ME T H O D S
IN
MO L E C U L A R BI O L O G Y
Series Editor John M. Walker School of Life Sciences University of Hertfordshire Hatfield, Hertfordshire, AL10 9AB, UK
For other titles published in this series, go to www.springer.com/series/7651
TM
Mouse Models for Drug Discovery Methods and Protocols
Edited by
Gabriele Proetzel and Michael V. Wiles The Jackson Laboratory, Bar Harbor, ME, USA
Editors Gabriele Proetzel The Jackson Laboratory 610 Main Street Bar Harbor ME 04609 USA
[email protected]
Michael V. Wiles The Jackson Laboratory 600 Main Street Bar Harbor ME 04609 USA
[email protected]
ISSN 1064-3745 e-ISSN 1940-6029 ISBN 978-1-60761-057-1 e-ISBN 978-1-60761-058-8 DOI 10.1007/978-1-60761-058-8 Library of Congress Control Number: 2009939644 © Humana Press, a part of Springer Science+Business Media, LLC 2010 All rights reserved. This work may not be translated or copied in whole or in part without the written permission of the publisher (Humana Press, c/o Springer Science+Business Media, LLC, 233 Spring Street, New York, NY 10013, USA), except for brief excerpts in connection with reviews or scholarly analysis. Use in connection with any form of information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed is forbidden. The use in this publication of trade names, trademarks, service marks, and similar terms, even if they are not identified as such, is not to be taken as an expression of opinion as to whether or not they are subject to proprietary rights. While the advice and information in this book are believed to be true and accurate at the date of going to press, neither the authors nor the editors nor the publisher can accept any legal responsibility for any errors or omissions that may be made. The publisher makes no warranty, express or implied, with respect to the material contained herein. Printed on acid-free paper springer.com
Preface The drug discovery process has shifted from essentially a trial and error approach toward unraveling disease and underlying molecular mechanisms with the goal to specifically target pathways and molecules. A key for understanding and exploring disease mechanisms is the availability of good models, preferably in vivo models as these more closely reflect the complexity of life. Each organism is a highly complex integrated system and is considerably more than the sum of its parts being the net result of an evolving genome and its interactions with the environment. This volume, Mouse Models for Drug Discovery: Methods and Protocols, attempts to illustrate The Mouse as an exceptionally versatile and sophisticated platform which can meet this challenge. With the development of inbred strains which are genetically invariant within a strain, it has become possible to use genetically defined animals and highly reproducible systems. This has allowed The Mouse to rise from being a pest, to a cute collectable pet, to an advanced and well-established tool for genetic and molecular research, becoming the premium instrument for drug discovery, validation, preclinical, and toxicological studies. The reasons for the mouse’s rise in prominence are many. In brief, mice are small, require relatively little space, have simple nutritional needs, a short generation time, and few special needs to reproduce. However, it is in the last 20 years with the advent of genetic engineering and the ease of manipulating the mouse’s genome that has made the mouse the most versatile mammalian experimental system. Further, the sequencing of the human and mouse genomes has clearly demonstrated that mouse and humans have direct gene and functional homologies for more than 90 % of their genes. With the KOMP and other international programs, all mouse genes will be available as null alleles (i.e., knockouts) by approximately 2015. Development and distribution of these resources is aided by advances in Assisted Reproductive Technologies. With all of this in mind, it can now be truly stated that The Mouse has become a respected, indispensable tool in biomedical research and is the most commonly used animal research model. With thousands of mouse models available covering practically all disease areas, it is beyond this volume to cover the whole field of mouse applications in the drug discovery process. In this volume, we have selected chapters which cover general background as well as a few specific disease topics with the idea of introducing those less familiar with mice as experimental model platforms. The chapters by Festing and Wiles cover general aspects of experimental design, inbred vs. outbred mice, and how to manage the risks working with live animals. The chapter by McNeish highlights a pharma approach as how to use genetically engineered mouse models for target identification and validation. Koentgen et al. has given a general overview of the many approaches used to genetically engineer mice. Rando et al. show the power of combining novel imaging tools with genetically engineered mice for drug discovery. Representatives of the young and rapidly developing field of humanized mice are provided in the chapters by Roopenian and Shultz, highlighting also the possibility of engraftment of human tissue into mouse, for example, in regenerative biology and stem cell research.
v
vi
Preface
As examples for the wide field of specific disease areas and mouse models, we have included type 1 and 2 diabetes (Serreze and Baribault), cardiovascular disease (Howles), arthritis (Tak), skin disorders (Sundberg), cancer (Talmadge, Surguladze, and Li), the use of behavioral models for depression and anxiety (Kalueff), neurodegenerative diseases (Janus), neuromuscular diseases (Burgess), and infectious diseases (Medina). We hope that this volume will stimulate those not familiar with the power of the mouse and its potential for the drug discovery process, and further, that it will encourage the development of new models as well as new ways in utilizing existing models. We also hope to promote the development of more standardized models and assays such that results can be more easily compared and reproduced. We like to thank all the contributors, their discussions, and their patience in making this an important volume on mouse models and drug discovery. Gabriele Proetzel and Michael V. Wiles
Contents Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
v
Contributors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
ix
1.
Improving Toxicity Screening and Drug Development by Using Genetically Defined Strains . . . . . . . . . . . . . . . . . . . . . . . Michael F.W. Festing
1
2.
The Sophisticated Mouse: Protecting a Precious Reagent . . . . . . . . . . . . . Michael V. Wiles and Rob A. Taft
23
3.
Genetically Engineered Mouse Models in Drug Discovery Research . . . . . . . Rosalba Sacca, Sandra J. Engle, Wenning Qin, Jeffrey L. Stock, and John D. McNeish
37
4.
Engineering the Mouse Genome to Model Human Disease for Drug Discovery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Frank Koentgen, Gabriele Suess, and Dieter Naf
55
Profiling of Drug Action Using Reporter Mice and Molecular Imaging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Gianpaolo Rando, Andrea Biserni, Paolo Ciana, and Adriana Maggi
79
Human FcRn Transgenic Mice for Pharmacokinetic Evaluation of Therapeutic Antibodies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Derry C. Roopenian, Gregory J. Christianson, and Thomas J. Sproule
93
5.
6.
7.
Development of Novel Major Histocompatibility Complex Class I and Class II-Deficient NOD-SCID IL2R Gamma Chain Knockout Mice for Modeling Human Xenogeneic Graft-Versus-Host Disease . . . . . . . . . . . . 105 Steve Pino, Michael A. Brehm, Laurence Covassin-Barberis, Marie King, Bruce Gott, Thomas H. Chase, Jennifer Wagner, Lisa Burzenski, Oded Foreman, Dale L. Greiner, and Leonard D. Shultz
8.
Bridging Mice to Men: Using HLA Transgenic Mice to Enhance the Future Prediction and Prevention of Autoimmune Type 1 Diabetes in Humans . 119 David V. Serreze, Marijke Niens, John Kulik, and Teresa P. DiLorenzo
9.
Mouse Models of Type II Diabetes Mellitus in Drug Discovery . . . . . . . . . 135 Helene Baribault
10.
Cholesterol Absorption and Metabolism . . . . . . . . . . . . . . . . . . . . . 157 Philip N. Howles
11.
Collagen-Induced Arthritis in Mice . . . . . . . . . . . . . . . . . . . . . . . . 181 Lisette Bevaart, Margriet J. Vervoordeldonk, and Paul P. Tak
12.
Skin Diseases in Laboratory Mice: Approaches to Drug Target Identification and Efficacy Screening . . . . . . . . . . . . . . . . . . . . . . . 193 John P. Sundberg, Kathleen A. Silva, Caroline McPhee, and Lloyd E. King
vii
viii
Contents
13.
Models of Metastasis in Drug Discovery . . . . . . . . . . . . . . . . . . . . . 215 James E. Talmadge
14.
Methods for Evaluating Effects of an Irinotecan + 5Fluorouracil/Leucovorin (IFL) Regimen in an Orthotopic Metastatic Colorectal Cancer Model Utilizing In Vivo Bioluminescence Imaging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 235 David Surguladze, Philipp Steiner, Marie Prewett, and James R. Tonra
15.
CML Mouse Model in Translational Research . . . . . . . . . . . . . . . . . . 253 Cong Peng and Shaoguang Li
16.
Mouse Models for Studying Depression-Like States and Antidepressant Drugs . . 267 Carisa L. Bergner, Amanda N. Smolinsky, Peter C. Hart, Brett D. Dufour, Rupert J. Egan, Justin L. LaPorte, and Allan V. Kalueff
17.
Virus-Delivered RNA Interference in Mouse Brain to Study Addiction-Related Behaviors . . . . . . . . . . . . . . . . . . . . . . . . . . . 283 Amy W. Lasek and Nourredine Azouaou
18.
Experimental Models of Anxiety for Drug Discovery and Brain Research . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 299 Peter C. Hart, Carisa L. Bergner, Amanda N. Smolinsky, Brett D. Dufour, Rupert J. Egan, Justin L. LaPorte, and Allan V. Kalueff
19.
Mouse Models of Neurodegenerative Diseases: Criteria and General Methodology 323 Christopher Janus and Hans Welzl
20.
Neuromuscular Disease Models and Analysis . . . . . . . . . . . . . . . . . . . 347 Robert W. Burgess, Gregory A. Cox, and Kevin L. Seburn
21.
Murine Model of Cutaneous Infection with Streptococcus pyogenes . . . . . . . . 395 Eva Medina
22.
Murine Model of Pneumococcal Pneumonia . . . . . . . . . . . . . . . . . . . 405 Eva Medina
23.
Murine Model of Polymicrobial Septic Peritonitis Using Cecal Ligation and Puncture (CLP) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 411 Eva Medina
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 417
Contributors NOURREDINE AZOUAOU • The Ernest Gallo Clinic and Research Center, University of California at San Francisco, Emeryville, CA, USA HELENE BARIBAULT • Department of Metabolic Disorders, Amgen, South San Francisco, CA, USA CARISA L. BERGNER • Department of Physiology and Biophysics, Georgetown University Medical School, Washington, DC, USA LISETTE BEVAART • Division of Clinical Immunology and Rheumatology, Academic Medical Center/University of Amsterdam, Amsterdam, The Netherlands; Arthrogen B.V., Amsterdam, The Netherlands ANDREA BISERNI • TOP s.r.l., Lodi, Italy MICHAEL A. BREHM • Department of Medicine, The University of Massachusetts Medical School, Worcester, MA, USA ROBERT W. BURGESS • The Jackson Laboratory, Bar Harbor, ME, USA LISA BURZENSKI • The Jackson Laboratory, Bar Harbor, ME, USA THOMAS H. CHASE • The Jackson Laboratory, Bar Harbor, ME, USA GREGORY J. CHRISTIANSON • The Jackson Laboratory, Bar Harbor, ME, USA PAOLO CIANA • Center of Excellence on Neurodegenerative Diseases and Department of Pharmacological Sciences University of Milan, Milan, Italy LAURENCE COVASSIN-BARBERIS • Department of Medicine, The University of Massachusetts Medical School, Worcester, MA, USA GREGORY A. COX • The Jackson Laboratory, Bar Harbor, ME, USA TERESA P. DILORENZO • Department of Microbiology & Immunology and Department of Medicine, Division of Endocrinology, Albert Einstein College of Medicine, Bronx, NY, USA BRETT D. DUFOUR • Department of Animal Sciences, Purdue University, West Lafayette, IN, USA RUPERT J. EGAN • Department of Physiology and Biophysics, Georgetown University Medical School, Washington, DC, USA SANDRA J. ENGLE • Genetically Modified Models Center of Emphasis, Pfizer Global Research and Development, Pfizer Inc., Groton, CT, USA MICHAEL F.W. FESTING • Understanding Animal Research, London, UK ODED FOREMAN • The Jackson Laboratory, Bar Harbor, ME, USA
ix
x
Contributors
DALE L. GREINER • Department of Medicine, The University of Massachusetts Medical School, Worcester, MA, USA PETER C. HART • Department of Physiology and Biophysics, Georgetown University Medical School, Washington, DC, USA PHILIP N. HOWLES • Department of Pathology and Laboratory Medicine, Center for Lipid and Arteriosclerosis Studies, Genome Research Institute, University of Cincinnati College of Medicine, Cincinnati, OH, USA CHRISTOPHER JANUS • Department of Neuroscience, Mayo Clinic Jacksonville, Jacksonville, FL, USA ALLAN V. KALUEFF • Department of Physiology and Biophysics, Stress Physiology and Research Center (SPaRC), Georgetown University Medical School, Washington, DC, USA; Department of Pharmacology, Tulane University Medical Center, New Orleans, LA, USA LLOYD E. KING, JR. • The Skin Disease Research Center, Department of Medicine, Division of Dermatology, Vanderbilt Medical Center, Nashville, TN, USA MARIE KING • Department of Medicine, The University of Massachusetts Medical School, Worcester, MA, USA FRANK KOENTGEN • Ozgene Pty. Ltd., Bentley DC, Western Australia, Australia JOHN KULIK • The Jackson Laboratory, Bar Harbor, ME, USA JUSTIN L. LAPORTE • Department of Physiology and Biophysics, Stress Physiology and Research Center (SPaRC), Georgetown University Medical School, Washington, DC, USA AMY W. LASEK • The Ernest Gallo Clinic and Research Center, University of California at San Francisco, Emeryville, CA, USA SHAOGUANG LI • Division of Hematology & Oncology, Department of Medicine, University of Massachusetts Medical School, Worcester, MA, USA ADRIANA MAGGI • Center of Excellence on Neurodegenerative Diseases and Department of Pharmacological Sciences Universit`a degli Studi di Milano, via Balzaretti 9, 20133 Milan, Italy JOHN D. MCNEISH • Regenerative Medicine Unit, Pfizer Global Research and Development, Pfizer Inc., Cambridge, MA, USA CAROLINE MCPHEE • The Jackson Laboratory, Bar Harbor, ME, USA EVA MEDINA • Infection Immunology Research Group, Department of Microbial Pathogenesis, Helmholtz Centre for Infection Research, Braunschweig, Germany DIETER NAF • Ozgene Pty. Ltd., Bentley DC, Western Australia, Australia MARIJKE NIENS • The Jackson Laboratory, Bar Harbor, ME, USA CONG PENG – Division of Hematology & Oncology, Department of Medicine, University of Massachusetts Medical School, Worcester, MA, USA
Contributors
xi
STEVE PINO • Department of Medicine, The University of Massachusetts Medical School, Worcester, MA, USA MARIE PREWETT • Imclone Systems, a wholly-owned subsidiary of Eli Lilly & Company, New York, NY, USA WENNING QIN • Genetically Modified Models Center of Emphasis, Pfizer Global Research and Development, Pfizer Inc., Groton, CT, USA GIANPAOLO RANDO • Center of Excellence on Neurodegenerative Diseases and Department of Pharmacological Sciences University of Milan, Milan, Italy DERRY C. ROOPENIAN • The Jackson Laboratory, Bar Harbor, ME, USA ROSALBA SACCA • Genetically Modified Models Center of Emphasis, Pfizer Global Research and Development, Pfizer Inc., Groton, CT, USA KEVIN L. SEBURN • The Jackson Laboratory, Bar Harbor, ME, USA DAVID V. SERREZE • The Jackson Laboratory, Bar Harbor, ME, USA LEONARD D. SHULTZ • The Jackson Laboratory, Bar Harbor, ME, USA KATHLEEN A. SILVA • The Jackson Laboratory, Bar Harbor, ME, USA AMANDA N. SMOLINSKY • Department of Physiology and Biophysics, Georgetown University Medical School, Washington, DC, USA THOMAS J. SPROULE • The Jackson Laboratory, Bar Harbor, ME, USA PHILIPP STEINER • Imclone Systems, a wholly-owned subsidiary of Eli Lilly & Company, New York, NY, USA JEFFREY L. STOCK • Pfizer Global Research and Development, Pfizer Inc., Genetically Modified Models Center of Emphasis, Groton, CT, USA GABRIELE SUESS • Ozgene Pty. Ltd., Bentley DC, Western Australia, Australia JOHN P. SUNDBERG • The Jackson Laboratory, Bar Harbor, ME, USA; The Skin Disease Research Center, Department of Medicine, Division of Dermatology, Vanderbilt Medical Center, Nashville, TN, USA DAVID SURGULADZE • Imclone Systems, a wholly-owned subsidiary of Eli Lilly & Company, New York, NY, USA ROB A. TAFT • The Jackson Laboratory, Bar Harbor, ME, USA PAUL-PETER TAK • Division of Clinical Immunology and Rheumatology, Academic Medical Center/University of Amsterdam, Amsterdam, The Netherlands; Arthrogen B.V., Amsterdam, The Netherlands JAMES E. TALMADGE • University of Nebraska Medical Center, Omaha, NE, USA JAMES R. TONRA • Imclone Systems, a wholly-owned subsidiary of Eli Lilly & Company, New York, NY, USA
xii
Contributors
MARGRIET J. VERVOORDELDONK • Division of Clinical Immunology and Rheumatology, Academic Medical Center/University of Amsterdam, Amsterdam, The Netherlands; Arthrogen B.V., Amsterdam, The Netherlands JENNIFER WAGNER • The Jackson Laboratory, Bar Harbor, ME, USA HANS WELZL • Division of Neuroanatomy and Behavior, Institute of Anatomy, University of Z¨urich, Z¨urich, Switzerland MICHAEL V. WILES • The Jackson Laboratory, Bar Harbor, ME, USA
Chapter 1 Improving Toxicity Screening and Drug Development by Using Genetically Defined Strains Michael F.W. Festing Abstract According to the US Food and Drugs Administration (Food and Drug Administration (2004) Challenge and opportunity on the critical path to new medical products.) “The inability to better assess and predict product safety leads to failures during clinical development and, occasionally, after marketing”. This increases the cost of new drugs as clinical trials are even more expensive than pre-clinical testing. One relatively easy way of improving toxicity testing is to improve the design of animal experiments. A fundamental principle when designing an experiment is to control all variables except the one of interest: the treatment. Toxicologist and pharmacologists have widely ignored this principle by using genetically heterogeneous “outbred” rats and mice, increasing the chance of false-negative results. By using isogenic (inbred or F1 hybrid, see Note 1) rats and mice instead of outbred stocks the signal/noise ratio and the power of the experiments can be increased at little extra cost whilst using no more animals. Moreover, the power of the experiment can be further increased by using more than one strain, as this reduces the chance of selecting one which is resistant to the test chemical. This can also be done without increasing the total number of animals by using a factorial experimental design, e.g. if the ten outbred animals per treatment group in a 28-day toxicity test were replaced by two animals of each of five strains (still ten animals per treatment group) selected to be as genetically diverse as possible, this would increase the signal/noise ratio and power of the experiment. This would allow safety to be assessed using the most sensitive strain. Toxicologists should also consider making more use of the mouse instead of the rat. They are less costly to maintain, use less test substance, there are many inbred and genetically modified strains, and it is easier to identify gene loci controlling variation in response to xenobiotics in this species. We demonstrate here the advantage of using several inbred strains in two parallel studies of the haematological response to chloramphenicol at six dose levels with CD-1 outbred, or using four inbred strains of mice. Toxicity to the white blood cell lineage was easily detected using the inbred strains but not using the outbred stock, clearly showing the advantage of using the multi-inbred strain approach. Key words: Toxicity testing, pre-clinical development, inbred strain, drug development, factorial experimental designs, statistics, signal/noise ratio.
G. Proetzel, M.V. Wiles (eds.), Mouse Models for Drug Discovery, Methods in Molecular Biology 602, DOI 10.1007/978-1-60761-058-8 1, © Humana Press, a part of Springer Science+Business Media, LLC 2010
1
2
Festing
1. Introduction It is now 50 years since Russell and Burch, in their classical book on The Principles of Humane Experimental Technique (1), suggested that toxicologists should use small numbers of animals of several inbred strains rather than using outbred stocks in toxicological screening. The aim of this chapter is to explain how 1. by using genetically defined (inbred and F1 hybrid) laboratory mice (preferably) or rats, toxicologists and pharmacologists can immediately improve the power of their experiments and reduce false-negative results. 2. by using small numbers of animals of several inbred strains, without increasing the number of animals used, falsenegative results from the chance use of a resistant strain will be reduced. 3. by taking account of genetic variation, toxicologists can apply modern genetic research and develop a better understanding of toxicological mechanisms and the genes in humans and animals associated with susceptibility to toxic and pharmaceutically active compounds. 1.1. Need to Improve Toxicity Screening
There is an urgent need to improve methods of toxicity screening. According to the FDA 2004 “Critical path” white paper “The traditional tools used to assess product safety – animal toxicology and outcomes from human studies – have changed little over many decades and have largely not benefited from recent gains in scientific knowledge. The inability to better assess and predict product safety leads to failures during clinical development and, occasionally, after marketing” (2). According to one study (3) the attrition rate of new chemical entities (excluding “me-too” drugs) is about 96% including 27% rejected for toxicity and 46% rejected due to lack of efficacy. Clinical trials are considerably more expensive than pre-clinical testing. Therefore, if the number of misleading results could be minimised the cost of developing new drugs would be substantially reduced. The FDA “Critical path initiative” (4) and the European Union Innovative Medicines Initiative (IMI) both aim to improve methods for developing new drugs (5), so this is an opportune time to explain how the testing of potential new drugs could be improved and made more cost-effective by a relatively simple change in the type of animals which are used.
1.2. Use of Mice and Rats in Drug Development
When methods of toxicity testing were first developed most biochemical and physiological determinations required large samples of tissue. Traditionally, toxicologists have used the rat for this
Inbred Strains in Toxicity Testing
3
reason. However, the majority assays have now been miniaturised for work on smaller species such as Caenorhabditis elegans and Drosophila and the small size of the mouse is no longer a limitation, except in a few special situations. Many protocols are also available for studying mouse biology such as the 270 standard operating procedures for phenotyping the mouse developed by the EUMORPHIA consortium (6). Many of these can be used directly by toxicologists and pharmacologists when studying the effects of xenobiotics. There is also an extensive literature on the characteristics of the many inbred and genetically modified mouse strains (www.informatics.jax.org and http://jaxmice.jax.org/). Mice are also less expensive to maintain and use less of the test agent than rats. Even the use of techniques such as telemetry is now available for the mouse and new methods of whole body imaging such as SPECT and SPECT/CT (7) actually favour smaller animals, i.e. the balance is changing. According to one anonymous toxicologist in a major drug firm (8): “Although the rat genome is now available, the knowledge about differences between mouse strains, the huge access to mouse knockouts and the lower cost for maintaining mice will make mice preferred as an experimental species. However, there will be situations where it is very difficult to make measurements in mice, and sometimes this can be done in rats instead”.
The National Institute of Environmental Health Sciences (NIEHS) has recognised the value of isogenic mouse strains in identifying genes associated with response to xenobiotics. As part of their Host Susceptibility Program “. . . NTP (National Toxicology Program) scientists will take chemicals identified as toxicants in the research and testing program and evaluate them in multiple genetically diverse isogenic mouse strains to determine which strains are particularly sensitive or insensitive to the chemicals causing toxicity and associated disease” (http://ntp.niehs.nih.gov). They go on to state that “Ultimately, the NTP expects to learn more about the key genes and pathways involved in the toxic response and the etiology of disease mediated by substances in our environment. Such an understanding of genes and environment interactions will lead to more specific and targeted research with testing strategies for the NTP scientists to use for predicting the potential toxicity of substances in our environment and their presumptive risk to humans and disease susceptibility”. However, although there are many advantages in using mice, the main principle discussed here, namely taking control of genetic variation in the test animals, is equally applicable to both rats and mice.
4
Festing
1.3. Strains and Stocks of Rats and Mice
The three major classes of stocks which can be used in biomedical research are the following: 1. Outbred stocks (the term “stock” is used for outbred and “strain” for inbred colonies) are usually closed colonies (i.e. no new genetic material is normally introduced) in which each animal is genetically different and unique (9). These stocks may be designated by such names as Sprague–Dawley rats or Swiss mice, or a designation such as CD-1 or CFW. Although there are internationally recognised nomenclature rules, these are not always used consistently. 2. Inbred strains are like immortal clones of genetically identical individuals. They are produced by at least 20 generations of brother × sister mating with each individual being derived from a single breeding pair in the 20th or a subsequent generation, resulting in isogenicity (i.e. all animals being genetically identical) and homozygosity of practically all genetic loci (>98.6%). They are designated by a code such as LEW rats or C57BL mice. Different sub-strains (i.e. branches of a strain which are, or are presumed to be, slightly different) are indicated by further codes following a forward slash, e.g. C57BL/6J (www.informatics.jax.org). Nomenclature rules for inbred strains are well established and widely used (see www.informatics.jax.org/mgihome/nomen/). 3. Mutant and genetically modified strains either have a spontaneous or induced mutation or the deliberate incorporation of foreign DNA into the genome, sometimes in such a way as to disrupt a gene. This class of stock is of great importance in drug development and genetic research and is likely to be used increasingly in the study of toxic mechanisms, but a full discussion is beyond the scope of this chapter. However, the genetic modification may be maintained on either a genetically heterogeneous or an inbred genetic background. The former is not recommended because the mutation phenotype is often highly dependent upon the genetic background, and this may change rapidly in a heterogeneous stock unless it is maintained in large numbers preferably with a maximum avoidance of inbreeding rotational scheme (10). A brief summary of the properties of inbred strains and outbred stocks is given in Table 1.1. Inbred strains (or F1s made from them) have many advantages as experimental animals when compared with outbred stocks. They are uniform for most characteristics of toxicological and pharmaceutical interest because all animals within each inbred strain are genetically identical. In most cases this leads to less “noise” so either smaller numbers of animals are used or alternatively the experiments will be more powerful with less chance of false-negative results. Inbred strains are genetically much more
Inbred Strains in Toxicity Testing
5
Table 1.1 Brief summary of differences between inbred strains and outbred stocks Inbred strains
Outbred stocks
Isogenic All animals genetically identical (see Note 1) Homozygous Breed true. Parents and offspring genetically identical
Heterogenic Each individual genetically different and unique Heterozygous Do not breed true. May carry recessive genes. Parents and offspring genetically different
Phenotypically uniform Phenotypic variation essentially only due to environmental factors
Phenotypically variable Phenotypic variation is due to both genetic and environmental factors
Identifiable Each individual can be authenticated by genetic markers. Genetic quality control easy
Not identifiable There is no set of markers which can be used to authenticate an outbred stock so genetic quality control of individuals is impossible
Genetically stable Genetic drift slow and only due to mutation
Genetically unstable Genetic drift can be rapid due to changes in gene frequency caused by selection, random drift and mutation
Consistent data Extensive, reliable, data on genotype and phenotype of common strains
Variable data No data on genotype and reliability of data on phenotype questionable because of lack of strain authentication
Multi-strain experiments common. Differences between strains often the starting point for identifying sensitivity genes
Single stock experiments done Most investigators only use a single stock so differences between stocks are not commonly seen and the investigator is unaware of genetic variation in response
Universal Internationally distributed in academic and commercial organisations. Investigators in different countries can use genetically identical animals
Limited Source often limited to one or more commercial breeders. Each colony is unique so investigators in different countries have no access to a particular genotype
Consistent naming Genetic nomenclature well established with extensive lists of individual strains and their properties
Inconsistent naming Genetic nomenclature rules often ignored. No listings of stock characteristics.
stable than outbred stocks. Selective breeding (inadvert or otherwise) is ineffective in altering an inbred strain, whereas it can cause rapid change in an outbred stock. Many of the more widely used inbred strains were exchanged between investigators before they were fully inbred, so there are major sublines such as C57BL/6 and C57BL/10 which differed initially as a result of residual heterozygosity. Further, since then these strains have also gradually drifted apart as a result of the accumulation of new mutations
6
Festing
(11). This is generally relatively slow, although of some practical importance if a new mutation should affect a character of particular interest to an investigator. Even this source of variation can be greatly reduced by maintaining foundation banks of frozen embryos and restoring the strains from the freezer regularly (12). Recovery of a single sibling (brother–sister or parent– child) pair of animals is sufficient to recover the whole strain as each mouse has all the alleles present in the strain. Embryo freezing can also be used to preserve outbred stocks, but as each mouse is genetically different and as the genetic loci are not homozygous, large embryo stocks sizes are needed in order not to lose genetic variation. Each inbred strain can be easily identified using genetic markers. There is extensive data on the characteristics of most mouse and rat strains, and many genetic modifications (GM) and mutations are maintained on an inbred background so as to reduce genetic drift and increase the sensitivity when comparing GM versus wild-type animals. In short, inbred strains are the nearest representation of a “standard analytical grade reagent” that is possible when doing experiments with mammals. More details are given in www.isogenic.info. In contrast, outbred stocks are phenotypically more variable than inbred strains, with the possible exception of some aspects of reproduction, they are genetically labile because the frequency of alleles at any one locus within a colony can change rapidly as a result of selection or random genetic drift (13). Genetic quality control is impossible at the individual level. For example, it is not even easy to distinguish genetically between Wistar and Sprague– Dawley rats because there is no standard “Sprague–Dawley” rat or “CFW” mouse and colonies with the same name can be genetically quite different (9). In fact scientists have rarely made a scientific case for using them (14) except in special cases where, for example they want to use selective breeding or fine-mapping of a gene.
1.4. A Brief Review of the (Supposed) Logic for Continued Use of Outbred Stocks
With a few exceptions, such as the NTP Carcinogenesis bioassay, toxicologists have always used outbred stocks and innovation in toxicity testing has not been encouraged by the drug regulators, so there has been little incentive to change. Moreover, there seems to have been no serious discussion of the topic for at least 30 years. A brief review in 1987 noted a couple of committees which had recommended the use of outbred stocks, but these did not include mammalian geneticists and biostatisticians and their reasons for recommending the use of outbred stocks were unconvincing (15). The most common justification can be summed up by a quote from an anonymous toxicologist in 2005:
Inbred Strains in Toxicity Testing
7
“The variability of toxicity obtained in less well defined animals is a strength in itself, not a problem, when trying to predict safety margin in the non-isogenic human population”.
However, this simple logic is faulty. The fundamental assumption in toxicity testing is that if the compound causes toxicity in mice or rats, then it may cause similar toxicity in humans. Therefore we should be designing experiments which have the best possible chance of detecting adverse effects in our test animals. Every good scientist knows that the fundamental rule in designing an experiment is to control all variables except for the treatment. Toxicologists go to great pains to use animals of uniform weight, maintaining them free of disease in a controlled environment and feed them a standardised diet. The very last thing we should be doing in a toxicity test is trying to model human genetic variation with a handful of variable animals. Clinical trials need to be large because humans are genetically diverse. The advantage of using genetically defined rodents is that variability can be rigidly controlled, so that experiments can be much smaller or statistically more powerful. It makes no sense to leave crucial genetic variation uncontrolled when it can be controlled easily by using a group of isogenic strains. In any case, the genetic variation present in a few hundred outbred mice or rats from a single colony is not remotely comparable to the genetic variation found in many millions of humans of several different races. The UK Committee on Toxicity stated (incorrectly) that toxicologists use inbred strains (16) and went on to say that “A potential disadvantage of such tight controls of experimental conditions is that this approach reduces the chance of detecting an adverse effect that occurs only in a sub-group of the experimental animals. The use of larger groups of more outbred animals might increase the chances of detecting such groups, but this could not be guaranteed”.
This is a bad suggestion. Searching for genetic variation in response to a known toxic compound and testing whether a compound is toxic require different experimental designs. In the former case several hundred to a few thousand genetically heterogeneous animals are all treated with the toxic compound and sensitive and resistant animals are genotyped at a large number of polymorphic gene loci to see if any correlate with the sensitivity. In the latter case, animals are randomised into treatment and control groups and differences between group means and proportions are studied. Any attempt to combine the two objectives into a single experiment is likely to result in failure to achieve either objective. A more thoughtful claim is that . . . it is more correct to test on a random-bred stock on the grounds that it is more likely that at least a few individuals will respond to the administration of an active agent in a group which is genetically heterogeneous (17).
8
Festing
This refers particularly to discrete outcomes such as presence/absence of a tumour. However, although the variation within a single outbred stock is greater than that in a single inbred strain, outbred stocks do differ in mean response (18). A different conclusion could easily be reached if a resistant rather than a sensitive one is used. So if using outbreds it would be better to use small numbers of several stocks rather than larger numbers of one stock. However, a small collection of inbred strains will usually have an even wider range of susceptibility, particularly if unrelated strains are used. This considerably improves statistical power both for measurement and for discrete outcomes (19). Other objections raised, such as it being impractical to use several inbred strains or that the strains may not be available, are unconvincing. Geneticists commonly use many strains without difficulty, and many standard strains of both mice and rats are available commercially. Until a few years ago some toxicologists also claimed that they did not want to improve methods of toxicity testing because they were already discarding too many potentially useful compounds. However, most now recognize that it is far more cost-effective to discard risky compounds early, as clinical trials are extremely expensive and the withdrawal of drugs after marketing is even more so especially if linked with a lawsuit. In conclusion, modern toxicologists now have the opportunity to have serious discussions with geneticists and statisticians about the relative merits of using inbred strains versus outbred stocks in toxicity screening. Their past reasons for using outbred stocks which at first appear justified use incorrect logic and information. The solution for cost-effective drug screening is very simple. Control the within-strain genetic variation using isogenic strains and use small numbers of several strains to represent a wide range of sensitivity. This will result in more powerful experiments with fewer false negatives without using any more animals. Any differences between strains will give some indication of genetic variation in response. These strain differences can be also used to explore toxic mechanisms by studying the biology of response in sensitive and resistant strains which leads into personalized medicine, especially as many inbred strains have been sequenced.
2. Current Methods of Toxicity Testing
Currently, the pre-clinical testing of new drugs involves a number of formal experiments required by the regulatory authorities (16). These include 28-day, 90-day and 2-year studies in rodents usually involving four dose levels (including the control) and both
Inbred Strains in Toxicity Testing
9
sexes, with ten animals per group, or a total of 80 animals for the short-term tests. Many outcomes are measured from these tests. A 90-day study in a small numbers of dogs may also be required. For carcinogenesis 2-year studies are usually done in rats with both sexes, four dose levels and 50 animals per group, or a total of 400 animals. Data on one other species, often the mouse, are also required. More recently novel methods such as the use of transgenic strains are also being used, e.g. Trp53 knockout mouse (B6.129S2-Trp53tm1Tyj/J) for short-term carcinogenicity testing (20). Reproductive toxicity studies and in some cases multi-generation studies are increasingly required. A number of other tests such as for ocular toxicity, skin sensitisation and DNA damage may also be required. The tests actually used are tailored depending upon on individual circumstances and will also differ between pharmaceutical and industrial/environmental chemicals. 2.1. Study Design Using Inbred Strains
A small change which would substantially improve the existing experimental design for the 28-day and 90-day studies would be to replace, for example ten outbred animals in each experimental group either by three animals of each of three different inbred strains or two animals of each of five inbred strains (2 × 5). The choice being largely dependent on getting a balance between statistical power, which favours more strains, and practicality, which favours fewer strains. The existing 28-day experiment is already a factorial design with two sexes and four doses. Adding two or more strains would make it an elegant three-way factorial design (21). In order to make the experiment as simple as possible, if a three-strain study was chosen, it could be run as three separate “mini experiments” each involving a single strain with three male and three female animals at each dose level. This would involve a total of 24 animals, and would be repeated three times, once for each of the three strains. The data would then be combined into a single analysis with 3 (strains) × 4 (doses) × 2 (sexes) × 3 (animals per sub-group) involving all 72 animals. A similar procedure could be used if there were five strains, with each mini-experiment involving 16 animals, which is repeated once with each of the five strains. Several other variants and options become available once the principle of using several inbred strains is accepted.
2.1.1. Choice of Strains
Inbred strains could be chosen on the basis of availability, knowledge of strain characteristics and absence of any biological properties which would preclude use of the strain, such as autoimmune disease or cancer. Strains which are genetically dissimilar could be chosen in order to maximise the chance of choosing at least one susceptible strain. For example inbred mouse strains have been classified into seven major families, based on the analysis
10
Festing
of single nucleotide polymorphisms, so each strain could be chosen from a different family (22). F1 hybrids, the first-generation cross between two inbred strains, are isogenic (i.e. all animals are genetically identical), they are more vigorous than inbred strains and make a good choice in these assays. One possible objection is that they tend to be intermediate between the parental strains for many characteristics. So if the aim is to test the compound against strains which differ as much as possible it may be better to use pure inbred strains. However, the use of one or two F1 hybrids should certainly be considered. A multi-strain toxicity test will nearly always reveal strain differences in response to test substances and possibly also in basal levels, as is shown in the example given below. Most characteristics of interest to toxicologists have a polygenic mode of inheritance, and the strain differences will usually be due to quantitative trait loci (QTLs). Again another reason to use inbred strains because detecting QTLs in small samples of outbred stocks is virtually impossible. If the strain differences are small, then compound safety could be assessed according to the most sensitive strain. If there are large strain differences, then these should probably be investigated in more detail. They may be due to a major gene which may also be segregating in humans allowing potentially patient selection. 2.1.2. Dose Finding
This could normally be done in one strain provided the top dose used in the assay was well below the maximum tolerated dose, as this might be acutely toxic to a more sensitive strain. A lot of the increased sensitivity from using inbred strains comes from the reduced noise making it possible to detect subtle effects, not because the animals show more marked symptoms.
2.2. Use of Inbred Strains in Biomarker Development
Toxicologists are actively searching for better biomarkers of toxicity which give earlier and more subtle indications of tissue damage. A good biomarker should give repeatable results with a high signal/noise ratio. Unfortunately, most of this research is currently being done using outbred stocks, which may give unrepeatable results both over time, in different laboratories and even when nominally identical outbred animals come from different breeders. This is a waste of resources. Figure 1.1 shows the percent responders to a synthetic polypeptide in successive samples of about 30 outbred CD (Sprague–Dawley) rats from a single breeder over a 2-year period (23). Seven inbred strains gave consistent results (not shown), so it was the animals, not the assay which varied, but this would have been unclear if inbred strains had not also been used. It would be difficult to develop a new biomarker if the test animals vary in this sort of manner because it would not be clear whether the protocols or the animals were at fault.
Inbred Strains in Toxicity Testing
11
100 90
Percent responders
80 70 60 50 40 30 20 10 0 1
3
5
7
9
11 13 15 17 Sample number
19
21
23
25
Fig. 1.1. Percent responders in successive samples, each of about 30 outbred CD (Sprague–Dawley) rats from the same breeder, over a 2-year period. Re-drawn from Simonian et al. (24).
2.3. Identifying Susceptibility Loci: Towards Personalised Medicine
In the future it is thought that drugs will be administered to people according to their individual genotype for responses to a particular drug, i.e. personalized medicine. However, many of the genes controlling sensitivity to xenobiotics have yet to be discovered. Identifying these genes remains difficult in humans. If mild toxicity is seen in a few people in early clinical trials there is no assurance that it is due to genetic sensitivity, and it would be impossible to identify the susceptibility genes with such small sample sizes. Further, it is already too late once a drug is in phase III trials as it has cost many hundreds of millions of dollars by this time. A better strategy is to identify susceptibility genes in inbred rodents before starting the clinical trials. These are initially detected as differences between the strains. Using increasingly powerful genomic tools it is becoming possible to identify the genes involved, particularly if susceptibility depends on one or a few major loci, although this requires further experiments (24). If candidate genes can be identified in mice or rats the published literature could be checked to see if the same loci are associated with any adverse reactions in humans and these loci could be monitored in the early clinical trials. A full discussion of how individual susceptibility genes could be detected, starting with strain differences is beyond the scope of this chapter. However, a first step would usually be to characterise a wider range of inbred strains. If these fall into two groups of susceptible and resistant, then the difference may be due to a single locus. This may be mapped by determining whether any
12
Festing
genetic makers in F2 crosses between susceptible and resistant strains are associated with susceptibility. Combining this sort of information with knowledge of loci which have already been characterised, gene expression data, the location and type of adverse reactions and the known characteristics of the inbred strains often including their full DNA sequence may make it possible to identify the genes. These sorts of investigations are not trivial and would not always be worthwhile, but methods are developing rapidly so this likely to become easier in the future. Further details are available on the Complex Trait Consortium web site http://www.complextrait.org/.
3. Example: A Multi-strain Assay
This example demonstrates the increased power from using inbred strains compared with outbred stocks in obtaining relevant dose–response curves. The data in Table 1.2a show white blood cell (WBC) counts in four inbred strains of mice with two mice of each strain at each dose level (48 mice), treated with chloramphenicol succinate at six dose levels, these data were extracted from previously published data (25). Originally there were eight mice of each inbred strain at each dose, but in order to obtain two comparable experiments two mice of each of the four inbred strains were chosen at random at each dose level. Table 1.2b shows the WBC counts in 47 CD-1 mice, with seven to nine mice per dose level in a parallel experiment conducted at the same time and under the same conditions. Differences between two of the inbred mice of the same strain in the same treatment group (say the first two CBA mice) in Table 1.2a are due to non-genetic variation. This could be the
Table 1.2a WBC counts (× 109 /L) in four inbred strains of mice in response to chloramphenicol succinate (mg/kg) Dose
CBA
0
1.6
C3H 0.5
2.1
2.2
BALB/c
C57BL
2.3
1.9
2.2
Mean 2.6
1.925
500
1.1
1.2
1.2
2.1
1.8
1.8
3.5
1.1
1.725
1000
0.7
1.1
1.2
1.0
3.1
2.1
1.7
1.0
1.488
1500
0.9
1.1
1.1
0.6
2.1
3.0
2.1
1.3
1.525
2000
1.7
1.4
1.3
1.2
1.7
2.1
1.5
1.0
1.488
2500
1.3
1.4
0.4
0.4
0.8
1.1
0.4
0.2
0.750
Inbred Strains in Toxicity Testing
13
Table 1.2b WBC counts in CD-1 mice, in response to chloramphenicol succinate (mg/kg) Dose
Mean
0
3.0
1.7
1.5
2.0
3.8
0.9
2.6
2.3
500
2.8
1.9
1.5
1.7
1.6
1.3
1.9
1.6
1.788
1.7
2.075
1000
3.4
1.8
1.2
1.9
2.6
3.6
1.3
1500
2.3
2.3
2.1
2.4
1.6
1.7
2.5
2000
2.3
3.6
1.0
1.8
2.6
1.6
1.6
2500
1.9
1.9
3.5
1.2
2.3
1.0
1.3
1.7
2.167 2.257 2.071
1.6
1.838
result of environmental and developmental differences between mice and measurement error, i.e. experimental noise. Differences between two mice of different strains (e.g. a CBA and a C3H mouse) in the same treatment group are due to non-genetic plus genetic variation. Differences between two mice of the same strain in different treatment groups are due to a treatment effect plus non-genetic variation. Thus in this case it is possible to estimate non-genetic variation, genetic variation and treatment variation. Differences between any two CD-1 mice in the same treatment group in Table 1.2b represent non-genetic plus genetic variation. Differences between two mice in different treatment groups represents non-genetic, genetic and treatment variation. Thus when examining these data it is not possible to estimate either the non-genetic variation or the genetic variation! The two sources of variation are “confounded” or inextricably mixed. As the genetic and non-genetic variations are added together the total variation is increased. Moreover, the greater the genetic variation is, and therefore the greater its importance, the greater the noise will be and the less chance there will be of detecting a treatment effect. This is absolutely the opposite of what is wanted and needed in today’s toxicology studies. The inbred strain data in Table 1.2a can be statistically analysed using a two-way analysis of variance. The result is shown in Table 1.3. All statistical analyses shown below used the MINITAB statistical package (MINITAB Inc., 3081 Enterprise Drive, State College, PA 16801-3008, USA.). Other dedicated statistical packages will give similar results. Readers can repeat the analyses themselves using the data in Table 1.2 if they wish. Differences between strains and between doses are highly significant (p < 0.01 in both cases). The strain × dose interaction is approaching significance p = 0.081, implying that there may be strain differences in response. This could be indicated by a
Festing
Table 1.3 Analysis of variance for WBC counts in the inbred mice Source
DF
SS
MS
F
p
6.83
0.002 0.003
Strain
3
5.2817
1.7606
Dose
5
6.3442
1.2688
4.92
(Linear
1
4.9031
4.9031
19.07
Deviations
0.000)
4
1.4411
0.3603
1.40
0.283
Strain × dose
15
7.2708
0.4847
1.88
0.081
Error
24
6.1900
0.2579
Total
47
25.0867
The Linear Analysis is taken from Table 1.4.
significant linear strain × dose interaction term, not calculated here in order to avoid getting too complicated. Mean WBC counts for all five strains (including CD-1) at each dose are shown graphically in Fig. 1.2. For the inbred strains, the regression of WBC counts on dose can be estimated using all the inbred data (ignoring strain differences). This gives the regression equation WBC=1.9512−0.0003743 × dose. The ANOVA for the regression calculations is given in Table 1.4. The regression is significant at p = 0.002. This level of significance is adequate, but it is underestimated because strains differ in control mean WBC counts, as can be seen in Fig. 1.2, so variation about
BALB/c C3H C57BL CBA CD-1
2.5 2.0 WB Ccounts
14
1.5 1.0 0.5 0.0 0
500
1000 1500 Dose
2000
2500
Fig. 1.2. Dose–response relationships for each strain. The thick dotted line is the regression of WBC on dose averaged across the four inbred strains. This is statistically significantly different from 0 (p = 0.002). The thick dashed line is the corresponding line for CD-1. This is not significantly different from 0 (p = 0.63). Note that BALB/c has low basal counts and it and CD-1 show no significant dose–response. Strains CBA and C3H are most sensitive, with the response of C57BL being not significant, possibly due to non-linearity.
Inbred Strains in Toxicity Testing
15
Table 1.4 Analysis of variance for regression (four strains, strain differences ignored) Source Regression Residual
DF
SS
MS
F
p
11.17
0.002
1
4.9031
4.9031
46
20.1835
0.4388
47
25.0867
error Total
the regression line is much higher than if this were not the case. The MS for regression in Table 1.4 (4.9031) should be compared with the error MS in Table 1.3 (0.2579) to correct for this effect. This is shown as the “linear” effect in Table 1.3 and results in a much higher significance level. This linear effect could also be calculated using orthogonal contrasts, described in some textbooks, e.g. Snedecor and Cochran (26), though the calculations are not available in many statistical packages. It is probably easier just to estimate the regression of WBC counts on dose separately for each strain. These are summarised in Table 1.5. The R2 values are the percent of variance due to regression, the “S” is the standard deviation about the regression line and the p values indicate the probability that a slope of this magnitude could have arisen by chance. From these data only strains CBA and C3H show a statistically significant response (see Note 2). In all cases the 12 inbred mice give a better fit to the dose–response line than the 47 mice outbred stock mice. There are also highly significant differences in mean counts between strains ranging from 1.23 in C3H to 2.01 in C57BL, although here it is really only the response to the test compound rather than basal levels within each strain which is of interest. Some QTL associated with baseline WBC counts in mice have already been identified (27). Related to this observation there are also racial differences in WBC counts in humans, with African Americans having lower counts than European Americans, and a locus accounting for about 20% of the variation in WBC counts has been identified on chromosome 1q (28). There does not seem to be any information on human genetic variation in the WBC response to chloramphenicol although very rarely it can cause a fatal pernicious anaemia. The CD-1 data can be analysed by a one-way ANOVA. The p value for differences between doses is 0.792; this does not even approach statistical significance. The difference between the control and the highest dose level averaged across the four inbred strains is 1.175 counts and the standard deviation is 0.508 counts, so the signal/noise ratio is 1.175/0.508 = 2.314. In the CD-1 the difference is
16
Festing
Table 1.5 Summary of individual regression calculations for each strain
Strain
N
R2
S
Regression coefficient × 1000
p
CBA
12
48.1
0.70
−0.72
0.012
C3H
12
69.0
0.37
−0.58
0.001
BALB/c
12
16.6
0.34
+0.16
0.188
C57BL
12
23.3
0.61
−0.36
0.112
CD-1
47
1.0
0.73
−0.06
0.634
. R2 is the proportion of variance accounted for by regression. S is the standard deviation of deviations from regression. A lower value indicates a better fit of the regression line. The p values indicate the probability that a slope as great as this could have arisen by chance
only 0.329 counts and the standard deviation is 0.746 counts, giving a signal/noise ratio of 0.329/0.746 = 0.442. Using power analysis, the sample size that would be needed in some future experiment to detect a signal/noise ratio of 2.314 with a 90% power and a 5% significance level using a two-sided t test would be six treated and six control mice. However, in order to detect a signal/noise ratio of 0.442 found with CD-1 mice, it would require 109 animals per group, i.e. 18 times more resources. Altogether eight primary haematological parameters were measured, and the signal/noise ratios (i.e. difference between control and treated means divided by the standard deviation) of each of these are shown in Figs. 1.3 and 1.4. A comparison of these clearly shows the increased sensitivity of the multi-strain assay across all outcomes as well as showing which outcomes are responding and at which dose levels. In summary, the assay using four inbred strains (48 mice) is substantially more powerful than the one based on outbred CD-1 (47 mice), showing a highly significant regression of WBC counts on dose of chloramphenicol, averaged across four strains. However, this is restricted to only two of the strains. This clearly shows that there is genetic variation both in baseline WBC counts and in response to chloramphenicol and leads to some human literature on basal WBC levels. In contrast, the outbred stock CD-1 fails to show any dose-related response and provides no information on the genetic control of WBC counts. It is known that chloramphenicol causes a dose-related bone marrow suppression in humans (29) which is detected in the WBC lineage only by using
Inbred Strains in Toxicity Testing
HCT HGB Retics LYMP Neuts PLT RBC WBC
1 0 –1 Signal/noise ratio
17
–2 –3 –4 –5 –6 –7 –8 0
500
1000 1500 Dose
2000
2500
Fig. 1.3. Signal/noise ratios [(treated mean–control mean)]/SD for eight haematological traits in the CD-1 outbred stock. The dotted line represents the signal/noise ratio that should be detectable in a future experiment with a group size of 20 animals and a 90% power using a two-sided t-test and a 5% significance level.
1
HCT HGB Retics LYMP Neuts PLT RBC WBC
0 Signal/noise ratio
–1 –2 –3 –4 –5 –6 –7 –8 0
500
1000
1500
2000
2500
Dose Fig. 1.4. Signal/noise ratios [(treated mean–control mean)]/SD for eight haematological traits averaged across the four inbred strains. The dotted line represents the signal/noise ratio that should be detectable in a future experiment with a group size of 20 animals and a 90% power using a two-sided t-test and a 5% significance level. The last two points for the reticulocytes have been truncated at −7.5 so that the scale for Figs. 1.3 and 1.4 are identical. Note the increased sensitivity compared with the CD-1 mice in Fig. 1.3.
inbred strains. Similar conclusions are reached when all eight primary haematological characters are considered.
18
Festing
4. Conclusion A fundamental rule when designing an experiment is to control all variables except for the one being studied, namely the treatment. Failure to do this results in an increased chance of a false-negative result due to a low signal/noise ratio. Where the experimental subjects can also be classified into different groups, such as male and female, or strain A and strain B which may respond differently, this also needs to be taken into account. This can best be done by replicating the experiment within each class, so part of the experiment can be done on strain A males and part on strain B males and so on with the females. This of course assumes that the character being studied is present in all groups. Obviously we cannot study the effect of a xenobiotic on testis in females! What is not so widely understood is that using inbred animals in this type of design does not require four times as many animals. This is because the data can be combined into a single so-called factorial statistical analysis which separates out the effect of sex, strain and treatment and any interactions between them. Factorial designs are very powerful. R.A. Fisher, the great statistician and geneticist stated that “. . . in a wide class of cases (by using factorial designs) an experimental investigation, at the same time as it is made more comprehensive, may also be made more efficient if by more efficient we mean that more knowledge and a higher degree of precision are obtainable by the same number of observations.” (30). It is well established that different inbred strains and outbred stocks can respond differently to a xenobiotic. The important implication is that if a single strain or stock is used it may be relatively resistant to the test compound, giving a false-negative result. Moreover, if only a single strain or stock is used, the importance of strain/stock variation never becomes apparent. If it is possible to include more than one stock/strain in an experiment without the need to increase the total number of animals, and still with the same number of animals in each treatment group, it should be obvious that this would be a good strategy. Additionally, within each branch of such an experiment it makes sense to control the inter-individual variation by using isogenic strains. Indeed, the greater the genetic component of variation is, the more important it is to control it in this way if falsenegative results are to be avoided. There are, of course, several questions that need to be addressed in such a strategy. For example, how many strains should be used and how should they be chosen? The answer is that if the total number of animals (the main determinant of cost) is not increased, this sets a limit on the number of strains to be
Inbred Strains in Toxicity Testing
19
used. Strains should then be chosen to be as genetically different as is practical, given the availability of such strains. In the numerical example comparing a multi-strain study with one using a single outbred stock, group size was eight animals in both cases, so within this limit it was possible to use two mice of each of four inbred strains. The numerical example showed clearly that for the white blood cell lineage the CD-1 stock was both more resistant to the effects of the test chemical, chloramphenicol, and more variable in response. This resulted in a false-negative result. In contrast, two of the inbred strains responded strongly to the compound (a true positive result), also making it clear that there is genetic variation in response. Had the BALB/c strain been the only one used, then again there would have been a false-negative result; i.e. the use of a single inbred strain is not a good idea for a screening type of experiment. These data raise the question – So why don’t toxicologists use such a multi-strain assay of this sort? It is obviously more powerful and provides additional information? Perhaps this reflects a conflict between intuition and science. Intuitively, most scientists assume that as humans are genetically variable, they should use genetically variable animals as models. However, in a controlled experiment, genetic variation is just like any other source of variation, it needs to be controlled, if it is not it will confound the result. This is equally true in a clinical trial. If genetic variation was controlled using monozygous twins, then this would enormously increase the power of a clinical trial. Unfortunately this is not practical in clinical trials, but it is practical when using laboratory rodents. Also, unfortunately the idea that by using a factorial experimental design it is possible to get more information from the same number of animals is not well known. This seems to be due to a failure for most scientists to understand the principles of combining experimental design and statistics. Fortunately, research is becoming increasingly multi-disciplinary with statisticians becoming an important component of the research team. Toxicologists and pharmacologists only need to understand the broad principles. It is not essential that we all know exactly how to do the statistical analyses of this type of experiment. That can be left to the statisticians. The FDA in their “critical path” white paper have highlighted the need for improved methods of toxicity testing. Presented here is one obvious way ahead. By changing the type of animals used, the whole power of modern genetic techniques will become available to the pharmaceutical industry. The NIEHS is already starting down this path using 15 fully sequenced inbred strains of mice. With the dramatic advances made in the science of genetics in the last few decades, now is the time for toxicologists and others in drug development to start “thinking genetics”.
20
Festing
5. Notes 1. F1 hybrids, the first-generation cross between two inbred strains are isogenic but not homozygous and as such they do not breed true, i.e. F1’s between any pair of inbred strains are isogenic, however, if interbred lead to highly variable phenotypes. 2. Note that the standard deviation S, a measure of noise, is higher in the outbred CD-1 stock than in any of the inbred strains. References 1. Russell, W. M. S. and Burch, R. L. (1959) The principles of humane experimental technique, Universities Federation for Animal Welfare (UFAW), Potters Bar, Herts. 2. Food and Drug Administration (2004) Challenge and opportunity on the critical path to new medical products. http://www.fda. gov/oc/initiatives/criticalpath/whitepaper. html. 3. Caldwell, G. W., Ritchie, D. M., Masucci, J. A., Hageman, W. and Yan, Z. (2001) The new pre-preclinical paradigm: compound optimization in early and late phase drug discovery. Curr Top Med Chem 1, 353–366. 4. Food and Drug Administration (2008) The FDA Critical Path Initiative. http://www. fda.gov/oc/initiatives/criticalpath/ 5. The Innovative Medicines Initiative (2008) Innovative Medicines Initiativw. http://imi.europa.eu/docs/imi-gb-006v215022008-research-agenda˙en.pdf 6. Brown, S. D., Chambon, P. and de Angelis, M. H. (2005) EMPReSS: standardized phenotype screens for functional annotation of the mouse genome. Nat Genet 37, 1155. 7. Franc, B. L., Acton, P. D., Mari, C. and Hasegawa, B. H. (2008) Small-animal SPECT and SPECT/CT: important tools for preclinical investigation. J Nucl Med 49, 1651–1663. 8. Petit-Zeman, S. (2004) Rat genome sequence reignites preclinical model debate. Nat Rev Drug Discov 3, 287–288. 9. Chia, R., Achilli, F., Festing, M. F. and Fisher, E. M. (2005) The origins and uses of mouse outbred stocks. Nat Genet 37, 1181– 1186. 10. Festing, M. F. W. (2003) Laboratory animal genetics and genetic quality control, in
11.
12. 13. 14.
15. 16.
17. 18.
19.
Handbook of laboratory animal science: essential principles and practices (Hau, J. and Van Hoosier, G. L., Jr., eds.), 2nd ed. CRC Press, Boca Raton, London, New York, pp. 173– 204. Stevens, J. C., Banks, G. T., Festing, M. F. and Fisher, E. M. (2007) Quiet mutations in inbred strains of mice. Trends Mol Med 13, 512–519. Taft, R. A., Davisson, M. and Wiles, M. V. (2006) Know thy mouse. Trends Genet 22, 649–653. Papaioannou, V. E. and Festing, M. F. (1980) Genetic drift in a stock of laboratory mice. Lab Anim 14, 11–13. Festing, M. F. W. (1999) Warning: the use of genetically heterogeneous mice may seriously damage your research. Neurobiol Aging 20, 237–244. Festing, M. F. (1987) Genetic factors in toxicology: implications for toxicological screening. Crit Rev Toxicol 18, 1–26. Committee on Toxicity and Food Standards Agency (2007) Variability and Uncertainty in Toxicology of Chemicals in Food, Consumer Products and the Environment. cot.food.gov.uk/pdfs/cotstatementworkshop 200703.pdf Arcos, J. C., Argus, M. F. and Wolf, G. (1968) Chemical induction of cancer. Academic Press, Inc., New York. Kacew, S., Ruben, Z., McConnell, R. F. and MacPhail, R. C. (1995) Strain as a determinant factor in the differential responsiveness of rats to chemicals. Toxicol Pathol 23, 701– 715. Felton, R. P. and Gaylor, D. W. (1989) Multistrain experiments for screening toxic substances. J Toxicol Environ Health 26, 399– 411.
Inbred Strains in Toxicity Testing 20. Floyd, E., Mann, P., Long, G. and Ochoa, R. (2002) The Trp53 hemizygous mouse in pharmaceutical development: points to consider for pathologists. Toxicol Pathol 30, 147–156. 21. Montgomery, D. C. (2004) Design and analysis of experiments, 6th ed., John Wiley & Sons, Inc., Hoboken, NJ. 22. Petkov, P. M., Ding, Y., Cassell, M. A., Zhang, W., Wagner, G., Sargent, E. E., et al. (2004) An efficient SNP system for mouse genome scanning and elucidating strain relationships. Genome Res 14, 1806– 1811. 23. Simonian, S. J., Gill, T. J., 3rd and Gershoff, S. N. (1968) Studies on synthetic polypeptide antigens. XX. Genetic control of the antibody response in the rat to structurally different synthetic polypeptide antigens. J Immunol 101, 730–742. 24. Churchill, G. A., Airey, D. C., Allayee, H., Angel, J. M., Attie, A. D., Beatty, J., et al. (2004) The Collaborative Cross, a community resource for the genetic analysis of complex traits. Nat Genet 36, 1133–1137. 25. Festing, M. F., Diamanti, P. and Turton, J. A. (2001) Strain differences in haematologi-
26. 27.
28.
29.
30.
21
cal response to chloramphenicol succinate in mice: implications for toxicological research. Food Chem Toxicol 39, 375–383. Snedecor, G. W. and Cochran, W. G. (1980) Statistical methods, 7th ed. Iowa State University Press, Ames, Iowa. Peters, L. L., Zhang, W., Lambert, A. J., Brugnara, C., Churchill, G. A. and Platt, O. S. (2005) Quantitative trait loci for baseline white blood cell count, platelet count, and mean platelet volume. Mamm Genome 16, 749–763. Nalls, M. A., Wilson, J. G., Patterson, N. J., Tandon, A., Zmuda, J. M., Huntsman, S., et al. (2008) Admixture mapping of white cell count: genetic locus responsible for lower white blood cell count in the Health ABC and Jackson Heart studies. Am J Hum Genet 82, 81–87. Feder, H. M., Jr., Osier, C. and Maderazo, E. G. (1981) Chloramphenicol: a review of its use in clinical practice. Rev Infect Dis 3, 479–491. Fisher, R. A. (1960) The design of experiments, 7th ed. Hafner Publishing Company, New York.
Chapter 2 The Sophisticated Mouse: Protecting a Precious Reagent Michael V. Wiles and Rob A. Taft Abstract Definable, genetically and environmentally, the humble mouse has become a reagent with which to probe the human condition. The information thus gained is leading to a greater understanding of interindividual variation in drug responses and disease processes and is forming the basis for personalized medicine. Inbred mice are the tool of choice as each strain is essentially clonal in nature creating a defined, uniform setting where the effects of genetic background and modifications can be evaluated coherently. However, the creation and characterization of novel mouse strains remain expensive and time consuming. Further, the continual maintenance of these valuable animals as live colonies is financially draining and carries continual potential risks, including disastrous loss due to fire, flood, disease, etc. There are also other more insidious disasters including genetic contamination and genetic drift, either of which can go undiscovered until their effects ruin experiments. With this in mind, we strongly recommend that all mouse strains be cryopreserved as a matter of standard mouse management. Cryopreservation is a powerful colony management tool, assuring strains are available upon demand, for example, for regulatory requirements, re-initiation of projects, collaborations, re-evaluation of data etc. However, it is essential that any cryopreservation approach be cost-effective for both strain closure and strain recovery. In this chapter, we describe the variables which can afflict an inbred mouse’s genetic background (and hence phenotype), options to consider for strain archiving, and describe how to economically store and recover strains by sperm cryopreservation. Key words: Genetically modified mice, inbred mice, sperm cryopreservation, embryo cryopreservation, IVF, in vitro fertilization, genetic drift, genetic contamination. Abbreviations: IVF in vitro fertilization
1. Introduction 1.1. The Reagent Grade Mouse
Since the time of alchemy those engaged in research have continually needed to further refine and define their basic experimental reagents, their scientific tools of the trade. The resulting progress
G. Proetzel, M.V. Wiles (eds.), Mouse Models for Drug Discovery, Methods in Molecular Biology 602, DOI 10.1007/978-1-60761-058-8 2, © Humana Press, a part of Springer Science+Business Media, LLC 2010
23
24
Wiles and Taft
has allowed alchemy to become science and has driven the scientific revolutions of the 20th and 21st centuries. The term “reagent grade” is usually only applied to chemicals where impurities are either vanishingly absent or have been defined. In the field of experimental biomedical and genetic research, the evolution of the mouse as a reagent has been similar to other reagents whose precise definition and refinement has continually improved. Now, with the sequencing of the mouse genome and improvements in genetic monitoring of inbred mouse strains we believe that the term reagent grade can justifiably be applied to the mouse (1). This particular reagent is, however, one of the most sophisticated scientific tools available. It is a living creature (and as such deserving respect), a highly complex mammal capable of independent survival and is the net result of millions of years of evolution. So as with any reagent, its basic attributes must be understood if it is to be used successfully, including purity and/or quality, and how to maintain its full functionality. Two things make the reagent grade mouse invaluable as a model organism. First, it is possible to introduce precise genetic modifications into the mouse genome at will. These genetic modifications including gene ablation, addition, and modulation enable the rapid examination of the effects of these modifications in a complete living organism (2, 3). Second, comparisons of the mouse and human genomes reveal that mice and humans share a common ancestor which diverged about 75 million years ago, as such these two species have maintained many similarities in gene function. Comparisons between the two species show 99% shared gene function (4). Combined, these attributes make the mouse a reagent of exquisite subtlety and sophistication, enabling us to understand gene interactions within the complete organism, test their interplay with the environment, and extrapolate these data to human biology. However, “living reagents” require continual maintenance if they are not to be lost or degraded. It is here that mouse management by cryopreservation can be used, providing versatile archiving of these valuable resources. Lastly, the scientific method is based on the altruistic notion that others can take data and information, repeat it, and build upon it. Without repeatability there is no scientific advancement. As Sir Isaac Newton reportedly said, “If I have seen further it is by standing on the shoulders of Giants.” Treating the mouse as a living reagent will help ensure the creation of a sound foundation on which others can stand upon, build, and hopefully see further. 1.2. Mice Change and be Can Lost, If Not Cared for
Mice are a key element in many biological experimental designs, but their origin and quality are often overlooked. The variability introduced by using ill-defined mice, for example, randomly crossed strains, e.g., CD-1, or incomplete inbred
Sophisticated Mouse
25
congenics, increases experimental noise within and among experiments. If this is not understood and part of an experimental approach, this genetic variation will lead to a lack of reproducibility and difficulty in interpretation (see Chapter 1 by Festing this volume). Inbred mice, defined by at least 20 generations of brother– sister matings, are the most defined mammal available for experimental manipulation, with individual mice within each inbred strain being essentially clonal in nature (>99 % homozygosed at all loci). This allows precise experimental comparisons within strains, between multiple inbred strains, and between genetically modified versus non-modified mice of the same genetic background (5). However, all life has an innate biochemical, evolutionary capability to change and mutate, generating variation; mice are the selective result of this evolutionary past and their current environment (6). To work with these complex reagents successfully requires an appreciation of this. Inbred mice are generally maintained as a continually breeding colony requiring precise control of breeding. If done incorrectly their genetic background will change. There are two major sources of genetic change reported in live mouse colonies: (i) fast, disastrous genetic contamination (one breeding cycle) and (ii) insidious genetic drift. Until recently genetic drift was viewed as a slow process, however, recently the phenomena of copy number variation (CNV) has been discovered as a rapid source of genetic variation and hence drift. Although its impact is not fully understood, it is thought to cause very rapid genetic drift (within one generation) via duplication or deletion of one to thousands of kilobases of DNA, potentially containing entire genes. The resulting varying copy number effectively changes gene dosage which can result in phenotypic shifts (7, 8). Appropriate use of cryopreservation can forestall the cumulative adverse effects of genetic drift, including CNV, and allow rapid restoration of strains if genetic contamination, disease, phenotypic shifts, etc. occur (for a review ref. 1). Additionally, cryopreservation increases mouse management options, facilitating more cost-effective colony management. In regard to mainly custom genetically modified strains, although it is tempting to believe that these strains are safe during active experimental work, all vivariums carry risks including the possibility of disease, breeding cessation, genetic contamination, and other disasters. For example, in June 2001 the tropical storm Allison caused the flooding of vivariums at the Texas Medical Center killing more than 30,000 mice and rats, causing incalculable losses (9). Strain backup is therefore a prudent management step and also facilitates dynamic cost management allowing strains to be closed down and only upon demand, rapidly re-initiated.
