Pancreatic Cancer: Methods and Protocols (Methods in Molecular Medicine)

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 ...

Author: Gloria H. Su

129 downloads 1084 Views 10MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form

DOWNLOAD PDF

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,


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).


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


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):


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.


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


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-


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.


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


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. 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.


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.


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.


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.


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


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.


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.


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).


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-


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.


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.


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,


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 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


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-


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).


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.


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.


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.


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.


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.


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


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.


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


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”).


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-


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


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).


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


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 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


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.


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.


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


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).


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.


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).


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


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.


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.


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-


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


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.


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


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 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


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?


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


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


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.


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 (

Pancreatic Cancer: Methods and Protocols (Methods in Molecular Medicine)

Renal Cancer: Methods and Protocols (Methods in Molecular Medicine)

Ovarian Cancer: Methods and Protocols (Methods in Molecular Medicine)

Colorectal Cancer: Methods and Protocols (Methods in Molecular Medicine)

Cancer Cell Culture: Methods and Protocols (Methods in Molecular Medicine)

Renal Cancer: Methods and Protocols (Methods in Molecular Medicine)

Colorectal Cancer: Methods and Protocols (Methods in Molecular Medicine)

Ovarian Cancer: Methods and Protocols (Methods in Molecular Medicine)

Prostate Cancer Methods and Protocols (Methods in Molecular Medicine)

Molecular Cardiology: Methods and Protocols (Methods in Molecular Medicine)

Hemoglobin Disorders: Molecular Methods and Protocols (Methods in Molecular Medicine)

Cancer Cytogenetics: Methods and Protocols (Methods in Molecular Biology)

Cancer Stem Cells: Methods and Protocols (Methods in Molecular Biology)

Cancer Nanotechnology: Methods and Protocols (Methods in Molecular Biology)

Celiac Disease: Methods and Protocols (Methods in Molecular Medicine)

Atherosclerosis: Experimental Methods and Protocols (Methods in Molecular Medicine)

Cystic Fibrosis: Methods and Protocols (Methods in Molecular Medicine)

Single Cell Diagnostics: Methods and Protocols (Methods in Molecular Medicine)

Diabetes Mellitus: Methods and Protocols (Methods in Molecular Medicine)

Adoptive Immunotherapy: Methods and Protocols (Methods in Molecular Medicine)

Antiviral Methods and Protocols (Methods in Molecular Medicine)

Aging Methods and Protocols (Methods in Molecular Medicine)

Wound Healing: Methods and Protocols (Methods in Molecular Medicine)

Meningococcal Vaccines: Methods and Protocols (Methods in Molecular Medicine)

Hypertension: Methods and Protocols (Methods in Molecular Medicine)

Opioid Research: Methods and Protocols (Methods in Molecular Medicine)

Muscular Dystrophy: Methods and Protocols (Methods in Molecular Medicine)

Hepatocellular Carcinoma: Methods and Protocols (Methods in Molecular Medicine)

DNA Vaccines - Methods And Protocols (Methods in Molecular Medicine)

Lectin Methods and Protocols (Methods in Molecular Medicine)

Celiac Disease: Methods and Protocols (Methods in Molecular Medicine)

Pancreatic Cancer: Methods and Protocols (Methods in Molecular Medicine)

Renal Cancer: Methods and Protocols (Methods in Molecular Medicine)

Ovarian Cancer: Methods and Protocols (Methods in Molecular Medicine)

Colorectal Cancer: Methods and Protocols (Methods in Molecular Medicine)

Cancer Cell Culture: Methods and Protocols (Methods in Molecular Medicine)

Renal Cancer: Methods and Protocols (Methods in Molecular Medicine)

Colorectal Cancer: Methods and Protocols (Methods in Molecular Medicine)

Ovarian Cancer: Methods and Protocols (Methods in Molecular Medicine)

Prostate Cancer Methods and Protocols (Methods in Molecular Medicine)

Molecular Cardiology: Methods and Protocols (Methods in Molecular Medicine)

Hemoglobin Disorders: Molecular Methods and Protocols (Methods in Molecular Medicine)

Cancer Cytogenetics: Methods and Protocols (Methods in Molecular Biology)

Cancer Stem Cells: Methods and Protocols (Methods in Molecular Biology)

Cancer Nanotechnology: Methods and Protocols (Methods in Molecular Biology)

Celiac Disease: Methods and Protocols (Methods in Molecular Medicine)

Atherosclerosis: Experimental Methods and Protocols (Methods in Molecular Medicine)

Cystic Fibrosis: Methods and Protocols (Methods in Molecular Medicine)

Single Cell Diagnostics: Methods and Protocols (Methods in Molecular Medicine)

Diabetes Mellitus: Methods and Protocols (Methods in Molecular Medicine)

Adoptive Immunotherapy: Methods and Protocols (Methods in Molecular Medicine)

Antiviral Methods and Protocols (Methods in Molecular Medicine)

Aging Methods and Protocols (Methods in Molecular Medicine)

Wound Healing: Methods and Protocols (Methods in Molecular Medicine)

Meningococcal Vaccines: Methods and Protocols (Methods in Molecular Medicine)

Hypertension: Methods and Protocols (Methods in Molecular Medicine)

Opioid Research: Methods and Protocols (Methods in Molecular Medicine)

Muscular Dystrophy: Methods and Protocols (Methods in Molecular Medicine)

Hepatocellular Carcinoma: Methods and Protocols (Methods in Molecular Medicine)

DNA Vaccines - Methods And Protocols (Methods in Molecular Medicine)

Lectin Methods and Protocols (Methods in Molecular Medicine)

Celiac Disease: Methods and Protocols (Methods in Molecular Medicine)

Recommend Documents