1. Rat tail collagen-type I (Collaborative Biomedical Products, cat. no. 354236). 2. Waymouth’s MB 752/1 media (Gibco BRL, cat. no. 51400-026).
Primary Explant Cultures of Pancreas 3. 4. 5. 6. 7. 8. 9. 10.
261
Fetal bovine serum (FBS; Sigma). 0.34 N NaOH. Collagenase P (Roche no. 1249002). Hank’s balanced salt solution (HBSS) (Gibco BRL) 14025-084. Soybean trypsin inhibitor (Sigma T-9003). Dexamethasone (Sigma D-2915). 24-Well tissue culture plates. Polypropylene mesh (105 µm and 500 µm) (Spectrum Laboratories Inc., SpectraMesh-autoclaved 7.5 cm2).
2.2. Adenoviral Infection 1. 2. 3. 4.
High-titer stock of adenovirus containing the gene of interest (109–1011 pfu/mL). Cell counter (Fisher Scientific no. 0267110). 14-mL Falcon culture tube (Becton Dickinson, cat. no. 2059). 1X Waymouth’s MB 752/1 media (Gibco BRL, cat. no. 51400-026).
2.3. Preparation and Culture of Embryonic Pancreas 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11.
1X Phosphate-buffered saline (PBS) (Bufluids, cat. no. P315-500). Medium 199 (BioWhittaker, cat. no. 12-119F). Fungizone (Invitrogen, cat. no. 15295-017). Penicillin G-streptomycin (Invitrogen, cat. no. 15140-122). FBS (Invitrogen, cat. no. 16140-071). 24-Well plates (BD Falcon, cat. no. 3047). 60 ´ 15 mm tissue culture dish (BD Falcon, cat. no. 3002). 0.4-µm Millicell-CM culture inserts (Millipore, cat. no. PICM 01250). Microdissecting forceps (Roboz, cat. no. RS-4903 and RS-4974). Microdissecting scissors (Roboz, cat. no. RS-5850). Mouth pipet.
3. Methods 3.1. Explant Cultures of Adult Pancreas 3.1.1. Preparation of Collagen Plates (for One Mouse Pancreas) 1. On ice, in tissue culture hood, mix 9 mL of rat tail collagen, 900 µL of unsupplemented 10X Waymouth’s media (filtered), 600 µL of 0.34 N NaOH (filtered). This is just an example. You need only 200 µL per well, the number of wells depending on the experiment you are doing. 2. Place two 24-well plates on ice. Then pipet 200 µL of collagen gel mixture into each well, ensuring that the gel covers the bottom of each well. Cover and place in incubator to solidify until ready to use. (It is easiest to get total coverage on the bottom of the well if you pipet into the center of each well.) Now place three 100 ´ 15 mm tubes in ice and fill each with 10 mL of HBSS. (These will be used to do a series of washes on the resected pancreas.)
262
Esni et al.
3.1.2. Harvesting Adult Pancreas 1. Anesthetize the mouse with isofluorane (see Note 1). 2. Place the animal in a supine position and tape down limbs, wipe abdomen with betadine or 70% ETOH, and make midline abdominal incision. 3. Mobilize and resect the pancreas from tail to head with sterile instruments. 4. Place the resected pancreas into 10 mL of HBSS on ice. 5. In the hood, transfer the pancreas to fresh 10 mL of cold HBSS twice. 6. During the final wash, mince pancreas in HBSS into 1- to 5-mm pieces using sterile scissors. 7. Centrifuge at 720g for 2 min at 4°C.
3.1.3. Isolation of Adult Exocrine Epithelium (on Ice) 1. Aspirate the supernatant. 2. Resuspend pancreas in 5 mL of cold HBSS containing 1 mg of collagenase P (freshly made and filtered). 3. Shake gently at 37°C for 10 min. Wrap the tube top in parafilm (to prevent leaking in water bath). After shaking, rinse tube in 70% EtOH. 4. Stop reaction by placing on ice and adding 5 mL of HBSS with 5% FBS. 5. Centrifuge at 720g for 2 min at 4°C, remove the supernatant, and resuspend the pellet with 10 mL of cold HBSS with 5% FBS. Centrifuge at 400g for 2 min at 4°C (repeat two times). 6. After the final wash, resuspend the pellet in 5 mL HBSS with 5% FBS. 7. Pipet 5 mL of cell suspension through a sterile 500-µm mesh. 8. Rinse mesh with 5 mL of cold HBSS with 5% FBS. 9. Pipet a 10-mL cell suspension through a 105-µm mesh. 10. Add 10 mL of cell suspension carefully (use a 5-mL pipet) to the top of 20 mL of HBSS with 30% FBS. 11. Centrifuge at 180g for 2 min at 4°C. 12. Aspirate the supernatant. 13. Resuspend the pellet in 10 mL of cold Waymouth’s media with 10% FBS, 1% penicillin/streptomycin, 100 µg/mL of trypsin inhibitor, and 1 µg/mL of dexamethasone (see Note 2). 14. Prepare the collagen gel suspension with 12 mL of rat tail collagen, 1200 µL of 10X media, and 800 µL of 0.34 N NaOH. 15. Mix the cell suspension and the collagen gel mixture together (equal volumes, depending on recovery). 16. Remove the 24-well collagen-coated gel plates from the incubator and place on ice. 17. Pipet 500 µL of the collagen gel/cell suspension mixture into each coated well. 18. Place the 24-well plates back into the 37°C incubator. 19. After 30 min the gel should be solidified. At room temperature add 0.5 mL of media with 10% FBS and 1% penicillin/streptomycin to each well. Media should be at room temperature or 37°C. At this stage the healthy epithelium consists primarily
Primary Explant Cultures of Pancreas
263
Fig. 1. Freshly harvested explants of mouse exocrine pancreas. Note predominance of acinar cells localized in intact acinar units, and near complete elimination large interlobular ducts and islet elements (A, ´100; B, ´400). Arrows indicate dark zymogen granules, characteristic of acinar cells. Phase-contrast photomicrograph of adenoviral-GFP infected acini preparation after 24 h of infection (C) with corresponding GFP expression (D). of intact exocrine acini, visualized as clumps with dark brown zymogen granules under a phase-contrast microscope (Fig. 1A). 20. Add growth factors/inhibitors, and so forth at this time (see Note 3). 21. Return the plates to the 37°C incubator.
3.1.4. Extended Culture of Adult Epithelium We usually allow our cultures to continue for 5 d, but longer periods are possible. After 7–8 d, the cultures become acidic and the collagen begins to break down, requiring passage of explants into fresh collagen gels. For longer culture periods, transfer to fresh collagen gels is required. 1. Prepare the collagen gel as above. Coat bottom of 60 ´ 15 mm culture dish with 1.5 mL of collagen gel, making sure entire surface is covered. Allow solidifying for 30 min at 37°C.
264
Esni et al.
2. Using sterile (bleached) Gelman forceps, place four gel discs into each culture disc, not allowing each other to touch. 3. Overlay with 1.5 mL of collagen gel mixture. Allow 30 min to solidify. 4. Overlay with 2 mL of media with appropriate supplements, growth factors, and so forth.
3.1.5. Adenoviral Infection of Adult Pancreatic Epithelium Methods for high-titer adenoviral preparation have been described elsewhere (21). A series of preliminary infections should be performed to determine the multiplicity of infection (MOI), ideally resulting in greather than 50% transduction efficiency with acceptable cytotoxicity as determined by morphological comparison with mock-infected cells. 1. Prior to suspending the epithelium in the collagen gel mixture, reserve an aliquot to be trypsinized and counted to determine cell number for MOI calculations. 2. Mix the suspended epithelial structures with predetermined MOI of adenovirus containing gene of interest in 1X Waymouth’s media in a 15-mL Falcon culture tube. 3. Incubate at 37°C for 1 h at CO2 incubator at 4°C (see Note 4). 4. Centrifuge the epithelium at 400g for 2 min. 5. Aspirate the supernatant and resuspend the cells with appropriate 1X Waymouth’s media. 6. Follow step 13 onward as described in Subheading 3.3. 7. Check by immunofluorescence or other reporter assay after 24–48 h for successful infection (Fig. 1B,C).
3.2. Explant Cultures of Embryonic Pancreas 3.2.1. Culturing Pancreatic Explants Derived from E9.5 (approx 20-Somite) Mouse Embryos Before culturing, prepare the pancreas culture media (PCM) with Medium 199, 10% FBS, 50 U/mL of penicillin G–streptomycin, and 1.25 µg/mL of Fungizone. This solution may be stored at 4°C up to 1 mo. It is recommended to warm the media to 37°C prior to use. The procedures described here for isolating and culturing pancreatic anlagen has been modified from original method described by others (19,22,23). 1. Add 400 µL of PCM in each well before placing the Millicell-CM culture insert into the well. 2. Kill the pregnant female(s) and transfer the uteri to sterile ice-cold PBS. 3. Transfer the uteri to a fresh dish containing ice-cold PBS in order to wash out blood and debris. 4. Isolate the embryos and transfer to a fresh dish.
Primary Explant Cultures of Pancreas
265
5. Be sure the embryos are covered by PBS and kept on ice. Transfer one embryo at time to a fresh dish and perform the dissection. 6. Lay down the embryo on bottom of the dish with its right side facing you as shown in Fig. 2B. 7. By using forceps, carefully cut and remove the limb buds and lateral body wall along the rostral–caudal axis. This will allow visualization of the inner organs (Fig. 2B,C). 8. Flip the embryo vertically and repeat the same procedure on the other side (Fig. 2D,E). In the beginning it may be difficult to localize the pancreatic buds. To do this, try to follow the gut along the caudal–rostral axis, and you will see that at a certain point it becomes broader. That is the beginning of the stomach, and the pancreatic anlagen are located just below that on each side of the gut (Fig. 2A,F). The next step is to separate the gut from the neural tube. 9. Hold the embryo by one pair of forceps and insert the tips of another pair between the gut and the neural tube rostral to the stomach. 10. Carefully open and close the forceps to separate the tissues from each other, and move caudally along the axis by opening and closing the tweezers repeatedly (Fig. 2F). 11. Detach the gut from the rest of the embryo (Fig. 2G). Before transferring the gut to the membrane try to remove as much of the heart tissues connected to the gut tube as possible (Fig. 2H,I). 12. Transfer the gut by mouth pipet and place it on the MilliCell-CM insert. 13. Maintain the cultures for 7 d in a humidified incubator at 37°C with 5–10% CO2. 14. Change the medium every other day by transferring the inserts to the next well (Fig. 3).
3.2.2. Isolating and Culturing Pancreatic Rudiments Derived from E10.5 Mouse Embryos Isolating and culturing segments of the gut containing pancreatic rudiments from E10.5 embryos is similar to the procedures described above. In addition, at this stage we are also able to perform a more defined dissection by isolating only the pancreatic rudiments (both the dorsal and the ventral rudiments) without associated foregut structures. The culturing conditions are the same as for E9.5 explants. Moreover, it is also possible to separate the pancreatic epithelium from its surrounding mesenchyme to culture the epithelium alone or to recombine the mesenchyme with epithelium from other explants. 1. Perform the dissection as described in steps 1–11 in Subheading 3.2.1. The isolated gut should be similar to the one shown in Fig. 4. 2. To isolate the intact dorsal pancreatic rudiment pinch the area indicated by arrows in Fig. 4. This will disconnect the pancreatic bud from the gut. Next try to cut the mesenchyme by pinching along the indicated line in Fig. 4. Now the intact dorsal pancreatic bud is isolated. If you wish to culture the intact bud, go directly to step 4.
266
266
Esni et al.
Primary Explant Cultures of Pancreas
267
Fig. 3. Simplified method for changing explant medium every other day. Transferring the inserts to the next well minimizes risk of dislodging or disrupting the explant during aspiration and replacement of media.
3. To separate the epithelium from the surrounding mesenchyme, try to grip the intact bud with one pair of tweezers and carefully remove the mesenchyme with another pair of tweezers. Broken lines in Fig. 4 indicate the border between epithelium and mesenchyme. 4. Transfer the tissue by mouth pipet and put it on the MilliCell-CM insert. If heterologous recombination of epithelium and mesenchyme from different genetic backgrounds is desired, at this point add the heterologous mesenchyme by mouth pipet and put it for example on top of a naked epithelium. To isolate the ventral pancreatic rudiment, follow the same procedure described above.
3.2.3. Adenoviral Infection of Embryonic Pancreatic Explants We have developed a method for infection of pancreatic buds by recombinant adenoviruses. By infecting the intact bud, only the surrounding mesenchyme is efficiently transduced to express adenoviral genes. This provides the opportunity to selectively manipulate the mesenchyme and characterize the Fig. 2. (Opposite page) Dissection of E9.5 foregut segment containing embryonic pancreatic anlagen. Whole mount immunohistochemistry, using antibodies for detection of Pdx-1 in E9.5 mouse embryo (20). (A) Different steps of procedures described in Subheading 3.2.1. (B–I). vb, Ventral bud; db, dorsal bud; h, heart; g, gut.
268
Esni et al.
Fig. 4. Whole mount immunohistochemistry of E10.5 foregut segment containing pancreatic anlagen, using antibodies against Pdx-1.
subsequent effects on the epithelium. However, it is also possible to isolate the epithelium from its mesenchyme, infect it, and then culture it with or without mesenchymal recombination (Fig. 5). These two approaches thus offer complementary techniques for selective genetic manipulation of either epithelium or mesenchyme. 1. Transfer the intact or naked pancreatic bud onto MilliCell inserts as described in Subheading 3.2.2. 2. Add recombinant adenovirus to PCM to obtain 1 ´ 107 fp/mL of diluted viral stock. Two hundred microliters is required for each bud. 3. After mixing add 200 µL into the MilliCell insert, and incubate for 24 h at 37°C with 5–10% CO2. Make sure that the tissue is submerged in the mixture. 4. After 24 h, carefully remove the infecting mixture by pipet, and transfer the membrane to a new well and culture as usual. If you are performing a heterologous recombination on the infected naked epithelium, follow the procedures described in Subheading 3.2.2.
4. Notes 1. Male mice between ages 4–8 wk give the best results. Prep should be done in the morning. The animal should be killed by 10 AM).
Primary Explant Cultures of Pancreas
269
Fig. 5. Ad-GFP infection of E10.5 dorsal pancreatic epithelium. Phase-contrast (A,C) images of pancreatic epithelium expressing GFP (B,D) with (C,D) or without (A,B) mesenchyme recombination. Adenoviral infection performed with 1 ´ 107 pfu/ mL for 24 h. GFP expression was detected on d 2 after infection.
2. The serum itself will induce a slight phenotypic switch secondary to the growth factors within the serum. If you wish, you can do this experiment without serum in the media. The results are still quite good. 3. We change the media and add new growth factor on the day following the prep. We then change the media every other day and add new growth factor, 0.1 mg/mL of trypsin inhibitor, and 1 µg/mL of dexamethasone, and so forth. 4. During incubation vortex very gently three or four times at regular interval to prevent cell aggregation at the bottom of the tube.
Acknowledgments This work was supported by NIH Grant DK56211-01 (to S. D. L.). Dr. Leach is also supported by the Paul K. Neumann Professorship in Pancreatic Cancer at Johns Hopkins University. The authors wish to thank Dr. Ingrid Meszoely and Dr. Anna Means for their assistance in developing these techniques.
270
Esni et al.
References 1. 1 Meszoely, I. M., Means, A. R., Scoggins, C. R., and Leach, S. D. (2001) Developmental aspects of early pancreatic cancer. Cancer J. 7, 242–250. 2. 2 Lowenfels, A. B., Maisonneuve, P., and Whitcomb, D. C. (2000) Risk factors for cancer in hereditary pancreatitis. International Hereditary Pancreatitis Study Group. Med. Clin. North Am. 84, 565–573. 3. 3 Lowenfels, A. B., Maisonneuve, P., Cavallini, G., et al. (1993) Pancreatitis and the risk of pancreatic cancer. International Pancreatitis Study Group. N. Engl. J. Med. 328, 1433–1437. 4. 4 Wagner, M., Luhrs, H., Kloppel, G., Adler, G., and Schmid, R. M. (1998) Malignant transformation of duct-like cells originating from acini in transforming growth factor transgenic mice. Gastroenterology 115, 1254–1262. 5. O’Reilly, L. A., Gu, D., Sarvetnick, N., et al. (1997) -cell neogenesis in an animal model of IDDM. Diabetes 46, 599–606. 6. 6 Fernandes, A., King, L. C., Guz, R., et al. (1997) Differentiation of new insulinproducing cells is induced by injury in adult pancreatic islets. Endocrinology 138, 1750–1762. 7. 7 Wang, R. N., Klöppel, G., and Bouwens, L. (1995) Duct- to islet-cell differentiation and islet growth in the pancreas of duct-ligated adult rats. Diabetologia 38, 1405–1411. 8. 8 Gu, D. and Sarvetnick, N. (1993) Epithelial cell proliferation and islet neogenesis in IFN-γ transgenic mice. Development 118, 33–46. 9. 9 Sharma, A., Zangen, D. H., Reitz, P., et al. (1999) The homeodomain protein IDX-1 increases after an early burst of proliferation during pancreatic regeneration. Diabetes 48, 507–513. 10. 10 Song, S. Y., Gannon, M., Washington, M. K., et al. (1999) Expansion of Pdx1expressing pancreatic epithelium and islet neogenesis in transgenic mice overexpressing TGF. Gastroenterology 117, 1416–1426. 11. 11 Yuan, S., Rosenberg, L., Paraskevas, S., Agapitos, D., and Duguid, W. P. (1996) Transdifferentiation of human islets to pancreatic ductal cells in collagen matrix culture. Differentiation 61, 67–75. 12. 12 Schmied, B. M., Liu, G., Matsuzaki, H., et al. (2000) Differentiation of islet cells in long-term culture. Pancreas 20, 337–347. 13. 13 Krakowski, M. L., Kritzik, M. R., Jones, E. M., et al. (1999) Pancreatic expression of keratinocyte growth factor leads to differentiation of islet hepatocytes and proliferation of duct cells. Am. J. Pathol. 154, 683–691. 14 Yuan, S., Duguid, W. P., Agapitos, D., Wyllie, B., and Rosenberg, L. (1997) Phe14. notypic modulation of hamster acinar cells by culture in collagen matrix. Exp. Cell Res. 237, 247–258. 15 Rooman, I., Heremans, Y., Heimberg, H., and Bouwens, L. (2000) Modulation of 15. rat pancreatic acinoductal transdifferentiation and expression of Pdx-1 in vitro. Diabetologia 43, 907–914.
Primary Explant Cultures of Pancreas
271
16. 16 Sandgren, E. P., Luetteke, N. C., Palmiter, R. D., Brinster, R. L., and Lee, D. C. (1990) Overexpression of TGF alpha in transgenic mice: Induction of epithelial hyperplasia, pancreatic metaplasia, and carcinoma of the breast. Cell 61, 1121–1135. 17. 17 Kim, J. H., Ho, S. B., Montgomery, C. K., and Kim, Y. S. (1990) Cell lineage markers in pancreatic cancer. Cancer 66, 2134–2143. 18. Sessa, F., Bonato, M., Frigerio, B., et al. (1990) Ductal cancers of the pancreas 18 frequently express markers of gastrointestinal epithelial cells. Gastroenterology 98, 1655–1665. 19. Gittes, G. K. and Galante, P. E. (1993) A culture system for the study of pancre19 atic organogenesis. J. Tissue Cult. Meth. 15, 23–28. 20. Gittes, G. K., Galante, P. E., Hanahan, D., Rutter, W. J., and Debas, H. T. (1996) Lineage-specific morphogenesis in the developing pancreas: Role of mesenchymal factors. Development 122, 439–447. 21. He, T. C., Zhou, S., da Costa, L. T., Yu, J., Kinzler, K. W., and Vogelstein, B. 21 (1998) A simplified system for generating recombinant adenoviruses. Proc. Natl. Acad. Sci. USA 95, 2509–2514. 22. Ahlgren, U., Jonsson, J., and Edlund, H. (1996) The morphogenesis of the pan22 creatic mesenchyme is uncoupled from that of the pancreatic epithelium in IPF/ PDX1-deficient mice. Development 122, 1409–1416. 23. Esni, F., Johansson, B. R., Radice, G. L., and Semb, H. (2001) Dorsal pancreas agenesis in N-cadherin-deficient mice. Dev. Biol. 238, 202–212.
272
Esni et al.
Zebrafish as a Research Model
273
18 Zebrafish as a Model for Pancreatic Cancer Research Nelson S. Yee and Michael Pack Summary Elucidation of basic mechanisms that regulate pancreatic organogenesis may help define molecular pathways involved in the development of exocrine pancreas cancer. The zebrafish has emerged as a powerful model for genetic dissection of the mechanisms underlying vertebrate organogenesis including formation of the pancreas. Unique properties of zebrafish enable genetic and embryological analyses not feasible using other vertebrate model organisms. The optical clarity of the zebrafish embryos allows visual detection of markers for pancreatic morphogenesis and cytodifferentiation by whole mount immunohistochemistry and RNA in situ hybridization. This feature, coupled with the accessibility of the externally fertilized zebrafish embryo and the small size and fecundity of adult zebrafish, facilitates large-scale forward genetic screens using chemical or insertional mutagenesis techniques. Furthermore, these properties allow high throughput studies that target functions of known genes via antisense or enforced expression studies. Together, such studies are predicted to identify novel genes, or known genes essential for pancreas development. Work in zebrafish is predicted to complement research performed using other vertebrate model organisms, and may help identify markers that define early stages of pancreatic tumorigenesis as well as potential targets for therapy. Key Words: Zebrafish; immunohistochemistry; in situ hybridization; neoplasms; pancreas development; pancreatic cancer.
1. Introduction 1.1. Developmental Studies Relevant to Pancreatic Cancer Research Despite recent advances in the diagnosis and treatment of many human malignancies, pancreatic ductal adenocarcinoma, the most common cancer of the pancreas, remains an oncologic challenge. The disease is associated with increasing incidence and high mortality. Typically, at the time of diagnosis, the tumor is advanced locally or has metastasized distantly, and there is no effective therapy (reviewed in ref. 1). One important factor contributing to the late diagnosis of pancreatic ductal adenocarcinoma is the lack of sensitive and speciFrom: Methods in Molecular Medicine, Vol. 103: Pancreatic Cancer: Methods and Protocols Edited by: G. Su © Humana Press Inc., Totowa, NJ
273
274
Yee and Pack
fic tumor markers to aid in its early detection. Unlike with colorectal carcinoma, there is no simple clinically feasible means to screen for the presence of nonmalignant precursors in high-risk patients. Identification of tumor markers in serum or stool, or that can be imaged in vivo, could in theory allow early diagnosis and curative therapy of this important disease. Attempts to identify tumor markers for pancreatic ductal adenocarcinoma have been made through molecular analyses of pathological specimens. These studies have revealed the genetic alterations associated with the progressive transformation of the pancreatic duct cell into an invasive carcinoma (reviewed in ref. 2). These alterations include activating mutations of oncogenes and inactivating mutations of tumor suppressor genes. Unfortunately, these molecular changes are nonspecific for pancreatic ductal adenocarcinoma. Recently, tumor markers have been identified by serial analysis of gene expression using human cancer cell lines and primary pancreatic cancer tissues (3–6). However, geneprofiling studies of pancreatic cancer are not designed to identify markers of early tumors or transformed cells. Despite their obvious clinical relevance, these studies are not suited to address fundamental questions regarding the identity of the cell of origin of pancreatic ductal adenocarcinoma, and the molecular events that lead to preneoplastic lesions in the pancreas. Developmental analyses of pancreatic organogenesis may provide the answers to some of these questions. This is partly because the initiation and progression of malignant neoplasms recapitulate certain aspects of normal development and growth. The protooncogenes and tumor suppressor genes as well as the growth factors and their receptors involved in the early transformation processes of pancreatic ductal adenocarcinoma have been implicated in the regulation of normal development of the exocrine pancreas (reviewed in ref. 7). Developmental studies focusing on the pancreatic stem cells and progenitors may help resolve questions regarding the identity of the cell of origin of pancreatic ductal adenocarcinoma. Although pancreatic ductal cells are generally believed to be the cell of origin of human pancreatic ductal adenocarcinoma, transdifferentiation or metaplasia of an acinar cell or islet cell into a duct-like cell may occur during pancreatic tumorigenesis (reviewed in ref. 8). Identification of factors that regulate cell fate specification of pancreatic progenitor cells as well as their proliferation and differentiation are likely to be essential to understanding the molecular events that normally prevent transformation of such pancreatic cells. 1.2. Recent Advances in Pancreas Development Recent studies in mammals and birds have provided insight into the genetic basis of pancreas development. These studies have identified molecules that
Zebrafish as a Research Model
275
regulate the early steps of the pancreatic program including specification of the gut endoderm, initiation of the pancreatic anlage, specification of endocrine and exocrine lineages, as well as pancreatic morphogenesis and growth. Each of these steps is regulated by the interplay between intrinsic factors expressed in the pancreatic epithelium and extrinsic factors produced in the adjacent mesenchymal tissues (reviewed in refs. 9 and 10). The mammalian pancreas is formed by the fusion of ventral and dorsal anlage that arise from the primitive gut tube (reviewed in ref. 11). Multipotent progenitors of three pancreatic lineages (endocrine, acinar, and ductular) reside within both ventral and dorsal pancreatic anlage. Formation of both dorsal and ventral pancreatic buds is dependent on exclusion of sonic hedgehog (shh) expression in the prepancreatic endoderm of the dorsal and ventral foregut (12–14). Activin and fibroblast growth factor-2 (FGF-2) secreted from the notochord have been implicated in repression of endodermal expression of shh (13). Dorsal bud development also requires endodermal expression of hlxb9 (15,16) and N-cadherin expression in the surrounding mesenchyme (17). Growth of both pancreatic buds is regulated by the homeodomain genes pdx-1 and pbx (18–24). Specification of endocrine and exocrine pancreatic cell fates follows formation of the dorsal and ventral buds. The Notch signaling pathway is crucial for controlling the decision of the pancreatic epithelial cells to choose between endocrine and exocrine fates (25). Lack of Notch signaling activity leads to down-regulation of hairy/Enhancer of split 1 (hes1) and thus elevated expression of the basic helix–loop–helix (bHLH) protein neurogenin 3 (ngn3), a marker for islet progenitor cells (21,26–30). On the other hand, active Notch signaling results in elevation of Hes1 and down-regulation of ngn3, and thus generation of exocrine cells. Gene targeting studies have also implicated the genes isl-1 and NeuroD/Beta2 as having an early role in pancreas specification (31,32). Numerous transcription factors have been shown to play a role in endocrine differentiation. Development of the various hormone-expressing cells (,,, PP) is regulated by the paired box proteins Pax 4 (33) and Pax 6 (34), and the homeobox transcription factors Nkx2.2 (35) and Nkx6.1 (36). A transcription factor cascade controlled by hnf1a also plays a regulatory role in maintaining differentiated function of pancreatic cells (37). Extrinsic factors play a modulatory role in endocrine growth and morphogenesis. Transforming growth factor1 (TGF-1) (38) or activin receptor–mediated TGF- signaling (39) promotes endocrine development, whereas activation of the Notch (25) and Hedgehog signaling pathways (40) antagonizes growth of endocrine pancreas in addition to its effect on pancreatic specification. In contrast to endocrine differentiation, far fewer genes that regulate exocrine differentiation have been identified. The bHLH protein p48, the DNAbinding subunit of pancreas-specific transcription factor 1 (PTF-1) that controls
276
Yee and Pack
gene expression in the exocrine pancreas, has recently been shown to be required for commitment of exocrine cell fate and differentiation (41). Mesenchymal factors including activin (42), follistatin (43), FGF (44,45), laminin (46), and epidermal growth factor (47) are also essential for proliferation and branching morphogenesis in the development of pancreatic exocrine tissue. Despite these recent advances in the genetic regulation of endocrine pancreas development, major questions regarding development of exocrine pancreas remain unanswered. How the expression and activity of PTF-1 is regulated during exocrine development is unclear. The target genes regulated by PTF-1 in the exocrine progenitors remain to be identified. The mechanisms that regulate the activity of the mesenchymal factors that influence exocrine growth and morphogenesis are also largely unknown. Most genetic studies have so far failed to address the molecular basis for specification of the ductular and acinar cell fate and their lineage commitment. This is partly because there are few known markers that specifically label the precursor or committed progenitors of these exocrine cell types. 1.3. Zebrafish as a Model to Study Pancreatic Organogenesis Developmental studies offer a tremendous opportunity to identify genes that regulate organ development and they can serve as an important method to identify cell markers. Targeted mutagenesis of genes expressed in the developing endocrine pancreas have been particularly rewarding in this regard. However, such studies entailed the disruption of known genes whose expression within the pancreas was often well established. Furthermore, endocrine differentiation resembles in many regards neuronal differentiation, which has been extensively studied (48). In contrast, far less is known about the nature of pancreatic exocrine development and it is likely that many of its essential regulators will be novel genes. The zebrafish is an animal model system that is uniquely suited to study the molecular regulation of vertebrate organ development (49,50). The optical clarity of the zebrafish embryo and larva allows the visual identification of mutations that affect organ morphology and function. Large-scale forward mutagenesis screens have identified numerous mutations affecting the development and physiology of cardiac, hematopoietic, genitourinary, and gastrointestinal organ systems (51,52). Molecular characterization of several of these genes through positional techniques and candidate gene analysis has been reported (53–55). Molecular analysis of zebrafish mutations, albeit scientifically rewarding, is a labor-intensive effort. However, sequencing of the zebrafish genome, which is projected to be completed by the end of 2005 (http://www.sanger.ac. uk/Projects/D_rerio/advblast_server.shtml), will greatly simplify the molecular characterization of zebrafish mutations and is expected to facilitate their high throughput analysis.