26
Wiles and Taft
1.3. Mouse Archiving – Options
In the management of any resource a key consideration is “Return on Investment,” i.e., in this case, it is of little value to cheaply store/archive mouse strains in a format which makes their recovery prohibitively expensive and/or unpredictable (see Table 2.1 for summary of approaches). While at the same time it is not viable to invest large sums into the archiving of strains if the likelihood of them ever being wanted at a later date is small or totally unknown. Mouse strains can be archived as embryos (2–8 cell), gametes (sperm, oocytes), or as sources of gametes (spermatogonial stem cells, ovaries), see Table 2.1. When looking at costs to cryopreserve and recover mouse strains, sperm is in general the most logical choice based on the ease of collection and the sheer numbers of sperm available from a single male (∼30 × 106 sperm/male). However, until very recently recovery of live born from C57BL/6 sperm, the most commonly used background, was less than 5%, making routine recovery expensive and unpredictable. Females can provide naturally ∼6–8 oocytes or upon superovulation up to 50 oocytes. However, this is highly strain dependent, for example, C57Bl/6J give high numbers, while 129 strains give very few embryos. Complicating this there is also evidence suggesting that embryo quality falls with the high oocytes yields (10, 11). Thus upon comparing the economics of the two methods, embryos and sperm, it is very apparent that the freezing of embryos is considerably more expensive due to the need for more resources; for example, with a C57BL/6 background, to produce ∼250 two-cell embryos for cryopreservation by IVF requires >15 females. If strains are never recovered or only recovered once or twice, then the bulk of this expense remains forever frozen. In contrast, cryopreserving sperm has a low initial cost as only few animals (1–3 males) and relatively little labor and materials are required (12). It is upon recovery from sperm by in vitro fertilization (IVF) that animals and labor are used, but then only the required number animals per recovery are used as the IVF process can be scaled to produce the desired number of offspring. A major disadvantage of sperm cryopreservation of a strain is that only half the genome is stored by this approach (i.e., the donor sperm is haploid!). This does not represent a major problem when a genetic modification is on a readily available standard background, as high-quality female animals (oocyte donors) are readily available from reputable providers for most standard inbred mice. Although it should be appreciated that genetic drift still occurs unless the supplier has addressed this issue by restoring the breeder stock from cryopreserved pedigreed embryos every few generations (1).
Very simple Needs only1–3 carrier males ∼30×106 sperm/male Inexpensive
Simple Inexpensive
Moderately simple Needs only 1–3 carrier males
Ovary cryopreservation
Two cell embryo cryopreservation (heterozygotes embryos generated via IVF)
Pros
Sperm cryopreservation
Cryopreservation method/gametes
Cryopreservation
Only half the genome is preserved Needs IVF to make embryos Needs female (wild type) oocyte donors. Strain dependent 5–50 oocytes/female Only half the genome is persevered
Only half the genome is preserved. Needs multiple donor females, i.e., one female has only two ovaries
Due to potential rejection needs appropriate recipient animals. Moderate level of surgical skill to implant. Low yield of offspring/ovary Off-spring will be heterozygotes Can transmit disease Not scalable
Simple recovery into Low yield of offspring pseu-dopregnant animals Offspring will be Can be used to achieve heterozygotes strain rederivation. Not scalable (based in initial investment)
Represents female linage
Moderately simple Requires IVF to recover strain Can be used to achieve Requires appropriate oocyte strain rederivation donor strain Highly strain reproducible Some strains and mutaHighly scalable tions adversely affect IVF success. Offspring will be heterozygotes
Only half the genome is preserved Needs IVF quality control
Cons
Pros
Recovery Cons
Table 2.1 Summary of approaches to manage mouse strains
$$$
$$
$
(continued)
Cumulative cost/ animal to cryopreserve and recover
Sophisticated Mouse 27
Simple Inexpensive
ES cellsb None
Only half the genome is preserved
None
None
Simple to implant Can be used to achieve rederivation of strain
$$$$$
$$$$
Need to make germline $$$$$ transmitting chimeras Upon germline transmission only half the ES derived offspring will be heterozygotes ES cells carry tissue culture associated genetic damage Not scalable
Requires a high level of resources/skill to recover Pathogen transmission could occur Offspring will be heterozygotes Associated with genetic damage Not scalable
Expensive resource to restock Not scalable
Cons
Cumulative cost/ animal to cryopreserve and recover
Note: all the above approaches are also subject to strain effects and to possible deleterious effects of gene modification or addition. a Intracytoplasmic sperm injection, ICSI although not strictly necessary a cryopreservation method – this approach has been used as method to archive and restore mouse strains. b Embryonic Stem Cell, ES cell are often generated as part of the process for a targeted genetic modification. Although a strain can be recreated from the original targeted ES line it requires the re-creation of germline chimeras and their successful germline breeding before the strain is recovered. Further, ES cells are known the mutate in culture.
Very simple Inexpensive
ICSIa
Needs a large colony of strain to be cryopreserved (expense) to provide embryos
Pros
Cons
Pros
Simple Provides “homozygous” storage
Recovery
Cryopreservation
Two to eight cell embryo cryopreservation (flushed homozygous embryos)
Cryopreservation method/gametes
Table 2.1 (continued)
28 Wiles and Taft
Sophisticated Mouse
1.4. How Safe Is Gamete Cryopreservation
29
Any approach for archiving gametes has to provide long-term secure storage. Most methods for archiving mouse strains cryopreserve embryos, sperm, ovaries, etc., in vapor phase or liquid nitrogen. There are many papers discussing the longevity of gametes in cryogenic storage; however, it is generally accepted that once samples pass through the glass transition temperature of water, ∼−137◦ C all biological activities cease. Although gamma radiation can still cause accumulative damage, simulation studies suggest that this is insignificant over ∼2000 years under normal background radiation levels (13, 14). Of much greater concern with long-term cryogenic storage is temperature variation, where gametes are exposed to temperature fluctuations above −137◦ C. The most likely causes of temperature variation (increase) is improper handling of the frozen gametes (e.g., while “rummaging” in the liquid nitrogen storage tanks), failure to fill liquid nitrogen tanks, i.e., they run dry, destruction of the storage facility due to fire etc., or the physical failure of the tanks vacuum (15). As such, it is strongly recommended that cryopreserved gametes be stored physically in at least two liquid nitrogen storage tanks and, additionally, that tanks be in two or more separate facilities as one part of a comprehensive approach to repository operation (16).
2. Materials 2.1. Cryopreservation of Mouse Sperm
1. Distilled water (Invitrogen, cat # 15230-238)
2.1.1. Cryoprotective Medium
3. 3% w/v skim milk (BD Diagnostics cat # 232100).
2.1.2. Consumables
1. 0.25 mL French straws (IMV cat # AAA201)
2. 18% w/v raffinose (Sigma cat # R7630) 4. MTG: 447 M monothioglycerol (Sigma cat # M6145)
2. Cassettes (Zander Medical Supplies, 145 mm 16980/0601) 3. Styrofoam box internal dimensions ∼35 cm × ∼30 cm, a Styrofoam float (piece should be approximately ∼2–3 cm thick and be cut to cover ∼80% of internal area of box) 4. Monoject insulin 1 mL syringe 2.1.3. Mice – Strain to be Cryopreserved
Two to three male mice, preferably 10–16 weeks old (see Note 1)
2.2. In Vitro Fertilization Method
1. Pregnant mare serum gonadotropin (PMSG)
2.2.1. Hormones for Superovulation
3. Sterile physiological saline
2. Human chorionic gonadotropin (hCG)
30
Wiles and Taft
2.2.2. Mice for Superovulation 2.2.3. In Vitro Fertilization
Five to ten female mice, 3 weeks or 6–12 weeks of age, depending on the strain (17) 1. MVF media: Research vitro Fert (K-RVFE-50) COOKS Mouse In Vitro Fert Fertilization medium (Cook MVF, Australia, see http://www.specialtyvet.net/ page/page/6095801.htm) or Human Tubal Fluid media (see 18): NaCl (FW 58.44, Sigma S-5886) KCl (FW 74.55, Sigma P-5405) MgSO4 . 7H2 O (FW 246.5, Sigma M-7774) KH2 PO4 (FW136.09, Sigma P5655) CaCl2 . 2H2 O (FW 147, Sigma C-7902) NaHCO3 (FW 84.01, Sigma S-5761) Glucose (FW180.16, Sigma G-6152) Na-pyruvate (FW110.0, Sigma P-4562) Na-lactate 60% syrup (FW 112.1, Sigma L-7900) Penicillin-G (FW 372.5, Sigma P-4687) Streptomycin sulfate (FW 1457.4, Sigma S-1277) Phenol red (5%) (Sigma P-0290) BSA (Equitech-Bio BAC62-0050) Fill with water-cell culture grade (Sigma 59900C)
5.9375 g 0.3496 g 0.0493 g 0.0504 g 0.3 g 2.1 g 0.5 g 0.0365 g 3.42 mL 0.075 g 0.05 g 0.20 mL 4.0 g
Weigh each component and dissolve in high quality water (cell culture grade, double glass distilled or reverse osmosis, and filtration, i.e., 18 M) in a 1 L volumetric flask, but withhold the BSA for addition later. Bring the volume up to 1 L. Measure the osmolarity (290 ± 5). Bubble gas (5% CO2 , 5% O2 , 90% N2 ) through the medium for ∼5 min, add the BSA to the media, mix gently to avoid frothing. Filter through a 0.2 m filter into sterile bottles. Gas the medium with a mix of 5% CO2 , 5% O2 , 90% N2 to displace air above the medium, cap tightly, and store at 4◦ C for not more than 2 weeks. Repeat gassing after every use to maintain a pH of 7.2–7.4. 2. Mixed gas (5% O2 , 5% CO2 , balanced with N2 ) 3. One large 60 × 100 mm Falcon Petri dish for every three females (BD Biosciences) 4. One small 35 × 10 Falcon Petri dish for each male (BD Biosciences) 4. Embryo-tested mineral oil 5. Phosphate Buffered Saline(PBS)
Sophisticated Mouse
31
3. Methods It is critical that all media are carefully prepared and have the correct pH and temperature, as well as batch-tested reagents. Where possible we suggest buying reagents readymade. Further, it is helpful to pay attention to details and efficient laboratory setup, e.g., small incubators, heated stages, or other devices that ensure proper temperature and pH stability. 3.1. Sperm Cryopreservation 3.1.1. Preparation of Cryoprotective Media (CPM)
This method has been successfully used for many different mouse strains and backgrounds (12). 1. Place ∼80 mL of bottled distilled water in a beaker. 2. Heat for ∼40 sec in microwave to ∼60–80◦ C (do not boil). 3. Place beaker on heated stir plate, add 18 g of raffinose, and heat and stir till solution clears (see Note 2). 4. Add 3 g of skim milk to the raffinose mixture and heat and stir until dissolved (see Note 3). 5. Transfer solution to volumetric flask and bring to 100 mL with bottled distilled water. 6. Add MTG now or after thawing (see Note 4) 7. Mix well and divide the solution into two 50 mL centrifuge tubes. 8. Centrifuge at 13,000 × g for 15 min at room temperature (∼22◦ C). 9. Filter through a 0.22 m cellulose filter (a prefilter may help the flow). 10. Verify that the osmolarity is in the range of 470– 490 mOsm. 11. Aliquot 10 mL of filtered cryoprotective media into labeled 15 mL conical tubes. 12. Cap and store at −20◦ C until ready for use (see Note 5).
3.1.2. Sperm Cryopreservation Setup
1. Thaw and warm CPM in 37.5◦ C water bath (see Note 6). While the media is warming. 2. Label and mark straws and affix to a 1 mL monoject syringe. 3. Fill Styrofoam box to a depth 6–9 cm of liquid nitrogen. 4. Place Styrofoam float into Styrofoam box. 5. Replace Styrofoam box lid to slow the evaporation of liquid nitrogen. 6. Place the lid from Petri dish on the warming tray and lean the bottom of a Petri dish against it so that one side of the
32
Wiles and Taft
Petri dish is elevated. This arrangement forces the CPM to collect on one side, making it easier to fill the straws. 7. If monothioglycerol was not added prior to freezing, add it now, to a final concentration of 477 M. Add 1 mL of CPM to the dish for each male from which sperm will be collected. 3.1.3. Sperm collection
1. Euthanize the males (1–3) and remove the cauda epididymides and vas deferentia, carefully removing the testicular artery to avoid contaminating the sperm with blood. 2. Release sperm into the CPM by making several cuts through the epididymides and vas deferentia using a beveled hypodermic needle while holding the tissues with a pair of forceps. 3. Remove tissue from the CPM after 10 min.
3.1.4. Sperm cryopreservation
1. Aspirate a 4.5 cm column of CPM into a French straw followed by a 2 cm column of air. 2. Aspirate a 0.5 cm column of sperm into the French straw then aspirate additional air until the column of CPM without sperm contacts the PVA powder in the cotton plug. 3. Seal the end of the French straw with a brief pulse from an instantaneous heat sealer. 4. Repeat this process until the desired number of straws has been filled (we suggest minimum of 20/strain). 5. Place five straws into one cassette. Repeat until four cassettes have been filled. 6. Place the cassettes in the liquid nitrogen-filled box on the float (i.e., in vapor phase) so that they are not touching. 7. Put the lid on the box for 10–30 min. 8. Plunge the cassettes into the liquid nitrogen. 9. After at least 10 min in liquid nitrogen the cassettes can be removed and rapidly placed into storage in liquid nitrogen (see Note 7).
3.2. In Vitro Fertilization (IVF) with Frozen Sperm
Once a strain is frozen as sperm we strongly recommend that 1–2 straws be used to assess the quality of the sperm post-thaw. Although various devices exist to measure sperm motility etc., the only relevant test for sperm function is an actual IVF. Additionally, fertilization rates vary widely among commonly available inbred strains; also the introduction of mutations and genetic modifications into a strain can have indirect and unanticipated effects on the quantity and quality of oocytes and sperm produced, as well their performance during IVF (12, 17). IVF can be difficult to establish – the quality of the reagents is crucial, also during the IVF process it is essential for repro-
Sophisticated Mouse
33
ducible success to maintain media pH and temperature and that all practical steps are taken to keep precise control of the culture conditions. 3.2.1. Superovulation
1. Inject females with PMSG 44–48 h prior to injection with hCG (see Note 8). 2. Inject females with hCG 13 h prior to oocyte collection (see Note 8).
3.2.2. IVF Setup
1. Prepare oocyte collection dish by adding 2 mL of PBS to a 25 mm × 10 mm dish and keep at 37.5ºC in air. 2. Prepare IVF dish by placing a 250 L drop of MVF medium (see Note 9) in the center of a 60 mm Petri dish. Place four additional 150 L drops of MVF medium around the 250 L drop. 3. Carefully add sufficient oil to cover the media and place in an incubator or sealed chamber filled with mixed gas (5% O2 , 5% CO2 , 90% N2 ) at least 1 h prior to IVF (see Note 10).
3.2.3. Thawing Sperm
1. Place the straw in a 37.5◦ C (clean) water bath. 2. Rapidly swirl the straw in the water until all ice has melted (about 30 sec). 3. Dry the straw with a paper towel. 4. Cut off the sealed end of the straw opposite the cotton plug. Using a metal rod, expel the sperm from the straw into the 250 L IVF drop. 5. Allow sperm to incubate at 37◦ C for 1 h prior to adding oocytes.
3.2.4. Oocyte Collection and IVF
1. Euthanize 2–5 superovulated females approximately 13 h post-hCG (see Note 11). 2. Remove the ovary, oviduct, and a small portion of the uterine horn and place in the dish containing PBS from one female at a time (see Note 12). 3. Repeat for all females. 4. Working under low magnification, identify the ampulla. Cumulus enclosed oocytes should be easily visible within the ampulla of the oviduct. Using a beveled hypodermic needle, open the ampulla to release the cumulus-enclosed oocytes. 5. Repeat until all oocytes have been released. 6. Using a 1 mL pipette (or a wide bore pipette tip) transfer the cumulus enclosed oocytes to the dish containing MVF
34
Wiles and Taft
medium and thawed sperm (∼10 L), transferring as little medium as possible. 7. Incubate at 37ºC for 4 h under mixed gas. 8. Using a finely drawn glass pipette with a diameter slightly larger than an oocyte. Wash the oocytes through the 150 L media drops to remove cumulus cells and sperm. 9. Culture overnight at 37º C under mixed gas (see Note 13). 10. Count and evaluate embryos the following morning. Embryos can now be cultured, transferred to a pseudopregnant animal, or cryopreserved (see Notes 14 and 15).
4. Notes 1. Variations in sperm quality among individual males within a strain are common. 2. This is a nearly saturated solution, heating the solution makes it easier to get the raffinose into solution. 3. The solution will be opaque after the addition of the skim milk, centrifugation at room temperature (13,000 × g for 15 min) is recommended. 4. Addition of MTG is recommended immediately prior to use. Alternatively, it can be added in advance and the solution stored at −80ºC for up to 3 months. Solutions containing MTG should not be stored at 4ºC for more than a few days. MTG is viscous and needs to be pipetted carefully. Making an MTG stock solution that is added to the media helps reduce the likelihood of errors. MTG diluted stock solution should be used only on the day it is made. 5. This solution can be stored for at least 6 months at −80ºC without MTG or up to 3 months −80ºC with MTG. 6. Water baths are a common source of bacterial contamination and also liquid nitrogen is not sterile. Straws should be carefully wiped to remove any moisture from the outside of the straw prior to cutting the end off before dispensing sperm to reduce the risk of contaminating the IVF. 7. It is essential that during the transfer from liquid nitrogen to long-term storage to handle the cassettes rapidly thus preventing any warming. 8. Typical doses are in the range of 2.5–5 i.u. per mouse. Optimal dose varies by strain, age, and weight of the mouse. Extending the oocyte collection window beyond 14 h post-hCG may reduce fertilization rates and compromise embryo quality due to oocyte aging.
Sophisticated Mouse
35
9. COOKS Mouse Vitro Fert is similar to Human Tubal Fluid media reported by Quinn (18, 19). 10. Oil can be washed and filtered. Oil should be stored in a dark cool place. 11. The response to superovulation is highly strain dependent (see 17) and some strains appear to be entirely refractory to superovulation. 12. In order to reduce the risk of changes in temperature and pH, oocyte collection should take no more than 5 min from euthanasia to oocyte collection (i.e., practice). Typically cervical dislocation is used to prevent possible exposure to agents that may affect oocyte or embryo quality and to reduce the time from euthanasia to oocyte collection. 13. The use of a low O2 culture environment may not improve fertilization rate, but appears to improve embryo quality (20, 21). 14. The laboratory environment can have a significant effect on the outcome of IVF (22, 23). Materials that release volatile organic compounds (VOC), cleaning/sanitizing agents, such as bleach and floor waxes, should be avoided. 15. Prior to embryo transfer, embryos should be washed following the IETS protocol if the sperm or oocytes were collected from animals with an unknown or unacceptable health status (24).
References 1. Taft, R. A., Davisson, M. and Wiles, M. V. (2006) Know thy mouse. Trends Genet 22, 649–653. 2. Doetschman, T., Gregg, R. G., Maeda, N., Hooper, M. L., Melton, D. W., Thompson, S., et al. (1987) Targetted correction of a mutant HPRT gene in mouse embryonic stem cells. Nature 330, 576–578. 3. Thomas, K. R., Folger, K. R. and Capecchi, M. R. (1986) High frequency targeting of genes to specific sites in the mammalian genome. Cell 44, 419–428. 4. Waterston, R., Lindblad-Toh, K., Birney, E., Rogers, J., Abril, J., Agarwal, P., et al. (2002) Initial sequencing and comparative analysis of the mouse genome. Nature 420, 520–562. 5. Bogue, M. A., Grubb, S. C., Maddatu, T. P. and Bult, C. J. (2007) Mouse phenome database (MPD). Nucl Acids Res 35, D643–D649.
6. Darwin, C. (1859) On the origin of species by means of natural selection, 1st ed. John Murray, London. 7. Cook Jr, E. H. and Scherer, S. W. (2008) Copy-number variations associated with neuropsychiatric conditions. Nature 455, 919–923. 8. Iafrate, A. J., Feuk, L., Rivera, M. N., Listewnik, M. L., Donahoe, P. K., Qi, Y., et al. (2004) Detection of large-scale variation in the human genome. Nat Genet 36, 949–951. 9. Sincell, M. (2001) Houston flood: research toll is heavy in time and money. Science 293, 589. 10. Fortier, A. L., Lopes, F. L., Darri¨ carrEre, N., Martel, J. and Trasler, J. M. (2008) Superovulation alters the expression of imprinted genes in the midgestation mouse placenta. Hum Mol Genet 17, 1653–1665.
36
Wiles and Taft
11. Wang, Y., Ock, S. A. and Chian, R. C. (2006) Effect of gonadotropin stimulation on mouse oocyte quality and subsequent embryonic development in vitro. Reprod BioMed Online 12, 304–314. 12. Ostermeier, G. C., Wiles, M. V., Farley, J. S. and Taft, R. A. (2008) Conserving, distributing and managing genetically modified mouse lines by sperm cryopreservation. PLoS ONE 3. 13. Whittingham, D. G. (1986) Principles of embryo preservation, (Ashwood-Smith, M. J., Farrant, J., eds.). Low Temperature Preservation in Medicine and Biology. Pitman Medical, Tunbridge Wells, pp. 65–83. 14. Lyon, M. F. (1981) Sensitivity of various germ-cell stages to environmental mutagens. Mutat Res 87, 323–345. 15. Tomlinson, M. (2008) Risk management in cryopreservation associated with assisted reproduction. Cryo Lett 29, 165–174. 16. International Society for, B. and Environmental, R. (2008) Best practices for repositories: Collection, storage, distribution and retrieval of biological materials for research. Cell Preserv Technol 6, 3–58. 17. Byers, S. L., Payson, S. J. and Taft, R. A. (2006) Performance of ten inbred mouse strains following assisted reproductive technologies (ARTs). Theriogenology 65, 1716–1726. 18. Quinn, P., Kerin, J. F. and Warnes, G. M. (1985) Improved pregnancy rate in human in vitro fertilization with the use of a medium
19.
20.
21.
22.
23.
24.
based on the composition of human tubal fluid. Fertil Steril 44, 493–498. Quinn, P. (1995) Enhanced results in mouse and human embryo culture using a modified human tubal fluid medium lacking glucose and phosphate. J Assist Reprod Genet 12, 97–105. Adam, A. A., Takahashi, Y., Katagiri, S. and Nagano, M. (2004) Effects of oxygen tension in the gas atmosphere during in vitro maturation, in vitro fertilization and in vitro culture on the efficiency of in vitro production of mouse embryos. Jpn J Vet Res 52, 77–84. Dumoulin, J. C., Vanvuchelen, R. C., Land, J. A., Pieters, M. H., Geraedts, J. P. and Evers, J. L. (1995) Effect of oxygen concentration on in vitro fertilization and embryo culture in the human and the mouse. Fertil Steril 63, 115–119. Cohen, J., Gilligan, A., Esposito, W., Schimmel, T. and Dale, B. (1997) Ambient air and its potential effects on conception in vitro. Hum Reprod 12, 1742–1749. Hall, J., Gilligan, A., Schimmel, T., Cecchi, M. and Cohen, J. (1998) The origin, effects and control of air pollution in laboratories used for human embryo culture. Hum Reprod 13 Suppl 4, 146–155. Stringfellow, D. A. and Seidel, S. M. (eds.) (1998) Manual of the International Embryo Transfer Society: procedural guide and general information for the use of embryo transfer technology, emphasizing sanitary procedures, 3rd ed. International Embryo Transfer Society, Illinois.
Chapter 3 Genetically Engineered Mouse Models in Drug Discovery Research Rosalba Sacca, Sandra J. Engle, Wenning Qin, Jeffrey L. Stock, and John D. McNeish Abstract Genetically modified mouse models have been proven to be a powerful tool in drug discovery. The ability to genetically modify the mouse genome by removing or replacing a specific gene has enhanced our ability to identify and validate target genes of interest. In addition, many human diseases can be mimicked in the mouse and signaling pathways have been shown to be conserved. In spite of these advantages the technology has limitations. In transgenic animals there may be significant heterogeneity among different founders. In knock-out animals the predicted phenotypes are not always readily observed and occasionally a completely novel and unexpected phenotype emerges. To address the latter and ensure that a deep knowledge of the target of interest is obtained, we have developed a comprehensive phenotyping program which has identified novel phenotypes as well as any potential safety concerns which may be associated with a particular target. Finally we continue to explore innovative technologies as they become available such as RNAi for temporal and spatial gene knock-down and humanized models that may better simulate human disease states. Key words: Drug discovery, knock-out, transgenic, phenotype, humanized mice, knock-in.
1. Introduction The discovery of new drug candidates can be a daunting effort considering the high attrition rate which results in only ∼10% of the compounds tested in clinical trials translating into a successful drug. With the cost of launching a new medicine approaching $1 billion, investment in technologies that impact the assessment of targets and lead compounds early in the drug discovery process is essential in reducing the high attrition rate. G. Proetzel, M.V. Wiles (eds.), Mouse Models for Drug Discovery, Methods in Molecular Biology 602, DOI 10.1007/978-1-60761-058-8 3, © Humana Press, a part of Springer Science+Business Media, LLC 2010
37
38
Sacca et al.
Transgenic technologies have become an important tool in influencing decision making by providing resources that play a key role in target identification, validation of function, as well as in vivo models for providing confidence in the efficacy and safety of the novel therapeutic (genetic versus chemical selectivity and toxicity). The mouse is by far the most widely used genetically modified animal in drug discovery. The availability of extensive information on mouse genetics, advances in molecular biology allowing precise manipulation of the mouse genome, and miniaturization of phenotyping technologies to assess CNS behaviors as well as tracking metabolic changes (1–3) have clearly contributed to establishing genetically modified mice (GeMM) as the single most valuable in vivo resource to modern drug discovery. Many human diseases can be mimicked in the mouse and signaling pathways have been shown to be conserved. Loss of function has been the prominent approach to understanding gene function across numerous laboratory species. The ability to inactivate or modify mammalian genes by homologous recombination in mouse embryonic stem (mES) cells has become a central tool in understanding gene function in association with potential therapeutic targets. Retrospective evaluation of murine knock-out (KO) phenotypes was shown to have a close correlation with the therapeutic effects on the targets of the top 100 selling drugs (4) clearly demonstrating the value of applying KO mice in the validation of novel targets identified through the sequencing of the human genome. Knock-in (KI) technology allowing pre-planned replacement mutations in the endogenous murine alleles has also been used extensively. These replacement modifications range from the change of single base pairs to the replacement of the complete murine gene (5,6). KI technology has allowed the generation of better predictive “humanized” animal models to assess the effect of the compound on the human target and also serve as important in vivo models for screening lead chemical material that is highly species selective. This latter example is often the case with important drug-able target gene families such as G-protein coupled receptors (GPCR). In this chapter, we provide examples of how investigators at Pfizer Inc. are currently utilizing GeMM to advance drug discovery programs, with emphasis on the specific strengths and limitations of the various technologies. Here, we will also present our comprehensive phenotyping program by which KO mice are assessed across various disease areas as a way to maximize the value obtained from these in vivo models. Finally, our current efforts in the generation of human–mouse chimeras as tools to address compound specificity and safety will be described.
Genetically Engineered Mouse Models
2. Applications of Genetically Modified Animals
2.1. Transgenics
39
The use of GeMM by the pharmaceutical industry as a whole has increased substantially over the past two decades. Many factors have influenced this rise in GeMM use within Pfizer Inc. and other companies including a better understanding by scientists of how and when to apply GeMM to discovery programs. Technology improvements including the availability of genomic sequence, PCR genotyping, inbred mouse ES cell lines, karyotype analysis, tetraploid complementation (7,8), and laser-assisted injection of eight-cell embryos (9), all have contributed to increasing the quality of the GeMM models, while decreasing development timelines. Shorter timelines provide models more in sync with drug discovery timelines thus increasing their potential to impact decision making. Arguably, the advancement in molecular biology using Red/ET Recombineering (10) that allows for rapid genetic engineering of highly sophisticated transgenic DNA vectors has had a significant impact in the reduction of GeMM development timelines. Recombineering makes nearly any genetic modification a realistic undertaking by allowing the development of the most appropriate GeMM model based on the needs of a drug discovery program. Greater accessibility to GeMM models has also been a factor for their increased use in drug discovery. In addition to internal efforts, Pfizer Inc. also uses external resources to obtain GeMM models including academic institutions The Jackson Laboratories Induced Mutant Resource (Bar Harbor, ME.), contract research organizations, such as Xenogen Corporation (acquired by Caliper Life Sciences; Alameda, CA.), and private collaborations with companies such as Lexicon Pharmaceuticals, Inc. (The Woodlands, TX.) and Deltagen (San Mateo, CA.). Transgenic mice created by the microinjection of purified linear DNA into one-celled, pronuclear stage embryos were first described by Gordon and Ruddle in 1980 (11). By 1982, Brinster and Palmiter demonstrated that precise DNA constructs could be developed that allowed tissue-specific transgene expression (12) and ushered in the era of gain of function modification of the mouse genome. To this day, the techniques used in the development of pronuclear transgenic mice are similar to those described in 1980. The methodologies and technical improvements for the generation of transgenic mice have been previously reviewed (13). In the early 1990s, Pfizer Inc. scientists began to develop pronuclear transgenic mice not only to validate targets and pathways but also to develop in vivo models of human pathophysiology and disease that recapitulate clinical conditions for the testing of novel therapeutic strategies.
40
Sacca et al.
An early, yet innovative approach to validating therapeutically relevant pathways using pronuclear transgenic mice was the expression of human Glucose Transporter 4 (GLUT4) gene in an inbred model of diabetes (db/db). These results demonstrated that expression of the GLUT4 transgene resulted in the alleviation of insulin resistance and significant improvement in glycemic control (14). A key area of unmet medical need is the identification of new therapies for Alzheimer’s disease (AD). AD is characterized by the deposit of neurofibrillary tangles and amyloid plaques in the brain. However, discovery of new medicines for AD is hindered by the lack of in vivo models that spontaneously display these pathologic hallmarks of the disease. Probably no area of transgenic disease model development has been more rigorously pursued than AD models (15). An important example is the Tg2576 transgenic mouse (16), one of the most widely used Alzheimer’s disease model across the world. This transgenic model of AD develops amyloid plaques by 9–11 months. To shorten the timeframe for plaque development a second transgenic was created using the neuron-specific Thy-1 promoter expressing a human presenilin G384A mutation, associated with early onset AD (17). Mice harboring both the Tg2576 and the G384A transgenes develop plaque as early as 6 months. Another example of a GeMM transgenic model displaying clinical like conditions including cardiac hypertrophy, fibrosis, and heart failure is the ␣-myosin heavy chain promoter 11hydroxysteroid dehydrogenase type 2 transgenic mouse (18). Eplerenone (INSPRATM , Pfizer Inc.), a selective aldosterone blocker, was tested in this model and proved to ameliorate the phenotype. This model revealed the deleterious consequences of inappropriate mineralocorticoid receptor activation in the heart and supported the notion that aldosterone blockade may provide additional therapeutic benefit in the treatment of heart failure. Beyond their use as models of human disease, transgenics are widely used as GeMM “tool” lines. Tissue-specific and/or ubiquitous Cre recombinase transgenic lines are commonly used for conditional knock-outs and excision of floxed antibiotic selection cassettes. When temporal transgene expression is essential, the tetracycline-inducible expression system has been utilized to study the acute effects of transgene expression to circumvent developmental compensation or lethality due to transgene expression during development. The development and application of inducible and/or conditional transgenic mice have previously been reviewed (19,20). Transgenic models have a broad spectrum of utilities across drug discovery, but they also have their limitations. Because transgene integration is random, occasionally the presence of multiple insertion sites, varied expression levels, and tissue distribution
Genetically Engineered Mouse Models
41
results in significant heterogeneity among founder lines and subsequent generations. This variability results in the necessity to thoroughly evaluate and characterize multiple mouse lines and although rare, the transgene integration may also disrupt an endogenous gene resulting in insertional mutagenesis. In addition, the expression of transgenes is at times reduced or lost after many generations of breeding, presumably due to epigenetic modification of the inserted transgenic DNA sequences (21). This can be avoided by cryopreservation of important transgenic mouse lines in their earliest generations. 2.2. Knock-outs
Since the isolation of mES cells in 1985 (22) and the subsequent ability to direct pre-planned mutagenesis of the murine genome by homologous recombination in the mES cells (23–25), investigators have been using this seminal technology to address all aspects of mammalian biology. Over the last 20 years, thousands of KO mice have been developed worldwide providing drug discovery scientists’ access to this critical resource to investigate the role of potential drug targets. KO mice have become an invaluable tool for determining target function, selectivity, and potential toxicity liabilities (26). KO mice within Pfizer Inc. have been utilized numerous times to validate or nullify hypotheses regarding confidence in a novel gene’s rationale as a potential therapeutic target. To improve the decision making regarding target validation using KO mice, we developed mES cells from the DBA/1lacJ mouse strain, an inbred mouse model of collagen-induced arthritis. This DBA mES cell line enabled us to evaluate the effect of single gene deletions on the resulting inflammatory phenotype. For example, KO mice for the 5-lipoxygenase-activating protein (FLAP) gene resulted in marked reduction in the collagen-induced arthritis compared to inbred, control littermates (Fig. 3.1) and provided in vivo data implicating FLAP and the leukotriene pathway in inflammation (27). The KO mouse provides an in vivo model with 100% inhibition of the target to understand differences in selectivity between genetic and pharmacological inhibition. An excellent example of which is the phosphodiesterase 10a (Pde10a) KO mouse. PDE10a emerged as a potential target for the treatment of psychosis based on conserved mouse and human striatal brain expression data, as well as activity of inhibitors in rodent models predictive of clinical antipsychotic activity. Efficacy of these inhibitors was absent in Pde10a KO mice, confirming that PDE10a inhibition underlies the antipsychotic-like activity of these compounds (28,29). KO mice are useful as models for important clinical conditions such as is the case with the Apolipoprotein E (ApoE) KO, a robust atherosclerosis model (30,31). Not only has the ApoE KO mouse been extensively used as a single GeMM disease model but also
42
Sacca et al.
Fig. 3.1. Collagen-induced arthritis in FLAP-deficient mice. DBA/1 mice were immunized with chick type II collagen on day 0 and 21. IL-1 was administered subcutaneously on days 45 and 46 to trigger an arthritic flare. Disease severity was scored by observation of the paws for redness and swelling. Open circles, +/+ mice; closed triangles, +/− mice; open squares, −/− mice. Results are mean ± SEM, n = three experiments.
additional genetic modifications have been introduced into the ApoE null background for evaluating the role of these new genes in atherosclerosis pathophysiology. KO mice have also been used to investigate whether a toxicity finding is the result of the chemical lead or has a mechanistic basis. 5-Lipoxygenase KO mice were used to prove that a chemical series of leukotriene B4 antagonists was responsible for the induction of hepatic enzymes, not an associated phenotype of leukotriene-deficient KO mice (32). Although KO mice are extremely useful as in vivo models for drug discovery scientists, they are not without their limitations. Because the gene inactivation is produced in the zygote, embryonic lethality is a reality for genes essential in development. Redundancy or compensatory effects due to altered expression of unmodified gene family members may also occur, all of which can complicate interpretation of the KO phenotype. The use of a tissue-specific conditional KO is one way to circumvent these issues, and as discussed earlier, many tissue-specific Cre tool GeMM lines are available for generating this type of model. Another option is to create an inducible KO using the
Genetically Engineered Mouse Models
43
tamoxifen-inducible Cre-ER system (33,34) allowing for temporal control of gene knock-down in all tissues of adult animals. This approach provides an opportunity for drug discoverer’s to more robustly reproduce the effect of a therapeutic molecule. However, these multiple genetic modifications substantially increase the time and costs for GeMM model delivery, making these complex approaches a consideration when initiating a new project. 2.3. Knock-In
Just as KO mice have been utilized to address a variety of needs for drug discovery scientists, KI models have also had many different applications to drug discovery programs. KI modifications can range from single base changes to complete replacement of a murine gene with an orthologous human gene. The need for more predictive in vivo models to evaluate a lead compound’s efficacy, pharmacokinetic profile, and toxicological properties has made humanized KI mice an important tool. For example, chemokine receptor-1 (CCR1) humanized KI mice were used to assess the activity of human-specific CCR1 antagonists and their ability to modulate inflammatory responses (35). In this model the entire mouse CCR1 gene was replaced with the CCR1 human ortholog. A complete gene replacement is not always necessary for the “humanization” of a target gene as illustrated by a model generated for the Thrombopoietin receptor (TPOr) program. Pfizer Inc. identified agonists for the thrombopoietin receptor (TPOr) that were shown to have selective human receptor binding. In vitro mouse/human receptor constructs identified a region of the protein responsible for this species specificity. A humanized TPOr KI mouse was developed in which only mouse exons 8-10, the region responsible for the specificity, were replaced with human exons 8-10. When treated with TPOr agonists, TPOr KI mice demonstrated a dosedependent increase in platelet numbers as compared to their wildtype littermates (Fig. 3.2, manuscript in preparation). Knock-in models have also been used to introduce point mutations to create mouse models of human disease alleles (36). Point mutant KI mice can also be used to inactivate essential regions of a protein, such as a catalytic kinase domain, or to alter a drug-binding site within the target protein. The ␣2␦1 R217A KI mouse was used to demonstrate the mechanism of action for the analgesic effects of gabapentin and pregabalin, showing that it is mediated by an interaction with the ␣2␦1 subunit, specifically the arginine amino acid at position 217 (37). Another important application of KI mice is the creation of fluorescent and bioluminescent reporter models which have a broad spectrum of utility. These models can be developed to tag specific genes, cell lineages, and biological pathways. In this section, we have tried to review a broad crosssection of genetically engineered mouse models and some specific
44
Sacca et al.
Fig. 3.2. The graph compares TPOr knock-in and wild-type responses to PF-356316902. Mice were dosed with PF-3563169-02 s.c. for two consecutive days and blood was analyzed on day 5. Each group is represented by five mice.
applications aligned with real drug discovery programs from the hundreds of possible examples. Clearly, the technologies for manipulating the murine genome allow for highly innovative genomic modifications limited only by one’s imagination. Most importantly, understanding the needs of a drug discovery project will help determine the appropriate genetically modified mouse to develop, as no one model type is appropriate for every situation. For example, a standard KO mouse may provide relatively quick and valuable information about mechanisms and pathways for a novel target with little background literature. On the other hand, a conditional or inducible KO may be essential to understanding the biology of a more mature target with known embryonic lethality or complex, interdependent phenotypes. Humanized knock-in mice are now becoming essential for developing biotherapeutics. Once an understanding of the needs of the project is attained and the most appropriate GeMM model is developed, then the focus is on assessing its phenotype.
3. Comprehensive Phenotyping of Genetically Modified Mouse Models
The greatest challenge to understanding targets in drug discovery is assigning in vivo function to the genes of interest and to do so in an efficient and cost-effective manner. As discussed, KO mice provide a powerful approach for defining gene function in the context of mammalian physiology (4,38) and are usually generated in response to a particular hypothesis. When the KO model confirms the predicted hypothesis, there is often little incentive
Genetically Engineered Mouse Models
45
to evaluate the model in additional assays in the rush to progress a drug discovery program. When evaluation of the KO mouse does not confirm the predicted phenotype or the mouse has an unexpected or unclear phenotype, the model may be set aside. In our experience, this is relatively common phenomenon as almost half of all KO mice generated do not present with the expected phenotype. In the past, KO mice are rarely examined in diverse physiological assays to comprehensively define the function of the gene, therefore failing to capitalize on the significant investment made in the target or the generation and initial characterization of the KO model. Phenotype Pfinder was developed as a tool for systematically phenotyping KO mice to maximize on our investment in the KO model and underlying gene target. At its inception Phenotype Pfinder was envisioned to be a panel of assays that combined speed and efficiency to deliver high-quality assessment of a broad spectrum of physiological parameters that would promote drug discovery. Individual assays (Table 3.1) were required to meet several criteria for inclusion in the program. All assays were expected to be decision making for the respective therapeutic area such that when the model returns a statistically significant finding, the therapeutic area would followup on the finding. On the other hand, if the results showed no significant difference between wild-type (WT) and KO mice, scientists were expected to have enough confidence in the results to discontinue efforts in that area. Biomedical statisticians were engaged to ensure that each assay used enough mice to identify statistically significant differences between WT and KO mice and that assays used the appropriate age and gender of mice. To ensure that each assay can detect either improved or impaired function, assays are required to be validated with positive and negative controls using either pharmacological or genetic validation strategies. Additionally, assays were tested for reproducibility and robustness across the variety of mouse background strains (C57BL/6, DBA/1lacJ, 129.B6) commonly used to generate KO models. Only those assays that were independent of the background strain were included. Highly labor-intensive assays (for example, those requiring surgical intervention) or time-intensive assays (for example, aging paradigms or oncology assays) were excluded from the program in order to manage cost and time to delivery of data. To further manage cost and efficiency, all assays were consolidated into a single platform at a single site and performed by a dedicated group of in vivo biologists. Each assay was validated independently and followed by testing of small combinations of assays. Those combinations for which results differed significantly from na¨ıve control data were eliminated. Successful combinations were assembled into larger panels and revalidated. Panels were combined and rigorously repeated to ensure
46
Sacca et al.
Table 3.1 Assays evaluated in Phenotype Pfinder Immunology/inflammation
Thrombosis
B-cell proliferation
Collagen-induced platelet aggregation
CD4 T-cell proliferation
Ex vivo coagulation
Cytokine production by CD4 T cells
Tail transection bleeding time
Delayed-type hypersensitivity
Thrombin–antithrombin III complex
FACS analysis of splenocytes
Frailty and osteoporosis
In vivo antibody production
Body composition by DEXA
KRN serum-induced arthritis
Calcein labeling for bone histopathology
LPS induced inflammation
Excised bone DEXA
Pulmonary inflammation
Rotarod
Thioglycollate-induced monocyte infiltration
Selected muscle weights
Gastrointestinal function
Dermatology
Colon:body weight ratio
Water retention test for sebum production
Longitudinal body weight assessment
Pain
Diabetes and obesity
Formalin-induced pain
Body composition by DEXA
Hot plate
Corticosterone level
Von Frey
Food intake
Neurodegeneration
Free fatty acids
Neurodegeneration: excitotoxicity
Leptin level
Neurodegeneration: lactacystin
Long-term monitoring in comprehensive cage monitoring system for metabolism and activity
Neurodegeneration: neurite outgrowth
Pre- and post-high-fat diet
Neurodegeneration: vitamin K
Adiponectin levels
Psycotherapeutics
Cholesterol distribution
Elevated plus maze
Insulin level
Irwin behavior test
Oral glucose tolerance test
Open field
Total cholesterol level
Pentylenetetrazole-induced seizure
Phospholipids B level
Startle and pre-pulse inhibition
Body composition by DEXA
Tail suspension
Regional fat depot assessment by CT
Sexual health
Weekly body weight
Female contact sexual behavior
Tissue glycogen level
Forced erection test
Cardiovascular function
Male contact sex behavior (continued)
Genetically Engineered Mouse Models
47
Table 3.1 (continued) Blood pressure
Male fertility
Heart rate
Testosterone
Angiotensin II and aldosterone levels
Cross-therapeutic assays
Bladder and renal function
Automated blood chemistries
Diuretic-induced stress test
Hematology
Urine chemistries
that upstream, less invasive assays did not interfere with or alter the response to the increasingly more invasive downstream assays. The comprehensive Phenotype Pfinder protocol evaluates 50 KO and 50 WT mice through a sequential combination of ∼50 assays in 15 disease areas in 17 weeks. If done individually and not at a centralized facility, these assays would consume more than 800 mice and require substantially more time and money for breeding, genotyping, and labor. To date, more than 100 unique KO mouse lines have been characterized in the Phenotype Pfinder protocol and have impacted numerous drug discovery projects. Raw data generated by the Phenotype Pfinder protocol is maintained in a dedicated database. The database automatically graphs the data, generates statistical analyses of assays, and flags those assays which show a statistically significant difference between WT and KO mice. The data can be searched and viewed by either specific KO line or assay. The database also offers a tool for analysis of historically accumulated control data sorted by genetic background. Information generated by Phenotype Pfinder has been used in a variety of ways. At its core, Phenotype Pfinder tests the hypothesis that a gene may be a previously unsuspected target in novel therapeutic areas and ensures that we would identify the serendipitous finding. For example, the previously mentioned Pde10a KO mice were generated at the request of the psychotherapeutics program. The KO line was evaluated in Phenotype Pfinder as part of a strategic decision to target specific gene families. Unexpectedly, the KO mice were completely resistant to weight gain on a highfat diet . Furthermore the KO mice showed increase basal energy expenditure and no change in food intake compared to WT mice. When WT mice on a high-fat diet were treated with a Pde10aspecific inhibitor, they consumed as much food as the vehicletreated cohort but failed to gain weight. KO mice treated with the Pde10a inhibitor also consumed as much food as the WT cohort and showed no additional weight loss thus strongly suggesting that the failure to gain weight on a high-fat diet was specifically
48
Sacca et al.
related to the loss of the Pde10a. By identifying a relevant phenotype in a different therapeutic area, Phenotype Pfinder allowed scientists to leverage the investment in knowledge, reagents, and tool compounds to rapidly explore a new drug discovery program. Conversely, the broad phenotypic picture afforded by Phenotype Pfinder can minimize investment in targets with previously unsuspected liabilities. HCN1 is a hyperpolarization-activated cation channel. It is known to play a role in the regulation of cell excitability and contributes to spontaneous rhythmic activity in both heart and brain. Data suggested that HCN1 may be a target for epilepsy and neuropathic pain. As hypothesized, Hcn1 KO mice display increased sensitivity to pentylenetetrazole (PTZ)-induced clonus seizures (39). Hcn1 KO mice also showed significant impairments in gait, body posture, coordination, and locomotion as well as changes in body composition, metabolism, bladder function, heart rate, and blood coagulation. The constellation and severity of the phenotypes reduced the attractiveness of the target. Phenotype Pfinder data can also build strategic information for hypothesis generation. G-protein coupled receptors (GPRs) are a class of protein that historically has been very amenable to drug activation or inhibition. Prior to the identification of a ligand for a GPR, so-called orphan status, it is often difficult to ascertain the GPR’s specific function. In 2005 when Gpr35 KO mice were evaluated in Phenotype Pfinder, the “orphan” GPR had been associated with expression in gastric cancer cells, a chromosomal deletion found in Albright hereditary osteodystrophy-like (AOH) syndrome and type II diabetes in a Mexican-American population (40–42). Male Gpr35 KO mice consumed less food per unit body weight, had higher leptin levels, a lower metabolic rate, and tended to be heavier than WT mice. These phenotypes support the association of GPR35 with an obesity and diabetes phenotype. No bone phenotypes were detected suggesting that deletion of GPR35 is not responsible for the bone phenotypes in AOH syndrome. Because Phenotype Pfinder is time constrained, no gastric cancers were detected. Subsequent work has identified kynurenic acid, an endogenous metabolite of tryptophan, and Zaprinast, a well-known cyclic guanosine monophosphate-specific phosphodiesterase inhibitor, as ligands for GPR35 (43,44). Recent work has shown that levels of kynurenic acid in rat intestine are sufficient to affect GPR35 thus supporting a link between GPR35, gastrointestinal function and a potential role in obesity and/or diabetes (45). Occasionally, Phenotype Pfinder can identify new pharmacological models. Map3K11 KO mice were reported to be healthy and viable (46). The only reported phenotype in the literature was a defect in the ability of embryonic fibroblasts from the
Genetically Engineered Mouse Models
49
Map3K11 KO to activate JNK in response to TNF␣. Studies using a Map3k11 inhibitor suggested that inhibition of JNK activity improved neuron survival and growth (47). Phenotype Pfinder studies showed no difference between WT and KO neurons in their response to glutamate or vitamin K-induced neurotoxicity. However, both before and after the Map3K11 KO mice were fed a high-fat diet, they showed statistically elevated systolic blood pressure compared to WT mice. When treated with an FDAapproved angiotensin I converting enzyme (ACE) inhibitor, the blood pressure of the Map3k11 KO mice fell to WT control levels (O. Buiakova, personal communication). This presents a new mouse model of high blood pressure which can be evaluated with pharmacological agents to determine clinical relevance. Systematically evaluating KO mice across a panel of diverse assays provides a cost-effective, time constrained method for maximizing the collection of information. It can lead to the identification of new phenotypes and therapeutic indications, confirm literature-reported phenotypes, and hypothesized effects, highlight potential target related safety effects so that they can be monitored or managed and draw attention to potential clinical biomarkers. Assembly of the data into an easily accessible, searchable, dedicated database allows the information to be continuously referenced and evaluated as new information becomes available. With the ever-increasing cost of drug discovery programs, it is essential to identify the most promising targets early, to terminate nonviable project quickly, and to exploit all available therapeutic applications around a single target. The identification of one high value, novel therapeutic application can justify the entire expense of a comprehensive KO mouse phenotyping program. It maximizes the investment in any one model and target and promotes better in vivo biology.
4. Future Directions of Genetically Engineered Models in Drug Discovery
With 99% of the mouse genes having homologues in humans (48), and the ability to readily manipulate the mouse genome, mice rightfully have been an important model organism to understand human physiology and diseases. However, the protein sequences of these homologues may differ from their human counterparts such that they do not bind compounds or biotherapeutic agents targeted against the human counterparts. In addition, mice differ from humans in many aspects, including reproduction, olfaction, behavior, metabolism, and immunity. Thus, to robustly model the human condition, it would be desirable to, at times, replace the mouse homologue with the entire human gene,
50
Sacca et al.
the entire gene family, or an entire organ/system. Some specific examples of variation of “humanization” will be described here. Deleting the mouse genes and replacing them with a human gene family residing over a million base pairs of genomic sequence has been accomplished and is exemplified by the creation of mouse models carrying the human immunoglobulin genes. The first fully human antibody directed against epidermal growth factor receptor (EGFR), panitumumab (VECTIBIX, Amgen Inc., Thousand Oaks, CA.), was generated by utilizing the XenoMouse of Abgenix Incorporated (Fremont, CA.) and recently won regulatory approval as treatment for advanced colorectal cancer (49). Several other important candidate gene families for humanization are the MHC loci for a human adaptive immune response for vaccine testing and the cytochorme P450 genes for drug metabolism. Single gene replacements from these gene families have been reported (50,51). In the last 20 years, naturally occurring mouse mutants were identified and genetically engineered mouse models created that allow the design and creation of sophisticated immunocompromised murine hosts that can support chimerism with cells from other species (52). Employing this new generation of immune compromised host mice (53), it is now routinely possible to achieve up to 70% chimerism of human CD45+ cells in the bone marrow and peripheral blood from these bone marrow engrafted models (54,55). Although these achievements are encouraging, major limitations must be overcome before the model can be used for testing and developing human vaccines. Specific technology improvements that are needed include humanized MHC genes to improve human T-cell selection and maturation in mouse thymus, retaining the Peyer’s patches and lymph nodes in these models to allow antigen presentation that normally occurs in these primary lymphoid organs and human cytokines critical for differentiation and maturation of human immune cells in a mouse host. These are daunting tasks but if successful will considerably extend the use of animal models for drug testing. In drug discovery, it is important to understand how a drug is metabolized in humans. In both the urokinase-type plasminogen activator (uPA) transgenic (56) and the fumarylacetoacetate hydrolase (Fah) knock-out models (57) bred onto an immunocompromised mouse background, liver engraftment in the resulting chimera can be achieved up to 90% with human hepatocytes. In addition to human hepatocytes becoming integrated into the recipient liver, they also express human liver-specific genes (CYP1A2 and 3A4) and respond to prototypical inducers of the cytochrome P450 genes. However, these models currently rely on organ donor or cadaver for the supply of human hepatocytes. This presents a major limitation in reproducibility in these human liver
Genetically Engineered Mouse Models
51
chimera models, as the donor hepatocytes likely differ in genetic background, age, environmental exposure during the course of life, and whether isolated fresh or revived from cryopreservation (57). In this respect, human embryonic stem cells or induced pluripotent stem (iPS) cells hold the promise to derive an unlimited supply of genetically homogeneous population of hepatocytes for engraftment (58). The goal for cell therapy is to restore or replace damaged cells by mobilizing host cells or by providing terminally differentiated cell types for diseases such as type 1 diabetes, Parkinson’s disease, Alzheimer’s disease, myocardial infarction, or muscle dystrophy. In each case, it would be desirable to have a relevant animal model to follow the mobilization or engraftment, migration, and differentiation of the human cells in the target organs in the host. This scheme will likely necessitate an immunodeficient background to avoid xenograft rejection, labeling of the donor cells to allow tracking and an in vivo environment to allow structural and functional integration of these cells. Genetically engineered models, combined with cell and tissue transplantation, will undoubtedly generate the next generation of humanized mice that promises to translate into better preclinical models for drug discovery and regenerative medicine research and development.
5. Conclusion In the decade of the 1980s, a number of powerful and reproducible technologies were described which allowed the precise manipulation of the murine genome. These methods resulted in the evaluation of gene function by classical approaches of gain and loss of function mutagenesis. In addition, homologous recombination technologies allowed for targeted replacement or humanized knock-ins into the murine genome, resulting in the development of human alleles in mouse models. Although, these technologies are of great importance to basic research, drug discoverers soon recognized that the application of transgenic technologies could deliver novel approaches to in vivo target validation and model development aimed at better decisions for the discovery of new medicines. In this chapter, we have selected a few of the many examples of pronuclear transgenic mice and mES cell-derived knock-out and knock-in mice that have been developed at Pfizer Inc. or by other researchers for specific applications from therapeutic idea generation to hypothesis testing. The ability to derive or obtain GeMM has become routine for most academic or applied laboratories; however, the ability to fastidiously evaluate the phenotype of the GeMM is less well refined.
52
Sacca et al.
Here, our efforts to ensure the careful evaluation of the primary KO mouse phenotypes and also maximize the return of the KO mouse investment by comprehensively evaluating for unsuspected phenotypes was described. The central aim of this chapter was to illustrate the types of problem statements and challenges that face investigators, and the application of innovative approaches using transgenic technologies that assist in our decision making. Improvements in technologies have always guided our application of GeMM to drug discovery. In the future, new technologies such as cellular reprogramming may lead to pluripotent, germline competent stem cells that allow genetic modification in other lab species and the expression of RNAi to introduce temporal or spatial gene knock-down that more closely resembles the effect of a drug. References 1. van der Staay, F. J. and Steckler, T. (2001) Behavioural phenotyping of mouse mutants. Behav Brain Res 125, 3–12. 2. Crawley, J. N. (1999) Behavioral phenotyping of transgenic and knockout mice: experimental design and evaluation of general health, sensory functions, motor abilities, and specific behavioral tests. Brain Res 835, 18–26. 3. Doevendans, P. A., Daemen, M. J., de Muinck, E. D. and Smits, J. F. (1998) Cardiovascular phenotyping in mice. Cardiovasc Res 39, 34–49. 4. Zambrowicz, B. P. and Sands, A. T. (2003) Knockouts model the 100 best-selling drugs– will they model the next 100? Nat Rev Drug Discov 2, 38–51. 5. Forlino, A., Porter, F. D., Lee, E. J., Westphal, H. and Marini, J. C. (1999) Use of the Cre/lox recombination system to develop a non-lethal knock-in murine model for osteogenesis imperfecta with an alpha1(I) G349C substitution. Variability in phenotype in BrtlIV mice. J Biol Chem 274, 37923– 37931. 6. Reichardt, H. M., Kaestner, K. H., Tuckermann, J., Kretz, O., Wessely, O., Bock, R., et al. (1998) DNA binding of the glucocorticoid receptor is not essential for survival. Cell 93, 531–541. 7. Nagy, A., Rossant, J., Nagy, R., AbramowNewerly, W. and Roder, J. C. (1993) Derivation of completely cell culture-derived mice from early-passage embryonic stem cells. Proc Natl Acad Sci U S A 90, 8424–8428. 8. Nagy, A., Gocza, E., Diaz, E. M., Prideaux, V. R., Ivanyi, E., Markkula, M., et al. (1990) Embryonic stem cells alone are able to sup-
9.
10.
11.
12.
13. 14.
15.
16.
port fetal development in the mouse. Development 110, 815–821. Poueymirou, W. T., Auerbach, W., Frendewey, D., Hickey, J. F., Escaravage, J. M., Esau, L., et al. (2007) F0 generation mice fully derived from gene-targeted embryonic stem cells allowing immediate phenotypic analyses. Nat Biotechnol 25, 91–99. Zhang, Y., Muyrers, J. P., Testa, G. and Stewart, A. F. (2000) DNA cloning by homologous recombination in Escherichia coli. Nat Biotechnol 18, 1314–1317. Gordon, J. W., Scangos, G. A., Plotkin, D. J., Barbosa, J. A. and Ruddle, F. H. (1980) Genetic transformation of mouse embryos by microinjection of purified DNA. Proc Natl Acad Sci U S A 77, 7380–7384. Palmiter, R. D., Chen, H. Y. and Brinster, R. L. (1982) Differential regulation of metallothionein-thymidine kinase fusion genes in transgenic mice and their offspring. Cell 29, 701–710. Wall, R. J. (2001) Pronuclear microinjection. Cloning Stem Cells 3, 209–220. Gibbs, E. M., Stock, J. L., McCoid, S. C., Stukenbrok, H. A., Pessin, J. E., Stevenson, R. W., et al. (1995) Glycemic improvement in diabetic db/db mice by overexpression of the human insulin-regulatable glucose transporter (GLUT4). J Clin Invest 95, 1512–1518. Higgins, G. A. and Jacobsen, H. (2003) Transgenic mouse models of Alzheimer s disease: phenotype and application. Behav Pharmacol 14, 419–438. Hsiao, K., Chapman, P., Nilsen, S., Eckman, C., Harigaya, Y., Younkin, S., et al. (1996) Correlative memory deficits, Abeta elevation,
Genetically Engineered Mouse Models
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
and amyloid plaques in transgenic mice. Science 274, 99–102. Cruts, M., Backhovens, H., Wang, S. Y., Van Gassen, G., Theuns, J., De Jonghe, C. D., et al. (1995) Molecular genetic analysis of familial early-onset Alzheimer s disease linked to chromosome 14q24.3. Hum Mol Genet 4, 2363–2371. Qin, W., Rudolph, A. E., Bond, B. R., Rocha, R., Blomme, E. A., Goellner, J. J., et al. (2003) Transgenic model of aldosteronedriven cardiac hypertrophy and heart failure. Circ Res 93, 69–76. Belteki, G., Haigh, J., Kabacs, N., Haigh, K., Sison, K., Costantini, F., et al. (2005) Conditional and inducible transgene expression in mice through the combinatorial use of Cre-mediated recombination and tetracycline induction. Nucleic Acids Res 33, e51. Bockamp, E., Maringer, M., Spangenberg, C., Fees, S., Fraser, S., Eshkind, L., et al. (2002) Of mice and models: improved animal models for biomedical research. Physiol Genomics 11, 115–132. Sutherland, H. G., Kearns, M., Morgan, H. D., Headley, A. P., Morris, C., Martin, D. I., et al. (2000) Reactivation of heritably silenced gene expression in mice. Mamm Genome 11, 347–355. Doetschman, T. C., Eistetter, H., Katz, M., Schmidt, W. and Kemler, R. (1985) The in vitro development of blastocyst-derived embryonic stem cell lines: formation of visceral yolk sac, blood islands and myocardium. J Embryol Exp Morphol 87, 27–45. Koller, B. H., Hagemann, L. J., Doetschman, T., Hagaman, J. R., Huang, S., Williams, P. J., et al. (1989) Germ-line transmission of a planned alteration made in a hypoxanthine phosphoribosyltransferase gene by homologous recombination in embryonic stem cells. Proc Natl Acad Sci U S A 86, 8927–8931. Thomas, K. R. and Capecchi, M. R. (1987) Site-directed mutagenesis by gene targeting in mouse embryo-derived stem cells. Cell 51, 503–512. Doetschman, T., Maeda, N. and Smithies, O. (1988) Targeted mutation of the Hprt gene in mouse embryonic stem cells. Proc Natl Acad Sci U S A 85, 8583–8587. Harris, S. (2001) Transgenic knockouts as part of high-throughput, evidence-based target selection and validation strategies. Drug Discov Today 6, 628–636. Griffiths, R. J., Smith, M. A., Roach, M. L., Stock, J. L., Stam, E. J., Milici, A. J., et al. (1997) Collagen-induced arthritis is reduced in 5-lipoxygenase-activating protein-deficient mice. J Exp Med 185, 1123–1129.
53
28. Siuciak, J. A., Chapin, D. S., Harms, J. F., Lebel, L. A., McCarthy, S. A., Chambers, L., et al. (2006) Inhibition of the striatum-enriched phosphodiesterase PDE10A: a novel approach to the treatment of psychosis. Neuropharmacology 51, 386–396. 29. Siuciak, J. A., McCarthy, S. A., Chapin, D. S., Fujiwara, R. A., James, L. C., Williams, R. D., et al. (2006) Genetic deletion of the striatum-enriched phosphodiesterase PDE10A: evidence for altered striatal function. Neuropharmacology 51, 374–385. 30. Plump, A. S., Smith, J. D., Hayek, T., AaltoSetala, K., Walsh, A., Verstuyft, J. G., et al. (1992) Severe hypercholesterolemia and atherosclerosis in apolipoprotein E-deficient mice created by homologous recombination in ES cells. Cell 71, 343–353. 31. Zhang, S. H., Reddick, R. L., Piedrahita, J. A. and Maeda, N. (1992) Spontaneous hypercholesterolemia and arterial lesions in mice lacking apolipoprotein E. Science 258, 468–471. 32. Beierschmitt, W. P., McNeish, J. D., Griffiths, R. J., Nagahisa, A., Nakane, M. and Amacher, D. E. (2001) Induction of hepatic microsomal drug-metabolizing enzymes by inhibitors of 5-lipoxygenase (5-LO): studies in rats and 5-LO knockout mice. Toxicol Sci 63, 15–21. 33. Seibler, J., Zevnik, B., Kuter-Luks, B., Andreas, S., Kern, H., Hennek, T., et al. (2003) Rapid generation of inducible mouse mutants. Nucl Acids Res 31, e12. 34. Hayashi, S. and McMahon, A. P. (2002) Efficient recombination in diverse tissues by a tamoxifen-inducible form of Cre: a tool for temporally regulated gene activation/inactivation in the mouse. Dev Biol 244, 305–318. 35. Gladue, R. P., Cole, S. H., Roach, M. L., Tylaska, L. A., Nelson, R. T., Shepard, R. M., et al. (2006) The human specific CCR1 antagonist CP-481,715 inhibits cell infiltration and inflammatory responses in human CCR1 transgenic mice. J Immunol 176, 3141–3148. 36. Guo, Q., Fu, W., Sopher, B. L., Miller, M. W., Ware, C. B., Martin, G. M., et al. (1999) Increased vulnerability of hippocampal neurons to excitotoxic necrosis in presenilin-1 mutant knock-in mice. Nat Med 5, 101–106. 37. Bian, F., Li, Z., Offord, J., Davis, M. D., McCormick, J., Taylor, C. P., et al. (2006) Calcium channel alpha2-delta type 1 subunit is the major binding protein for pregabalin in neocortex, hippocampus, amygdala, and spinal cord: an ex vivo autoradiographic study
54
38.
39.
40.
41.
42.
43.
44.
45.
46.
47.
Sacca et al. in alpha2-delta type 1 genetically modified mice. Brain Res 1075, 68–80. Zambrowicz, B. P., Turner, C. A. and Sands, A. T. (2003) Predicting drug efficacy: knockouts model pipeline drugs of the pharmaceutical industry. Curr Opin Pharmacol 3, 563–570. Zong, X., Eckert, C., Yuan, H., WahlSchott, C., Abicht, H., Fang, L., et al. (2005) A novel mechanism of modulation of hyperpolarization-activated cyclic nucleotidegated channels by Src kinase. J Biol Chem 280, 34224–34232. Vander Molen, J., Frisse, L. M., Fullerton, S. M., Qian, Y., Del Bosque-Plata, L., Hudson, R. R., et al. (2005) Population genetics of CAPN10 and GPR35: implications for the evolution of type 2 diabetes variants. Am J Hum Genet 76, 548–560. Okumura, S., Baba, H., Kumada, T., Nanmoku, K., Nakajima, H., Nakane, Y., et al. (2004) Cloning of a G-protein-coupled receptor that shows an activity to transform NIH3T3 cells and is expressed in gastric cancer cells. Cancer Sci 95, 131–135. Shrimpton, A. E., Braddock, B. R., Thomson, L. L., Stein, C. K. and Hoo, J. J. (2004) Molecular delineation of deletions on 2q37.3 in three cases with an Albright hereditary osteodystrophy-like phenotype. Clin Genet 66, 537–544. Wang, J., Simonavicius, N., Wu, X., Swaminath, G., Reagan, J., Tian, H., et al. (2006) Kynurenic acid as a ligand for orphan G protein-coupled receptor GPR35. J Biol Chem 281, 22021–22028. Taniguchi, Y., Tonai-Kachi, H. and Shinjo, K. (2006) Zaprinast, a well-known cyclic guanosine monophosphate-specific phosphodiesterase inhibitor, is an agonist for GPR35. FEBS Lett 580, 5003–5008. Kuc, D., Zgrajka, W., Parada-Turska, J., Urbanik-Sypniewska, T. and Turski, W. A. (2008) Micromolar concentration of kynurenic acid in rat small intestine. Amino Acids. Brancho, D., Ventura, J. J., Jaeschke, A., Doran, B., Flavell, R. A. and Davis, R. J. (2005) Role of MLK3 in the regulation of mitogen-activated protein kinase signaling cascades. Mol Cell Biol 25, 3670–3681. Handley, M. E., Rasaiyaah, J., Chain, B. M. and Katz, D. R. (2007) Mixed lineage kinases (MLKs): a role in dendritic cells, inflammation and immunity? Int J Exp Pathol 88, 111–126.
48. Waterston, R. H., Lindblad-Toh, K., Birney, E., Rogers, J., Abril, J. F., Agarwal, P., et al. (2002) Initial sequencing and comparative analysis of the mouse genome. Nature 420, 520–562. 49. Jakobovits, A., Amado, R. G., Yang, X., Roskos, L. and Schwab, G. (2007) From XenoMouse technology to panitumumab, the first fully human antibody product from transgenic mice. Nat Biotechnol 25, 1134–1143. 50. Gonzalez, F. J. and Yu, A. M. (2006) Cytochrome P450 and xenobiotic receptor humanized mice. Annu Rev Pharmacol Toxicol 46, 41–64. 51. Pascolo, S. (2005) HLA class I transgenic mice: development, utilisation and improvement. Expert Opin Biol Ther 5, 919–938. 52. Shultz, L. D., Ishikawa, F. and Greiner, D. L. (2007) Humanized mice in translational biomedical research. Nat Rev Immunol 7, 118–130. 53. Ito, M., Hiramatsu, H., Kobayashi, K., Suzue, K., Kawahata, M., Hioki, K., et al. (2002) NOD/SCID/gamma(c)(null) mouse: an excellent recipient mouse model for engraftment of human cells. Blood 100, 3175–3182. 54. Traggiai, E., Chicha, L., Mazzucchelli, L., Bronz, L., Piffaretti, J. C., Lanzavecchia, A., et al. (2004) Development of a human adaptive immune system in cord blood celltransplanted mice. Science 304, 104–107. 55. Ishikawa, F., Yasukawa, M., Lyons, B., Yoshida, S., Miyamoto, T., Yoshimoto, G., et al. (2005) Development of functional human blood and immune systems in NOD/SCID/IL2 receptor {gamma} chain(null) mice. Blood 106, 1565–1573. 56. Tateno, C., Yoshizane, Y., Saito, N., Kataoka, M., Utoh, R., Yamasaki, C., et al. (2004) Near completely humanized liver in mice shows human-type metabolic responses to drugs. Am J Pathol 165, 901–912. 57. Azuma, H., Paulk, N., Ranade, A., Dorrell, C., Al-Dhalimy, M., Ellis, E., et al. (2007) Robust expansion of human hepatocytes in Fah-/-/Rag2-/-/Il2rg-/- mice. Nat Biotechnol 25, 903–910. 58. Hay, D. C., Zhao, D., Fletcher, J., Hewitt, Z. A., McLean, D., Urruticoechea-Uriguen, A., et al. (2008) Efficient differentiation of hepatocytes from human embryonic stem cells exhibiting markers recapitulating liver development in vivo. Stem Cells 26, 894–902.
Chapter 4 Engineering the Mouse Genome to Model Human Disease for Drug Discovery Frank Koentgen, Gabriele Suess, and Dieter Naf Abstract Genetically engineered mice (GEM) have become invaluable tools for human disease modeling and drug development. Completion of the mouse genome sequence in combination with transgenesis and gene targeting in embryonal stem cells have opened up unprecedented opportunities. Advanced technologies for derivation of GEM models will be introduced and discussed. Key words: Transgenic mice, gene targeting, homologous recombination, humanized mice, knockout, reporter gene.
1. Introduction The mouse (Mus musculus) has in the course of a mere century completed a remarkable journey; the metamorphosis from ancient agricultural pest – the most plausible cause for domestication of the cat by despairing early farmers in the Fertile Crescent (1) – into the premier experimental model of human genetics and disease (2). Studies on mouse coat color inheritance first provided evidence for the applicability of Mendel’s laws to mammals (3). Clarence Little, a student of William Ernest Castle’s at the Bussey Institute of Harvard University, recognized the need for genetically homogeneous lines of mice and derived the first inbred strains, using them to demonstrate the genetic basis of cancer. Little established The Jackson Laboratory in Bar Harbor, Maine, and his inbred mice became essential tools for various fields of biomedical research. Cancer biology is among the many fields G. Proetzel, M.V. Wiles (eds.), Mouse Models for Drug Discovery, Methods in Molecular Biology 602, DOI 10.1007/978-1-60761-058-8 4, © Humana Press, a part of Springer Science+Business Media, LLC 2010
55
56
Koentgen, Suess, and Naf
that have greatly benefited from disease modeling in the mouse (4). Mice are attractive research models not only because they are relatively easy and inexpensive to breed but mainly because technologies of genetic engineering have become available that now enable us to manipulate the mouse genome virtually at will. In this introductory chapter, we will review and discuss state-of-the-art methodologies used in the production of genetically engineered mice (GEM) and explain some of the terminology used in the field.
2. Transgenic Mice 2.1. Pronuclear Transgenesis
The well-established field of mouse genetics underwent a revolutionary transformation in the winter of 1980–1981 when the powerful new tools of molecular biology were introduced (5). Five near-simultaneous publications demonstrated that cloned foreign genes could be inserted into the mouse genome by injection of DNA into the male pronucleus of fertilized oocytes, resulting in live mice that passed the foreign gene on to their offspring (6–10). Most notably, Wagner et al. (10) showed that such transgenic mice harboring a rabbit beta-globin gene faithfully expressed the foreign protein in erythrocytes. The ramifications were monumental; it became apparent that transgenic mice could be used to experimentally elucidate the function of human genes. The term “transgenic” broadly refers to any organism carrying a foreign gene in a stable, heritable form. Pronuclear DNA injection continues to be employed in the generation of transgenic mice, and the method has been adapted to many other mammals, including laboratory rats and agricultural livestock (11). The typical makeup of a transgenic mouse is exemplified by the TRAMP strain, which was designed to model human prostate cancer (12). With the aim of directing expression of a potent oncogene to epithelial cells of the prostate, the coding sequence of simian virus 40 (SV40) T antigen was placed under control of regulatory sequences borrowed from the prostate-specific promoter of the rat probasin gene (Fig. 4.1A). Pronuclear injection of the DNA construct resulted in transgenic mice that exhibited the expected phenotype; males developed epithelial prostatic neoplasias that progressed to adenocarcinomas by 3 months and culminated in metastatic cancer with a frequency approaching 100% (13). The histopathology of TRAMP mice closely resembled human prostate cancer (14), validating their use as preclinical models for the development of new combination immunotherapies (15) that have since progressed to phase II clinical trials (N. Greenberg, personal communication).
GEM Models for Drug Discovery
57
Fig. 4.1. Pronuclear and lentiviral transgenesis. (A) Scheme of a pronuclear transgene construct and pronuclear DNA microinjection. Basic pronuclear transgenesis constructs are plasmids containing a mammalian or viral promoter (gray arrow) followed by the coding sequence of the gene to be expressed (black box) from initiation to termination codon (black and white lollipop, respectively). Most pronuclear transgenes include an intron as splicing enhances translation of mRNAs. A polyadenylation signal (pA, hatched box) is used to terminate transcription. Linearized plasmid DNA is injected directly into one of the pronuclei of a fertilized oocyte. (B) Scheme of a lentiviral construct and perivitelline injection of virions. Lentiviral constructs are based on a rudimentary version of the HIV-1 genome that consists of the 5 and 3 long terminal repeats (LTR, stippled boxes) and the viral packaging signal (). A promoter and the coding sequence of the gene of interest are inserted as for pronuclear transgenes. Introns cannot be included as these would be spliced out in the packaging cell line after transcription of the DNA template into viral genomic RNA. Instead, a posttranscriptional regulatory element from woodchuck hepatitis virus (WPRE, light gray box) may be used to enhance transgene expression. Transcription is terminated by a viral polyadenylation signal in the 3 LTR. Lentiviral plasmid constructs are transiently transfected into packaging cell lines from which infectious virions are harvested. These are injected into the perivitelline space of fertilized oocytes.
The major advantage of pronuclear transgenesis is that the technology imposes virtually no limits on transgene size (16). Transgenic mice have been obtained even with yeast artificial chromosomes (YAC) carrying inserts as large as 1.3 Mb (17). Smith et al. (18) used a panel of YAC transgenic mice covering a total of 2 Mb of human chromosome 21q22.2 to model neurological aspects of Down syndrome. As valuable as the technology has proven to be, pronuclear transgenesis is plagued with significant drawbacks. Transgene integration into the genome is an uncontrolled, stochastic event, and transgenes tend to insert as head-to-tail concatemers rather than single copies (8, 19). One serious limitation of the technology is tissue specificity of transgene expression or – more often – the lack thereof. Host sequences flanking the insertion site can severely affect transgene expression patterns, a phenomenon known as the chromosome position effect (20). The sword cuts both ways; because most transgenes contain strong transcriptional regulatory elements, such as promoters and enhancers, they in turn may cause dysregulation of host genes in the vicinity of the
58
Koentgen, Suess, and Naf
integration site, sometimes (in 5–10% of the cases studied) resulting in homozygous lethality (21). Some transgenic strains have been observed to lose transgene expression over time, either gradually or catastrophically, as a consequence of epigenetic silencing (e.g., DNA methylation) or even loss of transgene copies. Another caveat follows from the random nature of transgene insertion: each transgenic founder (i.e., each mouse derived from an individual injected oocyte) is genetically unique. It is therefore necessary to breed and characterize multiple transgenic lines for each construct independently in order to control for phenotypic distortions caused by integration site and copy effects. Even individual progeny from the same founder may show variation in transgene expression (22). Finally, not all transgenic animals will transmit the transgene. 2.2. Lentiviral Transgenesis
An alternative to pronuclear transgenesis is the use of lentiviral vectors for transgene delivery. These vector systems were originally developed as tools of gene therapy for human disease. Lentiviruses are a clade of retroviruses that have the capacity to productively infect non-dividing cells (23). The most extensively studied member is human immunodeficiency virus 1 (HIV-1), from which many of the currently used vectors are derived. For a detailed account of lentiviral biology and the characteristics of lentiviral vector systems we refer the reader to a recent review by Buchschacher and Wong-Staal (24). For the purpose of the present discussion it is sufficient to know that lentiviral vectors are, in essence, rudimentary versions of the HIV genome from which all genes required for viral replication have been removed. The vector retains only the cis-acting regulatory sequences for transcription into the viral RNA genome and its encapsidation into viral particles, as well as a cloning site for insertion of the transgene. Virus is produced by transient transfection of the construct into packaging cell lines, which are engineered to provide the essential viral proteins for assembly of infectious particles (25). Virions are then harvested from the cell culture medium. Importantly, these are capable of completing one infectious cycle only and cannot replicate because they contain no viral genes. The types of transgenes used in lentiviral transgenesis are not dissimilar to the pronuclear ones described above. They usually contain a mammalian or viral promoter linked to the coding sequence of the gene of interest (Fig. 4.1B). There is, however, a natural limit to transgene size as virions can only encapsidate RNA molecules up to approximately 10 kb. Many retroviral vectors include a posttranscriptional regulatory element from woodchuck hepatitis virus to boost transgene expression (26). The preferred method for delivery of lentiviral vectors is microinjection into the perivitelline space of single cell or cleavage-stage embryos, which is a far less invasive procedure
GEM Models for Drug Discovery
59
than pronuclear DNA injection and hence results in higher survival rates, sometimes approaching 90% (27). Most lentiviral vectors are pseudotyped with the surface protein VSV-G of vesicular stomatitis virus, which enables them to infect a wide variety of cell types, including mouse embryos. Infected cells reverse transcribe the lentiviral RNA into DNA (provirus), which then inserts into the genome. Unlike pronuclear transgenes, proviruses integrate as single copies, but there can be multiple, independent integrations into different chromosomal sites. Proviruses integrate preferentially into actively transcribed loci (28, 29) and hence may disrupt host genes. Lentiviral transgenes appear to be expressed more predictably than pronuclear ones and are less prone to epigenetic silencing (27, 30, 31). Many of the caveats discussed above for pronuclear transgenes apply to lentiviral transgenesis as well (Table 4.2). Retroviral integration into transcribed regions of the genome may cause dysregulation or disruption of unknown host genes, and effects of the integration site on transgene expression have also been documented (32). Multiple transgenic founders need to be produced and characterized because the lentivirus will integrate into different loci in each embryo. In addition, it may be difficult to establish stable transgenic lines because transgenes that integrated into multiple sites on different chromosomes will segregate during breeding. A further complication has been observed with constructs carrying strong, ubiquitously active promoters; the transgene may be expressed at high levels in the packaging cell line, and its product may be toxic to the cells or interfere with virion assembly. Thus, it is advisable to use inducible or conditional systems wherever possible.
3. Embryonal Stem Cell Technology and Gene Targeting
In 2007, Mario Capecchi, Sir Martin Evans, and Oliver Smithies were awarded the Nobel Prize in Physiology or Medicine in recognition of “their discoveries of principles for introducing specific gene modifications in mice by the use of embryonic stem cells.” The advent of embryonic stem (ES) cell technology and gene targeting in the late 1980s (33) squarely placed the mouse in pole position as the model of choice for preclinical drug discovery. ES cells are derived from the inner cell mass of blastocysts (34, 35). This cell population normally gives rise to the embryo proper. Under appropriate conditions, ES cells can be cultured in vitro and will retain their pluripotency and undifferentiated state (36). When injected into a host blastocyst and returned to a foster mother, ES cells will repopulate the embryo and give rise to
60
Koentgen, Suess, and Naf
chimeric mice in which part of each organ system, including the germline, can be derived from the ES cells (37). It follows that any genetic modification engineered in vitro into the genome of ES cells can be propagated through the chimera’s germline (germline transmission) and passed on to its offspring, provided that the mutation does not adversely affect pluripotency and the cells’ ability to contribute to the germline. Doetschman et al. (38) first demonstrated that homologous recombination could be exploited for targeted modification of a specific gene in ES cells. Since those pioneering days, thousands of genes have been modified in the mouse through gene targeting. In the following, we will discuss the targeting strategies most commonly used today and technical refinements that have been made to enhance the technology’s versatility. 3.1. Gene Knockouts
Every gene targeting project starts with design and construction of a targeting vector, usually a plasmid that contains parts of the gene of interest and the mutation(s) to be introduced. This is a critically important step as poor vector design will inevitably result in disaster. Figure 4.2A shows the structure of a basic targeting vector. It consists of three essential components: two stretches of mouse genomic DNA that are identical in sequence to the gene to be targeted (the 5 and 3 “homology arms”) and a mini-transgene that serves as a positive selection marker (“selection cassette”). Upon transfection into ES cells, the vector’s homology arms guide it to the matching position in the mouse genome and drive homologous recombination, which copies vector sequences, including the (positive) selection cassette, into the endogenous gene. The resulting targeted allele is a constitutive genetic “knockout” if the selection marker is positioned to replace a functionally important segment of the gene (e.g., a crucial exon). Positive selection is necessary because transfection of ES cells and homologous recombination are frustratingly inefficient. The transfection method of choice is electroporation due to a high degree of reproducibility, but its efficiency is low (<10−3 ). Only a fraction of cells that have taken up vector DNA will undergo homologous recombination (3–5%) (39). Table 4.1 lists markers used for selection of transfected ES cells, the most common being neo, which confers resistance to the aminoglycoside antibiotic, G418. To add insult to injury, the majority of ES cell clones that emerge from positive selection harbor random integrations of the targeting vector either in addition to, or in absence of, the properly targeted allele. To minimize the number of clones with random integrations many laboratories use a positive-negative selection procedure (40) that relies on counter-selectable markers (Table 4.1) integrated into the vector backbone. This strategy
GEM Models for Drug Discovery
61
Fig. 4.2. Constitutive and conditional gene targeting. (A) Constitutive gene knockout. The targeting vector is designed to replace a functionally important exon (exon 2) with the neomycin (neo) selection marker. The selection cassette includes a promoter (gray arrow) that is active in ES cells and a polyadenylation signal (hatched box). A counter selection gene (e.g., diphtheria toxin subunit A, DTA) is included in the plasmid backbone (gray line) to eliminate cells that carry nonhomologous integrations of the vector. The 5 and 3 homology arms (5 HA and 3 HA, respectively) drive homologous recombination (indicated by gray X symbols), which results in the knockout (KO) allele. (B) Conditional gene targeting. The targeting vector is shown without a counter selection marker for sake of simplicity. The positive selection marker (neo) is inserted into an intron and flanked by FRT sites (white triangular flags). Black triangles depict loxP sites, which are positioned to flank (“floxed”) exon 2. Homologous recombination results in the targeted allele (tm). Flp-mediated recombination is used to excise the selection cassette in order to obtain a conditional (cond) allele, which is fully functional because intronic loxP and FRT sites do not disrupt transcription. Cre-mediated deletion of the floxed exon generates the knockout (KO) allele.
seeks to take advantage of the fact that homologous recombination eliminates targeting vector sequences outside the homology arms, whereas non-homologous integration usually incorporates the entire plasmid. Even with negative selection robust genetic screens must be conducted to unambiguously identify correctly targeted clones. PCR screens are widely used, at least for the initial screen (39), but Southern blotting may be preferable because the method is far less prone to contamination and erroneous results. Initial PCR screening followed by Southern blotting is also an option, combining the strength of both methods. 3.1.1. Conditional Gene Targeting
Targeting vectors of the type described above induce null alleles (gene “knockouts” [KO]), which are of limited use for disease modeling because knockouts are mostly recessive, thus
62
Koentgen, Suess, and Naf
Table 4.1 Markers for positive and negative selection of transfected ES cells Positive selection Marker
Selective agent
Resistance gene
neo
Geneticin (G418)
Neomycin 3 -phosphotransferase
hygro
Hygromycin B
Hygromycin phosphotransferase
puro
Puromycin
Puromycin N-acetyl-transferase
Negative selection Marker
Mechanism of action
HSV-tk
Counterselective Gene agent Ganciclovir (cytovene, Herpes simplex virus synthetic nucleo(HSV) thymidine kinase side analogue)
DTA
None
Inhibits translation by ADP ribosylation of eukaryotic elongation factor, eEF-2
Diphtheria toxin subunit A (Corynebacterium diphtheriae)
HSV-tk converts ganciclovir to a cytotoxic compound
showing no heterozygous phenotype, many cause embryonic lethality when mice are bred to homozygosity or the mutation’s effect can be masked, partly or fully by compensation during embryonic development. To study the role of a gene during development or in a specific tissue of the adult animal it is necessary to employ conditional targeting strategies. The principle of conditional gene targeting is depicted in Fig. 4.2B. The key tools are site-specific DNA recombinases of the integrase (Int) type and their binding sites (41). Int enzymes recognize short sequence motifs that consist of palindromic recombinase binding sites separated by a central core that gives the motif directionality. Two such recombinases have been modified for optimal efficiency in mammalian cells: Cre recombinase of bacteriophage P1 (42), which mediates recombination at lox sites, and Flp recombinase of yeast (43), which binds to FRT sites. Cre will excise any stretch of DNA that is flanked (“floxed”) by unidirectional lox sites. The floxed sequence is inverted if the sites are in opposing orientation. Flp acts analogously on sequences flanked (“flrted”) by FRT sites. Instead of replacing a functionally important exon with neo as described above, conditional vectors typically target the selection cassette into an intron and introduce lox sites on both sides of the exon of interest. Since no coding sequences of the gene are disrupted, the targeted allele should encode a normal product, and homozygous mice are in general viable and phenotypically normal. It is important, however, to note that selection cassettes are by no means inert pieces of DNA. Most of them contain strong promoters and polyadenylation signals, effectors of gene expression that are bound to raise havoc even if inserted into an
GEM Models for Drug Discovery
63
intron (44). In addition, some selection cassettes contain cryptic splice sites that may interfere with expression of the targeted allele (45). In order to obtain a truly conditional allele, it is therefore imperative that the selection cassette be removed from the targeted locus. A popular approach is to flank the exon(s) to be deleted with loxP sites and use FRT sites to bracket the selection marker, such that FlpE recombination will create the conditional allele by removing the selection cassette, and secondary recombination with Cre will induce the KO. This is most commonly achieved by breeding mice carrying the floxed allele with transgenic mice expressing the appropriate recombinase. 3.1.2. Tissue-Specific Knockouts
Rajewsky and colleagues (46) first demonstrated that matings of mice carrying a floxed allele with transgenic animals expressing Cre in a T-cell-specific manner produced offspring in which the targeted gene was knocked out only in the T cells. This strategy of tissue-specific gene inactivation exemplifies the immense versatility of conditional gene targeting; a single conditional mutant strain may be used in conjunction with any number of tissue-specific Cre transgenics to study gene function in the adult mouse, specifically in the cell type or tissue of interest, thereby circumventing the early lethal phenotype or unplanned developmental compensatory effects associated with many nonconditional KOs. To date, scores of Cre-expressing strains have been derived and characterized as invaluable tools for tissuespecific gene targeting (47). Efforts are underway to build a comprehensive online database of these strains, their organ- or cell type-specific Cre expression patterns, and availability (Cre-XMice; http://nagy.mshri.on.ca/cre/) (48).
3.1.3. Inducible Knockouts
Many human diseases of great pharmaceutical importance, such as cancer, arise from stochastic genetic events that occur in a particular tissue at a certain stage of development or late in adult life. Accurate modeling of such conditions has been achieved through creative combinations of conditional gene targeting and tissue-specific Cre variants whose activity can be controlled by administration of a non-toxic chemical compound. Fusion proteins of Cre and modified ligand binding domains of the estrogen (49–51) or progesterone (52) receptors have been devised that mediate recombination only in presence of synthetic steroids. A widely used version is the estrogen receptor fusion, Cre-ERT2 (53), which responds to 4-hydroxytamoxifen (4-OHT) but not to endogenous estrogen. In absence of the drug, Cre-ERT2 is sequestered in the cytoplasm, presumably as a complex with heat shock proteins (HSP) 70 and 90. Upon binding of 4-OHT, the chimeric recombinase dissociates from HSPs and translocates to the nucleus where it engages its DNA substrate and catalyzes the recombination event. Mouse strains that express CreERT2 under
64
Koentgen, Suess, and Naf
Fig. 4.3. (continued)
GEM Models for Drug Discovery
65
control of tissue-specific promoters have been used to combine temporal and spatial control of gene inactivation (47). 3.2. Targeted Knock-ins
Gene targeted mouse models would be of questionable utility to preclinical drug development if the technology were limited to inducing simple loss-of-function alleles (KO). Precise modeling of human disease often requires more subtle manipulation of the mouse genome. For instance, a single codon might need to be altered in the orthologue of a disease-associated gene to produce an antimorphic, dominant-negative protein (i.e., an aberrant protein that negatively affects function of the wild-type polypeptide). Fortunately, properly designed targeting vectors can deliver almost any imaginable mutation, and homologous recombination will “knock it into” the ES cell genome. It is important to emphasize the fundamental difference between knockouts and knock-ins; whereas loss-of-function KO alleles are mostly recessive, knock-ins tend to have a dominant phenotype. In the worst case, knock-in ES cells expressing a mutant gene product may lose their capacity to participate in normal embryonal development or populate the germline. For this reason it is far better to carry out knock-in projects conditionally, such that viable mice are obtained and expression of the mutant gene can be switched on only in the tissue(s) of interest or in a temporally controlled fashion. This also provides a far better experimental system. Figure 4.3 depicts three conditional strategies for point mutation knock-ins. All aim to satisfy a set of key requirements, namely (i) that the point-mutant allele be expressed from the
Fig. 4.3. Conditional mutation knock-ins. (A) The “mini-cDNA” method. The targeting vector introduces a partial cDNA (“mini-cDNA”) consisting of the exon to be targeted (e.g., exon 2) and all coding downstream exons (2+3). The mutant exon 2 (2∗ , white star) is inserted downstream of the mini cDNA. A polyadenylation signal (hatched box) terminates transcription after the mini cDNA, resulting in expression of the wild-type protein. Cre-mediated excision of the floxed mini cDNA leaves mutant exon 2 in its position, activating expression of the mutant protein. (B) The Cre/lox inversion method. The targeting vector delivers the mutant exon (2∗ ) into an intronic site. The mutant exon blends into the intron and has no coding capacity because it is in reverse orientation relative to the gene. Note that removal of the neomycin cassette (neo) is necessary to obtain a conditional allele (not shown) as the polyadenylation signal (hatched box) is expected to terminate transcription upstream of exon 3. Variant lox sites (lox66, lox71) are positioned to flank the wildtype and mutant exons. The lox sites are in head-to-head orientation, such that Cre-mediated recombination causes a chromosomal inversion. After the event, exon 2∗ is in sense orientation and the mutant protein is expressed. Recombination between lox66 and lox71 creates one wild-type (loxP) and one double-mutant ([lox]) site, preventing the reverse reaction. (C) Inversion with heterotypic lox sites. Lox511 (stippled triangular flag) and loxP (black triangular flag) are incompatible due to mismatches in the central spacer sequence; lox511 will recombine with lox511 but not loxP, and vice versa. Targeting vector design is similar to Fig. 4.3B, except that the wild-type and inverted exons are flanked with an array of heterotypic lox sites. Removal of the neomycin cassette (neo) by Flp recombination generates a conditional allele (cond) that only expresses the wild-type protein. Cre recombination is a two-step reaction that first creates an inversion through recombination at loxP (KI(transient)) and then deletes one loxP site by recombination at lox511. The resulting knock-in allele (KI(fixed)) cannot revert because there are no compatible lox sites. Only the mutant protein can be expressed.
66
Koentgen, Suess, and Naf
Fig. 4.4. Knock-ins and targeted transgenesis. (A) Transgene knocked into the coding sequence. The transgene open reading frame (ORF, striped box) is targeted into the initiation codon (black lollipop) of the host gene, replacing part of the endogenous coding sequence. A polyadenylation signal after the transgenic ORF terminates transcription, which is driven by the endogenous promoter. Although the neomycin resistance cassette is located downstream of the transgenic transcription unit, its removal by Flp recombination is recommended to avoid interference with transgene expression. (B) IRES-mediated transgene expression. The targeting vector inserts the transgene (Tg ORF, striped box) into the 3 nontranslated sequence (gray box) of the host gene. No host gene coding sequences are disrupted. An internal ribosome entry
GEM Models for Drug Discovery
67
endogenous promoter, (ii) only the wild-type protein be produced prior to Cre recombination, and (iii) the recombination event switches to expression of the mutant coding sequence. The “mini cDNA” approach (54) (Fig. 4.3A) replaces the exon to be mutagenized with a floxed partial cDNA that fuses it to downstream exons, ensuring expression of a wild-type gene product prior to Cre recombination. The mutagenized exon is positioned downstream of the selection marker. Cre-mediated recombination excises the mini-cDNA, leaving the point-mutant exon and an intronic lox site. The resulting allele can only express the mutant gene product as the wild-type exon has been eliminated. The alternative “inversion” strategies (Fig. 4.3B,C) exploit asymmetric recombination driven by variant lox sites (55). The pointmutant exon and its flanking splice sites are initially placed in opposite orientation relative to transcription of the targeted gene. Cre-mediated recombination inverts the floxed sequences, bringing the mutant exon into sense orientation. Reversion of the recombination event is prevented either through the use of selfinactivating lox variants, lox66 and lox71 (56) (Fig. 4.3B), or by a secondary deletion (“FLEx switch”, Fig. 4.3C) in an array of heterotypic lox sites (57, 58). Disadvantages of the inversion methods include the probability of disrupting intronic regulatory elements by insertion of inverted exons and reduced recombination efficiency of variant lox sites (54, 59). The cDNA strategy, on the other hand, may not be feasible if a point mutation is to be introduced into an early exon of a gene with a very large coding sequence, or if the conditional allele must express multiple splice variants. Of course, neither approach is strictly limited to point mutations. Knock-in vectors can also be used as transgenesis tools. In order to achieve truly tissue-specific expression of a transgene, it is best to knock its coding sequence into a locus that naturally exhibits the expression pattern of interest (Fig. 4.4A). Many tissue-specific Cre mice were generated this way. If the transgene is targeted into the initiation codon of the host locus, it will be expressed in lieu of the endogenous gene. Thus, the trans-
Fig. 4.4. (continued) site (IRES, stippled box) is inserted immediately after the endogenous stop codon (white lollipop) to enable translation of the transgene. Note that the 3 non-translated region and any potential regulatory sequences therein are preserved. Transcription of the bicistronic mRNA terminates at the natural polyadenylation signal (pA, gray arrow). (C) Conditional and tissue-specific ROSA26 transgenesis. Exons of the ROSA26 gene, Gt(ROSA)26Sor, are depicted as black boxes. Gray boxes represent exons of Thumpd3, which is transcribed in the opposite direction. The ROSA26 promoter is indicated as a black arrow. Gene targeting inserts a conditional transgene into the first ROSA26 intron. The transgene consists of an ORF (Tg, striped box), which is separated from the constitutive human UbiC promoter (stippled arrow) by a floxed transcriptional stop element (LSL). The transgene is only expressed from the knock-in allele (KI) after removal of LSL by Cre-mediated recombination. Although transgene transcription is driven by a ubiquitous promoter, UbiC, cell-type specificity is determined by the expression pattern of Cre.
68
Koentgen, Suess, and Naf
gene knock-in simultaneously knocks out the targeted gene and is expressed. Alternatively, the transgene open reading frame (ORF) can be knocked into the 3 non-translated region (3 NTR) of a gene such that no endogenous coding sequences are disrupted (Fig. 4.4B). The knock-in allele then transcribes a bicistronic mRNA. In order for the second ORF – the transgene – to be translatable, it must be preceded by an internal ribosome entry site (IRES). IRES sequences are translational regulatory elements that were first identified in poliovirus (60) and encephalomyocarditis virus (EMCV) (61). They allow for assembly of the ribosome complex and translational initiation independent of an mRNA 5 end cap structure, albeit generally with reduced efficiency (62, 63). These knock-in methods are particularly useful to tag certain populations of cells with fluorescent or bioluminescent marker proteins that are readily detectable by in vivo imaging technologies. Such applications are of considerable interest to the field of disease modeling because they enable direct, realtime observation of disease progression or remission in response to drug treatment. Since the first transgenic mice expressing green fluorescent protein (GFP) were obtained by pronuclear injection (64) or through the use of ES cells (65), a cornucopia of fluorescent markers has emerged (66) and many have been used successfully in mice (67). Paired with conditional gene targeting and Cre/lox recombination, fluorescent reporters can light up cells that have activated the genetic modification of interest. Variant recombinase sites can be used to switch between expression of different colors (68). 3.3. Humanized Mice
The generation of humanized mice by homologous recombination is an expansion of the gene knock-in technology. Instead of manipulating a gene to introduce subtle mutations that recapitulate a human disease-associated genotype, the mouse genome is engineered to express the entire human protein (or critical domains thereof) in lieu of the murine counterpart. Humanized mice hold great promise as models for preclinical drug validation, particularly if the drug of interest is a small molecule antagonist or a therapeutic antibody that shows poor cross-reactivity between human and rodent targets (69). For instance, Lee et al. (70) used homologous recombination in ES cells to replace most of the coding portion of murine C5ar (complement component 5a receptor) with the human sequence. Homozygous animals expressing only the humanized protein were obtained and used successfully to screen anti-C5AR monoclonal antibodies for their ability to reverse the progression of experimental arthritis in the humanized mouse model. Although expression of human genes in the mouse was achieved early on by standard pronuclear transgenesis, knock-in models are far superior because tissue-specific
GEM Models for Drug Discovery
69
and temporal expression of the humanized gene is governed by endogenous mouse sequences. As is true for gene targeting in general, there is no single strategy for humanizing mouse genes that would deliver a silver bullet. Careful evaluation of alternative approaches is essential for each project. In some cases it may be possible to replace the entire mouse gene – exons and introns – with the human one. In other instances, such as signaling transmembrane receptors or polypeptides that participate in formation of protein complexes, it may be desirable to retain some mouse sequences to ensure the humanized protein functions properly in the murine environment (70). 3.4. Targeted Transgenesis
Targeted gene knock-in technology provides an attractive alternative to pronuclear and lentiviral transgenesis. Many of the latter techniques’ caveats, as discussed earlier, can be avoided by integrating transgenes into a “safe haven” locus in the mouse genome by means of homologous recombination in ES cells (71). Two mouse genes have been used extensively as host loci for targeted transgene insertion, Hprt1 and Gt(ROSA)26Sor. The Hprt1 gene encodes the adenine salvage pathway enzyme, hypoxanthine guanine phosphoribosyl transferase 1. Defects in HPRT1 cause Lesch–Nyhan syndrome in humans (72), but the mouse gene is not essential as Hprt1 deficient mice are phenotypically normal (73, 74). Numerous transgenes have been targeted successfully into the ubiquitously expressed Hprt1 gene. In some cases, however, these transgenes showed unexpected expression patterns suggesting that the Hprt1 locus may influence activity or tissue specificity of promoter/transgene insertions (75, 76). Since Hprt1 is X chromosomal, transgenes targeted into this locus are subject to random X chromosome inactivation in females. The Gt(ROSA)26Sor gene, commonly referred to as ROSA26, was discovered in a gene trapping experiment in which ES cells were infected with retroviruses containing a promoterless -galactosidase reporter gene (77). X-gal (5-bromo-4-chloro-3indolyl-beta-D-galactopyranoside) staining of ROSA26 embryos showed ubiquitous expression of the gene throughout development (77, 78). The gene products of ROSA26 are two alternatively spliced non-coding RNAs of unknown function. ROSA26 partially overlaps with a protein coding gene (Thumpd3) on the opposite strand of mouse chromosome 6 (78) (Fig. 4.4C). The genomic structure of ROSA26 is conserved between mice and humans (79). ROSA26 gene trapped mice are fertile and phenotypically normal, indicating that the gene is not essential. Because of its autosomal location, ubiquitous expression, and the fact that the gene appears not to be subject to epigenetic inactivation, ROSA26 has become the locus of choice for targeted transgenesis. ROSA26 transgenesis has a number of advantages
70
Koentgen, Suess, and Naf
Table 4.2 Characteristics of transgenesis methods Method
Integration
Transgene copy number per genome
Genetics of founders
Breeding
Pronuclear injection
Random, multiple copies in single integration site
1–100 (not controllable)
Each founder is genetically unique
Must breed multiple lines. Transgene expression and/or copy number may be unstable
Lentiviral infection
Random, singlecopy integrations at multiple loci
1–10 (somewhat controllable by viral titer)
Each founder is genetically unique
Must breed multiple lines. Transgenes may segregate
Targeted transgenesis
Defined by targeted gene (e.g., ROSA26)
1
Each founder is genetically identical
Stable inheritance
over pronuclear DNA injection and lentiviral transgene delivery (Table 4.2). Because transgene integration is mediated by homologous recombination, a single copy of the transgene will be inserted. This eliminates instability issues that are often associated with repetitive transgene clusters resulting from pronuclear injection. Furthermore, transgene insertion occurs at a defined site in the genome, which greatly facilitates genotyping and identification of homozygous transgenic animals. Single site integration also means that targeted transgenes would not independently segregate during breeding. Each transgenic founder will be genetically identical, and it is therefore not necessary to breed and characterize multiple lines. Most importantly, targeted transgenesis in ROSA26 avoids unpredictable chromosome position effects, the Achilles heel of pronuclear and lentiviral transgenics. 3.4.1. Tissue-Specific Transgene Expression in ROSA26
The ROSA26 locus is actively transcribed in most cell types, and the mouse ROSA26 promoter has also been used to drive ubiquitous marker gene expression in transgenic rats and mice (80). Therefore, it is obvious that ROSA26 can serve as a host locus for widespread expression of transgenes. Just like the original -galactosidase gene trap, promoterless transgenes can be brought under control of the endogenous ROSA26 regulatory elements. Alternatively, transgenes can be knocked into ROSA26 together with a constitutive or activatable promoter of their own. Strathdee et al. (81) observed that a tetracyclin-inducible transgene targeted into ROSA26 was expressed at higher levels when inserted in the opposite orientation relative to the endogenous transcription unit, suggesting transcriptional interference from
GEM Models for Drug Discovery
71
the ROSA26 promoter. It is conceivable that the ROSA26 promoter may alter the activity or modify tissue specificity of ectopic regulatory sequences targeted into the locus. For instance, when inserted into ROSA26, the human skeletal actin (HSA) promoter directed transgene expression to skeletal muscle as expected, but activity was also observed in the embryonic heart (82). It is thus not clear that tissue specificity of transgene expression can be accomplished in this manner. An alternative to tissue-specific promoters controlling the transgene is to employ the Cre/lox recombination system. Fig. 4.4C depicts how a transgene can be targeted into ROSA26 such that its expression from the endogenous promoter – or a constitutive promoter included in the transgene construct – is squelched by a “Lox-STOP-Lox” (LSL) element, a potent transcriptional terminator flanked by lox sites (83, 84). Such conditional transgenes can then be activated by Cre-mediated excision of LSL. Tissue specificity is determined by the expression pattern of Cre, which can be tightly controlled by selecting an appropriate and well-characterized Cre strain.
4. RNA Interference in Mice
Gene targeting by homologous recombination is still a timeconsuming and fairly expensive process. RNA interference (RNAi) (85) has recently emerged as a tool for gene function studies, which – at least in theory – promises to be more rapid than gene knockouts (86). RNAi is an evolutionarily conserved mechanism of posttranscriptional gene regulation, which uses short interfering RNAs (siRNA) to guide degradation or translational repression of mRNAs containing complementary sequences (87, 88). RNAi results in a “knock-down” rather than a complete loss of gene function. In principle, functionally effective siRNAs can be designed against literally any target gene (89). They can be expressed in mammalian cells as small hairpin-shaped molecules (shRNA), which are cleaved into active siRNAs by the same cellular machinery that processes endogenous micro RNAs (miRNA) (90). To express shRNAs ubiquitously, RNA polymerase III (PolIII) promoters are often used, such as H1 or U6, because PolIII effectively transcribes small RNAs without polyadenosine tails. The first siRNA knock-down mice were generated by conventional pronuclear DNA microinjection (91) but the method yielded rather unimpressive results (92). Lentiviral delivery systems and transgenesis in ES cells have since been used with increasing success. For instance, Hou et al. (93) used lentiviruses carrying a PolIII-driven shRNA against the kidney-
72
Koentgen, Suess, and Naf
specific tight junction protein, claudin-16 (Cldn16) to generate a model for the renal disorder, familial hypomagnesemia with hypercalciuria and nephrocalcinosis (FHHNC). Lentiviral RNAi transgenesis also demonstrated a role for the Wnt co-receptor tyrosine kinase, Ryk in neurite outgrowth and axon guidance (94), and led to the identification of the phagosomal ion transporter, Nramp1 (Slc11a1) as a type 1 diabetes susceptibility locus in the nonobese diabetic (NOD) mouse model (95). Sustained high-level expression of shRNAs in mice may cause unexpected phenotypes, including death, which are believed to arise from oversaturation of cellular miRNA processing pathways (96). To minimize the risk of such non-specific effects inducible RNAi technologies were devised. For example, Amar et al. created a doxycycline-activatable (“Tet-ON”) system based on the reverse tetracycline transactivator (rtTA) (97). The doxycyclinecontrolled DNA binding moiety of rtTA was fused to multimers of a modified, PolIII-compatible transactivation domain borrowed from the human transcription factor, Oct-2. In the presence of doxycycline the rtTA-Oct2 protein binds to tetO sequences in a modified PolIII promoter and interacts with RNA polymerase III to activate transcription. In a different version, rtTA or the original tet repressor (tTA) (which dissociates from DNA upon binding of doxycycline) was joined to the Kr¨uppelassociated box (KRAB) domain found in many vertebrate chromatin remodeling factors (98). When bound to DNA, KRAB proteins initiate formation of multiprotein complexes that mediate heterochromatin formation, thereby shutting down expression of genes in the vicinity (99). Doxycycline releases the tTAKRAB polypeptide from the DNA, opening the chromatin for transcription. Unfortunately, tTA-KRAB causes irreversible chromatin silencing if it is allowed to bind to DNA during early stages of mouse embryonal development (98), which limits its utility for RNAi transgenesis. As with conditional gene targeting, Cre/lox recombination is probably the most tightly controllable – albeit irreversible – way of achieving cell type-specific or temporally regulated RNAi. As outlined above, shRNAs are palindromic, hairpin-shaped structures, and so are loxP sites. A single LoxP site intercalated between the sense and the antisense portions of an shRNA properly folds into a loop that does not disturb siRNA processing (100). Hitz et al. developed a conditional gene knock-down system by inserting a floxed transcriptional stop cassette into the shRNA loop, such that the RNA can only be transcribed after Cremediated excision of the stop sequence (101). Their method was used successfully to silence mitogen activated protein kinase (MAPK) signaling in the mouse brain and was later expanded to enable expression of multiple shRNAs from the same vector (102). Lentiviral delivery of cre-inducible shRNA genes has been
GEM Models for Drug Discovery
73
described (103). However, multicopy lentiviral integrations could result in unpredictable chromosomal rearrangements after Cre recombination. It is thus preferable to target a single copy of the shRNA into a defined chromosomal locus, such as ROSA26 (101, 102). Finally, we would like to caution that, although there is an expanding body of literature showing successful RNAi in transgenic animals, biological activity of a particular shRNA construct is not always predictable. From an animal ethics perspective, shRNAs should first be tested in an in vitro system if possible before generation of a transgenic animal is envisaged.
Acknowledgments We thank Drs. David Tarlinton and Louise McKenzie for critical reading of the manuscript, and Ozgene scientific staff for fruitful discussions. References 1. Driscoll, C.A., Menotti-Raymond, M., Roca, A.L., Hupe, K., Johnson, W.E., Geffen, E., Harley, E.H., Delibes, M., Pontier, D., Kitchener, A.C., Yamaguchi, N., O’Brien, S.J., and Macdonald, D.W. (2007) The Near Eastern origin of cat domestication. Science 317, 519–523. 2. Paigen, K. One hundred years of mouse genetics: an intellectual history. I. The classical period (1902-1980). Genetics 163, 1–7. 3. Cuenot, L. (1902) La loi de Mendel et l’heredite de la pigmentation chez les souris. Arch. Zool. Exp. Gen. 3, 27–30. 4. Van Dyke, T. and Jacks, T. (2002) Cancer modeling in the modern era: progress and challenges. Cell 108, 5–15. 5. Paigen, K. One hundred years of mouse genetics: an intellectual history. II. The molecular revolution (1981-2002). Genetics 163, 1227–1235. 6. Gordon, J.W., Scangos, G.A., Plotkin, D.J., Barbosa, J.A., and Ruddle, F.H. (1980) Genetic transformation of mouse embryos by microinjection of purified DNA. Proc. Natl. Acad. Sci. USA 77, 7380–7384. 7. Brinster, R.L., Chen, H.Y., Trumbauer, M., Senear, A.W., Warren, R., and Palmiter, R.D. (1981) Somatic expression of herpes thymidine kinase in mice following injection of a fusion gene into eggs. Cell 27, 223–231. 8. Constantini, F. and Lacy, E. (1981) Introduction of a rabbit beta-globin gene into the mouse germ line. Nature 294, 92–94. 9. Harbers, K., Jahner, D., and Jaenisch,
10.
11. 12.
13.
14.
15.
R. (1981) Microinjection of cloned viral genomes into mouse zygotes: integration and expression in the animal. Nature 293, 540–542. Wagner, T.E., Hoppe, P.C., Jollick, J.D., Scholl, D.R., Hodinka, R.L., and Gault, J.B. (1981) Microinjection of a rabbit beta-globin gene into zygotes and its subsequent expression in adult mice and their offspring. Proc. Natl. Acad. Sci. USA 78, 6376–6380. Jaenisch, R. (1988) Transgenic animals. Science 240, 1468–1474. Greenberg, N.M., DeMayo, F.J., Finegold, M.J., Medina, D., Tilley, W.D., Aspinall, J.O., Cunha, G.R., Donjacour, A.A., Matusik, R.J., and Rosen, J.M. (1995) Prostate cancer in a transgenic mouse. Proc. Natl. Acad. Sci. USA 92, 3439–3443. Gingrich, J.R., Barrios, R.J., Morton, R.A., Boyce, B.F., DeMayo, F.J., Finegold, M.J., Angelopoulou, R., Rosen, J.M., and Greenberg, N.M. (1996) Metastatic prostate cancer in a transgenic mouse. Cancer Res. 56, 4096–4102. Kaplan-Lefko, P.J., Chen, T.M., Ittmann, M.M., Barrios, R.J., Ayala, G.E., Huss, W.J., Maddison, L.A., Foster, B.A., and Greenberg, N.M. (2003) Pathobiology of autochthonous prostate cancer in a preclinical transgenic mouse model. Prostate 55, 219–237. Hurwitz, A.A., Foster, B.A., Kwong, E.D., Truong, T., Choi, E.M., Greenberg, N.M., Burg, M.B., and Allison, J.P. (2000) Com-
74
16.
17.
18.
19.
20.
21. 22.
23.
24.
25. 26.
27.
Koentgen, Suess, and Naf bination immunotherapy of primary prostate cancer in a transgenic mouse model using CTLA-4 blockade. Cancer Res. 60, 2444– 2448. Giraldo, P. and Montoliu, L. (2001) Size matters: use of YACs, BACs and PACs in transgenic animals. Transgenic Res. 10, 83–103. Schedl, A., Larin, Z., Montoliu, L., Thies, E., Kelsey, G., Lehrach, H., and Schutz, G. (1993) A method for the generation of YAC transgenic mice by pronuclear microinjection. Nucl. Acids Res. 21, 4783–4787. Smith, D.J., Stevens, M.E., Sudanagunta, S.P., Bronson, R.T., Makhinson, M., Watabe, A.M., O’Dell, T.J., Fung, J., Weier, H.U., Cheng, J.F., and Rubin, E.M. (1997) Functional screening of 2 Mb of human chromosome 21q22.2 in transgenic mice implicates minibrain in learning defects associated with Down syndrome. Nat. Genet. 16, 28–36. Wagner, E.F., Stewart, T.A., and Mintz, B. (1981) The human ß-globin gene and a functional viral thymidine kinase gene in developing mice. Proc. Natl. Acad. Sci. USA 78, 5016–5020. Wilson, C., Bellen, H.J., and Gehring, W.J. (1990) Position effects on eukaryotic gene expression. Annu. Rev. Cell. Biol. 6, 679– 714. Palmiter, R.D. and Brinster, R.L. (1986) Germ-line transformation of mice. Annu. Rev. Genet. 20, 465–499. Furth, P.A., Hennighausen, L., Baker, C., Beatty, B., and Woychick, R. (1991) The variability in activity of the universally expressed human cytomegalovirus immediate early gene 1 enhancer/promoter in transgenic mice. Nucl. Acids Res. 19, 6205–6208. Lewis, P.F. and Emerman, M. (1994) Passage through mitosis is required for oncoretroviruses but not for the human immunodeficiency virus. J. Virol. 68, 510–516. Buchschacher, G.L. and Wong-Staal, F. (2000) Development of lentiviral vectors for gene therapy of human disease. Blood 95, 2499–2504. Miller, A.D. (1990) Retrovirus packaging cells. Hum. Gene Ther. 1, 5–14. Zuffrey, R., Donello, J.E., Trono, D., and Hope, T.J. (1999) Woodchuck hepatitis virus posttranscriptional regulatory element enhances expression of transgenes delivered by retroviral vectors. J. Virol. 73, 2886– 2892. Lois, C., Hong, E.J., Pease, S., Brown, E.J., and Baltimore, D. (2002) Germline transmission and tissue-specific expression of trans-
28.
29.
30. 31.
32.
33. 34.
35.
36.
37.
38.
39.
genes delivered by lentiviral vectors. Science 295, 868–872. Imren, S., Farbry, M.E., Westerman, K.A., Pawliuk, R., Tang, P., Rosten, P.M., Nagel, R.L., Leboulch, P., Eaves, C.J., and Humphries, R.K. (2004) High level betaglobin expression and preferred intragenic integration after lentiviral transduction of human cord blood stem cells. J. Clin. Invest. 114, 953–962. Schroder, A.R., Shinn, P., Chen, H., Berry, C., Ecker, J.R., and Bushman, F. (2002) HIV-1 integration in the human genome favors active genes and local hotspots. Cell 110, 521–529. Dann, C.T. (2007) New technology for an old favorite: lentiviral transgenesis and RNAi in rats. Transgenic Res. 16, 571–580. Jahner, D., Stuhlmann, H., Stewart, C.L., Harbers, K., Lohler, J., Simon, I., and Jaenisch, R. (1982) De novo methylation and expression of retroviral genomes during mouse embryogenesis. Nature 298, 623– 628. Van den Brandt, J., Wang, D., Kwon, S.H., Heinkelein, M., and Reichardt, H.M. (2004) Lentivirally generated eGFP-transgenic rats allow efficient cell tracking in vivo. Genesis 39, 94–99. Capecchi, M.R. (1989) Altering the genome by homologous recombination. Science 244, 1288–1292. Evans, M.J. and Kaufman, M.H. (1981) Establishment in culture of pluripotent cells from mouse embryos. Nature 292, 154–156. Martin, G.R. (1981) Isolation of a pluripotent cell line from early mouse embryos cultured in medium conditioned by teratocarcinoma stem cells. Proc. Natl. Acad. Sci. USA 78, 7634–7636. Abbondanzo, S.J., Gadi, I., and Stewart, C.L. (1993) Derivation of embryonic stem cell lines. Methods Enzymol. 225, 803–823. Bradley, A., Evans, M., Kaufman, M.H., and Robertson, E. (1984) Formation of germ-line chimaeras from embryo derived teratocarcinoma cell lines. Nature 309, 255–256. Doetschman, T., Gregg, R.G., Maeda, N., Hooper, M.L., Melton, D.W., Thompson, S., and Smithies, O. (1987) Targeted correction of a mutant HPRT gene in mouse embryonic stem cells. Nature 330, 576–578. Koentgen, F. and Stewart, C.L. (1993) Simple screening procedure to detect gene targeting events in embryonic stem cells. Methods Enzymol. 225, 878–890.
GEM Models for Drug Discovery 40. Mansour, S., Thomas, K, and Capecchi, M.R. (1988) Disruption of the proto-oncogene int-2 in mouse embryo-derived stem cells: a general strategy for targeting mutations to non-selectable genes. Nature 336, 348–352. 41. Branda, C.S. and Dymecki, S.M. (2004) Talking about a revolution: the impact of site-specific recombinases on genetic analyses in mice. Dev. Cell 6, 7–28. 42. Sauer, B. and Henderson, N. (1989) Crestimulated recombination at loxP-containing DNA sequences placed into the mammalian genome. Nucl. Acids Res. 17, 147–161. 43. O’Gorman, S., Fox, D.T., and Wahl, G.M. (1991) Recombinase-mediated gene activation and site-specific integration in mammalian cells. Science 251, 1351–1355. 44. Lewandoski, M. (2001) Conditional control of gene expression in the mouse. Nat. Rev. Genet. 2, 743–755. 45. Forlino, A., Porter, F.D., Lee, E.J., Westphal, H., and Marini, J.C. (1999) Use of Cre/lox recombination system to develop a non-lethal knock-in murine model for osteogenesis imperfecta with an alpha1(I) G349C substitution. J. Biol. Chem. 274, 37923–37931. 46. Gu, H., Marth, J.D., Orban, P.C., Mossmann, H., and Rajewsky, K. (1994) Deletion of a DNA polymerase beta gene segment in T cells using cell type-specific gene targeting. Science 265, 103–106. 47. Furuta, Y. and Behringer, R.R. (2005) Recent innovations in tissue-specific gene modifications in the mouse. Birth Defects Res 75, 43–57. 48. Nagy, A. (2000) Cre recombinase: the universal reagent for genome tailoring. Genesis 26, 99–109. 49. Feil, R., Brocard, J., Mascrez, B., LeMeur, M., Metzger, D., and Chambon, P. (1996) Ligand-activated site-specific recombination in mice. Proc. Natl. Acad. Sci. USA 93, 10887–10890. 50. Danielian, P.S., Muccino, D., Rowitch, D.H., Michael, S.K., and MacMahon, A.P. (1998) Modification of gene activity in mouse embryos in utero by a tamoxifen-inducible form of Cre recombinase. Curr. Biol. 8, 1323–1326. 51. Schwenk F., Kuehn R., Angrand, P.-O., Rajewsky K., and Stewart, A.F. (1998) Temporally and spatially regulated somatic mutagenesis in mice. Nucl. Acids Res. 26, 1427–1432. 52. Kellendonk, C., Tronche, F., Monaghan, A.P., Angrand, P.O., Stewart, F., and Schuetz, G. (1996) Regulation of Cre
53.
54. 55.
56.
57.
58.
59.
60.
61.
62.
63.
75
recombinase activity by the synthetic steroid RU486. Nucl. Acids Res. 24, 1404–1411. Feil, R., Wagner, J., Metzger, D., and Chambon, P. (1997) Regulation of Cre recombinase activity by mutated estrogen receptor ligand-binding domains. Biochem. Biophys. Res. Commun. 237, 752–757. Skvorak, K., Vissel, B., and Homanis, G.E. (2006) Production of conditional point mutant knockin mice. Genesis 44, 345–353. Albert, H., Dale, E.C., Lee, E., and Ow, D.W. (1995) Site-specific integration of DNA into wild-type and mutant lox sites placed in the plant genome. Plant J. 7, 649–659. Zhang, Z. and Lutz, B. (2002) Cre recombinase-mediated inversion using lox66 and lox71: method to introduce conditional point mutations into the CREB-binding protein. Nucl. Acids Res. 30, e90. Schnuetgen, F., Doerflinger, N., Calleja, C., Wendling, O., Chambon, P., and Ghyselinck, N.B. (2003) A directional strategy for monitoring Cre-mediated recombination at the cellular level in the mouse. Nat. Biotechnol. 21, 562–565. Schnuetgen, F., De-Zolt, S., Van Sloun, P., Hollatz, M., Floss, T., Hansen, J., Altschmied, J., Seisenberger, C., Ghyselinck, N.B., Ruiz, P., Chambon, P., Wurst, W., and von Melchner, H. (2005) Genomewide production of multipurpose alleles for the functional analysis of the mouse genome. Proc. Natl. Acad. Sci. USA 102, 7221–7226. Bayascas, J.R., Sakamoto, K., Armit, L., Arthur, J.S.C., and Alessi, D.R. (2006) Evaluation of approaches to generation of tissuespecific knock-in mice. J. Biol. Chem. 281, 28772–28781. Pelletier, J. and Sonnenberg, N. (1988) Internal initiation of translation of eukaryotic mRNA directed by a sequence derived from poliovirus RNA. Nature 334, 320–325. Jang, S.K., Krausslich, H.G., Nicklin, M.J., Duke, G.M., Palmenberg, A.C., and Wimmer, E. (1988) A segment of the 5 nontranslated region of encephalomyocarditis virus RNA directs internal entry of ribosomes during in vitro translation. J. Virol. 62, 2636– 2643. Bochov, Y.A. and Palmenberg, A.C. (2006) Translational efficiency of EMCV IRES in bicistronic vectors is dependent upon IRES sequence and gene location. BioTechniques 41, 283–292. Mizuguchi, H., Xu, Z., Ishii-Watabe, A., Uchida, E., and Hayakawa, T. (2000) IRESdependent second gene expression is significantly lower than cap-dependent first gene
76
64.
65.
66. 67.
68.
69. 70.
71.
72.
73.
74.
75.
Koentgen, Suess, and Naf expression in a bicistronic vector. Mol. Ther. 1, 376–382. Okabe, M., Ikawa, M., Kominami, K., Nakanishi, T., and Nishimune, Y. (1997) Green mice as a source of ubiquitous green cells. FEBS Lett. 407, 313–319. Hadjantonakis, A.K., Gertsenstein, M., Ikawa, M., Okabe, M., and Nagy, A. (1998) Generating green fluorescent mice by germline transmission of green fluorescent ES cells. Mech. Dev. 76, 79–90. Shaner, N.C., Steinbach, P.A., and Tsien, R.Y. (2005) A guide to choosing fluorescent proteins. Nat. Methods 2, 905–909. Hadjantonakis, A.K., Dickinson, M.E., Fraser, S.E., and Papaiannou, V.E. (2003) Technicolour transgenics: imaging tools for functional genomics in the mouse. Nat. Rev. Genet. 4, 613–625. Livet, J., Weissmann, T.A., Kang, H., Draft, R.W., Lu, J., Bennis, R.A., Sanes, J.R., and Lichtman, J.W. (2007) Transgenic strategies for combinatorial expression of fluorescent proteins in the nervous system. Nature 450, 56–62. Hopkin, M. (2006) Can super-antibody drugs be tamed? Nature 440, 855–856. Lee, H., Zahra, D., Vogelzang, A., Newton, R., Thatcher, R., Quan, A., So, T., Zwirner, J., Koentgen, F., Padkjaer, S., Mackay, F., Whitfield, P.L., and Mackay, C.R. (2006) Human C5aR knock-in mice facilitate the production and assessment of antiinflammatory monoclonal antibodies. Nat. Biotechnol. 24, 1279–1284. Le, Y., Gagneten, S., Larson, T., Santha, E., Dobi, A., v. Agoston, D., and Sauer, B. (2003) Far-upstream elements are dispensable for tissue-specific proenkephalin expression using a Cre-mediated knock-in strategy. J. Neurochem. 84, 689–697. Nyhan, W.L. (1997) The recognition of Lesch-Nyhan syndrome as an inborn error of purine metabolism. J. Inherit. Metab. Dis. 20, 171–178. Williamson, D.J., Hooper, M.L., and Melton, D.W. (1992) Mouse models of hypoxanthine phosphoribosyltransferase deficiency. J. Inherit. Metab. Dis. 15, 665–673. Oberdoerffer. P., Kanellopoulou, C., Heissmeyer, V., Paeper, C., Borowski, C., Aifantis, I., Rao, A., and Rajewsky, K. (2005) Efficiency of RNA interference in the mouse hematopoietic system varies between cell types and developmental stages. Mol. Cell. Biol. 25, 3896–3905. Bronson, S.K., Plaehn, E.G., Kluckman, K.D., Hagaman, J.R., Maeda, N., and Smithies, O. (1996) Single-copy transgenic
76.
77.
78.
79.
80.
81.
82.
83.
84.
85.
86.
mice with chosen-site integration. Proc. Natl. Acad. Sci. USA 93, 9067–9072. Hatada, S., Kuziel, W., Smithies, O., and Maeda, N. (1999) The influence of chromosomal location on the expression of two transgenes in mice. J. Biol. Chem. 274, 948– 955. Friedrich, G. and Soriano, P. (1991) Promoter traps in embryonic stem cells: a genetic screen to identify and mutate developmental genes in mice. Genes Dev. 5, 1513–1523. Zambrowicz, B.P., Imamoto, A., Fiering, S., Herzenberg, L.A., Kerr, W.G., and Soriano, P. (1997) Disruption of overlapping transcripts in the ROSA geo 26 gene trap strain leads to widespread expression of beta-galactosidase in mouse embryos and hematopoietic cells. Proc. Natl. Acad. Sci. USA 94, 3789–3794. Irion, S., Luche, H., Gadue, P., Fehling, H.J., Kennedy, M., and Keller, G. (2007) Identification and targeting of the ROSA26 locus in human embryonic stem cells. Nat. Biotech. 25, 1477–1482. Kisseberth, W.C., Brettingen, N.T., Lohse, J.K., and Sandgren, E.P. (1999) Ubiquitous expression of marker transgenes in mice and rats. Dev. Biol. 128–138. Strathdee, D., Ibbotson, H., and Grant, S.G.N. (2006) Expression of transgenes targeted to the Gt(ROSA)26Sor locus is orientation dependent. PloS ONE 1, e4. Lai, K.M.V., Gonzalez, M., Poueymirou, W.T., Kline, W.O., Na, E., Zlotchenko, E., Stitt, T.N., Economides, A.N., Yancopoulos, G.D., and Glass, D.J. (2004) Conditional activation of Akt in adult skeletal muscle induces rapid hypertrophy. Mol. Cell. Biol. 24, 9295–9304. Lakso, M., Sauer, B., Mosinger, B., Lee, E.J., Manning, R.W., Yu, S.H., Mulder, K.L., and Westphal, H. (1992). Targeted oncogene activation by site-specific recombination in transgenic mice. Proc. Natl. Acad. Sci. USA 89, 6232–6236. Jackson, E.L., Willis, N., Mercer, K., Bronson, R.T., Crowley, D., Montoya, R., Jacks, T., and Tuveson, D.A. (2001) Analysis of lung tumor initiation and progression using conditional expression of oncogenic K-ras. Genes Dev. 15, 3243–3248. Fire, A., Xu, S., Montgomery, M.K., Kostas, S.S., Driver, S.E., and Mello, C.C. (1998) Potent and specific genetic interference by double-stranded RNA in Caenorhabditis elegans. Nature 391, 806–811. McManus, M.T. and Sharp, P.A. (2002) Gene silencing in mammals by small interfering RNAs. Nat. Rev. Genet. 3, 737–747.
GEM Models for Drug Discovery 87. Hannon, G.J. (2002) RNA interference. Nature 418, 244–251. 88. Novina, C.D. and Sharp, P.A. (2004) The RNAi revolution. Nature 430, 161–164. 89. Ui-Tei, K., Naito, Y., Takahashi, F., Haraguchi, T., Ohki-Hamazaki, H., Juni, A., Ueda, R., and Saigo, K. (2004) Guidelines for the selection of highly effective siRNA sequences for mammalian and chick RNA interference. Nucl. Acids Res. 32, 936–948. 90. Meister, G. and Tuschl, T. (2004) Mechanisms of gene silencing by double-stranded RNA. Nature 431, 343–349. 91. Hasuwa, H., Kaseda, K., Einarsdottir, T., and Okabe, M. (2002) Small interfering RNA and gene silencing in transgenic mice and rats. FEBS Lett. 532, 227–230. 92. Mitani, T. and Yokota, T. (2005) RNA interference as a tool for producing knockdown mice. J. Mamm. Ova Res. 22, 139–151. 93. Hou, J., Schan, Q., Wang, T., Gomes, A.S., Yan, Q.S., Paul, D.L., Bleich, M., and Goodenough, D.A. (2007) Transgenic RNAi depletion of claudin-16 and the renal handling of magnesium. J. Biol. Chem. 282, 17114–17122. 94. Lu, W., Yamamoto, V., Ortega, B., and Baltimore, D. (2004) Mammalian Ryk is a Wnt coreceptor required for stimulation of neurite outgrowth. Cell 119, 97–108. 95. Kissler, S., Stern, P., Takahashi, K., Hunter, K., Peterson, L.B., and Wicker, L.S. (2006) In vivo RNA interference demonstrates a role for Nramp1 in modifying susceptibility to type 1 diabetes. Nat. Genet. 38, 479–483. 96. Grimm D., Streetz, K.L., Jopling, C.L., Storm, T.A., Pandey, K., Davis, C.R., Mar-
97.
98.
99.
100.
101.
102.
103.
77
ion, P., Salzar, F., and Kay, M.A. (2006). Fatality in mice due to oversaturation of cellular microRNA/short hairpin RNA pathways. Nature 441, 537–541. Amar, L., Desclaux, M., Faucon-Biguet, N., Mallet, J., and Vogel, R. (2006) Control of small inhibitory RNA levels and RNA interference by doxycycline induced activation of a minimal RNA polymerase III promoter. Nucl. Acids Res. 34, e37. Szulc, J., Wiznerowicz, M., Sauvain, M.-O., Trono, D., and Aebischer, P. (2006). A versatile tool for conditional gene expression and knockdown. Nat. Methods 3, 109–116. Moosmann, P. Georgiev, O., Thiesen, H.J., Hagmann, M., and Schaffner, W. (1997) Silencing of RNA polymerases II and IIIdependent transcription by the KRAB protein domain of KOX1, a Kruppel-type zinc finger factor. Biol. Chem. 378, 669–677. Fritsch, L., Martinez, L.A., Sekhri, R., Naguibneva, I., Gerard, M., Vandromme, M., Schaeffer, L., and Harel-Bellan, A. (2004) Conditional gene knock-down by Cre-dependent short interfering RNAs. EMBO Rep. 5, 178–182. Hitz, C., Wurst, W., and Kuhn, R. (2007) Conditional brain-specific knockdown of MAPK using Cre/loxP regulated RNA interference. Nucl. Acids Res. 35, e90. Steuber-Buchberger, P., Wurst, W., and Kuhn, R. (2008) Simultaneous cre-mediated conditional knockdown of two genes in mice. Genesis 46, 144–151. Tiscornia, G., Tergaonkar, V., Galimi, F., and Verma, I.M. (2004) Cre recombinaseinducible RNA interference mediated by lentiviral vectors. Proc. Natl. Acad. Sci. USA 101, 7347–7351.
Chapter 5 Profiling of Drug Action Using Reporter Mice and Molecular Imaging Gianpaolo Rando, Andrea Biserni, Paolo Ciana, and Adriana Maggi Abstract Reporter mice associated to molecular imaging represent a major asset for the study of the spatio-temporal effects of drugs in living animals. The field is still relatively young and so far the number of animals genetically modified to express a given reporter gene ubiquitously and under the control of specific drugs is still limited. For a reporter animal the indispensable elements for the application to drug research and development are (i) the short life of the reporter enabling to have a clear view of the onset as well as the termination of drug effects, (ii) the generalized, drug-dependent activation of the reporter, and (iii) imaging modality suitable for high-throughput analysis. Because of its relative cheapness and ease to perform, in addition to all the above considerations, bioluminescence-based imaging is now regarded as the best imaging technology to be applied to the field of drug research. We show here the application of reporter mouse systems for drug screening in living animals in order to compare drug potency on target and specificity of action. Key words: Reporter mouse, pharmacodynamics, pharmacokinetics, bioluminescence imaging, ERE-Luc, PPRE-Luc, luciferase, brain imaging.
1. Introduction We recently generated reporter mice for the study of compounds active through intracellular receptors. In these transgenic mice, the luciferase reporter gene is under the control of a hormone responsive element (HRE) responsive promoter; the generalized expression of the transgene is guaranteed by the presence of insulator sequences flanking the construct (1, 2). These mice have been shown to be useful tools for the study of the physiological activity of endogenous ligands (2, 3). Since luciferase activity G. Proetzel, M.V. Wiles (eds.), Mouse Models for Drug Discovery, Methods in Molecular Biology 602, DOI 10.1007/978-1-60761-058-8 5, © Humana Press, a part of Springer Science+Business Media, LLC 2010
79
80
Rando et al.
can be measured in vivo by bioluminescence, these animal models may, however, also represent a novel relevant system for the screening of new chemical entities with potential pharmacological activity, enabling the rapid comparative study of drug potency and specificity in acute and long-term administration (4). However, the main limitation of in vivo imaging by bioluminescence imaging (BLI) is due to the restricted penetration through tissue layers of the photons generated by the activity of the enzyme on its substrate luciferin and to the fact that bioluminescence can be analyzed only in two dimensions. This limitation can be overcome by associating the initial analysis performed in living animals to the ex vivo analysis of luciferase in dissected organs by BLI and the quantitative measurement of luciferase enzymatic activity in tissue homogenates. In our experience, ex vivo imaging in dissected organs can be quite informative as it provides a clear view of the diffusion of the chemical entity examined in each specific organ. In a drug screening program the effects of drugs are initially studied after acute exposure using a classical time course and concentration–effect analysis. However, since most drugs are administered for prolonged periods of time, a long-term longitudinal study is advisable because it could reveal unexpected activities of the compound during the study due to downregulation of the target receptors, accumulation in specific organs, or others unpredictable factors. In this latter application the possibility to follow drug action in vivo is a major asset of in vivo imaging.
2. Materials 2.1. Transgenic Reporter Mice
2.2. In Vivo and Ex Vivo BLI
ERE-Luc and PPRE-Luc reporter mice (TOP s.r.l. Lodi Milan, Italy) were generated by pronuclear DNA injection of zygotes C57BL/6xDBA/2 F2 using standard procedures. Microinjected zygotes were reimplanted into pseudopregnant C57BL/6xDBA/2 F2 foster mothers to bring to term. Founders were backcrossed with C57BL/6 (Charles River, Wilmington, MA) wild type for 15 generations. 1. Bioluminescence imaging unit: Night Owl imaging unit (Berthold Technologies, Bad Wildbad, Germany) consisting of a Peltier cooled charge-coupled device (CCD) slow-scan camera equipped with a 25 mm/f 220 0.95 lens located in a light-tight chamber, or Xenogen IVIS Lumina System (Caliper Life Sciences, Hopkinton, MA, USA) consisting of a scientific grade thermoelectrically cooled CCD camera mounted on a light-tight imaging chamber.