Zebrafish as a Research Model
277
Genes identified in forward mutagenesis screens that regulate zebrafish pancreas development are for several reasons predicted to function similarly in mammals. First, pancreas anatomy and histology are similar in teleosts and mammals (56). The zebrafish pancreas consists of several principal islets that are located adjacent to the gallbladder as well as numerous smaller accessory islets that are embedded within exocrine tissue located in the intestinal mesentery. Zebrafish islets are organized in a nearly identical fashion to mammalian islets and they synthesize peptide hormones that can be localized immunohistochemically using antibodies raised to mammalian insulin, glucagon, and somatostatin (Table 1). Zymogen-rich acinar cells surround the principal and numerous accessory islets and express digestive enzymes such as amylase, peptidases, and lipases that also can be identified immunohistochemically using antibodies raised against mammalian proteins (Table 1), or biochemically (60). The exocrine cells are arranged into acini that are connected to the intestine through a complex ductular network. As in mammals, the zebrafish main pancreatic duct and hepatic duct are joined at their insertion site into the intestine. A second reason conserved genes are predicted to regulate zebrafish and mammalian pancreas development in a similar fashion is that there already exists a considerable amount of evidence that orthologous signaling pathways and transcription factors regulate pancreas development in these organisms. For example, shh plays an essential role during zebrafish and mammalian development, although their roles has been reversed. shh is a positive regulator of pancreas development in zebrafish (61,62), but a negative regulator in mammals and birds (12–14). Importantly, it is also known that shh is also a positive regulator of pancreas function in adult mammals (63), suggesting that this function of shh may have evolved from teleosts. Studies using morpholinos-induced gene knockdown have demonstrated a conserved role of the homeodomain protein PDX-1 in formation of the pancreatic exocrine and endocrine tissues in zebrafish larvae (59,64) and mammalian embryos (18,19). In addition, the zebrafish orthologs of isl-1 and other genes known to regulate mammalian pancreas development are also expressed in the developing zebrafish pancreas (Table 2), suggesting that their role have been evolutionarily conserved as well. For these reasons, zebrafish genes identified in forward mutagenesis screens for pancreatic phenotypes are predicted to play related roles during mammalian pancreas development. In the original morphology-based screens, mutations affecting the differentiation and maintenance of the exocrine pancreas and intestinal epithelium were recovered (56). In a recent mutagenesis screen, we also have identified mutations that specifically affect exocrine development using pancreatic cell-specific markers (Yee, N. S. and Pack, M., unpublished). Preliminary studies suggest that the affected genes regulate aspects of exocrine pancreas development that may relate to events that occur during cell transfor-
278
Yee and Pack
Table 1 Primary and Secondary Antibodies Used for Immunohistochemistry of Exocrine Enzymes and Islet Hormones in Zebrafish Pancreas A. Primary antibodies
Sources
Examples of use (ref.)
Antibodies against exocrine enzymes Rabbit anti-bovine carboxypeptidase A Rabbit anti-bovine carboxypeptidase A
Rockland, Inc. Chemicon
(56) (57)
Antibodies against islet hormones Guinea pig anti-porcine insulin Guinea pig anti-porcine insulin Guinea pig anti-porcine glucagon Rabbit anti-porcine glucagon Rabbit anti-porcine glucagon Mouse anti-porcine glucagon Rabbit anti-somatostatin-14 Rabbit anti-somatostatin-14
Linco Dako Linco Dako Biotrend Sigma Biotrend Genosys
(56) (58) (59) (59) (57) (57) (57) (58)
Note: Localization of pancreatic polypeptide in the zebrafish islet has been reported by some but not other investigators (57, 58, Yee, N. S. and Pack, M., unpublished).
B. Secondary antibodies Antibodies against rabbit IgG Fluorescein (FITC)-conjugated AffiniPure goat anti-rabbit IgG Alexa Fluor™ 488 goat anti-rabbit IgG conjugate TRIC-conjugated goat anti-rabbit IgG Alexa Fluor™ 546 goat anti-rabbit IgG conjugate Antibodies against guinea pig IgG FITC-conjugated goat anti-guniea pig IgG Alexa Fluor™ 488 goat anti-guniea pig IgG conjugate Texas Red™ dye-conjugated AffiniPure goat anti-guniea pig IgG Alexa Fluor™ 594 goat anti-guniea pig IgG conjugate Alexa Fluor™ 546 goat anti-guniea pig IgG conjugate
Sources
Examples of use (ref.)
Jackson ImmunoResearch Molecular Probes
(56)
Genosys Molecular Probes
(58) (57)
Sigma Molecular Probes
(58) (57)
Jackson ImmunoResearch Molecular Probes
(56)
Molecular Probes
(57)
(59)
(59)
Zebrafish as a Research Model
279
Table 2 Zebrafish Orthologs of Mammalian Genes Suitable for Use as Molecular Markers of Pancreas Development Zebrafish genes
Accession number for sequences (National Center for Biotech. Info.)
Source and examples (ref.)
Transcription factors as markers of specification of pancreatic fate PDX-1 AF036325 (65) Islet1 D21135 (57,66) Nk2.2 X85977 (57,67) Pax6.2 AF061252 (57,68) GATA6 AF191578 (56) Prox1 AF063018 (61,69) NeuroD AF036148 (62,70) Axial/FoxA2/fkd1 NM-130949 (57,71) FoxA3/fkd2/zffkh1 NM-131299 (57,72) Digestive enzymes as markers of exocrine differentiation Carboxypeptidase A AF376130 Trypsin AJ297822
(62) (57)
Islet hormones as markers of endocrine differentiation Preproinsulin AJ237750 Glucagon AJ133697 Somatostatin AJ238017
(58) (58) (58)
mation in the mammalian exocrine pancreas. Functional analysis and molecular characterization of these genes are predicted to advance our understanding of the molecular pathways regulating development of the exocrine pancreas as well the mechanisms that normally prevent pancreatic cell transformation. 1.4. Techniques Employed in Studying Developmental Biology of Zebrafish Pancreas Analysis of pancreas development in the zebrafish relies on visualization of the three pancreatic lineage cells at various developmental stages. The optical clarity of zebrafish embryos and larvae allow direct visualization of the pancreas in whole mounts using in situ hybridization and immunohistochemistry (Fig. 1A,B,D). However, at many time points, the pancreas of embryonic and larval zebrafish cannot be directly visualized because of its location relative to pigmented cutaneous cells. For this reason, mutations affecting exocrine development in the original morphology-based mutagenesis screens were recovered in association with mutations that perturb the intestinal epithelium, which is easily
280
Yee and Pack
Fig. 1. Zebrafish larval pancreas. (A) Brightfield lateral view of a 5 dpf larva (arrow points to pancreas). (B) Immunohistochemistry of 5 dpf exocrine pancreas (paraffin section: carboxypeptidase A, green; insulin, red; i, intestine). (C) Sagittal histological plastic section of a 5 dpf larva (e, exocrine pancreas; i, islet; y, yolk). (D) GATA-6 expression in the 4 dpf exocrine pancreas (whole-mount in situ hybridization of a PTUtreated larva). (Illustration is in color in e-book.)
visualized in zebrafish larvae (56). Chemical inhibition of the development of cutaneous pigmentation is, however, feasible in zebrafish. This technique, coupled with immunohistochemical detection of pancreas differentiation markers, has enabled us to identify mutations that specifically perturb exocrine development (Yee, N. S. & Pack, M., unpublished). Visualization of the anatomical relationship of the pancreas, liver, and intestine in whole-mount specimens of larval stage zebrafish is enhanced by manually removing the skin. Skin removal has the added benefit of eliminating most of the melanophores overlying the pancreas. Microscopic anatomy of the pancreatic islets and exocrine tissues is often required for the analysis of mutant and wild-type fish. Tissue resolution of embryonic and larval stage zebrafish is optimal in plastic, rather than paraffin sections (Fig. 1C). Specimens processed as whole mounts for in situ hybridization or immunohistochemistry can also be analyzed histologically (Fig. 2), provided care is taken to avoid loss of signal intensity.
Zebrafish as a Research Model
281
Fig. 2. Cell proliferation in the 48 hpf zebrafish pancreas. Plastic section of a wholemount specimen processed for BrdU and insulin immunohistochemistry. BrdU incorporation reveals proliferating cells (green) that surround the insulin-expressing islet cells (red). (Illustration is in color in e-book.)
The optical clarity of zebrafish embryos and larvae allow visualization of transcripts expressed in the pancreas in whole-mount specimens. Techniques for RNA in situ hybridization in zebrafish are modifications of standard protocols used with other model organisms. Identification of pancreatic precursors and differentiated pancreatic cells can be accomplished using the zebrafish orthologs of established vertebrate genes (Table 2). Such cells can also be identified immunohistochemically using antibodies raised against mammalian endocrine hormones and exocrine enzymes (Table 1). Cell proliferation within the developing pancreas can be monitored immunohistochemically by detection of microinjected 5-bromo-2'-deoxyuridine (BrdU) (Fig. 2), or using antibodies that recognize proliferating cell nuclear antigen (PCNA) or an M-phase specific phospho-histone (73). Molecular analyses of zebrafish pancreas development can be performed using a recently described gene knockdown technique that uses modified antisense oligonucleotides (morpholinos) that block gene translation (74). Antisense morpholinos injected into fertilized zebrafish embryos at the one-cell stage provide a highly specific inhibition of gene function that can phenocopy established mutations. This technique is useful to examine the role played by the zebrafish orthologs of vertebrate genes with established roles in pancreas development. Antisense
282
Yee and Pack
morpholinos are also a powerful way to evaluate the function of newly identified genes or genes not previously known to contribute to pancreas development. 1.5. Summary Developmental studies offer a unique opportunity to identify molecular determinants that regulate biological processes involved in human pancreatic cancer (75). Rapid gene identification in zebrafish, either through mutagenesis studies or gene knockdown techniques, may define gene targets or cellular processes that can be analyzed in a clinically relevant context. In this regard, work performed using zebrafish will complement related work from the mouse and other model systems. 2. Materials 2.1. Fish Husbandry 1. E3 medium: Stock 60X E3 medium is prepared by adding 17.59 g of NaCl, 0.76 g of KCl, 2.91 g of CaCl2, and 2.38 g of MgSO4 to 1 L of H2O (Millipore). Dilute 1:60 with H2O for use and keep at 28.5°C. 2. Tricaine: To prepare stock solution, dissolve 0.4 g of Tricaine powder (3-aminobenzoic acid ethyl ester, Sigma) in 97.9 mL of H2O. Add 2.1 mL of 1 M Tris HCl solution, pH 9. Adjust the pH to about 7. Store at 4°C. Make a 1:25 dilution in E3 medium prior to use. Tricaine stock solution can also be added directly to embryos and larvae in Petri dishes (final dilution 1:25). 3. Methylene blue: Stock solution 0.03 M methylene blue (Fisher) in H2O. Store at 4°C. Add 20 µL of 0.03 M methylene blue solution to 1 L of E3 medium prior to use and store at 28.5°C. 4. PTU/E3 medium: Prepare 10X stock solution (2 mM) of 1-phenyl-2-thiourea (PTU) by adding 304 mg of PTU powder in 1 L of E3 medium. Store at 28.5°C. Dilute 1:10 in E3 medium for use.
2.2. Whole-Mount Immunohistochemistry 1. PBST: Phosphate-buffered saline supplemented with 0.1% Tween-20 (Fisher). 2. Paraformaldehyde fixative: Prepare 4% paraformaldehyde solution by adding 4 g of paraformaldehyde (Fisher) in 100 mL of PBS in a chemical hood. Repeated heating at 65°C and vigorous mixing is necessary for dissolution. Store in 10-mL aliquots at 80°C. 3. 0.1% Collagenase: Dissolve collagenase (Sigma) in PBST prior to use. 4. Proteinase K: Stock solution (20 mg/mL) is prepared by dissolving proteinase K (Boehringer) in H2O, and then adding equal volume of glycerol (Fisher). Store at 20°C. Dilute with PBST prior to use. 5. 5% Goat serum: Dilute goat serum (GibcoBRL) with PBST; keep at 4°C. 6. Primary and secondary antibodies (see Table 1): In general, the antibodies can be diluted at 1:100 in 5% goat serum.
Zebrafish as a Research Model
283
7. Orbitron Rotator II (Boekel Industries, Inc., Model 260250). 8. Leica fluorescent stereomicroscope (Leica MZFLIII). 9. 3% Methylcellulose: Prepare by first heating 50 mL of E3 medium in a 500-mL beaker to 60°C. Add 1.5 g of methylcellulose (Sigma M-0387) and stir with a long glass rod. Place the mixture in a 20°C freezer. Every 30 min, repeat the stir–freeze cycle until the mixture freezes. Keep the mixture at 4°C overnight. On the next day, the solution should be clear and free from speckles of undissolved methylcellulose and is stored in 1 mL aliquots at 20°C. Prior to each use, the methylcellulose solution should be centrifuged at 20,000g for 5 min at room temperature (RT) to remove residual undissolved methylcellulose.
2.3. Histology 1. JB-4 Plus Catalyst (Polysciences). 2. Activated JB-4 Plus solution A: Add JB-4 Plus Catalyst to JB-4 Plus Monomer Solution A (Polysciences) at a final concentration of 1%. Shake vigorously at RT to dissolve catalyst in solution A. Keep at 4°C in a foil-wrapped bottle to protect from light (see Note 1). 3. JB-4 Plus Accelerator Solution B (Polysciences). 4. Polyethylene Molding Cup Trays (Polysciences). 5. Plastic block holder (Polysciences). 6. Microtome: Leica Model RM 2155. 7. Glass knives: Ted Pella, Inc. (Reading, CA). 8. Dumont cat. no. 55 forceps (Fine Sciences Tools, Inc.). 9. Zeiss Axioplan compound microscope. 10. Superfrost/Plus 25 75 1.0 mm microscope slides (Fisher). 11. Methylene blue-azure II dye solution (100 mL): Dissolve 0.130 g of methylene blue and 0.020 g of azure II in 50 mL of distilled water. Then add 10 mL of glycerol, 10 mL of methanol, and 30 mL of phosphate buffer, pH 6.9. The phosphate buffer is prepared by adding 1.067 g of Na2HPO4 and 0.908 g of KH2PO4 to distilled water up to 100 mL. 12. Permount™ mounting medium (Fisher). 13. Vectashield™ mounting medium with DAPI (Vector Laboratories).
2.4. RNA In Situ Hybridization 1. Phenol–chloroform–isoamyl alcohol (25:24:1, pH 6.7 ± 0.2) (Fisher Scientific). 2. Sodium acetate solution: 3 M sodium acetate dissolved in H2O, titrated to pH 5.2 using glacial acetic acid; autoclave. 3. 5X Transcription buffer (Promega or Boehringer). 4. RNA polymerase T7 or SP6 (Promega or Boehringer). 5. 10X Digoxigenin/RNA labeling mix (Enzo). 6. 10X Dithiothreitol (DTT). 7. RNase-free DNase. 8. Lithium chloride solution: 4 M LiCl in H2O.
284
Yee and Pack
9. Microspin™ S-400 HR columns (Amersham Pharmacia Biotech). 10. Proteinase K solution: Stock solution 20 of mg/mL of proteinase K (Boehringer) in 50% glycerol–50% H2O. Diluted with PBS-T prior to use. 11. Prehybridization buffer: 50% Formamide, 5X saline sodium citrate (SSC), 50 µg/ mL of heparin (stock solution 50 mg/mL of heparin in H2O, Sigma), 0.1% Tween20, 9.2 mM citric acid (stock solution 0.5 M citric acid titrated to pH 6 using 10 N NaOH), 500 µg/mL of tRNA (Boehringer Mannheim). Store at 20°C. 12. 2X SSC in H2O containing 0.1% Tween-20. 13. 0.2X SSC in H2O containing 0.1% Tween-20. 14. Anti-digoxigenin–alkaline phosphatase Fab fragments (Enzo): Dilute 1:2000 with 5% goat serum prior to use. 15. Predetection buffer: 100 mM Tris-HCl, pH 9.5, 100 mM NaCl, 50 mM MgCl2, 0.1% Tween-20. 16. 4-Nitroblue tetrazolium chloride (NBT) solution: To prepare a stock solution of 75 mg/mL of NBT in 70% DMF, mix 1 mL of 100 mg/mL NBT 70% DMF (Boehringer Mannheim) and 0.33 mL of 70% DMF in H2O. Store at 20°C. 17. 5-Bromo-4-chloro-3-indolyl-phosphate (BCIP) solution: 50 mg/mL of BCIP in DMF (Boehringer Mannheim). Store at 20°C. 18. Detection buffer contains 0.3375 mg of NBT and 0.175 mg of BCIP per milliliter of predetection buffer prepared prior to use. 19. 20 mM EDTA solution: Prepare 0.5 M stock solution EDTA in H2O, and dilute to 20 mM with PBS-T.
2.5. Embryo Injection 1. Agarose 2% solution in E3 medium. 2. Injection needles: 1.0-mm outer diameter thin-walled borosilicate capillary tubes with filament (World Precision Instruments). 3. Flaming/Brown Micropipette Puller (Sutter Instrument Co.). 4. Manual microinjector and semi-automatic microinjector (Medical Systems Research Products, Model PLI-90). 5. Phenol red solution: Prepare working solution by diluting two parts of 0.5% stock solution of phenol red in DPBS (Sigma) with one part of 0.1 M KCl. Centrifuge the stock solution of phenol red through a 0.22-µm filter unit (Millipore Ultrafree™MC) prior to use. 6. Bromodeoxyuridine (BrdU): Prepare a 30 mM stock solution of BrdU in water. A working solution of BrdU for injection is prepared by mixing nine parts of 600 µM BrdU solution with one part of working solution of phenol red (as in step 5). Store the stock and working solutions of BrdU at 80°C. 7. 3 N Hydrochloric acid (HCl) solution: Dilute 5 N HCl with H2O. 8. Mouse anti-BrdU IgG (Boehringer Mannheim): Use at a 1:100 dilution with 5% goat serum. 9. Fluorescein isothiocyanate (FITC) goat anti-mouse IgG (Boehringer Mannheim). 10. Morpholinos (Gene Tools): Design morpholinos by making antisense oligonucleotides as directed by the manufacturer (http://198.88.150.227/genetools). Prepare a
Zebrafish as a Research Model
285
1 mM stock solution in 1X Danieau buffer (58 mM NaCl, 0.7 mM KCl, 0.4 mM MgSO4, 0.6 mM Ca(NO3)2, 5.0 mM N-2-hydroxyethylpiperazine-N'-2-ethanesulfonic acid (HEPES), pH 7.6) and store at 80°C in aliquots.
3. Methods 3.1. Fish Husbandry Methods for raising and maintaining zebrafish stocks are described in detail in The Zebrafish Book (76); also available on line at http://zfin.org/zf_info/zfbook/ zfbk.html). To generate embryos, pairwise matings of adult fish are arranged in the evening. Male and female fish are housed in separate compartments of the mating chambers to prevent injuries that can occur to fish that are housed in close proximity. Fish are reunited at the beginning of the light cycle the following day and allowed to mate. Mating typically begins within 30–60 min of the onset of the light cycle and can continue for several hours. Embryos are recovered 1–4 h post-fertilization (hpf); fertilized eggs collected at later time periods are also acceptable if the water in the mating chamber is clean. Eggs are collected from mating chambers using a tea strainer and subsequently rinsed with E3 medium and then distributed to 10-cm Petri dishes (maximum 60 fertilized embryos per dish) containing E3 medium supplemented with 0.6 µM methylene blue to inhibit growth of contaminating fungi and other organisms. Developing embryos are raised in Petri dishes in a refrigerated incubator at 28.5°C. Embryos and embryo media must be inspected at least once per day. Embryos that are malformed or growth-delayed are discarded. The E3 medium for each Petri dish is also exchanged daily. Zebrafish embryos typically hatch on the second day post-fertilization (dpf; 48–72 hpf). Discard all ruptured chorions from the Petri dishes of each clutch so that the E3 medium does not support growth of contaminating organisms. This is typically done on the third dpf. Visual inspection of developing embryos and larvae is performed using a stereo dissecting microscope with illumination from a halogen light source. Tricaine (aminobenzoic acid ethyl ester) is used as a paralytic and anesthetic agent when prolonged observation of developing embryos older than 24 hpf is necessary. Embryos exposed to Tricaine must be placed in fresh E3 within 30 min to avoid toxicity. Visualization of embryonic and larval zebrafish digestive organs in wholemount preparations is often obscured by overlaying melanophores. Pigmentation can be inhibited chemically by raising embryos and larvae 0.2 mM PTU in E3 medium. PTU is added to the E3 medium at or before 24 hpf. Normal growth of PTU-treated embryos is dependent on daily exchange of PTU/E3 medium. Delayed growth is common even in well-maintained clutches. For this reason correlation of chronological and developmental age of PTU embryos and larvae with non–PTU-treated siblings is advised.
286
Yee and Pack
3.2. Whole-Mount Immunohistochemistry Methods for whole-mount immunohistochemistry of zebrafish embryos and larvae are similar to those used for the analysis of other model organisms. Embryos and larvae are typically fixed with paraformaldehyde, although other fixatives may be required to visualize specific tissue antigens. Fixed embryos are dehydrated with methanol and then rehydrated just prior to immunostaining. Dehydration improves tissue permeability. At stages beyond 24 hpf, supplemental permeabilization with collagenase or proteinase K is essential. Dehydrated embryos can be stored in 100% methanol at 20°C for extended periods of time. 3.2.1. Fixation 1. Fix anesthetized 5 dpf larvae in 1.6-mL tubes for 2 h in 4% paraformaldehyde supplemented with Tricaine (1:25 dilution) at RT, or at 4°C overnight. Embryos are gently rocked during fixation (see Notes 2 and 3). 2. Wash embryos three times in PBST (5 min per wash). 3. Dehydrate washed embryos using graded methanol (MeOH) solutions (50% MeOH/ PBST; 80% MeOH/PBST; 100% MeOH 2) and then store at 4°C.
3.2.2. Permeabilization 1. Rehydrate embryos prior to permeabilization by washing three times in PBST 2 min per wash. 2. Permeabilize rehydrated embryos by collagenase digestion by adding 1 mL of 0.1% collagenase solution to each 1.6-mL tube containing fixed embryos at RT for 45 min (see Notes 4 and 5). 3. Wash embryos three times with PBST following permeabilization. Embryos are gently agitated on a rocking platform during permeabilization and washes.
3.2.3. Primary Antibody Incubation 1. Incubate permeabilized larvae overnight at 4°C in a minimum of 100 µL of diluted primary antibodies. 2. Following the overnight incubation the primary antibody solution is removed and stored at 4°C for subsequent reuse. Embryos are then washed with PBS-T three times, 5 min per wash.
3.2.4. Secondary Antibody Incubation 1. Incubate larvae in fluorescent-conjugated secondary antibodies. Use a minimum of 100 µL diluted antibody solution per 1.6-mL tube. Incubate for a minimum of 2 h at RT or 4°C overnight. Tubes containing the embryos/larvae in secondary antibody solution are wrapped in aluminum foil to protect samples from light (see Note 6). 2. Following incubation, the samples are washed three times in PBST as described for the primary antibody incubation.
Zebrafish as a Research Model
287
3.2.5. Visualization of Whole-Mount Specimens 1. The specimen can be mounted in glycerol and examined under a fluorescent stereomicroscope. This can be accomplished by washing the specimen in graded glycerol– PBST solutions (50% glycerol–PBS-T; 70% glycerol–PBST; 95% glycerol–PBST 2; each wash 5 min). 2. To examine fluorescence in live larvae, the specimen can be mounted in 3% methylcellulose and viewed using a fluorescent stereomicroscope. For the latter, one anesthetized embryo/larva is place on a depression slide. Using a pipet, remove most of the liquid carryover. Next, cover the specimen with a small amount of methylcellulose supplemented with Tricaine. Arrange the specimen in the desired orientation.
3.3. Histology Histological sections of zebrafish embryos and larvae can supplement analyses of whole-mount specimens. To optimize tissue resolution and orientation, glycol methacrylate, as opposed to paraffin, is used as the embedding compound. 3.3.1. Infiltration 1. Serially dehydrate the larvae with graded ethanol (EtOH) solutions (50% EtOH– PBST, 70% EtOH–PBST and 95% EtOH; each wash 5 min at RT). 2. Infiltrate dehydrated larvae in activated JB-4 Plus solution A (500 µL per 1.6-mL tube, maximum of 20 larvae per tube) overnight at 4°C. Alternatively, specimens can be infiltrated at RT for 30 min to 1 h.