Profiling of Drug Action
81
2. Brain matrix sectioning device for adult mouse, coronal and sagittal, 1 mm spacing (Ted Pella, Redding, CA, USA). 3. Syringes and micro-dissection tools, including forceps, scissors, and blades (Biological Instruments, Besozzo, Italy). 4. Anesthetic solution for mice: 78% v/v ketamine (Ketavet 50, Intervet, Peschiera Borromeo, Italy), 15% v/v Xilazine (Rompun 2% solution, Bayer, Leverkusen, Germany), and 7% v/v water for injections (Galenica Senese, Monteroni d’Arbia, Italy). Alternatively, gas anesthesia using isofluorane (Isofluorane-Vet, Merial, Lyon, France) and Xenogen XGI-8 Gas Anesthesia System (Caliper Life Sciences) with the following setting: vaporizer value 2.5%; oxygen flow 1.5 L/min in the induction chamber and 0.25 L/min to the mice during the in vivo imaging in the CCD camera. 5. Luciferin: D-luciferin powder (beetle luciferin potassium salt, Promega, Madison, WI, USA) resuspended 25 mg/mL in water for injection (Galenica Senese). 2.3. Luciferase Enzymatic Assay
1. Tissue homogenizer (TissueLyser, Qiagen, Hilden, Germany), including 5 mm inox beads (Qiagen, Hilden, Germany) and bead dispenser (Qiagen, Hilden, Germany) 2. Refrigerated centrifuge and rotor for plates (Rotanta 460R, Hettich Zentrifugen, Germany) 3. Microplate luminometer (Glomax, Promega Madison, WI, USA) 4. Microplate absorbance reader with filter at 595 nm (EM8600, Bio-Rad Laboratories, Segrate, Italy) 5. P20, P200, P1000 single pipettors (Pipetman Starter Kit, Gilson, Middleton, WI, USA) and P20 and P200 multichannel pipettors (Gilson, Middleton, WI, USA) 6. 1.2 mL deep well strips racked (Starlab, Milano, Italy) 7. Transparent 96-well microplates (Corning, Torino, Italy) 8. White (not transparent) 96-well microplates (Corning, Torino, Italy) 9. Phosphate lysis buffer. Prepare 50 mL by adding 25 mL K-phosphate buffer 0.2 M (Sigma-Aldrich, St Louis, MO, USA), 50 L DTT 1 M (Sigma), 1 mL EGTA 0.2 M (Sigma), 0.4 mL sterile EDTA 0.5 M (Sigma), and bidistilled water to 50 mL. Keep at 4 ◦ C 10. Recombinant firefly luciferase (QuantiLum, Promega) 11. Luciferase Assay System (Promega): composed of (i) trycine buffer to dilute recombinant luciferase and (ii) luciferin buffer, containing a patented mix of luciferase substrates and coenzymes including D-luciferin, ATP, Mg2+, CoA
82
Rando et al.
12. Bradford Protein Assay Kit (Pierce, Rockford, IL), including protein standards (Pierce) 13. Bovine serum albumin (BSA), lyophilized powder (Sigma) 2.4. Analysis of Imaging Data
Usually, imaging workstations and luminometers include proprietary software to operate the devices and allow the direct export of raw data and basic measurements to common spreadsheet software, including Microsoft Excel. More complex image analyses can be successfully performed with ImageJ or Matlab software packages. Statistical analyses were performed with GraphPad Prism version 5.01 for Windows, GraphPad Software (San Diego, California, USA). Given the possibility to longitudinally image reporter mice over long periods of time, imaging data sets tend to expand very rapidly; in these cases the large amount of data are best stored in Microsoft Access or Filemaker databases which allow easier data retrievals and comparisons.
3. Methods 3.1. Experimental Determination of the Linearity Range of Photon Emission as Measured by CCD Camera
Although the luciferase assay is claimed to be linear over a wide range of magnitude, different imaging workstations may differ in term of linearity and sensitivity. Artificial light standards are available on the market to evaluate the detector performances; however, given the wide spectral differences between different luciferases utilized as reporter genes, it is strongly advised to test the range of linearity of the CCD camera with the same luciferase reporter utilized in the mouse model under study, as follows: 1. A solution of recombinant firefly luciferase is serially diluted (1:5) with luciferase buffer containing 1 mg/mL BSA (see Note 1). 2. 20 L of each dilution, e.g., use 1:5 dilutions in the range 1,000,000 to 0.01 pg of luciferase are plated in triplicate in transparent or black microplates and 100 L of D-luciferin buffer are added. 3. After 5 min, the time necessary to obtain a steady state of the enzymatic reaction, photon emission from the microplate is measured with the CCD camera using the same setting (i.e., height of the camera, integration time, binning) adopted for mouse imaging acquisition. 4. Photon emission from each single well is acquired and the values are expressed with the same unit of measurement adopted for mouse imaging acquisition (i.e., cts/cm2 sec).
Profiling of Drug Action
83
5. In order to calculate the CCD camera linear dynamic range of the photon emission, the average measurements of each dilution are plotted against the logarithm of the corresponding amount of recombinant luciferase. If care is taken to avoid overlapping of signals from wells with high and low luciferase content, a typical sigmoid curve is obtained (Fig. 5.1). Photon emission should be considered proportional to the amount of luciferase only in the linear part of the curve. In our experience, we never found in reporter mice values of photon emission which were able to saturate the CCD camera sensor. Conversely, particular care must be taken in handling and interpreting raw measurements lower than the minimum threshold of linearity; in this case, the photon emissions are not directly proportional to the amount of luciferase, and arithmetical processing like fold-induction normalization should be avoided.
Fig. 5.1. Photon emission (as measured in cts/cm2 s using the CCD camera) was plotted against log10 concentration of recombinant luciferase used in the enzymatic assay. The range of linearity in this study was between 250 and 1700 cts/cm2 s.
3.2. Evaluation of Substrate D-Luciferin Distribution by In Vivo Imaging: Time Course
Any semi-quantitative evaluation of photon emission derived from the luciferin/luciferase enzymatic reaction by in vivo imaging requires prior study of the diffusion to all organs of the enzyme substrate, D-luciferin, at a concentration sufficient to saturate the reporter enzymatic activity (4). Indeed the diffusion of the substrate may change depending on the strain of mouse used or the specific formulation of the substrate. Here follows the protocol we adopted to this aim: 1. Five reporter mice are anesthetized using a subcutaneous (s.c.) injection of 50 L ketamine–xylazine solution or by gas anesthesia exposing mice for 2 min to isofluorane. They are then injected intraperitoneal (i.p.) with 80 mg/kg Dluciferin. 2. Mice are placed in the light-tight chamber and a gray scale photo of the animals is taken with dimmed light.
84
Rando et al.
Fig. 5.2. Time course of luciferin distribution in ERE-Luc reporter mice. Luciferin (80 mg/kg) was injected i.p. and then imaging was carried out with a series of 5 min imaging sessions. Left panel: in vivo imaging of one, representative, individual. Right panel: mean ± sem of photon emission measured in the chest area at different times after substrate injection to a group of five animals.
3. In order to be able to follow photon emission in time, photon emission is then measured with a series of 5 min imaging sessions (Fig. 5.2). Generally, photon emission is maximal between 10 and 25 min after the i.p. injection and then gradually decreases. 3.3. Evaluation of Substrate D-Luciferin Distribution by In Vivo Imaging: Dose–Response
Once the window of time is established which guarantees optimal distribution of the substrate to the different tissues (typically 20 and 25 min), this time is selected for defining the optimal substrate dosage (see Note 2). To this aim: 1. Five reporter mice/experimental group are anesthetized as described in Section 3.2. 2. Mice are injected i.p. with increasing concentrations of luciferin (10, 25, 50, and 100 mg/kg).
D-
3. At the time point previously selected as suitable for the optimal distribution of the substrate (20 min) animals are subjected to a 5 min imaging session (Fig. 5.3).
Fig. 5.3. Dose–response of luciferin distribution in PPRE-Luc reporter mice. Luciferin was injected i.p. and then imaging was carried out 20 min after substrate injection. Data in the right graph represents the mean ± sem of photon emission measured in the chest and abdomen area of groups of five animals.
Profiling of Drug Action
85
4. Plotting the photon emission from the required specific body area establishes the concentration of luciferin sufficient to carry out future studies. 3.4. Screening of Compounds with Unknown Pharmacological Activity: Time Course
Time course and concentration–effect studies are greatly facilitated by in vivo imaging with reporter mice. The initial pilot experiments can be carried out with a very limited number of animals (e.g., 2–3) to establish the best time and concentration at which each compound is most active. Once the relative potency of the chemical entities under study is established, additional investigation can be carried out to evaluate the specificity of action of the given drug in each tissue. 1. Groups of two to three mice are treated with: (a) the same concentration of the compounds of interest, (b) the reference ligand of known activity, and (c) the vehicle. 2. Whole body photon emission in each animal is measured at different times after ligand injection (typically at 1, 3, 6, 12, 24 h). 3. For a comparative analysis of the maximal effect of each compound in target tissues, photon emission is measured in selected body areas as shown in Fig. 5.4. Data are expressed as photon counts/time/surface unit to assess the relative effect of the treatments in the different body areas. These experiments allow the assessment of relative potency for each compound on the target in different organs and to define their action over time after a single administration.
Fig. 5.4. Comparative potency of different estrogen receptor (ER) compounds in adult ovx (ovariectomized) ERE-Luc mice. Mice were ovariectomized to decrease the levels of endogenous estradiol. Animals were treated with a concentration of 5 g/kg of reference compound (17-beta-estradiol) or compounds of unknown activity (500 g/kg). BLI was carried out 6 h after s.c. injection of each compound. Photon emission was measured in the body segmented as depicted in the left panel. Data were expressed as percentage of photon emission vs vehicle to be able to assess the relative potency of the treatments in the different body areas. Bars represent mean ± sem of a minimum of five animals/group.
86
Rando et al.
3.5. Screening of Compounds with Unknown Pharmacological Activity: Dose–Response
Once the time is defined where the activity of each given compound is highest, a dose–response curve is carried out in order to establish the concentration necessary to provide the best signal to noise ratio in the desired target tissue. The dose–response curve is usually carried out with groups of 5–8 animals and establishes the dose with the highest potency of the compound of interest on the target; generally the study is carried out by measuring the drug effect in different organs at the time selected as in described in Section 3.4. At the end of the in vivo imaging experiment, animals are euthanized and photon emission is measured as defined in Section 3.6, tissues are also frozen for the quantitative assessment of luciferase content as specified in Section 3.7 or a more precise analysis of the effect of the compounds in the study. Once the dose is defined at which the compound has the desired effect on its target, a further time-course study can be performed to verify the drug pharmacodynamics at the desired concentration of the compound. This study is very important as it verifies the extent to which the chemical entity is active on its target in a certain tissues and can highlight any unexpected activities in other tissues where the activity of the drug may cause undesired effects. For compounds which may have mixed agonist/antagonist activity in different organs (SERMs, SPPARMs, others) the study provides data on the agonist potency in each tissue; antagonist activity can then be estimated in the presence of a known agonist.
3.6. Ex Vivo Semi-quantitative Measurement of Luciferase Activity
Present technology for bioluminescence-based in vivo imaging can only be consistently carried out in reporter mice in two spatial dimensions, thus the definition of the organ/tissue contributing to the photon emission as measured in vivo is limited (see Note 3). Furthermore, signaling from the most deeply placed organs is significantly reduced by photon scattering and absorption by the intervening tissues. Therefore, to better evaluate the contribution of each organ to photon emissions measured in living mice, the analysis can be refined at the end of the in vivo imaging session by euthanizing the mice and dissecting all organs for additional more detailed analysis. 1. At the end of the in vivo imaging session, animals are euthanized, tissues are dissected, and placed in a pre-ordered manner in a Petri dish cooled on a bed of ice. The dissection must be carried out quite rapidly (maximally 15 min) to ensure that the substrate luciferin remains in the tissue at sufficient concentration and the luciferase is not catabolized by postmortem biochemical events. 2. Photon emission is measured for each tissue with CCD camera exposure of 1–5 min.
Profiling of Drug Action
87
Fig. 5.5. Left panel: photon emission was measured by CCD camera ex vivo in tissues dissected (left panel) after administration of a compound X. Quantitative measurement of photon emission (right panel) shows increased accumulation of luciferase in white adipose tissue and liver after treatment.
3. Photon emission in each organ is then measured as photon/sec/cm2 /steradiant as shown in Fig. 5.5. 3.7. Ex Vivo Semi-quantitative Measurement of Luciferase Activity in Brain
Traditional optical imaging techniques cannot be satisfactorily applied to the study of the central nervous system because of a series of methodological limitations. Among these are (i) luciferin has a very low permeability through the blood–brain barrier, (ii) the skull prevents bioluminescence to be efficiently detected by the CCD camera, and (iii) no three-dimensional reconstruction of the brain is practical, making it impossible to correctly attribute signals to the respective emitting areas of the brain. Therefore, we developed a functional method to precisely localize and quantify nuclear receptor transcriptional activity in selected mouse brain regions. With this technique (see Fig. 5.6), we are able to observe for the first time that the estrous cycle significantly influences estrogen receptors (ER) activity in brain areas, such as limbic areas which are not related to reproduction (5) and the existence of a sex dimorphism in the brain peroxisome proliferator-activated receptors (PPAR) signaling in response to nutritional changes (6). 1. Animals are anesthetized with s.c. injection of ketamine– xylazine solution or gas anesthesia using isofluorane. 2. Luciferin (3 L of 26.5 mg/mL luciferin solution in NaCl 0.9%) is administered intracerebro-ventricularly (i.c.v.) to anesthetize animals. The injection is carried out according to specific stereotaxic coordinates (bregma, −0.25 mm; lateral, 1 mm; depth, 2.25 mm) by the use of a Hamilton syringe rotated on the coronal plate about 3◦ from the orthogonal position. 3. 20 min after administration of luciferin animals are killed by cervical dislocation.
88
Rando et al.
Fig. 5.6. Effect of treatment with 17-beta-estradiol and ICI 182,780 (ICI) on photon emission from ERE-Luc male brain. Upper panels: optical imaging of representative brains of treated ERE-Luc mice. In the control group, animals received corn oil s.c. as 17-beta-estradiol vehicle, and ethanol i.c.v. as ICI vehicle; in the ICI group, mice received corn oil s.c. and ICI 182,780 1.3 mg/kg i.c.v.; in the 17-beta-estradiol group, mice received 17-beta-estradiol 50 g/kg s.c. and ethanol i.c.v. Upper low left panel: electronic grid delimiting specific brain areas to be used for photon emission counting, created according to Paxinos and Franklin, superimposed to a black-and-white photograph of brain sections. Lower panel: comparative analysis of photon emission in brain areas. Numbers on the x-axis correspond to the areas indicated in the brain sections. Data are expressed as the average photon emission in the area unit (cts/sec/mm2 ) and represent the mean ± sem (n = 6–10). Statistical differences (∗vehicle vs. 17-beta-estradiol; ◦ vehicle vs. ICI) are reported: ◦ /∗ p < .05; ◦◦ /∗∗ p < .01; ◦◦◦ /∗∗∗ p < .001. Abbreviations: Acb, nucleus accumbens; Amy, Amygdala; Arc, arcuate nucleus; CPu, caudate putamen; Hipp, hippocampus; Hyp, hypothalamus; MCx, motor cortex; PCx, parietal cortex; PyrCx, pyriform cortex; Sn, septal nuclei; Tha, thalamus.
4. Brains are rapidly dissected and sectioned by means of a “brain matrix” device and placed in a pre-ordered position on a Petri dish cooled on ice. 5. Sections are immediately visualized by the CCD camera in the light-tight chamber. 6. Grayscale images of the sections are first taken with dimmed light. 7. Photon emission is measured over a 15 min exposure time.
Profiling of Drug Action
89
8. Pseudocolor images associated with photon emission are generated by the CCD image processor and transferred via video cable to a PCI frame grabber using the camera software (or equivalent). 9. For co-localization of the bioluminescent photon emission, grayscale and pseudocolor images are merged using the appropriate software. 10. Luminescence of the brain slices is expressed as the integration of photon emission per time unit (cts/sec); to be able to compare the extent of photon emission in brain nuclei characterized by different areas, the data are expressed as cts/sec/mm2 . Photon emission in selected brain areas is quantified by means of a grid generated with the aid of a brain atlas. 3.8. Ex Vivo Quantitative Assessment of Luciferase Enzymatic Activity in Tissue Extracts
1. Tissues (or selected brain regions) are manually dissected, immediately frozen on dry-ice, and stored at −80◦ C. 2. Tissues are homogenized by TissueLyser using 300 L of the phosphate lysis buffer and putting a stainless steel bead in each 1.2 mL polyethylene microtube. 3. Homogenates are centrifuged at 5900 xg for 30 min 4◦ C. 4. 20 L of the supernatant is transferred to a white opaque 96-well plate for luminescence quantification by luminometer, with an integration time of 10 sec, after machine-driven injecting 100 L of luciferase assay reagent. 5. Light measurements are recorded by the luminometer software GloMax and protein concentrations in the supernatants are measured using a Bradford Protein Assay Kit following the manufacturer’s protocols and analyzed with an EM680 microplate reader. 6. The luminescence data, normalized over protein content of each sample, are expressed as relative light units (RLU) per g of protein.
3.9. Profiling Drug Activity During Repeated Administration of Compounds with Potential Pharmacological Activity
Most of the drugs currently in study aim at the treatment of chronic disorders and need to be administered for a prolonged period of time. The efficacy of the treatment is evaluated mainly by the study of the model disease and evaluating recovery from the pathological status. The availability of appropriate models where the direct effect of a treatment on the drug target can be measured directly tremendously improves the predictability of preclinical studies. Furthermore, the whole body molecular imaging followed by the study of the molecular events elicited by the treatment of the experimental animal may reveal unexpected effects in tissues; for example, where the repeated administration of the compound may cause toxicity (e.g., liver or others). For the
90
Rando et al.
Fig. 5.7. Left panel: daily imaging of an exemplificative mouse individual treated with tamoxifen at the daily dose of 0.5 mg/day s.c. Right panel: photon emission by abdomen in mice treated daily for 21 days with vehicle (controls); 17-beta estradiol (5.5 g/kg/day s.c.); tamoxifen al low dosage (5.5 g/kg/day) or at high dosage (0.5 mg/kg/day). Luciferase expression appears to become more elevated at the end of the treatment with tamoxifen indicating, in this tissue, an unpredicted, potential effect of the treatment during prolonged treatments with the drug.
evaluation of the effects of a given compound in chronic studies the investigation is carried out in the following steps: 1. Groups of five to eight animals are caged separately and treated daily by the selected route of administration. 2. Mice are subjected to bioluminescence imaging by CCD camera daily (Fig. 5.7), preferably in the early afternoon. This avoids the confounding effects due to food consumption that occurs mostly during the night (see Notes 4 and 5). 3. Calculate and plot data according to the description provided in Section 3.4.
4. Notes 1. BSA stabilizes luciferase at lower concentrations. 2. Luciferin dosage must be sufficient to saturate the enzymatic reaction both in vivo and ex vivo in order to obtain meaningful data when comparing photon emission among the different experimental animals. 3. The major limitation of in vivo bioluminescent imaging is that embodies only bi-dimensional information. To overcome this limitation most current CCD camera imaging systems can perform several CCD acquisitions at different camera angle reconstructing a tridimensional reconstruction using specific informatics tools (mice body atlas and particular algorithms). Another way to overcome this limitation is the generation of multimodal reporter mice equipped with
Profiling of Drug Action
91
a second reporter that can be used in nuclear imaging to obtain all the information related to the third dimension. 4. Animal diet may influence in vivo imaging. Thus mice must be kept under standard alimentary condition previously tested to be sure that there are no confounding modulations of reporter expression due to the feed. For instance, alfalfa derivatives present in soy-based diets may increase abdominal autofluorescence (7). 5. The pattern of photon emission observed in longitudinal studies can be influenced by contextual aspects that must be taken into consideration: a. Animals may become resistant to anesthetics after longterm studies, and some anesthetics (e.g., isofluorane) have also been described to inhibit the activity of luciferase (8, 9). b. The toxicity of D-luciferin after long-term exposures has not been fully characterized. However, we failed to observe major anatomical defects (e.g., liver and kidney weight) after treating animals with 50 mg/kg of Dluciferin every day, for up to 3 months. c. The compound or drug tested may interfere with the luciferin–luciferase reaction through altering the pharmacokinetics of the substrate; e.g., differences in ABCG2/BCRP expression and function have been recently described to modify bioluminescent output in vivo (10). Hence, drugs potentially affecting multidrugresistant genes may alter the pattern of photon emission by influencing the bioavailability of D-luciferin. d. Analysis of the data in longitudinal studies can be a hurdle. One hundred animals can be reasonably processed in a day by 2–3 trained scientists. However, this casts a great operational complexity to the data analysis; this usually includes time consuming identification and quantification of several anatomical signal areas belonging to different mice. In addition, such analysis requires expert and trained personnel able to appropriately identify the specific areas of photon emission where drug effects can be measured. Algorithms enabling the identification (segment) of region of interests, allowing the identification and quantification of photon emitting areas from the mouse body are needed. A recently developed algorithm has been designed to recognize the different parts of the animals using the exact shape of the photon emitting area (11). Compared to the manual methods offered by current optical imaging software (Fig. 5.3), the use of such algorithms is expected to save time and to allow
92
Rando et al.
a more reliable operator-independent calculation of the drug effects in throughput screening experiments with reporter mice.
Acknowledgment The authors are grateful to the European Community for the significant and continued contribution to the work here shown with grant support: EDERA (EC-QLK4-CT-2002-02221); NoE DIMI (LSHB CT 2005 512146); NoE EMIL (LSHC CT 2004 503569); Strep EWA (LSHM CT 2005 518245); IP CRESCENDO (LSHM CT 2005 018652). References 1. Ciana P, Di Luccio G, Belcredito S, Pollio G, Vegeto E, Tatangelo L, Tiveron C, Maggi A. (2001) Engineering of a mouse for the in vivo profiling of estrogen receptor activity. Mol Endocrinol. 15(7), 1104–13. 2. Ciana P, Biserni A, Tatangelo L, Tiveron C, Sciarroni AF, Ottobrini L, Maggi A. (2007) A novel peroxisome proliferatoractivated receptor responsive elementluciferase reporter mouse reveals gender specificity of peroxisome proliferatoractivated receptor activity in liver. Mol Endocrinol. 21(2), 388–400. 3. Ciana P, Raviscioni M, Mussi P, Vegeto E, Que I, Parker MG, Lowik C, Maggi A. (2003) In vivo imaging of transcriptionally active estrogen receptors. Nat Med. 9(1), 82–6. 4. Biserni A, Giannessi F, Sciarroni AF, Milazzo FM, Maggi A, Ciana P. (2008) In vivo imaging reveals selective peroxisome proliferator activated receptor modulator activity of the synthetic ligand 3-(1-(4-chlorobenzyl)3-t-butylthio-5-isopropylindol-2-yl)-2,2dimethylpropanoic acid (MK-886). Mol Pharmacol. 73(5), 1434–43. 5. Stell A, Belcredito S, Ciana P, Maggi A. (2008) Molecular imaging provides novel insights on estrogen receptor activity in mouse brain. Mol Imaging. 7(6), 283–92.
6. Ibarra C, Bertocchi I, Ciana P, Biserni A, Eva C, Maggi A. (2009) Gender-dimorphic activity of brain PPARs in response to changes in the nutritional status (unpublished results). 7. Bouchard MB, MacLaurin SA, Dwyer PJ, Mansfield J, Levenson R, Krucker T. (2007) Technical considerations in longitudinal multispectral small animal molecular imaging. J Biomed Opt. 12(5), 051601. 8. Szarecka A, Xu Y, Tang P. (2007) Dynamics of firefly luciferase inhibition by general anesthetics: Gaussian and anisotropic network analyses. Biophys J. 93(6), 1895–905. 9. Cui K, Xu X, Zhao H, Wong S. (2008) A quantitative study of factors affecting in vivo bioluminescence imaging. Luminescence 23, 292–95. 10. Zhang Y, Bressler JP, Neal J, Lal B, Bhang HE, Laterra J, Pomper MG. (2007) ABCG2/BCRP expression modulates Dluciferin based bioluminescence imaging. Cancer Res. 67(19), 9389–97. 11. Rando G, Casiraghi E, Arca S, Campadelli P, Maggi A. (2009) Automatic segmentation of mouse images. In: Stereology and Image Analysis. Ecs10-Proceedings of the 10th European Congress of ISS, (V.Capasso et al. Eds.), The MIRIAM Project Series, ESCULAPIO Pub. Co., p.60.
Chapter 6 Human FcRn Transgenic Mice for Pharmacokinetic Evaluation of Therapeutic Antibodies Derry C. Roopenian, Gregory J. Christianson, and Thomas J. Sproule Abstract Therapeutic monoclonal antibodies are widely recognized to be a most promising means to treat an increasing number of human diseases, including cancers and autoimmunity. To a large extent, the efficacy of monoclonal antibody treatment is because IgG antibodies have greatly extended persistence in vivo. However, conventional rodent models do not mirror human antibody pharmacokinetics. The key molecule responsible for the extended persistence antibodies is the major histocompatibility complex class I family Fc receptor, FcRn. We describe human FcRn transgenic mouse models and how they can be exploited productively for the preclinical pharmacokinetic evaluation of therapeutic antibodies. Key words: Monoclonal antibodies, Fc-fusion proteins, FcRn, transgenic mice, pharmacokinetics, serum half-life.
1. Introduction 1.1. Antibodies
As realized by Frank Macfarlane Burnet (1), antibodies are considered nature’s magic bullets. The antigen binding sites permit the antibody to bind its ligand with extraordinary specificity. Sequences in the Fc fragment then focus events, such as the activation of complement and Fc receptors, to the antigenic moiety (Fig. 6.1). This coupling of specificity to effector function is the basis for humoral immunity. Of the various antibody classes (IgM, IgA, IgE, and IgG), IgG has features that make it the preferred antibody for therapeutic development. IgG is a highly stable tetrameric molecule, comprised of two light chains and two heavy chains. It is the dominant antibody class in circula-
G. Proetzel, M.V. Wiles (eds.), Mouse Models for Drug Discovery, Methods in Molecular Biology 602, DOI 10.1007/978-1-60761-058-8 6, © Humana Press, a part of Springer Science+Business Media, LLC 2010
93
94
Roopenian, Christianson, and Sproule
Fig. 6.1. Rendition of an IgG antibody molecule and its interaction sites. V, variable regions; L, light chain; VH1 and CH 1-4, constant domains of the heavy chain. Sites of C1q, FcgR, and FcRn binding are indicated.
tion because it has greatly extended persistence. For example, the half-life of IgG1 in human circulation is 10–20 days, 10 times the half-life for antibodies of other isotypes. IgG is also distinguished from IgM, IgA, and IgE in that it readily passes from the circulation to extravascular tissues. Like the effector functions associated with IgG, it is the Fc fragment of IgG that is responsible for the increased serum persistence and extracellular access. 1.2. FcRn Protects IgG Antibodies from Catabolic Destruction
FcRn is quite distinct from conventional Fc receptors and is an evolutionary offshoot of major histocompatibility complex (MHC) class I proteins (2, 3). Like other MHC class I family members, FcRn forms an obligate heterodimer with the 2 microglobulin (2 m) light chain. FcRn binds the CH2-CH3 hinge region of the Fc fragment in an acidic environment (pH ∼6–6.5) with nanomolar affinity, but demonstrates a precipitous affinity drop at neutral pH ∼7.2 (4–6). The steady-state localization is primarily in the early endosomes. In cells, such as vascular endothelial cells, extracellular proteins (including IgG) are taken up by fluid phase endocytosis. IgG/FcRn complexes are then recycled to the plasma membrane leading to the dissociation of IgG at neutral pH and release into the extracellular environment (Fig. 6.2) (reviewed in 7–9). In this way, FcRn is thought to selectively “protect” IgG from catabolic elimination by recycling the IgG cargo either apically (back to the bloodstream) or
Human FcRn Transgenic Mice
95
Fig. 6.2. Model for how FcRn rescues IgG from catabolism by recycling and transcytosis. IgG and many other soluble proteins are present in extracellular fluids. Vascular endothelial cells are active in fluid phase endocytosis of blood proteins. Material taken up by these cells enters the endosomes where FcRn is found as an integral membrane protein. The IgG then binds FcRn in this acidic environment. This binding results in transport of the IgG to the apical plasma membrane for recycling into the circulation, or to the basolateral membrane for transcytosis into the extracellular space. Exposure to a neutral pH in both locations then results in the release of IgG. The remaining soluble proteins are channeled to the lysosomal degradation pathway.
basolaterally (across endothelial or epithelial barriers). As such, FcRn increases both the serum persistence and the extravascular bioavailability of IgG antibodies. In contrast, proteins that do not bind FcRn in this acid pH-dependent manner are not rescued by FcRn and thus channeled into the lysosomal compartment where they are catabolized. The failure to be rescued by FcRn thus explains the abbreviated serum half-life of most proteins in the circulation compared to IgG.
96
Roopenian, Christianson, and Sproule
1.3. Therapeutic Antibodies
As of 2008, 25 therapeutic monoclonal antibodies (mAbs) mAbs had been approved for clinical use in the United States, and with over 400 antibodies being in preclinical and clinical development further increase of antibody therapies is assured (10, 11). As a general rule, the Fc fragment is a key component of therapeutic mAb design because it extends their pharmacokinetics. Inclusion of the Fc from IgG is also a key component of other bioactive proteins where prolongation of pharmacokinetics is desired, e.g., the tumor necrosis factor receptor (TNFR) fusion protein etanR ercept (Enbrel ) (12). Thus for both therapeutic antibodies and Fc-fusion proteins, the FcRn interaction is a generalized way to exploit FcRn protection to achieve the benefits of extended persistence in vivo. There are substantial constraints on the species source of the Fc fragment that is incorporated into the therapeutic mAb. The first therapeutic mAb, OKT3, exhibited very rapid clearance when administered to humans due to being a standard mouse mAb. A primary cause for OKT3’s abbreviated serum persistence is because mouse IgG Fc acts in a species-specific manner by neither binding nor being protected by human FcRn (13). It has thus become standard to “chimerize” therapeutic mAbs by replacing the mouse Fc with the human IgG Fc counterpart (most commonly of the IgG1 subclass) to harness the benefits of FcRn protection while reducing immunogenicity in humans. A further modification to reduce immunogenicity is to more fully “humanize” the mAb by the replacement of certain amino acids within the variable region. There are also continuing adaptations to the above “first generation” mAb therapeutics to create more effective designer mAbs. These include (i) amino acid substitutions in the Fc region to maximize the interaction with human FcRn and prolong the pharmacokinetics, (ii) substitutions that eliminate FcRn interaction and minimize half-life, and (iii) modifications in the Fc region that alter binding to conventional Fc receptors or complement, with the goal of enhancing or eliminating effector functions without interfering with FcRn binding (reviewed in 7, 14, 15).
1.4. Conventional In Vivo Model Systems for Therapeutic mAb Evaluation
Conventional rodent model systems have proven problematic as they do not reliable model the pharmacokinetics of humanized mAb and Fc-fusion proteins. In contrast to the failure of mouse mAbs to be protected by human FcRn, humanized mAbs have an abnormally high affinity for mouse and rat FcRn, resulting in an artificially prolonged serum persistence (13, 16). This fact has greatly diminished the preclinical utility of standard mice for therapeutic mAb development and testing. The alternative cynamologous monkey model has proven to be reliable, but it is hampered by considerable expense and ethical concerns that limit its routine use.
Human FcRn Transgenic Mice
97
2. Materials 2.1. hFCRN Mouse Models
1. B6.Cg-Fcgrt
Tg(FCGRT)276Dcr (The Jackson Laboratory, Bar Harbor, stock number 004915); abbreviated B6.mFcRn−/− hFCRN Tg276 2. B6.Cg-Fcgrt Tg(FCGRT)32Dcr (The Jackson Laboratory, stock number 004919); abbreviated B6.m FcRn−/− hFCRN Tg32 3. B6.129X1-Fcgrt/Dcr (The Jackson Laboratory, stock number 003982); abbreviated FcRn−/− mice
2.2. Monitoring Mouse Genotypes and FCRN Expression
1. Wild-type murine genotypic FcRn forward primer (FcRn-F): GGGATGCCACTGCCCTG; reverse primer (FcRn-R): CGAGCCTGAGATTGTCAAGTGTATT 2. Targeted murine genotypic FcRn forward primer (T-Fc Rn-F): GGAATTCCCAGTGAAGGGC; reverse primer (T-FcRn-R): CGAGCCTGAGATTGTCAAGTGTATT 3. Human FCRN forward genotypic primer (hFcRn-F): AGCCAAGTCCTCCGTGCTC; reverse primer (hFcRn-R): CTCAGAGATGCCAGTGTTCC 4. Antibodies against human FCRN: ADM31 and ADM32 (Dr. Roopenian, The Jackson Laboratory) 5. AlexaFluor #A20006)
647
label
(Invitrogen,
Carlsbad,
USA,
6. FACSCalibur Flow Cytometer (Becton Dickinson) 2.3. Antibodies and Materials for Monitoring Serum Kinetics and Half-Lives
1. Purified human IgG antibody (GammaGard, Baxter Laboratories)
2.4. Blood Collection
1. Capillary tubes, 75 l volume, heparinized (Globe Scientific, 51608)
R 2. Humanized mAb Herceptin (trastuzumab; Genentech, Inc.)
2. Capillary tubes, 25 l volume (Drummond Scientific, 251000-0250) 3. Heparin, sodium salt (Sigma, H0777) at 10,000 U/ml in 0.85% NaCl solution 4. Microcentrifuge tubes, 1.5 ml (USA Scientific #1615-5500) 2.5. ELISA Reagents
1. 96-well high binding ELISA plates (Greiner Bio-One, 655061) 2. Mouse anti-human IgG-Fc (Southern Biotech, Birmingham, Alabama; 9040-01)
98
Roopenian, Christianson, and Sproule
3. Dulbecco’s phosphate-buffered saline (DPBS), pH 7.2 (Hyclone; SH30013) 4. Manifold for plate washing (Bel-Art, Pequannock, NJ; Vaccu-Pette/96) 5. Bovine serum albumin (BSA, Fitzgerald Industries International, Concord, MA; 30-AB75) 6. Mouse anti-human kappa-AP (Southern Biotech; 9220-04) 7. P-nitrophenyl phosphate (Sigma; N2765) 8. Substrate buffer: 20 mM sodium bicarbonate, 24 mM sodium carbonate, 7 mM magnesium chloride hexahydrate, pH 8.6
3. Methods 3.1. The hFCRN Transgenic Model
To overcome the limitations in both monkey and conventional rodent models, we generated human (h) FcRn transgenic mice with the idea that they would prove to be an effective surrogate for preclinical evaluation of therapeutic mAbs destined for human use (16–18). The key elements of this model are (i) the lack of mouse (m) FcRn accomplished by gene inactivation, (ii) the transgenic expression of hFCRN, and (iii) a homogenous genetic background C57BL/6 J (B6) that reduces biological noise caused by genetic variation in mixed backgrounds. To produce these mice, we first inactivated the FcRn gene by conventional gene targeting techniques (17). The resulting 129X1-derived null (−) allele of FcRn (formally designated Fcgrt tm1dcr ) was backcrossed for 11 generations onto C57BL6/J (abbreviated B6) resulting in the congenic strain B6.129X1Fcgrt/Dcr (B6.mFcRn−/−). We also produced a number of hFCRN transgenic (Tg) lines of mice on the B6 background, including B6.hFCRN Tg276. This strain carries a hFCRN cDNA transgene from a B6-derived cDNA construct driven by a tissue ubiquitous CMV promotor/chicken -actin enhancer. Another strain, B6.hFCRN Tg32, carries a genomic hFCRN cosmid fragment, including ∼11 kb encoding the transcription product along with ∼10 kb of 5 and 3 non-coding human sequence (16, 17). Both of these transgenic founder lines were produced by microinjection of the respective hFCRN plasmids into B6 zygotes. To produce mice lacking mouse FcRn but expressing the human FcRn transgene, we crossed hFcRn Tgs to mFcRn−/− mice and established lines of B6.mFcRn−/− hFCRN Tg mice (see Note 1).
Human FcRn Transgenic Mice
99
3.2. Tracking the hFCRN Transgenes and Monitoring Their Expression
FcRn−/− mice can be genotyped by PCR using primers designed to distinguish the targeted from the wild-type FcRn alleles. The primers for FcRn-F and FcRn-R yield a 248-bp wild-type product, while the primers for T-FcRn-F and T-FcRn-R yield a 378bp targeted allele product. Tracking the hFCRN transgenes can be carried out by PCR-based genotyping. hFcRn-F and hFcRn-R primers yield an amplified hFCRN product of ∼300 bp for the cDNA Tg276 and 740 bp for the genomic Tg32. Immunofluorescence monitoring of the protein expression of hFCRN can be carried out using mAbs ADM31 or ADM32, both of which are highly specific for hFCRN (see Note 2). Expression of hFCRN on leukocytes from peripheral blood, spleen, or lymph nodes can be evaluated using conventional flow cytometric procedures. To assess cell surface expression levels, single cell suspensions at approximately 106 /50 l PBS with 1% BSA and 0.05% NaN3 can be stained with 200 ng of ADM31-AlexaFluor 647 (labeled with Alexa Fluor 647 carboxylic acid, succinimidyl ester) and data acquired on a FACSCalibur Flow Cytometer. Expression in solid tissues can be evaluated using conventional immunofluorescence imaging on frozen sections.
3.3. Use of hFCRN Transgenic Models to Evaluate Antibody Kinetics In Vivo
Measurement of the serum concentrations of administered antibodies is a general tool to evaluate their persistence in circulation. This is usually performed by introducing a sufficient amount of the test antibody either by the intravenous or by the intraperitoneal routes (see Note 3), in a quantity that can be easily detected and quantified in serum samples, even after a two log reduction in concentration. The antibody “tracer” can be labeled with radioisotope which permits direct quantification in serum samples. To minimize radioactive isotope use, we use an antibody tracer that is unmodified or labeled with biotin or other derivation chemistries, and then determine its serum concentrations by ELISA techniques. We commonly inject 100 g of the test antibody in a 200 l volume of phosphate-buffered saline into each mouse intraperitoneally (i.p.) (see Note 4). This amount can vary depending on the goals of the experiment and the sensitivity of the detection method. A minimum of five inbred mice, sex-matched and age-matched, at 8–16 weeks of age are recommended for each antibody to be tested. A typical kinetic profile is divided into alpha and beta phases. The alpha phase occurs rapidly (in the first 24 h) and is considered to be the period in which the administered antibody reaches equilibrium. The beta phase then continues for several days in a direct log–linear relationship of serum concentration to time. It is this beta phase that is considered to be the most reliable indicator of antibody half-life in circulation. To establish beta phase kinetics, we take blood samples (25–75 l depending on the frequency of sampling) at 1–4 days following administration, followed by
100
Roopenian, Christianson, and Sproule
several additional samples over a period of days to weeks, depending on the persistence of the test antibody (see Note 3). A least five blood samples per mouse is recommended. To illustrate the use of the hFcRn-humanized models to monitor the serum clearance of human IgG, we show an experiment in which we administered 100 g of purified human IgG on day 0 to B6.mFcRn−/− hFCRN Tg276 and to B6.mFcRn−/− hFCRN Tg32 mice (carrying one or two copies of the respective transgenes). We used five mice in each group and collected 75 l of blood on days 4, 11, 18, 25, and 32 (Fig. 6.3A). For comparison, in Fig. 6.3B, we show results of an experiment in which 100 g human IgG (hIgG) was administered to five mice lacking FcRn (B6.mFcRn−/− mice). On day 0, 25 l blood was collected, and then on days 3, 5, 6, 7, 9, and 11. The beta phase clearance results are plotted as concentration of tracer remaining as a percent of that observed at the first timepoint. From these data it is apparent that the serum persistence and the computed serum half-life of hIgG administered to the mice is dependent on the presence of hFCRN and is influenced by the number of copies of the transgenes; i.e., it is shorter in hFCRN hemizygous mice and extended in homozygous mice. It is also apparent that hIgG has a reduced half-life in B6.mFcRn−/− hFCRN Tg276 mice as compared with B6.mFcRn−/− hFCRN Tg32 mice (see Note 5). In general, we find that B6.mFcRn−/− A
B
C
Fig. 6.3. Beta phase clearance of human IgG in hFcRn transgenic mice. Hundred microgram of purified human IgG (A and B) or Herceptin (C) was administered through the intraperitoneal route. Blood samples were taken from the retroorbital sinus at the time points indicated by the symbols. The concentrations of hIgG and Herceptin were determined by ELISA techniques from plasma prepared from these samples, as described in Section 3. (A) Clearance of human IgG in transgenic mice: (open triangles) B6.mFcRn−/− hFCRN Tg276 heterozygotes; (closed triangles) B6.mFcRn−/− hFCRN Tg276 homozygotes; (open circles) B6.mFcRn−/− hFCRN Tg32 heterozygotes; (closed circles) B6.mFcRn−/− hFCRN Tg32 homozygotes; B6.mFcRn−/− hFCRN Tg276 homozygotes. (B) Clearance of hIgG in FcRn−/− mice. (C) Clearance of Herceptin in B6.mFcRn−/− hFCRN Tg32 homozygous mice. Data are presented as the average percent of IgG or Herceptin remaining normalized to the first bleed.
Human FcRn Transgenic Mice
101
hFCRN Tg276 hemizygous mice are best suited to detect subtle differences in mAb persistence and that B6.mFcRn−/− hFCRN Tg32 homozygous mice are best suited to achieve prolonged serum persistence. Thus one can select the model that is best suited for a specific application. To illustrate the use of the transgenic mice for determination of the pharmacokinetics of therapeutic mAb, we administered the humanized therapeutic Herceptin mAb into eight B6.mFcRn−/− hFCRN Tg32 homozygous mice. As shown in Fig. 6.3C, two copies of transgene hFCRN Tg32 resulted in prolonged serum persistence of Herceptin, similar to that observed for purified human IgG. Further illustration of the use of the hFCRN transgenic mice to analyze serum persistence of therapeutic mAbs can be found in Petkova et al. (16). These results show that the hFCRN transgenic model is well suited to evaluate the pharmacokinetics of human therapeutic antibodies. 3.4. Determination of Serum Concentrations and Half-Life of hIgG and Herceptin
1. hIgG-specific ELISA of plasma samples are performed using 96-well high-binding ELISA plates coated with mouse anti-human IgG-Fc diluted to 0.5 g/100 l/well in Dulbecco’s phosphate-buffered saline pH 7.2, overnight at 4ºC. 2. Plates are hand washed using a manifold with 96 nozzles connected to a 30 ml syringe, or multichannel pipette, two times with 300 l per well of DPBS with 0.05% Tween 20 and 0.05% sodium azide (ELISA Wash). 3. The ELISA Wash is removed by flicking plates into a sink and blotting the plates on paper towels. 4. Plates are blocked with 300 l/well of ELISA Block (ELISA Wash with 1% bovine serum albumin [BSA]) for a minimum of 1 h at room temperature. 5. Plasma or serum samples are diluted 1/200 with ELISA Block. 6. HIgG are diluted in a two-fold series from 1000 ng/ml down to 1 ng/ml to act as standards. 7. ELISA Block is removed by flicking and blotting as before (see Step 3). 8. 100 l of diluted standards and plasma as well as diluent only (to act as blanks) are transferred into the coated and blocked ELISA plates. 9. The plates are incubated at 37o C for 1 h. 10. The plates are washed three times with 300 l/well with ELISA Wash (see Step 4).
102
Roopenian, Christianson, and Sproule
11. 100 l/well of mouse anti-human kappa-AP diluted 1/1000 in diluent is then added and incubated at 37o C for 1 h. 12. Plates are washed as in Step 4 three times with 300 l/well. 13. 100 l/well of 1 mg/ml p-nitrophenyl phosphate in substrate buffer is added. 14. The plates are read A405 nm at 30 min, or when the maximum optical density exceeds 1. 15. Data are plotted as the percent tracer remaining compared to the first tracer concentration determination. Half-life is calculated using the following formula: t1/2 = (log 0.5/(log Ae /A0 )) × t where t 1/2 is the half-life of the tracer, A e is the amount of tracer remaining, Ao is the original amount of tracer at day 0, and t is the elapsed time. 3.5. Analysis of Albumin Pharmacokinetics
While this chapter focuses on issues related to FcRn and its control over the pharmacokinetics of therapeutic mAbs, there is also strong evidence that FcRn operates in a similar manner to extend the serum persistence of albumin. Indeed, serum albumin has a prolonged half-life because it binds FcRn in an acid-dependent manner in a distinct docking site from IgG and is recycled similarly to IgG (reviewed in (19)). Albumin-conjugated therapeutic proteins are being developed that take advantage of this route to prolong the pharmacokinetics (20). The human FCRN transgenic models described above have also been used to address the serum persistence of human albumin (18) and should prove to be an effective general model for the preclinical evaluation of albuminbased therapeutics.
3.6. Analysis of Therapeutic mAbs That Are Immunogenic in Mice
The beta phase clearance kinetics of human IgG and therapeutic antibodies is typically monophasic and abides to a log–linear relationship of serum concentration with time after administration. However, in rare cases we have found that certain administered mAbs are immunogenic to the mouse and elicit mouseanti-human antibodies. In this case, there is a bi-phasic kinetic, typically appearing at days 5–6 after antibody administration and resulting in a precipitous antibody loss. This complication can be overcome by the use of immunodeficient hFcRn transgenic mice, described in Section 3.7.
3.7. Use of the Model for Efficacy Testing of Therapeutic mAbs
Use of the humanized FCRN model is proving to be an effective preclinical surrogate to evaluate the serum persistence of therapeutic mAbs. With modification, this approach may also assist in the testing of the efficacy. For example, there is a need to
Human FcRn Transgenic Mice
103
evaluate the efficacy of therapeutic mAbs directed against neoplastic tissues. Typically, the rodent models used are conventional SCID, nude, and Rag1-deficient mice. However, as we described in Section 3.1., the presence of mouse FcRn makes the mouse unsuitable for reliable modeling of the pharmacokinetics of human therapeutic mAbs. We have intercrossed the SCID and Rag1-null mutations into B6.mFcRn−/− hFCRN Tg276 and B6.mFcRn−/− hFCRN Tg32 mice, making it possible to transplant normal or neoplastic tissue into these mice and then to evaluate both the pharmacokinetics and the efficacy of therapeutic mAb treatment. Studies are underway to evaluate these models.
4. Notes 1. The standard B6.mFcRn−/− stock and FCRN-humanized B6.mFcRn−/− stocks show no evidence of failure to thrive when maintained in conventional, specific pathogen-free conditions. 2. Specific antibodies against the human FCRN are used in regular intervals for quality control and for ascertaining the level of hFCRN expression in the transgenic strains. We recommend checking hFCRN protein expression levels in the transgenic colony at least once a year confirm expression. 3. All animal experimental procedures (including handling, housing, husbandry, blood collection, and drug treatment) must be conducted in accordance with national and institutional guidelines for the care and use of laboratory animals. 4. We have not observed a difference in  phase clearance kinetics when comparing tail vein i.v. with i.p. injections. 5. As a general rule, the hFCRN copy number in our strains correlates with the hFCRN expression levels, with B6.mFcRn−/− hFCRN Tg32 homozygous mice having the highest hFCRN protein expression. References 1. Burnet, F. M., Freeman, M., Jackson, A. V. and Lush, D. (1941) The Production of Antibodies. Macmillan and Company Limited, Melbourne. 2. Simister, N. E. and Mostov, K. E. (1989) An Fc receptor structurally related to MHC class I antigens. Nature 337, 184–187. 3. Ahouse, J. J., Hagerman, C. L., Mittal, P., Gilbert, D. J., Copeland, N. G., Jenkins,
N. A., et al. (1993) Mouse MHC class Ilike Fc receptor encoded outside the MHC. J Immunol 151, 6076–6088. 4. Raghavan, M., Gastinel, L. N. and Bjorkman, P. J. (1993) The class I major histocompatibility complex related Fc receptor shows pHdependent stability differences correlating with immunoglobulin binding and release. Biochemistry 32, 8654–8660.
104
Roopenian, Christianson, and Sproule
5. Burmeister, W. P., Gastinel, L. N., Simister, N. E., Blum, M. L. and Bjorkman, P. J. (1994) Crystal structure at 2.2 A resolution of the MHC-related neonatal Fc receptor. Nature 372, 336–343. 6. Burmeister, W. P., Huber, A. H. and Bjorkman, P. J. (1994) Crystal structure of the complex of rat neonatal Fc receptor with Fc. Nature 372, 379–383. 7. Roopenian, D. C. and Akilesh, S. (2007) FcRn: the neonatal Fc receptor comes of age. Nat Rev Immunol 7, 715–725. 8. Ghetie, V. and Ward, E. S. (2002) Transcytosis and catabolism of antibody. Immunol Res 25, 97–113. 9. Ghetie, V. and Ward, E. S. (1997) FcRn: the MHC class I-related receptor that is more than an IgG transporter. Immunol Today 18, 592–598. 10. http://www.bioportfolio.com/cgi-bin/ acatalog/Therapeutic Monoclonal Antibodies Report 2008-2023.html. 11. Deckert, P. M. (2009) Current constructs and targets in clinical development for antibody-based cancer therapy. Curr Drug Targets 10, 158–175. 12. Jarvis, B. and Faulds, D. (1999) Etanercept: a review of its use in rheumatoid arthritis. Drugs 57, 945–966. 13. Ober, R. J., Radu, C. G., Ghetie, V. and Ward, E. S. (2001) Differences in promiscuity for antibody-FcRn interactions across species: implications for therapeutic antibodies. Int Immunol 13, 1551–1559. 14. Presta, L. G. (2008) Molecular engineering and design of therapeutic antibodies. Curr Opin Immunol 20, 460–470.
15. Liu, X. Y., Pop, L. M. and Vitetta, E. S. (2008) Engineering therapeutic monoclonal antibodies. Immunol Rev 222, 9–27. 16. Petkova, S. B., Akilesh, S., Sproule, T. J., Christianson, G. J., Al Khabbaz, H., Brown, A. C., et al. (2006) Enhanced half-life of genetically engineered human IgG1 antibodies in a humanized FcRn mouse model: potential application in humorally mediated autoimmune disease. Int Immunol 18, 1759– 1769. 17. Roopenian, D. C., Christianson, G. J., Sproule, T. J., Brown, A. C., Akilesh, S., Jung, N., et al. (2003) The MHC class I-like IgG receptor controls perinatal IgG transport, IgG homeostasis, and fate of IgG-Fccoupled drugs. J Immunol 170, 3528–3533. 18. Chaudhury, C., Mehnaz, S., Robinson, J. M., Hayton, W. L., Pearl, D. K., Roopenian, D. C., et al. (2003) The major histocompatibility complex-related Fc receptor for IgG (FcRn) binds albumin and prolongs its lifespan. J Exp Med 197, 315–322. 19. Anderson, C. L., Chaudhury, C., Kim, J., Bronson, C. L., Wani, M. A. and Mohanty, S. (2006) Perspective–FcRn transports albumin: relevance to immunology and medicine. Trends Immunol 27, 343–348. 20. Osborn, B. L., Olsen, H. S., Nardelli, B., Murray, J. H., Zhou, J. X., Garcia, A., et al. (2002) Pharmacokinetic and pharmacodynamic studies of a human serum albumininterferon-alpha fusion protein in cynomolgus monkeys. J Pharmacol Exp Ther 303, 540–548.
Chapter 7 Development of Novel Major Histocompatibility Complex Class I and Class II-Deficient NOD-SCID IL2R Gamma Chain Knockout Mice for Modeling Human Xenogeneic Graft-Versus-Host Disease Steve Pino, Michael A. Brehm, Laurence Covassin-Barberis, Marie King, Bruce Gott, Thomas H. Chase, Jennifer Wagner, Lisa Burzenski, Oded Foreman, Dale L. Greiner, and Leonard D. Shultz Abstract Immunodeficient mice have been used as recipients of human peripheral blood mononuclear cells (PBMC) for in vivo analyses of human xeno-graft-versus-host disease (GVHD). This xeno-GVHD model system in many ways mimics the human disease. The model system is established by intravenous or intraperitoneal injection of human PBMC or spleen cells into unconditioned or irradiated immunodeficient recipient mice. Recently, the development of several stocks of immunodeficient Prkdcscid (scid) and recombination activating 1 or 2 gene (Rag1 or Rag2) knockout mice bearing a targeted mutation in the gene encoding the IL2 receptor gamma chain (IL2rγ ) have been reported. The addition of the mutated IL2rγ gene onto an immunodeficient mouse stock facilitates heightened engraftment with human PBMC. Stocks of mice with mutations in the IL2rγ gene have been studied in several laboratories on NOD-scid, NOD-Rag1null , BALB/c-Rag1null , BALB/c-Rag2null , and Stock-H2d -Rag2null strain backgrounds. Parameters to induce human xeno-GVHD in H2d -Rag2null IL2rγ null mice have been published, but variability in the frequency of disease and kinetics of GVHD were observed. The availability of the NOD-scid IL2rγ null stock that engrafts more readily with human PBMC than does the Stock-H2d Rag2null IL2rγ null stock should lead to a more reproducible humanized mouse model of GVHD and for the use in drug evaluation and validation. Furthermore, GVHD in human PBMC-engrafted scid mice has been postulated to result predominately from a human anti-mouse major histocompatibility complex (MHC) class II reactivity. Our recent development of NOD-scid IL2rγ null β2mnull and NOD-scid IL2rγ null Abnull stocks of mice now make it possible to investigate directly the role of host MHC class I and class II in the pathogenesis of GVHD in humanized mice using NOD-scid IL2rγ null stocks that engraft at high levels with human PBMC and are deficient in murine MHC class I, class II, or both classes of MHC molecules. Key words: scid, Rag, humanized mice, GVHD, immunodeficient mice, MHC class I, MHC class II.
G. Proetzel, M.V. Wiles (eds.), Mouse Models for Drug Discovery, Methods in Molecular Biology 602, DOI 10.1007/978-1-60761-058-8 7, © Humana Press, a part of Springer Science+Business Media, LLC 2010
105
106
Pino et al.
1. Introduction 1.1. Hematopoietic Stem Cell Transplantation and Graft-Versus-Host Disease
Hematopoietic stem cell (HSC) transplantation has the potential to be curative for a variety of human diseases. These include diseases of the blood, metabolic disorders, autoimmune diseases, and solid tumors (1–5). Of particular importance is the use of HSC transplantation for the treatment of hematological malignancies (6–8). In this setting, the beneficial graft-versus-leukemia (GVL) effects versus the detrimental GVHD effects are commonly intertwined (9). The beneficial effects of GVL appear to correlate directly with the severity of chronic GVHD and these two phenomena are difficult to study as independent phenomena in patients (10–12). A small animal model of human immunity that accurately mimics the clinical presentation of human GVHD would provide a model for the studies of GVL as well as GVHD and for testing the efficacy of human-specific therapeutics prior to their use in the clinic. Such a model system would also provide a tool for investigating the underlying mechanisms involved in the pathogenesis of human GVHD and to discover ways to dissociate GVHD from GVL effects.
1.2. Immunodeficient Mice Engrafted with Human Lymphohematopoietic Cells
The first description of engrafting immunodeficient mice with human PBMC was reported in 1988 (13). In those studies, CB17 mice harboring the Prkdcscid (scid) mutation were injected intraperitoneally with 10–100 × 106 human PBMC, and low levels of human PBMC engraftment were observed. However, instead of developing GVHD, these mice developed EBVassociated lymphomas (14). Most of the human T cells that did engraft became anergic (15, 16) and few engrafted mice developed GVHD (17). Over the ensuing two decades, genetically altered immunodeficient mice were found to support heightened engraftment with human lymphohematopoietic cells, with a major advance being the generation of NOD-scid mice (18). Additional hereditable modifications included the generation of NOD-scid mice genetically deficient in beta-2 microglobulin (β2mtm1Unc , abbreviated as β2mnull ) (19). In the absence of 2m, natural killer (NK) cell numbers and activity are severely depressed due to the lack of MHC class I expression (19). This removed a major impediment to human lymphohematopoietic cell engraftment. Although NOD-scid β2mnull mice support higher levels of human PBMC engraftment than do CB17-scid or NOD-scid mice (19), development of symptoms of GVHD remained variable (20). In this and other model systems developed during this period, the low frequency and poor engraftment observed, and the high numbers of human PBMC required for engraftment and
Humanized Mouse Models for GVHD
107
disease expression, precluded widespread use of these mice for studies of xeno-GVHD. 1.3. Immunodeficient Stocks of IL2r␥null Mice
A major breakthrough in modeling human immunity and human xeno-GVHD reactivity in immunodeficient mice followed the development by several research groups of immunodeficient mice bearing a targeted mutation in the gene encoding the IL2 receptor gamma chain (IL2rg), abbreviated as (IL2rγ ) (21). The IL2 receptor gamma chain is important for high affinity signaling through the IL2, IL4, IL7, IL9, IL15, and IL21 cytokine receptors. Without these receptor signals, both adaptive and innate immunities are severely impaired (22). We and others are now building on the foundation of immunodeficient IL2rγ null mice that are available to knockout additional genes and to introduce new genes that will facilitate the engraftment of functional human immune systems. On the BALB/c strain background, this is being pursued using BALB/c embryonic stem cells to genetically manipulate the genome. For example, knocking in human genes into the BALB/c ES cells can rapidly introduce human genes of particular importance for the enhancement of human hematolymphopoietic cell engraftment without requiring two or more years of backcrossing the mutated genes made in B6/129 ES cells onto the BALB/c strain background. Development of NOD ES cells, currently underway in a number of laboratories, when available will provide an important tool for rapidly manipulating the NOD genome, facilitating the development of new stocks of NOD-scid IL2rγ null mice for human cell engraftment. Using a Stock-H2d mouse strain bearing both the Rag2null and the IL2rγ null targeted mutations, it was shown that there was improved engraftment of human PBMC as compared with NODscid mice. Unfortunately, even extensive preconditioning of the Stock-H2d -Rag2null IL2rγ null recipient with total body irradiation and macrophage depletion using clodronate-containing liposomes did not lead to reproducible development of GVHD symptoms in all of the mice (23).
1.4. Engraftment of NOD-scid IL2r␥null Mice with Human PBMC
The independent development of NOD-scid mice bearing a targeted mutation in the IL2rγ gene was reported by two separate groups. The laboratory of Ito and colleagues developed a stock bearing a truncated IL2rγ (Il2rgtm1Sug , abbreviated as IL2rγ trunc ) gene (24). We have developed a stock bearing a complete null IL2rγ mutation (Il2rgtm1Wjll , abbreviated as IL2rγ null ) (25). We have recently described the parameters of human PBMC engraftment into unconditioned NOD-scid IL2rγ null mice (26). We observed that injection of as few as 1–5 × 106 human PBMC led to engraftment and that all mice were successfully engrafted when injected with 10 × 106 or more human PBMC. At about
108
Pino et al.
30–40 days after PBMC engraftment, we consistently observed that all mice were losing weight and developing symptoms of GVHD (26). Our observations here are consistent with the recent report of the gradual development of phenotypic changes resembling GVHD over the course of 4–8 weeks observed following injection of 10 × 106 human PBMC into unconditioned NODscid IL2rγ null mice (27). These data suggest that heightened engraftment of functional human PBMC in NOD-scid IL2rγ null mice, particularly that of CD8+ T cells, leads to the development of xenoreactivity, aggression against host tissues, and development of GVHD. More recently, we have shown that the NOD-scid IL2rγ null mouse model supports higher levels of human hematolymphoid engraftment than does the CBy.Cg-Rag1tm1Mom Il2rgtm1wjl /Sz (abbreviated as BALB/c-Rag1null IL2rγ null ) stock of mice (28). We have subsequently shown that NOD.129S7(B6)Rag1tm1Mom Il2rgtm1Wjll (abbreviated as NOD-Rag1null IL2rγ null ) mice also engraft with human PBMC at high levels that are comparable to those observed in NOD-scid IL2rγ null mice (29). These latter data document that the difference in human PBMC engraftment in NOD-scid IL2rγ null and BALB/c.129P2Rag2tm1Mnz (abbreviated as BALB/c-Rag2null IL2rγ null ) mice is not due to the effects of the scid versus Rag1null or Rag2null genes but rather to other polymorphic background modifiers. One of these important background modifiers has been reported to be the Sirpa (also termed Shps-1 and Ptpns1) gene. The SIRP-␣ protein is a potent regulator of interactions between the hematopoietic cells and the bone marrow microenvironment. The NOD SIRP-␣ protein is more similar to the human SIRP␣ protein than that of the C57BL/6 SIRP-␣ protein, shows enhanced binding to the human CD47 ligand, and its expression on mouse macrophages leads to better support of human hematopoiesis (30).
2. Role of Mouse MHC Class I and Class II in GVHD Following Engraftment of NOD-scid IL2r␥null Mice with Human PBMC
It has been suggested that mismatches at the human HLA class I loci are the most important predictor of human GVHD following hematopoietic stem cell transplantation (9). In humanized mice, there is contradictory evidence on the role of MHC class I and class II in the pathogenesis of GVHD. Human CD4 T-cell clones derived from CB17-scid mice engrafted with PBMC exhibited reactivity against host MHC class II (15). However, the predominant human cell population engrafting in CB17-scid mice was the CD8 T-cell population (31), which would more
Humanized Mouse Models for GVHD
109
likely be driven to proliferate by mouse MHC class I rather than MHC class II. To re-examine the role of mouse MHC class I and II in the engraftment and development of xeno-GVHD following human PBMC engraftment, we developed two new stocks of NOD-scid IL2rγ null mice. 2.1. Development of NOD-scid IL2r␥null Mice That Are Also Deficient in Mouse MHC Class I or MHC Class II
The development of the NOD.Cg-β2mtm1Unc Prkdcscid Il2rgtm1Wjl/ Sz (abbreviated as NOD-scid IL2rγ null β2mnull ) genetic stock was initiated by crossing NOD-scid/scid IL2rγ null females with NOD-scid/scid β2mnull males. The NOD-scid IL2rγ null / Y +/β2mnull F1 males were backcrossed to NOD-scid/scid IL2rγ null females. The NOD-scid/scid IL2rγ null +/β2mnull female offspring were mated with their NOD-scid/scid IL2rγ null /γ +/β2mnull sibs to produce homozygous mice at each of the mutated genes. The homozygous NOD-scid IL2rγ null B2mnull offspring were then intercrossed to establish and maintain this strain. The development of the NOD.Cg-Prkdcscid H2-Ab1tm1Gru Il2rgtm1Wjl /Sz (abbreviated as NOD-scid IL2rγ null Abnull ) genetic stock was initiated by first crossing NOD-scid/scid IL2rγ null females with NOD-scid/scid Abnull males. The NODscid/scid IL2rγ null /Y +/Abnull F1 males were backcrossed to NOD-scid/scid IL2rγ null females. The NOD-scid/scid IL2rγ null +/Abnull female offspring were mated with their NODscid/scid IL2rγ null /γ +/Abnull sibs to produce homozygous mice at each of the mutated genes. The homozygous NOD-scid IL2rγ null Abnull offspring were intercrossed to establish and maintain this strain. Both NOD and 129 strains of mice have a genetic deficiency in I-E␣ expression. The crossing of the Abnull allele, made using 129 strain ES cells, onto the NOD strain background therefore results in a deficiency of both I-A and I-E expression and complete lack of MHC class II expression.
2.2. Phenotype of Cells Developing in NOD-scid IL2r␥null Mice Deficient in Mouse MHC Class I or MHC Class II
We have previously shown that NOD-scid IL2rγ null mice lacking NK cells, mature T lymphocytes and mature B lymphocytes are excellent recipients of human PBMC (26). To determine if deletion of either the mouse MHC class I or class II altered the phenotype of the murine cell subsets in NOD-scid IL2rγ null mice, we performed flow cytometry analyses. As shown in Tables 7.1 and 7.2 and Fig. 7.1, Panels A and B, there were no MHC class I positive cells in the bone marrow and spleens of NODscid IL2rγ null β2mnull mice. Similarly, there were no MHC class II positive cells in the bone marrow and spleens of NOD-scid IL2rγ null Abnull mice. We also confirmed the absence of mature mouse CD4 and CD8 T cells, mature cell-surface Ig-positive B cells, and LGL+ /CD122+ NK cells in the spleens and bone marrow of NOD-scid IL2rγ null β2mnull and NOD-scid IL2rγ null
110
Pino et al.
Table 7.1 Flow cytometry analyses of spleen cells from NOD-scid IL2r␥null β 2Mnull and NODscid IL2rγ null Abnull mice showing percentages of leukocyte populationsa NOD-scid IL2rγ null (N=5)
NOD-scid IL2r␥null β 2Mnull (N=6)
NOD-scid IL2rγ null Abnull (N=5)
Spleen surface marker
NOD +/+ (N=1)
CD3+ /CD4+
26.1
0.0±0.0
0.0±0.0
0.0±0.0
10.6
0.0±0.0
0.0±0.0
0.0±0.0
45.0
0.2±0.1
0.0±0.0
0.0±0.0
+
CD3 /CD8
+
B220+ /IgK, light chain+ +
B220 /IgK, light chain
−
1.3
3.1±0.7
0.9±0.2
1.6±0.5
GR-1+ /Mac-1+
6.3
17.3±1.0
10.9±2.0∗
21.1±0.9
GR-1- /Mac-1+
5.1
9.6±0.9
5.9±0.4∗
11.0±1.2
2.8
45.4±2.6
25±1.8
∗
44.3±1.7
4.2
1.3±0.1
1.2±0.1
0.7
0.1±0.0
0.3±0.1
Ter 119
+
DX5+ /CD122+ +
+
LGL /CD122 d+
94.2
65.3±2.2
b+
83.6
28.8±3.5
50.9
3.6±0.4
H-2 K H-2D g7+
I-A
1.0±0.1
1.5±0.1 0.1±0.0 ∗∗
N/A
0.2±0.0∗∗ 36.8±1.6 0.0±0.0∗∗
6.4±0.0
are expressed as mean ± SE. NOD-scid IL2rγ null β2Mnull or NOD-scid IL2rγ null Abnull mice versus NOD-scid IL2rγ null . ∗∗ p<0.001 NOD-scid IL2rγ null β2Mnull or NOD-scid IL2rγ null Abnull mice versus NOD-scid IL2rγ null . a Data
∗ p<0.05
Table 7.2 Flow cytometry analyses of bone marrow cells from NOD-scid IL2rγ null β 2Mnull and NOD-scid IL2rγ null Abnull mice showing percentages of leukocyte populationsa Bone marrow surface marker
NOD +/+ (N=1)
CD3+ /CD4+ +
CD3 /CD8
+
B220+ /IgK, light chain+ +
B220 /IgK, light chain
-
NOD-scid IL2rγ null (N=5)
NOD-scid IL2r␥null β 2Mnull (N=6)
NOD-scid IL2rγ null Abnull (N=5)
0.5
0.0±0.0
0.0±0.0
0.0±0.0
0.5
0.0±0.0
0.0±0.0
0.0±0.0
5.0
0.0±0.0
0.0±0.0
0.0±0.0
14.9
1.6±0.1
0.8±0.0
1.3±0.2
+
48.8
57.9±0.3
67.7±1.4
52.1±2.8
GR-1- /Mac-1+
6.9
9.6±0.4
5.9±0.4
+
GR-1 /Mac-1 Ter 119
+
10.4
14.7±0.5
6.5±0.7
DX5 /CD122
1.8
1.9±0.1
1.3±0.0
LGL+ /CD122+
0.41
0.2±0.0
0.1±0.0
+
+
d+
92.8
85.8±2.0
b+
68
72.2±0.6
H-2 K H-2D g7+
I-A
12.6
2.9±0.2
0.3±0.0 N/A 4.3±0.6
9.2±0.3 ∗
18.3±1.0 1.4±0.1 0.2±0.0
∗∗
0.3±0.1∗∗ 69.2±1.2 0.3±0.0∗∗
are expressed as mean ± SE. p<0.05 NOD-scid IL2rγ null β2Mnull or NOD-scid IL2rγ null Abnull mice versus NOD-scid IL2rγ null . ∗∗ p<0.001 NOD-scid IL2rγ null β2Mnull or NOD-scid IL2rγ null Abnull mice versus NOD-scid IL2rγ null . a Data ∗
Humanized Mouse Models for GVHD
111
Fig. 7.1. Flow cytometric analysis of MHC class I and II expression. Panel A. Analysis of MHC class I (H-2 Kd ) and MHC class II (I-Ag7 ) in spleen and bone marrow cells. While nearly all NOD +/+ splenocytes and bone marrow cells express MHC class I (H-2 Kd ), <1% of NOD-scid IL2rγ null β 2mnull bone marrow cells or splenocytes are MHC class I positive. Approximately 50% of NOD +/+ splenocytes and 13% of NOD +/+ bone marrow cells express MHC class II. NOD-scid IL2rγ null as well as NOD-scid IL2rγ null β 2mnull mice have reduced percentages of MHC class II positive cells, consistent with a lack of mature B cells. Panel B. Analysis of MHC class I (H-2 Kd , NOD-derived and H-2Db , C57BL/6-derived) and MHC class II (I-Ag7 , NOD-derived and I-Ab , C57BL/6-derived) expression in splenic cells. Different sets of antibodies were used because genetic backcrossing of the MHC class II targeted mutation in 129 ES cells crossed onto the C57BL/6 J strain and subsequently crossed onto the NOD strain results in the crossing of the entire MHC region onto NOD. Thus, NOD MHC class II (I-A)null mice express the 129/C57BL/6 class I haplotype (H-2b ) and fail to express either the NOD MHC class I (H-2 Kd ) or the NOD MHC class II (I-Ag7 ) antigens. NOD +/+ mice express H-2 Kd and H-2Db MHC class I alleles. They also express I-Ag7 and do not express the I-Ab. allele. NOD-scid IL2rγ null mice express the H-2 Kd and H-2Db MHC class I alleles. Expression is duller because of the absence of MHC class I-bright lymphocytes. These mice have low numbers of I-Ag7 expressing cells and lack expression of I-Ab on their cells. NOD-scid IL2rγ null Abnull mice lack MHC class II expression and exhibit dull H-2Db expression with no expression of H-2 Kd .
Abnull mice. NOD-scid IL2rγ null β2mnull mice also exhibited decreased percentages of splenic Gr-1+ /Mac1+ granulocytes, Gr-1− /Mac1+ macrophages, and Ter-119+ erythroid cells compared with NOD-scid IL2rγ null mice. There were also reduced percentages of Ter-119+ erythroid cells in the bone marrow of NOD-scid IL2rγ null β2mnull mice compared with NODscid IL2rγ null mice. 2.3. Absence of Thymic Lymphomas in NOD-scid IL2r␥null β 2mnull Mice
NOD-scid and NOD-scid β2mnull mice develop thymic lymphomas and die at an early age (19). NOD-scid mice have a median life span of 36 weeks (18), while in NOD-scid β2mnull mice the median life span is shortened to 24 weeks (19). In contrast, NOD-scid IL2rγ null mice do not develop thymic lymphomas (25). The absence of thymic lymphoma development suggests that signaling through IL2rγ receptor-dependent cytokines such as IL-2 or IL-7 is required for the thymic lymphomagenesis. Because of the increased incidence and accelerated kinetics of thymic lymphoma development in NOD-scid β2mnull mice, we investigated whether NOD-scid IL2rγ null β2mnull mice
112
Pino et al.
developed thymic lymphomas and die at an early age. We first examined the histology of the thymus from 10-week-old NODscid IL2rγ null β2mnull mice and compared it with age-matched NOD-scid β2mnull mice. As shown in Fig. 7.2A,B, the thymus of young adult NOD-scid IL2rγ null β2mnull mice is a hypoplastic, degenerate thymic rudiment with a cystic formation. In contrast, the thymus of age-matched NOD-scid β2mnull mice displays thymic lymphomas showing a highly proliferative thymus containing numerous mitotic cells. In addition, although NODscid β2mnull mice do not survive beyond 8 months of age due to thymic lymphomas, NOD-scid IL2rγ null β2mnull mice show an increased life span with no gross or histological evidence of thymic lymphomas at 8 months, with mice surviving past 14 months of age (unpublished observations). A cohort of NOD-scid IL2rγ null β2mnull retired breeders is being aged for evaluating thymic lymphoma development. Four of these mice euthanized at 12–18 months of age were found to have no histological evidence of thymic lymphoma. 2.4. Development of Hemochromatosis in NOD-scid IL2r␥null 2mnull Mice
We have previously reported that the lack of 2-microglobulin led to iron accumulation in the liver of NOD-scid β2mnull mice (19), consistent with previous reports showing that 2microglobulin is required for appropriate regulation of iron absorption (32, 33). To determine whether the addition of the IL2rγ null gene to NOD-scid β2mnull mice altered this phenotype, we stained livers from 10-week-old NOD-scid IL2rγ null β2mnull mice for the presence of iron. As shown in Fig. 7.2C, we observed the accumulation of iron deposition in the liver. In contrast, there was no detectable iron accumulation in the liver of age-matched NOD-scid IL2rγ null mice (Fig. 7.2D). These observations document that the accumulation of iron in the liver of 2-microglobulin knockout mice does not require common gamma chain-dependent receptor signaling.
2.5. Engraftment of NOD-scid IL2r␥null 2mnull or NOD-scid IL2r␥null Abnull Mice with Human PBMC
The next logical step in the development of this model system is to determine the proliferative response of human PBMC to antigen-presenting cells obtained from NOD-scid IL2rγ null mice deficient in MHC class I or class II. In preliminary studies, we have observed that the deficiency of MHC class I severely reduces the ability of mouse antigen-presenting cells to stimulate in vitro proliferation of human CD8 T cells and similarly, the deficiency of MHC class II reduces their ability to stimulate in vitro proliferation of human CD4 T cells. These data suggest that xeno-GVHD in immunodeficient mice may be dependent on both MHC class I and class II expression of the murine immunodeficient host. Furthermore, the in vitro proliferative data suggest that the development of GVHD in NOD-scid IL2rγ null mice will likely be medi-
Humanized Mouse Models for GVHD
113
Fig. 7.2. Histopathological sections of tissues from NOD-scid IL2rγ null β 2mnull and control mice. Panel A. Thymic lobe from 10-week-old NOD-scid IL2rγ null β 2mnull female. Thymus is composed of epithelial and stromal cells and is completely devoid of lymphocytes. Note the cyst lined by a single layer of occasionally ciliated epithelial cells present within the thymic parenchyma (branchial arch remnant). Panel B. Thymic lobe from 10-week-old NOD-scid β 2mnull male. A thymic lymphoma is present. Thymic architecture is replaced by dense sheets of neoplastic lymphocytes. Note large atypical lymphocytes with abundant mitotic figures (insert). Panel C. Liver from 10-week-old NOD-scid IL2rγ null β 2mnull male. Hepatocytes show intracytoplasmic deposition of iron. Panel D. Liver from 20-week-old NOD-scid IL2rγ null male. There are no detectable iron depositions (A and B, H&E; C and D, Gomori’s iron) (A and B, ×100, insert ×400; C and D, ×200).
ated by both human CD4 and CD8 T cells targeting MHC class II or class I, respectively. If so, these MHC class I or class IIdeficient recipients may permit the ability to study the role of each cell subset in the absence of the confounding contribution of the other cell subset to the development of the disease. This hypothesis is currently being tested. This would make these excellent model systems to test the efficacy of therapeutics to modulate the activity of either cell subset in the absence of the functional contribution of the other cell subset to the disease process. 2.6. Human Immune Responses in PBMC-Engrafted Mice
Previous attempts to measure immune function of human cells following PBMC engraftment of immunodeficient mice have been met with limited success (34). The lack of a vigorous primary immune response mediated by human PBMC engrafted into earlier humanized mouse models was associated with host murine NK cell activity and other innate immune function as well as an overwhelming xenogeneic GVHD response of the
114
Pino et al.
human T cells. The development of NOD-scid IL2rγ null mice provided an immunodeficient mouse lacking host NK cells, facilitating human PBMC engraftment. Although NOD-scid IL2rγ null mice support survival and improved function of engrafted human T cells, the human cells mount a severe xenogeneic GVHD following engraftment into these NK cell-deficient hosts. Our findings that removal of mouse MHC class I activity greatly reduces human in vitro T-cell proliferative responses to murine cells should also lead to reduced xenogeneic GVHD following engraftment with human PBMC, an hypothesis we are currently testing. This may now allow immunization of PBMCengrafted NOD-scid IL2rγ null β2mnull mice with experimental vaccines to provide a model system for the assessment of the ability of these vaccines to elicit a protective response against infection with human pathogens in the absence of an overwhelming xeno-GVHD.
3. Conclusions The use of immunodeficient mice for the in vivo study of human xeno-GVHD holds the promise of understanding the pathogenesis of human GVHD in vivo in a small animal model without putting patients at risk. This model will be critical for understanding the role of host MHC class I and class II in the pathogenesis of GVHD, as well as opening the possibility of setting up an allogeneic GVHD model system with the emergence of NOD-scid IL2rγ null mice deficient in MHC class I or II that also transgenically express human HLA molecules (35). Of particular importance will be the use of this model system for evaluating the efficacy of human-specific therapeutics on the pathogenesis of GVHD in vivo. This model will permit the in vivo study of human-specific drugs to test their ability to ameliorate or prevent the disease process without putting patients at risk. NOD-scid IL2rγ null mice have also been engrafted with blood obtained from patients with human acute myelogenous leukemia and chronic myelogenous leukemia as well as other hematopoietic tumors. This may provide “personalized” animal models of primary human tumors that can be used to screen drugs for their in vivo effectiveness on the tumor and tumor stem cells prior to their use in the patient. The combination of optimal immunodeficient recipients and the analyses of human immunity and xenoGVHD reactions in vivo should provide insights into the mechanisms of human immunity and the pathogenesis of GVHD and
Humanized Mouse Models for GVHD
115
other human diseases and their response to drugs in ways not possible in the clinic.
Acknowledgment We thank Jean Leif, Linda Paquin, Amy Cuthbert, Celia Hartigan, Amy Sands, Allison Ingalls, and Candy Knoll for their technical assistance. We also thank the participation of the Clinical Trials Unit at the University of Massachusetts Medical School. This work was supported by National Institutes of Health Grants AI46629, DK072473, CA34196, an institutional Diabetes Endocrinology Research Center (DERC) grant DK32520, RR-07068, the Beta Cell Biology Consortium, and grants from the Juvenile Diabetes Foundation, International, including a grant through the nPOD program. The contents of this publication are solely the responsibility of the authors and do not necessarily represent the official views of the National Institutes of Health. References 1. Alderuccio, F., Siatskas, C., Chan, J., Field, J., Murphy, K., Nasa, Z. and Toh, B. H. (2006). Haematopoietic stem cell gene therapy to treat autoimmune disease. Curr. Stem Cell Res. Ther. 1: 279–287. 2. Passweg, J. and Tyndall A. (2007). Autologous stem cell transplantation in autoimmune diseases. Semin. Hematol. 44: 278– 285. 3. Ringden, O. and Le Blanc, K. (2005). Allogeneic hematopoietic stem cell transplantation: state of the art and new perspectives. APMIS 113: 813–830. 4. Welniak, L. A., Blazar B. R. and Murphy W. J. (2007). Immunobiology of allogeneic hematopoietic stem cell transplantation. Annu. Rev. Immunol. 25: 139–170. 5. Demirer, T., Barkholt, L., Blaise, D., Pedrazzoli, P., Aglietta, M., Carella, A. M., Bay, J. O., Arpaci, F., Rosti, G., Gurman, G., Niederwieser, D. and Bregni, M. (2008). Transplantation of allogeneic hematopoietic stem cells: an emerging treatment modality for solid tumors. Nat. Clin. Pract. Oncol. 5: 256–267. 6. Bhatia, M. and Walters, M. C. (2008). Hematopoietic cell transplantation for thalassemia and sickle cell disease: past, present and future. Bone Marrow Transplant. 41: 109–117.
7. Lebensburger, J. and Persons, D. A. (2008). Progress toward safe and effective gene therapy for beta-thalassemia and sickle cell disease. Curr. Opin. Drug Discov. Devel. 11: 225–232. 8. Pinto, F. O. and Roberts, I. (2008). Cord blood stem cell transplantation for haemoglobinopathies. Br. J. Haematol. 141: 309–324. 9. Kawase, T. Matsuo, K., Kashiwase, K., Inoko, H., Saji, H., Ogawa, S., Kato, S., Sasazuki, T., Kodera, Y. and Morishima, Y. (2008). HLA mismatch combinations associated with decreased risk of relapse: implications for molecular mechanism. Blood. On line publication, December, 2008. 10. Gale, R. P., Horowitz, M. M. and Butturini, A. (1991). Autotransplants in acute leukaemia. Br. J. Haematol. 78: 135–137. 11. Porter, D. and Levine, J. E. (2006). Graftversus-host disease and graft-versus-leukemia after donor leukocyte infusion. Semin. Hematol. 43: 53–61. 12. Foss, F. M. (2006). The role of purine analogues in low-intensity regimens with allogeneic hematopoietic stem cell transplantation. Semin. Hematol. 43: S35–S43. 13. Mosier, D. E., Gulizia, R. J., Baird, S. M. and Wilson, D. B. (1988). Transfer of a functional human immune system to mice with
116
14.
15.
16.
17.
18.
19.
20.
21.
22.
Pino et al. severe combined immunodeficiency. Nature 335: 256–259. Mosier, D. E., Baird, S. M. and Kirven, M. B. (1990). EBV-associated B-cell lymphomas following transfer of human peripheral blood lymphocytes to mice with severe combined immune deficiency. Curr. Top. Microbiol. Immunol. 166: 317–323. Tary-Lehmann, M., Lehmann, P. V., Schols, D., Roncarolo, M. G. and Saxon, A. (1994). Anti-SCID mouse reactivity shapes the human CD4+ T cell repertoire in huPBL-SCID chimeras. J. Exp. Med. 180: 1817–1827. Tary-Lehmann, M. and Saxon, A. (1992). Human mature T cells that are anergic in vivo prevail in SCID mice reconstituted with human peripheral blood. J. Exp. Med. 175: 503–516. Hoffmann-Fezer, G., Gall, C., Zengerle, U., Kranz, B. and Thierfelder, S. (1993). Immunohistology and immunocytology of human T-cell chimerism and graft-versushost disease in SCID mice. Blood 81: 3440– 3448. Shultz, L. D., Schweitzer, P. A., Christianson, S. W., Gott, B., Schweitzer, I. B., McKenna, S., Mobraaten, L., Rajan, T. V., Greiner, D. L. and Leiter, E. H. (1995). Multiple defects in innate and adaptive immunologic function in NOD/LtSz-scid mice. J. Immunol. 154: 180–191. Christianson, S. W., Greiner, D. L., Hesselton, R. M., Leif, J. H., Wagar, E. J., Schweitzer, I. B., Rajan, T. V., Gott, B., Roopenian, D. C. and Shultz, L. D. (1997). Enhanced human CD4+ T cell engraftment in b2-microglobulin-deficient NODscid mice. J. Immunol. 158: 3578–3586. Nervi, B., Rettig, M. P., Ritchey, J. K., Wang, H. L., Bauer, G., Walker, J., Bonyhadi, M. L., Berenson, R. J., Prior, J. L., Piwnica-Worms, D., Nolta, J. A. and DiPersio, J. F. (2007). Factors affecting human T cell engraftment, trafficking, and associated xenogeneic graftvs-host disease in NOD/SCID beta2mnull mice. Exp. Hematol. 35: 1823–1838. Shultz, L. D., Ishikawa, F. and Greiner, D. L. (2007). Humanized mice in translational biomedical research. Nat. Rev Immunol. 7: 118–130. Cao, X., Shores, E. W., Hu-Li, J., Anver, M. R., Kelsall, B. L., Russell, S. M., Drago, J., Noguchi, M., Grinberg, A., Bloom, E. T., Paul, W. E., Katz, S. I., Love, P. E. and Leonard, W. J. (1995). Defective lymphoid development in mice lacking expression of the common cytokine receptor gamma chain. Immunity 2: 223–238.
23. van Rijn, R. S., Simonetti, E. R., Hagenbeek, A., Hogenes M. C., de Weger, R. A., Canninga-van Dijk, M. R., Weijer, K., Spits, H., Storm, G., van Bloois, L., Rijkers, G., Martens, A. C. and Ebeling, S. B. (2003). A new xenograft model for graftversus-host disease by intravenous transfer of human peripheral blood mononuclear cells in RAG2-/- gammac-/- double-mutant mice. Blood 102: 2522–2531. 24. Ito, M., Hiramatsu, H., Kobayashi, K., Suzue, K., Kawahata, M., Hioki, K., Ueyama, Y., Koyanagi, Y., Sugamura, K., Tsuji, K., Heike, T. and Nakahata, T. (2002). NOD/SCID/gamma(c)(null) mouse: an excellent recipient mouse model for engraftment of human cells. Blood 100: 3175–3182. 25. Shultz, L. D., Lyons, B. L., Burzenski, L. M., Gott, B., Chen, X., Chaleff, S., Gillies, S. D., King, M., Mangada, J., Greiner, D. L., and Handgretinger, R. (2005). Human lymphoid and myeloid cell development in NOD/LtSz-scid IL2rg null mice engrafted with mobilized human hematopoietic stem cell. J. Immunol. 174: 6477–6489. 26. King, M., Pearson, T., Shultz, L. D., Leif, J., Bottino, R., Trucco, M., Atkinson, M. A., Wasserfall, C., Herold, K. C., Woodland, R. T., Schmidt, M. R., Woda, B. A., Thompson, M. J., Rossini, A. A. and Greiner, D. L. (2008). A new Hu-PBL model for the study of human islet alloreactivity based on NOD-scid mice bearing a targeted mutation in the IL-2 receptor gamma chain gene. Clin. Immunol. 126: 303–314. 27. Golovina, T. N., Mikheeva, T., Suhoski, M. M., Aqui, N. A., Tai, V. C., Shan, X., Liu, R., Balcarcel, R. R., Fisher, N., Levine, B. L., Carroll, R. G., Warner, N., Blazar, B. R., June, C. H. and Riley, J. L. (2008). CD28 costimulation is essential for human T regulatory expansion and function. J. Immunol. 181: 2855–2868. 28. King, M., Pearson, T., Rossini, A. A., Shultz, L. D. and Greiner, D. L. (2008). Humanized mice for the study of type 1 diabetes and beta cell function. Ann. NY. Acad. Sci. 1150: 46–53. 29. Pearson, T., Shultz, L. D., Miller, D., King, M., Laning, J., Fodor, W., Cuthbert, A., Burzenski, L., Gott, B., Lyons, B., Foreman, O., Rossini, A. A. and Greiner, D. L. (2008). Non-obese diabetic-recombination activating gene-1 (NOD-Rag1 null) interleukin (IL)-2 receptor common gamma chain (IL2r gamma null) null mice: a radioresistant model for human lymphohaematopoietic engraftment. Clin. Exp. Immunol. 154: 270–284.
Humanized Mouse Models for GVHD 30. Takenaka, K., Prasolava, T. K., Wang, J. C., Mortin-Toth, S. M., Khalouei, S., Gan, O. I,. Dick, J. E. and Danska, J. S. (2007). Polymorphism in Sirpa modulates engraftment of human hematopoietic stem cells. Nat. Immunol. 8: 1313–1323. 31. Mosier, D. E. (1991). Adoptive transfer of human lymphoid cells to severely immunodeficient mice: models for normal human immune function, autoimmunity, lymphomagenesis, and AIDS. Adv. Immunol. 50: 303–325. 32. Rothenberg, B. E. and Voland, J. R. (1996). Beta2 knockout mice develop parenchymal iron overload: a putative role for class I genes of the major histocompatibility complex in iron metabolism. Proc. Natl. Acad. Sci. USA 93: 1529–1534.
117
33. Junghans, R. P. and Anderson, C. L. (1996). The protection receptor for IgG catabolism is the b2 -microglobulin-containing neonatal intestinal transport receptor. Proc. Natl. Acad. Sci. USA 93: 5512–5516. 34. Torbett, B. E., Picchio, G. and Mosier, D. E. (1991). hu-PBL-SCID mice: a model for human immune function, AIDS, and lymphomagenesis. Immunol. Rev. 124: 139– 164. 35. Banuelos, S. J. Shultz, L. D., Greiner, D. L., Burzenski, L. M., Gott, B., Lyons, B. L., Rossini, A. A. and Appel, M. C. (2004). Rejection of human islets and human HLAA2.1 transgenic mouse islets by alloreactive human lymphocytes in immunodeficient NOD-scid and NOD-Rag1 null Prf1 null mice. Clin. Immunol. 112: 273–283.
Chapter 8 Bridging Mice to Men: Using HLA Transgenic Mice to Enhance the Future Prediction and Prevention of Autoimmune Type 1 Diabetes in Humans David V. Serreze, Marijke Niens, John Kulik, and Teresa P. DiLorenzo Abstract Similar to the vast majority of cases in humans, the development of type 1 diabetes (T1D) in the NOD mouse model is due to T-cell mediated autoimmune destruction of insulin-producing pancreatic  cells. Particular major histocompatibility complex (MHC) haplotypes (designated HLA in humans and H2 in mice) provide the primary genetic risk factor for T1D development. It has long been appreciated that within the MHC, particular unusual class II genes contribute to the development of T1D in both humans and NOD mice by allowing for the development and functional activation of -cell autoreactive CD4 T cells. However, studies in NOD mice have revealed that through interactions with other background susceptibility genes, the quite common class I variants (Kd , Db ) characterizing this strain’s H2g7 MHC haplotype aberrantly acquire an ability to support the development of  cell autoreactive CD8 T-cell responses also essential to T1D development. Similarly, recent studies indicate that in the proper genetic context some quite common HLA class I variants also aberrantly contribute to T1D development in humans. This chapter will focus on how “humanized” HLA transgenic NOD mice can be created and used to identify class I-dependent  cell autoreactive CD8 T-cell populations of clinical relevance to T1D development. There is also discussion on how HLA transgenic NOD mice can be used to develop protocols that may ultimately be useful for the prevention of T1D in humans by attenuating autoreactive CD8 T-cell responses against pancreatic  cells. Key words: Type 1 diabetes (T1D), autoimmunity, “humanized” NOD mice, T cells, HLA transgenics.
1. Introduction The development of type 1 diabetes (T1D) in both humans and the NOD mouse model is due to the aberrant autoimmune G. Proetzel, M.V. Wiles (eds.), Mouse Models for Drug Discovery, Methods in Molecular Biology 602, DOI 10.1007/978-1-60761-058-8 8, © Humana Press, a part of Springer Science+Business Media, LLC 2010
119
120
Serreze et al.
destruction of insulin-producing pancreatic  cells by T lymphocytes (reviewed in 1–3). In both genera, multiple susceptibility genes (designated Idd in mice; IDDM in humans) interactively contribute to T1D development. However, while polygenically controlled, particular major histocompatibility complex (MHC) haplotypes provide the primary risk factor for T1D development. The MHC (designated H2 in mice and HLA in humans) encodes two primary types of gene products termed class I and class II molecules (reviewed in 4). Class I molecules present peptides primarily derived from intracellular proteins to CD8 T cells that usually exert cytotoxic functions. Virtually all cells express MHC class I molecules. In contrast, MHC class II expression is largely limited to a specialized subset of hematopoietically derived antigen presenting cells (APC) that include B cells, macrophages, and dendritic cells (DC). MHC class II molecules expressed by APC display peptides largely derived from internalized and processed extracellular proteins to CD4 T cells, which produce cytokine molecules that amplify other components of the immune response, including cytotoxic CD8 T cells. Within the MHC, specific combinations of HLA-DQ and DR class II variants provide a large component of T1D susceptibility in humans by mediating the selection and functional activation of  cell autoreactive CD4 T cells (5, 6). Similarly, transgenic analyses (7–11) have demonstrated that T1D development in NOD mice requires that APC homozygously express the unusual H2-Ag7 MHC class II gene product (homolog of human DQ8) in the absence of H2E MHC class II molecules (homolog of human DR). However, while MHC class II-restricted CD4 T-cell responses clearly contribute to T1D development in both humans and NOD mice, there is compelling evidence that MHC class I-restricted CD8 T cells also play an essential pathogenic role. Because they lack expression of MHC class II molecules, pancreatic  cells cannot be directly recognized by autoreactive CD4 T cells contributing to T1D development. Furthermore, studies in NOD mice have shown that purified CD4 T cells from young pre-diabetic female donors (which would ultimately demonstrate a disease frequency of ∼90%) cannot independently transfer T1D to T- and B-cell deficient NOD-scid recipients (12). This ruled out the possibility that autoreactive MHC class II-restricted CD4 T cells could independently initiate T1D development by destroying pancreatic  cells through the release of some soluble cytotoxic factor(s). However, like most other cell types, pancreatic  cells do express MHC class I molecules. In order to bind antigenic peptides and then be transported to, and expressed in a stable fashion on the cell surface, MHC class I molecules must dimerize with 2-microglobulin (2m) (4). As a result, mice carrying a β2m allele inactivated by homologous recombination (β2mnull ) fail to express detectable levels of cell surface
Bridging Mice to Men
121
MHC class I molecules, and hence cannot generate CD8 T cells (13, 14). NOD.β2mnull mice lacking MHC class I expression and CD8 T cells are completely resistant to T1D (15–17). These results proved that while the Kd and/or Db class I gene products encoded by the H2g7 MHC haplotype are common variants also characterizing many strains lacking autoimmune proclivity, when expressed in NOD mice they aberrantly mediate CD8 T-cell responses essential to T1D development. It was recently found that due to strong interactive effects provided by a polymorphic gene in the Idd7 locus on chromosome 7, that when the common class I variants characterizing the H2g7 MHC haplotype are expressed in NOD mice but not a C57BL/6 (B6) background strain, they aberrantly lose the ability to mediate the deletion of autoreactive diabetogenic CD8 T cells during their development in the thymus (18). Similarly, a recent quite definitive study (19), coupled with earlier suggestive reports (20–29), revealed that when expressed in some genetic contexts, including the presence of particular MHC class II molecules, certain quite common class I variants can also contribute to T1D development in humans. It is difficult to directly determine the mechanistic role played by MHC class I genes in the initiation and amplification of diabetogenic T-cell responses in humans. However, insights to the types of immune responses that are controlled by various human MHC class I alleles have been gained through their transgenic expression in mice. Indeed, the antigenic peptides presented to murine CD8 T cells by a previously described HLA-A2.1 class I transgene product overlap those presented to human CD8 T cells by endogenously encoded HLA-A2.1 molecules (30, 31). Interestingly, HLA-A2.1 has been implicated as one common human MHC class I variant (present in ∼40% of Caucasians) that when expressed in the right genetic context can aberrantly exert diabetogenic functions in some individuals (19). Thus, this chapter will focus on how possible diabetogenic roles played by HLAA2.1, and perhaps other human class I variants, can be assessed through their transgenic expression in NOD mice. There will also be discussion of how such “humanized” NOD transgenic mouse strains might be used to identify approaches to block the development or function of HLA class I mediated diabetogenic T-cell responses.
2. Material 2.1. Transgenic Production
1. Pregnant mare serum gonadotropin (PMSG) 2. Human chorionic gonadotropin (hCG) 3. Freund’s adjuvant
122
Serreze et al.
2.2. Mouse Strains
1. NOD/ShiLtJ (The Jackson Laboratory, stock 001976) 2. NOD.CB17-Prkdcscid /J (The Jackson Laboratory, stock 001803) 3. NOD.β2mnull mice: NOD.129P2(B6)-B2mtm1Unc /J (The Jackson Laboratory stock 002309) 4. NOD.β2mnull .HHD mice: NOD.129P2(B6)-B2mtm1Unc Tg(HLA-A/H2-D/B2M)1Dvs/DvsJ (The Jackson Laboratory, stock 006611)
3. Methods 3.1. Direct Introduction of Transgenes into NOD Zygotes: Advantages and Keys to Success
A wide variety of transgenes have been expressed in NOD mice, either by repeated crossing of non-NOD transgenic strains to NOD mice or by direct introduction of transgenes into NOD zygotes. In the former case, because at least 20 susceptibility loci contribute to T1D development in NOD mice (2), extensive backcrossing is required to insure that the transgenic strain is fixed to homozygosity for each of the NOD-derived Idd loci. This process can be facilitated somewhat using a “speed congenic” approach (32), but it remains nonetheless expensive and time consuming. Furthermore, difficulties can arise if a transgene has integrated within or near an Idd gene. For example, when a knockout allele for interferon-␥ receptor alpha chain was transferred by breeding from 129 to NOD mice, an original study concluded the engineered mutation contributed to T1D resistance (33). However, more extensive backcrossing and further analysis revealed that the T1D resistance phenotype was actually mediated by a 129 derived gene(s) flanking the knockout allele (34). A similar scenario could also arise in the case of a transgene. For these reasons, direct introduction of transgenes into NOD zygotes is the method of choice to assess their possible effects on T1D pathogenesis. Differences in the reproductive biology of NOD mice compared to strains more commonly used for the production of transgenic mice must be taken into account in order for this procedure to be successful. Some of these differences have been noted previously (35). The Jackson Laboratory routinely creates transgenic mice on the NOD background. Transgenic models have been made on NOD/ShiLtJ, NOD/ShiLtDVS, and NOD.CB17-Prkdcscid /J. NOD strains can be used successfully for transgenic mouse production following many of the same basic procedures as those used for more common strains such as FVB and B6. There are, however, some important considerations when using NOD mice.
Bridging Mice to Men
123
We have found that using mature females as embryo donors is a key deviation from standard transgenic mouse production practices. The mature females typically produce fewer oocytes compared to pre-pubescent females, but the oocytes that are released tend to yield more developmentally synchronized populations of zygotes as well as zygotes of similar quality and “injectibility.” 1. Females 8–12 weeks of age are superovulated by intraperitoneal (IP) injection of 5 I.U. each of PMSG and hCG delivered 47–48 h apart (see Note 1). 2. Superovulated females are mated to sexually mature stud males immediately after hCG administration. The females are examined the following morning for the presence of copulation plugs and are separated into two groups: plugged and non-plugged. All females are used for zygote collection regardless of plug status. We have found that many females without obvious copulation plugs yield fertilized oocytes. 3. Mature NOD stud male mice are singly housed and mated no more than once each week. Injection of Freund’s adjuvant may be given to delay onset of the diabetic phenotype (see Note 2). 4. No special considerations are taken regarding the collection, culture, microinjection, and surgical transfer of NOD embryos beyond those associated with standard strains used for transgenic mouse creation (for details see 36). One can expect the release of 15–20 oocytes per superovulated female. As many as 30 oocytes per female can be obtained from certain NOD strains. Fertility averages 50% for the NOD strains that we have used to make transgenic models. The timing of hormone administration and matings to yield zygotes with pronuclei large enough to microinject will vary depending on factors such as the light/dark cycle of the vivarium and the desired time of DNA microinjection (morning or afternoon). The Jackson Laboratory most often utilizes a 14/10 h light/dark cycle with lights on at 06:00 and lights off at 20:00. PMSG is administered to NOD embryo donors at 10:00 and hCG and mating occurs at 09:00. Plugs are checked the following morning by 08:00 and microinjection occurs between 09:00 and 14:00. DNA-microinjected NOD zygotes are transferred to the oviducts of 0.5 d post coitum pseudopregnant recipient females immediately after microinjection. 5. Typically, 20–25 microinjected zygotes are transferred unilaterally to each recipient female. Litter sizes of 5–7 pups can be expected. Rates of NOD transgenesis will vary due to various factors surrounding the construction and
124
Serreze et al.
preparation of the transgene, but production rates of 5–15% are not uncommon. 3.2. Human HLA-A2.1 Class I Molecules Transgenically Expressed in NOD Mice Mediate Diabetogenic CD8 T-cell Responses
The genetic association studies described earlier implicated HLAA2.1 as one common human MHC class I variant that may aberrantly exert diabetogenic functions in some individuals. Thus, we assessed whether an HLA-A2.1 transgene product expressed in NOD mice could mediate autoreactive CD8 T-cell responses against pancreatic  cells. This was indeed found to be the case, and when added to the responses mediated by endogenous murine H2g7 MHC class I molecules, the result was a significantly accelerated rate of T1D development in our NOD.HLA-A2.1 transgenic strain (37). The accelerated rate of T1D development in NOD.HLA-A2.1 mice was not a generic effect of transgenically expressing any human class I variant in addition to the endogenous murine molecules, since introduction of the B27 allele actually suppressed disease onset (37). Similarly, the B27 class I variant was also recently found to exert a T1D protective effect in humans (19). These studies provided the first functional demonstration that specific human MHC class I molecules can mediate diabetogenic immune responses in addition to those previously known to be elicited by particular class II variants. Furthermore, the finding that HLA-A2.1 facilitates T1D development in both humans and NOD mice, while HLA-B27 is protective in both, further supports the usefulness of HLA-transgenic NOD strains as models for the human disease. We subsequently introduced, directly into NOD mice, a chimeric monochain transgene construct, designated HHD, that encodes human 2m covalently linked to the ␣1 and ␣2 domains of HLA-A∗0201, and the ␣3, transmembrane, and cytoplasmic domains of murine H-2Db . The HHD transgene was then transferred to the NOD.β2mnull strain. Due to its covalent linkage, the human 2m encoded by the HHD construct cannot stabilize the expression of murine class I molecules in a trans-acting fashion. Hence, NOD.β2mnull .HHD mice express human HLAA2.1, but not murine class I molecules (38). Most importantly, while albeit present at lower levels than in standard NOD mice, the NOD.β2mnull .HHD strain generates a sufficient array of  cell autoreactive CD8 T cells to induce T1D (38). Due to its ability to preferentially interact with murine CD8, the inclusion of a murine class I MHC ␣3 domain in the HHD construct facilitates T-cell development and antigen recognition (39). It should be noted that any human HLA class I coding sequence can be incorporated into the HHD transgene vector. For this reason it should be possible to use the HHD transgene system to generate NOD background mice that could be used to evaluate the potential diabetogenic function of any chosen human HLA class I variant.
Bridging Mice to Men
3.3. NOD.2mnull .HHD Mice Can Be Used to Predict HLA-A2.1-Restricted β -Cell Antigens Targeted by CD8 T Cells in Human T1D Patients
125
Since all of their CD8 T cells must be HLA-A2.1 restricted, we reasoned the NOD.β2mnull .HHD strain might provide an excellent model for identifying  cell autoantigens that are presented by this class I variant to pathogenic CD8 T cells in human T1D patients. Indeed, CD8 T cells could be isolated from NOD.β2mnull .HHD mice that were specifically cytotoxic to human HLA-A2.1 positive islet cells (38). This demonstrated murine and human HLA-A2.1-positive islets present one or more peptides in common. Previous studies employing a tetramerbased technology approach found that peptides derived from islet-specific glucose-6-phosphatase catalytic subunit-related protein (IGRP 206-214 epitope), insulin (B chain 15-23 epitope), and dystrophia myotonica kinase (DMK 138-146 epitope) represented the targets of up to ∼60% of H2g7 class I-restricted diabetogenic CD8 T cells in standard NOD mice (40). While the proportion of islet derived CD8 T cells that recognized these antigenic epitopes varied greatly in individual NOD mice, usually the most frequent target was IGRP 206-214, followed by insulin B 15-23 and DMK 138-146 (40). Thus, we have been carrying out studies to determine if HLA-A2.1-restricted T-cell lines propagated from the pancreatic islets of NOD.β2mnull .HHD mice can also recognize IGRP-, insulin-, or DMK-derived peptides. To date a series of islet-derived CD8 T-cell lines from 12to 13 week-old female NOD.β2mnull .HHD mice have been used to screen peptide libraries consisting of all possible 8- to 11mer sequences that can be derived from IGRP or preproinsulin (both preproinsulin 1 and 2). Three peptide sets from each of these  cell proteins were found to stimulate at least a subset of islet-derived CD8 T cell lines from NOD.β2mnull .HHD mice, as assessed by ELISPOT analysis of IFN␥ production. The individual peptides comprising each positive set were then separately screened for autoantigenic activity. These analyses revealed that the three individual IGRP-derived peptides recognized by isletderived CD8 T cells from NOD.β2mnull .HHD mice consisted of amino acid residues 228-236, 265-273, and 337-345 (38 and Table 8.1). The 228-236 peptide appeared to represent the immunodominant epitope recognized by IGRP autoreactive CD8 T cells in NOD.β2mnull .HHD mice (38 and Table 8.1). The IGRP 265-273 epitope was also frequently targeted, while IGRP 337-345 represented a less frequent target (38 and Table 8.1). For insulin epitopes, we found amino acid residues 3-11 from the leader sequence and 5-14 from the B chain of Ins1, as well as an A chain 2-10 epitope common to both Ins1 and Ins2 were HLA-A2.1 restricted targets of islet-derived CD8 T cell lines from NOD.β2mnull .HHD mice (41 and Table 8.1). The A chain 2-10 peptide was found to be the most frequently targeted epitope of insulin autoreactive CD8 T cells in NOD.β2mnull .HHD mice
126
Serreze et al.
Table 8.1 Murine and human IGRP and insulin peptides recognized in a HLA-A2.1-restricted fashion by islet-derived CD8 T cells from NOD.ß2mnull .HHD mice
A2-restricted murine autoantigenic β cell peptide
#NOD.β 2mnull .HHD CD8 T-cell lines recognizing murine peptide
Homologous human peptide∗
Human peptide recognized by NOD.β 2mnull .HHD CD8 T cells?
IGRP228-236
FGIDLLWSV
14/16
LNIDLLWSV
YES
IGRP265-273
VLFGLGFAI
8/16
VLFGLGFAI
YES
IGRP337-345
ALIPYCVHM
4/16
AFIPYSVHM
NO
INS1 L3-11
LLVHFLPLL
9/22
LWMRLLPLL
nd
INS1 B5-14
HLCGPHLVEA
17/22
HLCGSHLVEA
YES
INS1/2 A2-10
IVDQCCTSI
20/22
IVEQCCTSI
YES
nd, not determined. ∗ Bold and underlined letters indicate amino acid differences distinguishing human from murine IGRP or insulin sequences.
(41 and Table 8.1). Ins1 B 5-14 was also a frequent target (41 and Table 8.1). The Ins1 L3-11 epitope was the least frequent target (41 and Table 8.1). It should be added that most isletderived CD8 T-cell lines from NOD.β2mnull .HHD mice included specificities that recognized at least one peptide each of IGRP or insulin origin. The studies described above identified the murine variants of IGRP or insulin peptides that are recognized in an HLA-A2.1-restricted fashion by islet-derived CD8 T cells from NOD.β2mnull .HHD mice. However, it remained possible that the human counterparts of these peptides might not bind to HLAA2.1 class I molecules and be presented to pathogenic CD8 T cells in T1D patients. This turned out not to be a concern for the human (h) IGRP 265-273 peptide since its sequence is identical to the murine homologue (38 and Table 8.1). Conversely, the human and murine variants of IGRP 228-236 and IGRP 337-345 differed by two amino acids each (38 and Table 8.1). The hIGRP 337-345 variant was found to bind very poorly to HLA-A2.1 molecules, which probably accounted for its inability to stimulate islet-derived CD8 T cells from NOD.β2mnull .HHD mice (38). In contrast, the hIGRP 228-236 variant demonstrated strong HLA-A2.1 binding and was recognized by islet-derived CD8 T cells from NOD.β2mnull .HHD mice. The murine Ins1 B 5-14 and Ins1/2 A 2-10 peptides differ from their human counterparts by a single amino acid (41 and Table 8.1). However, the human homologues of the Ins1 B 5-14 and Ins1/2 A 2-10 peptides are also recognized by islet-derived CD8 T cells from NOD.β2mnull .HHD mice (41 and Table 8.1). These results indicated the human Ins A 2-10, Ins B 5-14, IGRP 228-236, and/or
Bridging Mice to Men
127
IGRP 265-273 peptides may represent relevant autoantigenic targets of pathogenic CD8 T cells in HLA-A2.1 expressing T1D patients. Partly based on our analyses of NOD.β2mnull .HHD mice, there has been a recent acceleration in efforts by other investigators to identify populations of autoreactive MHC class Irestricted CD8 T cells potentially contributing to T1D development in humans (42–47). One such study (42), that was of particular interest from our perspective, reported that T1D patients can generate HLA-A2.1-restricted CD8 T cells against the IGRP 228-236 or IGRP 265-273 peptides which we had originally identified to be pathologically relevant targets in NOD.β2mnull .HHD mice. Our work describing insulin peptides recognized by islet-derived CD8 T cells from NOD.β2mnull .HHD mice was published more recently than that focused on IGRP responses. However, we hope our study describing insulin epitopes recognized by islet-derived CD8 T cells from NOD.β2mnull .HHD mice also prompts clinical investigators to screen HLA-A2.1 expressing human T1D patients for the presence of similar pathogenic effectors. 3.4. Use of NOD.2mnull .HHD Mice for Identifying Means to Block HLA-A2.1-Restricted Diabetogenic T-Cell Responses
“Humanized” NOD.β2mnull .HHD mice provide a model system for developing strategies to inhibit the generation or function of HLA-A2.1-restricted T cells contributing to T1D development that might ultimately be translatable to individuals expressing this class I variant and deemed to be at high future disease risk. Several different approaches could be envisioned for attenuating diabetogenic HLA-A2.1-restricted CD8 T-cell responses in NOD.β2mnull .HHD mice that might ultimately be suitable for clinical use. Previous studies have demonstrated that in standard NOD mice, IGRP autoreactive CD8 T cells can be induced to undergo deletion by appropriately dosed injections of a soluble antigenic peptide, and this is sufficient to inhibit T1D development (48). On the other hand, it has also been reported that in standard NOD mice, a CD8 T-cell response against (pro)insulin epitopes must initially be established to allow for the subsequent appearance of IGRP-specific effectors (49). An ability to block such “epitope spreading” could at least partly explain why protocols that induce T-cell tolerance to (pro)insulin can also efficiently inhibit T1D development in standard NOD mice (49–51). However, at this point it is unknown whether in the presence of human HLA-A2.1, rather than murine H2g7 class I molecules expressed by standard NOD mice, T1D development still requires the initial generation of autoreactive CD8 T-cell responses against (pro)insulin as a prerequisite to trigger a subsequent appearance of IGRP-specific effectors. The NOD.β2mnull .HHD mouse strain should provide a resource making it possible to determine the relative T1D protective effect that may be achieved by
128
Serreze et al.
ablating autoreactive HLA-A2.1-restricted CD8 T cells that recognize (pro)insulin and/or IGRP epitopes through a tolerogenic antigenic peptide administration approach. Tolerogenic delivery of peptide epitopes to steady-state DCs using an antibody to the DC endocytic receptor DEC-205 has also shown promise in inducing CD8 T-cell deletion in NOD mice and should also be considered for this purpose (52). Through the use of mixed hematopoietic chimera systems, multiple investigators have found that the development of diabetogenic T cells from precursors in the bone marrow (BM) of standard NOD mice is blocked when they are forced to mature in the presence of APC expressing MHC molecules other than those encoded by the H2g7 haplotype (53–62). Thus, provided a relatively benign pre-conditioning protocol is ultimately developed, hematopoietic chimerization by APCs expressing dominantly protective MHC molecules could conceivably provide a means for blocking progression to overt T1D in humans deemed to be a high future disease risk. An obviously important consideration is what array of MHC molecules APC must express to most efficiently block the development and/or functional activation of diabetogenic T cells. We have found that APC expressing MHC variants that elicit a strong cross-reactive allogeneic response by a mature autoreactive CD8 cell clonotype (AI4) contributing to T1D in standard NOD mice will also have the ability to mediate the negative selection of these pathogenic effectors during thymic development (63). Thus, a similar analysis of allogeneic cross reactivity could identify MHC haplotypes that might block the development of HLA-A2.1-restricted diabetogenic T cells which are normally generated in NOD.β2mnull .HHD mice. As an initial “proof of principal” we have identified murine MHC haplotypes that provide strongly cross-activating allogeneic ligands for CD8 T cells from NOD.β2mnull .HHD mice which have previously responded to priming with a cocktail of the HLA-A2.1restricted peptides Ins1/2 A 2-10, Ins1 B 5-14, IGRP 228-236, and IGRP 265-273 (unpublished). Using a partial BM chimerization approach (61), it will then be determined if APC expressing the potentially appropriate allogeneic MHC molecules have the capacity to block development of any of these HLA-A2.1restricted (pro)insulin and/or IGRP autoreactive CD8 T-cell populations that are normally generated in NOD.β2mnull .HHD mice. The NOD.β2mnull .HHD mouse strain is also undergoing further modifications so it can be used in a system to assess the capacity of human APC expressing various HLA molecules to block the development of HLA-A2.1-restricted diabetogenic CD8 T-cell responses. Other investigators have found that human hematopoietic stem cells (HSC) engrafted in a NOD-scid strain also carrying a functionally inactivated IL-2 receptor gamma
Bridging Mice to Men
129
chain gene (NOD-scid.IL2rγ null ) give rise to functional populations of all tested lineages of lymphoid and myeloid cells (64). NOD-scid.IL2rγ null mice can also be engrafted with human peripheral blood mononuclear cells (65), further increasing the utility of models based on this strain. The HHD transgene and the β2mnull mutation have recently been transferred to the NODscid.IL2rγ null strain (unpublished). Development of functional human immune cell populations occurs most readily in NODscid.IL2rγ null mice engrafted at 1–3 days of age with 3–10 × 104 of the test HSC (marked by CD34 expression) after receiving a low 100R pre-conditioning dose of irradiation (66). This protocol could be utilized to separately or co-engraft NODscid.IL2rγ null .β2mnull .HHD mice with the human HSC of choice plus NOD.β2mnull .HHD BM. This will make it possible to determine if human APC expressing particular allogeneic HLA molecules have the capacity to temper the development of HLAA2.1-restricted diabetogenic CD8 T cells that would normally differentiate from precursors in NOD.β2mnull .HHD BM. Such information would be invaluable in determining what donor type cells would be most likely to confer strong T1D protection in any hematopoietic chimerization protocol ultimately approved for clinical use. 3.5. Conclusions and Future Directions
Studies to date have demonstrated the value of using “humanized” HLA transgenic NOD mice to identify pancreatic  cell antigenic peptides that are targets of pathogenic autoreactive T cells in T1D patients. As depicted in Fig. 8.1, NOD HLA transgenic mice could additionally provide a model system for initially developing clinically translatable protocols for blocking the development or function of autoreactive T-cell populations also contributing to T1D in humans. The efficacy of various potential T1D intervention protocols may be enhanced when initiated at the earliest possible stage of disease development. Thus, there is also a need to more clearly identify individuals at future T1D risk at earlier prodromal stages of disease development than now possible. Additional markers that might ultimately be useful in assessing an individual’s future risk for T1D development may be circulating levels of various autoantigen-specific T-cell populations first identified in humanized HLA transgenic NOD mice. To date, the currently available NOD.β2mnull .HHD mouse strain has made it possible to identify T-cell populations that contribute to T1D by recognizing  cell autoantigens presented by the human HLA-A2.1 class I variant. However, there is now strong evidence that when expressed in the proper genetic context, other human HLA class I molecules, such as the A24 and B39 variants, can also mediate diabetogenic CD8 T-cell responses (19). For this reason there is also a need to utilize the HHD transgenic approach
130
Serreze et al.
Fig. 8.1. Possible approaches for how NOD HLA transgenic mice can provide a model system for initially developing clinically translatable means for blocking the development or function of autoreactive T-cell populations contributing to T1D development in human patients.
to generate NOD background mice only expressing these additional human HLA class I variants, or any others subsequently found to potentially contribute to T1D development. This broadening array of HLA transgenic NOD mouse strains should not only expand identification of T-cell populations contributing to, and also potentially providing markers for future risk of T1D in humans, but may also allow for the initial development of intervention protocols that may prevent disease onset in otherwise susceptible individuals.
4. Notes 1. Females at 6–7 weeks of age have also been used successfully, but zygote quality is not as consistent as from older females. Some of the sexually mature females may be in proestrus or estrus at the time of gonadotropin administration and injection of exogenous gonadotropins may not successfully induce superovulation. In our experience, however, the quality of zygotes produced by superovulating mature females (compared to traditional use of prepubescent females) outweighs the loss of overall zygote production due to failure to superovulate. 2. If stud males are not routinely used for mating to superovulated females, we find it advantageous to “practice” mate the stud males to any strain of female on-hand 1–2 weeks prior to the next transgenic NOD experiment. This helps ensure the highest plug and fertility rates when it comes time to mate to superovulated NOD females.
Bridging Mice to Men
131
References 1. Anderson, M. S. and Bluestone, J. A. (2005) The NOD mouse: a model of immune dysregulation. Annu. Rev. Immunol. 23, 447– 485. 2. Serreze, D. V. and Leiter, E. H. (2001) Genes and pathways underlying autoimmune diabetes in NOD mice, in Molecular Pathology of Insulin Dependent Diabetes Mellitus (von Herrath, M. G., ed.). Karger Press, New York, pp. 31–67. 3. Onengut-Gumuscu, S. and Concannon, P. (2006) Recent advances in the immunogenetics of human type 1 diabetes. Curr. Opin. Immunol. 18, 634–638. 4. Mellman, I. (2007) Private lives: reflections and challenges in understanding the cell biology of the immune system. Science 317, 625– 627. 5. Aly, T. A., Ide, A., Jahromi, M. M., Barker, J. M., Fernando, M. S., et al. (2006) Extreme genetic risk for type 1A diabetes. Proc. Natl. Acad. Sci. USA 103, 14074–14079. 6. Erlich, H., Valdes, A. M., Noble, J., Carlson, J. A., Varney, M., et al. (2008) HLA DR-DQ haplotypes and genotypes and type 1 diabetes risk: analysis of the type 1 diabetes genetics consortium families. Diabetes 57, 1084– 1092. 7. Hanson, M. S., Cetkovic-Cvrlje, M., Ramiya, V. K., Atkinson, M. A., MacLaren, N. K., et al. (1996) Quantitative thresholds of MHC class II I-E expressed on hematopoietically derived APC in transgenic NOD/Lt mice determine level of diabetes resistance and indicate mechanism of protection. J. Immunol. 157, 1279–1287. 8. Lund, T., O Reilly, L., Hutchings, P., Kanagawa, O., Simpson, E., et al. (1990) Prevention of insulin-dependent diabetes mellitus in non-obese diabetic mice by transgenes encoding modified I-A -chain or normal I-E a-chain. Nature 345, 727–729. 9. Miyazaki, T., Uno, M., Uehira, M., Kikutani, H., Kishimoto, T., et al. (1990) Direct evidence for the contribution of the unique I-Anod to the development of insulitis in non-obese diabetic mice. Nature 345, 722–724. 10. Singer, S. M., Tisch, R., Yang, X. -D. and McDevitt, H. O. (1993) An Abd transgene prevents diabetes in nonobese diabetic mice by inducing regulatory T cells. Proc. Natl. Acad. Sci. USA 90, 9566–9570. 11. Slattery, R. M., Kjer-Nielsen, L., Allison, J., Charlton, B., Mandel, T., et al. (1990) Prevention of diabetes in non-obese diabetic IAk transgenic mice. Nature 345, 724–726.
12. Christianson, S. W., Shultz, L. D. and Leiter, E. H. (1993) Adoptive transfer of diabetes into immunodeficient NOD-scid/scid mice: relative contributions of CD4+ and CD8+ T lymphocytes from diabetic versus prediabetic NOD.NON-Thy 1a donors. Diabetes 42, 44– 55. 13. Zijlstra, M., Bix, M., Simister, N. E., Loring, J. M., Raulet, D. H., et al. (1990) 2microglobulin deficient mice lack CD4- 8+ cytolytic T cells. Nature 344, 742–746. 14. Koller, B. H., Marrack, P., Kappler, J. W. and Smithes, O. (1990) Normal development of mice deficient in b2m, MHC class I proteins, and CD8+ T cells. Science 248, 1227–1230. 15. Serreze, D. V., Leiter, E. H., Christianson, G. J., Greiner, D. and Roopenian, D. C. (1994) MHC class I deficient NOD-B2mnull mice are diabetes and insulitis resistant. Diabetes 43, 505–509. 16. Wicker, L. S., Leiter, E. H., Todd, J. A., Renjilian, R. J., Peterson, E., et al. (1994) 2microglobulin-deficient NOD mice do not develop insulitis or diabetes. Diabetes 43, 500–504. 17. Sumida, T., Furukawa, M., Sakamoto, A., Namekawa, T., Maeda, T., et al. (1994) Prevention of insulitis and diabetes in beta(2)microglobulin-deficient non-obese diabetic mice. Int. Immunol. 6, 1445–1449. 18. Serreze, D. V., Choisy-Rossi, C.-M., Grier, A., Holl, T. M., Chapman, H. D., et al. (2008) Through regulation of TCR expression levels, an Idd7 region gene(s) interactively contributes to the impaired thymic deletion of autoreactive diabetogenic CD8 Tcells in NOD mice. J. Immunol. 180, 3250– 3259. 19. Nejentsev, S., Howson, J. M. M., Walker, N. M., Szeszko, J., Field, S. F., et al. (2007) Localization of type 1 diabetes susceptibility to the MHC class I genes HLA-B and HLAA. Nature 450, 887–892. 20. Demaine, A. G., Hibberd, M. L., Mangles, D. and Millward, B. A. (1995) A new marker in the HLA class I region is associated with the age at onset of IDDM. Diabetologia 38, 622–628. 21. Fennessy, M., Metcalfe, K., Hitman, G. A., Niven, M., Biro, P. A., et al. (1994) A gene in the HLA class I region contributes to susceptibility to IDDM in the Finnish population. Diabetologia 37, 937–944. 22. Honeyman, M. C., Harrison, L. C., Drummond, B., Colman, P. G. and Tait, B. D. (1995) Analysis of families at risk for insulindependent diabetes mellitus reveals that HLA
132
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
Serreze et al. antigens influence progression to clinical disease. Mol. Med. 1, 576–582. Mizota, M., Uchigata, Y., Moriyama, S., Tokunaga, K., Matsuura, N., et al. (1996) Age dependent association of HLA-A24 in Japanese IDDM patients. Diabetologia 39, 371–373. Nakanishi, K., Kobayashi, T., Murase, T., Naruse, T., Nose, Y., et al. (1999) Human leukocyte antigen-A24 and-DQA1∗0301 in Japanese insulin-dependent diabetes mellitus: independent contributions to susceptibility to the disease and additive contributions to acceleration of beta-cell destruction. J. Clin. Endocrinol. Metab. 84, 3721–3725. Nejentsev, S., Reijonen, H., Adojaan, B., Kovalchuk, L., Sochnevs, A., et al. (1997) The effect of HLA-B allele on the IDDM risk defined by DRB1∗04 subtypes and DQB1∗0302. Diabetes 46, 1888–1892. Nejentsev, S., Gombos, Z., Laine, A.-P., Veijola, R., Knip, M., et al. (2000) Non-class II HLA gene associated with type 1 diabetes maps to the 240-kb region near HLA-B. Diabetes 49, 2217–2221. Robles, D. T., Eisenbarth, G. S., Wang, T., Erlich, H. A., Bugawan, T. L., et al. (2002) Identification of children with early onset and high incidence of anti-islet autoantibodies. Clin. Immunol. 102, 217–224. Tait, B. D., Colman, P. G., Morahan, G., Marchinovska, L., Dore, E., et al. (2003) HLA genes associated with autoimmunity and progression to disease in type 1 diabetes. Tissue Antigens 61, 146–153. Undlien, D. E., Lie, B. A. and Thorsby, E. (2001) HLA complex genes in type 1 diabetes and other autoimmune diseases. Which genes are involved? Trends Genet. 17, 93–100. Geluk, A., van Meijgaarden, K. E., Fraken, K. L. M. C., Drijfhout, J. W., D Souza, S., et al. (2000) Identification of major epitopes of Mycobacterium tuberculosis AG85B that are recognized by HLA-A0201 restricted CD8+ T cells in HLA-transgenic mice and humans. J. Immunol. 165, 6463–6471. Shirai, M., Arichi, T., Nishioka, M., Nomura, T., Ikeda, K., et al. (1995) CTL responses of HLA-A2.1-transgenic mice specific for Hepatitis C viral peptides predict epitopes for CTL of humans carrying HLA-2.1. J. Immunol. 154, 2733–2742. Serreze, D. V., Chapman, H. D., Varnum, D. S., Hanson, M. S., Reifsnyder, P. C., et al. (1996) B lymphocytes are essential for the initiation of T cell mediated autoimmune diabetes: analysis of a new “speed congenic”
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
43.
stock of NOD.Ignull mice. J. Exp. Med. 184, 2049–2053. Wang, B., Andre, I., Gonzalez, A., Katz, J. D., Aguet, M., et al. (1997) Interferon-g impacts at multiple points during progression of autoimmune diabetes. Proc. Natl. Acad. Sci. USA 94, 13844–13849. Kanagawa, O., Xu, G., Tevaarwerk, A. and Vaupel, B. A. (2000) Protection of nonobese diabetic mice from diabetes by gene(s) closely linked to IFN-g receptor loci. J. Immunol. 164, 3919–3923. Leiter, E. H. (1997) The NOD mouse: a model for insulin-dependent diabetes mellitus. Curr. Protocols Immunol. 24, 15.19.11– 15.19.23. Nagy, A., Gertsenstein, M., Vintersten, K. and Behringer, R. (2003) Manipulating the Mouse Embryo: A Laboratory Manual, third ed ed. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, New York. Marron, M. P., Graser, R. T., Chapman, H. D. and Serreze, D. V. (2002) Functional evidence for the mediation of diabetogenic T cell responses by human HLA-A2.1 MHC class I molecules through transgenic expression in NOD mice. Proc. Natl. Acad. Sci. USA 99, 13753–13758. Takaki, T., Marron, M. P., Mathews, C. E., Guttman, S. T., Bottino, R., et al. (2006) HLA-A∗0201-restricted T cells from humanized NOD mice recognize autoantigens of potential clinical relevance to type 1 diabetes. J. Immunol. 176, 3257–3265. Irwin, M. J., Heath, W. R. and Sherman, L. A. (1989) Species-restricted interactions between CD8 and the alpha 3 domain of class I influence the magnitude of the xenogeneic response. J Exp Med 170, 1091–1101. Lieberman, S. M., Takaki, T., Han, B., Santamaria, P., Serreze, D. V., et al. (2004) Individual nonobese diabetic mice exhibit unique patterns of CD8 T cell reactivity to three islet antigens including the newly identified widely expressed dystrophia myotonica kinase. J. Immunol. 173, 6727–6734. Jarchum, I., Baker, J. C., Yamada, T., Takaki, T., Marron, M. P., et al. (2007) In vivo cytotoxicity of insulin specific CD8+ T cells in HLA-A∗0201 transgenic NOD mice. Diabetes 56, 2551–2560. Mallone, R., Martinuzzi, E., Blancou, P., Novelli, G., Afonzo, G., et al. (2007) CD8+ T cell responses identify -cell autoimmunity in human type 1 diabetes. Diabetes 56, 613–621. Ouyang, Q., Standifer, N. E., Qin, H., Gottlieb, P. A., Verchere, C. B., et al. (2006) Recognition of HLA-class I restricted -cell
Bridging Mice to Men
44.
45.
46.
47.
48.
49.
50.
51.
52.
53.
54.
epitopes in type 1 diabetes. Diabetes 55, 3068–3074. Panagiotopoulos, C., Qin, H., Tan, R. and Verchere, C. B. (2003) Identification of a  cell specific HLA class I restricted epitope in type 1 diabetes. Diabetes 52, 2647–2651. Pinske, G. G. M., Tysma, O. H. M., Bergen, C. A. M., Kester, M. G. D., Ossendorp, F., et al. (2005) Autoreactive CD8 T cells associated with  cell destruction in type 1 diabetes. Proc. Natl. Acad. Sci. USA 102, 18425–18430. Standifer, N. E., Ouyang, Q., Panagiotopoulos, C., Verchere, C. B., Tan, R., et al. (2006) Identification of novel HLAA∗0201-restricted epitopes in recent onset type 1 diabetic subjects and antibody positive relatives. Diabetes 55, 3061–3067. Toma, A., Haddouk, S., Briand, J. P., Camoin, L., Gahery, H., et al. (2005) Recognition of a subregion of human proinsulin by class I restricted T cells in type 1 diabetic patients. Proc. Natl. Acad. Sci. USA 102, 10581–10586. Han, B., Serra, P., Amrani, A., Yamanouchi, J., Maree, A. F., et al. (2005) Prevention of diabetes by manipulation of anti-IGRP autoimmunity: high efficiency of a low affinity peptide. Nat. Med. 11, 645–652. Krishnamurthy, B., Dudek, N., McKenzie, M. D., Purcell, A. W., Brooks, A. G., et al. (2006) Responses against islet antigens in NOD mice are prevented by tolerance to proinsulin but not IGRP. J. Clin. Invest. 116, 3258–3265. Jaeckel, E., Lipes, M. A. and von Boehmer, H. (2004) Recessive tolerance to preproinsulin 2 reduces but does not abolish type 1 diabetes. Nat. Immunol. 5, 1028–1035. Nakayama, T., Abiru, N., Moriyama, H., Babaya, N., Liu, E., et al. (2005) Prime role for an insulin epitope in the development of type 1 diabetes in NOD mice. Nature 435, 220–223. Mukhopadhaya, A., Hanafusa, T., Jarchum, I., Chen, Y. G., Iwai, Y., et al. (2008) Selective delivery of beta cell antigen to dendritic cells in vivo leads to deletion and tolerance of autoreactive CD8+ T cells in NOD mice. Proc. Natl. Acad. Sci. U S A 105, 6374– 6379. Chen, Y.-G., Silveira, P., Osborne, M. A., Chapman, H. D. and Serreze, D. V. (2007) Cellular expression requirements for inhibition of type 1 diabetes by a dominantly protective major histocompatibility complex haplotype. Diabetes 56, 424–430. Ikehara, S., Ohtsuki, H., Good, R. A., Asamoto, H., Nakamura, T., et al. (1985)
55.
56.
57.
58.
59.
60.
61.
62.
133
Prevention of type 1 diabetes in nonobese diabetic mice by allogeneic bone marrow transplantation. Proc. Natl. Acad. Sci. USA 82, 7743–7747. Ildstad, S. T., Chilton, P. M., Xu, H., Domenick, M. A. and Ray, M. B. (2005) Preconditioning of NOD mice with anti-CD8 mAb and co-stimulatory blockade enhances chimerism and tolerance and prevents diabetes while depletion of a TCR+ and CD4 T cells negates the effect. Blood 105, 2577– 2584. LaFace, D. W. and Peck, A. B. (1989) Reciprocal allogeneic bone marrow transplantation between NOD mice and diabetesnonsusceptible mice associated with transfer and prevention of autoimmune diabetes. Diabetes 38, 894–901. Li, H., Kaufman, C. L., Boggs, S. S., Johnson, P. C., Patrene, K. D., et al. (1996) Mixed allogeneic chimerism induced by a sublethal approach prevents autoimmune diabetes and reverses insulitis in nonobese diabetic (NOD) mice. J. Immunol. 156, 380–388. Mathieu, C., Castells, K., Bouillon, R. and Waer, M. (1997) Protection against autoimmune diabetes in mixed bone marrow chimeras. J. Immunol. 158, 1453–1457. Nikolic, B., Takeuchi, Y., Leykin, I., Fudaba, Y., Smith, R. N., et al. (2004) Mixed hematopoietic chimerism allows cure of autoimmune diabetes through allogeneic tolerance and reversal of autoimmunity. Diabetes 53, 376–383. Serreze, D. V. and Leiter, E. H. (1991) Development of diabetogenic T cells from NOD/Lt marrow is blocked when an allo-H-2 haplotype is expressed on cells of hematopoietic origin, but not on thymic epithelium. J. Immunol. 147: 1222–1229. Serreze, D. V., Osborne, M. A., Chen, Y.-G., Chapman, H. D., Pearson, T., et al. (2006) Partial versus full allogeneic hemopoietic chimerization is a preferential means to inhibit type 1 diabetes as the latter induces generalized immunosuppression. J. Immunol. 177, 6675–6684. Seung, E., Iwakoshi, N., Woda, B. A., Markees, T. G., Mordes, J. P., et al. (2000) Allogeneic hematopoietic chimerism in mice treated with sublethal myeloablation and anti-CD154 antibody: absence of graft-versus-host disease, induction of skin allograft tolerance, and prevention of recurrent autoimmunity in isletallografted NOD/Lt mice. Blood 95, 2175–2182.
134
Serreze et al.
63. Serreze, D. V., Holl, T. M., Marron, M. P., Graser, R. T., Johnson, E. A., et al. (2004) MHC class II molecules play a role in the selection of autoreactive class I restricted CD8 T cells that are essential contributors to type 1 diabetes development in NOD mice. J. Immunol. 172, 871–879. 64. Shultz, L. D., Ishikawa, F. and Greiner, D. L. (2007) Humanized mice in translational biomedical research. Nat. Rev. Immunol. 7, 118–130.
65. King, M., Pearson, T., Shultz, L. D., Leif, J., Bottino, R., et al. (2008) A new Hu-PBL model for the study of human islet alloreactivity based on NOD-scid mice bearing a targeted mutation in the IL-2 receptor gamma chain gene. Clin. Immunol. 126, 303–314. 66. Ishikawa, F., Yasukawa, M., Lyons, B., Yoshida, S., Miyamoto, T., et al. (2005) Development of functional human blood and immune systems in NOD/SCID/IL2 ␥ chainnull mice. Blood 106, 1565–1573.
Chapter 9 Mouse Models of Type II Diabetes Mellitus in Drug Discovery Helene Baribault Abstract Type II diabetes is a fast-growing epidemic in industrialized countries. Many recent advances have led to the discovery and marketing of efficient novel therapeutic medications. Yet, because of side effects of these medications and the variability in individual patient responsiveness, unmet needs subsist for the discovery of new drugs. The mouse has proven to be a reliable model for discovering and validating new treatments for type II diabetes mellitus. We review here the most common mouse models used for drug discovery for the treatment of type II diabetes. The methods presented focus on measuring the equivalent end points in mice to the clinical values of glucose metabolism used for the diagnostic of type II diabetes in humans: i.e., baseline fasting glucose and insulin, glucose tolerance test, and insulin sensitivity index. Improvements on these clinical values are essential for the progression of a novel potential therapeutic molecule through a preclinical and clinical pipeline. Key words: Type II diabetes mellitus, drug discovery, glucose tolerance test, insulin tolerance test, insulin secretion, insulin sensitivity, diet-induced obesity, leptin, insulin, NEFA. Abbreviations: DEXA dual energy X-ray absorptiometry; MRI magnetic resonance imaging; DIO diet-induced obesity; GSIS glucose-stimulated insulin secretion; GTT glucose tolerance test; ITT insulin tolerance test; NEFA non-esterified fatty acid; T2DM type II diabetes mellitus; D-PBS Dulbecco’s phosphate-buffered saline; STZ streptozotocin; PK pharmacokinetics; PD pharmacodynamics; ED50 dose providing 50% efficacy.