3.3.2. Embedding 1. Prior to embedding, prepare a foundation on which to place the embedded specimens by cutting a blank methacrylate block into two halves; the tapered halfblock is placed in the bottom of the molding tray. The foundation, which serves as a platform on which to rest the immunostained larvae, raises the specimens from the bottom of the molding tray. This facilitates cross-sectioning of the block. 2. The blank block is made by adding a mixture of 750 µL of activated JB-4 Plus solution A and 50 µL of JB-4 Plus solution B (vortex-mix vigorously) to a molding tray covered with a piece of plastic at RT overnight. Blank blocks are cut with a coping saw or related tool. 3. Place the foundation in the bottom of the molding tray and add enough of the activated JB-4 A and JB-4 B solution (15:1 v/v; vortex-mix vigorously) to fill the chamber completely. Transfer the larvae previously infiltrated with activated JB4 Plus solution A onto the foundation, and align them in parallel with their heads touching one wall of the molding chamber. Allow the JB-4 A/B solution at the bottom of the molding well to polymerize (15–20 min at RT) (see Note 7). 4. Once the specimens are fixed in the block, cover the molding well with a piece of plastic. Allow to block to harden, in the dark, at RT. Remove the block when solid (minimum 3 h) and attach to a block holder using super-glue for cross-sectioning (5-µm sections) (Fig. 3A).
288
Yee and Pack
Fig. 3. Histological sectioning. (A) Plastic block adhered to a block holder. Note the larvae being aligned in parallel with their heads against one side of the block. (B) Microtome with plastic block aligned such that the face of the block where the larval heads lie is parallel to the knife-edge.
3.3.3. Sectioning 1. Align the block on the microtome so that the face of the block is parallel to the knife edge. Secure the block so that it rests within 0.25 mm of the knife edge. Be certain that the top and bottom edges of the block are equidistant from the knife before beginning sectioning (Fig. 3B). 2. Gently advance the block toward the knife in 5-µm increments. Remove sections from the glass knife with watchmakers’ forceps. Drop sections into a dish of deionized water. Sections will unfold and float of the surface of the deionized water (see Note 8). 3. Collect sections on glass slides. Insert slide into deionized water and gently guide the section onto the slide as it is removed from the water. Place the slide in an upright position to remove excess water while cutting the next section (see Note 9). 4. Allow the slides to dry at room temperature in the dark. A standard glass slide can hold seven or eight plastic sections. 5. For light microscopic analysis, sections can be stained in methylene blue–azure II solution for 5 min at RT, rinsed in two changes of distilled water, and rinsed in a stream of distilled water for 1 min, air-dried, and then mounted in Permount™ mounting medium and cover slipped. 6. For fluorescent microscopic analysis, sections are mounted with a glycerol-based mounting medium (Vectashield™) supplemented with DAPI and cover slipped.
3.4. RNA In Situ Hybridization Standard protocols are followed for in situ hybridization with antisense RNA probes (77) (see Note 10). For larvae older than 48 hpf, care must be taken to ensure adequate tissue permeabilization (see Note 11).
Zebrafish as a Research Model
289
3.4.1. Generation of Antisense Riboprobe 1. Linearize 10 µg of plasmid with the appropriate restriction enzyme in a 50-µL reaction. One microliter of the reaction mixture is used to ensure complete digestion by agarose gel electropheresis. To the remaining 49 µL of DNA solution, add 50 µL of H2O and mix. 2. Extract proteins with 100 µL of Tris-buffered phenol–chloroform–isoamyl alcohol. Vortex-mix vigorously and centrifuge for 2 min at RT. Transfer 80 µL of the upper aqueous layer containing DNA to a fresh tube. 3. Precipitate the linearized plasmid by adding 8 µL of 3 M sodium acetate, pH 5.2, solution and 200 µL of 100% ethanol to the extracted DNA solution. Mix and keep at 20°C overnight. 4. Centrifuge at 12,000g (Eppendorf) at 4°C for 30 min, and discard the supernatant. Wash the DNA pellet twice with 1.5 mL of 80% ethanol. 5. Centrifuge at 12,000g (Eppendorf) at 4°C for 5 min. Discard the supernatant, and air-dry the DNA precipitate. Add 18 µL of H2O to dissolve the DNA, and store at 20°C. 6. To generate a digoxigenin-labeled riboprobe, a 20-µL transcription reaction mixture contains the following: 2 µL of linearized plasmid, 4 µL of 5X transcription buffer, 2 µL of 10X digoxigenin/RNA labeling mix, 2 µL of RNA polymerase, 1 µL of RNase inhibitor, 2 µL of 10X DTT, and 7 µL of H2O. 7. Mix the transcription reaction and incubate at 37°C for 2 h. Add 1 µL of RNasefree DNAse to digest the vector. Incubate at 37°C for 30 min. Use 1 µL of the transcription product for agarose gel electrophoresis to check for RNA. 8. To precipitate the riboprobe, add 2.5 µL of 4 M LiCl solution and 80 µL of 95% ethanol to the remaining 19 µL of transcription product. Mix and keep at –20°C overnight. Centrifuge at 12,000g for 30 min at 4°C, and discard the supernatant (see Note 12). 9. Wash the RNA precipitate with 1 mL of 80% ethanol and centrifuge at 12,000g for 10 min at 4°C. Air-dry the RNA precipitate. Redissolve the RNA in 100 µL of H2O, to which add 1 µL RNase inhibitor, and store the riboprobe at 80°C in small aliquots. 10. For preadsorption of riboprobe (to reduce background due to nonspecific binding), add 20 µL of riboprobe to 1.5 mL of prehybridization buffer in a 1.6-mL tube containing twenty larvae, incubate at 68°C overnight, and then collect the supernatant and store at 80°C.
3.4.2. In Situ Hybridization 3.4.2.1. REHYDRATION 1. Rehydrate fixed larvae that are stored at 20°C by washing larvae three times with PBST using 1.5 mL each wash for 2 min at RT. 3.4.2.2. PERMEABILIZATION 1. Incubate rehydrated embryos in proteinase K solution for 25 min at RT. Then wash the larvae three times with PBST using 1.5 mL each wash for 5 min at RT (see Note 11).
290
Yee and Pack
2. Following permeabilization, refix larvae in 1.5 mL of 4% PFA at RT for 20 min. Wash three times with PBST using 1.5 mL each wash for 5 min at RT.
3.4.2.3. PREHYBRIDIZATION
AND
HYBRIDIZATION (SEE NOTE 13)
1. Add 200 µL of prehybridization buffer and incubate at 68°C for 2–4 h. 2. Remove prehybridization buffer, add 200 µL of preadsorbed riboprobe (preheated to 68°C) and incubate at 68°C overnight. 3. Collect the supernatant and store at –80°C for reuse. 4. Wash larvae three times with 2X SSC (+ 0.1% Tween) using 1.5 mL each wash for 30 min at 68°C. 5. Wash three times with 0.2X SSC (+ 0.1% Tween) using 1.5 mL each wash for 30 min at 68°C. 6. Add 200 µL of 5% goat serum and incubate at RT for 1 h. 7. Remove supernatant, add 200 µL of diluted antidigoxigenin–alkaline phosphatase Fab fragments, and incubate at RT for 2 h. 8. Wash five times briefly with PBST using 1.5 mL each wash. Then wash with PBST overnight at 4°C.
3.4.2.4. HISTOCHEMICAL DETECTION
OF
DIGOXIGENIN PROBES
1. Wash hybridized specimens twice briefly with predetection buffer using 1.5 mL each wash. Incubate specimens in 1.5 mL of predetection buffer at RT for 1 h. 2. Remove supernatant, add 1.5 mL of detection buffer; incubate at RT and protect from light. 3. Examine the larvae at 1-h intervals to ensure optimal signal to background ratio. Overnight incubation in detection solution is often required for detection of lowlevel transcripts. 4. Stop the reaction by adding 1 mL of 20 mM EDTA solution. Wash overnight in PBST to reduce nonspecific background color. 5. Clear embryos in graded glycerol–PBST solutions (50% glycerol–PBST; 80% glycerol–PBST; 95% glycerol–PBST; 5 min each wash) to improve signal detection if desired. Specimens can be stored in either glycerol or PBST at 4°C. Visual inspection of the specimen is performed using a stereo dissecting microscope with overhead illumination from a halogen light source.
3.5. Embryo Injection Microinjection of zebrafish embryos and larvae is performed using a stereomicroscope and agarose ramps. The ramps serve to immobilize the embryos as they are punctured by the microinjection needle. 3.5.1. BrdU Injection 1. Microinjection needles are generated using a Flaming/Brown Micropipette Puller. Pulled pipets are opened by brushing the closed tip against the edge of razor blade using a stereomicroscope. Tip diameter will vary. Pressure and time of injection are varied to regulate injection volume. Fine control of BrdU dosage is not crucial.
Zebrafish as a Research Model
291
Fig. 4. Microinjection of embryos. (A) An agarose ramp as injection platform is prepared by placing a glass slide with the edge rested in the middle of the Petri dish cover. (B) Setup for embryo injection using semiautomatic microinjector, glass needle and injection platform. Note the red-colored solution in the needle tip pointing toward the edge where embryos are aligned.
2. Injection platform: Add 6 mL of 2% agarose solution to the cover of a 6-cm Petri dish. To create the ramp, rest the edge of a standard glass microscope slide in the middle of the Petri dish cover (Fig. 4A). When the agarose hardens, store at 4°C for an additional 1 h. Remove slide and wrap ramp in plastic to prevent dehydration of agarose. Store at 4°C. 3. Anesthetize larvae and transfer them to the injection tray filled with E3 supplemented with Tricaine (1:25 dilution of stock solution). 4. Under the Leica dissecting microscope, inject BrdU into the yolk (Fig. 4B). We typically inject a volume that fills a small portion of the yolk. 5. Incubate the microinjected larvae at 28.5°C for 1 h. 6. Fix with 4% PFA at 4°C overnight. 7. Wash three times with PBST using 1.5 mL each wash for 5 min at RT. 8. Dehydrate with ice-cold methanol and store at 20°C. 9. Prior to immunohistochemistry, rehydrate the larvae with PBST by washing three times, 1.5 mL each wash for 2 min at RT. 10. Add 500 µL of proteinase K and incubate at RT for 25 min. 11. Wash three times with PBST using 1.5 mL each wash for 5 min at RT. 12. Add 200 µL of 3 N HCl and incubate at RT for 20 min in order to denature the DNA strands. 13. Wash three times with PBST using 1.5 mL each wash for 5 min at RT. 14. Add 100 µL of diluted mouse anti-BrdU IgG, and incubate at 4°C overnight. 15. Wash three times with PBST using 1.5 mL each wash for 5 min at RT. 16. Add 100 µL of diluted FITC goat anti-mouse IgG, and incubate at 4°C overnight. 17. Wash three times with PBST using 1.5 mL each wash for 5 min at RT.
292
Yee and Pack
18. Examine whole-mount specimens under fluorescent stereomicroscope. Alternatively, the specimens can be analyzed by histological sectioning (see Subheading 3.3.) or confocal microscopy for resolution at the cellular level.
3.5.2. Morpholino Injection 1. Prepare morpholinos solutions for injection by mixing eight parts of Danieau buffer, one part of 1 mM morpholinos stock solution, and one part of diluted phenol red solution (see Note 14). 2. Set up crosses the evening before morpholinos injection as outlined in Subheading 3.1. 3. Inject one-cell stage embryos aligned on the injection ramp discussed in Subheading 3.5.1. Inject 1 µL of diluted morpholinos solutions supplemented with phenol red intracellularly, or alternatively, insert the injection needle through the yolk and inject the morpholinos solution at the base of the cell (see Notes 15 and 16). 4. Place the injected embryos in E3 medium and incubate at 28.5°C until further analysis. 5. For all experiments the following control sibling embryos are analyzed in conjunction with morpholinos-injected embryos: uninjected embryos, embryos injected with Danieau buffer, embryos injected with morpholinos against scrambled RNA sequence. 6. Coinjection of sense RNA and morpholinos to confirm specificity of morpholinos effects.
4. Notes 1. According to the manufacturer, JB-4 Plus solution A is carcinogenic and teratogenic and should be handled with gloves and adequate ventilation. Activated JB4 Plus solution A may be discolored over time and it should be discarded properly after a month. 2. In general, each 1.6-mL tube should contain a maximum of 20 larvae. 3. Agitate larvae on a rocking platform during all washes and incubations. Care should be taken to ensure the embryos/larvae are kept in solutions at all time. 4. Proteinase K can be used to permeabilize embryos and larvae as an alternative to collagenase. Typically, 125 ng/µL of proteinase K (25 min at RT) is used for 5 dpf larvae. Store proteinase K solution at 20°C in 50% glycerol. 5. The skin of 5 dpf can be peeled as an alternative to collagenase or proteinase K permeabilization. Incubation of peeled embryos with a low concentration of proteinase K (10 ng/µL; 20 min at RT) may be required for antigen detection. 6. Immunohistochemical detection of the pancreas can also be performed using alkaline phosphatase- or horseradish peroxidase-conjugated secondary antibodies. 7. Specimens in molding wells are observed using a stereomicroscope as the block polymerizes. This facilitates reorientation of the specimens. Keep the heads of each specimen against the wall of the molding well; this ensures that the sections will be taken at or near the same position from each larva.
Zebrafish as a Research Model
293
8. To avoid curling of plastic sections, hold leading edge of section with watchmaker’s forceps as the block is advanced onto the glass knife. 9. Keep forcep tips apposed while advancing sections onto glass slides so that they do not adhere to the forceps. 10. Take the general precautions with RNA work including wearing gloves and using RNase-free water for all solutions. 11. The concentration of proteinase K used for permeabilization must be titrated. For permeabilizaton of 5 dpf larvae use approx 125 ng/µL of proteinase K. Proteinase K is stored at −20°C in 50% glycerol (20 mg/mL) to avoid loss of enzymatic activity that can occur with repeated freeze–thaw cycles. 12. As an alternative to precipitation, unincorporated nucleotides can be removed using a Microspin™ S-400 HR column. 13. Prehybridization, hybridization, and post-hybridization washes of larvae can be performed in a heat block or incubator. The embryos/larvae become fragile following hybridization at 68°C, so they need to be handled gently with care to avoid physical damage. 14. The working morpholinos stock solution can be stored at 4°C and heated for 5 min at 65°C prior to use. Dilution should be done prior to each use. 15. Although microinjection of morpholinos into one-cell stage embryos is desirable, use of two-cell stage embryos is acceptable. 16. Establish a dose–response relationship to determine the optimum dose for a given morpholinos. The volume of solution injected may be determined using a calibrator. Injection of 1 ng of morpholinos per embryo may be used as a general guideline.
Acknowledgments The authors’ research has been supported in part by grants from the National Institutes of Health to N. S. Y (K08 DK 060529, T32 CA 09615, and T32 DK 07066) and M. P. (DK 61142). References 1. 1 Kern, S., Hruban, R., Hollingsworth, M. A., et al. (2001) A white paper: The product of a pancreas cancer think tank. Cancer Res. 61, 4923–4932. 2. 2 Bardeesy, N., Sharpless, N. E., DePinho, R. A., and Merlino, G. (2001) The genetics of pancreatic adenocarcinoma: A roadmap for a mouse model. Semin. Cancer Biol. 11, 201–218. 3. Zhou, W., Sokoll, L. J., Bruzek, D. J., et al. (1998) Identifying markers for pancreatic cancer by gene expression analysis. Cancer Epidemiol. Biomark. Prev. 7, 109–112. 4. 4 Argani, P., Rosty, C., Reiter, R. E., et al. (2001) Discovery of new markers of cancer through serial analysis of gene expression: Prostate stem cell antigen is overexpressed in pancreatic adenocarcinoma. Cancer Res. 61, 4320–4324.
294
Yee and Pack
5. 5 Ryu, B., Jones, J., Blades, N. J., et al. (2002) Relationships and differentially expressed genes among pancreatic cancers examined by large-scale serial analysis of gene expression. Cancer Res. 62, 819–826. 6 Iacobuzio-Donahue, C. A., Maitra, A., Shen-Ong, G. L., et al. (2002) Discovery 6. of novel tumor markers of pancreatic cancer using global gene expression technology. Am. J. Pathol. 160, 1239–1249. 7. 7 Sakorafas, G. H., Tsiotou, A. G., and Tsiotos, G. G. (2000) Molecular biology of pancreatic cancer; oncogenes, tumor suppressor genes, growth factors, and their receptors from a clinical perspective. Cancer Treat. Rev. 26, 29–52. 8. 8 Reddy, J. K. and Rao, M. S. (1995) Progress in pancreatic cancer: Implications of phenotypic and molecular plasticity. Lab. Invest. 72, 383–385. 9. 9 Edlund, H. (2001) Developmental biology of the pancreas. Diabetes 50, S5–S9. 10 Kim, S. K. and Hebrok, M. (2001) Intercellular signals regulating pancreas devel10. opment and function. Genes Dev. 15, 111–127. 11. 11 Slack, J. M. W. (1995) Developmental biology of the pancreas. Development 121, 1569–1580. 12. 12 Deutsch, G., Jung, J., Zheng, M., Lora, J., and Zaret, K. S. (2001) A bipotential precursor population for pancreas and liver within the embryonic endoderm. Development 128, 871–881. 13. 13 Hebrok, M., Kim, S. K., and Melton, D. A. (1998) Notochord repression of endodermal Sonic hedgehog permits pancreas development. Genes Dev. 12, 1705–1713. 14 Kim, S. K. and Melton, D. A. (1998) Pancreas development is promoted by cyclo14. pamine, a Hedgehog signaling inhibitor. Proc. Natl. Acad. Sci. USA 95, 13036–13041. 15. 15 Li, H., Arber, S., Jessell, T. M., and Edlund, H. (1999) Selective agenesis of the dorsal pancreas in mice lacking homeobox gene Hlxb9. Nat. Genet. 23, 67–70. 16. 16 Harrison, K. A., Thaler, J., Pfaff, S. L., Gu, H., and Kehrl, J. H. (1999) Pancreas dorsal lobe agenesis and abnormal islets of Langerhans in Hlxb9-deficient mice. Nat. Genet. 23, 71–75. 17. 17 Esni, F., Johansson, B. R., Radice, G. L., and Semb, H. (2001) Dorsal pancreas agenesis in N-cadherin-deficient mice. Dev. Biol. 238, 202–212. 18. 18 Jonsson, J., Carlsson, L., Edlund, T., and Edlund, H. (1994) Insulin-promoterfactor 1 is required for pancreas development in mice. Nature 371, 606–609. 19 Offield, M. F., Jetton, T. L., Labosky, P. A., et al. (1996) PDX-1 is required for 19. pancreatic outgrowth and differentiation of the rostral duodenum. Development 122, 983–995. 20. 20 Stoffers, D. A., Zinkin, N. T., Stanojevic, V., Clarke, W. L., and Habener, J. F. (1997) Pancreatic agenesis attributable to a single nucleotide deletion in the human IPF1 gene coding sequence. Nat. Genet. 15, 106–110. 21. 21 Grapin-Botton, A., Majithia, A. R., and Melton, D. A. (2001) Key events of pancreas formation are triggered in gut endoderm by ectopic expression of pancreatic regulatory genes. Genes Dev. 15, 444–454. 22 Swift, G. H., Liu, Y., Rose, S. D., et al. (1998) An endocrine-exocrine switch in 22. the activity of the pancreatic homeodomain protein PDX1 through formation of a trimeric complex with PBX1b and MRG1 (MEIS2). Mol. Cell. Biol. 18, 5109–5120.
Zebrafish as a Research Model
295
23. 23 Dutta, S., Gannon, M., Peers, B., Wright, C., Bonner-Weir, S., and Montminy, M. (2001) PDX:PBX complexes are required for normal proliferation of pancreatic cells during development. Proc. Natl. Acad. Sci. USA 98, 1065–1070. 24. 24 Kim, S. K., Selleri, L., Lee, J. S., et al. (2002) Pbx1 inactivation disrupts pancreas development and in Ipf1-deficient mice promotes diabetes mellitus. Nat. Genet. 30, 430–435. 25. 25 Apelqvist, A., Li, H., Sommer, L., et al. (1999) Notch signalling controls pancreatic cell differentiation. Nature 400, 877–881. 26. 26 Jensen, J., Pedersen, E. E., Galante, P., et al. (2000) Control of endodermal endocrine development by Hes-1. Nat. Genet. 24, 36–44. 27. 27 Jensen, J., Heller, R. S., Funder-Nielsen, T., et al. (2000) Independent development of pancreatic α- and β-cells from neurogenin3-expressing precursors. Diabetes 49, 163–176. 28. 28 Gradwohl, G., Dierich, A., LeMeur, M., and Guillemot, F. (2000) Neurogenin3 is requuired for the development of the four endocrine cell lineages of the pancreas. Proc. Natl. Acad. Sci. USA 97, 1607–1611. 29. 29 Schwitzgebel, V. M., Scheel, D. W., Conners, J. R., et al. (2000) Expression of neurogenin3 reveals an islet cell precursor population in the pancreas. Development 127, 3533–3542. 30. 30 Gu, G., Dubauskaite, J., and Melton, D. A. (2002) Direct evidence for the pancreatic lineage: NGN3+ cells are islet progenitors and are distinct from duct progenitors. Development 129, 2447–2457. 31. 31 Ahlgren, U., Pfaff, S. L., Jessell, T. M., Edlund, T., and Edlund, H. (1997) Independent requirement for ISL1 in formation of pancreatic mesenchyme and islet cells. Nature 385, 257–260. 32. 32 Naya, F. J., Huang H. P., Qiu, Y., et al. (1997) Diabetes, defective pancreatic morphogenesis, and abnormal enteroendocrine differentiation in BETA2/neuroDdeficient mice. Genes Dev. 11, 2323–2334. 33. 33 Sosa-Pineda, B., Chowdhury, K., Torres, M., Oliver, G., and Gruss, P. (1997) The Pax4 gene is essential for differentiation of insulin-producing beta cells in the mammalian pancreas. Nature 386, 399–402. 34. 34 St-Onge, L., Sosa-Pineda, B., Chowdhury, K., Mansouri, A., and Gruss, P. (1997) Pax6 is required for differentiation of glucagons-producing alpha-cells in mouse pancreas. Nature 387, 406–409. 35. 35 Sussel, L., Kalamaras, J., Hartigan-O’Connor, D. J., et al. (1998) Mice lacking the homeodomain transcription factor Nkx2.2 have diabetes due to arrested differentiation of pancreatic beta cells. Development 125, 2213–2221. 36. 36 Sander, M., Sussel, L., Conners, J., et al. (2000) Homeobox gene Nkx6.1 lies downstream of Nkx2.2 in the major pathway of β-cell formation in the pancreas. Development 127, 5533–5540. 37. 37 Boj, S. F., Parrizas, M., Maestro, M. A., and Ferrer, J. (2001) A transcription factor circuit in differentiated pancreatic cells. Proc. Natl. Acad. Sci. USA 98, 14481–14486. 38. 38 Sanvito, F., Herrera, P.-L., Huarte, J., et al. (1994) TGF-β1 influences the relative development of the exocrine and endocrine pancreas in vitro. Development 120, 3451–3462.
296
Yee and Pack
39. 39 Kim, S. K., Hebrok, M., Li, E., et al. (2000) Activin receptor patterning of foregut organogenesis. Genes Dev. 14, 1866–1871. 40. 40 Hebrok, M., Kim, S. K., St-Jacques, B., McMahon, A. P., and Melton, D. A. (2000) Regulation of pancreas development by hedgehog signaling. Development 127, 4905–4913. 41. 41 Krapp, A., Knofler, M., Lederman, B., et al. (1998) The bHLH protein PTF1-p48 is essential for the formation of the exocrine and the correct spatial organization of the endocrine pancreas. Genes Dev. 12, 3752–3763. 42. Ritvos, O., Tuuri, T., Eramaa, M., et al. (1995) Activin disrupts epithelial branch42 ing morphogenesis in developing glandular organs of the mouse. Mech. Dev. 50, 229–245. 43. Miralles, F., Czernichow, P., and Scharfman, R. (1998) Follistatin regulates the 43 relative proportions of endocrine versus exocrine tissue during pancreatic development. Development 125, 1017–1024. 44. Miralles, F., Czernichow, P., Ozaki, K., Itoh, N., and Scharfmann, R. (1999) Sig44 naling through fibroblast growth factor receptor 2b plays a key role in the development of the exocrine pancreas. Proc. Natl. Acad. Sci. USA 96, 6267–6272. 45. 45 Bhushan, A., Itoh, N., Kato, S., Thiery, J. P., Bellusci, S., and Scharfmann, R. (2001) Fgf10 is essential for maintaining the proliferative capacity of epithelial progenitor cells during early pancreatic organogenesis. Development 128, 5109–5117. 46. 46 Crisera, C. A., Kadison, A. S., Breslow, G. D., Maldonado, T. S., Longaker, M. T., and Gittes, G. K. (2000) Expression and role of laminin-1 in mouse pancreatic organogenesis. Diabetes 49, 936–944. 47. Miettinen, P. J., Huotari, M. A., Koivisto, T., et al. (2000) Impaired migration and 47 delayed differentiation of pancreatic islet cells in mice lacking EGF-receptors. Development 127, 2617–2627. 48. 48 Lewis, J. (1996) Neurogenic genes and vertebrate neurogenesis. Curr. Opin. Neurobiol. 6, 3–10. 49. 49 Streisinger, G., Walker, C., Dower, N., Knauber, D., and Singer, F. (1981) Production of clones of homozygous diploid zebrafish (Brachydanio rerio). Nature 291, 293–296. 50. 50 Kimmel, C. B. (1989) Genetics and early development of zebrafish. Trends Genet. 5, 283–288. 51. Haffter, P., Granato, M., Brand, M., et al. (1996) The identification of genes with 51 unique and essential functions in the development of the zebrafish, Danio rerio. Development 123, 1–36. 52. Driever, W., Solnica-Krezel, L., Schier, A. F., et al. (1996) A genetic screen for 52 mutations affecting embryogenesis in zebrafish. Development 123, 37–46. 53. 53 Donovan, A., Brownlie, A., Zhou, Y., et al. (2000) Positional cloning of zebrafish ferroportin1 identifies a conserved vertebrate iron exporter. Nature 403, 776–781. 54. Zhong, T. P., Rosenberg, M., Mohideen, M. A. P. K., Weinstein, B., and Fishman, 54 M. C. (2000) Gridlock, an HLH gene required for assembly of the aorta in zebrafish. Science 287, 1820–1824.
Zebrafish as a Research Model
297
55. 55 Kupperman, E., An, S., Osborne, N., Waldron, S., and Stainier, D. Y. (2000) A sphingosine-1-phosphate receptor regulates cell migration during vertebrate heart development. Nature 406, 192–195. 56. Pack, M., Solnica-Krezel, L., Malicki, J., et al. (1996) Mutations affecting devel56 opment of zebrafish digestive organs. Development 123, 321–328. 57. 57 Biemar, F., Argenton, F., Schmidtke, R., Epperlein, S., Peers, B., and Driever, W. (2001) Pancreas development in zebrafish: Early dispersed appearance of endocrine hormone expressing cells and their convergence to form the definitive islet. Dev. Biol. 230, 189–203. 58. Argenton, F., Zecchin, E., and Bortolussi, M. (1999) Early appearance of pancre58 atic hormone-expressing cells in the zebrafish embryo. Mech. Dev. 87, 217–221. 59. 59 Yee, N. S., Yusuff, S., and Pack, M. (2001) Zebrafish pdx1 morphant displays defects in pancreas development and digestive organ chirality, and potentially identifies a multipotent pancreas progenitor cell. Genesis 30, 137–140. 60. Farber, S. A., Pack, M., Ho, S. Y., et al. (2001) Genetic analysis of digestive 60 physiology using fluorescent phospholipid reporters. Science 292, 1385–1388. 61. 61 Roy, S., Qiao, T., Wolff, C., and Ingham, P. W. (2001) Hedgehog signaling pathway is essential for pancreas specification in the zebrafish embryo. Curr. Biol. 11, 1358–1363. 62. diIorio, P. J., Moss, J. B., Sbrogna, J. L., Karlstrom, R. O., and Moss, L. G. (2002) 62 Sonic hedgehog is required early in pancreatic islet development. Dev. Biol. 244, 75–84. 63. Thomas, M. K., Lee, J. H., Rastalsky, N., and Habener, J. F. (2001) Hedgehog 63 signaling regulation of homeodomain protein Islet Duodenum Homeobox-1 expression in pancreatic β-cells. Endocrinology 142, 1033–1040. 64. 64 Huang, H., Liu, N., and Lin, S. (2001) Pdx-1 knockdown reduces insulin promoter activity in zebrafish. Genesis 30, 134–136. 65. Milewski, W. M., Duguay, S. J., Chan, S. J., and Steiner, D. F. (1998) Conserva65 tion of PDX-1 structure, function, and expression in zebrafish. Endocrinology 139, 1440–1449. 66. Inoue, A., Takahashi, M., Hatta, K., Hotta, Y., and Okamoto, H. (1994) Developmental regulation of islet-1 mRNA expression during neuronal differentiation in embryonic zebrafish. Dev. Dynam. 199, 1–11. 67. Barth, K. A. and Wilson, S. W. (1995) Expression of zebrafish nk2.2 is influenced 67 by sonic hedgehog/vertebrate hedgehog-1 and demarcates a zone of neuronal differentiation in the embryonic forebrain. Development 121, 1755–1768. 68. Nornes, S., Clarkson, M., Mikkola, I., et al. (1998) Zebrafish contains two Pax6 68 genes involved in eye development. Mech. Dev. 77, 185–196. 69. 69 Glasgow, E. and Tomarev, S. I. (1998) Restricted expression of the homeobox gene prox 1 in developing zebrafish. Mech. Dev. 76, 175–178. 70. 70 Korzh, V., Sleptsova, I., Liao, J., He, J., and Gong, Z. (1998) Expression of zebrafish bHLH genes ngn1 and nrd defines distinct stages of neural differentiation. Dev. Dynam. 213, 92–104.
298
Yee and Pack
71. 71 Strahle, U., Blader, P., Henrique, D., and Ingham, P. W. (1993) Axial, a zebrafish gene expressed along the developing body axis, shows altered expression in Cyclops mutant embryos. Genes Dev. 7, 1436–1446. 72. 72 Odenthal, J. and Nusslein-Volhard, C. (1998) fork head domain genes in zebrafish. Dev. Genes Evol. 208, 245–258. 73. 73 An, M., Luo, R., and Henion, P. D. (2002) Differentiation and maturation of zebrafish dorsal root and sympathetic ganglion neurons. J. Comp. Neurol. 446, 267–275. 74. Nasevicius, A. and Ekker, S. C. (2000) Effective targeted gene ‘knockdown’ in 74 zebrafish. Nat. Genet. 26, 216–220. 75. 75 Yee, N. S., Furth, E. E., and Pack, M. (2003) Clinicopathologic and molecular features of pancreatic adenocarcinoma associated with Peutz-Jeghers syndrome. Cancer Biol. Ther. 2, 38–47. 76. Westerfield, M. (2000) The Zebrafish Book. Guide for the Labotratory Use of Zebrafish (Danio rerio). 4th ed. Eugene, OR: University of Oregon Press. 77. Jowett, T. (1999) Analysis of protein and gene expression, in Methods in Cell Biology (Detrich, H. M., Westerfield, M., and Zon, L. I., ed.), San Diego: Academic Press, pp. 63–85.
Whole Cell Pancreatic Cancer Vaccine
299
19 Development of a Cytokine-Modified Allogeneic Whole Cell Pancreatic Cancer Vaccine Dan Laheru, Barbara Biedrzycki, Amy M. Thomas, and Elizabeth M. Jaffee Summary The management of patients with pancreatic cancer is a multidisciplinary approach that presents enormous challenges to the clinician. Overall 5-yr survival for all patients remains < 3%. Symptoms of early pancreas cancer are nonspecific. As such, only a fraction of patients are candidates for surgery. While surgical resection provides the only curative option, most patients will develop tumor recurrence and die of their disease. To date, the clinical benefits of chemotherapy and radiation therapy have been important but have led to modest improvements. Tumor vaccines have the potential to specifically target the needle of pancreas cancer cells amidst the haystack of normal tissue. The discovery of pancreas tumor-specific antigens and the subsequent ability to harness this technology has become an area of intense interest for tumor immunologists and clinicians alike. Without knowledge of specific antigen targets, the whole tumor cell represents the best source of immunizing antigens. This chapter will focus on the development of whole tumor cell vaccine strategies for pancreas cancer. Key Words: Pancreas cancer; immunotherapy; tumor antigens; genetically modified whole cell allogeneic vaccine.
1. Introduction Surgical resection provides the only curative option for patients with pancreatic cancer. Unfortunately, only 10–20% of patients are candidates for definitive surgical resection. Even in this select population, the 5-yr survival is only 15–20% for resectable disease with a median survival of 15–19 mo (1–6). To date, the clinical benefits of chemotherapy and/or radiation therapy have been relatively modest (7,8). More effective therapies for all stages of pancreatic adenocarcinoma are needed.
From: Methods in Molecular Medicine, Vol. 103: Pancreatic Cancer: Methods and Protocols Edited by: G. Su © Humana Press Inc., Totowa, NJ
299
300
Laheru et al.
Immunotherapy has the potential to provide a non-cross-resistant mechanism of antitumor activity that can be integrated with surgery, radiation, and chemotherapy. A major advantage of immune-based therapies is their ability to target a tumor cell specifically relative to the normal cell of origin, thereby minimizing nonspecific toxicities. Antitumor immune responses can be diverse and consist of both antibody and cellular responses. In the case of solid tumors and in particular, pancreatic cancer, specific targets of the immune response that serve as tumor rejection antigens have not been defined. However, a few candidate antigens have been identified against which cellular and antibody responses can be induced. Because the majority of these antigens are intracellular, vaccine strategies should be able to induce predominantly T-cell rather than antibody responses. Without knowledge of specific antigen targets, the whole tumor cell represents the best source of immunizing antigens. Therefore, this chapter focuses on the development of whole tumor cell vaccine strategies. However, with the recent sequencing of the human genome (9,10) and the development of rapid methods for identifying genes that are differentially expressed by tumor cells, many more candidate immune targets are expected to be uncovered that may serve as immunogens for the development of vaccine strategies that aim at both the treatment an prevention of pancreatic cancer (11–18). This chapter (1) reviews the important features of an effective antitumor immune response; (2) discusses the results of some of the more promising strategies that are currently under clinical consideration with a focus on a cytokine-modified allogeneic whole cell vaccine approach; and (3) provides a detailed description of the materials and methods employed for the development of a granulocyte-macrophage colonystimulating factor (GM-CSF) modified allogeneic whole cell pancreatic cancer vaccine. It is important to point out that these methods can be modified to develop whole cell vaccines that express other cytokines and costimulatory molecules that are known to be important for activating tumor-specific immune responses. 1.1. Features of the Immune System that Are Advantageous for Cancer Immunotherapy The immune system comprises a number of cell types that, when activated, are extremely efficient at recognizing and killing their target. In particular, Bcells and T-cells each possess vast arrays of clonally distributed antigen receptors that enable them to recognize foreign antigens and to discriminate self from nonself. It has been estimated that both T- and B-cells can express more than a million different antigen-specific receptors through recombination of the genes encoding for their receptor at the time of maturation in the thymus and bone marrow, respectively. Therefore, the immune system should have an unlimited capability to recognize antigenic differences between normal and
Whole Cell Pancreatic Cancer Vaccine
301
malignant cells, whether they are in the form of the product of a new genetic alteration, a reactivated embryonic gene, or an overexpressed gene. The B-cell receptor, which is a surface immunoglobulin with the ability to bind antigens on soluble molecules, recognizes free antigenic determinants (Fig. 1). Therefore, special antigen processing is not required for B-cell receptor or soluble antibody binding to its antigen. In contrast, the T-cell receptor recognizes fragments of the antigenic protein bound to HLA class I and II molecules on another cell. This peptide–HLA complex is formed as a result of fragmentation of proteins within specialized cellular compartments and subsequent association with a binding site on the HLA molecule (Fig. 1). Two forms of T-cell antigen processing exist (19–24). Professional antigen presenting cells (macrophages, B-cells, and dendritic cells) have the ability to capture extracellular proteins that are released by the tumor through secretion, shedding, or tumor lysis. These proteins are subsequently internalized via endocytosis and processed through the exogenous pathway. These proteins are taken up into low-pH vesicles (the lysosomal compartment), where they undergo fragmentation. Peptide fragments (10–25 amino acids in length) then bind to the HLA class II protein, prior to expression of the complex on the cell surface. This complex is recognized exclusively by CD4+ helper T-cells in the context of a second costimulatory molecule such as B7 (25,26). In the presence of both of these signals, activated CD4+ T-cells can amplify the CD8+ T-cell response. In addition, memory CD4+ T-cells are generated and play the key role in the maintenance of protective immunity. Presentation of antigen on HLA class II and the ability to express costimulatory molecules are the specialized function of these “professional” antigen presenting cells that derive from hematopoietic precursors in the bone marrow. In contrast to professional antigen-presenting cells, pancreatic and most solid tumors derive from epithelial cells rather than hematopoietic cells. Therefore, pancreatic cancer cells cannot process and present antigen through the exogenous pathway. However, all cells including tumor cells have the ability to process and present antigens that derive from cellular proteins through the endogenous pathway (Fig. 1) (27,28). Any protein within a tumor cell can gain access to the cytosol and undergo enzymatic degradation into 8–10 amino acid fragments by specialized machinery (the proteasome). The peptide fragments are subsequently transported into the endoplasmic reticulum via the transporter associated with antigen processing (TAP) where they bind to HLA class I molecules and are transported to the cell surface for recognition by CD8+ T cells. CD8+ T cells exclusively recognize antigen in this way. In general, CD4+ T-cells provide helper or regulatory function while CD8+ T-cells carry out direct tumor lysis. A few candidate pancreatic antigens recognized by B- and T-cells have already been identified and are listed in Table 1.
302
Laheru et al.
Fig. 1. Tumor antigen processing and presentation. This model demonstrates that there are many types of proteins or glycoproteins expressed by tumor cells that can be recognized by T-cells of the immune system. The tertiary structure of the surface mucins or surface glycoproteins that are the product of reactivated embryonic genes, or the tertiary structure of overexpressed growth factor receptors such as HER-2/neu, can be recognized by soluble antibodies secreted by B-cells. However, these same proteins or glycoproteins can be internalized into endocytic vacuoles, and then trafficked either into the lysosomal antigen processing pathway to generate MHC class II bound tumor antigens via the exogenous pathway, or into the cytoplasm where they will be degraded into MHC class I bound tumor antigens via the endogenous pathway. Tumor cells do not usually express MHC class II. Therefore, most tumor antigens will not be presented for MHC class II presentation to CD4+ helper T-cells. However, the surface tumor antigens that will enter the cytoplasm will be degraded into small peptide fragments by the proteasomes (A,B). The small peptide fragments will then be transported into the endoplasmic reticulum (ER) where they will bind to MHC class I (C) and subsequently be presented on the tumor’s surface bound to MHC class I molecules, and recognized
Whole Cell Pancreatic Cancer Vaccine
303
Table 1 Candidate B- and T-Cell Pancreatic Targets Category of antigena
T-Cell antigens
B-Cell antigens
Reactivated embryonic genes Mutated oncogene/suppressor genes Tissue-specific antigens Glycoproteins
CEA k-ras, p53 HER-2/neu MUC-1
CEA p53 HER-2/neu MUC-1
a Categories
are defined based on already defined categories of human melanoma antigens (78).
1.2. Immunotherapy in Practice The goal of immune-based therapies is to either recruit and activate tumorspecific T-cells that have the ability to directly lyse a tumor cell or employ monoclonal antibodies that can target tumor-specific antigens and either directly lyse the tumor or lyse the tumor via delivery of a cytotoxic agent. Both approaches are attractive for several reasons (1). The activation of tumor-specific T-cells or the use of monoclonal antibodies act via a mechanism that is distinct from chemotherapy or radiation therapy and would represent a non-cross-resistant treatment with an entirely different spectrum of toxicities (2). The immune system is capable of recognizing a broad diversity of potential antigens while orchestrating selective as well as specific cytotoxic responses. This feature may be essential in recognizing and eliminating a heterogeneous tumor population while avoiding normal tissue toxicity (3). Preclinical animal models using both forms of immunotherapy have been able to eliminate small burdens of established tumors, a situation that corresponds to the state of minimal residual disease commonly found after resection of human tumors (29,30). Immunotherapy can be broadly divided into passive and active therapeutic approaches (Table 2). Passive immunotherapy mainly involves the use of unlabeled or labeled monoclonal antibodies that are specifically raised against tumor antigens. Antibodies have so far been the most successful form of immunotherapy clinically. They are being employed as diagnostic tools, as prognostic indicators, and for the treatment of cancer. Advantages include specific targeting of tumor cells while sparing normal tissue, relative ease of administration, and low toxicity profile. The major disadvantages include the absence of T-cell Fig. 1. (Continued) by the cytotoxic T cell (D) Any protein within a tumor cell, whether it is presented on the cell surface, expressed in the nucleus, cytoplasm, or endoplasmic reticulum, can gain access to the cytoplasm for processing and presentation on MHC class I molecules, and be recognized by cytotoxic T cells. b2M, b2-Microglobulin; TAP, transporter associated with antigen processing.
304
Laheru et al. Table 2 Different Categories of Immunotherapy Passive immunotherapy Unlabeled monoclonal antibodies Radioimmunotherapy Antibody-directed immunotoxins T-cell adoptive transfer Active nonspecific immunotherapy Whole tumor cells/tumor lysate mixed with bacterial adjuvant Systemic cytokines Active specific immunotherapy (vaccines) Genetically modified whole tumor cells Protein/peptide/carbohydrate based antigen vaccines Dendritic cell-based antigen vaccines DNA-based vaccines Recombinant viral-based antigen vaccines
activation which therefore precludes T-cell-mediated cytotoxic killing and the generation of memory immune responses. In addition, a potential limiting factor in its use involves tumor heterogeneity. Specifically, all tumor cells within a proliferating mass may not express the antigen being targeted by the antibody. A number of monoclonal antibodies have been used for solid tumors including advanced pancreatic cancer (31–34). However, these results provide additional rationale for developing vaccine approaches that can induce natural tumor specific antibody responses in the patient. Active immunotherapy is typically divided into nonspecific and specific processes. Nonspecific therapy attempts to augment an immune response without actually targeting a specific tumor antigen. Examples include the use of BCG, interleukin-2 (IL-2), and levamisole. In contrast, active specific or vaccine therapy, targets specific tumor antigens as a result of the induction of antigen-specific B-cell- or T-cell-mediated immune responses. Active specific therapy can also generate antigen specific memory T-cell responses that are capable of being reactivated if tumor cells expressing the same antigen profile recur. Furthermore, the induction of cellular immune responses has the added benefit of allowing natural access to the microenvironment of the tumor. Preclinical studies have already shown that T-cell-mediated vaccine therapy can induce antitumor immune responses that are potent enough to eradicate colorectal tumors (35, 36). Translation of these vaccine approaches into therapies for patients with pancreatic and colorectal tumors are in early phases of clinical development. Exam-
Whole Cell Pancreatic Cancer Vaccine
305
ples of the different vaccine approaches that are currently undergoing clinical testing are discussed below. 1.2.1. Peptide- and Protein-Based Vaccines Although significant progress has been made in understanding the biology of pancreatic cancer at the genetic level, specific pancreatic tumor antigens that can serve as rejection targets have not yet been identified. However, point mutations in a variety of oncogenes (K-ras) or tumor suppressor genes (p53, p16, DPC4, BRCA2, Her-2/neu) have been associated with different histologically defined precursor lesions (15–18,37–40) and are being studied as candidate immune targets. Mutated K-ras is a particularly attractive immune target because it is mutated in more than 90% of pancreatic adenocarcinomas (41– 44). The ras p21 proto-oncogenes including K-ras, N-ras, and H-ras encode proteins that are important for regulating cellular events including growth and differentiation. Point mutations at codons 12, 13, and 61 have been identified in many cancers including pancreatic adenocarcinoma (43,44). These mutations encode distinct proteins that are potential immunogens. The major advantage of a protein- or peptide-based vaccine is the ability to deliver high doses of the potential immunogen safely and at a relatively modest cost. However, there are also several limitations to vaccine approaches that employ peptides and proteins. First, the vaccine approaches that will be most successful at optimally priming with the peptide and/or protein have not yet been determined. Second, proteins that are identified as a candidate immunogen based on the criteria that they are overexpressed in pancreatic adenocarcinoma may turn out not to be the most relevant target of the immune response. Mutated K-ras–based vaccines have been the most extensively studied peptide- or protein-based vaccine approach in patients with pancreatic adenocarcinoma. In preclinical models, vaccination with mutant K-ras peptides induces both major histocompatibility complex (MHC) class I and II restricted T-cell responses. K-ras–specific T-cell responses have also been elicited in patients treated with vaccines that consist of k-ras peptides that contain a point mutation at codon 12 (45–49). Heat shock proteins (HSP) are ubiquitous and highly conserved cellular proteins that are up-regulated during cell stress. They are thought to bind to cellular proteins that become damaged when a cell experiences stress, thereby facilitating the protein’s refolding to an active conformation. In the nonstressed environment, HSPs are thought to have multiple functions including helping newly synthesized polypeptides fold, assisting in protein transport, and associating with peptides generated during protein degradation. They are also thought to stimulate macrophage and dendritic cell activation and assist in representation of peptides. HSPs as vaccine have already been used in clinical trials (50,51).
306
Laheru et al.
1.2.2. Glycoproteins as Antigens Other antigen-targeted immunologic approaches have been tested in patients with pancreatic adenocarcinoma. Mucin-1 (MUC-1) is a glycosylated transmembrane protein that is uniquely characterized by an extracellular domain that consists of a variable number of tandem repeats of 20 amino acids rich in proline, serine, and threonine residues (52). Although normally present lining ductal epithelial surfaces including the gastrointestinal tract, MUC-1 is overexpressed on the cell surface of many cancers including pancreatic adenocarcinoma (52,53). There is also evidence to suggest that MUC-1 protein expression is associated with an increased risk for metastasis and poor prognosis (54). As a glycoprotein with altered expression in pancreatic cancers, MUC-1 is considered another attractive candidate target for immunologic manipulation (52,53,55,56). There is also evidence to suggest that the alterations in the glycosylation of mucin may initiate the events that generate the antitumor immune response. Presumably in normal mucin secreting cells, the bulky carbohydrate side chains block the presentation of potential peptide to T-cells. As a consequence of under glycosylation of mucin thought secondary to decreased glycosyl transferase activity in malignant cells, new peptide epitopes are exposed. Data from animal and phase I clinical studies have demonstrated that HLA-unrestricted T-cells isolated from patients with MUC-1 expressing tumors can recognize these exposed epitopes and can induce MUC-1-specific delayed-type hypersensitivity (DTH) responses (55). More recently, immunization with MUC-1 peptide has been shown to induce an HLA-A2-restricted T-cell response (56). Carcinoembryonic antigen (CEA) is another glycoprotein that is overexpressed in pancreatic cancers. Although it is known to be a member of the immunoglobulin supergene family, the exact function of CEA is unclear (57). A CEA vaccine approach has also already been tested in a clinical trial (58). Additional potential glycoprotein targets have recently been identified by gene expression analysis (15–18). These candidate targets will likely be tested as immunogens in clinical trials in the near future. 1.2.3. Whole Tumor Cell Vaccines Currently, one major limitation of defined antigen-based vaccines is the lack of identified pancreatic tumor antigens that are the known targets of the immune response. Until a panel of pancreatic tumor-specific antigens is discovered, the whole tumor cell represents the best source of immunogens. In addition, at this time it is not clear which if any of the currently studied antigen-based vaccine approaches are most effective at delivering these immunogens for activation of antitumor immunity. A whole tumor cell vaccine approach involves the use of autologous or allogeneic tumor cells to stimulate an immune response. The whole tumor cell can
Whole Cell Pancreatic Cancer Vaccine
307
provide multiple antigens for immunization and therefore it is the immune system that selects which antigens are relevant. However, studies aimed at dissecting antitumor immune responses have confirmed that most tumors are not naturally immunogenic (59). Evidence from preclinical models suggests that the failure of the immune system to reject spontaneously arising tumors is unrelated to the absence of sufficiently immunogenic tumor antigens. Instead, the problem is derived from the immune system’s inability to appropriately respond to these antigens (60,61). The importance of the local release of stimulatory cytokines to provide an immunological boost and attract other immune cells has been extensively examined. These findings have led to the concept that a tumor cell can become more immunogenic if engineered to secrete immune activating cytokines. Tumor cells genetically modified to secrete immune activating cytokines have been extensively studied for their ability to induce systemic antitumor immune responses (60). Preclinical studies have shown that these vaccines can induce immune responses potent enough to cure mice of preestablished tumor (62,63). In one comparison study of 10 cytokines, GM-CSF was most potent, generating systemic immunity dependent on both CD4+ and CD8+ T-cells (64). GM-CSF is known to be involved in the recruitment and differentiation of bone marrow derived dendritic cells and dendritic cells are known to be the most efficient “professional” antigen-presenting cells (APCs) at activating Tcells (65,66). In addition, GM-CSF is produced by activated CD4+ T-helper cells further supporting the concept that this cytokine may function by priming immune effector cells (60,66,67). Studies aimed at optimizing this cytokinesecreting tumor vaccine approach confirmed that GM-CSF secretion must be at the site of relevant tumor antigen. Simple injection of soluble GM-CSF along with the appropriate tumor cells does not provide sustained local levels required to provide a sufficient immunologic boost (68). This information suggested that the mere presence of GM-CSF was not sufficient. Rather, the sustained release and duration of GM-CSF secretion appeared to be critical for priming the immune response. Furthermore, high levels must be sustained for several days. In the preclinical data, it appeared that a minimum of 35 ng/106 cells/24 h is necessary to generate effective antitumor immunity (68,69). A GM-CSF secreting tumor vaccine was first tested in patients with renal cell carcinoma. Lethally irradiated autologous renal cell carcinoma cells transduced with the GM-CSF gene were prepared and tested in a phase I trial of patients with metastatic renal cell carcinoma. Although the maximally tolerated dose (MTD) as well as the dose of maximal bioactivity could not be determined secondary to technical difficulty of expanding each patient’s tumor cells beyond 4 ´ 107 cells, a dose of 4 ´ 107 cells resulted in post-vaccination DTH responses against autologous tumor cells that were similar to those measured in
308
Laheru et al.
mouse studies testing this vaccine approach. Other immune parameters including histologic evaluation of the vaccine biopsy and DTH sites revealed similar immune infiltrates when compared with preclinical models. The vaccine was well tolerated at all vaccine doses tested. The most common side effects were local induration and erythema at the vaccine site (70). A similar spectrum of toxicities were subsequently observed in autologous prostate and melanoma studies and in allogeneic prostate vaccine studies (71,72). While the use of autologous tumor cells may preserve unique antigens expressed by each patient’s cancer, the development of an autologous vaccine requires extensive processing, in vitro expansion, and regulatory testing be performed for each individual patient vaccine. In the case of metastatic disease, the development of autologous tumor vaccine would also require the ability to obtain adequate tissue. These limitations preclude the use of autologous cellular vaccines for most cancers including pancreatic adenocarcinoma. Recent data support the immunologic rationale for using allogeneic cells as the source of immunogen. First, studies evaluating human melanoma antigens have demonstrated that most antigens identified so far are shared among at least 50% of other patient melanomas, regardless of whether or not they share the same HLA type (72–74). In addition, there are data to support that the professional APCs that present immunogen to specific T-cells in the context of HLA are host derived. Therefore, the vaccine cells do not need to be HLA compatible with the host’s immune system as long as they can release cellular proteins (the tumor antigens) for uptake by the professional APCs (macrophages and dendritic cells) that are attracted to the vaccine site by GM-CSF. Tumor antigens are taken up by the APCs’ exogenous processing pathway. However, these antigens have been shown to also gain access to the cytosol for processing onto HLA class I through the recently defined cross-priming mechanism (60,61). Taken together, the data suggest that relevant tumor antigens can be delivered by an allogeneic tumor and still sufficiently mount an effective CD4+ and CD8+ T-cell response against a tumor. The results of a phase I study using irradiated allogeneic pancreatic tumor cell lines transfected with GM-CSF (75) as adjuvant treatment administered in sequence with adjuvant chemoradiation in patients with resected adenocarcinoma of the pancreas was recently reported (76). This was the first GM-CSF secreting vaccine study to escalate the vaccine dose to 5 ´ 108 GM-CSF secreting cells. However, toxicities remained mostly limited to grade I/II local reactions at the vaccine site. In addition, there were self-limited systemic rashes, including one documented case of Grover’s syndrome (77). Systemic GM-CSF levels were evaluated as an indirect measure of the longevity of vaccine cells at the immunizing site. This pancreatic vaccine study is the first GM-CSF vaccine clinical trial to measure low but detectable serum GM-CSF levels in patients.
Whole Cell Pancreatic Cancer Vaccine
309
As was observed in preclinical studies (64,69), GM-CSF levels peaked at 48 h following vaccination. In addition, serum GM-CSF levels could be detected for up to 96 h following vaccination. These data, together with data from preclinical models, would suggest that serum GM-CSF levels may serve as a biomarker of immune response. Followup studies are ongoing to determine if these promising effects on immune activation will translate into a true clinical benefit for patients with pancreatic cancer. 1.3. Features of Gene Transfer Systems 1.3.1. Rationale for Choosing a Retroviral Vector for Gene Transfer into Tumor Cells Gene transfer into tumor cells can be accomplished by a variety of methods involving either naked DNA or the use of viral vectors. There are many methods for transferring naked DNA into cells, including (1) coprecipitation with calcium phosphate (79); (2) the use of electroporation, which exposes cells to rapid pulses of high voltage current, thereby providing a physically induced opening in the cell membrane for entry of DNA (80); (3) direct introduction of DNA into cells by microinjection (81); and (4) encapsidation of DNA into liposomes (82). Use of any of these methods will often result in the introduction of multiple copies of the cytokine gene randomly into the host cell’s genome. Several of these methods can result in transient expression of the gene (for 24– 72 h) by as many as 50% of the cells in the transfected population, because the transfected DNA can exist free in the cell nucleus for a short time (83). However, stable gene expression, which requires integration of the transfected DNA into the host’s genome, usually occurs in fewer than 1% of the cells within the population undergoing transfection. To achieve stable integration and expression of the DNA into a high proportion of the cell population, it is often necessary to select for the minority of cells in the transfected population that have successfully retained the foreign DNA. This can be accomplished by cotransfecting DNA that encodes for a selectable marker, which will allow cells expressing its product to survive in growth media that contains a substrate for the gene’s product, which is normally toxic to most mammalian cells. In this way, in vitro selection of those cells that have successfully incorporated the transferred DNA can be accomplished. However, although in vitro selection will enhance the number of gene-transduced tumor cells in the cell population to nearly 100%, it is at the theoretical expense of antigen expression loss among that tumor cell population. In theory, loss of particular antigenic populations of tumor cells will decrease the effectiveness of the vaccine whether it be an allogeneic or autologous vaccine approach.