G. Proetzel, M.V. Wiles (eds.), Mouse Models for Drug Discovery, Methods in Molecular Biology 602, DOI 10.1007/978-1-60761-058-8 9, © Humana Press, a part of Springer Science+Business Media, LLC 2010
135
136
Baribault
1. Introduction The house mouse, mus musculus, is characterized by its ability to live in close association with people. In fact, most have been dependant on human shelter and activity and have migrated along with human population for over 10,000 years. They are referred to as commensal animals because they share related food supplies. Mice like humans are omnivorous and overall are considered a good model to study the regulation of dietary intake and nutrient metabolism in humans. In addition, a wealth of information resources and experimental approaches for mouse genetics are available for the study of biological processes involved in human diseases (1). Type II diabetes mellitus (T2DM) is defined clinically by either of these two glucose values: a fasting plasma glucose higher than 126 mg/dL or a 2 h postload glucose level exceeding 200 mg/dL, in a glucose tolerance test (GTT). These clinical tests are used to diagnose diabetes in humans and also constitute the basic criteria used in mouse models of type II diabetes, albeit in the mouse baseline glucose values are higher. Early-stage type II diabetes is caused by a resistance to insulin, resulting in a decrease in glucose uptake mainly by hepatocytes and adipocytes [for review, (2)]. Consequently, the pancreas compensates for insulin resistance by secreting more insulin causing the individual to become hyperinsulinemic. At later stages in the disease, a degeneration of the insulin-secreting beta cells in pancreatic islets is observed and patients may become dependent on external administration of insulin. Chronic high-blood glucose levels are cytotoxic and cause diabetic nephropathy, retinopathy, and other secondary diseases including peripheral neuropathy and arterial diseases. Successful current therapies for type II diabetes include exenatide (Byetta), a synthetic version of exendin-4 and a GLP-1 analog which acts as a glucose-dependent insulin secretagogue (3); the sulfonylureas, a class of glucose independent insulin secretagogues; metformin (Glucophage), which regulates hepatic glucose output through poorly understood mechanisms; thiazolidinedione (TZD), such as rosiglitazone (Avandia), an insulin sensitizer acting as an agonist of PPAR-␥ in adipocytes; and acarbose, which prevents the digestion of complex carbohydrates. While these are effective, many side effects such as weight gain, hypoglycemia, and pancreatitis have been reported for these treatments and new therapies are desirable. To address this need current drug discovery efforts are targeting mechanisms of both early- and late-stage disease, including food intake, gastric emptying, insulin secretion, insulin sensitivity, hepatic
T2DM and drug discovery in mice
137
glucose output, and approaches to protect and increase beta-cell mass (4). This chapter will present methods that are widely used for screening and validating new therapeutic drugs using in vivo mouse models. These models can roughly be divided in two types: first, genetic models where animals develop symptoms of diabetes, even in absence of environmental changes, and second, diet-induced obesity, which results in increased insulin resistance (see Note 1). 1.1. Genetic Models
Some of the most widely used genetic models of type II diabetes are B6.V-Lepob /J, B6.Cg-m +/+ Leprdb /J, and BKS.Cg-m +/+ Leprdb /J. These strains carry single gene spontaneous mutations in either the leptin (Lepob ) or the leptin receptor (Leprdb ) genes in an inbred C57BL/6J or a C57BLKS/J background. Mice homozygous for the Lepob mutation are hyperphagic, gain weight becoming rapidly obese, and are hyperglycemic at a young age. In a C57BL/6J background hyperglycemia is only transient. Baseline glucose decreases and insulin increases steadily. Mice become normoglycemic, yet hyperinsulinemic by 14–16 weeks of age, with the disease progressing similarly in males and females. In comparison with the C57BL/6J background, the time course and symptoms on a C57BLKS/J background are more severe, with mice becoming severely diabetic by 6 weeks of age, suffering pancreatic islet degeneration and renal complications, resulting in lethality sometimes seen as early as 16–20 weeks of age. Because of this, the use of these mutant strains on the C57BL/6J background is particularly useful to mimic early stages of T2DM. The same mutations in the C57BLKS/J background are more useful to study the effects of therapeutics on advanced stages of the disease. While they are widely used, a concern with these genetic models is that mutations in the leptin gene and its receptor are rare occurrences in humans. Moreover, leptin administration in humans is ineffective in treating T2DM, except for a rare population of patients with mutations in their leptin and leptin receptor genes. In fact, diabetic patients develop hyperleptinemia and leptin resistance. For those reasons, these mutant mouse strains, while convenient, may have shortcomings in modeling all physiological aspects of T2DM in humans.
1.2. Diet-Induced Obesity
Obesity is strongly associated with type II diabetes, as a highfat diet is a major cause of insulin resistance both in humans and in several mouse strains. AKR/J, DBA/2J, and BTBR T+ tf/J strains are very responsive to high-fat diets, C57BL/6J can be considered a strain of intermediate susceptibility, while A/J and Balb/cJ mice are diet-induced “diabetes-resistant” (5–7). For reasons that are still unclear, males develop more insulin
138
Baribault
resistance in response to the diet-induced obesity (DIO) regimen (8). C57BL/6J is one of the most commonly used mouse strains for diabetes studies. Diet-induced obesity in this strain causes hyperglycemia, hyperinsulinemia, and the development of a fatty liver. One shortcoming of this model, however, is that diet alone is insufficient for the symptoms to progress to a later stage disease, such as beta-cell degeneration and diabetic nephropathy. To mimic late-stage T2DM symptoms observed in humans, DIO-C57BL/6J mice can be treated with low doses of streptozotocin (STZ). While mice treated with high doses of STZ are considered a model of type I diabetes, when the drug is administered at low doses to DIO mice, it mimics the partial loss of islet cells in the advanced stages of T2DM (9). In recent years, recombinant mouse strains have also been developed that model advanced stages of T2DM, such as NONcNZO10/LtJ mice (10).
2. Material 2.1. Mice
1. C57BL/6J males, 4–6 weeks old (The Jackson Laboratory, stock 000664) (see Notes 2 and 3). 2. Inventoried (DIO) C57BL/6J males, fed for 12 weeks with D12492i, 60 kcal% fat diet, (Research Diets, Inc.) (Jackson Laboratory, stock 000664) (see Note 4). 3. BKS.Cg-m +/+ Leprdb /J males, 3–4 weeks old, (The Jackson Laboratory, stock 000642), (see Note 5). 4. B6.V-Lepob/J, 3–4 weeks old males (The Jackson Laboratory stock 000632), (see Note 6).
2.2. Feeding and Dosing
1. 10 kcal% fat diet (standard diet), (Research Diets, D12450B) 2. 45 kcal% fat diet (Research Diets, D12451i) (see Note 7) 3. 60 kcal% fat diet (Research Diets, D12492i) (see Note 7) 4. 60 kcal% fat diet supplemented with rosiglitazone R (Avandia ; GlaxoSmithKline), custom-order (Research Diets) (see Note 8) 5. Vehicles 5.1. Dulbecco’s phosphate-buffered saline (D-PBS), 1× without calcium and magnesium (Mediatech, 21-031CM) 5.2. Sodium chloride injection solution, saline: NaCl 0.9%, 250 mL, Baxter IV solutions (VWR, 68000-342)
T2DM and drug discovery in mice
139
5.3. 1% hydroxypropyl methylcellulose (Alfa Caesar, 900465-3). Mix overnight using a magnetic stirrer R ; GlaxoSmithKline). To prepare a 6. Rosiglitazone (Avandia 1 mg/kg solution, put 50 mg rosiglitazone in a mortar and add 500 l Tween-80 (Alfa Caesar, 9005-65-6). Mix using the pestle. Add 50 ml 1% methylcellulose and mix with a 10 ml pipette. The solution may remain cloudy. The ◦ solution can be stored at 4 C for the duration of the experiment. Mix well before dosing. Dose per oral gavage (p.o.) at 5 mg/kg (100 l for a mouse of 20 g)
7. Metformin, 1,1-dimethylbiguanide hydrochloride (Sigma, D5035). Prepare a 200 mg/10 ml metformin solution in D-PBS. Dose via intraperitoneal (i.p.) injection at 100 mg/kg (100 l for a mouse of 20 g) 8. Exendin-4 (California Peptide Research, 507-77). Prepare a 1 mg/ml solution in D-PBS. Dilute in D-PBS 100-fold for dosing via i.p. injection at 10 g/kg (200 l for a mouse of 20 g) 9. Disposable sterile animal feeding needles, Popper & Sons, 20 G ×1 1/2 in. (VWR, 20068-666) 10. Tuberculin syringe with 27 G × 1/2 in. needle (BectonDickinson, 309623) 2.3. Glucose Tolerance, Insulin Tolerance, and Glucose-Stimulated Insulin Secretion Tests
1. Heated hard pad (Hallowell EMC, 000A2787A) 2. T/Pump Gaymar (Hallowell EMC, 000A3458) 3. Micro-hematocrit centrifuge: AutocritTM Ultra-3 (Becton Dickinson, 420575) 4. Clean cages with water bottles 5.
D -(+)-glucose
99.5% (Sigma, G7528)
6. Human insulin: (NovoNordisk)
Novolin
R,
100
U/ml,
10ml
7. Tuberculin syringe with 27 G × 1/2 in. needle, BectonDickinson, cat. nr. 309623 8. Scalpel 9. Mineral oil, Sigma, cat. nr. M3516 10. Paper towels 11. Accu-chek active glucometer, Roche Diagnostics (Battery: 3-volt lithium type CR 2032) 12. Accu-chek active test strips (Roche Diagnostics, 3146332) 13. Batteries, Eveready(R) button cells, ECR2032, 3 V Lithium cell (Sigma, B0653) 14. SurePrepTM capillary tubes, 75 mm, self-sealing, Becton Dickinson, cat. nr. 420315
140
Baribault
15. MicroWell 96-well polystyrene plates, round bottom (nontreated), sterile (Nunc, Sigma, P4241) 16. Miniature file set: in-vinyl pouch, with eight files of different profiles (Fine Science Tools, 30076-14) 17. 20–200 l pipette with tips 18. Alumina-Seal (Diversified Biotech, ALUM-1000) 2.4. Insulin ELISA
1. Insulin (Mouse) Ultrasensitive (Alpco Diagnostics, EIA 80INSMSU-E10) 2. Titer Plate Shaker (Lab-line instruments, 5246) 3. EL406TM Combination Washer Dispenser (BioTek Instruments) 4. Microplate Spectrophotometer, SpectraMax Plus 384 (Molecular Devices)
2.5. NEFA
1. HR Series NEFA-HR(2) kit (Wako Chemicals GmbH, 99475409), which include the following: 1.1. Color reagent A (999-346691) 1.2. Solvent A (995-34791) 1.3. Color reagent B (991-34891) 1.4. Solvent B (993-35191) 1.5. NEFA standard (oleic acid) solution (276-76491) 1.6. Control sera (410-00101 and 416-00202) 2. Clear flat bottom 96-well plates 3. Polyester films for ELISA and incubation (VWR, 60941120)
2.6. Body Composition
1. PIXImusTM densitometer (GE Medical Systems, LUNAR, Madison WI) 2. Lunar PIXImus trays (GE Medical Systems, 30950) 3. Anesthetic solution: mix ketamine (Ketaset, 100 mg/ml NADA # 45-290, Fort Dodge) with xylazine (AnaSed, 20 mg/ml, NADA # 139-236, Lloyd Laboratories) and water for a final concentration of ketamine:xylazine 10:1 mg/ml. 4. EchoMRI-100TM (Echo Medical System, Houston, TX)
2.7. Terminal Blood Collection and Tissue Collection for Histology
1. One cc syringes with 25 G × 5/8 needle, Becton-Dickinson, cat. nr. 305122 2. Microtainer serum separator tubes (BD Diagnostics, 365956) 3. Nunc MicroWell 96-well polystyrene plates, Sigma, P4241, round bottom (nontreated), sterile
T2DM and drug discovery in mice
4. Formalin solution, HT501128
neutral
buffered,
10%,
141
Sigma,
5. 50 mL conical tubes, VWR, cat. nr. 21008-178 6. Standard dissection tools: scissors and forceps
3. Methods Circulating glucose and insulin levels are the key values for a diagnosis of type II diabetes. Obesity and elevated levels of nonesterified fatty acids (NEFA) are known to cause insulin resistance and diabetes. Comorbidity of T2DM and dyslipidemia are common in animal models and in clinical populations and therefore, cholesterol, triglycerides, inflammation markers, and blood pressure are often measured within the same experiments. However, for the purpose of this chapter, we will cover only values directly linked to T2DM. Although the protocols for the genetic and DIO models are largely overlapping, they have a number of important differences: e.g., a high-fat diet is required for C57BL/6J mice to develop symptoms of T2DM, while genetic models develop the disease spontaneously. When differences apply, they will be detailed below. Otherwise, the following protocols apply to both types of models. 3.1. Experimental Design
1. Choose a positive control. These need to be chosen depending on the mechanism of action being investigated: for example, 1.1. Insulin secretagogues: exendin-4 is a long acting homolog of GLP-1, which induces insulin secretion and thereby lowers glucose. 1.2. Insulin resistance: rosiglitazone is an insulin sensitizer that targets PPAR-␥. 1.3. Hepatic glucose output: metformin is an orphan drug which lowers blood glucose by interfering with hepatic glucose output. Its mechanism of action is poorly understood. 2. Choose a non-toxic vehicle for the test therapeutic molecule(s). It may be different from the vehicle used for in vitro experiments (see Note 9). 3. Choose a negative control. This will be the vehicle used to deliver the positive control and test compounds: e.g., exendin-4 is soluble in D-PBS, therefore, one group of mice
142
Baribault
injected with D-PBS alone will be included as a negative control. 4. Choose a range of doses to be injected to determine a pharmacokinetic/pharmacodynamic (PK/PD) relationship. Six different concentrations per compound should be used to establish an ED50 value. For protein therapeutics, this may range from 0.1 ng/kg to 10 g/kg. For small molecules this can range from 1 ng/kg to 10 mg/kg. 5. Use 10–15 mice per group to obtain statistically significant results. For genetic models, where 80% of the mice will be used (see pre-selection in Section 3.3), adjust the number of mice accordingly. For example, 100 B6.V-Lepob /J mice will be sufficient to test one compound at six doses that include a positive and a negative control, using 10 mice per group. 6. Choose the injection schedule. Some compounds, such as rosiglitazone, are effective at lowering glucose levels after multiple injections over several days, but not in an acute, 1-day injection setting. Other compounds such as exendin-4 are active in an acute setting. When studying novel compounds, both types of study – acute and chronic dosing – should be performed to determine the properties of the molecule tested. The schedule of injections will depend on the pharmacokinetic properties of the test compounds. 7. Choose a mode of administration. Therapeutic compounds can be administered intraperitoneally (i.p.), orally (p.o.), subcutaneously (s.c.), or intravenously (i.v). Prolonged methods of administration also include osmotic pumps (e.g., Alzet mini-pumps). Alternatively, transgenic delivery of nucleic acids is useful for proof of concept studies: e.g., germline transgenic mice, viral expression, or systemic delivery of naked DNA via tail vein injection. The optimal mode of administration for a given compound will depend on its solubility and on whether the therapeutic is a small molecule, a protein, siRNA, or DNA. For example, rosiglitazone is not water-soluble and its best modes of administration are via oral gavage or mixed in the diet (see Note 10). 8. Choose method and schedule for blood collection. Common survival blood collection methods for metabolic profiling are tail-nick, tail snip, saphenous vein, submandibular (cheek), and retroorbital bleeding. With the tail nick or tail snip blood collection methods, 75 l samples of blood can be collected up to four times in a 1-day experiment, not exceeding a total of 250 l, and are used as default blood collection procedures. This usually provides up to 30 l serum samples and is sufficient to measure glucose, insulin, and other metabolic markers. In some cases, larger
T2DM and drug discovery in mice
143
volumes are needed (up to 250 l). Then, consider alternatives such as saphenous or submandibular vein collection or retroorbital bleeding (see Note 11). For terminal blood collection, exsanguination via cardiac puncture or via the vena cava and decapitation are commonly used. It is worth noting that CO2 asphyxiation and many anesthesia methods can interfere significantly with baseline glucose and insulin measurements. 3.2. Animal Housing
1. Keep mice housed under standard specified pathogen-free conditions with food and water ad libitum and a 12:12 dark:light cycle. 2. Keep mice group-housed (see Note 12). 3. For C57BL/6 J, feed mice with 60% fat diet. For genetic models, B6.V-Lepob/J and BKS.Cg-m +/+ Leprdb /J, feed mice with standard diet, D12450B. 4. Perform daily assessment of the bedding conditions. While this is a standard procedure in most vivaria, some strains of diabetic mice such as BKS.Cg-m +/+ Leprdb /J mice need additional attention because of related kidney problems, resulting in polyuria. Even with a HEPA-filtered ventilated caging system, mice caging may need to be changed two to three times weekly as the diabetic condition worsens. 5. Identify individual mice with ear notches several days before performing an experiment as the procedure may have a transient effect on glucose levels.
3.3. Randomization and Pre-selection of Mouse Cohorts
Pre-selection of mouse cohorts based on their baseline glucose and body weight is particularly important for genetic models as baseline glucose can span a range of several hundred units even for mice of the same gender and age: e.g., 200–600 mg/dL (see Note 13 and Fig. 9.1). 1. Measure mouse body weight. 2. Measure baseline glucose: insert glucometer strip into glucometer. Using gloves, pick up mouse by the tail. Wrap its body in a paper towel restraining it loosely with one hand, leaving the tail exposed. With the other hand, cut approximately 1 mm off the tail tip with a scalpel or perform a nick of the tail vein on the side of the tail close to the tip. Put one drop of blood onto a glucometer strip and take reading. 3. Sort mice by body weight values. Exclude mice with outlier body weight values. 4. Sort glucose values in increasing order. 5. Select mice with mid-range values.
144
Baribault 500
baseline glucose
400 300 200 100 0
Fig. 9.1. Selection of mice based on baseline glucose value. Baseline glucose was measured on 100 mice and sorted in increasing order. Each bar represents the glucose value for one mouse. Values ranged from 115 to 430 mg/dl. Mice with glucose values higher than 275 and lower than 140 were excluded (mice over the black rectangles). Selection of 80 mice in the middle range was done in a semi-quantitative manner (mice at either end of the glucose value range that seem to deviate from the mean the most substantially were discarded) and used for randomization (mice over the open/white rectangle).
6. Distribute mice of similar baseline glucose values evenly among experimental groups. For example, for three groups, assign the group to mice in sequential order: e.g., A, B, C, C, B, A, A, B, C, C, B, A. 3.4. Tail Snip Blood Collection for Multiple Measurements: Baseline Glucose, Insulin, and NEFA
1. Insert glucometer strip into glucometer. 2. Using gloves, pick up mouse by the tail and cut approximately 1 mm or less off the tail tip or perform a nick of the tail vein on the side of the tail close to the tip as before. If the tail tip was previously cut, remove the scab off the tail tip with the scalpel. 3. Put one drop of blood on glucometer strip and take reading. 4. Put a minute amount of mineral oil on index finger and thumb. 5. Rub the sides of the tail on each side with index finger and thumb from the base toward the tail tip. Collect the blood into a capillary tube. Repeat this step until the tube is full. 6. Spin the capillary tubes in a micro-hematocrit centrifuge at 10,000 rpm for 8 min. 7. Cut the capillary tubes at the separation between serum and red blood cells with the sharp edge of a file. 8. Use a pipette with a 200 l tip to gently blow the serum out of the capillary tube into a 96-well plate.
3.5. Insulin ELISA
1. Reagent preparation: dilute the enzyme conjugate concentrate (11×) with 10 parts of enzyme conjugate buffer. Dilute insulin controls with 0.6 ml of distilled water. Dilute wash buffer concentrate with 20 parts distilled water.
T2DM and drug discovery in mice
145
2. Designate wells for standards, controls, and unknown samples. Pipette 5 l of each in their respective wells. 3. Add 75 l of enzyme conjugate. Seal the plate with the polyester film provided. Place the plate on an orbital microplate shaker, shaking at 700–900 rpm for 2 h at room temperature. 4. Remove sealing film and wash the plate six times with wash buffer with a microplate washer. 5. Add 100 l of substrate to each well. Reseal the plate with a polyester film and incubate for 30 min on an orbital microplate shaker at room temperature. 6. Remove sealing film and add 100 l of stop solution to each well. Gently mix to stop the reaction, remove bubbles before reading with the microplate reader. 7. Program the location of the standards, controls, and unknown samples into an absorption plate reader, with the absorbance at 450 nm with a reference wavelength of 650 nm. Read the plate within 30 min following the addition of the stop solution (see Note 14). 3.6. NEFA (Non-esterified Free Fatty Acids)
Free fatty acids are elevated in the plasma of obese patients and are known to cause muscle and liver insulin resistance. The Wako HR series NEFA-HR(2) is an in vitro enzymatic colorimetric method assay for the quantitative determination of non-esterified fatty acids (NEFA) in serum. Perform the assay on serum collected from mice fasted for a period greater than 4 h, but less than 16 h. Perform the test on samples immediately after collection, without freezing. Also note that hemolysis in the serum samples may interfere with the assay. 1. Prepare reagent solutions: add 10 ml solvent A into one vial of color reagent A and mix gently. Add 20 ml solvent B into one vial of color reagent B and mix gently. Solvent A and B solutions are stable at 2–10ºC for 5 days. 2. Prepare standard solutions: stock solution is 1 mEq/L. The test is linear from 0.01 to 4.00 mEq/L. Carry out a serial dilution 1:1 of the standard with water to obtain concentrations of 0.5, 0.25, 0.125 mEq/L. 3. Add samples to 96-well plates in duplicates. Add 10 l of standard stock solution in two of the wells to obtain a reading of 2 mEq/L. Then add 5 l of the 1, 0.5, 0.25, and 0.125 mEq/L of the standard solutions in subsequent wells, respectively. Add 5 l of water in two wells to serve as blanks. Add 5 l of two control samples. Add 5 l of unknown samples. If values greater than 2 mEq/L are expected, dilute the samples in PBS.
146
Baribault
4. Add 50 l of reagent A solution to each well. Add polyester film to seal the plate. Mix and incubate at 37ºC for 10 min. 5. Add 100 l of reagent B solution to each well. Mix and incubate at 37ºC for 10 min. Seal the plate with a polyester film. Mix and incubate at 37ºC for 10 min. 6. Program the location of the standards, controls, and unknown samples into an absorption plate reader. Set the wavelength to 550 nm. Read the plate. 3.7. Glucose Tolerance Test (GTT)
Glucose tolerance tests can be conducted on mice fed ad libitum or following a fasting period. Because food intake in mice occurs mainly during the dark period of the dark-light cycle, we prefer to fast mice during daytime at the beginning of the light period, from 6:00 AM to 10:00 AM. This more closely mimics an overnight fast in humans. Fasting mice also reduces the range of baseline glucose readings and can reveal significant differences between experimental groups that would not reach significance in non-fasted animals. Glucose can be administered by intraperitoneal injection (i.p.) or by oral gavage (p.o.). Injection is usually faster than gavage and allows for handling a larger number of mice; however, it bypasses a potential effect from stimulating incretins in enteroendocrine cells. 1. To initiate a fasting period transfer mice to clean cages without food but with water ad libitum of 4 h. 2. Weigh mice. If mice are group-housed, draw bars on tails with a black marker pen for quick identification during the test: e.g., zero, one, two, three, or four bars, respectively. Marks on the tail can be read faster than ear notches. These marks will fade after a few days. 3. Prepare a 10% (w/v) glucose solution in distilled water. Glucose will be administered at a concentration of 1 g/kg (glucose/body weight) in a 10 ml/kg volume: e.g., 250 l for a mouse of 25 g (see Note 15). 4. Preload 1 ml syringes with the glucose solution. Ensure that all air bubbles have been removed by tapping on the side of the syringes and expressing the air. 5. Shortly before the 4-h fasting is complete, take a measurement of the baseline glucose level. Insert a glucometer strip into a glucometer. Wrap the mouse in a paper towel and restrain loosely with one hand. With the other hand, cut approximately 1 mm off the tail tip with a scalpel or perform a nick of the tail vein on the side of the tail close to the tip. Put one drop of blood on glucometer strip and take reading. Record glucose value. The range of readable
T2DM and drug discovery in mice
147
glucose values is 20–600. If the glucometer indicates the value “Hi”, i.e., off scale, use 600 as a value (see Note 16). 6. Upon completion of the 4 h fast, inject mice with glucose, intraperitoneally, or by oral gavage. Up to 40 mice can be injected i.p. per person during a 15 min period, or 20 by oral gavage. 7. Repeat blood glucose measurements at 15, 30, 60, and 90 min. 8. Add food to mouse cages. 3.8. Glucose-Stimulated Insulin Secretion (GSIS)
Glucose-stimulated insulin secretion (GSIS) is central to normal control of metabolic fuel homeostasis, and its impairment is a key factor in beta-cell failure in T2DM. Some targets may show a phenotype only with GSIS and not in a GTT (11). GSIS can be performed along with a GTT. The peak of insulin secretion occurs approximately at 7–8 min after glucose injection. Blood is typically collected at 7.5, 15, and 30 min after glucose administration. GSIS procedure is more time consuming than a standard GTT and fewer mice can be handled/person. For example, five to eight mice can be handled per person for a GSIS experiment (one mouse per 1–2 min) compared to 40 mice per person in a standard GTT experiment (one mouse per 20–25 sec). To obtain statistical significance data, it may be necessary to stagger multiple groups of mice (see Note 17). 1. Prepare glucose, weigh mice, and transfer to clean cages. 2. Put cages on heating pads at 37◦ C. Keeping the mice warm will help blood flow for multiple blood sampling. 3. Put a drop of mineral oil on index finger and thumb. 4. Rub the sides of the tail on each side with index finger and thumb from the base toward the tail tip. Collect the blood into a capillary tube. Repeat this step until the tube is halffull (approximately 40 l). 5. Administer glucose to mice (1 mg/kg in 10 ml/kg volume), t =0. 6. At 7.5, 15, 30, and 60 min after glucose injection, repeat the blood collection procedure. 7. Spin the capillary tubes into a micro-hematocrit centrifuge at 10,000 rpm for 8 min. 8. Cut the capillary tubes at the separation between serum and red blood cells with the sharp edge of a file. 9. Use a pipette with a 200 l tip to gently blow the serum out of the capillary tube into a 96-well plate. 10. Samples may be frozen at −20◦ C for later measurements. Cover the plates with Alumi-Seal.
148
Baribault
11. Measure levels of insulin according to the manufacturer’s protocol. 3.9. Insulin Tolerance Test (ITT)
Insulin sensitivity can be measured directly by injecting insulin in mice and measuring its effect on circulating glucose levels. The dose of insulin to be administered ranges from 1 to 5 U/kg and is adjusted depending on the model system used. For example, lean C57BL/6J mice will have their baseline glucose reduced substantially at 1 U/kg, while mice fed on a high-fat diet will often need to be treated with 1.5 or 2.0 U/kg, and ob/ob mice will sustain up to 5U/kg. 1. Provide access to food and water ad libitum (see Note 18). 2. Weigh mice. 3. Calculate injection volumes using a volume of 10 ml/kg. 4. Dilute insulin in saline to the desired concentration: 1–5 U/10 ml. Keep on ice during preparation. Mix gently and preload syringes. 5. Measure baseline glucose by tail tip or tail snip blood collection, according to procedure described above. 6. Inject mice i.p. with insulin. 7. Measure glucose 15, 30, 60, 90, and 120 min after insulin injection. If the glucometer indicates “Lo,” mice should be injected immediately with glucose to prevent loss of consciousness due to hypoglycemia.
3.10. Insulin Sensitivity
Measuring glucose alone is not sufficient to evaluate the state of sugar metabolism. For example, as ob/ob mice age, they become normoglycemic yet they are severely insulin resistant. Further, their pancreatic islets become hypertrophic, and the levels of circulating insulin exceed 10 ng/mL, thereby compensating for the insulin resistance. The “gold standard” for calculating an insulin sensitivity index is the hyperinsulinemic–euglycemic clamp. It measures the amount of glucose necessary to compensate for an increased insulin level without causing hypoglycemia. This method involves the catheterization of carotid arteries and jugular veins in rodents. It is particularly challenging in mice because of their small size and is not amenable to large-scale studies. Consequently, alternatives that correlate closely to clamp results are often used in determining insulin sensitivity indices. The direct measurement of insulin tolerance, insulin tolerance test (ITT), is often used. Alternatively, simple surrogate indexes for insulin sensitivity/resistance are available (e.g., QUICKI, HOMA, 1/insulin, Matusda index) that are derived from blood insulin and glucose concentrations under fasting conditions (steady state) or after an oral glucose load (dynamic).
T2DM and drug discovery in mice
149
The homeostasis model assessment-insulin resistance (HOMA-IR) can be calculated with the following formula: HOMA-IR index = fasting glucose (mmol/L) × fasting insulin (mU/L)/22.5 (12) 3.11. Body Composition Measurements
3.11.1. Magnetic Resonance Imaging
Chemical carcass analysis is considered the “gold standard” for accurate whole body composition analysis (13). It is, however, terminal and time consuming. The adiposity index can also be measured by dissecting and weighing of fat depots in individual animals (14). This method is also terminal and less accurate. The collection of visceral fat required can be particularly challenging as it is often spread throughout internal organs. Two imaging systems, dual energy X-ray absorptiometry scanning (DEXA) and magnetic resonance imaging (MRI), allow for longitudinal studies of whole body composition. DEXA measures bone mineral density and content, fat content, and lean content in anesthetized mice. Echo MRI from Echo Medical System, Houston, TX, is used to measure whole body composition parameters such as total body fat, lean mass, body fluids, and total body water in live mice without the need for anesthesia or sedation (15). The MRI technology is more rapid, less than a minute to scan one mouse, than DEXA which takes about 5 min per mouse. 1. Prior to each run, calibrate the system using a calibrated standard provided by Echo Medical System. 2. Record mouse body weights. 3. Place each mouse into an appropriate size tube and place into the MRI machine. 4. EchoMRI software records spectra on each mouse. 5. The output information is expressed as lean tissue mass, fat mass, and free body fluids in grams.
3.11.2. DEXA
1. Measure the animals body weight 2. Run a quality control cycle using a phantom QC provided by the manufacturer on a daily basis. 3. Administer 100 mg:10 mg/kg ketamine:xylazine anesthetic solution i.p. (10 ml/kg). Wait until effective, i.e., no reflex response to a toe pinch. 4. Lift the label off the tray. Set the mouse on the mild adhesive and take measurements 5. Record bone mineral density (BMD), bone mineral content (BMC), lean mass, fat mass, and percentage body fat.
3.11.3. Terminal Blood and Tissue Collection
Terminal blood collection is often performed upon completion of an experiment, as it is useful to perform measurements of additional metabolic markers on cohorts that showed compound
150
Baribault
efficacy. Immediately following blood collection tissue collection can also be performed at the same time from the same animal. The cytoarchitecture of fat, liver, pancreas, and kidney are seriously affected by T2DM. As mentioned in Section 1, the pancreatic islets of Langerhans become hypertrophic in the early stages of the disease and subsequently undergo degeneration. Adipocytes in fat tissues become hypertrophic. Fatty livers can also be observed by gross morphology and the hepatocytes show accumulation of lipids at the histological level. Finally, in the later stages of the disease, glucotoxicity induces irreversible damage to kidney tubules. Consequently, those are the four main tissues to be examined in histology. 1. Prepare the syringes: break the vacuum seal of 1 ml syringes by pulling the plunger slightly and push back. Fit the syringes with 25 G × 5/8 inch needles. 2. Before beginning the terminal collection procedures have all serum separation tubes for blood collection labeled and all tubes for tissue collection labeled and filled with 5–10 ml 10% formalin (3.7% formaldehyde). Organs from the same mouse can be pooled in the same tube as they can simply be embedded in a single block for histological analysis. 3. Euthanize the animal by CO2 asphyxiation. 4. Put the animal on a paper towel on its back and palpate the sternum with a finger of one hand. 5. With the other hand, insert a needle slightly below and to the left of the sternum, with a mild inclination, aiming into the left cardiac ventricle (see Note 19). 6. Draw blood slowly so as not to collapse the ventricle. Usually 0.3–1.0 ml can be obtained in 10–30 sec. Once the blood starts flowing in, the depth and the angle of the needle may need to be slightly adjusted to maintain the flow of blood into the needle. 7. To avoid hemolysis remove the needle from the syringe and push the blood into a serum separation tube. Do not exceed the volume indicator on the tube (frosted area). 8. Centrifuge for 10 min at 10,000 rpm in an Eppendorf centrifuge at 4◦ C. 9. Transfer the supernatant (serum) to a clean tube (see Note 20). 10. Make an incision through the skin and peritoneal membrane to open the abdomen. 11. Cut a lobe of the liver and put in formalin. 12. Collect examples of visceral fat, e.g., the adipose tissue surrounding the intestine.
T2DM and drug discovery in mice
151
13. Pull out the spleen and the pancreas which is loosely attached to the spleen and cut out both. Put both organs on a paper towel. Separate the pancreas and put in formalin. 14. Collect one kidney and add to the formalin tube. 15. Proceed using standard histology procedures for paraffin embedding and sectioning (see Note 21). 16. Stain sections with hematoxylin/eosin (see Note 22).
4. Notes 1. The methods presented focus on in vivo pharmacology as it applies for drug discovery. Many more models of T2DM are used for target discovery and elucidating the mode of action of such potential targets but are not presented here including isolation and in vitro primary cultures of pancreatic islets, energy expenditure measurements with the use of metabolic chambers, gastric bypass and production of genetically engineered gain of function, and loss of functions mutants. 2. While the progression of T2DM symptoms is slower in C57BL/6J than in other strains more susceptible to diabetes, it is still widely used for in vivo pharmacology. In part this is because C57BL/6J mice have been used and characterized extensively for their response to known therapeutics. Moreover, many genetically engineered mutant mice are on a C57BL/6J background and are often utilized for off-target effect studies of novel compounds. 3. For reasons that are still unclear, C57BL/6J females do not respond as strongly to a high-fat diet as males do. 4. Insulin resistance increases continuously over 20 weeks of feeding a high-fat diet. Therefore, the period of high-fat diet feeding is a compromise between the time efficiency and the window necessary to see an effect. 5. Glucose increases rapidly in these mice and reaches 600 mg/dl by 6–8 weeks of age. If the availability of BKS.Cg-m +/+ Leprdb /J males is limited, females of the same age can be used. KK.Cg-Ay/J mice (stock 002468) may also be used. 6. B6.V-Lepob/J mice become normoglycemic as they age because of compensatory hyperinsulinemia. We have used these mice successfully up to 12 weeks for pharmacological studies.
152
Baribault
7. The most commonly used diets are 45 kcal% fat diet (Research Diets, D12451i) and 60 kcal% fat diet (Research Diets, D12492i). The “i” in “D124xxi” refers to the irradiated form of the diet. Many vivaria use the irradiated form for sterility purposes, but this is not required. Sixty percent fat diet induces a more severe insulin resistance. However, the high-fat pellets crumble more easily, falling in the bedding and make it more difficult to perform accurate food intake measurements. Also, under these regimes more skin lesions can be observed, presumably because of increased subcutaneous fat, greasy fur from the diet, and fighting. Use of high-fructose/sucrose diets are not commonly used in mice. While a high-carbohydrate diet is efficient at inducing insulin resistance in rats, effects vary strongly between mouse strains. It has little effect on C57BL/6J mice (cf. Research Diets web site). 8. Food intake can vary substantially by strain and diet within a range of 2–6 g per mouse daily. Therefore the concentration of rosiglitazone in the food may need to be adjusted. Formulation is custom made by the vendor. 9. Most peptides and proteins are water-soluble, and therefore D-PBS or saline can be used as a vehicle. Often, however, limited information is available about the solubility properties of novel small molecules, and the choice of a non-toxic vehicle is more difficult. For example, 200 l of a 5% ethanol solution is equivalent to one beer in humans and may affect behavior. A solution of 20% cyclodextrin has no known side effects in vivo, but in rare cases, some compounds are trapped in the solution and therefore mice have no exposure to the compound. Some vehicles used for in vitro studies can be toxic in live mice. Some vehicles such as methylcellulose have no side effect when given p.o., but are toxic if administered i.v. Access to information about the pharmacokinetics properties of a test compound can help in the choice of a vehicle. 10. When mixing a compound to the diet, pilot experiments should be conducted to verify that food intake is not affected by the change of palatability in the diet. This is counterintuitive, as a compound reducing food intake might be considered a benefit for a metabolic disease. However, unspecific effects on metabolism due to diet unpatability should be tested and excluded early on. 11. Retroorbital bleeding can cause blindness in animals and is increasingly discouraged by animal welfare committees. However, alternatives such as saphenous vein blood collection are cumbersome. Submandibular vein blood collection is rapid but causes glucose values to increase.
T2DM and drug discovery in mice
153
12. Mice can be either group-housed or single-housed. Singlyhoused mice also allow for food intake measurements at the individual level. Also mice fed with a high-fat diet are more prone to develop lasting skin lesions if they are injured from aggressive interactions. Therefore in many studies, mice are single-housed. On the other hand, mice are social animals and being single-housed increases their level of stress and decreases food intake. Additionally, when cage space is a restricted resource, group-housing may be preferable. In either case, singly- or group-housed, all mice in an experimental cohort should be housed in the same manner. 13. (a) In addition to stochastic individual differences two factors can contribute to widening the range of body weight and baseline serum glucose value. First, B6.V-Lepob /J, B6.Cg-m +/+ Leprdb /J, and BKS.Cg-m +/+ Leprdb /J mice are generated by mating heterozygous parents. Homozygous mutant mice are selected on the bases of their phenotype, i.e., increased body weight, and not on DNA analysis. Occasionally, although infrequent (less than 1%), stochastic events can lead to the selection of non-homozygous animals in the experimental cohort delivered, when the sorting is performed at a very young age. In addition, the date of birth used for shipment is usually a “bin” of the stated date of birth, and if often uses animals born 3–4 days before or after the stated date. On a 3-week-old mouse this can lead to significant weight differences. Quick progression of the disease state in these strains can contribute significantly to widening the range of body weight and baseline serum glucose value. (b) Limits are chosen somewhat arbitrarily to eliminate mice that deviate the most from the average values, while keeping enough mice for cohort size providing statistically significant values. 14. Values may differ slightly depending on the settings chosen in the template used for reading. The test is designed to use a “Blank” value rather than using “standard values = 0” in the template for the wells corresponding to the “zero standard.” The test is also designed to provide a linear correlation of absorbance and concentration when using log–log scales. 15. Higher glucose doses can be used (e.g., 2 g/kg) can be used. However, if mice are highly hyperglycemic and insulin resistant, many of the glucose values collected during a GTT will exceed the detection limit of the glucometer, 600 mg/dl. 16. Alternative glucose measurement methods are available. However, their relevance to glucose metabolism is of limited value considering that when glucose levels exceed
154
Baribault
600 mg/dl, glucose is eliminated through urine rather than by glucose uptake in peripheral tissues. 17. As soon as the blood has been collected for the 30 min time point in the first group, the test can be performed on a second group. The second group will therefore undergo a fast of ∼4:45 h instead of 4 h. This is a small difference and is largely overridden by the injection of glucose. When using several groups, fast animals from all groups and measure baseline glucose for mice in all groups first, then perform the glucose injection in the mice from the first group. 18. GTT and ITT can be conducted either in fasted or nonfasted conditions. This will affect the baseline glucose values and their S.E.M., but has little effect on glucose values later in the tests in comparison to the effects of glucose and insulin. Fasting causes an increase in appetite signals and other molecules involved in metabolism. Depending on the test molecule tested, this might interfere with the drug target and study results. 19. Right and left are defined from the mouse perspective, i.e., “left” refers to the mouse’s left side. 20. The serum should be yellowish. A reddish color is indicative of erythrocytes lysis (hemolysis) which may interfere with clinical chemistry assays based on colorimetric values. If this happens, you may need to adjust the speed at which the blood is collected and processed or other steps that may cause sheer and red blood cell lysis. 21. All samples can be embedded in a single block for histological analysis as a multi-tissue block. Histology services are offered by many commercial providers, e.g., IDEXX, a veterinary service with locations worldwide. Hematoxylin– eosin is sufficient to reveal pancreatic islet hyperplasia or degeneration, pancreatitis, liver steatosis, adipocyte hyperplasia, and diabetic nephropathies. Additional staining can be requested, such as immunostaining with anti-insulin or anti-glucagon antibodies, if additional information is needed about drugs mechanisms of action.
5. Acknowledgments I am grateful to Jonitha Gardner, Laura Hoffman, Cheryl Loughery, Drs. Jiangwen Majeti, Alykhan Motani, and Wen-Chen Yeh for scientific discussions and critical review of the manuscript.
T2DM and drug discovery in mice
155
References 1. Peters, L. L., Robledo, R. F., Bult, C. J., Churchill, G. A., Paigen, B. J., et al. (2007) The mouse as a model for human biology: a resource guide for complex trait analysis. Nat Rev Genet 8, 58–69. 2. Saltiel, A. R. (2001) New perspectives into the molecular pathogenesis and treatment of type 2 diabetes. Cell 104, 517–529. 3. De Leon, D. D., Crutchlow, M. F., Ham, J. Y. and Stoffers, D. A. (2006) Role of glucagon-like peptide-1 in the pathogenesis and treatment of diabetes mellitus. Int J Biochem Cell Biol 38, 845–859. 4. Keller, M. P., Choi, Y., Wang, P., Belt Davis, D., Rabaglia, M. E., et al. (2008) A gene expression network model of type 2 diabetes links cell cycle regulation in islets with diabetes susceptibility. Genome Res 18, 706–716. 5. Naggert, J. K., Svenson, K. L., Smith, R. V., Paigen, B. and Peters, L. L. (2006) Diet effects on bone mineral density and content, body composition, and plasma glucose, leptin and insulin levels MPD:143. Mouse Phenome Database (http://phenome.jax.org/pubcgi/phenome/mpdcgi?rtn=projects/list), The Jackson Laboratory, Bar Harbor, Maine. 6. Alexander, J., Chang, G. Q., Dourmashkin, J. T. and Leibowitz, S. F. (2006) Distinct phenotypes of obesity-prone AKR/J, DBA2J and C57BL/6 J mice compared to control strains. Int J Obes (Lond) 30, 50–59. 7. Clee, S. M. and Attie, A. D. (2007) The genetic landscape of type 2 diabetes in mice. Endocr Rev 28, 48–83. 8. Nishikawa, S., Yasoshima, A., Doi, K., Nakayama, H. and Uetsuka, K. (2007)
9.
10.
11.
12.
13. 14.
15.
Involvement of sex, strain and age factors in high fat diet-induced obesity in C57BL/6 J and BALB/cA mice. Exp Anim 56, 263–272. Luo, J., Quan, J., Tsai, J., Hobensack, C. K., Sullivan, C., et al. (1998) Nongenetic mouse models of non-insulin-dependent diabetes mellitus. Metabolism 47, 663–668. Cho, Y. R., Kim, H. J., Park, S. Y., Ko, H. J., Hong, E. G., et al. (2007) Hyperglycemia, maturity-onset obesity, and insulin resistance in NONcNZO10/LtJ males, a new mouse model of type 2 diabetes. Am J Physiol Endocrinol Metab 293, E327–336. Kebede, M., Alquier, T., Latour, M. G., Semache, M., Tremblay, C., et al. (2008) The fatty acid receptor GPR40 plays a role in insulin secretion in vivo after high-fat feeding. Diabetes 57, 2432–2437. Buchner, D. A., Burrage, L. C., Hill, A. E., Yazbek, S. N., O’Brien, W. E., et al. (2008) Resistance to diet-induced obesity in mice with a single substituted chromosome. Physiol Genom 35, 116–122. Brommage, R. (2003) Validation and calibration of DEXA body composition in mice. Am J Physiol Endocrinol Metab 285, E454–459. Gregoire, F. M., Zhang, Q., Smith, S. J., Tong, C., Ross, D., et al. (2002) Dietinduced obesity and hepatic gene expression alterations in C57BL/6 J and ICAM-1deficient mice. Am J Physiol Endocrinol Metab 282, E703–713. Tinsley, F. C., Taicher, G. Z. and Heiman, M. L. (2004) Evaluation of a quantitative magnetic resonance method for mouse whole body composition analysis. Obes Res 12, 150–160.
Chapter 10 Cholesterol Absorption and Metabolism Philip N. Howles Abstract Inhibitors of cholesterol absorption have been sought for decades as a means to treat and prevent cardiovascular diseases associated with hypercholesterolemia. Ezetimibe is the one clear success story in this regard, and other compounds with similar efficacy continue to be sought. In the last decade, the laboratory mouse, with all its genetic power, has become the premier experimental model for discovering the mechanisms underlying cholesterol absorption and has become a critical tool for preclinical testing of potential pharmaceutical entities. This chapter briefly reviews the history of cholesterol absorption research and the various gene candidates that have come under consideration as drug targets. The most common and versatile method of measuring cholesterol absorption is described in detail along with important considerations when interpreting results, and an alternative method is also presented. In recent years, reverse cholesterol transport has become an area of intense new interest for drug discovery since this process is now considered another key to reducing cardiovascular disease risk. The ultimate measure of reverse cholesterol transport is sterol excretion and a detailed description is given for measuring neutral and acidic fecal sterols and interpreting the results. Key words: Cholesterol absorption, phytosterols, cholesterol excretion, reverse cholesterol transport, ACAT, CEL, inhibitors, bile acids, fecal sterols, dual label, obesity, cardiovascular disease.
1. Introduction Cardiovascular diseases (CVD) are the leading cause of death in the United States and other “Western” societies despite the great success of statin drugs for lowering LDL cholesterol. Thus, finding new targets for pharmaceutical intervention aimed at reducing the risk for CVD has remained a high priority and the focus of decades of research. Since the advent of gene targeting by homologous recombination, a host of mouse models have been G. Proetzel, M.V. Wiles (eds.), Mouse Models for Drug Discovery, Methods in Molecular Biology 602, DOI 10.1007/978-1-60761-058-8 10, © Humana Press, a part of Springer Science+Business Media, LLC 2010
157
158
Howles
generated and widely used for the study of cholesterol and lipoprotein metabolism with the goal of understanding the molecular physiology and etiology of CVD. The result has been identification of numerous proteins, enzymes, and metabolic pathways, other than HMG CoA reductase and the cholesterol synthesis pathway, that are potential drug targets for the treatment and prevention of atherosclerosis and CVD. 1.1. Cholesterol Absorption
In this light, cholesterol absorption has received intense focus for several decades. Although the various statins lower LDL by decreasing endogenous cholesterol synthesis, another approach to prevent excess cholesterol accumulation is to reduce absorption of dietary cholesterol. Doing so also prevents reabsorption of biliary cholesterol, which can have a major impact on overall cholesterol metabolism since recirculation of biliary cholesterol represents a large portion of the cholesterol that transits through the intestine. For recent reviews on mechanisms of cholesterol and lipid absorption, see ref. (1–3). The search for intestinal cholesterol transporters extended for many years, beginning with a debate about whether or not it was even a protein-facilitated process (4, 5). The pancreatic enzyme carboxyl ester lipase (CEL, also called cholesterol esterase) was believed to be important to this process (6, 7) and several companies devoted considerable resources to the development and testing of compounds to inhibit CEL, with mixed results (8–10). These efforts were abandoned in the mid-1990s, however, after studies with gene-knockout mice demonstrated that the enzyme was important only for absorption of cholesteryl ester (11, 12), which is a minor component of dietary cholesterol and is present at very low levels in bile. Interestingly, CEL is also found in liver where it has been shown to affect HDL metabolism (13). Thus, it may ultimately play an important role in cholesterol metabolism and may yet prove to be a useful drug target for CVD treatment (Camarota and Howles, unpublished). Intestinal acyl-CoA:cholesterol acyltransferase (ACAT-2, also present in liver), which esterifies free cholesterol with palmitic or oleic acid, is another enzyme that was identified early on as a potential target to inhibit cholesterol absorption because most cholesterol in chylomicrons is esterified before being secreted by enterocytes (6, 14). As for CEL, various inhibitors of this enzyme were also developed and tested with mixed results (10, 15–17). However, the importance of ACAT-2 was later confirmed by studies of gene-knockout mice, which exhibit markedly reduced cholesterol absorption and atherosclerosis when fed Western diet (18). Nonetheless, progress in developing effective ACAT inhibitors has been slow, in part because of concerns about the potential for deleterious systemic effects resulting from inhibition of the more widely expressed ACAT-1 (19). Despite these
Cholesterol Absorption and Metabolism
159
difficulties as well as mixed results in animal and clinical trials, interest remains for the development of ACAT inhibitors, especially ones that could be restricted to act on ACAT-2 (20–22). In 1997, Schering-Plough reported dramatic inhibition of cholesterol absorption by a compound, later named ezetimibe, that was developed in the process of modifying known ACAT inhibitors to generate novel compounds with improved pharmacodynamics and pharmacokinetics (23). Interestingly, while the drug blocked absorption by up to 90%, it was no longer an effective ACAT inhibitor (23). Later, this group used database mining and gene-knockout mice to identify that the protein essential for cholesterol absorption is NPC1L1 (Niemann–Pick disease C, type 1-like protein 1) and that this protein is the likely target of ezetimibe (24, 25). The development of this drug and identification of the putative target gene constituted major breakthroughs in the field, both for research focused on cholesterol and lipoprotein metabolism and for the treatment of hypercholesterolemia and CVD. Ezetimibe (ZetiaTM ) has become widely used clinically, especially for patients whose LDL cholesterol responds poorly to statins or who cannot tolerate the side effects of these drugs (26). As patent expiration looms for the various statins, major pharmaceutical companies are eagerly seeking new targets and new compounds for lowering plasma cholesterol and reducing CVD risk. Particularly appealing are compounds that would resemble ezetimibe in action and pharmacokinetics. Ezetimibe is glucuronidated by the intestine (and liver), which increases its potency for blocking cholesterol absorption, in part because the glucuronidated form has a high affinity for the intestinal brush border and is poorly reabsorbed after the first pass through the enterohepatic circulation (23). Thus, compounds with an intestinal site of action that are either not absorbed, or concentrate at the site of action, or have good efficacy at low circulating levels are desired because the risk of off-target effects is reduced (especially if not absorbed). In this regard, one promising target for drug development is the phospholipase-A2 secreted by pancreatic acinar cells (PLA2) and/or the phospholipase-B made by intestinal epithelial cells. It has been demonstrated that decreasing intestinal PLA2 activity with inhibitors or by gene ablation results in decreased cholesterol absorption (27, 28) and cell culture models have also shown that intact phospholipids inhibit cholesterol absorption, apparently by inhibiting micelle formation (29). Data from knockout mice also suggest that there may be other health benefits of inhibiting PLA2 in the intestine (28). Thus, this enzyme is an ideal target for decreasing CVD since unabsorbed inhibitors should be efficacious with minimal risk of systemic side effects. Studies with gene-knockout mice (30) as well as chemical inhibitors (31) have shown that decreasing activity of
160
Howles
pancreatic triglyceride lipase (PTL) also decreases cholesterol absorption. Data indicate that this most likely results from incomplete digestion and delayed absorption of triglycerides (30, 31). Lumenal cholesterol partitions into the oil phase and is carried distally in the small intestine where its absorption is less efficient because of fewer transporter molecules and/or because of reduced bile salt concentration (see model in 1). Orlistat (tetrahyR drolipstatin), currently marketed as alli , is a potent inhibitor of PTL and is sold as an anti-obesity treatment. While decreased cholesterol absorption is a beneficial side effect, targeting lipases is a poor strategy for doing so since steatorrhea can occur, which tends to reduce compliance (32). After postprandial absorption, lumenal cholesterol is incorporated into chylomicrons (large triglyceride-rich lipoproteins) by enterocytes and secreted into intestinal lymphatics. Secretions into this network collect in the thoracic lymph duct and finally reach the circulation where this duct opens into the left subclavian vein. As chylomicrons are pumped throughout the body, the dietary lipids are rapidly hydrolyzed and taken up by various tissues, as are the remnant particles, with a circulating half-life of only minutes in mice. For this reason, many in vivo investigations of cholesterol absorption and various inhibitor evaluations were performed using rats with cannulated lymph and bile ducts so as to assay intestinal output before it entered the circulation and was metabolized (4, 6, 7, 14, 16, 17, 27). While extremely powerful, this method has its limitations, and questions sometimes arise about how physiological the absorption process is under these conditions. From the perspective of this book, the need for special surgical skills and equipment limits the number of animals per experimental group, so it is not suitable for screening studies. This problem is further magnified because lymph and bile duct cannulation is at least an order of magnitude more difficult in mice (one tenth the size) than in rats. These problems were circumvented in most of the studies on knockout and drug-treated mice described above, as well as several on rats and other animal models, by using the alternative method(s) for measuring cholesterol absorption described in detail in this chapter. This method was first pioneered and validated by Quint˜ao et al. (33) and remains widely used today because it is simple, accurate, noninvasive, reproducible, inexpensive, and does not require sophisticated equipment or specialized technical or surgical skills. One of the many difficulties in studying nutrient absorption is that it is sometimes necessary to achieve total recovery of 24 h fecal output. This can be avoided, however, if a non-absorbable analog of the nutrient in question is available. The method also requires that the marker and nutrient can be obtained with different radiolabels, typically 3 H on one and 14 C on the other.
Cholesterol Absorption and Metabolism
161
The marker and nutrient are mixed at a predetermined ratio before being given to the animals. Feces are collected for 24 h and the ratio of excreted marker to excreted nutrient is determined. The percent absorption is calculated from the difference in marker:nutrient ratio between the starting material and the fecal extract. The great advantage of this method is that analysis can be performed on only a portion of the feces, the amount being dictated primarily by the specific activity of the starting material. While it might seem reasonable to use a generic marker such as polyethylene glycol, which is completely eliminated without absorption (used to verify integrity of the epithelial barrier), it is important for the marker to have physical properties similar to the nutrient in question because of the complexity of the postprandial intestinal milieu – a thick slurry of mixed micelles, oil and water phases, and suspended particles. The marker should partition among the phases similarly to the analyte of interest and should have similar intestinal transit times. Thus, sugars must be used to trace sugars, sterols to trace sterols, etc. For cholesterol absorption, one can take advantage of the fact that plant sterols (phytosterols) are poorly absorbed by mice and humans despite close structural similarity to cholesterol. Beta sitosterol, for example, differs from cholesterol only by the addition of an ethyl group to carbon 24 of the sterol side chain (see Fig. 10.1). This reduces absorption to ∼6% (34, 35) as compared to 60–80% for cholesterol. Sitostanol, which has a saturated C5–C6 bond in addition to the ethyl group, is only absorbed at ∼3% (34, 35) and is widely used as a non-absorbed sterol marker (Fig. 10.1).
HO
cholesterol
HO
sitostanol
Fig. 10.1. Structures of cholesterol and a commonly used phytosterol. Adding a branch to the side chain and removing the double bond in the sterol B ring combine to reduce absorption of sitostanol ∼20-fold as compared to cholesterol.
1.2. Cholesterol Excretion and Reverse Cholesterol Transport
Recent years have seen increased focus on enhancing reverse cholesterol transport (RCT), as well as on lowering LDL cholesterol levels, as a strategy for reducing CVD risk. RCT is the HDLmediated process by which excess cholesterol is transported from peripheral tissues to the liver for disposal as biliary cholesterol and bile salts. In particular, there has been emphasis on finding ways to specifically raise HDL cholesterol based on the premise that doing so facilitates RCT by increasing vehicular capacity for the process. This was one of the goals of the CETP (cholesteryl ester transfer protein) inhibitor Torcetrapib developed by Pfizer. While
162
Howles
this drug did successfully raise HDL, unacceptable side effects caused it to fail in phase III clinical trials (36). Also, related studies yielded equivocal results with respect to whether or not RCT was increased by CETP inhibition (37, 38). Thus, targets and treatment modalities to increase RCT represent unmet needs of great potential for drug discovery. Ultimately, what matters is that excess cholesterol be eliminated from the body. Thus, measuring cholesterol excretion has become an important assay for evaluating new potential drug targets and new compounds, and a method for doing so is described in detail below. The assay has the advantage of being completely noninvasive to the animals although it is slightly more demanding technically and requires a gas chromatograph. The method is a modification of that described by Post et al. (39) and has been validated and used extensively by us with good success and reproducibility. In addition, a reasonably physiological method has been recently developed and is now often used to estimate cholesterol flux from macrophage foam cells (as in atherosclerotic plaque) to plasma HDL, then to liver, bile, and feces (40). Briefly, the method involves radiolabeling macrophage cells with cholesterol in culture before administering them to mice by intraperitoneal or subcutaneous injection. Movement and excretion of the radiolabel are measured over the course of 3–5 days to assay the different steps of RCT. Similarly, the relative contribution of different lipoprotein classes to the RCT process in any given model can be assayed by injecting (i.v.) radiolabeled HDL, LDL, or VLDL and quantitating its appearance in the liver, bile and feces over time. Although the experimental details for labeling cells and lipoproteins will not be discussed in this chapter, the final determination of RCT in each of these assays involves measurement of sterol excretion according to the method described here.
2. Materials 2.1. Mice and Diets
C57BL/6 is the most commonly used strain in cardiovascular research so it is the most characterized with respect to cholesterol absorption and excretion. However, mouse strains do differ with respect to cholesterol absorption (41). Thus, it is very important to have accurate control groups for all absorption experiments. Mice should be matched for age, sex, genetic background, and even diet history (unless that is a parameter being studied). For example, while many investigators will purchase control mice from commercial suppliers or will select them from a “wild-type” colony or collection, the purist will use control mice from the
Cholesterol Absorption and Metabolism
163
same breeding colony, if not the same cohort, from which the test mice are derived. Inbred mice give the most consistent results and require fewer animals per treatment group (∼8) but outbred mice can also be used if group size is increased (at least 12). Diet effects on cholesterol absorption and excretion measurements are usually not problematic but should be considered. Standard rodent chows from commercial suppliers work well in most instances. However, lot-to-lot variability in some micronutrients, such as phytoestrogens, may complicate interpretation of results from a serial study that spans several months. Semi-purified diets avoid this possibility. The caveat with semi-purified diets is that most have little or no fiber. Mice eat much less of these diets, have reduced gut motility, and less fecal output. 2.2. Radiolabeled Sterols (see Note 1)
1. 3 H-sitostanol or 3 H--sitosterol
2.3. Lipids (see Note 2)
1. 10 mg/ml phosphatidylcholine (egg PC) dissolved in ethanol, stored at –20◦ C in glass screw-capped vial
2.
14
C-cholesterol
2. 10 mg/ml cholesterol (Sigma-Aldrich) dissolved in ethanol, stored at –20◦ C in glass screw-capped vial 3. Triglyceride: triolein (Sigma-Aldrich) or vegetable oil, such as olive, canola, or safflower oil (grocery brands) stored at –20ºC in glass screw-capped vial to reduce oxidation 4. Filter sterilized 150 mM sodium taurocholate (SigmaAldrich) in water 2.4. Solvents
2.5. Special Supplies
Standard solvents, preferably HPLC grade: ethanol, methanol, petroleum ether, chloroform, hexane, diethyl ether 1. Gavage needles: 20 or 22 G with 1.25 or 2.25 mm ball, for small or large (>35 g) mice, respectively 2. Hamilton syringes: 250, 100, and 10 l with needles 3. Wire platforms for animal cage bottoms 4. 35 ml glass centrifuge tubes with screw caps 5. Ceramic mortar and pestle (two or more)
2.6. Equipment for Sample Preparation and Analysis
1. Needle manifold for drying samples with a nitrogen stream 2. Probe sonicator with narrow tip and variable, pulsatile energy delivery 3. Lyophilizer for drying diet–drug admixtures and feces
2.7. Optional Equipment for Sample Analysis
1. Silica gel G thin layer chromatography (TLC) plates 2. TLC tanks 3. Crystalline iodine (I2 ) for staining
164
Howles
4. Phosphomolybdic acid (10% in ethanol) solution with an atomizer for staining TLC plates 5. Biological oxidizer (e.g., from Harvey instruments) 2.8. Materials Specific to Measurement of Fecal Sterol Mass and RCT
1. Ethanol solutions of 5␣-cholestane, lathosterol, coprostanol, desmosterol, sitosterol, stigmasterol, sitostanol 2.
14
C-taurocholate
3. C18 Bond Elut columns (Varian Inc.) 4. Colorimetric bile acid assay kit (Trinity Biotech PLC) 5. Gas chromatograph (GC) with correct column and configuration to detect neutral sterols
3. Methods 3.1. Cholesterol Absorption 3.1.1. Preparation of the Sterol Mix
1. Dry the appropriate amounts of each radiolabel (usually supplied in ethanol or toluene) under a stream of N2 in a glass tube or glass scintillation vial. Typically, the non-absorbed marker is used at 0.2–0.5 Ci per mouse and the cholesterol at 0.5–1.0 Ci per mouse. 2. Either of two solutions can be used to redissolve the radiolabeled sterols. The first, oil, is simpler to prepare and administer to the mice. However, it is not strictly a physiological representation of dietary cholesterol, which is largely present with phospholipids in cell membranes, and the oil:cholesterol ratio is much greater than would occur in most diets. The second solution, a lipid emulsion, is more tedious to prepare and is less stable but is more physiologically accurate. 2.1. Oil: dissolve the radiolabeled sterols in a volume (l) of triglyceride equal to 100–200 times the number of doses being prepared. Because of viscosity and surface tension, the triglyceride is most accurately measured with a syringe or other positive displacement device rather than a pipette or pipettor. After adding the oil to the tube or vial, warm it to 37◦ C for 15 min and mix it thoroughly by vortexing or with the syringe. The oil can be pure triolein, or vegetable oil such as olive, canola, or safflower oil. 2.2. Emulsion: for each 1 ml of final suspension, add to the tube or vial from Step 1 above, 10 mol of phosphatidylcholine, and 5 mol of cholesterol from the stock solutions, plus 50 mol (48 l) of triolein (or other oil, see above) and dry with a N2 stream at the
Cholesterol Absorption and Metabolism
165
same time as the radiolabel. The organic solvents dry rapidly leaving the lipids partially dissolved in the oil. Add 1 ml of sterile water or saline to the tube and vortex vigorously to start emulsification. Sonicate for 5–10 min on a 2 sec, 50% duty cycle (1 sec on, 1 sec off) with the tube suspended in ice for cooling. Adjust burst energy to maximize sonication but avoid cavitation. Once prepared, the emulsion should be used within a few hours. Monitor its appearance and resonicate if phase separation becomes evident. Alternatively, the emulsion can be stabilized by adding sodium taurocholate to 10 mM (final concentration). This natural bile salt acts as a detergent to prevent phase separation. 3. Transfer 5 l of the material prepared in Section 2.1 or 2.2 into a scintillation vial to determine the isotope ratio in the material that will be administered to the animals. For Section 2.2, it is important to sample when the mixture is thoroughly emulsified. 3.2. Animal Preparation and Treatment
1. Drug pretreatment: either of two approaches can be used for testing drug efficacy. Comparison of a treated group to controls is more common, takes less time and may be preferred if there are concerns about effects due to prolonged vehicle exposure. The alternative is to measure absorption (or excretion) in the same animals before and after drug treatment. This offers the advantage of requiring fewer animals and is valid as long as drug treatment is not lengthy (2–4 weeks) and the animals are adults (10–12 weeks) at the onset of the experiment. Before testing for cholesterol absorption or excretion, considerations must be given to how and how long to administer the candidate compound to the mice prior to testing. This decision will be driven largely by the proposed method of action of the drug candidate. If it acts directly on the small intestine or on pancreatic enzymes in the intestinal lumen, acute treatment, 1–3 days may suffice. If the compound acts systemically and a steady state plasma level must be achieved it may be necessary to pretreat the mice for 1–2 weeks. If the drug is injectable, Steps 1.1 and 1.2 are not relevant. However, oral delivery is preferred for pharmaceuticals (people prefer pills over injections). Also, if the drug is designed to block cholesterol absorption, direct delivery to the site of action, the intestinal lumen and/or the apical surface of enterocytes, is likely to be necessary or at least preferred. 1.1. Drug delivery by diet admix: in most cases, compounds are supplied as powder or crystal. Therefore, an appropriate amount of diet (chow or semi-purified diet
166
Howles
of desired composition) is thoroughly ground (coffee grinder or other kitchen grinder works well). The appropriate amount of compound, ground to fine powder with a smooth-bottom mortar and pestle if at all crystalline, is added and mixed by further grinding. Diet is fed as powder or is repelleted by adding water, compacting with a syringe, and lyophilizing to remove the added water (lyophilize to original weight). The drug:diet ratio is determined by measuring the amount of food eaten per day (2–4 g per day for a 25 g mouse, depending on diet composition) and the desired dose. The latter typically ranges from 1 to 100 mg per kg body weight, but may vary widely. The number of days to feed the compound will depend on its mechanism of action but is typically 3–14 days. Form of the diet, powder or repelleted, should be the same for control mice. 1.2. Drug delivery by oral gavage: for this method, the compound is dissolved in oil (e.g., olive oil) or water, as appropriate, or can be given as a suspension if care is taken to dispense equal doses. The delivery volume should be 100–200 l per mouse. This is measured and dispensed using a Hamilton syringe or similar positive displacement device. The concentration of compound should be adjusted so that the entire daily dose is given in a single gavage. Drug should be given to the mice at the same time each day, typically near the end of the dark cycle when their stomachs are relatively empty. The final dose can (or should) be co-administered with the radiolabeled test mixture, depending on mechanism of action. 2. Acclimation of mice: house mice singly in cages with wire platforms and no bedding for 1 or 2 days prior to testing absorption. This allows the animals to acclimate to the stress of the wire platforms (no bedding) as well as to the stress of being singly housed (no nestling with cage mates). The wire platforms greatly reduce caprophagy and simplify fecal collection. Food and water are supplied ad libitum. 3. Fasting of mice: fast mice during the light cycle on the day of the absorption test. Remove food shortly after the end of the dark cycle. Because this is a short fast (∼8 h) during the time when eating is minimal, it does not stress the animals. Fasting is not strictly necessary, but it minimizes stomach content when the test mixture is administered, which aids delivery by gavage. 4. Treatment of mice with radiolabeled sterol mix: the radiolabeled sterol mix is given to the mice 1–3 h before the
Cholesterol Absorption and Metabolism
167
beginning of the dark cycle. Volume should be 100–200 l and the dose of isotope as described above (see Section 3.1.1). Sedation is not usually necessary but inexperienced handlers may wish to lightly anesthetize with isoflurane to prevent struggling while inserting the gavage tube. With the head tipped back, the tube should slide down the esophagus and into the stomach with little resistance. It is important to insert the tube into the stomach to prevent aspiration of the lipids. For a 25 g mouse, ∼3/4 of the shaft of a 20 G, 2 in. gavage needle is inserted in the mouth when the ball is in the stomach. 5. Feces collection: return the mice to their respective cages and collect feces for 24 h (see Note 3). Longer collection times have the potential to underestimate absorption since some absorbed cholesterol will end up in HDL and subsequently be taken up by the liver, transported to the bile, and secreted into the intestine. However, semi-purified diets are usually very low in fiber, which may reduce intestinal motility. If necessary, sterol transit time is measured in a preliminary experiment where only radiolabeled phytosterol is given and total fecal output is collected on days 1, 2, and 3. Sterols are quantitatively extracted (see below Section 3.1.3) and the time for 90% excretion determined. 6. Tissue analysis: after the 24 h fecal collection, animals are euthanized and any tissues of interest (e.g., for toxicology analysis, such as plasma, liver, kidney, heart, spleen) are harvested. Alternatively, if the mice are precious, they can be used again since the procedure is noninvasive. If the systemic presence of radiolabeled cholesterol does not compromise the second use, little delay is necessary. However, the mice will continue to excrete radiolabeled sterols at detectable levels for 2–3 weeks so bedding must be collected and segregated as contaminated waste. 3.2.1. Extraction of Sterols from Fecal Samples
1. Wire cage bottoms are removed and all fecal output is collected, lyophilized overnight or until dry, and weighed. Normally, 0.8–1.5 g feces can be collected for a 25 g mouse and should be similar for all mice in both control and drugtreated groups. Mice with significantly more or less (e.g., <0.6 or >2.0 g) are noted. If outliers are only in the treated group, consideration is given to side effects or drug toxicity that may affect eating habits or gut motility. 2. Grind the dried material to a coarse powder with mortar and pestle and weigh ∼0.5 g from each mouse into conical glass centrifuge tubes for processing. Save the remainder for reanalysis if necessary. Short-term storage (1–2 weeks) of
168
Howles
lyophilized material can be at room temperature but longterm storage is best at −20ºC. 3. Add 5–10 ml of water to rehydrate the material. Vortex intermittently and heat at 60◦ C for ∼30 min to speed the process. If heated, cool tubes to room temperature before proceeding. 4. Add an equal volume (5–10 ml) of hexane or petroleum ether to the tubes and cap tightly. Shake vigorously for 15 sec to thoroughly mix the phases. Vortexing will suffice but is less effective. 5. Centrifuge the tubes at 2500g for 10 min or until phases are well separated. Recover the organic layer (top) into a scintillation vial or other glass vessel. Take care not to transfer any of the interface or sub-phase. 6. Repeat the extraction twice and combine organic layers. 7. Transfer 0.5 ml of the combined extract to a fresh scintillation vial, evaporate the solvent, resuspend the residue in 0.25 ml methanol or ethanol, add scintillation cocktail, and count using a 3 H, 14 C dual-label program. 8. If cpm values are low (should be several thousand for each isotope), repeat Step 7 using 1 or more ml of the organic fractions. If additional volume is needed, attention should be given to possible quenching, especially in the 3 H channel, due to color in the samples. This is best avoided by including enough isotopes in the gavage dose so small portions (≤ 0.5 ml) of the extract are sufficient for counting. However, there are three possible solutions: 8.1. If the available scintillation counter does not have an adequate quench correction function, it may be necessary to process a non-radioactive fecal sample as described above and add an equivalent amount of the hexane extract to the counting vials for the starting sterol mix (from Section 2.1 or 2.2 above) so it is quenched similarly to the fecal samples. 8.2. A second way to resolve quench problems is to separate the sterols from the pigments by thin layer chromatography (TLC). Briefly, samples are concentrated by evaporation and redissolved in a small volume (≤100 l) of chloroform or hexane and spotted onto a pre-dried, silica gel G, TLC plate along with 5 g of cholesterol standard. Components are resolved with a mobile phase comprised of petroleum ether, diethyl ether, and acetic acid at a 300:200:1 ratio. Sterols migrate 1/4 to 1/3 of the way to the top of the plate, while pigments stay at or near the origin. Spots are visualized with iodine vapor, marked with a pencil, scraped into scintillation
Cholesterol Absorption and Metabolism
169
vials, and counted directly after soaking overnight in scintillation cocktail. 8.3. A third way to avoid pigment interference is to use a biological oxidizer. This specialized apparatus burns samples (dried onto pieces of filter paper) and collects the 3 H and 14 C directly into separate scintillation cocktails as water and CO2 , respectively. 3.2.2. Data Analysis
1. Percent absorption is calculated as 100 × (1 − (14 C/3 H feces/14 C/3 H dose)). 2. Cholesterol absorption by control mice is generally reported in the range of 60–80%. Variations within this range depend in part on small differences in method, sample processing, or animal manipulation. If results are notably above or below this range (less than 40% or more than 80%) for the control/untreated group, the technique used at each of the above steps should be carefully reviewed to determine where faults might lie. Focus especially on problems with counting efficiency and quenching, as noted in Step 8 of Section 3.1.3, and also on sampling of the emulsion in Section 2.2. 3. Attention should also be given to the standard deviation within groups. A typical value is 10–20% of the group average. Greater variation may indicate inconsistent processing of samples or, for the treated group, variation in drug dosing.