310
Laheru et al.
Currently, the most efficient method of stable gene delivery into mammalian cells is through the use of viral vectors, which infect their target cell by binding specific cell surface receptors. Most viral vectors are constructed so that they contain the sequences encoding for the expression of the gene and all of the genetic signals including the promoters, enhancers, splicing signals, and signals for polyadenylation of RNA transcripts, all of which are necessary for the transcription and ultimate translation of the inserted cytokine gene sequences. Often, the vectors will also contain selectable markers (84). Potential adverse consequences following viral infection can include: (1) damage or death to the host cell; (2) the activation of other latent viruses integrated into the host’s genome; (3) the activation of silent host genes such as protooncogenes; or (4) transformation of the defective viral vector from replication incompetent to replication competent by recombination with host gene sequences (which results in the production of helper virus). All of these possibilities should be considered in choosing a viral vector system for gene transfer. Retroviruses have been the most commonly employed vectors for the preparation of cytokine-secreting tumor vaccines, for study in murine models, and in human vaccine therapy trials. There are at least two reasons for this. First, as mentioned above, retroviruses usually do not enter into a lytic cycle of viral replication and therefore do not kill their host cell soon after viral infection. Second, retroviruses can infect most mammalian cells and integrate into the host genome, which is a critical requirement for efficient gene transfer and expression in a stable and heritable fashion. Most of the retroviral vectors employed in cytokine-secreting tumor vaccine studies have been developed from either avian or murine retroviruses. A key feature of these retroviral vectors is their incompetence to replicate following transduction into the host cell. Details of the mechanisms of infection, replication, integration, and gene expression of these viral vectors have already been described in significant detail in the literature (85–90). However, it is important to note that host cell replication and DNA synthesis are required for provirus integration, and thus, efficient gene transfer is restricted to replicating cells. This chapter focuses on the retrovirus gene transfer system for several reasons: (1) it is still one of the most efficient gene transfer systems for replicating tumor cells; (2) these viruses have already been used clinically for this purpose and have not been demonstrated to cause clinically significant side effects. However, the lentiviruses are a newer gene transfer system that appears to have equivalent gene transfer efficiency but to a wider range of cells (91,92). In particular, lentiviruses do not require a cell be replicating for successful gene transfer. In addition, lentiviruses infect bone-marrow-derived cells more efficiently than other viral vector systems. However, these viruses have not been tested in the clinics as of yet. There is still ongoing discussion over the increased safety con-
Whole Cell Pancreatic Cancer Vaccine
311
cerns associated with the gene transfer system. For gene transfer to pancreatic tumor cells, retroviruses should be sufficient. 1.3.2. Structure of the Retroviral Vector The transduction procedures that will be described employ the MFG retroviral vector system, which we have extensive experience using and which has been tested in the clinics. However, other similar and equally effective retroviral vectors are also available for this use (93). The structure of MFG has recently been described (90). In brief, in this vector, Moloney murine leukemia virus (Mo-MuLV) long terminal repeat (LTR) sequences are used to generate both a full-length viral RNA (for encapsidation into virus particles) and a subgenomic mRNA (analogous to the Mo-MuLV env mRNA) which is responsible for the expression of inserted sequences. The vector retains both sequences in the viral gag region shown to improve the encapsidation of viral RNA and the normal 5' and 3' splice sites necessary for the generation of the env mRNA. Protein coding sequences are inserted between the NcoI and BamHI sites in such a way that the cDNA sequences encoding the gene of interest are cloned into the downstream site such that the cDNA insert’s AUG at the exact position relative to the 3' splice acceptor site where the env gene starts in the original virus. It is therefore expressed as a subgenomic transcript off of the 5' LTR. No selectable marker exists in the vector. This feature, together with deletion of sequences in the 5' portion of the gag-region intron, results in high-titer production of viral particles by the packaging line as well as uniformly high levels of expression regardless of the particular gene cloned into the vector. 1.3.3. Construction of the Packaging and the Producer Lines The other critical component of the retroviral vector system is a cell line that produces the viral proteins that are required for encapsidation of the viral RNA. This cell line, referred to as the retrovirus Packaging line, is produced by transfection of proviral DNA containing the retroviral genes necessary for the synthesis of the viral proteins, into a fibroblast cell line. The most commonly employed fibroblast cell line is the murine NIH 3T3 cell line. After transfection of the proviral DNA vector into the packaging cell line, this line is then referred to as the retrovirus Producer cell line. The first generation of packaging lines contained the stable introduction of a mutant Moloney murine leukemia virus proviral genome that contained a deletion of the encapsidation sequence (psi sequence) (94–99). Most of the producer cell line clones derived from these packaging cell lines have been shown to produce retroviral titers in the range of 106 colonyforming units (cfu)/mL, levels that allow for high-efficiency gene transfer (97). These titers are comparable to low normal titers obtained for replication-competent retroviruses.
312
Laheru et al.
1.4. Choice of Pancreatic Tumor Cells for Vaccine Production As discussed above, autologous or allogeneic tumor lines can theoretically be used for this purpose. The major benefit of autologous tumor cells has to do with the ability to present antigens that are unique to the patient as well as common among the majority of pancreatic tumors. Unfortunately, it is technically difficult and expensive to isolate tumor cells from every pancreatic cancer patient. In addition, 20% or fewer of patients presenting with pancreatic cancer are surgical candidates. For these two reasons, we have chosen to use allogeneic tumor cells as the source of immunogen. Based on data from human melanoma, it appears that the majority of T-cell-specific tumor antigens identified for that disease are shared by at least 50% of other patients’ tumors. Therefore, we have taken the approach of mixing pancreatic tumor lines that derived from two dif-ferent patients (72–74). There are available from the American Tissue Culture Collection (ATCC, www.atcc.org) a significant number of pancreatic tumor lines that derive from either primary pancreatic tumors or from metastases. These lines have been characterized and could be a good source of tumor lines for this purpose. In choosing a line several parameters should be considered: (1) rate of growth because these cells will have to be expanded to large numbers for a vaccine trial; (2) number of known pancreatic-cancer-specific antigens. This second parameter is important for several reasons. First, expression of at least one pancreatic-cancer-specific antigen will provide a marker for stability testing following gene transfer and cell expansion. Second, these expressed pancreatic antigens can serve as surrogate targets for immune monitoring studies as part of evaluating efficacy. Alternatively, we have described methods for establishing new pancreatic cancer cell lines from primary tumor specimens (100). 2. Materials 2.1. Retroviral Vector Producer Lines The MFG retroviral producer cell lines were originally obtained from R. C. Mulligan (Whitehead Institute for Biomedical Research, Cambridge, MA). Both the amphotropic and ecotropic retroviral producer cell lines, CRIP and CRE, respectively, are grown in Dulbecco’s modified Eagle’s medium (DMEM) with high glucose (4500 g/L), supplemented with 10% bovine calf serum, penicillin (100 U/mL final concentration), streptomycin (100 µg/mL final concentration), L-glutamine (2 mM final concentration), and gentamicin (50 µg/mL final concentration), at 37°C, and 10% CO2. Trypsin (0.25%)–EDTA (0.1%) is used for passaging the cell lines.
Whole Cell Pancreatic Cancer Vaccine
313
2.2. Tumor Cell Lines All tumor cell lines to be transduced should be maintained in their optimal growth media before and after the transduction procedure is performed to enhance the proliferation capacity of the cell population. 2.3. Retroviral Gene Transfer to Tumor Cells Tumor cells and retroviral supernatant that has been prepared as described in Subheading 3. In addition, DEAE-dextran 10 mg/mL stock solution prepared by dissolving 1 g into 100 mL of the producer line growth media (DMEM + 10% calf serum), and filtered through a 0.45-µm filter. Store in sterile 5-mL aliquots at 4°C for up to 6 mo. Use sterile tumor growth media and sterile 1X phosphate-buffered saline (PBS). 3. Methods 3.1. Maintenance of Retroviral Vector Producer Lines 1. Grow the retroviral producer cell line in culture to confluency in large flasks (162cm2 or greater). These cells grow in an adherent monolayer (see Note 1). 2. When the cells have reached confluency, remove the supernatant, and incubate the cells with enough trypsin-EDTA to cover the bottom of the flask (usually 2–3 mL), at room temperature until the cells become nonadherent (usually 1–2 min) (see Note 2). 3. Quench the trypsin with at least three to four volumes of the growth media containing the calf serum. 4. Centrifuge for 10 min at 1500 rpm, at 4°C (see Note 3). 5. Remove the supernatant and count the cells. 6. Replate the cells at about a 1:10 dilution of the number of cells in the confluent flask (about 2 ´ 106 cells per 162-cm2 flask). 7. Split the cell lines 1:10 every 3–4 d, or when each flask reaches confluency (see Notes 4 and 5).
3.2. Preparation of Retroviral Supernatants 1. Two days prior to transduction, trypsinize the producer cells, wash them once, and replate them at a density of 2 ´ 106 cells per 100-mm culture dish (see Note 6). 2. One day prior to transduction, remove the media, and add 10 cc of fresh media to the cells. 3. On the day of transduction, collect a 24-h retroviral supernatant and filter through an 0.45-µm filter to remove contaminating retroviral producer cells (see Notes 7–11).
3.3. Maintenance of Tumor Cell Lines 1. Grow the tumor cells in culture in optimal tumor growth media to confluency (see Note 12).
314
Laheru et al.
2. On the day prior to transduction, replate the cells at a density of 2–5 ´ 105 cells per 75-cm2 culture dish (see Note 13).
3.4. Performing Retroviral Gene Transfer to Tumor Cells 1. Incubate the freshly collected retroviral supernatant with the 10 µg/mL final concentration of the transduction enhancer DEAE-dextran for approx 10 min at room temperature, so that the retrovirus will bind to the enhancer prior to exposure to the tumor cells (see Note 14). 2. Remove the growth media from the tumor cells and replace it with 10 cc of retroviral supernatant containing the enhancer. 3. Incubate the cells at 37°C for 24 h (see Notes 15 and 16). 4. Following incubation of the tumor cells with the retroviral supernatant, remove the supernatant, and wash the cells twice with sterile PBS to rinse away residual retroviral supernatant (see Notes 17–19). 5. Add 10 cc of tumor growth media and allow the cells to grow for 48 h (see Notes 20 and 21).
3.5. Testing for Cytokine Gene Product 1. At 48 h following transduction, remove the growth media, and add 10 cc of fresh tumor growth media. 2. Collect a 24-h supernatant for evaluation of cytokine secretion. To do this, remove the supernatant, centrifuge or filter through a 0.45-µm filter to remove the cells, and aliquot the supernatant into three 1-mL sterile aliquots that can be stored frozen at -70°C until the time of testing for cytokine secretion. Three separate aliquots should be stored so that repeat testing can be performed without multiple freeze– thaw cycles (which might reduce the concentration of the gene product). 3. Following collection of the cell supernatant, take up the cells and count them. Record the total number of cells that contributed to the production of the cytokine over the 24 h. This number will be used to calculate the concentration of cytokine secretion per given number of tumor cells after the concentration of cytokine in the supernatant is determined (see Notes 22–33).
4. Notes 1. Producer lines derived from the NIH 3T3 fibroblast cell line grow in an adherent monolayer. They do best when they are plated at a threshold density of about one tenth the flask’s total cell capacity. Plating the cells at a lower density may result in loss of the cell line. 2. Exposure to trypsin results in the rapid release of the producer cells from the tissue culture flask. Be aware that overexposure to trypsin will result in significant cell death. 3. The MFG producer line usually grows in media supplemented with bovine calf serum. Substitution of fetal bovine serum may result in a change in growth kinetics and viral particle production. The growth requirements recommended by the
Whole Cell Pancreatic Cancer Vaccine
4.
5.
6.
7.
8.
9.
315
laboratory in which the producer line originated should always be used to culture and cryopreserve the producer line being employed. It is advantage to initially expand enough of the producer cells to allow for freezing down of a large stock of aliquots for two reasons. First, prolonged passage in culture may increase the possibility of recombination events within the producer cells which may result in the production of helper virus. Second, there is the theoretical concern that prolonged passage in culture may result in a decrease in the population of producer cells capable of efficiently producing the viral particles. The MFG producer cell lines freeze well in 90% calf serum + 10% dimethyl sulfoxide (DMSO). Recovery of viable producer cells will be severely compromised if these cells are frozen in other types of serum. Each producer line should be frozen in the same type of serum that is used for in vitro growth. These cells can be stored long term in liquid nitrogen. Retroviruses are difficult to titer because they do not form plaques. It is therefore recommended that every retroviral producer line be tittered for transduction efficiency using an easily transducible cell line. NIH 3T3 cells are a good choice for comparison with other murine cell lines. For the transduction of human primary cultures, a human cell line may be a more appropriate cell line for comparison. Tittering can be accomplished by using the transduction procedure described above, and by performing serial two- to fivefold dilutions of the retroviral supernatant prior to exposure of retrovirus to the cell line. Dilutions of the retroviral supernatant should be made with the cell line’s growth media for best results. Most supernatants are optimal either undiluted, or between a 1:2 and 1:10 dilution. Before assuming that insufficient transduction rates are due to low-titer supernatants as a result of a poor supernatant collection, it is important to first determine if the producer cells themselves are still capable of producing high quantities of retroviral particles. Because the producer cells themselves also express the gene encoded by the retroviral vector, a convenient way to evaluate the producer line for production of the vector is to assay the cells for expression of the gene product. However, in vitro loss of high titer producer lines due to long-term culture can easily be avoided by routinely thawing a fresh aliquot of producer cells every 3–4 wk. If expression is at the expected level, then the problem is more likely to be the result of a low-titer retroviral supernatant resulting from suboptimal supernatant collection. There are two major causes of low-titer retroviral supernatants: (1) inadequate retroviral supernatant collection due to insufficient numbers of producer cells or overgrowth of producer cells; and (2) suboptimal growth conditions for retroviral supernatant collection, in particular, a bad lot of calf serum, use of the wrong media and supplements, inadequate CO2 concentration during incubation, and so forth. A recent study performed to evaluate the improvement of retroviral vector production observed that the growth of 21/22 producer cell lines at 32°C for up to 2 wk after the cells reached 100% confluence increased vector titers (102). Growth of the producer lines at 32°C is thought to increase the stability of the viral particles.
316
10.
11.
12.
13.
14.
15.
Laheru et al.
In addition, improved vector production may be caused by the decreased metabolism of the producer cells at this lower temperature. For optimal transduction efficiencies, freshly collected retroviral supernatants should be used. Although it is possible to store the supernatants at 4°C for several days, and to freeze these supernatants at -70°C for several weeks, the efficiency of transduction may decrease by as much as 50% following thaw of the supernatant. Producer lines must be frozen in the same type of serum used for in vitro growth unless otherwise advised. The substitution of other serum may not support the growth of these cells well, and may result in significant cell death during freezing and storage. Most proliferating cell lines can be transduced with a retroviral vector. However, the efficiency of transduction will depend on the percentage of cells that are actively proliferating at the time of exposure to the retrovirus, as integration into the host genome is required for stable expression of the transferred gene. Therefore, the growth conditions for each cell line being transduced should be optimized before attempting this transduction procedure. Most long-term cell lines already have defined growth conditions that support optimal growth. However, it is now possible to transduce many primary, short-term cancer cell lines, and conditions for optimizing their growth may already have been described. The actual density of cells in the flask should be optimized for every tumor cell type, keeping in mind that the cells will need to be able to proliferate maximally for at least 48 hours following transduction to allow for optimal integration and expression of the gene. For cells with a 48 to 72-h doubling time, adequate transduction can be achieved by plating the cells at a density that will result in approximately one-third confluency of the flask on the day of transduction. Both a negative and a positive control group should be included in each transduction experiment, to confirm that gene expression is the result of gene transfer. A good negative control group is to incubate a flask of each cell type to be transduced with retroviral producer cell growth media containing the enhancing polymer alone. An adequate positive control group would include the transduction of any easily transducible cell line with the same lot of retroviral supernatant used to transduce the test tumor cells. Transduction efficiency can be enhanced by the addition of polymers to the retroviral supernatant just prior to exposure of the target cells to the retroviral vector. Enhanced gene transfer is thought to occur via a charge-mediated mechanism that affects virus binding to or penetration of the target cell. The polycations protamine, polybrene, and DEAE-dextran are routinely used for this purpose (103,104). In addition, liposome-forming compounds such as N-[1-(2,3-dioleoyloxy)]-N,N, N-trimethylammonium propane methylsulfate (DOTAP) (Boehringer Mannheim, Indianapolis, IN) have also been successfully used to enhance retroviral gene transfer and may be less toxic to the host cell than other enhancers. Liposome-forming agents probably enhance gene transfer into the host cell by first forming stable interactions with the virus, then adhering to the cell surface, followed by fusing with the cell membrane and releasing the virus into the cell cytoplasm (105). Because
Whole Cell Pancreatic Cancer Vaccine
317
Table 3 Commonly Employed Transduction Enhancing Reagents Concentration range a (final concentration in retroviral supernatant)
Enhancer
Target cell
DEAE-dextran (Sigma, St. Louis, MO) Polybrene (Sigma, St. Louis, MO or Aldrich, Milwaukee, WI) Protamine sulfate (Lilly, Indianapolis , IN) DOTAP (Boehringer Mannheim, Indianapolis, IN)
Murine tumor cell lines Human tumor cell lines Murine tumor cell lines Human tumor cell lines
5–10 10–100 5–10 10–100
µg/mL µg/mL µg/mL µg/mL
Murine tumor cell lines Human tumor cell lines Murine tumor cell lines Human tumor cell lines
5–10 5–100 5–10 10–100
µg/mL µg/mL µg/mL µg/mL
a Final
concentration in retroviral supernatant.
most enhancers are toxic to the cell lines at high concentrations, yet higher concentrations of polymers may be required for enhanced transduction efficiency to some cell lines, it is recommended that a titer of the enhancer be performed on each new batch of enhancers used, to determine the least toxic, most enhancing concentration of the polycation or lipid compound. Table 3 illustrates recommended ranges of polycation and lipid reagent concentrations for the commonly employed transduction enhancers. 16. Longer incubation times will increase the number of proliferating cells that are exposed to the retroviral vector, and therefore may increase the efficiency of transduction. Hardy tumor cell lines may tolerate the retroviral supernatant containing low concentrations of enhancer for 24–48 h without significant cell death. However, primary human tumor cultures may not tolerate a change in the growth media for more than several hours. Therefore, it is best to perform a pilot study evaluating the rate of tumor cell death over time when cells are exposed to the retroviral supernatant containing the enhancer, to optimize the transduction procedure. 17. There is also evidence to suggest that the efficiency of retroviral transduction can be improved by a 90-min centrifugation at 2500 rpm, and 32°C, prior to an overnight incubation (at 32°C) of the tumor cells with the retroviral supernatant (102). However, some tumor cells may not tolerate an overnight incubation at 32°C. 18. It is often useful to perform the initial transduction studies on new tumor cell lines using the retroviral vector containing a marker gene (e.g., the lacz gene that expresses the cytoplasmic enzyme b-galactosidase, which will turn the transduced cell’s cytoplasm blue when exposed to the substrate bluogal or Xgal). Marker genes can be used quantitatively to determine the number of tumor cells in the
318
19.
20.
21.
22.
23.
Laheru et al.
transduced population that are capable of expressing the transferred gene (the transduction efficiency of the vector for a particular tumor cell line). It is not uncommon to have a high titer retroviral supernatant. If this is the case, the supernatant can be diluted 1:5 or 1:10 (depending on titer) with target cell growth media, prior to the transduction procedure, to decrease target cell toxicity from the retroviral supernatant. In fact, be aware that a dilution of a high titer retroviral supernatant may be necessary because higher titer supernatants may contain inhibitors against successful retroviral transduction. Freezing of large stocks of the transduced tumor cells is recommended to prevent loss of gene expression, as well as to prevent in vitro selection with loss of antigen expression. Transduced tumor cell lines can be frozen down and stored in liquid nitrogen long term without loss of gene expression. Controlled-rate freezing is recommended to prevent a significant decrease in viability following thawing. A cheap and efficient way of control-rate freezing is to immerse the freezing vial of cells in a propanol bath (Nalgene Cryo 1°C Freezing Container), and to place the apparatus into a -70°C freezer overnight. This will freeze the cells at approx 1°C per minute. The cells can then be placed into liquid nitrogen for long-term storage. However, more controlled and regulated freezing methods should be used for vaccine intended for clinical use. Primary human tumor lines are more difficult to transduce than long-term established lines. However, with the increasing applications of gene therapy to the clinics, there is an increasing need for improved methods of gene transfer to these cells. The most important criterion for efficient gene transfer to primary human tumor cultures is to optimize the growth conditions for maximally proliferation capacity. In addition, increasing the concentration of transduction-enhancing polymer may result in improved transduction efficiency. It is often beneficial to initially screen the different enhancing polymers for the upper limits of concentration of polymer, and incubation time, that each primary tumor cell line can tolerate, before significant cell death is observed. Following transduction of tumor cells with the cytokine gene, the transduced cells should be evaluated for the total quantity of cytokine produced and for the quantity of cytokine that is bioactive. The total quantity of cytokine produced is best determined by enzyme-linked immunosorbent assay (ELISA). ELISA kits are now commercially available for quantitation of most murine and human cytokines (Genzyme, Endogen, R&D systems). Although these kits are expensive, they usually have a sensitivity of 1–4 pg/mL, and are specific for the cytokine being tested. Bioassays are also available for many murine and human cytokines. Although they are often not as sensitive or specific as ELISA, they provide important information concerning the function of the cytokine being secreted by the tumor cells. Cell lines for bioassay of common murine and human cytokines are listed in Table 4. For all of these assays, serial dilutions are made of the tumor cell supernatants collected as described above. Most of the bioassays rely on cell lines that are growth factor dependent. In these assays, the degree of proliferation of the cell
Whole Cell Pancreatic Cancer Vaccine
319
Table 4 Common Bioassays Used to Quantitate Cytokine Production Cytokine
Bioassay (ref. no.)
Human Murine Human Murine Human Murine Human Murine Human Murine Human Murine Human Murine Human Murine Human Murine
CTLL-2 (106) CTLL (107) TF-1 cells (108) NFS 60 (109) PHA-activated peripheral (110) blood mononuclear cells CT4S or HT-2 cells (108,111) TF-1 cells (108) same T 1165.85.2.1 cells (112) same PHA-activated peripheral (110) blood mononuclear cells same TF-1 cells (108) NFS-60 cells (109) Antiviral assay (113) Antiviral assay (107) Cytotoxic assay (114) Cytotoxic assay (114)
IL-2 IL-2 IL-3 IL-3 IL-4 IL-4 IL-5 IL-5 IL-6 IL-6 IL-7 IL-7 GM-CSF GM-CSF interferon-g interferon-g TNF-a TNF-a
lines in the presence of the serially diluted cytokines is determined by [3H]thymidine incorporation. A recombinant standard is also run along with the test samples to accurately quantitate the cytokine in the test samples. Because several cytokines may stimulate the same cell line, duplicate curves are often run for each sample, one curve in the presence of cytokine blocking antibody, to evaluate the percent of proliferation that is specifically the result of that cytokine. The exceptions are the tumor necrosis factor a (TNF-a) assay, which is a cytotoxic assay; and the interferon-g assay, which is an antiviral assay. The exact procedures for performing these assays can be found in the references listed in Table 4. 24. Genes encoding cytokines are currently one of the most commonly employed genetically altered tumor vaccine strategies, in preclinical models, and in clinical trials. However, other gene modified vaccine strategies, including tumor cell surface expression of MHC class I and II molecules, and costimulatory cell-surface molecules (e.g., B7), are also under investigation. Successful gene transfer of these gene products can be assayed using cell surface staining with monoclonal antibodies specific for the gene product, and analyzed by standard flow cytometric methods. 25. Evaluation of vector copy number should be considered, particularly in cases of suboptimal gene product expression, to determine if the problem is at the level of transcription, or the result of inadequate transduction. Vector copy number can be
320
26.
27. 28.
29. 30.
31. 32. 33.
Laheru et al.
evaluated by Southern blot hybridization using standard procedures. It is also important to perform this procedure for characterizing cells that will be used clinically. If the problem is the result of inadequate transduction and all of the transduction conditions have been optimized, it is possible to improve significantly on the transduction efficiency by subjecting the transduced cells to one or two more rounds of transduction. Sometimes alternative gene transfer vectors are required. Although beyond the scope of this review, there are some general principles to consider when developing vaccine cells for clinical use. It is always a good idea to be familiar with FDA recommendations and to consider a pre-IND meeting to discuss the process to be used. Some general principles to consider are indicated in items 29–33. All procedures should be done under the sterile conditions. Well characterized vector systems should be used to produce transduced cells. There are specific requirements for each vector system. However, all vectors require complete sequencing, a well-documented construction history, sterility testing, and a BLAST search to confirm that it does not contain sequences with potential oncogene activity. Repeat sequencing of the vector should be performed after insertion into the cells to confirm it was inserted without new mutations or new sequences that could be potentially oncogenic. In addition, copy number should be determined. Antibiotic selection markers within vectors are of potential concern because of the potential for antibiotic resistance that they can confer long term. Antibiotics used for selection of genetically modified cells will probably need to be completely eliminated from the final vaccine formulation. Specific parameters should be identified for demonstrating stability of the genetically modified tumor cells. Examples include expression of genes or surface antigens, HLA typing, levels of cytokine production, and so forth.
References 1. Evans, D. B., Abbruzzese, J. L., and Rich, T. A. (1997) Cancer of the Pancreas, in Principles and Practice of Oncology, 5th ed. (DeVita, V. T., Hellman, S., and Rosenberg, S. A., eds.), Philadelphia: J. B. Lippincott, pp. 1054–1087. 2. Bastidas, J. A., Poen, J. C., and Niederhuber, J. E. (2000) Pancreas, in Clinical Oncology (Abeloff, M. D., Armitage, J. O., Lichter, A. S., and Niederhuber, J. E., eds.), Philadelphia: Churchill Livingstone, pp. 1749–1783. 3. 3 Lillemoe, K. D., Yeo, C. J., and Cameron, J. L. (2000) Pancreatic cancer: state of the art care. CA Cancer J. Clin. 50, 241–268. 4. 4 Conlon, K. C., Klimstra, D. S., and Brennan, M. F. (1996) Long term survival after curative resection for pancreatic ductal adenocarcinoma. Ann. Surg. 223, 273–279. 5. 5 Yeo, C. J., Cameron, J. L., Sohn, T. A., et al. (1997) Six hundred fifty consecutive pancreaticoduodenectomies in the 1990s: Pathology, complications and outcomes. Ann. Surg. 226, 248–260.