3.2.3. Interpretation of Results: Alternative Explanations and Additional Controls
Before accepting the result that absorption is, or is not, different between groups or drug treatments, it may be important to perform one or more control experiments to determine if confounding effects are masking the true result. Three particular areas of concern should be investigated. First, the possibility that the treatment is decreasing activity of the sterol transporter comprised of the two proteins, ABCG5 and ABCG8, that are responsible for preventing phytosterol absorption. All sterols are absorbed by enterocytes and can be incorporated into chylomicrons and transported to the circulation. However, the G5/G8 transporter is present on the apical surface of intestinal epithelial cells and pumps plant sterols back into the lumen all along the small intestine so that the net absorption of phytosterols is very low (less than 10%, as mentioned previously). A simple way to assess this possibility is to calculate the total recovery of radiolabeled phytosterol from 24 and 48 h fecal collections after isotope gavage. This number can be extrapolated from the data generated above by correcting for the fraction of extract counted and the fraction of total fecal material extracted. Recovery of the marker should be ∼90% after 24 h and ≥95% by 48 h. The important
170
Howles
determination is whether or not there is a difference in recovery between groups or treatments. If phytosterol recovery differs significantly between groups, an alternative dual-label method can be used to measure cholesterol absorption. In this case, a colloidal suspension of 3 Hcholesterol (or 14 C) is given intravenously (injected into tail, saphenous, or jugular vein) and an oral dose of 14 C-cholesterol (or 3 H), prepared as described above, is given at the same time. Heparinized or citrated blood samples (100 l) are taken at 48 h and sterols are extracted from whole blood with hexane and counted essentially as described above. Absorption is calculated as 100 × (% of oral dose in plasma aliquot/% of injected dose in plasma aliquot). Thus, it is important to accurately measure the exact doses given for each animal. This method is described in detail in the original paper by Zilversmit and Hughes (42). It has been extensively used and gives results comparable to the method described here. It is technically more demanding, however, because intravenous injections are more challenging in the mouse than in the rat, the species for which the method was originally developed. A second factor that can give rise to misleading cholesterol absorption data is different rates of biliary cholesterol excretion. This is especially the case if mice are fed a standard chow or a semi-purified diet without cholesterol, and/or if the radiolabeled sterols are given in oil rather than as a lipid emulsion. Under these circumstances, the great majority of cholesterol mass in the intestinal lumen is derived from biliary cholesterol. If the latter differs between groups or treatments, the specific activity of the radiolabel (dpm/mg cholesterol) will also differ. Thus, the group having higher biliary cholesterol may appear to have lower absorption because of greater dilution of the radiolabel with unlabeled cholesterol from bile. If there are reasons to suspect differences in biliary cholesterol secretion, it may be necessary to measure cholesterol concentrations in gall bladder bile and/or daily neutral sterol excretion. To obtain gall bladder (GB) bile, animals are fasted for 10–12 h, typically overnight. Fasting interrupts enterohepatic circulation, resulting in greater volume of bile in the GB. After euthanization, GB is exposed and the bile duct is gripped firmly with surgical forceps (smooth, not serrated) where it enters the GB. The organ is removed by cutting the bile duct and then snipping the ligament and connective tissue that holds it in place. Great care must be taken to not pull to hard on the GB and to not nick it with the scissors or forceps. Surprisingly, the amount of connective tissue varies between strains so that GB are easily harvested from FVB/N mice but are more firmly attached in C57BL/6 animals. During and after removal, a firm grip is maintained with the forceps to prevent leakage. The GB is briefly rinsed in a dish of
Cholesterol Absorption and Metabolism
171
PBS and gently blotted to remove blood and tissue fluids, then suspended over the lip of a small (250 l) microcentrifuge tube and punctured with a needle to allow the bile to drain into the tube. Store frozen until assay. There is usually enough material to measure lipid composition (bile acids, cholesterol, phospholipids) with standard colorimetric kits (≤1 l needed for each assay). In addition to biliary cholesterol levels, it is important to take note of bile salt concentrations, since these are the detergents which suspend dietary lipids in micelles and deliver them to the intestinal epithelium for absorption by enterocytes. Differences in bile salt concentration alone could lead to differences in cholesterol absorption. A third consideration is whether or not the treatment has affected fat digestion or absorption. Particular attention should be given to this possibility if bile acid secretion is reduced because micelle capacity will be decreased as mentioned above. If undigested or unabsorbed lipids are present in the lumen, sterols will partition into the oil phase, preventing or delaying their absorption (1, 30). Changes in fat absorption can be measured acutely by giving radiolabeled triolein or tripalmitin (labeled on the fatty acid, not the glycerol) with an oral gavage of olive or other vegetable oil (dosing must be precise). At the same time, mice are also given Poloxamer 407 (1 mg/g body weight; Pluronic F-127, BASF Corp.) by intraperitoneal injection to prevent postprandial hydrolysis and clearance of the absorbed lipid (43). Plasma levels of radiolabel are monitored hourly for 4–6 h to determine the rate and extent of absorption, and whether or not it is delayed in the treated group. A less invasive and more physiological method for measuring fat absorption is also available which uses sucrose polybehenate (a component of OlestraTM ) mixed with the diet as an unabsorbable marker (44). The fatty acid composition of the diet, including the marker behenic acid, is determined by gas chromatography. After 3 days of feeding, fatty acid composition of feces is determined and compared to that of the diet using the same equation shown above for cholesterol absorption. The power of this method is that analysis requires only a few fecal pellets and timing of their collection is not critical; no isotopes are used so the mice can be used for other studies; data are acquired for individual as well as total fatty acids. The latter point can be important because absorption mechanisms and efficiencies vary according to chain length and saturation of individual fatty acids (45). 3.2.4. Interpretation of Results: Benefit Assessment
While assessing the contributions of alternative scenarios described above is important to understanding the mechanism of drug action, assessing the benefit or risk posed by the overall effect is a slightly different process. With respect to the first case described above, increased phytosterol absorption due to
172
Howles
inhibition of ABCG5/G8 is undesirable since it is likely to cause sitosterolemia – a disease with clinical presentations that include elevated LDL and accelerated atherosclerosis. If other methods show that cholesterol absorption is decreased independent of the effect on phytosterols, it may be possible to alter drug structure to be more selective. Effects of drug treatment on bile composition could be beneficial or deleterious. For example, if biliary cholesterol secretion is increased in conjunction with a decrease or no change in cholesterol absorption, the overall effect may be positive. This result would suggest that reverse cholesterol transport (RCT) is increased by the treatment. This would be a very desirable effect as explained above. Increased bile salt secretion could also be beneficial for similar reasons. On the other hand, decreased biliary cholesterol or bile salt concentrations are cause for concern since either could be indicative of some degree of hepatotoxicity and/or cholestasis. At the very least, they would indicate decreased RCT and increased CVD risk. As mentioned above, if drug treatment decreases cholesterol absorption independently of effects on biliary lipids, structural alterations of the compound may improve specificity and/or decrease its absorption so its effect is limited to the intestine. While not altogether desirable, it is not necessarily deleterious if drug treatment decreases fat digestion or absorption as well as cholesterol absorption. As described previously, this is the R mechanism of action of the anti-obesity product alli . However, if decreased cholesterol absorption is due solely to decreased fat absorption, the efficacy is likely to be poor or highly variable and dependent on diet composition. The primary concern with this outcome is that fat-soluble vitamins will partition into the oil phase and be sequestered, reducing their absorption. 3.3. Quantification of Sterol Excretion and Reverse Cholesterol Transport
1. Collect total fecal output for exactly 3 days from mice pre-acclimated to, and individually housed in, cages with wire platforms as described above. Mice should be weighed before and after the collection period.
3.3.1. Sample Collection and Processing
2. Collect fecal pellets and separate from food particles and other detritus, and lyophilize 12–24 h or until a constant weight is reached. Record exact weight and grind to a fine powder. 3. Measure 0.5 g of powder into tube and add 40 g 5␣-cholestane plus 0.02–0.05 Ci of 14 C-taurocholate (see Note 4). 4. Add 5 ml methanol and 1 ml 10 N NaOH to each tube and heat at 80–90ºC for 2 h with intermittent vortexing. Except when vortexing, caps should be slightly loose to
Cholesterol Absorption and Metabolism
173
release pressure. As needed, add methanol to restore original volume. 5. Cool to room temperature, add 10 ml hexane or petroleum ether, tighten caps, and shake vigorously for 15 sec. Centrifuge at 1000g for 10 min to separate phases and remove the organic layer (top) to a fresh tube or scintillation vial. Repeat the extraction twice and combine organic layers. Save the aqueous layer (bottom) in the original tube for bile acid analysis. 6. Evaporate the organic solvent under a nitrogen stream and redissolve the residue in 1–3 ml hexane. Transfer 100 l to a GC vial and cap for analysis. The exact volumes for this step have to be determined empirically and are dictated by the sensitivity and injection settings of individual GC instruments. 7. Prepare GC vials to generate standard curves for 5␣cholestane, cholesterol, coprostanol, and lathosterol. Put appropriate amounts of standard from the ethanol stocks into GC vials, dry, and resuspend in 100 l hexane. Sensitivity varies between instruments but 0.3–10 g (in halflog steps) is a good initial range. Also prepare vials of the various phytosterols (see Section 2.8) to verify peak identification and sufficient separation between peaks to be quantitated (see Note 5). Determine area under the curve (AUC) for all peaks of interest. 8. Filter the aqueous layer from Step 5 through Whatman paper and collect the effluent in a glass tube. Wash the original tube and filter 2–3 times with 5 ml methanol, then dry the filtrate with a nitrogen stream. 9. While the material is drying, prewash the C18 BondElut columns with 3 ml methanol followed by 3 ml water. 10. Resuspend the dried filtrates in 6 ml water and load 3 ml onto the columns by gravity flow and wash three times with 5 ml water. Centrifuge the columns (with plastic collection tube) for 5 min at 50g to remove all the water. Effluents can be discarded. 11. Elute bile acids into a fresh tube with 5 ml methanol. After all 5 ml enters the column, centrifuge as above to maximize sample recovery. 12. Evaporate solvent with nitrogen stream and resuspend in 1 ml methanol. Centrifuge 5 min at 2000g to remove any insoluble residue. Count 20–50 l to determine sample recovery relative to original 0.5 g of fecal powder. Be mindful of potential color quench (see above).
174
Howles
13. Measure bile acids with a standard colorimetric kit using 5–20 l of the solution from 12. Volumes may need to be adjusted due to interference of color or turbidity in the samples or precipitates that may form when reagents and sample are mixed. 3.3.2. Data Analysis
1. The mass of internal standards and analytes of interest in each sample are determined from the appropriate standard curves (mass vs. AUC). The total mass of excreted cholesterol and related sterols is calculated from the 5␣-cholestane peak (i.s.) in each sample: sterol mass per 0.5 g feces = sterol massGC × 40/i.s. massGC . The total excreted per 3 days is then calculated from the total fecal mass collected. Data are finally reported as mass (or moles) excreted per day per gram body weight. 2. The mass of bile acids per 0.5 g feces = mass in 1 ml methanol × dpm 14 C-taurocholate added/dpm recovered in 1 ml methanol. As for neutral sterols, the total excreted per 3 days is calculated from the total fecal mass collected and data are reported as mass (or moles) excreted per day per gram body weight.
3.3.3. Interpretation of Sterol Excretion/Reverse Cholesterol Transport Results
Total neutral sterol excretion by mice (male, C57BL/6) fed a chow diet is typically ∼80 ± 15 nmol per day per gram body weight and bile acid excretion is ∼50 ± 10 nmol per day per g body weight. On semi-purified diets, which are low in fiber and reduce gut motility, neutral sterol excretion is reduced as much as 50% and bile acid excretion can be reduced as much as 5-fold or more (unpublished results). Thus, the diet used in a given study must be chosen carefully. Since human diets have a tendency to be low in fiber, a chow diet may partially mask potential treatment benefits because sterol excretion is already quite high. GC data should be evaluated carefully. The main sterols of interest are cholesterol and lathosterol, the latter being a latestage intermediate in cholesterol synthesis. Coprostanol is formed by the action of colonic bacteria on cholesterol and may vary considerably between mice. Peaks for coprostanol and other minor animal-derived sterols appear very close to, and may overlap with, that of cholesterol. Lathosterol is usually resolved between the cholesterol and the first phytosterol peak. Neutral sterol excretion should be reported as the sum of cholesterol, its precursors, and its derivatives. Evaluation excluding precursors is also appropriate. If bile acid excretion is different, further experiments should be done to determine if bile acid recovery by the intestine or bile acid pool size is changed. The former can initially be assessed by western blot detection of the transporter ASBT. Pool size is
Cholesterol Absorption and Metabolism
175
measured by harvesting liver and intestine from fasted mice and processing them as above starting with 5–10 volumes of methanol + NaOH per gram of tissue. Alternatively, bile acids can be extracted by finely mincing the tissues and soaking them in 5–10 volumes of ethanol at 80ºC for 2–4 h. The resulting solution is then concentrated and assayed as in Steps 8–13 above. Radioactive taurocholate should be used as internal standard in either case. A change in bile acid pool size or intestinal transport is interesting mechanistically in that it indicates greater bile acid synthesis. However, the clinically relevant result is change in sterol excretion. If sterol excretion is increased (neutral or acidic), further investigation is warranted to test for excretion of macrophage cholesterol (40) and prevention or regression of arterial lesions. Particular attention should be given to lathosterol levels in control vs. treated groups. If sterol excretion is significantly elevated (>20%) an even greater difference between groups may be detected for lathosterol, indicating that cholesterol synthesis is elevated in response to treatment. This should be confirmed by direct measurement (46). If true, the result indicates that the drug is having a physiologically significant impact on cholesterol metabolism that may be beneficial and that even greater benefit could be achieved by combined therapy with a statin.
4. Conclusion Cardiovascular disease remains a major cause of morbidity and mortality in the United States and elsewhere despite the efficacy of current therapeutic treatments. Thus, new and better targets for pharmaceutical intervention are needed in addition to modification of diet and lifestyle. This chapter has presented methods for analyzing the first and last stages of cholesterol metabolism – absorption and excretion – in part because absorption is the first line of defense against hypercholesterolemia and excretion is the ultimate measure of drug efficacy with respect to reverse cholesterol transport. These methods highlight two processes that are receiving greater attention with respect to drug development and target identification. In addition to these assays, lipoprotein and plasma cholesterol dynamics should be routinely evaluated as part of any drug screening process. This includes fractionation of plasma to determine distribution of cholesterol among the different lipoprotein classes; studies to determine rates of turnover of HDL, LDL, and VLDL; and clearance of these particles by liver and other tissues. Key studies would be the use of appropriate animal models to test for effects on plaque deposition or
176
Howles
regression. Several widely used methods for such analyses are described throughout the literature. Genetically engineered and inbred mouse models have been, and remain, central tools in unraveling the complexities of cholesterol and lipoprotein metabolism, and the pathologies associated with abnormalities in these processes, resulting in identification of several realized and potential drug targets. While absorption has been well studied in many mouse models, sterol excretion as a measure of RCT received less attention until recent years. Thus, new drug targets for affecting this process, as well as absorption, may be identified by analyzing available mutants as well as new ones that will be created as these pathways continue to be investigated.
5. Notes 1. The chemical purity of these radiolabeled compounds is very high from most sources. If in question, the purity can be checked and restored by standard TLC and/or HPLC methods. 14 C-labeled phytosterols are less available. 2. Lipid stocks in ethanol may precipitate with time at –20ºC and must be completely redissolved at room temperature or 37ºC before use. Triglycerides (triolein, vegetable oils) are also stored at –20ºC to protect them from light and air. If oxidized (color in triolein, “stale” odor in vegetable oils), they should be discarded. All lipids should always be stored in glass, not plastic, if possible. Fresh olive, safflower, and canola oils (grocery) are equally suitable to triolein for the purpose of these methods. 3. Some protocols call for 48 or 72 h feces collection. This is not necessary since ≥ 90% of the unabsorbed marker sterol is recovered within 24 h by mice eating a chow diet. 4. The cholestane and taurocholate serve as internal standards. They should be dispensed with a Hamilton syringe or other positive displacement device to insure accurate delivery to each sample. Let solvents evaporate before proceeding. The exact amount of taurocholate used should also be dispensed into a scintillation vial to accurately determine the number of dpm added to each sample. 5. There is a lot of plant material in standard rodent chows so peaks from plant sterols will be much larger than those of animal sterols in such samples. Chromatographs from mice fed semi-purified diets are simpler and more easily analyzed.
Cholesterol Absorption and Metabolism
177
References 1. Hui, D. Y. and Howles, P. N. (2005) Molecular mechanisms of cholesterol absorption and transport in the intestine. Semin. Cell Dev. Biol. 16, 183–192. 2. Levy, E., Spahis, S., Sinnett, D., Peretti, N., Maupas-Schwalm, F., et al. (2007) Intestinal cholesterol transport proteins: an update and beyond. Curr. Opin. Lipidol. 18, 310–318. 3. Hui, D. Y., Labont´e, E. D., and Howles, P. N. (2008) Development and physiological regulation of intestinal lipid absorption III. Intestinal transporters and cholesterol absorption. Am. J. Physiol. Gastrointest. Liver Physiol. 294, 839–843. 4. Westergaard, H. and Dietschy, J. M. (1976) The mechanism whereby bile acid micelles increase the rate of fatty acid and cholesterol uptake into the intestinal mucosal cells. J. Clin. Invest. 58, 97–108. 5. Thurnhofer, H. and Hauser, H. (1990) Uptake of cholesterol by small intestinal brush border membrane is protein-mediated. Biochemistry 29, 2142–2148. 6. Borja, C. R., Vahouny, G. V., and Treadwell, C. R. (1964) Role of bile and pancreatic juice in cholesterol absorption and esterification. Am. J. Physiol. 206, 223–228. 7. Gallo, L. L., Clark, S. B., Myers, S., and Vahouny, G. V. (1984) Cholesterol absorption in rat intestine: Role of cholesterol esterase and acyl coenzyme A:cholesterol acyl transferase. J. Lipid Res. 25, 604–612. 8. Fernandez, E. and Borgstr¨om, B. (1989) Effects of tetrahydrolipstatin, a lipase inhibitor, on absorption of fat from the intestine of the rat. Biochim. Biophys. Acta 1001, 249–255. 9. McKean M. L., Commons, T .J., Berens, M. S., Hsu, P. L., Ackerman, D. M., et al. (1992) Effect of inhibitors of pancreatic cholesterol ester hydrolase (PCEH) on 14 C-cholesterol absorption in animal models. FASEB J. 6, A1388. 10. Krause, B. R., Sliskovic, D. R., Anderson, M., and Homan, R. (1998) Lipid-lowering effects of WAY-121,898, an inhibitor of pancreatic cholesteryl ester hydrolase. Lipids 33, 489–498. 11. Howles, P. N., Carter, C. P., and Hui, D. Y. (1996) Dietary free and esterified cholesterol absorption in cholesterol esterase (bile saltstimulated lipase) gene-targeted mice. J. Biol. Chem. 271, 7196–7202. 12. Weng, W., Li, L., van Bennekum, A. M., Potter, S. H., Harrison, E. H., et al. (1999)
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
Intestinal absorption of dietary cholesteryl ester is decreased but retinyl ester absorption is normal in carboxyl ester lipase knockout mice. Biochemistry 38, 4143–4149. Camarota, L. M., Chapman, J. M., Hui, D. Y., and Howles, P. N. (2004) Carboxyl ester lipase cofractionates with scavenger receptor BI in hepatocyte lipid rafts and enhances selective uptake and hydrolysis of cholesteryl esters from HDL3 . J. Biol. Chem. 279, 27599–27606. Chaikoff, I. L., Bloom, B., Siperstein, M. D., Kiyasu, J. Y., Reinhardt, W. O., et al. (1952) C14 -cholesterol I: Lymphatic transport of absorbed cholesterol-4-C14 . J. Biol. Chem. 194, 407–412. Heider, J. G., Pickens, C. E., and Kelly, L. A. (1983) Role of acyl CoA:cholesterol acyltransferase in cholesterol absorption and its inhibition by 57-118 in the rabbit. J. Lipid Res. 24, 1127–1134. Clark, S. B. and Tercyak, A. M. (1984) Reduced cholesterol transmucosal transport in rats with inhibited mucosal acyl CoA:cholesterol acyltransferase and normal pancreatic function. J. Lipid Res. 25, 148–159. Gallo, L. L., Wadsworth, J. A., and Vahouny, G. V. (1987) Normal cholesterol absorption in rats deficient in intestinal acyl coenzyme A:cholesterol acyltransferase activity. J. Lipid Res. 28, 381–387. Buhman, K. K., Accad, M., Novak, S., Choi, R. S., Wong, J. S., et al. (2000) resistance to diet-induced hypercholesterolemia and gallstone formation in ACAT2-deficient mice. Nat. Med. 6, 1341–1347. Yagu, H., Kitamine, T., Osuga, J., Tozawa, R., Chen, Z., et al. (2000) Absence of ACAT-1 attenuates atherosclerosis but causes dry eye and cutaneous xanthomatosis in mice with congenital hyperlipidemia. J. Biol. Chem. 275, 21324–21330. Leon, C., Hill, J. S., and Wasan, K. M. (2005) Potential role of acyl-coenzyme A:cholesterol transferase (ACAT) inhibitors as hypolipidemic and antiatherosclerosis drugs. Pharm. Res. 22, 1578–1588. Rudel, L. L., Lee, R. G., and Parini, P. (2005) ACAT2 is a target for treatment of coronary heart disease associated with hypercholesterolemia. Arterioscler. Thromb. Vasc. Biol. 25, 1112–1118. Lada, A. T., Davis, M., Kent, C., Chapman, J., Tomoda, H., et al. (2004) Identification of ACAT1- and ACAT2-specific inhibitors using a novel, cell-based fluorescence assay:
178
23.
24.
25.
26.
27.
28.
29.
30.
31.
32. 33.
Howles individual ACAT uniqueness. J. Lipid Res. 45, 378–386. VanHeek, M., France, C. F., Compton, D. S., McLeon, R. L., Yumibe, N. P., et al. (1997) In vivo metabolism-based discovery of a potent absorption cholesterol inhibitor, SCH58235, in the rat, and rhesus monkey through the identification of the active metabolites of SCH48461. J. Pharmacol. Exp. Therap. 283, 157–163. Altmann S. W., Davis, H. R., Zhu, L., Yao, X., Hoos, L. M., et al. (2004) NiemannPick C1 like 1 protein is critical for intestinal cholesterol absorption. Science 303, 1201– 1204. Garcia-Calvo, M., Lisnock, H. G., Bull, B. E., Hawes, D. A., Burnett, M. P., et al. (2005) The target of ezetimibe is NiemannPick C1-like 1 (NPC1L1). Proc. Natl. Acad. Sci. U.S.A. 102, 8132–8137. Ziajka, P. E., Reis, M., Kreul, S., and King., H. (2004) Initial low-density lipoprotein response to statin therapy predicts subsequent low-density lipoprotein response to the addition of ezetimibe. Am. J. Cardiol. 93, 779–780. Richmond, B. L., Boileau, A. C., Zheng, S., Huggins, K. W., Gramholm, N. A., et al. (2001) Compensatory phospholipid digestion is required for cholesterol absorption in pancreatic phospholipase A(2)-deficient mice. Gastroenterology 120, 1193–1202. Huggins, K. W., Boileau, A. C., and Hui, D. Y. (2002) Protection against diet-induced obesity and insulin resistance in group 1B PLA2 deficient mice. Am. J. Physiol. Endocrinol. Metab. 283, E994-E1001. Homan, R. and Hamelehle, K. L. (1998) Phospholipase A2 relieves phosphatidylcholine inhibition of micellar cholesterol absorption and transport by human intestinal cell line Caco-2. J. Lipid Res. 39, 1197–1209. Huggins, K. W., Camarota, L. M., Howles, P. N., and Hui, D. Y. (2003) Pancreatic triglyceride lipase deficiency minimally affects dietary fat absorption but dramatically decreases dietary cholesterol absorption in mice. J. Biol. Chem. 278, 42899–42905. Mittendorf, B., Ostlund, R. E., Patterson, B. W., and Klein, S. (2001) Orlistat inhibits dietary cholesterol absorption. Obes. Res. 9, 599–604. Drew, B. S., Dixon, A. F., and Drew, J. B. (2007) Obesity management: update on orlistat. Vasc. Health Risk Manag. 3, 817–821. Quint˜ao, E., Grundy, S. M., and Ahrens, E. H. (1971) An evaluation of four methods for measuring cholesterol absorption by the intestine in man. J. Lipid Res. 12, 221–232.
34. Sanders, D. J., Minter, H. J., Howes, D., and Hepburn, P. A. (2000) The safety evaluation of phytosterol esters. Part 6. The comparative absorption and tissue distribution of phytosterols in the rat. Food Chem. Toxicol. 38, 485–491. 35. Igel, M., Giesa, U., Lutjohann, D., and von Bergmann, K. (2003) Comparison of the intestinal uptake of cholesterol, plant sterols, and stanols in mice. J. Lipid Res. 44, 533– 538. 36. Kastelein, J. J., van Leuven, S. I., Burgess, L., Evans, G. W., Kuivenhoven, J. A., Barter, P. J., Revkin, J. H., Grobbee, D. E., Riley, W.A., Shear, C. L., Duggan, W. T., Bots, M. L. and RADIANCE 1 investigators. (2007) Effect of torcetrapib on carotid atherosclerosis in familial hypercholesterolemia. N. Engl. J. Med. 356, 1620–1630. 37. Forrester, J. S., Makkar, R., Shah, P. K. (2005) Increasing high-density lipoprotein cholesterol in dyslipidemia by cholesteryl ester transfer protein inhibition. Circulation 111, 1847–1854. 38. Tchoua, U., D Souza, W., Mukhamedova, N., Blum, D., Niesor, E., Mizrahi, J., Maugeais, C., Sviridov, D. (2008 The effect of cholesteryl ester transfer protein overexpression and inhibition on reverse cholesterol transport. Cardiovasc. Res. 77, 732–739. 39. Post, S. M., de Crom, R., van Haperen, R., van Tol, A., and Princen, H. M. (2003) Increased fecal bile acid excretion in transgenic mice with elevated expression of human phospho lipid transfer protein. Arterioscler. Thromb. Vasc. Biol. 23, 892–897. 40. Zhang, Y. Z., Zanotti, I., Reilly, M. P., Glick, J. M., Rothblat, G. H., et al. (2003) Overexpression of apolipoprotein A-I promotes reverse cholesterol transport from macrophages to feces in vivo. Circulation 108, 661–663. 41. Carter C. P., Howles, P. N., and Hui, D. Y. (1997) Genetic variation in cholesterol absorption efficiency among inbred strains of mice. J Nutr. 127, 1344–1348. 42. Zilversmit, D. B. and Hughes, L. B. (1974) Validation of a dual-isotope plasma ratio for measurement of cholesterol absorption in rats. J. Lipid Res. 15, 465–473. 43. Millar, J. S., Cromley, D. A., McCoy, M. G., Rader, D. J., and Billheimer, J. T. (2005) Determining hepatic triglyceride production in mice: comparison of poloxamer 407 with Triton WR-1339. J. Lipid Res. 46, 2023– 2028. 44. Jandacek, R. J., Heubi, J. E., and Tso, P. (2004) A novel, noninvasive method for
Cholesterol Absorption and Metabolism the measurement of intestinal fat absorption. Gastroenterology 127, 139–144. 45. LaBont´e, E. D., Camarota, L. M., Rojas, J. C., Jandacek R. J., Gilham, D.E., et al. (2008) Reduced absorption of saturated fatty acids and resistance to diet-induced obesity and diabetes by ezetimibe-treated and
179
Npc1l1-/- mice. Am. J. Physiol. Gastrointest. Liver Physiol. 295, G776-G783. 46. Osono, Y., Woollett, L. A., Herz, J., and Dietschy, J. M. (1995) Role of the low density lipoprotein receptor in the flux of cholesterol across the tissues of the mouse. J. Clin. Invest. 95, 1124–1132.
Chapter 11 Collagen-Induced Arthritis in Mice Lisette Bevaart, Margriet J. Vervoordeldonk, and Paul P. Tak Abstract Collagen-induced arthritis (CIA) in mice is an animal model for rheumatoid arthritis (RA) and can be induced in DBA/1 and C57BL/6 mice using different protocols. The CIA model can be used to unravel mechanisms involved in the development of arthritis and is frequently used to study the effect of new therapeutics. The development of a CIA model in C57BL/6 mice recently enabled researchers to use knockout mice on this background for arthritis research. In this chapter, the protocol for induction of arthritis in both mice strains is described, including the monitoring of clinical arthritis and paw swelling in the mice during the experiment. Furthermore, protocols for decalcification of paws and for the detection of collagen-specific antibodies in mice sera are described. Key words: Arthritis, CIA, collagen, DBA/1, C57BL/6.
1. Introduction Rheumatoid arthritis (RA) is a debilitating disease which afflicts about 1% of the population. It is characterized as a chronic, immune-mediated inflammatory disease with destruction of cartilage and bone in the joints. Treatment strategies vary from the use of conventional disease-modifying anti-rheumatic drugs (DMARDs) to biologicals (e.g., TNF-␣ or CD20-specific antibodies). Unfortunately, not all patients respond to treatment and there is still a need for new drug development. Collagen-induced arthritis (CIA) is a frequently used animal model to study the effect of new therapeutics for RA. Similarities between CIA and RA include symmetrical joint involvement, synovial hyperplasia, inflammatory cell infiltrates, and the presence of autoantibodies G. Proetzel, M.V. Wiles (eds.), Mouse Models for Drug Discovery, Methods in Molecular Biology 602, DOI 10.1007/978-1-60761-058-8 11, © Humana Press, a part of Springer Science+Business Media, LLC 2010
181
182
Bevaart, Vervoordeldonk, and Tak
to type II collagen (1). Even production of rheumatoid factor has been reported. CIA is induced by injection of an emulsion of collagen in complete Freund’s adjuvant (CFA). This initiates an immune response directed toward collagen, involving T-and B-cell activation and antibody production. Antibodies recognize collagen in the joints and instigate the local autoimmune response in which monocytes, granulocytes, and T cells are attracted to the joints. This further fuels the inflammatory process resulting in production of cytokines and inflammatory mediators (2, 3). CIA in DBA/1 mice is induced via an initial intradermal (i.d.) injection of an emulsion consisting of bovine type II collagen in CFA. Mice are intraperitoneally (i.p.) injected with bovine type II collagen 21 days later which boosts the immune system, resulting in chronic inflammation in both the hind and the front paws. In C57BL/6 mice, CIA can be induced by two i.d. injections with chicken type II collagen in CFA, on day 0 and day 21 (4). Compared to CIA in DBA/1 mice, CIA on the C57BL/6 background shows a reduced severity in clinical scores and paw swelling and a decreased incidence, approximately 70–80% in C57BL/6 compared to 90–100% in DBA/1 mice. The presence of a CIA model in C57BL/6 mice enables researchers to investigate the role of specific genes, receptors, or immune mediators such as cytokines in knockout or transgenic mice, which often are on a C57BL/6 background. In this way, opportunities arise to study new targets for drug development.
2. Materials 2.1. CIA in DBA/1 Mice
1. Male DBA/1 mice (Harlan [see Note 1]), 8–12 weeks old. All animal experiments have to be done in accordance with institutional and national guidelines and regulations. 2. Bovine type II collagen 2 mg/ml (Chondrex Inc., cat. no. 2002-2). 3. Complete Freund’s Adjuvant, CFA (Mycobacterium tuberculosis, heat inactivated); 4 mg/ml (Chondrex Inc.; cat. no. 7001). 4. Incomplete Freund’s Adjuvant, IFA (Chondrex Inc.; cat. no. 7002).
2.2. CIA in C57BL/6 Mice
1. Male or female C57BL/6 mice (C57BL/6J, Charles River or The Jackson Laboratory), 8–12 weeks old (see Note 2). Try to use animals of one sex. All animal experiments have
Collagen-Induced Arthritis in Mice
183
to be done in accordance with institutional and national guidelines and regulations. When using knockout mice, use wild-type (WT) littermates as control (see Note 3). 2. Acetic acid 100% (Merck, cat. no. 1.00063.1000). 3. 0.1M acetic acid. Dilute acetic acid to a concentration of 0.1M and filter through a 0.2 m filter with use of a syringe. Keep the solution on ice and add acid slowly to the water! 4. Chicken type II collagen, 5 mg (Sigma, cat. no. C9301). Dissolve the complete content of the vial in 2.5 ml sterile filtered 0.1 M acetic acid. Place the vial on a rotator and dissolve the collagen overnight at 4◦ C. 5. Complete Freund’s Adjuvant, CFA (M. tuberculosis, heat inactivated); 5 mg/ml (Chondrex Inc.; cat. no. 7023). 2.3. Preparation of CFA/Collagen Emulsion
1. 14 ml round bottom tubes (Falcon, cat. no. 35-2057)
2.4. Injection of the CFA Emulsion
1. 100 l glass syringes (Hamilton 1710 TLLX with stop; P/N 81022/01)
2. Homogenizer (for example, IKA; model T10 basic, with dispersing tool S10N-5G, IKA, cat. no. 3304000)
2. 25G needles 3. Blades 4. Tweezers 5. Anesthetics (for example, isofluorane) 2.5. Measurement of Arthritis
1. Caliper (for example, Kroeplin L¨angenmesstechnik, cat. no. SO247)
2.6. Sacrificing Mice
1. Eppendorf tubes 2. Tweezers and scissors 3. Anesthetics 4. 4% Formalin 5. Ethanol (Merck, cat. no. 1.00983.2500)
2.7. Decalcifcation of Mice Paws
1. 70% alcohol 2. EDTA (Titriplex II, Merck, cat. no. 1.08418) 3. Sodium hydroxide (NaOH) pellets GR for analysis (Merck, cat. no. 1.06498) 4. Millipore water 5. 4M HCl 6. Tissue embedding and processing cassettes
184
Bevaart, Vervoordeldonk, and Tak
7. Decalcification solution: a. Add 300 g of EDTA to 1800 ml millipore water. b. Add approximately 40 g of NaOH pellets to increase the pH. This will help to dissolve the EDTA. c. After dissolving of the EDTA, set the pH tot 7.5 with use of 4M HCl. 8. Beaker with a grid and a stir bar 2.8. Determining Levels of Anti-collagen IgG1 and IgG2a Antibodies
1. High binding ELISA plates (for example, 96-well RIA/EIA plate, high binding, flat bottom, Costar, cat. no. 3590). 2. Bovine or chicken collagen, see Section 2.1.2 or 2.2.4. 3. PBS. 4. PBS/1%BSA, BSA used should be suitable for ELISA application (for example, Sigma Aldrich, cat. no. A7030). 5. Wash buffer: 0.05% Tween-20 in 1 × PBS. To prepare wash buffer add 500 l Tween-20 (Merck, cat. no. 8.22184) to 1 liter of 1 × PBS. 6. Mouse anti-type II collagen IgG1 antibody solution (10 ng, Chondrex Inc., cat. no. 20306). 7. Mouse anti-type II collagen IgG2a antibody solution (10 ng, Chondrex Inc., cat. no. 20307). 8. Rat anti-mouse IgG1-HRP (1 ml, BD Pharmingen, cat. no. 559626). 9. Rat anti-mouse IgG2a-HRP (1 ml, BD Pharmingen, cat. no. 553391). 10. Sodium acetate buffer, pH 5.5. Prepare by adding 1.3 ml 100% acetic acid to 198.7 ml H2 O. Place the H2 O on ice and slowly add the acetic acid. Be sure to add acid to water! Set pH at 5.5. 11. TMB (3,3 –5,5 -tetramethylbenzidine, Merck, cat. no. 1.08622) in DMSO. Dissolve 100 mg TMB in 10 ml DMSO (dimethylsulfoxide, Merck, cat. no. 1.02950). 12. 3% H2 O2 13. 1M H2 SO4
3. Methods 3.1. Preparation of Bovine Type II Collagen/CFA Emulsion for Injection in DBA/1 Mice
1. Per mouse a volume of 100 l is injected. Before preparation of the emulsion, calculate the total volume of emulsion needed and multiply that by 2 to provide excess for practical losses. A significant portion of the emulsion will be lost due to dead needle volume and refilling of the syringe. For example, when 40 mice need to be injected,
Collagen-Induced Arthritis in Mice
185
Fig. 11.1. Preparation of the collagen/CFA emulsion with the use of the homogenizer.
40 mice × 100 l × 2 = 8 ml of emulsion needs to be prepared. 2. Dilute the CFA 4 mg/ml 1:1 in IFA, which will result in CFA with a concentration of 2 mg/ml. Place the tube on ice. In the example, add 2 ml of IFA to 2 ml of CFA. 3. Add bovine type II collagen drop-wise to the CFA (2 mg/ml) while homogenizing (see Fig. 11.1). Keep the tube on ice! In the example, add 4 ml of bovine type II collagen. 4. Homogenize the emulsion until this appears white and viscous. 5. Perform a test to determine whether the emulsion is well prepared. Pipet 20 l of the emulsion on water. The emulsion must not disperse for at least 40 s but form a nice round-shaped drop. Since the CFA 2 mg/ml is mixed 1:1 with bovine type II collagen 2 mg/ml, the final concentration of M. tuberculosis present in the emulsion is 1 mg/ml (100 g per mouse as 100 l is injected). The final concentration of bovine collagen II is 1 mg/ml (100 g per mouse as 100 l is injected).
186
Bevaart, Vervoordeldonk, and Tak
3.2. Preparation of Chicken Type II Collagen/CFA Emulsion for Injection in C57BL/6 Mice
1. Prepare chicken type II collagen the day before injection of mice as mentioned in Section 2.2.4. 2. Per mouse a volume of 100 l is injected. Before preparation of the emulsion, calculate the total volume of emulsion needed and multiply that by 2 to provide excess for practical losses. A significant portion of the emulsion will be lost due to dead needle volume and refilling of the syringe. For example, when 40 mice need to be injected, 40 mice × 100 l × 2 = 8 ml of emulsion needs to be prepared. 3. Transfer the amount of CFA (concentration 5 mg/ml) that is need to a 14 ml round bottom tube. In the example, pipet 4 ml CFA. 4. Add the chicken type II collagen drop-wise to the CFA (5 mg/ml) while homogenizing (see Fig. 11.1). Keep the tube on ice! In the example, add 4 ml of chicken type II collagen. 5. Homogenize the emulsion until this appears white and viscous. 6. Perform a test to determine whether the emulsion is well prepared. Pipet 20 l of the emulsion on water. The emulsion must not disperse for at least 40 s but form a nice round-shaped drop. Since the CFA 5 mg/ml is mixed 1:1 with chicken type II collagen 2 mg/ml, the end concentration of M. tuberculosis present in the emulsion is 2.5 mg/ml (250 g per mouse as 100 l is injected). The end concentration of chicken collagen is 1 mg/ml (100 g per mouse as 100 l is injected).
3.3. Injection of the Emulsion in Mice
1. Fill the glass syringe with the emulsion. Due to its consistency, this may take some time. 2. Anesthetize the mice (for example, with oxygen 2 l/min, 3% isofluorane). 3. Shave the injection area at the base of the tail. Be careful not to damage the skin. 4. A total of 100 l emulsion is injected at the base of the tail with the glass syringe and 25G needles (see Note 4). This volume is divided over two injection sites. When injected properly, a white spot should be visible at the sites of injection (see Fig. 11.2). 5. Clean the sites of injection with a wipe soaked in ethanol. 6. Put the mice back in their cages where they can recover from anesthesia. Treatment with compounds or drugs can be done either prophylactic or therapeutic (see Note 5). CIA is reported to be
Collagen-Induced Arthritis in Mice
187
Fig. 11.2. Intradermally injected collagen/CFA emulsion.
a good predictive model for outcome of therapy in the clinical situation in combination with the adjuvant arthritis model in rats (5). 3.4. Boost Injection in DBA/1 Mice
At day 21 after the first injection, mice are injected intraperitoneally (i.p.) with bovine type II collagen diluted in saline. A 1:2 dilution of bovine collagen in saline results in a final concentration of 1 mg/ml. Each mouse is injected i.p. with 100 l of this solution, resulting in a boost injection of 100 g bovine type II collagen per mouse.
3.5. Boost Injection in C57BL/6 Mice
At day 21 after the first injection, mice are injected i.d. at the base of the tail close to the previous injection sites with a chicken type II collagen/CFA emulsion. This emulsion is made in exactly the same way as described in Section 3.2 and injected as explained in Section 3.3.
3.6. Monitoring of Arthritis
Arthritis is monitored using a clinical arthritis scoring system and paw swelling is measured using a caliper. Both measurements should be obtained at least three times weekly; if the treatment period is only one week before sacrificing, daily measurements are advised to obtain a sufficient amount of data. Scoring has to be started 5 days before the booster injection for baseline measurements of non-inflamed joints. Typically, signs of arthritis will start after the boost injection; however, paw swelling and clinical arthritis scores are also observed shortly before the boost (see Note 3).
3.6.1. Caliper Score
Hold the mice by the scruff of the neck and tail, and measure paw thickness of both hind ankle joints (Fig. 11.3). Be sure to position the caliper in the same way during each measurement.
188
Bevaart, Vervoordeldonk, and Tak
Fig. 11.3. Paw swelling measurement.
Table 11.1 Assignment of clinical scores in DBA/1 and C57BL/6 mice score DBA/1 mice
C57BL/6 mice
0
Normal
Normal
1
Swelling of 1 joint (wrist/ankle or digit)
Mild swelling
2
Swelling of 2 joints or more
Moderate swelling (or mild swelling + 1 or 2 swollen joints)
3
Swelling of all joints
Swelling of all joints
4
Swelling of all joints and ankylosis (= joint stiffness)
Joint distortion and/or rigidity and dysfunction
3.6.2. Clinical Score
Swelling and inflammation develop differently in the paws of DBA/1 and C57BL/6 mice. Table 11.1 shows how clinical scores are assigned per mouse strain. The score is determined for all four paws. A maximal score of 4 can be reached per paw, resulting in a maximum total score of 16 per mouse (see Note 6).
Collagen-Induced Arthritis in Mice
3.7. Sacrificing of Mice
189
1. Mice are anesthetized (for example, with oxygen 2 l/min, 3% isofluorane). 2. Blood is withdrawn from mice by heart puncture and collected in Eppendorf tubes. Typically, 500 l of blood is obtained from diseased animals. When the blood is needed for specific purposes EDTA tubes or heparin tubes can be used. Blood should be stored on ice if cytokines levels have to be determined at a later time point. 3. Mice are sacrificed by cervical dislocation. 4. Hind paws are cut of at the hips, as it is of great importance to keep the knee joint intact for histological analysis. Skin is removed. 5. Paws are placed in tubes containing 4% formalin. 6. Spin the Eppendorf tubes containing blood at 10,000 rpm, 4◦ C for 10 min. 7. Supernatant is transferred to new Eppendorf tubes after which the tubes are centrifuged again. 8. Supernatant is transferred to new, labeled Eppendorf tubes which are stored at −80◦ C. Make aliquots of serum per animal to prevent multiple freeze–thaw cycles. 9. After 16–24 h, formalin is replaced by 70% ethanol. It is important to replace the formalin within this time frame to preserve the material for further immunohistochemical analysis (see Note 7). Paws can be stored in 70% ethanol for years until further processing. For further processing, paws need to be decalcified (see below) before they are embedded in paraffin. Before decalcification, X-rays of the hind paws can be made for radiologic analysis in which demineralization and erosions of bone are scored. For further details see ref. (6). 10. Depending on the research question, other organs such as spleen and lymph nodes can be removed from the mice for (immunological) analysis, fixed in 4% formalin, and subsequently in 70% ethanol before paraffin embedding. Alternatively, the tissue, placed in aluminum cryotubes, can be snap frozen immediately in liquid nitrogen after dissection. This procedure should be performed if RNA, DNA, or protein needs to be isolated from the tissue. Freshly isolated spleens and draining lymph nodes can also be used for ex vivo stimulation with collagen type II or other stimuli of interest. These procedures are described in several research articles (7–9).
3.8. Decalcification of Mice Paws
1. Transfer mouse hind paws to labeled tissue embedding and processing cassettes.
190
Bevaart, Vervoordeldonk, and Tak
2. Put a stir bar in a 2 l beaker, place a grid within the beaker and fill this with decalcification solution. Transfer the cassettes with the hind paws to the beaker and place it at 4◦ C on a rotator so that the fluid is stirred continuously. 3. Refresh the decalcification fluid two times per week. 4. Decalcification will take 4–6 weeks. Decalcification can be monitored with use of X-rays to check whether the bone has demineralized. 5. Once decalcification is complete, decalcification fluid can be replaced by 70% ethanol. 6. Embed paws with paraffin and make serial 5 m sagittal sections of whole hind paws. Hematoxylin and eosin (to study synovitis), safranin-O (to determine cartilage integrity), or cytokine staining can be performed. For information and scoring, see ref. (6). 3.9. Anti-collagen Antibodies ELISA
1. Coat an ELISA plate overnight with 5 g/ml bovine (DBA/1 mice) or chicken collagen (dissolved according to Section 2.2.4) (C57BL/6 mice) diluted in PBS, 50 l/well. Make sure to prepare different plates to detect either IgG1 or IgG2a collagen-specific antibodies. Coat 2 wells for each sample and standard, including two negative controls (also see Note 8). 2. Incubate overnight at 4◦ C, in the dark. 3. Wash the plate three times with 0.05% Tween-20 in PBS (PBS-Tween), 100 l/well. 4. Block for 1 h with PBS/1% BSA, 100 l/well. 5. Remove the PBS/1%BSA. 6. Prepare standards and serum dilutions. 6.1. Use a standard curve of anti-collagen IgG1 or IgG2a, with a high concentration of 10 ng/ml and a 1:2 serial dilution in PBS/1% BSA (seven steps in total). Incubate two wells per plate with PBS/1% BSA only as a negative control. 6.2. Concentrations of anti-collagen antibodies can vary between mice, so it is important to test a range of concentrations of serum from each mouse in one ELISA. A starting dilution of 1:50 is recommended, with seven 4-fold dilution steps. 6.3. Pipet all standards and samples in duplicate, add 50 l/well. 7. Incubate standard and samples for 1 h at room temperature. 8. Wash the plate three times with PBS-Tween, 100 l/well.
Collagen-Induced Arthritis in Mice
191
9. Add the detection antibody, 50 l/well, diluted in PBS/1% BSA. Use either rat anti-mouse IgG1-HRP 1:1000 dilution or rat anti-mouse IgG2a-HRP 1:1000 dilution. 10. Incubate 1 h at room temperature. 11. Wash the plate three times with PBS-Tween, 100 l/well. 12. Make substrate: add 5.5 ml sodium–acetate buffer, 5.5 l 3% H2 O2 , and 55 l TMB DMSO together. Prepare the substrate just prior to usage and be aware that substrate is light sensitive! 13. Add 50 l substrate solution/well and incubate for 20 min at RT, in the dark! 14. Switch on the ELISA reader and prepare a sample template. 15. Add 25 l/well 1 M H2 SO4 to stop the reaction. 16. Read the plate at 450/540 nm. Make a 4PL curve fit and generate a standard curve with both x-axis and y-axis linearly. Determine concentration of collagen-specific antibodies based on the standard curve.
4. Notes 1. The authors of the article only have experience with mice obtained from Harlan. However, other groups also used DBA/1 obtained from the Jackson Laboratory. 2. For the CIA model in C57BL/6, mice between 12 and 16 weeks of age can also be used although this might not be optimal. 3. Mice can develop less arthritis in a clean(er) animal facility. Other factors, like stress, may also influence the development of arthritis. When moving to a new facility, first test whether the model still works before conducting new experiments. 4. Be sure to inject the collagen/CFA emulsion intradermally, as depth of injection (e.g., subcutaneous vs. intradermal) can influence the immune response. 5. Treatment tested in a so-called prophylactic setting is usually started on days 16–21, whereas therapeutic treatment is often started at day 33. 6. The clinical score better reflects the development of CIA in C57BL/6 mice than the caliper score.
192
Bevaart, Vervoordeldonk, and Tak
7. Fix paws in formalin for no more than 24 h. This is to prevent loss of (cytokine-) epitopes and to ensure good cartilage integrity. 8. In the collagen antibody-ELISA, normal mouse serum can be included as an additional negative control in the same dilution range as sera from mice with CIA. This serum should not react in this ELISA.
Acknowledgments The authors would like to thank Dr. Cristina Lebre for providing Fig. 11.2. References 1. Joe, B. and Wilder, R. L. (1999) Animal models of rheumatoid arthritis. Mol Med Today 5, 367–369. 2. Kannan, K., Ortmann, R. A. and Kimpel, D. (2005) Animal models of rheumatoid arthritis and their relevance to human disease. Pathophysiology 12, 167–181. 3. Luross, J. A. and Williams, N. A. (2001) The genetic and immunopathological processes underlying collagen-induced arthritis. Immunology 103, 407–416. 4. Campbell, I. K., Hamilton, J. A. and Wicks, I. P. (2000) Collagen-induced arthritis in C57BL/6 (H-2b) mice: new insights into an important disease model of rheumatoid arthritis. Eur J Immunol 30, 1568–1575. 5. Hegen, M., Keith, J. C., Jr., Collins, M. and Nickerson-Nutter, C. L. (2008) Utility of animal models for identification of potential therapeutics for rheumatoid arthritis. Ann Rheum Dis 67(11):1505–1515.
6. van Maanen, M. A., Lebre, M. C., van der Poll, T., Larosa, G. J., Elbaum, D., Vervoordeldonk, M. J., et al. (2009) Stimulation of nicotinic acetylcholine receptors attenuates collagen-induced arthritis in mice. Arthritis Rheum 60, 114–122. 7. Khoury, M., Escriou, V., Courties, G., Galy, A., Yao, R., Largeau, C., et al. (2008) Efficient suppression of murine arthritis by combined anticytokine small interfering RNA lipoplexes. Arthritis Rheum 58, 2356–2367. 8. Inglis, J. J., Simelyte, E., McCann, F. E., Criado, G. and Williams, R. O. (2008) Protocol for the induction of arthritis in C57BL/6 mice. Nat Protoc 3, 612–618. 9. Young, D. A., Hegen, M., Ma, H. L., Whitters, M. J., Albert, L. M., Lowe, L., et al. (2007) Blockade of the interleukin21/interleukin-21 receptor pathway ameliorates disease in animal models of rheumatoid arthritis. Arthritis Rheum 56, 1152–1163.
Chapter 12 Skin Diseases in Laboratory Mice: Approaches to Drug Target Identification and Efficacy Screening John P. Sundberg, Kathleen A. Silva, Caroline McPhee, and Lloyd E. King Abstract A large variety of mouse models for human skin and adnexa diseases are readily available from investigators and vendors worldwide. While the skin is an obvious organ to observe lesions and their response to therapy, actually treating and monitoring progress in mice can be challenging. This chapter provides an overview on how to use the laboratory mouse as a preclinical tool to evaluate efficacy of a new compound or test potential new uses for a compound approved for use for treating an unrelated disease. Basic approaches to handling mice, applying compounds, and quantifying effects of the treatment are presented. Key words: Skin, alopecia areata, atopic dermatitis, chronic proliferative dermatitis, full thickness skin grafts, hair, xenograft.
1. Introduction Domestic animals have been incredibly useful for discovering novel approaches to combat major disease problems in humans for centuries. Without question, the most notorious infectious disease of all time with a prominent skin lesion was smallpox. Descriptions by Barron (1) remind us of the severity of the skin disease, even though the viral infection involved many other organs. Jenner’s controversial work, first published in 1798, based on interspecies transmission of cowpox or, more likely, horsepox, from its natural host to human caretakers who subsequently were immune to smallpox infection lead to the development of vaccines (derived from the Latin word vaccinus, which means relating to a cow) G. Proetzel, M.V. Wiles (eds.), Mouse Models for Drug Discovery, Methods in Molecular Biology 602, DOI 10.1007/978-1-60761-058-8 12, © Humana Press, a part of Springer Science+Business Media, LLC 2010
193
194
Sundberg et al.
(2–4). Today we use laboratory mice as the leading biomedical tool and have amassed a variety of molecular methods to analyze mice with spontaneous or genetically engineered diseases that very closely resemble human diseases. These small mammals can effectively be used for preclinical trials as new drugs are developed, to identify potential new uses for approved drugs, or using large-scale gene array technologies with sophisticated software we can now actually predict the best model for drug testing or drugs to test on a particular mouse model. The skin represents a somewhat unique organ system in that with topical administration the potential for ingestion by the mice exists, resulting in undesired systemic delivery. By contrast, effects can easily be seen without special manipulation to rapidly determine if there is efficacy of a compound or not. There are many reviews on constitutes a mouse model for human diseases (5–7), as well as models for specific diseases (8– 22). It is beyond the scope of this chapter to discuss the pros and cons of each of these potential models, especially since a recent scan of the public literature and databases revealed over 1200 genetically engineered and spontaneous mouse mutations that have skin diseases that potentially serve as models for specific human diseases. Information on these are and where to find them is discussed below. Further, this chapter will not describe the usefulness of the mouse to study wound healing, and the reader is referred to other reviews on such (23, 24). In this chapter, we will use several specific models to illustrate how mouse models for dermatology drug studies can be effectively utilized to address primary questions and methodologies applicable to most, if not all of these mouse models. The days of matching mouse spontaneous or experimentally induced diseases with clinically similar human diseases are gone. It is unlikely that all of the complex diseases, such as psoriasis, will ever be adequately mimicked by mouse models as they lack some of the key anatomic structures or response features (see Table 12.1 and 21, 25). However, some complex genetic diseases, like alopecia areata, which appears to be a cell-mediated autoimmune disease that involves the classical lymphocyte co-stimulatory cascade mechanism found in many cell-mediated autoimmune diseases may can be modeled in mice and used for screening compounds with a variety of applications (26, 27). Furthermore, many human diseases are actually groups of very similar diseases lumped together based on clinical features. Now that the molecular bases for many human skin diseases are unraveled, it is possible to accurately match them with the specific mouse homologue(s). Nevertheless, these mouse models are extremely useful when common molecular targets are the focus of testing, which are not always easy to define. If the goal is to alter proliferative, scaly skin disease, albeit psoriasis or ichthyosis, panels of mutant
Skin Diseases in Laboratory Mice
195
Table 12.1 Differences between human and mouse skin Criteria
Human
Mouse
Hair cycle pattern
Mosaic pattern
Wave pattern
Predominant cycle stage
Anagen
Telogen
Hair follicle size
Large
Small
Hair types: head/trunk
Terminal and vellus
Guard, auchene, awl, and zigzag
Unique hair types
No homologous structure
Vibrissae
Relative hair density
Low
High
Epidermal thickness
High
Low
Rete ridges
Prominent
Non-existent
Interfollicular epidermal pigmentation
Yes
No
Hair bulb/fiber pigmentation
Yes
Yes
mice with this basic phenotype are readily available, often with the specific mutated gene known. Efficacy screening for a drug can be correlated with the responding scaly skin mouse model(s) and the target(s) (mutated gene or dysrupted molecular network) with the homologous human disease (28). If the target is known, such as filaggrin, a major risk factor for atopic dermatitis and modifier of other skin diseases (29–38), then a mouse with defects in filaggrin expression, such as the flaky tail mutant mouse (ft), can be used (15, 39). Gene expression profiling is an approach to identify dysregulated genes in response to drug treatment. Data are available through public repositories, the Gene Expression Omnibus at NIH (http://www.ncbi.nlm.nih.gov/geo/) and the ArrayExpress at EBI (http://www.ebi.ac.uk/microarray-as/ae/). References for web sites and repositories worldwide for mice can be found in the Genetically Engineered Mice Handbook (40). A good source is the International Mouse Strain Resource (IMSR) which is available online at (http://www.informatics. jax.org/imsr/index.jsp). Many repository databases provide summaries on the mice and their potential uses. Textbooks continue to serve a useful purpose providing detailed images of the gross and histologic lesions seen in mutant mice (40, 41). Some web sites are highly specialized, such as the Mouse Tumor Biology Database (http://tumor.informatics.jax.org/mtbwi/index.do) which focuses on cancer models, or more generalized mouse
196
Sundberg et al.
pathology databases such as the European Mouse Pathology Consortium web site (Pathbase.net) (42–46). At times it may be difficult to maintain the mutant mouse strain stock or to generate enough sample material. One possibility to overcome this problem is to use full thickness skin grafts, a simple surgical approach, to expand the limited mouse resources (47, 48). More importantly, the most appropriate animal model may not be the mouse but the human. In that case, xenografts can be used in which human skin, or even the skin of any other species (19, 49), is grafted onto immunodeficient mice can become the model of choice (50–52). Yet another reason for doing skin grafts is that rare, low penetrance, complex polygenic diseases, such as alopecia areata, can be reproduced by grafting skin from affected mice to young na¨ıve normal histocompatible mice (17, 53). This chapter will cover approaches to test compounds on mouse skin disease models and evaluating skin to identify potential gene dysregulation that can ultimately suggest FDA approved drugs currently available for treating disease. Quantifying changes, as well as selecting the most appropriate changes to measure, can be a daunting, yet critical aspect of these types of studies. A variety of approaches will be discussed to detail how this can all be done.
2. Material 2.1. Application of Drugs
1. Gavage tube (Instech Laboratories, Inc., Plymouth Meeting, PA) 2. Cherry flavored compounding syrup 3. Elizabethan collars Massachusetts)
(Harvard
Apparatus,
Holliston,
4. Hilltop Chambers (Hill Top Research, Miamiville, Ohio) 5. 3 M Coban wrap 6. Wound clips 7. Nonstick pads, non-adhering dressing (Johnson & Johnson) R 8. ALZET Osmotic Pumps (Durect Corp., Cupertino, CA)
9. Silastic tubing (Dow-Corning Corp., Midland, MI) 10. Custom-made pelleted implants (Brookwood Pharmaceuticals; http://www.brookwoodpharma.com/drug-loadedimplants.html)
Skin Diseases in Laboratory Mice
2.2. Quantification of Drug Response 2.2.1. Hair Regrowth/Repigmentation Evaluation
197
1. Isoflurane 2. Tribromoethanol 3. Stick-on type rulers scene.com/ecpi/references.shtml)
(crime-
4. Dermlite (San Juan Capistrano, CA) (http://www. dermlite.com/) 5. TrichoScan Software (Tricholog, Freiburg, Germany) (www.tricholog.de)
2.2.2. Histopathology
1. Formaldehyde, Fekete’s acid–alcohol–formalin, Bouin’s solution or other tissue fixative 2. 70% ethanol for storage of tissues after fixation 3. Hematoxylin/eosin stain 11. Toluidine blue stain 12. Bromodeoxyuridine (5-bromo-2-deoxyuridine, BrdU) 13. Caspase-3-specific antibody: cleaved caspase-3 (Asp175) antibody (9661, Cell Signaling, Danvers, MA)
2.2.3. Scanning Electron Microscopy
1. Fixation buffer: 2.5% glutaraldehyde in 0.1 M cacodylate buffer 2. Nylon Mesh (Sefar Filtration, Inc., Depew, NY) 14. Hitachi S3000N VP Scanning Electron Microscope (Hitachi Science Systems, Japan) 15. EDAX X-ray microanalysis system (Mahwah, New Jersey)
2.2.4. Transmission Electron Microscopy
1. Extraction buffer: 0.1 mM sodium phosphate buffer (pH 7.9), 2% SDS, 10 mM dithiothreitol 2. Karnovsky’s fixative: 16% paraformaldehyde solution, 50% glutaraldehyde, 0.2 M sodium phosphate buffer 3. 1% osmium tetroxide 4. Dehydration solutions: specimens are transferred through graded solvents (50–60% in distilled water) up to 100% in the solvent. Solvents such as ethanol, methanol, or acetone can be used 5. Epoxy resin combination: Araldite, Embed 812 (Electron Microscopy Sciences, Hatfield, PA) 6. TEM JEOL JEM-1230 (JEOL)
198
Sundberg et al.
3. Methods 3.1. Application of Drugs
There are various ways to administer drugs to rodents in order to test for efficacy and toxicity. The most common approaches are summarized below with references to provide more specific details on how to conduct these types of studies (54).
3.1.1. Oral Routes
Mice can be dosed using a gavage tube or mixing the drug with jelly or cherry syrup in order to allow the rodents to eat the measured portions.
3.1.2. Topical Routes
Although topical applications are the most practical for human patients, it can be problematic for mice depending upon the vehicle. Compounds in volatile vehicles (such as acetone) can be applied with a micropipette, allowed to spread over a marked area (site can be tattooed to ensure that the compound is repeatedly applied to the same site and spread over the same unit area of skin), and then allowed to dry (55). Small amounts can be applied repeatedly allowing for evaporation between applications to maintain volume in a defined area. Aqueous or ointment vehicles are more problematic primarily because they cannot be easily contained. We have tried a variety of stick-on bubble chambers, compression bandages, and “Elizabethan collars” for mice that are available from various vendors with variable results. Elizabethan collars are made of plastic, the neck openings are lined with padding and they close with a Velcro fastener (see Note 1). The Elizabethan collars prevent mice from accessing the drug administration site. Hilltop Chambers are molded plastic chambers that are flexible and conform to the skin. Within the chamber is a pad which holds the test sample. The chambers are secured to the mouse using Coban wrap which is secured by 9 mm wound clips. The compounds/drugs are applied with a pipette under the bandages (see Note 2). Bandaging the drug application site can be a simple and very effective alternative. A nonstick type pad is placed over the site and a small custom cut vest-shaped elastic bandage is used to hold the pad in place. The compound is applied under the bandage daily and the bandages are changed weekly or when untoward effects are noted, e.g., ulcers, swelling (56).
3.1.3. Injectable Routes
Several routes of administration are available. Frequency will depend on route, volume, and tolerance of the compound, all depending upon Institutional Animal Care and Use Committee (IACUC) approval.
Skin Diseases in Laboratory Mice
199
1. Intravenous injections. The ventral coccygeal vein is a common site for intravenous injections. This is located on the ventral side of the mouse tail running along the midline. The mouse can be placed in a restraint device with the tail protruding. This can be a plexiglas box or tube with a slot the width of the tail at one end. The mouse is allowed to enter the device while the tail is held. Warming the mouse or its tail will increase the flow of blood. Small gauge (30 G) needles are used on syringes to inject through the skin and into the vein. Albino strains are easier for venipuncture than pigmented strains as the red vessel can be seen easily through the finely haired skin of the tail. Mouse truncal skin usually lacks interfollicular epidermal pigmentation but this is not the case for the tail skin which also has a very thick epidermis. 2. Subcutaneous injections. The mouse truncal skin is very loosely attached to the underlying fascia and musculature and can be lifted easily. A small gauge needle can then be inserted into the tent of skin over the back and neck once it is raised digitally. 3. Intradermal injections. It is commonly believed that due to the extremely thin dermis and epidermis of the mouse truncal skin, injections into the “dermis” are the same as subcutaneous. However, it is possible to stretch the dorsal or ventral skin and, using a very small gauge needle, position it within the layers of the skin instead of through the skin to produce small blisters with the injected material. 4. Intramuscular injections. Generally, intramuscular injections are made with small gauge needles into the epaxial muscles on either side of the lumbar vertebral column or the quadriceps femoris muscle on the ventral side of either rear leg. Due to the small size of mice, very small volumes (e.g., 0.05 mL) should be injected. 5. Intralesional injections. Small gauge needles can be used on syringes to inject drugs into a neoplasm or other raised abnormality affecting a defined area of the skin. The volume used will depend upon the size and number of lesions being treated. 6. Intraperitoneal injections. The mouse is manually restrained with the head and body tilted downward. A small gauge needle is inserted into the caudal left abdominal quarter, thereby avoiding injection into the cecum on the right side. 3.1.4. Osmotic Pumps
R ALZET Osmotic Pumps are miniature, infusion pumps for the continuous dosing of drugs to mice. These minipumps can be surgically implanted intraperitoneally or subcutaneously to transport
200
Sundberg et al.
drugs from 1 day to 6 weeks. Longer durations may be accomplished through serial implantations of pumps. These minipumps provide a convenient and reliable method for controlled agent delivery in vivo. The doses are constant and accurate and the variables are minimized thereby producing constant results. By using the minipumps the stress and handling are reduced in the mice and this is especially true when giving daily doses. 3.1.5. Slow-Release Subcutaneous Implants
Silastic capsules, consisting of 10 mm silastic tubing packed with drug or hormone with sealed ends (insertion of 3 mm glass beads), can be implanted through a small incision over the dorsal thoracic midline into the subcutaneous fat (57, 58). A number of companies will formulate drugs into various pelleted implants and these companies can be found online (e.g., Brookwood Pharmaceuticals).