Whole Cell Pancreatic Cancer Vaccine
321
6. 6 Sohn, T. A., Yeo, C. J., Cameron, J. L., et al. (2000) Resected adenocarcinoma of the pancreas-616 patients: Results, outcomes, and prognostic indicators. J. Gastrointest. Surg. 4, 567–579. 7. 7 Burris, H. A., Moore, M. J., Cripps, M. C., et al. (1997) Improvements in survival and clinical benefit with Gemcitabine as first-line therapy for patients with advanced pancreatic cancer: A randomized trial. J. Clin. Oncol. 15, 2403–2413. 8. McGinn, C. J. and Zalupski, M. M. (2001) Combined modality therapy in pan8 creatic cancer: Current status and future directions. Cancer J. 7, 338–348. 9. 9 Lander, E. S., Linton, L. M., Birren, B., et al. (2001) Initial sequencing and analysis of the human genome. Nature 409, 860–921. 10. Venter, J. C., Adams, M. D., Myers, E. W., et al. (2001) The sequence of the human 10 genome. Science 291, 1304–1351. 11. 11 Velculescu, V. E., Zhang, L., Vogelstein, B., and Kinzler, K. W. (1995) Serial analysis of gene expression. Science 270, 484–487. 12. Zhang, L., Zhou, W., Velculescu, V. E., et al. (1997) Genome expression profiles 12 in normal and cancer cells. Science 276, 1268–1272. 13. 13 Lal, A., Lash, A. E., Altschul, S. F., et al. (1999) A public database for gene expression in human cancers. Cancer Res. 59, 5403–5407. 14. Lash, A. E., Tolstoshev, C. M., Wagner, L., et al. (2000) SAGEmap: A public 14 gene expression resource. Genome Res. 10, 1051–1060. 15. Iacobuzio-Donahue, C. A., Maitra, A., Shen-Ong, G. L., et al. (2002) Discovery of novel tumor markers of pancreatic cancer using global gene expression technology. Am. J. Pathology 160, 1239–1249. 16. Argani, P., Iacobuzio-Donahue, C., Ryu, B., et al. (2001) Mesothelin is over16 expressed in the vast majority of ductal adenocarcinoma of the pancreas: Identification of a new pancreatic cancer marker by serial analysis of gene expression (SAGE). Clin. Cancer Res. 7, 3862–3868. 17. Ryu, B., Jones, J., Blades, N. J., et al. (2002) Relationships and differentially expressed genes among pancreatic cancers examined by large scale serial analysis of gene expression. Can. Res. 62, 819–826. 18. Argani, P., Rosty, C., Reiter, R. E., et al. (2001) Discovery of new markers of 18 cancer through serial analysis of gene expression: Prostate stem cell antigen is overexpressed in pancreatic adenocarcinoma. Cancer Res. 61, 4320–4324. 19. Germain, R. N. (1986) Immunology: The ins and outs of antigen processing and 19 presentation. Nature 322, 687–689. 20. 20 Steinman, R. M. (1991) The dendritic cell system and its role in immunogenicity. Annu. Rev. Immunol. 9, 271–296. 21. 21 Pieters, J. (2000) MHC class II restricted antigen processing and presentation. Adv. Immunol. 75, 159–208. 22. Solheim, J. C. (1999) Class I MHC molecules: Assembly and antigen presentation. 22 Immunol. Rev. 172, 11–19. 23. 23 Hammerling, G. J., Vogt, A. B., and Kropshofer, H. (1999) Antigen processing and presentation—towards the millennium. Immunol. Rev. 172, 5–9.
322
Laheru et al.
24. 24 Pardoll, D. M. (2002) Spinning molecular immunology into successful immunotherapy. Nat. Rev. 2, 227–238. 25. Chen, L., Ashe, S., Brady, W. A., et al. (1992) Costimulation of antitumor immu25 nity by the B7 counterreceptor for the T lymphocyte molecules CD28 and CTLA4. Cell 71, 1093–1102. 26. Schwartz, R. H. (1992) Costimulation of T lymphocytes: The role of CD28, 26 CTLA-4, and B7/BB1 in interleukin-2 production and immunotherapy. Cell 71, 1065–1068. 27. Lechler, R., Aichinger, G., and Lightstone, L. (1996) The endogenous pathway 27 of MHC class II antigen presentation. Immunol. Rev. 151, 51–79. 28. 28 Ostrand-Rosenberg, S. (1994) Tumor immunotherapy: The tumor cell as an antigen presenting cell. Curr. Opin. Immunol. 6, 722–727. 29. 29 Golumbek, P., Lazenby, A., Levitsky, H. I., et al. (1991) Treatment of established renal cancer by tumor cells engineered to secrete interleukin-4. Science 254, 713–716. 30. 30 Dranoff, G., Jaffee, E. M., Golumbek, P., et al. (1993) Vaccination with irradiated tumor cells engineered to secrete murine GM-CSF stimulates potent, specific and long lasting antitumor immunity. Proc. Natl. Acad. Sci. USA 90, 3539–3543. 31. 31 Nishihara, T., Sawada, T., Yamamoto, A., et al. (2000) Antibody-dependent cytotoxicity mediated by chimeric monoclonal antibody Nd2 and experimental immunotherapy for pancreatic cancer. Jpn. J. Cancer Res. 91, 817–824. 32. Bruns, C. J., Harbison, M. T., Davis, D. W., et al. (2000) Epidermal growth fac32 tor receptor blockade with C225 plus gemcitabine results in regression of human pancreatic carcinoma growing orthotopically in nude mice by antiangiogenic mechanisms. Clin. Cancer Res. 6, 1936–1948. 33. Green, M. C., Murray, J. L., and Hortobagyi, G. N. (2000) Monoclonal antibody 33 therapy for solid tumors. Cancer Treat. Rev. 26, 269–286. 34. 34 Tempero, M. (1998) Biologic therapy of gastrointestinal cancer. Cancer Treat. Res. 98, 227–237. 35. Foon, K. A., Yannelli, J., and Bhattacharya-Chatterjee, M. (1999) Colorectal can35 cer as a model for immunotherapy. Clin. Cancer Res. 5, 225–236. 36. 36 Offringa, R., Vierboom, M. P., van der Burg, S. H., Erdile, L., and Melief, C. J. (2000) p53: A potential target antigen for immunotherapy of cancer. Ann. NY Acad. Sci. 910, 223–233. 37. Abbruzzese, J. L. (2000) Molecular diagnosis of pancreatic and biliary cancer: 37 Ready for broad implementation? Cancer J. 6, 282–284. 38. 38 Saforafas, G. H., Tsiotou, A. G., and Tsiotos, G. G. (2000) Molecular biology of pancreatic cancer; oncogenes, tumor suppressor genes, growth factors, and their receptors from a clinical perspective. Cancer Treat. Rev. 26, 29–52. 39. Hruban, R. H., Wilentz, R. E., and Kern, S. E. (2000) Genetic progression in the 39 pancreatic ducts. Am. J. Pathol. 156, 1821–1825. 40. 40 Hahn, S. A. and Kern, S. E. (1995) Molecular genetics of exocrine pancreatic neoplasms. Surg. Clin. North Am. 75, 857–869.
Whole Cell Pancreatic Cancer Vaccine
323
41. 41 Bos, J. L. (1989) ras oncogenes in human cancer: A review. Cancer Res. 49, 4682–4689. 42. Flanders, T. Y. and Foulkes, W. D. (1996) Pancreatic adenocarcinoma: Epidemi42 ology and genetics. J. Med. Genet. 33, 889–898. 43. 43 Hruban, R. H., Van Mansfeld, A. D., Offerhaus, G. J., et al. (1993) K-ras oncogene activation in adenocarcinoma of the pancreas. Am. J. Pathol. 143, 545–554. 44. Gjertsen, M. K., Bakka, A., Breivik, J., et al. (1995) Vaccination with mutant ras 44 peptides and induction of T-cell responsiveness in pancreatic carcinoma patients carrying the corresponding ras mutation. Lancet 346, 1399–1400. 45. Bergmann-Leitner, E. S., Kantor, J. A., Shupert, W. L., Schlom, J., and Abrams, 45 S. I. (1998) Identification of a human CD8+ T lymphocyte neo-epitope created by a ras codon 12 mutation which is restricted by the HLA-A2 allele. Cell Immunol. 187, 103–116. 46. Khleif, S. N., Abrams, S., Allegra, C., et al. (1997) The generation of CD4+ and CD8+ T cell responses from patients vaccinated with mutant ras peptides corresponding to the patient’s own ras mutation. Proc. ASCO 1566. 47. Wojtowicz, M. E., Hamilton, M. J., Bernstein, S., et al. (2000) Clinical trial of mutant ras peptide vaccination along with IL-2 or GM-CSF. Proc. ASCO 1818. 48. 48 Gjertsen, M. K., Buanes, T., Rosseland, A. R., et al. (2001) Intradermal ras peptide vaccination with granulocyte-macrophage colony stimulating factor as adjuvant: Clinical and immunological responses in patients with pancreatic adenocarcinoma. Int. J. Cancer 92, 441–450. 49. Wang, X. Y., Kaneko, Y., Repasky, E., and Subjeck, J. R. (2000) Heat shock 49 proteins and cancer immunotherapy. Immunol. Invest. 29, 131–137. 50. 50 Janetzki, S., Blachere, N. E., and Srivastava, P. K. (1998) Generation of tumor specific cytotoxic T lymphocytes and memory T cells by immunization with tumor derived heat shock protein gp96. J. Immunother. 21, 269–276. 51. Lewis, J. J., Janetzki, S., Livingston, P. O., et al. (1999) Pilot trial of vaccination with autologous tumor derived gp96 heat shock protein–peptide complex (HSPPC96) in patients with resected pancreatic adenocarcinoma. Proc. ASCO 1687. 52. Finn, O. J., Jerome, K. R., Henderson, R. A., et al. (1995) MUC-1 epithelial 52 tumor mucin-based immunity and vaccines. Immunol. Rev. 145, 61–89. 53. Apostopopoulos, V. and McKenzie, I. F. (1994) Cellular mucins: Targets for immunotherapy. Crit. Rev. Immunol. 14, 293–309. 54. Mukherjee, P., Ginardi, A. R., Madsen, C. S., et al. (2000) Mice with sponta54 neous pancreatic cancer naturally develop MUC-1 specific CTLs that eradicate tumors when adoptively transferred. J. Immunol. 165, 3451–3460. 55. Brossart, P., Heinrich, K., and Stuhler, G. (1999) Identification of HLA-A2 restricted T cell epitopes derived from the MUC-1 tumor antigen for broadly applicable vaccine therapies. Blood 12, 4309–4317. 56. Ramanathan, R. K., Lee, K., Mckolanis, J., et al. (2000) Phase I study of a MUC1 synthetic vaccine admixed with SB-AS2 adjuvant in resected and locally advanced pancreatic cancer. Proc. ASCO 1791.
324
Laheru et al.
57. 57 Hammarstrom, S. (1999) The carcinoembryonic antigen (CEA) family: Structures, suggested functions and expression in normal and malignant tissues. Semin. Cancer Biol. 9, 67–81. 58. 58 Marshall, J. L., Hoyer, R. J., Toomey, M. A., et al. (2000) Phase I study in advanced cancer patients of a diversified prime and boost vaccination protocol using recombinant vaccinia virus and recombinant nonreplicating Avipox virus to elicit anti-carcinoembryonic antigen immune responses. J. Clin. Oncol. 18, 3964–3973. 59. Restifo, N. P. (2000) Cancer vaccines: Basic principles, in The Principles and Practice of the Biologic Therapy of Cancer, 3rd ed. (Rosenberg, S. A., ed.), Philadelphia: Lippincott Williams & Wilkins, pp. 571–584. 60. Pardoll, D. M. and Jaffee, E. M. (2000) Cancer vaccines: Clinical applications, in The Principles and Practice of the Biologic Therapy of Cancer, 3rd ed. (Rosenberg, S. A., ed.), Philadelphia: Lippincott Williams & Wilkins, pp. 647–662. 61. Greten, T. F. and Jaffee, E. M. (1999) Cancer vaccines. J. Clin. Oncol. 17, 1047– 61 1060. 62. 62 Fearon, E. R., Itaya, T., Hunt, B., Vogelstein, B., and Frost, P. (1988) Induction in a murine tumor of immunogenic tumor variants by transfection with a foreign gene. Cancer Res. 48, 2975–2980. 63. Golumbek, P., Lazenby, A., Levitsky, H. I., et al. (1991) Treatment of established 63 renal cancer by tumor cells engineered to secrete interleukin-4. Science 254, 713– 716. 64. 64 Dranoff, G., Jaffee, E. M., Golumbek, P., et al. (1993) Vaccination with irradiated tumor cells engineered to secrete murine GM-CSF stimulates potent, specific and long lasting anti-tumor immunity. Proc. Natl. Acad. Sci. USA 90, 3539–3543. 65. Inaba, K., Steinman, R., Pack, M., et al. (1992) Identification of proliferating 65 dendritic cell precursors in mouse blood. J. Exp. Med. 175, 1157–1167. 66. 66 Huang, A. Y., Golumbek, P. T., Ahmadzadeh, M., et al. (1994) Role of bone marrow derived cells in presenting MHC class I restricted tumor antigens. Science 264, 961–965. 67. Nakazaki, Y., Tani, K., Lin, Z. T., et al. (1998) Vaccine effect of granulocyte67 macrophage colony stimulating factor or CD80 gene transduced murine hematopoietic tumor cells and their cooperative enhancement of anti-tumor immunity. Gene Ther. 5, 1355–1362. 68. Golumbek, P. T., Azhari, R., Jaffee, E. M., et al. (1993) Controlled release biodegradable cytokine depots: A new approach to cancer vaccine design. Cancer Res. 53, 1–4. 69. 69 Jaffee, E. M., Abrams, R. A., Cameron, J. L., et al. (1998) A phase I trial of lethally irradiated allogeneic pancreatic tumor cells transfected with the GM-CSF gene for the treatment of pancreatic adenocarcinoma. Hum. Gene Ther. 9, 1951–1971. 70. Simons, J. W., Jaffee, E. M., Weber, C., et al. (1997) Bioactivity of human GM70 CSF gene transduced autologous renal vaccines. Cancer Res. 57, 1537–1546. 71. 71 Simons, J. W., Mikhak, B., Chang, J. F., et al. (1999) Induction of immunity to prostate cancer antigens: Results of a clinical trial of vaccination with irradiated
Whole Cell Pancreatic Cancer Vaccine
72. 72
73. 73
74. 74
75. 75
76.
77. 77 78.
79. 79 80. 80
81. 81 82. 82
83. 83 84.
85.
86. 86
325
autologous prostate tumor cells engineered to secrete granulocyte-macrophage colony stimulating factor using ex vivo gene transfer. Cancer Res. 59, 5160–5168. Soiffer, R., Lynch, T., Mihm, M., et al. (1998) Vaccination with irradiated autologous melanoma cells engineered to secrete human granulocyte-macrophage colony-stimulating factor generates potent antitumor immunity in patients with metastatic melanoma. Proc. Natl. Acad. Sci. USA 95, 13141–13146. Cox, A. L., Skipper, J., Chen, Y., et al. (1994) Identification of a peptide recognized by five melanoma specific human cytotoxic T cell lines. Science 264, 716–719. Kawakami, Y., Eliyahu, S., Delgado, C. H., et al. (1994) Cloning of the gene coding for a shared human melanoma antigen recognized by autologous T cells infiltrating into tumor. Proc. Natl. Acad. Sci. USA 91, 3515–3519. Jaffee, E. M., Schutte, M., Gossett, J., et al. (1998) Development and characterization of a cytokine-secreting pancreatic adenocarcinoma vaccine from primary tumors for use in clinical trials. Can. J. Sci. Am. 4, 194–203. Jaffee, E. M., Hruban, R., Biedzycki, B., et al. (2001) A novel allogeneic GMCSF secreting tumor vaccine for pancreatic cancer: A phase I trial of safety and immune activation. J. Clin. Oncol. 19, 145–156. Davis, M. P., Dinneen, A. B., Landa, N., et al. (1999) Grover’s disease: Clinicopathologic review of 72 cases. Mayo Clin. Proc. 74, 229–234. Boon, T. and Van Den Eynde, B. J. (2000) Cancer vaccines; cancer antigens, in The Principles and Practice of the Biologic Therapy of Cancer, 3rd ed. (Rosenberg, S. A., ed.), Philadelphia: Lippincott Williams & Wilkins. pp. 493–504. Graham, F. L. and Van Der Eb, A. J. (1973) A new technique for the assay of infectivity of human adenovirus 5 DNA. Virology 52, 456–467. Potter, H., Weir, L., and Leder, P. (1984) Enhancer-dependent expression of human kappa immunoglobulin genes introduced into mouse pre-B lymphocytes by electroporation. Proc. Natl. Acad. Sci. USA 81, 7161–7165. Capecchi, M. R. (1980) High efficiency transformation by direct microinjection of DNA into cultured mammalian cells. Cell 22, 479–488. Felgner, P. L., Gadek, T. R., Holm, M., et al. (1987) Lipofection: A highly efficient, lipid-mediated DNA-transfection procedure. Proc. Natl. Acad. Sci. USA 84, 7413–7417. Banerji, J., Rusconi, S., and Schaffner, W. (1981) Expression of a beta-globin gene is enhanced by remote SV40 DNA sequences. Cell 27, 299–308. Kingston, R. E. (1993) Introduction of DNA into mammalian cells, in Current Protocols in Molecular Biology, vol. 1 (Ausubel, F. M., Brent, R., Kingston, R. E., et al., eds.). Hoboken, NJ: John Wiley and Sons. Mulligan, R. C. (1991) Gene transfer and gene therapy. principles, prospects, and perspective, in Etiology of Human Diseases at the DNA Level (Lindsten, J. and Pettersson, U., eds.), New York: Raven Press. Danos, O. and Mulligan, R. C. (1988) Safe and efficient generation of recombinant retroviruses with amphotropic and ecotropic host ranges. Proc. Natl. Acad. Sci. USA 85, 6460–6464.
326
Laheru et al.
87. 87 Mann, R., Mulligan, R. C., and Baltimore, D. (1983) Construction of a retrovirus packaging mutant and its use to produce helper-free defective retrovirus. Cell 33, 153–159. 88. Miller, D. A. and Buttimore, C. (1986) Redesign of retrovirus packaging cell 88 lines to avoid recombination leading to helper virus production. Mol. Cell. Biol. 6, 2895–2902. 89. Armentano, D., Sheau-Fung, Y., Kantoff, P., von Ruden, T., Anderson, W. F., 89 and Gilboa, E. (1987) Effect of internal viral sequences on the utility of retroviral vectors. J. Virol. 61, 1647–1650. 90. 90 Lindemann, D., Patriquin, E., Feng, S., and Mulligan, R. C. (1997) Versatile retroviral vector systems for regulating gene expression in vitro and in vivo. Mol. Med. 3, 466–476. 91. Uberla, K. (2002) Lentivirus vector based on simian immunodeficiency virus. Development and use. Methods Mol. Med. 69, 351–360. 92. Srinivasakumar, N. (2002) Packaging cell system for lentivirus vectors. Preparation and use. Methods Mol. Med. 69, 275–302. 93. Miller, D. A., Miller, D. G., Garcia, V. J., and Lynch, C. M. (1993) Use of retro93 viral vectors for gene transfer and expression. Methods Enzymol. 217, 581–599. 94. 94 Mann, R., Mulligan, R. C., and Baltimore, D. (1983) Construction of a retrovirus packaging mutant and its use to produce helper-free defective retrovirus. Cell 33, 153–159. 95. Miller, A. D., Law, M. F., and Verma, I. M. (1985) Generation of helper-free 95 amphotropic retroviruses that transduce a dominant acting, methotrexate-resistant dihydrofolate reductase gene. Mol. Cell Biol. 5, 431–437. 96. Mann, R. and Baltimore, D. (1985) Varying the position of a retrovirus pack96 aging sequence results in the encapsidation of both unspliced and spliced RNAs. J. Virol. 54, 401–407. 97. Danos, O. and Mulligan, R. C. (1988) Safe and efficient generation of recombi97 nant retroviruses with amphotropic and ecotropic host ranges. Proc. Natl. Acad. Sci. USA 85, 6460–6464. 98. Miller, D. A. and Buttimore, C. (1986) Redesign of retrovirus packaging cell 98 lines to avoid recombination leading to helper virus production. Mol. Cell. Biol. 6, 2895–2902. 99. Bosselman, R. A., Hsu, R. Y., Bruszewski, J., Hu, F., Martin, F., and Nicholson, 99 M. (1987) Replication-defective chimeric helper proviruses and factors affecting generation of competent virus: Expression of Muloney murine leukemia virus structural genes via the metallothionein promoter. Mol. Cell Biol. 7, 1797–1806. 100. Jaffee, E. M., Schutte, M., Gossett, J., et al. (1998) Development and characterization of a cytokine secreting pancreatic adenocarcinoma vaccine from primary tumors for use in clinical trials. Cancer J. Sci. Am. 4, 194–203. 101. Small, J. and Scangos, G. (1983) Recombination during gene transfer into mouse 101 cells can restore the function of deleted genes. Science 219, 174–176. 102. 102 Kotani, H., Newton, P. B., Zhang, S., et al. (1994) Improved methods of retroviral vector transduction and production for gene therapy. Hum. Gene Ther. 5, 19–28.
Whole Cell Pancreatic Cancer Vaccine
327
103. 103 Cornetta, K. and Anderson, F. (1989) Protamine sulfate as an effective alternative to polybrene in retroviral-mediated gene transfer: Implications for human gene therapy. J. Virol. Methods 23, 187–194. 104. Wilson, J. M., Jefferson, D. M., Chowdhury, J. R., Novikoff, P. M., Johnston, 104 D. E., and Mulligan, R. C. (1988) Retrovirus-mediated transduction of adult hepatocytes. Proc. Natl. Acad. Sci. USA 85, 3014–3018. 105. Leventis, R. and Silvius, J. R. (1990) Interactions of mammalian cells with lipid 105 dispersions containing novel metabolizable cationic amphiphiles. Biochem. Biophys. Acta 1023, 124–132. 106. Gearing, A. J. H. and Bird, C. B. (1987) Production and assay of interleukin 2, in Lymphokines and Interferons, A Practical Approach (Clemens, M. J., Morris, A. G., and Gearing, A. J. H., eds.), Oxford: IRL Press, p. 291. 107. Coligan, J. E., Kruisbeck, A. M., Margulies, D. H., Shevach, E. M., and Strober, W. (1991) Current Protocols in Immunology New York: Greene Wiley-Interscience. 108. Kitamura, T., Tojo, A., Kuwaki, T., et al. (1989) Identification and analysis of 108 human erythropoietin receptors on a factor-dependent cell line, TF-1. Blood 73, 375–380. 109. Holmes, K. L., Palaszynski, E., and Fredrikson, T. N. (1985) Correlation of cell109 surface phenotype with the establishment of interleukin3-dependent cell lines from wild-mouse murine leukemia virus-induced neoplasms. Proc. Natl. Acad. Sci. USA 82, 6687–6691. 110. Yokota, T., Otsuka, T., Mosmann, T., et al. (1986) Isolation and characterization 110 of a human interleukin cDNA clone, homologous to mouse B-cell stimulatory factor 1, that expresses B-cell stimulatory factor 1, that expresses B-cell- and Tcell-stimulating activities. Proc. Natl. Acad. Sci. USA 83, 5894–5898. 111. Fernandez-Botran, R., Krammer, P. H., Diamanttein, T., Uhr, W., and Vitetta, 111 E. S. (1986) B cell-stimulatory factor 1 (BSF-1) promotes growth of helper T cell lines. J. Exp. Med. 164, 580–593. 112. Nordan, R. P., Pumphrey, J. G., and Rudikoff, S. (1987) Purification and NH2112 terminal sequence of a plasmacytoma growth factor derived from the murine macrophage cell line P388D1. J. Immunol. 139, 813–817. 113. Meager, A. (1987) Quantification of interferons by anti-viral assays and their standardization, in Lymphokines and Interferons, A Practical Approach (Clemens, M. J., Morris, A. G., and Gearing, A. J. H., eds.), Oxford: IRL Press, p. 129. 114. Mathews, N. and Neale, M. L. (1987) Cytotoxicity assays for tumor necrosis factor and lymphotoxin, in Lymphokines and Interferons, A Practical Approach (Clemens, M. J., Morris, A. G., and Gearing, A. J. H., eds.), Oxford: IRL Press, p. 221.
328
Laheru et al.
Linkage Analysis in Pancreatic Cancer
329
20 Overview of Linkage Analysis Application to Pancreatic Cancer Alison P. Klein
Summary Linkage analysis has aided in the identification of genes involved in many diseases, including several cancers. It relies on using family-based data to detect genetic loci that may harbor disease predisposing genes. Although linkage studies were first designed to find the genes responsible for simple Mendelian diseases (diseases caused by alterations in a single gene), today it is more common for investigators to use linkage analysis to locate genes involved in complex diseases (diseases caused by the independent and joint effects of multiple genes often in conjunction with environmental factors), such as pancreatic cancer. During the past decade linkage analysis has been key step in the identification of several cancer genes, including BRCA2 and STK11, which additional studies have shown also carry an increased risk of pancreatic cancer. However, these known genes explain very little of the observed familial aggregation of pancreatic cancer. While the foundations of linkage analysis are relatively straightforward, the actual implementation of linkage studies, especially for complex diseases such as pancreatic cancer, can be quite difficult. This chapter focuses on the basics of linkage analysis for qualitative traits (affected/unaffected) as could be applied to the study of pancreatic cancer. Key Words: Pancreatic cancer; linkage analysis; LOD score; identical-by-descent (IBD); genetic heterogeneity.
1. Introduction Linkage analysis is a powerful tool that has aided in the identification of genes involved in many diseases. Although linkage studies were first designed to find the genes responsible for simple Mendelian diseases (diseases caused by alterations in a single gene), today it is more common for investigators to use linkage analysis to locate genes involved in complex diseases (diseases caused
From: Methods in Molecular Medicine, Vol. 103: Pancreatic Cancer: Methods and Protocols Edited by: G. Su © Humana Press Inc., Totowa, NJ
329
330
Klein
by the independent and joint effects of multiple genes often in conjunction with environmental factors), such as pancreatic cancer. During the past decade linkage analysis has been key step in the identification of several cancer genes, including BRCA2 (1) and STK11 (2,3), which additional studies have shown also carry an increased risk of pancreatic cancer (4,5). However, these known genes explain very little of the observed familial aggregation of pancreatic cancer. Additional linkage studies designed to localize pancreatic cancer susceptibility genes, such as that recently reported by Eberle et al. (6), who reported evidence of significant linkage to chromosome 4q32–34 in a single family with an aggregation of pancreatic cancer and pancreatic insufficiency, may help to find the gene(s) that cause the familial aggregation of this disease. Although the foundations of linkage analysis are relatively straightforward, the actual implementation of linkage studies, especially for complex diseases such as pancreatic cancer, can be quite difficult. Thus the wise investigator will include a statistical geneticist or a genetic epidemiologist with experience in conducting linkage studies as part of the investigative team from the initial stage of the data collection. In addition to performing the analysis, the statistical geneticist can provide advice on which families should be targeted for recruitment to maximize the probability of detecting disease-causing genes. This chapter focuses on the basics of linkage analysis for qualitative traits (affected/unaffected) as could be applied to the study of pancreatic cancer. Numerous methods also exist for examining linkage for quantitative or continuous traits. However, linkage analysis for quantitative traits is beyond the scope of this chapter. In linkage analysis, as with any statistical analysis, there is the possibility of both type I and type II errors occurring. The type I error, typically denoted by the p-value, is the probability of a false-positive finding, detecting linkage when linkage is not present. A type II error is the probability of a false-negative, declaring there is no linkage when true linkage is present. Conversely, statistical power is the probability of detecting linkage when there is true linkage present and is equal to 1- type II error rate. It is important to consider the probability of both false-positive and false-negative results during both study design and analysis. 2. Materials Linkage analysis relies on using family based data to detect genetic loci that may harbor disease-predisposing genes. There are two major components to the data used for linkage studies. The first are data on families in which there is an aggregation of the disease. These data must include family structure information and precisely defined phenotype data. The second component is the genetic marker data for the individuals within the families.
Linkage Analysis in Pancreatic Cancer
331
2.1. Family Structure and Phenotype Data Family ascertainment is extremely difficult and costly, especially for a rare and rapidly fatal disease such as pancreatic cancer. Given that it is estimated that only about 10% of individuals with pancreatic cancer have a family history of the disease (7) and the rarity of pancreatic cancer, finding families in which there are two cases of pancreatic cancer is difficult. Families recruited for linkage studies can be comprised of just a pair of affected siblings or more extended pedigree data. For a rare disease such as pancreatic cancer recruiting extended pedigrees can be more powerful than recruitment of only affected siblings, because studying more distant pair of relatives (i.e., grandchild–grandparent pairs, cousin pairs) can be more informative for linkage than just sibling pairs. Therefore the same power can be obtained by studying a fewer number of extended pedigrees. For a rapidly fatal disease, such as pancreatic cancer, obtaining biological specimens from affected individuals, to obtain DNA for genotyping, is more problematic than identifying families with multiple affected individuals. Although blood samples from spouses and children can be used to “re-create” the genotypes of deceased family members, there is an inherent loss of information when using “re-created” genotypes that can cause a reduction in power. An additional source of DNA is archival tissue, such as pathological specimens. However, it is difficult to obtain the specimens, and the quality and quantity of DNA available from the specimens are reduced. The creation of registries for families in which there is an aggregation of disease under study, in this case pancreatic cancer, can provide a valuable pool of families for linkage studies. Although only a small proportion of the families who enroll in such registries will be informative for linkage studies, the registries provide a valuable resource of data for a variety of studies examining the etiology, both genetic and nongenetic, of pancreatic cancer, in addition to linkage studies. Detailed collection of phenotype data for all family members is a crucial component to any linkage study and can allow for future stratification of the data. For pancreatic cancer, ideally all affected family members will have pathological confirmation of their disease; however, this is frequently unavailable. Confirmation of the cancer could also be obtained from medical records or death certificates; however, confirmation obtained in this manner can lead to a greater degree of misclassification which can lead in turn to a reduction in the ability to detect the disease causing loci. In addition to the family structure data and precise phenotypic data, data on important covariates should also be collected, including, but not limited to, ageat-onset data for all cases and current age or age-at-death for all unaffected family members, ethnicity, and important environmental risk factors, such as the amount and duration of cigarette smoking for each individual in the dataset.
332
Klein
2.2. Genetic Marker Data Most linkage studies performed today utilize microsatellite marker data. Although, techniques of genotyping are outside of the scope of this chapter, it is important that a high-quality control standard for the genotyping is maintained, because genotyping errors can lead to a reduction in power (8,9). The usefulness of the current marker map is attributable to the availability of a fairly accurate marker map and the high polymorphism of these microsatellite markers. To assess if there is linkage between a disease and a marker loci, one must be able to track the transmission of both the genetic marker and the disease phenotype from generation to generation. To determine which grandparental marker allele is passed from the parent to the offspring, the parent must be heterozygous at that marker locus. Therefore, only heterozygous parents are informative for linkage studies, and the more polymorphic a marker locus is the more informative it is for linkage analysis. The most commonly used measure of a given markers polymorphism is its heterozygosity, which is equal to 1 - S p2i, where pi is the population frequency of each of the marker alleles (10). 3. Methods 3.1. Foundation of Genetic Linkage Genetic linkage occurs when the alleles at two chromosomal loci are inherited together, that is, passed as a single unit from parent to offspring. Linkage occurs in violation of Mendel’s second law, the law of independent assortment which states that traits are determined by discrete factors (genes) that are passed independently from parents to offspring. Genetic linkage occurs when the two chromosomal loci are located physically close to one another on the same chromosomes. Linkage analysis examines whether a disease phenotype is inherited jointly with a genetic marker locus, thereby indicating that the disease locus is physically located nearby the marker locus. The probability that the alleles at two genetic loci (or at the disease locus and at the marker locus) are inherited jointly is denoted as the recombination fraction q, which is a function of the genetic distance between the two loci. When two loci are located on separate chromosomes or far apart on the same chromosome such that they are inherited independently, they are said to be unlinked, with the probability of recombination, q = ½ (i.e., due to chance alone the alleles at the disease locus and marker locus will be inherited jointly one half of the time). Therefore loci are said to be linked when q <1/2. For short map distances, q = 0.01 (1% recombination between the two loci) corresponds to the loci being located approximately 1 cM apart. However, this one-to-one correspondence does not hold up for larger map distances (approx >10 cM) (10). The closer the genetic marker is to the disease locus the greater power there is to detect linkage.
Linkage Analysis in Pancreatic Cancer
333
3.2. Parametic Linkage Analysis The traditional linkage analysis is the LOD score method, which was originally proposed by Morton in 1955 (11). The term LOD score means the “logarithm of the odds.” This method was developed to test for the presence of genetic linkage between two loci in sequential sample of families and is the log to the base 10 of the ratio of the likelihood of the probability that the disease and marker loci are linked (q < 0.5, the alternative hypothesis) to the probability that the disease and marker loci are unlinked (q = 0.5, the null hypothesis). Thus, LOD score, Z, is the ratio of the likelihood of the observed pedigree data at q = x where x < 1/2 to the likelihood of the observed pedigree data at q = 1/2: Z(q) = log10
(
Likelihood q
)
Likelihood 0.5
To determine the likelihoods for the overall sample of families at a given q, the likelihood, for each of the independent families would be multiplied together. However, since the family-specific LOD scores are log10 of the likelihood, the LOD scores for each of the independent families may be added together, given the mathematical properties of logarithms. The LOD score is then maximized with respect to q over a given family or set of families. Thus calculating, Zmax(q), the LOD score maximized with respect to q, is essentially computing the maximum likelihood estimate of q (12,13). For simple Mendelian diseases, a total LOD score >3 is typically viewed as being strong evidence in favor of linkage and is equivalent to odds of 1000:1 in favor of linkage. A LOD score of 3 is equal to a c2 of approx 13.8 with one degree of freedom. This corresponds to a p-value of approx 0.0001 (14). To a calculate LOD score for a given set of families, one must specify a genetic model for the disease including the number of genetic loci, the mode of inheritance at each locus (dominant, recessive, codominant), frequency of the diseasecausing allele, penetrance of the disease for each genotype, phenocopy rate, and marker-allele frequencies. The probability that an individual who has inherited a high-risk genotype develops disease is referred to as penetrance. For simple Mendelian diseases, there is typically 100% penetrance; however, for complex diseases, not every individual who inherits the disease predisposing genotype develops disease. In addition, in complex diseases there are also individuals who do not carry the “high-risk” genotype but go on to develop disease. These individuals are referred to as phenocopies. One method used to determine a genetic model for a particular disease or trait is segregation analysis. Segregation analysis is a process by which a series of statistical models, both genetic models and nongenetic models, are fit to the observed pedigree data. If a genetic model provides an adequate fit to the observed pedigree data, then the results
334
Klein
support the involvement of genes in the etiology of the disease. This genetic model can then be used in the linkage analysis. Given that a previous segregation analysis of pancreatic cancer supported the involvement of a rare dominant gene in the etiology of pancreatic cancer (15), this model may be used in parametric linkage analysis of pancreatic cancer. In general, parametric linkage methods are fairly robust to misspecification of the disease-allele frequency, and somewhat robust to misspecification of the penetrance and phenocopy rate. Misspecification of the mode of inheritance can greatly reduce the power of model based linkage analysis (14,16). Misspecification of the marker allele frequency may lead to incorrect linkage findings (10,14); however, this is not a limitation restricted to parametric linkage analysis and is discussed in more detail below. The major weakness of parametric linkage analysis is the need to specify an underlying genetic model for the disease, which includes the number of loci involved, the mode of inheritance of the disease (dominant, recessive, or codominant), frequency and penetrance of the disease-causing allele, and the phenocopy rate. For many complex diseases, there is no established genetic model. In addition, the results of segregation analysis may not be generalizable to other populations, such that the genetic model obtained in a segregation analysis of one set of families may not necessarily be correct for another set of families. Nonparametric methods overcome this need for a prespecified genetic model. However, the power of nonparametric methods is not a great as the power of the parametric methods when the genetic model is correctly specified. In addition, parametric methods have been developed that jointly perform linkage and segregation analysis in which both the recombination fraction and genetic model parameters are estimated (16). There are a variety of programs available for parametric linkage analysis, such as LODLINK, GENEHUNTER, LINKAGE, FASTLINK, and VITESSE. However, these programs implement the underlying theory of parametric linkage analysis in different ways, and it is therefore important to understand the strengths and weakness of each of the different programs before applying them to any given dataset. 3.2.1. Heterogenity Traditional parametric linkage analysis has been extended to allow for genetic heterogeneity. Genetic heterogeneity can arise when there are multiple major genes that act to increase susceptibility to disease. For example, both BRCA1 and BRCA2 act to increase susceptibility of early-onset breast cancer. Under the assumption of homogeneity, most linkage methods allow for only a single major gene. Families that are not linked to the particular marker of interest provide evidence against linkage in that region, while the families that are linked
Linkage Analysis in Pancreatic Cancer
335
to that marker locus provide evidence in favor of linkage. The combination of these negative and positive linkage signals often results in failure to obtain significant evidence in favor of any linkage to the region. There are several commonly used approaches to allow for the presence of genetic heterogeneity. Given that some diseases are found at higher frequency in population isolates, which have derived from small founder populations, and that therefore may have less genetic heterogeneity (perhaps only a single gene acts to increase susceptibility to that disease in this population), then one strategy is to limit recruitment to that population. Thus, although, in the general population there may be genetic heterogeneity, the investigators may choose to study a population in which it is likely that only a single gene acts to increase the risk of disease. This strategy may work well for relatively rare diseases, but tends to be less successful for common ones, as several different susceptibility genes may then have been present in the small group of founders. In addition to limiting the amount of heterogeneity present during study recruitment, heterogeneity can also be allowed for in the analysis of the data. Analytic methods that allow for heterogeneity are more powerful than methods that assume homogeneity when heterogeneity is actually present. Another way of allowing for heterogeneity in the data is to divide the families based on observable clinical criteria, such as age at onset or the occurrence of additional cancers of a specific type in the family (i.e., aggregation of breast and ovarian cancer), under the assumption that families who meet the criteria (i.e., early-onset families) are genetically distinct from other families. When there is some evidence to support the criteria by which the families are stratified, this method can yield greater power, because the p-values do not need to be adjusted for additional tests. However, frequently there are no clinical criteria by which families can be stratified. Therefore, analytical methods have been developed that allow for a mixture of families that are linked to the locus and families that are unlinked, where a denotes the proportion of families that are linked to the marker locus. Thus the LOD score can be maximized over both the genetic distance between the marker locus and the disease (q), but also over the proportion of linked families (a). This statistics is denoted as the HLOD, for heterogeneity LOD score. It is important to note that although this is a valid test for linkage, the derived estimate of a can be biased and thus not correctly reflect the proportion of linked families (17,18). In addition, by estimating both the recombination fraction (q) and the proportion of linked families (a), a correction between 0.24 and 0.47 to the LOD score criteria for significant linkage (19) should be made, such that the corresponding p-value will be adjusted for the estimation of the additional parameter.
336
Klein
3.3. Nonparametic Linkage Analysis All nonparametric linkage analysis, also typically referred to as model-free linkage analysis, rely on assessing the proportion of alleles shared among pairs of relatives. Penrose (20) developed the first method, which examined allele sharing among siblings as a test for linkage. These methods examined alleles shared identical-by-state (IBS) among sibling pairs, that is, sibling who carry the same allele at the same locus. These methods were later extended, first by Haseman and Elston (21) to examine alleles shared identical-by-descent (IBD). IBD means that these alleles have been transmitted to each individual in the relative pair from a common ancestor; in the case of sibling pairs, from either the father or the mother of the sibling pair. Today there are numerous statistical tests for linkage that have been developed based on allele sharing not only among sibling pairs but among various sets of relative pairs (see McPeek [22] for a review of affected pair test statistics). Each of these tests has it own advantages and limitations and careful consideration of the statistical theory underlying each of the methods should be given before choosing the method and interpreting the results. Although, many of these tests are limited to testing for linkage among affected sibling pairs only (ASP), several tests are also available for examining linkage among more distant relative pairs. These methods can include both affected and unaffected relative pairs. Limiting these tests to affected pairs can increase power to detect linkage when there is incomplete penetrance. However, inclusion of affected pairs in which at least one of the cases is not due to inheritance of a gene can reduce the power of these methods that utilize affected relatives only (10,23). For a late age-at-onset disease such as pancreatic cancer, considering affected relatives only may have the benefit that family members who have inherited the predisposing gene but have not yet developed disease will not provide evidence against linkage. Although a variety of statistical tests are used to determine allele sharing among relative pairs, the biological underpinnings of these methods are quite similar. For simplicity let us examine sharing among sibling pairs. Siblings can share zero, one, or two alleles IBD at a given marker locus. The expected probability of siblings sharing zero, one, and two alleles is ¼, ½, ¼, respectively, at any locus. When a marker locus is linked to the disease locus, affected pairs of siblings are expected to share one or two copies of the susceptibility allele at the disease locus and thus, on average, the probability that an affected sibpair share one or two disease locus alleles IBD should be greater than ½ and ¼, respectively. When examining sibling pairs in which one sibling is affected and the other unaffected, if there is linkage, then the observed sharing of one or two marker alleles IBD may be less than expected under the null hypothesis of no linkage.
Linkage Analysis in Pancreatic Cancer
337
The computer program GENEHUNTER, which is commonly used for nonparametric linkage analysis, uses the NPL statistic as a test for linkage. The NPL statistic is the sum of the weighted standardized IBD sharing within a pedigree. IBD sharing is calculated in two ways, denoted, Spairs and Sall, where Spairs is the number of pairs of alleles shared IBD among all pedigrees members and Sall is weighted toward pedigrees in which there are three or more affected individuals who share an allele IBD. Thus, the IBD sharing is standardized for a given pedigree (i) by the following formula: Zi =
Si - µ i si
where µ i and s i denote mean and variance of the expected sharing for a given pedigree (24). However, the NPL score is conservative when genotype data are missing on some family members, marker heterozygosity is low, and markers are widely spaced. Alternative statistics that are less conservative have been proposed (25). In addition to GENEHUNTER there are numerous other programs that perform nonparametric linkage analysis. 3.4. Two-Point and Multipoint Linkage Analysis The previous methods we have discussed are referred to as two-point linkage analysis, meaning we are assessing evidence for linkage between a single genetic marker and purported disease locus. Multipoint linkage analysis differs from traditional two-point linkage analysis in that several adjacent markers are simultaneously examined for evidence of linkage to a disease locus. Multipoint analysis can be much more powerful than two-point analysis when the map order of the marker loci is correctly known. However, incorrect map order and/or intermarker distances can cause a grave reduction in the power of multipoint analysis. Many of the same test statistics mentioned previously can be applied to multipoint analysis. Some of the programs currently available that are commonly used for multipoint linkage analysis include GENEHUNTER, ALLEGRO, and FASTLINK. 3.5. Genome Scans, Candidate Regions, and Fine Mapping There are two main approaches to linkage studies, the candidate region approach and the full genome scan. When there is strong evidence, including evidence from linkage studies, cytogenetic analysis, and so forth, that a particular chromosomal region may harbor disease causing genes, investigators may wish to perform linkage analysis in this region. However, for many diseases, there is little a priori evidence to support the involvement of a particular region, or it is likely that multiple genes located in different areas of the genome may be
338
Klein
involved in disease susceptibility. Thus, markers throughout the genome are assessed for the presence of linkage to the disease. This is commonly referred to as a genome scan. The microsatellite marker sets used today are comprised of approximately 400 markers evenly spaced throughout the genome, approximately 10 cM apart. However, denser microsatellite maps with markers spaced approximately every 5 cM and very dense SNP (single-nucleotide polymorphism) sets may soon be available. Additional markers are often typed in regions that give evidence in favor of linkage in the genome-wide scan analysis, thus increasing the marker density, and thereby increasing the power to detect loci and to more precisely localize the disease causing gene. This process of genotyping additional markers in suggestive regions is referred to as fine-mapping. Finemapping can allow for the creation of shared haplotypes for affected family members, and thereby pinpointing where key recombination events have occurred and thus narrowing the chromosomal region in which the disease causing gene could be located. 3.6. Incorporating Covariates For a disease with strong known environmental risk factors, the power to detect linkage can be increased when the environmental risk factor(s) can be incorporated into the model. Environmental covariates can be incorporated into parametric linkage analysis, through (1) their incorporation into segregation analysis models, the results of which are directly used in the linkage analysis, or (2) through the establishment of different estimates of penetrance for those exposed and unexposed to environmental risk factors. Mandal et al. (26) showed that there was a significant increase in power for both the Hasemen–Elston sib-pair linkage test and several affected sib-pair tests when individuals who were unexposed to the environmental risk factor were considered to have an unknown phenotype. 4. Notes Although this chapter gives an overview of the variety of analytical methods available for linkage studies, each with its own strengths and limitations, it is important to note that once the expense and difficulty in recruiting family members to participate in a linkage study, collecting samples for DNA, and the expense of genotyping have been incurred, a variety of analytical techniques is employed for each dataset, to reduce the chance that a region containing a “true” disease-predisposing loci is not detected. Depending on the size and structure of the families used in the analysis, some programs may offer significant advantages or disadvantages over others. In addition, evidence for linkage observed
Linkage Analysis in Pancreatic Cancer
339
by a variety of methods may be less likely to be a “false”-positive result. However, there is always the trade-off between power to detect “true” loci and the risk of obtaining false-positive results. By not detecting “true” loci we can potentially greatly delay the detection of important disease causing loci but by detecting false-positives we can expend years of effort and large amounts of research funds on fruitless positional cloning studies. Given the wide variety of statistical tests available, and the numerous different test statistics reported in the literature, it is important to focus on the p-value corresponding to the test, not the score itself. This is because the p-value for each of the tests is providing information of the probability of a false-positive result, and the p-value corresponding to a LOD score of 3 is different from the corresponding p-value of an NPL score of 3. In addition to considering the probability of a false-positive result for a single marker (the nominal significance level), the experiment-wide significance level (the probability of a false-positive result among any of the markers tested in a given study) must also be considered. Lander and Kruglyak (27) proposed the following guidelines to assess whether there is evidence for linkage in a genome-wide scan. These reporting guidelines are generally used in the literature today. For declaring significant linkage, they propose the probability of a false-positive result should be one in every five genome scans, which corresponds to a LOD score of 3.3. For declaring suggestive linkage, they suggest a probability of a false-positive result should be 1 in every genome scan, which corresponds to a LOD score of 1.9. In addition, when there is a previous evidence of linkage, through a single or several linkage studies, they suggest the p-value of 0.01 to confirm (replicate) linkage to a region in which there is previous evidence for linkage. Confirmation that a chromosomal region is linked to a disease in an independent study helps further establish that a disease-causing gene is located in that region. However, if the results of a subsequent linkage study do not confirm the previously reported results, this does not necessarily indicate that the original linkage reported was a false-positive result. There are many reasons why a reported linkage may not be confirmed by a future study. One important reason is genetic heterogeneity: the genes that are important determinants of disease in the first population are not the same genes that are involved in the second population, possibly due to ethnic differences, differences in phenotype definition (i.e., a gene that causes early-onset cancers may not important when studying a collection of late age-at-onset families) and differences in environmental exposures (i.e., if a gene acts to cause cancer in a group of smokers but not in nonsmokers, it may not be detected among families with a smaller proportion of smokers). Lack of statistical power is another reason why some linkage results
340
Klein
are not confirmed, in that linkage studies have limited power to detect genes of modest effect. If a gene has a smaller effect in the families used to replicate the initial finding, the linkage results may not be confirmed. Linkage studies have aided in identification of some cancer-predisposing genes, which additional studies have shown also carry an increased risk of pancreatic cancer (e.g., BRCA2). It is hoped that careful application of these methods will lead to the identification or confirmation of additional pancreatic cancer loci. Acknowledgments I would like to thank Dr. Joan E. Bailey-Wilson for her helpful suggestions and editorial comments in preparing this manuscript. References 1. 1 Wooster, R., Neuhausen, S. L., Mangion, J., et al. (1994) Localization of a breast cancer susceptibility gene, BRCA2, to chromosome 13q12-13. Science 265, 2088–2090. 2. 2 Hemminki, A., Tomlinson, I., Markie, D., et al. (1997) Localization of a susceptibility locus for Peutz-Jeghers syndrome to 19p using comparative genomic hybridization and targeted linkage analysis. Nat. Genet. 15, 87–90. 3. 3 Amos, C. I., Bali, D. , Thiel, T. J., et al. (1997) Fine mapping of a genetic locus for Peutz-Jeghers syndrome on chromosome 19p. Cancer Res. 57, 3653–3656. 4. 4 Goggins, M., Schutte, M., Lu, J., et al. (1996) Germline BRCA2 gene mutations in patients with apparently sporadic pancreatic carcinomas. Cancer Res. 56, 5360–5364. 5. 5 Su, G. H., Hruban, R. H., Bansal, R. K., et al. (1999) Germline and somatic mutations of the STK11/LKB1 Peutz-Jeghers gene in pancreatic and biliary cancers. Am. J. Pathol. 154, 1835–1840. 6. 6 Eberle, M. A., Pfutzer, R., Pogue-Geile, K. L., et al. (2002) A new susceptibility locus for autosomal dominant pancreatic cancer maps to chromosome 4q32-34. Am. J. Hum. Genet. 70, 1044–1048. 7. 7 Klein, A. P., Hruban, R. H., Brune, K. A., Petersen, G. M., and Goggins, M. (2001) Familial pancreatic cancer. Cancer J. 7, 266–273. 8. 8 Douglas, J. A., Boehnke, M., and Lange, K. (2000) A multipoint method for detecting genotyping errors and mutations in sibling-pair linkage data. Am. J. Hum. Genet. 66, 1287–1297. 9. 9 Abecasis, G. R., Cherny, S. S., and Cardon, L. R. (2001) The impact of genotyping error on family-based analysis of quantitative traits. Eur. J. Hum. Genet. 9, 130–134. 10. Ott, J. (1999) Handbook of Human Genetic Linkage. Baltimore, MD: The Johns Hopkins University Press. 11. 11 Morton, N. (1955) Sequential tests for the detection of linkage. Am. J. Hum. Genet. 7, 277–318. 12. Khoury, M. J., Beaty, T. H., and Cohen, H. (1993) Fundamentals of Genetic Epidemiology. Baltimore, MD: The Johns Hopkins University Press.
Linkage Analysis in Pancreatic Cancer
341
13. 13 Ott, J. (1974) Estimation of the recombination fraction in human pedigrees: Efficient computation of the likelihood for human linkage studies. Am. J. Hum. Genet. 26, 588–597. 14. Xu, J., Meyers, D., and Pericak-Vance, M. A. (1998) Lod score analysis, in Approaches to Gene Mapping in Complex Human Diseases (Pericak-Vance, M. A. and Haines, J. L., eds.), New York: Wiley-Liss, pp. 253–271. 15. 15 Klein, A. P., Beaty, T. H., Bailey-Wilson, J. E., Brune, K. A., Hruban, R. H., and Petersen, G. M. (2002) Evidence for a major gene influencing risk of pancreatic cancer. Genet. Epidemiol. 23, 133–149. 16. Clerget-Darpoux, F., Bonaiti-Pellie, C., and Hochez, J. (1986) Effects of mis16 specifying genetic parameters in lod score analysis. Biometrics 42, 393–399. 17. 17 Whittemore, A. and Halpern, J. (2001) Problems in the definition, interpretation, and evaluation of genetic heterogeneity. Am. J. Hum. Genet. 68, 457–465. 18. Hodge, S., Vieland, V., and Greenberg, D. (2002) HLODs remain powerful tools 18 for detecting linkage in the presence of genetic heterogeneity. Am. J. Hum. Genet. 70, 556–559. 19. Abreu, P., Hodge, S., and Greenberg, D. (2002) Quantification of type 1 error 19 probabilities for heterogeneity lod scores. Genet. Epidemiol. 22, 156–169. 20. 20 Penrose, L. (1935) The detection of autosomal linkage in data which consist of pairs of brothers and sisters of unspecified parantage. Ann. Eugen. 18, 120–144. 21. Haseman, J. K. and Elston, R. C. (1972) The investigation of linkage between a 21 quantitative trait and a marker locus. Behav. Genet. 2, 3–19. 22. 22 McPeek, M. S. (1999) Optimal allele-sharing statistics for genetic mapping using affected relatives. Genet. Epidemiol. 16, 225–249. 23. Bishop, D. T. and Williamson, J. A. (1990) The power of identity-by-state methods 23 for linkage analysis. Am. J. Hum. Genet. 46, 254–265. 24. 24 Kruglyak, L., Daly, M. J., Reeve-Daly, M. P., and Lander, E. S. (1996) Parametric and nonparametric linkage analysis: A unified multipoint approach. Am. J. Hum. Genet. 58, 1347–1363. 25. 25 Kong, A. and Cox, N. J. (1997) Allele-sharing models: LOD scores and accurate linkage tests. Am. J. Hum. Genet. 61, 1179–1188. 26. Mandal, D. M., Sorant, A. J., Pugh, E. W., et al. (1999) Environmental covariates: 26 Effects on the power of sib-pair linkage methods. Genet. Epidemiol. 17(Suppl. 1), S643–S648. 27. Lander, E. and Kruglyak, L. (1995) Genetic dissection of complex traits: Guidelines for interpreting and reporting linkage results. Nat. Genet. 11, 241–247.
342
Klein
Index
343
Index A acinar, 70, 72, 114–117, 178, 218, 221, 225–231, 234–235, 259–260, 263, 274–277 cell carcinoma, 70, 72, 218, 228 activin, 275 adenocarcinoma, 1–3, 7, 68–70, 72, 74, 91, 103–104, 110, 116, 175, 176, 178, 185–187, 221, 224, 228–229, 231, 259–260, 273–274, 299, 305–306, 308 adenoma, 1, 3–8, 199–201 adenovirus (see viral/virus) Affymetrix, 131, 133, 175–177, 181–186 allele/allelic, 137–140, 143,151, 221, 232, 234, 250–251, 253–254, 332–334, 336–337 imbalance, 137, 140 loss, 143 allogeneic, 299–300, 306, 308–309, 312 amylase, 70, 115, 229, 277 anchoring enzyme, 162, 164, 167 animal model (also see carcinogen, hamster, gene-targeted, mouse, rats, transgenic, and zebrafish), 114, 217–218, 220, 224, 228, 234–236, 276, 303 antibody/antibodies, 10, 34, 37, 43, 45, 58–60, 75, 77–79, 83, 90, 99, 116, 126, 128, 130–131, 177, 207–210, 213, 215, 222, 267–268, 277–278, 281, 283, 286, 292, 300–304, 319 antigen/antigenic, 10, 59, 68, 77, 116–117, 130, 207–208, 215, 228, 281, 286, 292, 299–309, 312, 318, 320 retrieval 59, 77–78, 130 antigen presenting cells (APC), 307–308
antisense (also see morpholinos and RNA), 281–282, 284, 288–289 APC, 70, 72–73, 235, 5-aza-2’-deoxycytidine (5-aza-dC), 123–125, 127–129, 133–134 B bacteria, 17, 128, 131, 171, 209, 211, 213–214, 248, 304 basic helix-loop-helix (bHLH), 275 B-cells, 209, 215, 300-304 β-catenin, 70, 72–74, 228, 235 β-galactosidase, 231–232, 317 bioinformatics, 191, 194 biomarker 59, 89, 189-191, 195, 309 biotin, 10, 37, 45, 75–79, 82–83, 128, 134, 146, 162, 164, 166, 172, 177, 180 bisulfite modification, 123–124, 126–127, 130–131 BLAST search, 276, 320 blood, 16, 19, 43, 59–60, 68, 74, 109, 114, 264, 319, 331 bombesin, 115, bovine pituitary extract, 115, 118 BRCA2, 144, 150, 305, 329–330, 334, 340 BrdU (5-bromo-2’-deoxyuridine), 281, 284, 290–291 C cancer vaccine (see vaccine) carbachol, 115 carbohydrate antigen (CA), 68, 72 carboxypeptidase A, 278–280
343
344 carcinoembryonic antigen (CEA), 68, 72, 222, 303, 306 carcinogen/carcinogenic, 40, 57, 113, 217–221, 224, 228–232, 235, 260, 292 carcinoma in situ, 2, 4–6, 8, 229–230 caudal–rostral axis, 265 cDNA, 18, 60, 63, 80, 117, 119, 113, 144, 158, 161–162, 164, 166–168, 172–173, 176, 179–180, 207–210, 214–215, 311 cell aggregation, 269 concentration, 20, 34, 42 culture, 97, 114–116, 128, 133–134, 166, 173, 178, 210, 221 differentiation (see differentiation) enrichment, 50 fixed, 92, 97 frozen, 125, 118, 315, 316, 318 line, 91–92, 97, 103-104, 106, 109, 114, 117, 125, 129, 151, 175–176, 178, 186–187, 190, 203, 208, 210, 214–215, 219–220, 223, 260, 274, 308, 311–319 lysis, 34, 43 microdissection (see microdissection) proliferation, 114 purification, 16 separation, 33, 59, 114 transplant, 217–223, 235 types, 15–16, 20–21, 114, 117, 221, 224–225, 231, 234, 260 xenograft (see xenograft) cellular/cellularity, 20, 25, 27, 29, 41, 59, 103, 115, 117, 143–144, 178, 190, 217–218, 220–221, 224–225, 235, 259, 282, 292, 300–301, 304–305, 308 chimera/chimeric, 232–233 cholera toxin, 115–116 chronic pancreatitis (also see pancreatitis), 260 chymotrypsin, 70, 115
Index clinical, 1, 3, 67, 70, 74, 90, 92, 137, 273–274, 282, 299–300, 303–310, 318–320 collagen, 114–116 260–264 collagenase, 33, 42, 115, 119, 261, 262 combined bisulfite restriction analysis (COBRA), 124–126, 128 CpG island, 123–127, 129–132, 134, 144 Cre/loxP, 227, 231–232, 234, 245–249, 251, 253–254, 256 cryostat sectioning (see tissue sectioning) cystic fribrosis transmembrane conductance regulator (CFTR), 115 cytokeratin, 68, 72, 75, 114–116, 119, 225, 227 cytokeratin 19 (CK19), 114–116, 119, 225, 227–229, 231 cytokine, 33, 299–300, 304, 307, 309–310, 314, 318–320 cytologic/cytology, 3–7, 21, 33, 58–59, 62, 202–203 D DAB (3’,3-diaminobenzidine), 37, 45, 60, 75–76, 78–79, 82–83, 215 delayed-type hypersensitivity (DTH), 306–308 dendritic cells, 301, 304–305, 307–308 deparaffination, 78 development (see pancreatic development or tumor development) dexamethasone, 115, 261–262, 269 diagnosis/diagnostic, 1, 4, 6, 19, 67–68, 91, 93–94, 97, 123–124, 133, 189, 191, 199–201, 218, 273, 303 differentiation, 68, 70, 72, 74–75, 114–116, 126, 143, 208, 223, 231, 259–260, 274–277, 279–280, 305, 307 digital, 137–140
Index DNA amplification, 7, 63, 79, 82, 127, 132, 134, 137, 127, 147–149, 154, 157, 169,172, 203 collection, 331, 338 denature, 130, 291 digestion (dsDNA), 39, 61, 127, 129, 131–133, 144, 153, 155, 167–168, 170 electroporation, 309 extraction, 7–9, 39, 53, 153–154, 156–157, 167, 168–171, 214, 289 isolation, 19, 38–39, 37, 52–53, 57, 103, 126, 151 ligation, 165, 167–169, 171, 173, 179, 214 methylation (also see methylation), 126–127, 129, 131–132, 134, 221 microinjection, 224, 309 probe, 79–80 proviral, 311 purification, 40, 127, 130, 134, 138, 140 quality, 39, 57, 134, 137–138, 166, 173 quantity, 20, 28, 35, 39, 47, 60, 126, 140, 151, 157, 171, 214 sequencing (also see sequencing), 162, 164–165, 172–173, 209 transfection, 309 DNase, 41, 56, 283, 289 dorsal, 265, 267, 269, 275 DPC4 (also see SMAD4), 9–10, 68–70, 72, 74, 305 ductal, 1–7, 68, 70, 72, 74, 104, 113–119, 203, 218, 221, 223–226, 228–231, 259–260, 273–274, 306 ductal epithelial cells (DEC), 113–117 E E-cadherin, 123–124, 222
345 elastase (EL), 224–230, 234 electrophoresis, 41, 133, 149, 156, 164, 166, 173, 189–190, 193, 203, 247, 289 embryo/embryonic, 124, 224, 232–233, 245, 259–261, 264–265, 267, 276–277, 279–282, 284–287, 289-293, 301–303, embryonic stem cell (ES cell), 227, 232–233, 245–247, 249–250, 254, 256, 289–293 endocrine cell /lineage, 68, 114, 116–117, 231, 259, 274–277, 279 hormone, 281 neoplasm, 72, 74–75 endoplasmic reticulum (ER), 19, 301–303 endoscopic retrograde pancreatography (ERP), 200–201, 203 enzyme-linked immunosorbent assay (ELISA), 318 epidermal growth factor (EGF), 115, 118, 276 epithelial/epithelium, 3–7, 9, 33, 50, 68, 72, 90, 115–119, 175–176, 178, 186, 200–202, 208, 225, 229–231, 259–260, 262–265, 267–269, 275, 277, 279, 301, 306 (epithelial aggregate separation and isolation (EASI), 7, 9, 50–51, 59, 62 epithelial membrane antigen (EMA), 68, 72 Epstein-Barr virus (see viral/virus) exocrine, 68, 72, 114, 217, 220, 224, 228, 231, 262–263, 274–281 explant, 114–117, 259–261, 263–265, 267 F familial, 260, 329–330 fibroblast, 114–117, 311, 314
346 fibroblast growth factor-2 (FGF-2), 275 Ficoll, 34, 43, 59, 115–116, fine-needle microdissection, 7 fluorescence/fluorescent, 19, 24–25, 89, 140, 185, 264, 283, 286–288, 292 follistatin, 275 formalin-fixed (also see tissue), 7, 9, 18, 35–36, 41, 43, 56–59, 62, 77, 79, 92, 97, 109, 128, 130 frozen (also see tissue), 9, 16, 18, 20, 33–36, 39, 41, 44–45, 47, 53, 58–63, 77, 125, 176, 178–179, 187, 210, 214 frt/flp, 231, 235, 246–247, 250–251, 253–254 G γ-glutamyl transferase, 115 glycosyl transferase, 306 gene/genetic activation, 310 alteration (also see mutation), 1–4, 144–145, 220, 229, 231, 234, 259 chip, 175–176, 183–185 cloning, 150, 207, 209, 311 disease-causing, 330–331, 334, 337–339 expression, 79–80, 123–124, 127, 129, 133–134, 161–162, 175–176, 178, 186–187, 189, 217, 224–225, 227–232, 235, 245, 274–275, 306, 309, 310, 315–316, 318–320 heterogeneity, 232, 267, 334–335, 339 identification 144, 150, 176, 178, 209, 229, 282, inactivation, 68, 123–124 knockdown, 277, 281–282 knock-out (also see knockout mice), 232, 245, 250, 253 linkage (see linkage)
Index marker (also see biomarker, microsatellite, molecular marker), 199–200, 232, 317, 330, 332–339 mutation (also see gene alteration and mutation), 74, 200, 217–218, 219, 233–234 regulation, 224, 225–226, 228, 231, 245, 274–276 sequencing (see sequencing) silencing, 123–124 targeting (also see gene-targeted), 217, 232, 234, 245–247, 250, 253, 255–256, 275–276, 281 therapy (also see immunotherapy), 318 transfer, 309–314, 316, 318–320 translation, 190, 281 GeneChip Array, 133, 182–184, 219, 232, 234 gene-targeted (also see RMDI and RMCE), 217, 219, 232, 234, 245–247, 250, 253, 255–256 genome, 94, 123–125, 127, 129, 133, 138, 143–148, 150–152, 157, 161–162, 181, 276, 300, 309– 311, 316 candidate region, 123–124, 127, 143, 337 fine mapping, 138, 337–338 gene-targeting (see gene-targeted and gene targeting) scanning (also see RLGS), 124, 127, 129, 132–134, 177, 181, 183–185, 337–339 genotype/genotyping, 15–16, 331–332, 338 granulocyte-macrophage colonystimulating factor (GM-CSF), 300, 307–309, 319 green fluorescent protein (GFP), 220, 263, 269 growth factor, 33, 115, 118, 218, 225, 228, 235, 260, 263–264, 269,
Index
347 274–276, 302, 318
H hairy/Enhancer of split 1 (Hes1), 275 hamster, 114, 217, 220–224 Hedgehog signaling, 275 hematopoietic/hematopoiesis, 276, 301 H&E (hematoxylin and eosin) staining, 19–20, 44, 47, 51, 58, 60–61, 67, 69–70, 73, 75–76, 78–79, 82, 89, 94–95, 178, 187 hepatocyte growth factor, 115 HER-2/neu, 218, 302–303, 305 hereditary nonpolyposis colorectal carcinoma (HNPCC), 70 heterozygous/heterozygosity, 144, 151, 232, 234, 332, 337 HLA, 301, 306, 308, 320 hlxb9, 275 hnf1α, 275 HPLC (high performance liquid chromatography), 124–126, 128, 134, 152, 176–177 H-ras, 305 histology, 35–36, 41, 57–59, 70, 74, 277, 283, 287 homozygous deletion, 143–144, 150–152 hormone, 75, 115, 225, 275, 277–279, 281 hTERT, 117, 119, 200 human papillomavirus (HPV) (see viral/virus) hydrocortisone, 115 I identical by decent (IBD), 336–337 inducible transgene, 226, 231, 234 informative, 140, 331–332 inherited, 70, 332-333, 336 in situ hybridization, 67–68, 70, 76, 79–83, 89–90, 279–281, 283, 288–289 insulin, 33, 75, 115, 117, 277–281 insulin-like growth factor-1, 115 interleukin (IL), 115, 304, 319
image analysis, 22, 24, 28, 32, 48, 69, 90, 92–94, 138, 183–185, 194, 204, 269, 273 immortalization, 113, 117, 119, 199–200, 259–260 immunization, 299–301, 306–307 immunodeficient, 103, 217–220 immunohistochemical (IHC) labeling/ immunohistochemistry, 9–10, 16, 20, 42, 45, 59–60, 67–68, 70, 72–77, 82–83, 89, 94, 97–98, 123, 124–126, 128, 130, 209, 264, 267–268, 277, 279–281, 286–287, 291–292 immunoscreening, 207, 212 immunotherapy, 230, 300, 303–304 imprinting, 124 interleukin-6 (IL-6), 319 IPMN (intraductal papillary mucinous neoplasm), 2–6, 9–10, 72 intraepithelial, 1–2, 4–5, 68, 91, 260 invasive, 2–7, 10, 68–69, 74, 91, 200, 221, 229, 231, 260, 274 isl-1, 275, 277 islet cell, 68, 114–117, 221, 228, 231, 259–260, 263, 274–275, 277–281 K K-ras, 68, 200, 234, 303, 305 Ki-67, 74–75 kinetic enrichment, 126, 143, 146–147, 149, 158 knockout mice, 218, 227–229, 232, 234, 245, 249–251, 253 conditional, 234–235, 245, 249–251 L lacZ, 317 laminin-1/nidogen matrix, 116, 275 LCM (laser capture microdissection) (also see microdissection), 7–9,
348 21, 23, 25, 28–29, 38–40, 45, 48–52, 54, 58–60, 62–63 linkage/linkage analysis, 144, 329–340 lipase, 70, 72, 74, 277 LOD score, 333–335, 339 long terminal repeat (LTR), 311 M macrophage, 222, 300–301, 305, 308 magnetic-bead separation, 34, 43 mass spectrometry, 189–190, 193, 196 Matrigel, 104, 109, 114, 116 medullary carcinoma, 68–70 Mendelian disease, 329, 333 mesenchymal, 219, 260, 268, 275–276 neoplasm, 68, 75 metaplasia, 2, 228–229, 231, 234, 259–260, 274 metastasis/metastatic, 110, 218–220, 222–223, 229, 273, 306, 312 methylation (also see DNA methylation), 123–134, 144, 221 MHC, 302–303, 305, 319 microarray DNA 144, 189 oligonucelotide, 129, 131–133, 175–187 methylation-specific oligonucelotide (MSO), 124–126, 129, 131–133 tissue (also see tissue), 89–100 microdissection (also see fine-needle microdissection, LCM), 7–10, 15–16, 18–22, 28–33, 35–38, 40, 42–51, 55, 59–63, 126, 138, 151, 178–179, 187, 261 microenvironment, 304 microinjection, 281, 284, 290–291, 293, 309, microsatellite, 68, 70, 137–138, 145, 332, 338 microscope/microscopic, 3–4, 8–10, 20–22, 28, 30, 33–34, 37–38, 47–51, 57–59, 61, 70, 72, 79–81, 83, 89, 93, 114, 263,
Index 283, 285, 287, 288, 290–292 mismatch repair gene, 70 mitogen, 115 MLH1, 70 molecular beacon, 138–140 marker, 279 profiling/analysis, 7, 16–63, 137–138, 161, 274, 276, 281 Moloney murine leukemia virus (Mo-MuLV) (see viral/virus) morpholinos, 277, 281–282, 284, 292–293 mouse/mice, 79, 103–111, 184, 220–224, 229–235, 245, 247, 256, 259–265, 267, 278, 282, 284, 291, 308 chimera (see chimera/chimeric) gene-targeting (see gene-targeted) model (also see animal model), 229, 231, 234–235, 260 transgenic (see transgene/transgenic) M-phase specific phosphor-histone, 281 MSH2, 70 mucin-1 (MUC-1), 115, 119, 222, 225, 227, 230, 231, 235, 303, 306 mucinous cystic neoplasm (MCN), 2–3, 5–10, 72, 74, 91 multiplicity of infection (MOI), 264 multi-tissue tumor block (MTTB), 89 mung bean nuclease (see DNase) mutagenesis, 276–277, 279, 282 mutation, 3, 74, 123–124, 143, 158, 200, 208–209, 217–219, 221, 224, 227, 229, 233, 245, 247, 274, 276–277, 279–281, 305, 320 N N-cadherin, 275 neoplasia, 1–4, 68, 91, 217, 220–221, 224, 228, 230, 234, 260 neoplasm, 2–10, 91, 67–68, 72, 74–75, 199–200, 203, 217–221, 225, 228–230, 260, 274
Index NeuroD/Beta2, 275, 279 neuron-specific enolase (NSE), 72, 75 ngn3, 275 nitrocellulose, 209–212, 215 nkx2.2, 275 nkx6.1, 275 Notch signaling, 275 N-ras, 305 nuclease endonuclease, 144, 146–148, 151–152, 156–157 mung bean nuclease, 147, 149, 152, 156–157 nuclease P1, 131 ribonuclease H, 176 S1 nuclease, 157
O oligonucleotide, 80, 123–124, 126, 129, 131, 138, 161, 175–176, 185, 200, 222, 250, 281, 284 oncogene, 117, 143, 217, 225, 228, 231–232, 234–235, 274, 303, 305, 310, 320 organogenesis (see pancreatic organogenesis) ortholog, 277, 279, 281
P p16, 123–124, 218, 232, 234, 305 p48, 275 p53, 200, 208, 218, 221, 225, 232, 234, 303, 305 pancreas/pancreatic adenoma (see adenoma) cancer, 1–7, 259–260, 273–274, 282, 329–331, 334, 336, 340 classification, 1–7, 10, 331 nomenclature, 1–2, 4, 10 carcinoma, 1–7, 42, 68–70, 74, 127, 190–191, 199–201, 229–231, 235
349 development, 259–260, 273–277, 279–282 ductal adenocarcinoma (also see adenocarcinoma), 1–3, 68–70, 72, 74, 221, 224, 231, 259–260, 273–274 ductal epithelial cells (also see ductal epithelial cells), 4, 113–114, 116–119, 225, 229, 231, 260 isolate, 265, 267–268 juice, 189, 191, 193, 195, 199–204 organogenesis, 274, 276 pancreas-specific transcription factor-1 (PTF-1), 275–276 pancreatitis, 113, 175–176, 178, 186–187, 199–201, 260 pancreatoblastoma, 72–73 PanIN (pancreatic intraepithelial neoplasia), 1–5, 9–10, 68, 260 paraffin (also see tissue embedding), 7, 9–10, 35–36, 41, 44–46, 55, 57–60, 75–79, 81, 89–90, 92– 93, 95–99, 125–126, 128, 130, 134, 280, 287 pax 4, 275 pax 6, 275 PCNA (proliferating cell nuclear antigen), 281 PCR, 7–9, 38, 50–52, 54, 60, 79–80, 82–83, 137–141, 143, 147–149, 153–158, 162, 164–166, 169–170, 172–173, 179, 181, 183, 203, 246–247, 249–250 methylation-specific (MSP), 124–134 quantitative (see quantitative PCR) PDX-1, 225, 231, 267–268, 275, 277, 279 penetrance, 333–334, 336, 338 peptide mapping, 189–196 phenol-chloroform extraction, 38–40, 53–56, 62, 76, 80, 153, 165, 176–177, 179, 180, 187, 214, 283, 289
350 phenol-enhanced reassociation technique (PERT), 145 phenotype/phenotyping, 15–16, 20, 33, 59, 70, 116–117, 217–221, 228–229, 234–235, 259, 277, 330–332, 338–339 PicoGreen, 140 platelet-derived growth factor, 115 polymorphism/polymorphic, 137–138, 151–152, 332 precursor, 1–3, 7, 9–10, 72, 82, 113, 229, 259–260, 273, 276, 281, 301, 305 primary tumor, 104, 107, 151, 221, 223 proliferation/proliferative/proliferate, 3, 74–75, 103, 114–115, 143, 199, 221, 274, 276, 281 protease, 60, 77, 193, 195, 199, 203 inhibitor 37, 44, 60–61, 63, 191, 199–200, 203 proteinase K, 39, 52, 57–58, 76, 81, 83 protein analysis, 19, 50, 63, 68, 70, 73, 189–196, 203 identification, 189–196 isolation, 16, 19–20, 28, 35–37, 39, 41, 44, 47, 50, 57, 60–61, 190–196, 289 stability 35, 57, 59–60, 68, 77, 190 proteomics, 189–191 PTEN, 144, 150 p-value, 330, 333, 335, 339
Q quality control, 15–63, 98, 186, 331–332 quantitative PCR, 134, 138, 204
R rat, 114, 116, 217, 219–224, 260–262 Rb, 123–124, 228
Index receiver-operating characteristic (ROC), 140 recombinase-mediated cassette exchange (RMCE), 245, 250, 254–255 recombinase-mediated DNA insertion (RMDI), 245, 253–254 representation differential analysis (RDA), 125–127, 129, 132–134, 143–158 restriction enzyme digestion (also see DNA digestion), 126–127, 129, 131–132, 162, 250, 289 restriction landmark genome scanning (RLGS), 125, 127, 133 retrovirus (see virus) riboprobe, 79–81, 83, 289–290 risk, 18, 28, 44, 47, 60, 211, 217, 219, 228, 260, 267, 273, 329–331, 333, 335, 338–340 RNA analysis, 50, 58, 59, 63 cRNA, 133, 177, 180–181, 183 isolation, 9, 19–21, 35–36, 40–41, 44, 47, 54–56, 60–61, 63, 133, 166, 176, 178–179, 209–210, 214 mRNA, 89, 162, 164, 166, 178, 185, 189-190, 200, 209–210, 215 polymerase, 179–180, 283, 289 precipitation, 289 probe (also see riboprobe), 79–80, 288 quality, 57, 59–60, 63, 83, 166, 172–173, 214 quantity, 60–61, 63, 81, 173, 180, 189–190, 210 sense/antisense, 288, 292 tRNA, 284 total RNA, 41, 54, 56, 63, 133, 166, 172–173, 176, 178–179, 210, 215 RNase, 18, 40, 45, 52, 54–55, 60, 76, 79, 81–82, 177, 180, 199, 203, 283, 289, 293
Index inhibitor, 37, 41, 54, 56, 60, 76, 81, 180, 199 RPA, 115
S secretin, 115–116, 199, 201, 203 segregation analysis, 333–334, 338 selenium, 115 senescence, 117 SEREX (serological analysis of recombinant tumor cDNA expression libraries), 207–210, 213, 215 serial analysis of gene expression (SAGE), 161–173 serous cystadenoma, 72, 234 serum, 33, 40, 77, 106, 117–118, 128, 130, 153, 167, 177, 189, 191, 207–215, 261, 269, 273, 282–284, 290 sequencing, 132, 134, 139, 161–173, 209, 214, 276 sequential probability ratio test (SPRT), 140 single nucleotide polymorphism (SNP), 137–141 site-specific recombinase, 245–247, 250, 254 SMAD4 (also see DPC4), 9–10, 234–235 solid-pseudopapillary neoplasm, 74 somatostatin, 72, 115, 277–279 sonic hedgehog (shh) 275, 277 statistic, 330, 333, 335–337, 339 stem cell, 116, 231 STK11, 329–330 stroma/stromal, 3, 5, 8–9, 51, 59, 62, 69–70, 72, 80, 103, 111, 175–176, 178, 186-187, 220, 223, 229 desmoplasia, 70 subtractive hybridization, 127, 143, 147, 149–151, 155, 161 sucrose infusion, 35, 43
351 surface-enhanced laser desorption ionization (SELDI), 189–191, 193–196 SV40, 117, 224 T tag/ditag, 162–173 tagging enzyme, 162, 168 T-cells, 103, 208, 219, 222, 300–308, 312 telomerase, 117, 199–204 telomerase amplification protocol (TRAP), 200–201, 203 TGFα, 224–225, 228, 234–235, 260 TGFβ, 275 tissue, 274–277, 280 antigen, 59, 77, 286 collection, 15, 19, 32, 47–48, 50, 92– 95, 178 culture, 113–119, 259–269 disaggregation, 33, 42 embedding (also see paraffin), 7, 9, 36, 44, 75–77, 79, 89, 125 fixation, 7, 34–36, 43, 77, 79, 109 freezing/frozen, 9, 20–21, 32, 36, 41, 44, 77, 125, 210 microarray (TMA)/array, 10, 82, 89–100 permeability, 286, 288 processing, 15–63 resolution, 280, 287 scrapping, 7, 9, 33, 42 sectioning (also see frozen section), 7, 10, 36, 44, 89, 93, 96–97, 187 -specific, 226–228, 230, 234, 245 staining (also see staining, immunohistochemistry), 7, 9–10, 19–20, 36–37, 44–46, 93, 187 storage, 15, 60, 99 transdifferentiation, 260, 274 transferring, 115 transgene/transgenic (also see inducible transgene), 217–219, 224, 228–231, 234–235, 256
352 TRIzol, 38, 133, 164, 166, 173, 176, 178–179 trypsin, 70, 74, 104, 106, 115–118, 193, 195–196, 264, 279 inhibitor 115, 118, 261–262, 269 tumor antigen, 207–208, 302–308, 312 development, 10, 68–75, 217–236 genome, 143–144, 150–152 immunology, 207–215, 299–320 marker, 190, 259–260, 273–274 samples, 151–152, 178–179 tumorigenesis, 114, 274 tumor necrosis factor α (TNF-α), 319 tumor-suppressor gene, 68,123–124, 126, 143–144, 150, 217, 232–235, 274, 305 two-dimensional polyacrylamide gel electrophoreses (2D-PAGE), 189–191, 193–196 V vaccine, 207, 299–300, 304–310, 312, 318–320 vector (also see viral/virus), 119, 132, 162, 210, 214, 227, 232, 246–247, 250, 253, 256, 289, 309–313, 315–320 VEGF, 222–223 ventral, 265, 267, 275 VHL, 123–124 vimentin, 115–116 viral/virus adenovirus, 261, 264, 267–268 antiviral assay, 319 DNA, 311 Epstein-Barr, 70
Index Herpes simplex virus (HSV), 232, 247, 250, 254 human papillomavirus (HPV), 117– 118 infection, 310 Moloney murine leukemia virus (Mo-MuLV), 311 oncogene, 117, 225 particle, 311, 314–315 producer, 312–313, 315–316 protein, 311 recombinant, 314 retrovirus, 119, 309–318 RNA, 70, 311 vector, 119, 309–313, 315–320
W whole cell vaccine, 299–300 whole-mount, 280–282, 286–287, 292 workflow management, 15, 21 X X-chromosome inactivation, 124 xenogeneic, 217, 219 xenograft, 103–111, 151–152 Y yeast artificial chromosome (YAC), 144 Z zebrafish, 273–293 zymogen, 263, 277