3.2. Quantification of Drug Response
Determination of efficacy of a drug in the mouse models appears to be superficially very simple: the mouse resolves the abnormal clinical phenotype and does not die due to the treatment. However, to prove this unequivocally or to determine a dose–response curve can be technically difficult. One must first define the specific goal of the study. With skin this can be as simple as hair growth promotion or even just hair growth on a bald mouse. Wound healing, resolution of scaly skin diseases, etc., can be quantified to various degrees. One must also understand the biological and anatomic differences between mice and humans to appreciate the drug effects (Table 12.1). This begins with the facts that mice have a hair cycle that is in prolonged telogen (humans are in prolonged anagen), the hair cycles in a wave pattern (mosaic pattern in humans), and mouse truncal skin lacks interfollicular epidermal pigmentation (humans have pigment in their epidermis). This translates to the fact that mice have pink skin naturally. The pigment of the skin is really the pigmentation of actively growing, late anagen stage hair follicles where the bulb and hair fiber are heavily pigmented (Fig. 12.1). These observations are well known tools for following induction and patterning of the hair cycle in pigmented mice. Scoring systems by grey tone intensity have been developed for monitoring hair cycle. C57BL/6J (black) and C3H/HeJ (agouti) are commonly used production strains for these types of studies (59).
3.2.1. Hair Regrowth/Repigmentation Evaluation
While a number of magnification and photography tools are available to dermatologists to visualize and record changes in human skin, these can be more problematic in mice because they are relatively small, the body, and therefore the skin, curves making flat sites difficult to find, and they are very active. To avoid, minimize,
Skin Diseases in Laboratory Mice
201
Fig. 12.1. Hair cycle-related skin pigmentation. Hair cycle stage can be estimated by skin color in pigmented (A, C3H/HeJ) but not lightly pigmented or albino (B, BALB/cByJ) mice. Note the light colored (in live mice this is pink vs. gray to black) skin correlates histologically with telogen, while dark skin is due to hair follicles in late anagen. Unlike humans most of the skin pigmentation in mice is in the hair follicles during the actively growing, anagen stage.
or altogether circumvent these issues, a number of approaches can be taken. 3.2.2. Gross photography
How does one stop a very active mouse from moving? This can be simply done at the end of the study by euthanasia using IACUC approved protocols. These are recommended by the American Veterinary Medical Association and the protocols are regularly reviewed and revised (http://www.avma.org/issues/ animal˙welfare/euthanasia.pdf). Currently, a commonly used and IACUC approved euthanasia method is carbon dioxide gas asphyxiation. Alternative survival methods include anesthesia by inhalation (isoflurane) or injectable (tribromoethanol) anesthetics which can be used to temporarily immobilize the mice (60, 61). While isoflurane is commonly used for repeated anesthesias, tribromoethanol is contraindicated for repeated use (60). Another approach for partial immobilization of mice is to use a 50–50% O2 /CO2 mixture in a sealed container (see Note 3). Another alternative is to place the mice in a restraint device in which there is a black concave area for
202
Sundberg et al.
them to insert their heads (62). The side walls are adjustable to keep the mice from moving laterally. The mice are naturally frightened by flood lamps and will stand still without any anesthesia. This does not work in strains that are blind, which is a major issue since many of the commonly used inbred strains, such as C3H/HeJ and FVB/NJ, carrying mutations causing skin diseases also carry the retinal degeneration 1 (e.g., Pde6brd1 , phosphodiesterase 6B, cGMP, rod receptor, beta polypeptide) mutation (63). Another approach is to evaluate efficacy on the ventral abdomen of the mice. It is common practice to handle and restrain mice by picking them up by pinching lightly the skin behind their ears and grabbing the tail with the small finger of the same hand. This restains, immobilizes, and stretches the ventral skin so a photograph can be taken by an assistant of the live mouse. For whole mouse photographs, a high-quality single lens reflex type digital camera provides ease of use while generating high-quality photographs. A regular 50–55 mm macro lens is adequate for whole body images but a 100 mm macro lens allows closer evaluation of the skin surface. Repeat photographs at regular time intervals can be taken of the same area if the skin is tattooed at the start of the project. A fixed ruler should be placed in the field at the same height as the area of skin being photographed as a fixed internal standard for comparison and morphometric analyses. The animal identification number and date can be written on the ruler, especially if disposable, stick-on type rulers are used. Many modern cameras permit data to be added directly to the image (see Note 4). Several companies sell devices that are designed specifically to photograph human skin in a narrow field with or without magnification. These may have associated software to enable specific types of quantitative analyses to be done in a standardized manner. Dermlite has been used for evaluating melanomas and other skin lesions in humans and has potential for fine analysis of drug response in a variety of mouse skin disease models (64–66). This unit attaches to a hand held camera. Tricholog uses a small cylindrical camera that lays directly on the skin surface but the image is visualized, focused, and stored when attached by a cable to a portable computer. The TrichoScan software allows counting and measuring hair fiber size. 3.2.3. Histopathology
Histopathology of mouse skin is preferably done by a board certified veterinary anatomic pathologist with experience interpreting mouse skin. Misinterpretation can be made. For example, modified sebaceous glands are found in the genital region (preputial and clitoral glands) of mice, structures which humans lack. Normal aging changes of these glands have been misinterpreted as sebaceous gland tumors (67) or teratomas (68,69). Normality vs. disease can usually be seen quickly by a pathologist experienced
Skin Diseases in Laboratory Mice
203
with the specific model. This work can be done on routinely fixed and processed hematoxylin and eosin-stained paraffin sections (5– 6 m). Regardless, quantification or semi-quantitative data on specific structures are useful to assess drug response. Most pathologists are used to a very simple but effective method based on severity of the disease. The commonly used adjectives can have numbers added (normal, 0; mild, 1; moderate, 2; severe, 3; and extreme, 4). These can be used effectively when multiple criteria are used to generate a score per criteria as well as a total score. Data are summarized in a spread sheets for all criteria and in this way can be sorted quickly for analysis. This works for all organ systems (70). Real quantitative scores can be generated by counting the number of mitotic figures, specific cell types (e.g., mast cells using a toluidine blue stain), per high power field or whatever microscopic field is most appropriate. The area of the field can be obtained easily by measuring the diameter with a micrometer (A = πr2 ). Image analysis programs, such as NIH Image (http://rsb.info.nih.gov/nih-image/) or a calibrated ocular micrometer, can be used to rapidly generate quantitative data. It is critical to only choose areas of the slide where the entire lengths of hair follicles are present to serve as an internal standard for orientation. This is extremely important for the epidermis because it is very thin in normal mice. Routine measurements can include the interfollicular epidermal thickness (and that of each layer, if appropriate), length of the hair follicle, dermal thickness, and hypodermal fat layer thickness (keep in mind that this normally varies with the hair cycle) (71–73). These are illustrated in Figs. 12.2 and 12.3 (5).
Fig. 12.2. Quantitative measurements of skin. Routine measurements of skin, in late anagen in this figure, include interfollicular epidermal thickness (circled area), hair follicle length (L), dermal thickness (D), hypodermal fat layer thickness (H, normally varies with the hair cycle), and full thickness (FT) from the surface of the stratum corneum to the top of the panniculus carnosus muscle.
Proliferation (mitotic) rate or death (apoptosis of keratinocytes in the epidermis) are other useful criteria. Both can be
204
Sundberg et al.
Fig. 12.3. Epidermal measurements, mitotic figures, and apoptotic keratinocytes in a chronic proliferative dermatitis mutant (Sharpincpdm /Sharpincpdm ) mouse. Routine hematoxylin- and eosin-stained paraffin histologic sections can be used to determine proliferation rates based on mitotic index (number of mitotic figures, circled in the figure, in the stratum basale per 1000 cells) or the presence and numbers of apoptotic epidermal keratinocytes (dotted arrows) when present. Epidermal thickness can be measured at high dry magnification (40×) to include the malpigian, living cell, layer (M), the stratum corneum thickness (SC), or the full thickness of the epidermis (M+SC).
evaluated using a simple hematoxylin- and eosin-stained paraffin section (Figs. 12.3 and 12.4). Mitotic figures, one criterium for proliferation rate, can be seen in the basal cell layer. We found that hematoxylin alone provides optimal contrast to visualize mitotic figures. In traditional hematoxylin- and eosin-stained slides, apoptotic keratinocytes are brightly eosinophilic with dark blue to black shrunken nuclei (pyknotic or karyorrhectic nuclei). Proliferation rates can be quantified by injecting mice with 50 g/g body weight of bromodeoxyuridine (BrdU) intraperitoneally 2 h before necropsy (time interval is critical) and then labeling the tissues with an antibody against BrdU by immunohistochemistry or immunofluorescence (Fig. 12.4). The numbers of positive cells (those with nuclear labeling indicating they incorporated the BrdU when they were in the S phase of the cell cycle) per millimeter of skin, per 1000 basal cells, or other criteria can be used (74). Likewise, activated caspase 3-specific antibodies can be used to identify cells undergoing apoptosis (Fig. 12.5). Terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) is another approach to evaluate apoptosis but this method is not specific in that cells undergoing necrosis may also be positive. The hair cycle can be roughly graded at the gross level in shaved or alopecic mice by mapping color changes in the skin of pigmented mice from pink (follicles in telogen) to increasing
Skin Diseases in Laboratory Mice
205
Fig. 12.4. Bromodeoxyuridine (BRDU) labeling of cells in “S” phase. Routine paraffin sections from mice injected 1 h prior to euthanasia with BRDU have brown nuclei when labeled by immunohistochemistry using diaminobenzidine as a chromogen. These cells were synthesizing DNA at the time of necropsy and therefore incorporated the label. Counting the number of positive nuclei in basal cells per 1000 basal cells yields the proliferation rate. The boxed area in A is enlarged in B to illustrate the large number of positive (dark) nuclei in the skin of an adult chronic proliferative dermatitis mutant mouse (Sharpincpdm /Sharpincpdm ).
Fig. 12.5. Determination of apoptosis. Apoptotic keratinocytes can be confirmed using activated caspase 3-specific antibodies using immunohistochemistry. The boxed area in A is enlarged in B to illustrate the large number of positive (dark) cells (arrows) in the skin of an adult chronic proliferative dermatitis mutant mouse (Sharpincpdm /Sharpincpdm ).
darkening of areas of skin as follicles proceed into the later stages of anagen. The reason for this is described above as illustrated (Figs. 12.1 and 12.6). At the microscopic level this can be done two ways. The major hair cycle stages (anagen, catagen, or telogen) can be estimated (as a percentage of the total skin section) to demonstrate major shifts associated with a mutant mouse phenotype or drug treatment (75). Alternatively, if very subtle differences between the test and control groups need to be identified, scoring hairs based on the finely divided grading of all stages and substages of the hair cycle are a complicated but functional approach. This system, developed by the Paus laboratory (76), divides anagen into six stages, catagen into eight stages, and telogen into one stage (Table 12.2). Although exogen as a stage has been discussed, little is available defining histologic criteria for the purpose of scoring (77).
206
Sundberg et al.
Fig. 12.6. Estimating changes in hair cycle. A simple means to estimate changes in hair cycle in a semi-quantitative manner is to estimate the percent of hair follicles in anagen, catagen, or telogen in histologic sections, entering the data into an Excel type spread sheet, and graph results. The section of skin from the mouse in A is all in telogen. By contrast, the second section (B) is approximately 50% in late anagen, 45% in catagen, and 5% in telogen. These data can be graphically presented in various types of bar graphs (C, D).
Fig. 12.7. Scanning electron microscopy reveals details of hair fibers. Normal hairs from an adult C57BL/6J examined as a whole mount (A) illustrates density of mouse hairs and the nature of the normal skin surface. Manually plucked hairs illustrate the structural differences between some of the hair fiber types (B). Higher magnification of boxed area in B reveals the regular cuticular scale patterns on these hair fibers (C). These approaches illustrate details of hair fiber structure and density (80).
All of the types of data that can be collected, as listed above, can be stored in databases for summary and analysis. One such database, the Mouse Disease Information System (MoDIS), is freely available online (http://research.jax.org/
Skin Diseases in Laboratory Mice
207
Table 12.2 Stages of the mouse hair cycle (76, 77) Anagen
Catagen
Telogen
Exogen
I
I
One stage
Not defined anatomically
II
II
IIIa, b, c
III
IV
IV
V
V
VI
VI VII VIII
faculty/sundberg/registration.php) and it will automatically spell organ names and diseases processes as word strings are typed in. This database allows recording of semi-quantitative histological scores and can be modified easily for specific tasks. More importantly, when used online and linked to Pathbase.net, it is possible to confirm the definition of terms used and search for photomicrographs that can serve as a “virtual second opinion” (78, 79). 3.2.4. Scanning Electron Microscopy
To perform these assays, one must collect hair samples from the mice. This is easily done from adults where hair follicles are in prolonged telogen and exogen stage. Hair fibers are easily removed without pain to the mouse simply by plucking. Because of the hair cycle stage, it is rare that damage occurs to the fibers no matter how defective they are. Alternatively, 1 cm2 biopsies can be removed, laid flat on a firm nylon mesh, and fixed routinely in a glutaraldehyde based fixative. Hair should be removed from the same location on each mouse and the same hair fiber types should be compared at the same location on each hair. We routinely examine hairs from the dorsal, interscapular region of the trunk. We found variations in sulfur levels in normal and mutant hairs from three commonly used inbred strains along the length of the hairs and by hair fiber types although they were not significant in normal mice (81,82). Hair fibers are mounted with double-stick tape on aluminum stubs, sputter coated with a 4 nm layer of gold, and examined at 20 kV at a working distance of approximately 15 mm in our hands on a Hitachi S3000N VP Scanning Electron Microscope (83). Hair fibers are assessed for sulfur content by weight using an EDAX X-ray microanalysis system. Samples are examined for an average of at least 300 live seconds to ensure a comprehensive reading is obtained (36).
208
Sundberg et al.
3.2.5. Transmission Electron Microscopy
In order to evaluate the cellular structure within hair fibers, the fibers can be processed for transmission electron microscopy (TEM). Due to the hard consistency of hairs in plastic blocks, the hairs are easily removed during sectioning which makes evaluation impossible. In order to section the hair fibers they must first be extracted. 1. Ten or more hair fibers per mouse are incubated in 2 mL extraction buffer at room temperature. 2. When the hair became swollen, in comparison to a sample incubated in parallel without dithiothreitol (2–2.5 h) it is immersed in Karnovsky’s fixative. 3. Hair is postfixed with osmium tetroxide overnight at 4◦ C. 4. Hair is dehydrated in graded series of ethanol. 5. Hair is embedded in an epoxy resin combination. 6. Blocks are orientated visually to produce longitudinal and cross sections of the hair fibers. 7. 80 nm sections are prepared. 8. Sections are then examined by a high-performance high contrast TEM (JEOL JEM-1230) (see 82, 84, 85). These TEM studies can be further evaluated by proteomic analysis of hairs to determine specific changes that result from various treatments (84).
3.3. Alopecia Areata
Alopecia areata is a relatively common autoimmune skin disease that affects humans, mice, rats, horses, dogs, cattle, and even a feather form in chickens (26). Although the disease occurs spontaneously, mice have a low frequency of disease. Full thickness skin grafts provided a reproducible and predictable model (53). This mouse has been used effectively to test drugs known to work on humans with alopecia areata by all methods discussed in this chapter (27).
3.4. Chronic Proliferative Dermatitis
The chronic proliferative dermatitis mutant mouse is one of a number of mouse models proposed for psoriasis. This spontaneous mutant was recently shown to be caused by a mutation in the Sharpin gene (86). This mouse model was used to screen recombinant human cytokines in which recombinant interleukin 12 (IL-12) but not interleukin 11 (IL-11) effectively corrected the skin disease (28).
4. Notes 1. When we used Elizabethan collars, every one of the mice managed to slip out of the collars within a 24 h period making these collars unreliable.
Skin Diseases in Laboratory Mice
209
2. Bandages and chambers in our studies were on for 5 days and off for 2 days to allow for any irritation or inflammation of the skin caused by the wound clip to heal. 3. The mice do rapidly recover so the photographer needs to be ready when the mouse is removed from the container. 4. Remember to keep the information off the site of interest. This way, the data markers can be cropped out of the images used in the final report.
Acknowledgments This work was supported by grants from the National Alopecia Areata Foundation (NAAF). References 1. Barron, J. (1838) The life of Edward Jenner, with illustrations of his doctrines and selections from his correspondence. Henry Colburn Publisher, London. 2. Jenner, E. (1798) An inquiry into the causes and effects of the variole vaccinae, a disease discovered in some of the western counties of England, particularly Gloucestershire and known by the name of the cow-pox. Sampson Low, London. 3. Jenner, E. (1959) An inquiry into the causes and effects of the variola vaccinae, a disease discovered in some of the western countries of England, particularly Gloucestershire, and known by the name of cowpox., in Classics of medicine and surgery (Camac, L. N. B., ed.). Dover, New York, pp. 213–240. 4. Stedman, T. L. (2005) Stedman’s medical dictionary for the health professions and nursing, 5th ed., Lippincott, Williams and Wilkins, London. 5. Sundberg, J. P. (1991) Mouse mutations: animal models and biomedical tools. Lab Anim 20, 40–49. 6. Sundberg, J. P. (1993) Conceptual evaluation of animal models as tools for the study of diseases in other species. Lab Anim 21, 48–51. 7. Sundberg, J. P. and King, L. E. (2000) Comparative pathology and animal model development, in Systematic approach to evaluation of mouse mutations (Sundberg, J. P. and Boggess, D., eds.). CRC Press, Boca Raton, pp. 101–109. 8. Sundberg, J. P. and Shultz, L. D. (1991) Inherited mouse mutations: models for the
9.
10.
11.
12. 13. 14.
15.
16.
study of alopecia. J Invest Dermatol 96, 95–96S. Sundberg, J. P. and King, L. E. (1994) Mouse mutations as animal models and biomedical tools for dermatological research. Prog Dermatol 28, 1–12. Sundberg, J. P., Beamer, W. G., Shultz, L. D. and Dunstan, R. W. (1990) Inherited mouse mutations as models of human adnexal, cornification, and papulosquamous dermatoses. J Invest Dermatol 95, 62s–63s. Sundberg, J. P. (1994) Handbook of mouse mutations with skin and hair abnormalities. Animal models and biomedical tools. CRC Press, Inc, Boca Raton, Florida. Sundberg, J. P., HogenEsch, H. and King, L. E. (1995) Mouse models for scaly skin diseases. CRC Press, Inc, Boca Raton. Sundberg, J. P. and L E King, J. (1996) Mouse models for the study of human hair loss. Dermatol Clin 14, 619–632. Sundberg, J. P., Beamer, W. G., Uno, H., Neste, D. V. and King, L. E. (1999) Androgenetic alopecia: in vivo models. Exp Mol Pathol 67, 118–129. Presland, R. B., Boggess, D., Lewis, S. P., Hull, C., Fleckman, P. and Sundberg, J. P. (2000) Loss of normal profilaggrin and filaggrin in flaky tail (ft/ft) mice: an animal model for the filaggrin-deficient skin disease ichthyosis vulgaris. J Invest Dermatol 115, 1072–1081. Dunnwald, M., Zuberi, A. R., Stephens, K., Le, R., Sundberg, J. P., Fleckmann, P., et al. (2003) The ichq mutant mouse, a model for the human skin disorder harlequin ichthyosis:
210
17.
18.
19.
20. 21.
22.
23.
24.
25. 26.
27.
28.
Sundberg et al. mapping, keratinocyte culture, and consideration of candidate genes involved in epidermal growth regulation. Exp Dermatol 12, 245–254. Sundberg, J. P., McElwee, K. J. and King, L. E. (2004) Spontaneous and experimental skin-graft-transfer mouse models of alopecia areata, in Animal models of human inflammatory skin diseases (Chan, L. S., ed.). CRC Press, Boca Raton, pp. 429–449. Sundberg, J. P. and King, L. E. (2003) Mouse alopecia areata models: an array of data on mechanisms and genetics. J Invest Dermatol Sym Proc 8, 173–175. Sundberg, J. P., King, L. E. and Bascom, C. (2001) Animal models for male pattern (androgenic) alopecia. Eur J Dermatol 11, 321–325. Porter, R. M. (2003) Mouse models for human hair loss disorders. J Anat 202, 125–131. Gudjonsson, J. E., Johnston, A., Dyson, M., Valdimarsson, H. and Elder, J. T. (2007) Mouse models of psoriasis. J Invest Dermatol 127, 1292–1308. Plikus, M. V., Sundberg, J. P. and Chuong, C. -M. (2007) Mouse skin ectodermal organs, in The mouse in biomedical research (Fox, J. G., Davisson, M. T., Quimby, F. W., Barthold, S. W., Newcomer, C. E. and Smith, A. L., eds.), 2nd ed. Elsevier, Amsterdam, pp. 691–730. Fang, R. C. and Mustoe, T. A. (2008) Animal models of wound healing: utility in transgenic mice. J Biomater Sci Polym Ed 19, 989– 1005. Schneider, M. R., Werner, S., Paus, R. and Wolf, E. (2008) Beyond wavy hairs: the epidermal growth factor receptor and its ligands in skin biology and pathology. Am J Pathol 173, 14–24. Schon, M. P. (2008) Animal models of psoriasis: a critical appraisal. Exp Dermatol 17, 703–712. King, L. E., McElwee, K. J. and Sundberg, J. P. (2008) Alopecia areata, in Dermatologic immunity: current directions in autoimmunity (Nickoloff, B. J. and Nestle, F. O., eds.). Karger, Basel, pp. 280–312. Sun, J., Silva, K. A., McElwee, K. J., King, L. E. and Sundberg, J. P. (2008) The C3H/HeJ mouse and DEBR rat models for alopecia areata: preclinical drug screening tools. Exp Dermatol 17, 793–805. HogenEsch, H., Torregrossa, S. E., Boggess, D., Sundberg, B. A., Carroll, J. and Sundberg, J. P. (2001) Increased expression of type 2 cytokines in chronic proliferative dermatitis (cpdm) mutant mice and
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
resolution of inflammation following treatment with IL-12. Eur J Immunol 31, 734–742. Palmer, C. N., Irvine, A. D., TerronKwiatkowski, A., Zhao, Y., Liao, H., Lee, S. P., et al. (2006) Common loss-offunction variants of the epidermal barrier protein filaggrin are a major predisposing factor for atopic dermatitis. Nat Genet 38, 441–446. Zhao, Y., Terron-Kwiatkowski, A., Liao, H., Lee, S. P., Allen, M. H., Hull, P. R., et al. (2007) Filaggrin null alleles are not associated with psoriasis. J Invest Dermatol 127, 1878–1882. Sandilands, A., Smith, F. J., Irvine, A. D. and McLean, W. H. (2007) Filaggrin’s fuller figure: a glimpse into the genetic architecture of atopic dermatitis. J Invest Dermatol 127, 1282–1284. Sandilands, A., Terron-Kwiatkowski, A., Hull, P. R., O’Regan, G. M., Clayton, T. H., Watson, R. M., et al. (2007) Comprehensive analysis of the gene encoding filaggrin uncovers prevalent and rare mutations in ichthyosis vulgaris and atopic eczema. Nat Genet 39, 650–654. Palmer, C. N., Ismail, T., Lee, S. P., TerronKwiatkowski, A., Zhao, Y., Liao, H., et al. (2007) Filaggrin null mutations are associated with increased asthma severity in children and young adults. J Allergy Clin Immunol 120, 64–68. Nomura, T., Sandilands, A., Akiyama, M., Liao, H., Evans, A. T., Sakai, K., et al. (2007) Unique mutations in the filaggrin gene in Japanese patients with ichthyosis vulgaris and atopic dermatitis. J Allergy Clin Immunol 119, 434–440. Morar, N., Cookson, W. O., Harper, J. I. and Moffatt, M. F. (2007) Filaggrin mutations in children with severe atopic dermatitis. J Invest Dermatol 127, 1667–1672. McGrath, J. A. and Uitto, J. (2008) The filaggrin story: novel insights into skinbarrier function and disease. Trends Mol Med 14, 20–27. Hamada, T., Sandilands, A., Fukuda, S., Sakaguchi, S., Ohyama, B., Yasumoto, S., et al. (2008) De novo occurrence of the filaggrin mutation p.R501X with prevalent mutation c.3321delA in a Japanese family with ichthyosis vulgaris complicated by atopic dermatitis. J Invest Dermatol 128, 1323–1325. Betz, R. C., Pforr, J., Flaquer, A., Redler, S., Hanneken, S., Eigelshoven, S., et al. (2007) Loss-of-function mutations in the filaggrin gene and alopecia areata: strong risk factor for a severe course of disease in patients
Skin Diseases in Laboratory Mice
39.
40. 41.
42.
43.
44.
45.
46.
47.
48.
49.
50.
comorbid for atopic disease. J Invest Dermatol 127, 2539–2543. Fallon, P. G., Sasaki, T., Sandilands, A., Campbell, L. E., Saunders, S. P., Mangan, N. E., et al. (2009) A homozygous frameshift mutation in the mouse Flg gene facilitates enhanced percutaneous allergen priming. Nat Genet 41(5):602–608. Sundberg, J. P. and Ichiki, T. (2005) Genetically engineered mice handbook. CRC Press, Boca Raton. Ward, J., Mahler, J., Maronpot, R. and Sundberg, J. P. (2000) Pathology of genetically engineered mice. Iowa State University Press, Ames. Naf, D., Krupke, D. M., Sundberg, J. P., Eppig, J. T. and Bult, C. J. (2002) The mouse tumor biology database: a public resource for cancer genetics and pathology of the mouse. Cancer Res 62, 1235–1240. Krupke, D., Begley, D., Sundberg, J., Bult, C. and Eppig, J. (2008) The mouse tumor biology database. Nat Rev Cancer 8(6):459–465. Schofield, P. N., Bard, J. B. L., Boniver, J., Covelli, V., Delvenne, P., Ellender, M., et al. (2004) Pathbase: a new reference resource and database for laboratory mouse pathology. Radiat Prot Dosimetry 112, 525–528. Schofield, P. N., Bard, J. B. L., Booth, C., Boniver, J., Covelli, V., Delvenne, P., et al. (2004) Pathbase: a database of mutant mouse pathology. Nucl Acids Res 32, D512–515. Schofield, P. N., Bard, J. B. L., Rozell, B. and Sundberg, J. P. (2005) Computational pathology: challenges in the informatics of phenotype description in mutant mice, in Handbook on genetically engineered mice (Sundberg, J. P. and Ichiki, T., eds.). CRC Press, Boca Raton, pp. 61–81. Sundberg, J. P., Dunstan, R. W., Roop, D. R. and Beamer, W. G. (1994) Full-thickness skin grafts from flaky skin mice to nude mice: maintenance of the psoriasiform phenotype. J Invest Dermatol 102, 781–788. Sundberg, J. P., Boggess, D., Hogan, M. E., Sundberg, B. A., Rourk, M. H., Harris, B., et al. (1997) Harlequin ichthyosis. A juvenile lethal mouse mutation with ichthyosiform dermatitis. Am J Pathol 151, 293–310. Herbst, L. H., Sundberg, J. P., Shultz, L. D., Gray, B. and Klein, P. A. (1998) Tumorigenicity of green turtle fibropapillomaderived fibroblast lines in immunodeficient mice. Lab Anim Sci 48, 162–167. Baker, B. S., Brent, L., Valdimarsson, H., Powles, A. V., Al-Imara, L., Walker, M., et al. (1992) Is epidermal cell proliferation in psoriatic skin grafts on nude mice driven by T-
51. 52.
53.
54.
55.
56.
57.
58.
59. 60.
61.
62.
211
cell derived cytokines? Brit J Dermatol 126, 105–110. Nickoloff, B. J. (1999) Animal models of psoriasis. Expert Opin Investig Drugs 8, 393–401. Nickoloff, B. J. (2000) The search for pathogenic T cells and the genetic basis of psoriasis using a severe combined immunodeficient mouse model. Cutis 65, 110–114. McElwee, K. J., Boggess, D., King, L. E. and Sundberg, J. P. (1998) Experimental induction of alopecia areata-like hair loss in C3H/HeJ mice using fullthickness skin grafts. J Invest Dermatol 111, 797–803. Boggess, D., Silva, K. A., Landel, C., Mobraaten, L. and Sundberg, J. P. (2004) Approaches to handling, breeding, strain preservation, genotyping, and drug administration for mouse models of cancer, in Mouse models of human cancer (Holland, E. C., ed.). John Wiley & Sons, Inc., Hoboken, NJ, pp. 3–14. Binder, R. L., Gallagher, P. M., Johnson, G. R., Stockman, S. L., Smith, B. J., Sundberg, J. P., et al. (1997) Evidence that initiated keratinocytes clonally expand into multiple existing hair follicles during papilloma histogenesis in SENCAR mouse skin. Mol Carcinog 20, 151–158. Freyschmidt-Paul, P., Happle, R., McElwee, K. J. and Hoffmann, R. (2003) Alopecia areata: treatment of today and tomorrow. J Invest Dermatol Sym Proc 8, 12–17. Jones, R. C. and Datko, L. J. (1977) Administration of anti-inflammatory compounds in silastic capsules Inflammation Res 7, 555–562. Kumar, D., Farooq, A. and Laumas, K. R. (1981) Fluid-filled silastic capsules: a new approach to a more constant steroidal drug delivery system. Contraception 23, 261–268. Uno, H. (1991) Quantitative models for the study of hair growth in vivo. Ann NY Acad Sci 642, 107–124. Meyer, R. E. and Fish, R. E. (2005) A review of tribromoethanol anesthesia for production of genetically engineered mice and rats. Lab Anim (NY) 34, 47–52. Chu, D. K., Jordan, M. C., Kim, J. K., Couto, M. A. and Roos, K. P. (2006) Comparing isoflurane with tribromoethanol anesthesia for echocardiographic phenotyping of transgenic mice. J Am Assoc Lab Anim Sci 45, 8–13. Binder, R. L. (1996) Nonstressful restraint device for longitudinal evaluation and photography of mouse skin lesions during
212
63.
64.
65.
66.
67.
68.
69. 70.
71. 72. 73.
74.
Sundberg et al. tumorigenesis studies. Lab Anim Sci 46, 350–351. Smith, R. S. and Sundberg, J. P. (2002) Strain background disease characteristics, in Systematic evaluation of the mouse eye. Anatomy, pathology, and biomethods (Smith, R. S., John, S. W. M., Nishina, P. M. and Sundberg, J. P., eds.). CRC Press, Boca Raton, pp. 67–75. Argenziano, G., Fabbrocini, G., Carli, P., DeGiorgi, V. and Delfino, M. (1997) Epiluminescence microscopy: criteria of cutaneous melanoma progression. J Am Acad Dermatol 37, 68–74. Argenziano, G., Fabbrocini, G., Carli, P., DeGiorgi, V. and Delfino, M. (1999) Clinical and dermatoscopic criteria for the preoperative evaluation of cutaneous melanoma thickness. J Am Acad Dermatol 40, 61–68. Argenziano, G., Fabbrocini, G., Carli, P., DeGiorgi, V., Sammarco, E. and Delfino, M. (1998) Epiluminescence microscopy for the diagnosis of doubtful melanocytic skin lesions. Comparison of the ABCD rule of dermatoscopy and a new 7-point checklist based on pattern analysis. Arch Dermatol 134, 1563–1570. Nakamura, Y., Fukami, K., Yu, H., Taenaka, K., Kataoka, Y., Shirakata, Y., et al. (2003) Phospholipase C delta 1 is required for skin stem cell lineage commitment. EMBO J 22, 2981–2991. Fu, L., Pelicano, H., Liu, J., Huang, P. and Lee, C. C. (2002) The circadian gene Period2 plays an important role in tumor suppression and DNA damage response in vivo. Cell 111, 41–50. Fu, L., Pelicano, H., Liu, J., Huang, P. and Lee, C. C. (2002) Erratum. Cell 111, 1055. Bleich, A., Mahler, M., Most, C., Leiter, E. H., Lieber-Tenorio, E., Elson, C. O., et al. (2004) Refined histopathologic scoring systems improves power to detect colitis QTL in mice. Mammal Genom 15, 865–871. Chase, H. B. and Eaton, G. J. (1959) The growth of hair follicles in waves. Ann NY Acad Sci 83, 365–368. Chase, H. B., Montagna, W. and Malone, J. D. (1953) Changes in the skin in relation to the hair growth cycle. Anat Rec 116, 75–82. Chase, H. B., Rauch, H. and Smith, V. W. (1951) Critical stages of hair development and pigmentation in the mouse. Physiol Zool 24, 1–10. Smith, R. S., Martin, G. and Boggess, D. (2000) Kinetics and morphometrics, in Systematic approach to evaluation of mouse mutations (Sundberg, J. P. and Boggess,
75.
76.
77.
78.
79.
80.
81.
82.
83.
84.
85.
D., eds.). CRC Press, Boca Raton, FL, pp. 111–119. Sundberg, J. P., Rourk, M., Boggess, D., Hogan, M. E., Sundberg, B. A. and Bertolino, A. (1997) Angora mouse mutation: altered hair cycle, follicular dystrophy, phenotypic maintenance of skin grafts, and changes in keratin expression. Vet Pathol 34, 171–179. Muller-Rover, S., Handjiski, B., vanderVeen, C., Eichmuller, S., Foitzik, K., McKay, I. A., et al. (2001) A comprehensive guide for the accurate classification of murine hair follicles in distinct hair cycle stages. J Invest Dermatol 117, 3–15. Milner, Y., Sudnik, J., Filippi, M., Kizoulis, M., Kashgarian, M. and Stenn, K. (2002) Exogen, shedding phase of the hair growth cycle: characterization of a mouse model. J Invest Dermatol 119, 639–644. Sundberg, J. P., Sundberg, B. A. and Schofield, P. N. (2008) Integrating mouse anatomy and pathology ontologies into a diagnostic/phenotyping database: tools for record keeping and teaching. Mammal Genom 19, 413–419. Sundberg, B. A., Schofield, P. N., Gruenberger, M. and Sundberg, J. P. (2009) A data capture tool for mouse pathology phenotyping. Vet Pathol, in press. Sundberg, J. P. and Hogan, M. E. (1994) Hair types and subtypes in the laboratory mouse, in Handbook of mouse mutations with skin and hair abnormalities: animal models and biomedical tools (Sundberg, J. P., ed.). CRC Press, Boca Raton, FL. Mecklenburg, L., Paus, R., Halata, Z., Bechtold, L. S., Fleckman, P. and Sundberg, J. P. (2004) FOXN1 is critical for onychocyte terminal differentiation in nude (Foxn1μ) mice. J Invest Dermatol 123, 1001–1011. Giehl, K. A., Potter, C. S., Wu, B., Silva, K. A., Rowe, L., Awgulewitsch, A., et al. (2009) Hair interior defect (hid) in AKR/J mice maps to mouse Chromosome 1 Clin Exp Dermatol 34(4):509–517. Bechtold, L. S. (2000) Ultrastructural evaluation of mouse mutations, in Systematic characterization of mouse mutations (Sundberg, J. P. and Boggess, D., eds.). CRC Press, Boca Raton, 121–129. Rice, R. H., Rocke, D. M., Tsai, H.-S., Lee, Y. J., Silva, K. A. and Sundberg, J. P. (2009) Distinguishing mouse strains by proteomic analysis of pelage hair. J Invest Dermatol, submitted. Rice, R. H., Wong, V. J., Pinkerton, K. E. and Sundberg, J. P. (1999) Crosslinked features of mouse pelage hair resis-
Skin Diseases in Laboratory Mice tant to detergent extraction. Anat Rec 254, 231–237. 86. Seymour, R. E., Hasham, M. G., Cox, G. A., Shultz, L. D., Hogenesch, H., Roopenian, D. C., et al. (2007) Spontaneous mutations
213
in the mouse Sharpin gene result in multiorgan inflammation, immune system dysregulation and dermatitis. Genes Immun 8, 416–421.
Chapter 13 Models of Metastasis in Drug Discovery James E. Talmadge Abstract By definition, animal models provide only an approximation of clinical reality. One reason for this, for example, is that although metastases are the primary cause of mortality from neoplasia, by are rarely considered a target in drug discovery and development. Due to the impact of metastasis on clinical disease, we posit that metastasis should be considered in drug discovery, in addition, to more traditional biologic concepts, including drug pharmacology and toxicity. Drug discovery and developmental studies can incorporate orthotopic and spontaneous metastasis models (syngeneic and xenogeneic) with their inherent host–tumor microenvironmental interactions, in addition to confirmatory autochthonous and/or genetically engineered models (GEMs). This requires a rational and hierarchical approach using models of metastatic disease optimally using resected, orthotopic primary tumors and clinically relevant outcome parameters. In this chapter, we provide protocols for models of metastasis that can be used in translational and drug discovery studies. Key words: Metastasis, autochthonous tumors, genetically engineered tumor models, orthotopic tumors, spontaneous metastasis.
1. Introduction Mouse models are critical for the discovery and development of novel therapeutics; however, research has been minimally successful in decreasing the age-adjusted death rate for cancer. In 2003, for the first time since 1930, when epidemiological records were initiated, fewer people (<85 years old) died of cardiac disease as compared to cancer (1). This historic change was due to a 60%, 70%, and 0% decrease in mortality by heart disease, stroke and cancer, respectively. Tumor initiation, progression and metastasis, in contrast, represent a complex, multifactorial process that G. Proetzel, M.V. Wiles (eds.), Mouse Models for Drug Discovery, Methods in Molecular Biology 602, DOI 10.1007/978-1-60761-058-8 13, © Humana Press, a part of Springer Science+Business Media, LLC 2010
215
216
Talmadge
selects tumor variants from a heterogeneous primary tumor (2,3). Further, therapeutic intervention provides an additional selective pressure that results in refractory tumor populations (4). Rodent toxicologic studies can be used to reliably identify safe initial doses for Phase I studies; nonetheless, no animal model to date has been identified that can predict clinical efficacy. Recent studies have suggested that animal models using orthotopically implanted, syngeneic tumors are more predictive than ectopic tumors (5–8). Further there has also been interest in genetically engineered models (GEMs), in part, because of the orthotopic primary tumors and relevance to molecular therapeutics (9). However, GEMs use artificial promoters that can influence the affected cell type, vary expression based on the genetic background (10) and decrease cellular heterogeneity, which in turn can affect tumor progression and metastasis. Critical to studies of therapeutic intervention using animal models is the incorporation of the mechanistic concepts of tumor induction, progression and metastasis. Thus, discovery strategies that incorporate approaches and mechanistic considerations that include the metastatic process and an assessment of clinically relevant outcome parameters may result in the discovery of more effective cancer therapeutics.
2. Materials 2.1. Experimental Metastasis
1. Tumor cell line, e.g., B16 melanoma 2. Medium for tumor cell culture, e.g., Minimal Essential Medium Eagle (MEM) (Gibco, Invitrogen) 3. 10X Trypsin/EDTA: dilute to final concentration 1× with PBS (Sigma-Aldrich) 4. Standardized tumor cell suspension, radiolabeled 5. Hank’s balanced salt solution, Ca++ and Mg++ free (CMFHBSS) (Sigma-Aldrich, H-2387) 6. Trypan blue stain 7. Mouse vice 8. Heat lamps
2.2. Spontaneous Metastasis
1. Standardized tumor cell suspension 2. Hank’s balanced salt solution, Ca++ and Mg++ free (CMFHBSS) 3. Methoxyfluorane or xylazine/ketamine 4. Nail and hair removal cream
Models of Metastasis in Drug Discovery
217
5. Betadine (Purdue Products, 67618-150-01) 6. PBS, sterile (Cellgro/Fisher, 21-040CM) 7. Electric clippers 8. MikRon Wound Clips (Clay Adams, 427631) 9. Fentanyl 10. Tylenol elixir 11. Methoxyfluorane or xylazine/ketamine 12. PBS 13. Scissors, blunt, and curved iris 14. Electric clippers 15. Scalpel (sterile) 16. #11 scalpel blade (Personna, 9-73-0411) 17. MikRon Wound Clips (Clay Adams, 427631) 18. 5–0 nylon surgical suture 2.3. Tumor Evaluation
1. Bouin’s solution: formaldehyde solution, glacial acetic acid, and a saturated solution of trinitrophenol, or 2. Formalin 3. Indian ink
3. Methods 3.1. Experimental Metastasis
Mouse models are critical for the discovery and development of novel therapeutics and to understand the mechanism(s) of metastasis (see Note 1). Commonly, two different models of metastasis are utilized; the first is experimental or artificial metastases, in which tumor cells are injected intravenously or via the left ventricle both of which circumvent part of the metastatic process. As a generality, the first capillary bed encountered by the injected cells provides the site for experimental metastasis. Thus, following lateral tail vein injection, pulmonary metastases are observed in rodents. This contrasts with the profile of metastases observed following injection of the left ventricle which results in hepatic metastases, as well as, potentially bone marrow and brain metastases. Alternatively, models of spontaneous metastasis from a primary tumor at either a subcutaneous site or better an orthotopic site may be used. Because of rapid primary tumor growth it is often resected to allow time for metastases to grow. Therefore, multiple metrics, including number of metastases, metastatic sites, relapse rate, time to recurrence, tumor growth rate, survival and
218
Talmadge
evaluation of angiogenesis, spontaneous metastasis, histopathology, immunohistochemistry and immunomodulation may all be considered in study design (see Note 2 and 3). 3.2. Preparation of Adherent Cells for Experimental Metastasis
Adherent tumor cell cultures can be used in vitro and in vivo to characterize tumor growth, metastasis, regulation of the immune system, and gene expression. The standardization of the cell suspension for injection is a process critical to generate reproducible metastasis. The number of tumor cells injected affects lung colony formation, as does cellular viability, the tumor embolus size and composition. The number of experimental metastases is proportional to the number of viable emboli injected into the circulation (11, 12). However, the number of experimental metastases is also dependant on the size of tumor emboli (number of tumor cells per clump). Therefore, viability and clump (embolus) size influence the outcome of experimental metastasis assays (13). 1. To avoid clumping of cells, the cell preparations must be free of serum and the tumor cells should be injected in a Ca++ - and Mg++ -free balanced salt solution (BSS), which also serves to decrease clumping. 2. Trypsinization in excess of 1 min has been reported to decrease lung colony formation and should therefore be avoided (13). In other studies it has been shown that comparing the incidence of metastasis from the injection of one predetermined dose of tumor cells does not allow an analysis of their relative metastatic capacities. Reproducible and meaningful results require studies that introduce increasing numbers of viable tumor cells admixed with a constant number of non-tumorigenic (X-irradiated) carrier cells (15) (see Note 4). 3. Tumor cell cultures should be split 48 h before use and re-fed 24 h before harvest. 4. Withdraw medium and discard by decanting or aspirating. 5. Add calcium and magnesium free-balanced salt solution (CMF-HBSS) (∼5 ml/25 cm2 ) to the side of the flask opposite the cells so as to avoid dislodging cells, rinse cells, and discard rinse (see Note 5). 6. Add 1× trypsin/EDTA (∼2 ml/25 cm2 ; 3 ml/75 cm2 or 7 ml/150 cm2 ) to the side of the flask opposite the cells. Turn the flask over to cover the monolayer completely. Leave ∼1 min and withdraw the trypsin, making sure that the monolayer has not detached. Using trypsin at 4◦ C helps prevent this (see Note 6). 7. Incubate the cells briefly (1–2 min) at room temperature or 37◦ C and then “tap gently” to remove most of the cell monolayer. Do not leave cells in undiluted trypsin any
Models of Metastasis in Drug Discovery
219
longer than needed, as excessive trypsinization can impact metastatic properties. 8. Add CMF-HBSS to the flask to disperse cells by repeated pipetting over the monolayer surface. In addition, this dilutes the trypsin, reducing degradation of the tumor cell membrane. 9. Finally, pipette the cell suspension up and down a few times, with the tip of the pipette resting on the bottom corner of the flask. This will vary from one cell line to the next. A single cell suspension is desirable to ensure an accurate cell count and uniform lung color formation. 10. Pool the harvested cells into 50 ml centrifuge tubes, fill the tubes with CMF-HBSS, and centrifuge at 1000 rpm for 10 min (250–300 xg). 11. Wash cells 1X in CMF-HBSS and re-suspend in a small volume of CMF-HBSS. 12. Perform a viable cell count using trypan blue stain exclusion and adjust the viable cell concentration to the desired number per milliliter with CMF-HBSS. 3.3. Procedure for Intravenous Injection
1. Fill the syringe with a slightly greater volume than necessary. 2. Tilt the needle upward and draw in air until the meniscus can be seen. 3. Tap or shake the barrel sharply several times in order to dislodge any air bubbles clinging to the sides of the barrel. 4. Small quantities of material should be ejected from the needle until no more bubbles are seen. If needed express saline/air into a cotton ball (see Note 7). 5. For intravenous injections (i.v.), test animals are enclosed in adjustable restraining cages with openings that permitted access to the tail (Fig. 13.1). 6. For injection volumes of less than 100 l, a 100 l syringe should be employed; for volumes exceeding 100 l, a 250 l syringe should be employed, resulting in the most accurate injection volume. A 27 G needle is used, optimally with a length of 3/4 in. The i.v. injection of tumor cells is most accurate when 200 l of a tumor cell suspension in CMF-HBSS. If critical, the 0.25 ml glass syringes are used, although a 1 cc syringe which allows the injection of five mice per filling can also be used, albeit with a small loss in accuracy. In this instance a cage of five mice can be injected with a single syringe (see Note 8). 7. We have found the use of reusable stainless steel needles to be optimal as it is our experience that these are sharper and the needed length can be obtained.
220
Talmadge
8. Immediately prior to dosing, the mice may be weighed, tumors measured and proper dosage calculated (if relevant) on an mg/kg basis. The mice are warmed for 3– 10 min under two heat bulbs, which cover one cage. The distance of the lamp above the cage can vary but 4–8 in. is generally sufficient and this distance and the mouse strain determine the duration. Alternative strategies to dilate the peripheral blood vessels can be used; however, we have found this approach most satisfactory. Thus, placing the tail in pre-heated water or swabbing the tail with xylene are either time consuming or unpleasant/toxic. 9. If skilled operators and a need to inject large numbers of mice are combined, a system can be established whereby one person fills syringes, a second skilled operator injects, and a third person moves cages. If this approach is utilized a series of heat bulbs can be established covering four to six linear feet. As it takes 3–5 min to warm the mice and a skilled operator can inject five mice in 1 min it is possible to have three to five boxes of mice warming, reducing the delay associated with warming. Varying the number of mouse boxes being warmed can be used to compensate for skill levels. The heated animals should be to the left of a right-handed operator and a catch box established to the right. Immediately following injection, the mice are moved into the catch cage. When warming the mice it is necessary to remove all feed and water bottles to prevent shade during the warming of the animals. Furthermore, retention of cage tops is useful as the mice become “bouncy” as they warm. 10. Optimally, an open restraint is used such that a mouse is placed into it and gently pulled to the back of the device
Fig. 13.1. Photo of i.v. mouse tail vein injection.
Models of Metastasis in Drug Discovery
221
(Fig. 13.1). To inject, a mouse is grasped by the tail, taken out of the warming cage and placed into the restraint. The mouse is held in the box, by pulling back on the tail which extends outside through an opening in the restraining device. The tail is grasped between the thumb and the middle finger, wiped clean with 70% alcohol in a cotton ball, and rotated 90º to position the lateral tail vein. The swabbing with a 70% alcohol-saturated cotton ball both cleans the tail superficially and clarifies the vein. The vein is stabilized with the index finger. Slight pressure will straighten the tail and further dilate the lateral vein. The cotton ball is dropped and the syringe picked-up. The needle should be placed on the syringe such that the bevel is perpendicular to the cross supports on the syringe which limits the need to assess where the bevel is. 11. Holding the needle and syringe approximately parallel to the tail, the needle is inserted bevel up into the lateral tail vein. The needle is forced gently through the skin (at a slight angle) and then immediately “threaded” (parallel) into the vein. With experience there is the feeling of being in the groove when threading the needle into the vein. The injection is started slowly; if the needle is not in the vein, tissue resistance is felt. Liquids entering the vein can be observed by clarification of the blood within the vessel during the injection, which should last 5–10 sec. 12. If the mice are to receive therapy at some point following the establishment of pulmonary metastases, one may wish to inject starting at the distal end of the tail slowly moving with multiple injections toward the base of the tail alternating between lateral vessels. This is particularly critical if therapeutic strategies involve the use of chemotherapeutic agents which can be tissue toxic resulting in areas of necrosis and indeed loss of the tail. In our experience with an experienced operator, every other day injections can be undertaken for four weeks with a high degree of success. It should be expected that with repeated injections “hitting” the vein can become increasingly challenging. 3.4. Enumeration of Pulmonary Tumor Colonies 3.4.1. Enumeration of Pigmented Pulmonary Tumor Colonies
Tumor colonies that differ in color from the lung parenchyma (melanoma) can be counted with the aid of a dissecting microscope. 1. Mice are killed, their lungs removed, rinsed in tap water, and fixed in formalin.
222
Talmadge
2. Extrapulmonary metastases should be noted during necropsy and counted. When an organ containing extensive secondary foci is found, it should also be removed, rinsed, and fixed for subsequent enumeration of tumor foci. 3. Most pulmonary tumor colonies in mice are in the pleura. Therefore, pigmented colonies (such as melanomas) can be counted without lung dissection. However, it may be more facile, especially when there are extensive pulmonary metastases, to dissect the lungs into the individual four or five lobes which are then enumerated individually using a dissecting microscope. 3.4.2. Enumeration of Not-Pigmented Pulmonary Tumor Colonies
Tumor colonies that are not pigmented and are minimally different in their color from the pulmonary parenchyma are more difficult to enumerate. The following method is one of several that can be used to induce contrast between tumor colonies and lung parenchyma. 1. Animals suspected of having tumor colonies in an organ are killed, and their lungs or other organs removed and rinsed in tap water. 2. The lungs are then placed in a beaker containing Bouin’s solution. 3. The lung parenchyma will turn yellow and 24 h later, the white tumor colonies are readily distinguished from the yellow lung parenchyma. 4. The lungs are rinsed in water to remove excess Bouin’s solution, and the tumor colonies counted with the aid of a dissecting microscope (see Note 9). Murine lungs with greater than 300 colonies should be reported as >300 or too numerous to count.
3.5. Spontaneous Metastasis from a Primary Tumor
These spontaneous metastasis protocols can be viewed as a generic approach. For example, rather than an injection of the footpad (Long John Silver assay), tumor cells can also be injected into pinna resulting in the Van Gogh assay (16, 17). The general approach is similar although the resection of the popliteal lymph node is irrelevant when the pinna is injected. However, there are challenges to the resection of the pinna such that an electrocautery unit may be needed to reduce bleeding and re-growth of the primary tumor at the surgical site. At least in the case of the B16 melanoma, successful resection of a primary tumor, in the absence of a recurrence in the pinna can be challenging (18). A spontaneous metastasis assay can also be undertaken with the establishment of a primary tumor in the subrenal capsule, such that the kidney is resected with vascular bleeding addressed via the use of an electrocautery unit. Equally, primary tumors can be established in the spleen although ligation of the vasculature may
Models of Metastasis in Drug Discovery
223
be used rather than an electrocautery unit. Other orthotopic or primary tumor sites can be utilized with only minor modification. In the case of spleen and renal primary tumors, a control that includes organ injection, return of the organ into the body cavity, and immediate resection. This provides a control for leakage of cells at the injection site and metastasis by extension. 3.6. Spontaneous Metastasis from a Primary Foot Pad Tumor 3.6.1. Footpad Injection
1. Prepare injection aliquots of 0.05 ml of tumor cells in CMFHBSS. Typically 50,000–150,000 cells are injected using a 0.25 cc or tuberculin syringe. In the case of human cells it may be necessary to inject a higher number of cells depending on the cell line. 2. Injections are best accomplished with two individuals. One individual will hold the inverted mouse and the second operator grasps the leg and injects. Either individual can use thumb and forefinger to occlude retrograde flow from the injection toward the body (Fig. 13.2). 3. Grasp the leg at the ankle or as we prefer the toes and insert the syringe needle (3/8 in., 27 G) into the proximal side of the central pad. 4. Make sure the injection flows toward the nails, not the body.
Fig. 13.2. Photo of mouse footpad injection. A 1 cc tuberculin syringe is shown, which is suboptimal for volume accuracy.
5. The grip between the thumb and the forefinger at the “ankle” is used to reduce reflux of the injected fluid and cells outside of the central pad.
224
Talmadge
6. Once the injection bolus has been delivered, allow the needle to remain in place for 5–10 sec to allow absorption of the fluid/cells. 3.6.2. Foot Pad Tumor Resection
1. Anesthetize mice with methoxyfluorane or an intraperitoneal (i.p.) injection of 0.2 ml of a mixture of xylazine (10 mg/kg) and ketamine (80 mg/kg). 2. The tumor-bearing leg is shaved with electric clippers, as needed Nair can be used for additional removal of hair. 3. The surgical site is cleaned with betadine and followed by a rinse with 70% alcohol. 4. Aseptic techniques are used throughout with surgery undertaken in a biological safety cabinet. 5. Using a heavy pair of scissors, amputate the tumor-bearing leg just above the popliteal lymph node. It is critical to resect the popliteal lymph node as this is a common site of recurrence if it is not successfully removed. 6. Draw the wound together using wound clips and return the animal to its cage. 7. Sterile PBS (0.3 ml) is given i.p. to compensate for mild blood loss during surgery. 8. A pre-emptive analgesic, using a subcutaneous injection of 50 ug/kg fentanyl can be given immediately postoperative as indicated. This may be given q4h 2×, followed by a Tylenol elixir added to the drinking water (1 ml elixir per 100 ml water) in a rodent drinking bottle for 2–3 days post procedure.
3.7. Spontaneous Metastasis Assay from a Primary Mammary Fat Pad Tumor Resection
1. The tumor region (fourth inguinal nipple) is shaved with electric clippers and as needed Nair can be used for additional removal of hair. 2. Mice are lightly anesthetized with methoxyfluorane or an i.p. injection of 0.2 ml of a mixture of xylazine (10 mg/kg) and ketamine (80 mg/kg). 3. The surgical site is cleaned with betadine followed by a rinse with 70% alcohol. 4. Aseptic techniques are used throughout with surgery undertaken in a biological safety cabinet. 5. The initial incision is made with a sterile scalpel using a #11 blade and an elliptic incision made around the tumor. 6. The tumor is dissected free from the abdominal wall and the skin and adhering tumor reflected over the body using forceps or forceps and curved iris scissors. 7. As needed the tumor can be blunt dissected free from the fascia using blunt scissors.
Models of Metastasis in Drug Discovery
225
8. During reflection of the tumor from the body, major vessels will be cauterized (using a standard battery-operated surgical cauterizer) as they appear. The goal is to reduce blood loss and more importantly to prevent a surgical shower of tumor cells. 9. A margin of at least 5 mm is maintained from the tumor during surgery. 10. Once freed from the fascia and reflected over the body, the tumor and remaining skin are cut free of the fascia. 11. Prior to closing, the abdominal wall and local regional area will be cleansed with 70% alcohol to kill any tumor cells released during surgery. 12. After tumor removal, the skin is closed with sterile, stainless steel wound clips. In most cases, primary orthotopic tumors have not invaded into the abdominal wall and so this remains intact. 13. If it is necessary to excise part of the abdominal wall, it will be closed with continuous sutures, using 5–0 nylon surgical suture material prior to closing the outer skin using wound clips. 14. PBS (0.3 ml) is given i.p. to compensate for mild blood loss during surgery. 3.8. Genetically Engineered Tumor Models (GEMs)
GEMs provide rigorous models for drug discovery and development. However, one of the challenges with these models is to dissociate tumor prophylaxis from tumor therapy. Depending upon the model one can have premalignant disease or malignant disease that is subclinical when one initiates therapy. At the time of gross neoplastic disease, as is the case clinically, little time may remain with which to intercede against the tumor. Thus, one may initiate therapy once a gross lesion is observed, such that rather than entering cohorts of animals into therapy, animals go on to protocol as they develop tumors which provides a timing challenge to such models. Further, many of these models are characterized by the development of multiple primary tumors, and one can follow the growth of multiple lesions and assess tumor multiplicity in addition to survival as outcomes. Over the past 20 years, GEMs have contributed to our understanding of the molecular pathways responsible for the initiation, progression, and metastasis of cancer cells, and extended our understanding of the mechanistic role that oncogenes and tumor suppressor genes have in these processes. The initial GEMs were murine models that overexpressed viral and cellular oncogenes (19). Subsequent studies utilized genes targeted to mouse embryonic stem cells providing oncogene-bearing transgenic mice (knock-in) or loss of function, i.e., gene knockout (KO)
226
Talmadge
mice. In addition to the use of transgene overexpression models, conditional strategies have been developed that allow controlled gene expression in both a tissue- and a temporal-specific manner (20). Thus, tetracycline (tet)-regulated (21) or CRE-inducible alleles can regulate the timing, duration, and tissue compartment of gene expression or inactivation. Further, these technologies can be combined, resulting in GEMs with specific cancers that overexpress or lack genes of interest in all cells or in a specific tissue compartment and/or developmental stage (refer Chapter 4 by Naf). GEMs have been primarily used to study specific therapeutic questions relevant to the affected gene and to study interactions between tumor cells and their microenvironment. They are potentially more representative of specific human tumor histotypes than transplanted xenografts due to their in situ and autochthonous origin (22). However, GEMs have limitations, including expense, time commitment, intellectual property restrictions (23), and species-specific differences, resulting in different mutant phenotypes in human and mouse (24). Further, no one transgenic model is representative of all the different forms of even one tumor histotype; just as one human tumor cannot represent another human tumor of the same histotype. In addition, the transgenes in GEMs are driven by artificial promoters, which may influence the cell type affected. Further the different genetic background from different mouse strains can affect the transgene expression. By their very nature, GEMs cannot incorporate the heterogeneity inherent to tumor initiation, progression, and metastasis (25) and systemic disease is rarely observed in GEMs (26, 27). 3.9. Autochthonous Tumor Models
Autochthonous tumors include spontaneously occurring tumors and chemical, viral or physical carcinogen-induced tumors. The advantages of autochthonous tumors include orthotopic growth, tumor histology devoid of transplantation introduced changes, and metastasis via lymphatic and vascular vessels surrounding and within the primary tumor (28). Despite such positive properties, autochthonous tumor models have not been widely used as an animal model. Autochthonous tumor models have an inherent variability in the time to and frequency of tumor induction, and number of tumor(s) induced. Thus, the number of animals required for a study is large (29). These models require a timeframe of several months to a year for a single experiment, as opposed to weeks with transplanted xenograft models (28– 30). As such, autochthonous tumor models are best reserved for confirmation studies (28), although in the ‘‘post-genome era,’’ autochthonous models have to an extent been replaced by GEMs.
Models of Metastasis in Drug Discovery
227
3.10. Outcome Criteria for Tumor Models
Our ultimate goal is to cure patients of their tumors. However, more realistic goals in clinical oncology are to improve survival, quality of life and prevent disease recurrence (31). Given the differences between animal models and clinical realities, it is generally easier to obtain positive outcomes in animal models then clinically and so rigorous criteria are required in the use of animals in drug discovery. Given the diversity in tumor models different strategies are used with the different tumor models (see Note 10).
3.10.1. Experimental Metastasis
The assessment of the number of experimental metastases, typically pulmonary metastases, following i.v. injection of a single cell suspension of tumor cells or collagenase-digested primary tumor provides the most common outcomes. Multiple strategies may be used for this enumeration including 1. Enumeration of gross metastases following fixation in Bouin’s fixative or injection of Indian ink to increase contrast between the tumor and the parenchyma. 2. Sites of metastasis, typically with a focus on the lung, liver, marrow and brain. 3. Incorporation of 125 I cpm into the target organ (typically pulmonary) following i.p. injection of 125 Iudr. 4. Weight of organ with metastatic foci. 5. Number of neoplastic tissue culture colonies formed following culture of cells obtained from collagenase-dissociated target organ with metastatic foci. 6. Number of drug resistant colonies formed following culture of cells obtained from collagenase-dissociated organs with metastatic foci composed of tumor cells transfected with a selective vector. 7. The frequency or number of tumor cells can be identified by flow cytometry from collagenase-dissociated organs with metastatic foci. A viral vector with a florescent transgene can be used to mark tumor cells and used to identify tumor cells. Alternatively, antibodies to cytokeratin and permeabilized cells can be used to identify mammary or other epithelial origin tumor cells. Antibodies to CD45 can be used to identify hematopoietic origin cells, while antibodies to CD31 can be used to identify vascular cells. 8. Median survival of mice bearing experimental metastasis.
3.10.2. Spontaneous Metastases from a Primary Tumor
Perhaps the most common model used to assess outcomes uses syngeneic mice given a subcutaneous injection of a transplantable, syngeneic tumor or xenogenic tumor injected into nude or SCID mice. These tumor models allow a broad range of outcome strategies in response to therapeutic intervention. Alternatively, models
228
Talmadge
of spontaneous metastasis from a primary tumor at a “natural” or orthotopic primary tumor site can be used. In orthotopic studies, end points, which match the study context, accessibility of the implantation site, type of implantation and therapeutic or mechanism under study may be utilized. Therefore, multiple metrics, including number of metastases, metastatic sites, time to recurrence, tumor growth rate, survival and evaluation of angiogenesis, spontaneous metastasis and histopathology, immunohistochemistry and immunomodulation, may all be considered as part of the study design. Tumor growth rate, as assessed by repeated measure of tumor burden, is analyzed using a repeated measures test (32). 1. Using a digital caliper, measure the longest dimension of the tumor and then perpendicular to this measurement such that the two horizontal axes are measured. Alternatively, one can image tumors formed by tumor cells transfected with a retroviral vector containing a florescent transgene using bioluminescence imaging system or wild-type tumors using ultrasound, PET, or MRI imaging all of which provide digital data. However, assessment of such digital data is a preclinical and clinical challenge (33). 2. Tumor volume is determined by the formula: where a = the larger axis and b = the smaller axis using the formula for a prolated sphere where volume = 0.5×a×b2 . 3. Time to reach a predetermined tumor burden, an important parameter at institutes with aggressive IACUC committees. 4. Overall survival (Kaplan–Meier survival curve and nonparametric statistical analysis). 5. Number of pulmonary or other organ metastases as assessed using any of the techniques identified in Section 3.4. 6. Sites of metastases. 7. Frequency of mice which develop metastases. 8. Frequency of mice with hepatic and or marrow metastatic foci. 9. Frequency of mice which develop a gross tumor following injection of a defined number of tumor cells. 3.10.3. Spontaneous Metastasis Following Resection of a Primary Tumor
Clinical observations have suggested that the organ environment can influence the response of tumors to therapy. For example, in women with breast cancer, lymph node and skin metastases are more sensitive to chemotherapeutic intervention than metastases in either the lung or the bone (34). Similarly, orthotopic implantation of human tumor cells from surgical specimens into nude mice has shown that colon carcinomas injected into the wall of the colon, results in more clinically relevant outcomes then subcutaneously injected colon carcinomas. Orthotopic implanta-
Models of Metastasis in Drug Discovery
229
tion of tumor cells results in rapid growth of local tumors and in several tumor models, distant metastasis. There is also a striking, site-specific variation in response to chemotherapy. In one study, colon carcinoma cells were implanted into different anatomic locations of nude mice including mice injected into the subcutis (ectopic site), spleen (leading to experimental liver metastasis), or cecum (growth at the orthotopic site) (35). Tumor-bearing mice were treated with doxorubicin and subsequently evaluated for responses. Tumors grown within the subcutis tissue showed an 80% inhibition of growth after two i.v. injections of doxorubicin, compared to about 40% inhibition of the intracecal tumors and less than 10% inhibition of lesions in the liver (35). Thus, it was concluded that subcutis tumor models were not representative of the primary tumor site (36). In addition, clinically we treat well established and frequently advanced metastatic disease, whereas conventional subcutis xenograft models are of recent origin (1–14 days) and rarely have metastatic disease (37). Thus, orthotopic tumor models may provide a better model to assess the morphology and the growth characteristics of clinical disease (5–7, 38) and to be more representative of a primary tumor with respect to tumor site and metastasis (8). One of the obvious advantages of orthotopic models is that targeting processes involved in local invasion (e.g., angiogenesis) can be undertaken at a clinically relevant site (36). Since the early studies showing orthotopic transplantation of colon tumors and metastasis to the liver (39), tumor xenografts have been grown orthotopically in mice. However, despite the clinical relevance of orthotopic models, their utilization is hindered by a need for a high level of technical skill, time and cost. Therapeutic efficacy is also more difficult to assess with orthotopic models in contrast to the relative ease of subcutis tumor measurements (36). Tumor prophylactic models, wherein therapy is initiated prior to tumor challenge, are not clinically realistic. Similarly, models that measure tumor growth delay, i.e., the measurement of time required to reach a predetermined median tumor volume, have not been shown to predict survival in mice, much less humans. Treatment of mice bearing gross tumors typically assess the therapeutic response based on slowed tumor growth kinetics as opposed to tumor regression. In contrast, the Response Evaluation Criteria in Solid Tumors (RECIST) criteria classically used for evaluating efficacy in human clinical trials requires at least 50% shrinkage in tumor size in order to be considered a response. Thus, outcome criteria used with animal models and humans differ. Clinically and in rodent studies, survival provides a rigorous and consistent end point for the evaluation of treatment efficacy. However, preclinical studies that monitor survival at the termination of therapy are inadequate due to the lack of post-treatment follow-up. Survival must be followed after treatment to assess the
230
Talmadge
complete life expectancy to and if tumor re-growth occurs. In spite of a delay in tumor growth over the treatment period, a rebound effect post-treatment can occur and is indicative of lower overall efficacy. It should be noted that a chronic slow tumor growth rate, i.e., cytostasis, can be considered a relevant outcome as compared to tumor regression if our goal is to delay tumor progression. Thus, in addition to the outcomes listed for experimental metastasis and spontaneous metastasis from a subcutaneous primary tumor, outcomes focused on murine models using metastasis from mice with orthotopic primary tumors following resection provide a still more relevant model. Given that the primary tumor can induce mortality prior to the development of gross metastatic disease the prolonged survival obtained by surgical resection of the primary disease results in growth of secondary disease as occurs clinically. Thus, models of resected primary tumors allow relevant clinical outcomes to be used including frequency of primary tumor relapse and percent surgical cure. 3.10.4. Outcomes from GEM and Autochthonous Tumor Models
In addition to the prolonged time to assessment of outcomes, GEM and autochthonous tumor models provide additional measures of outcomes. In addition to those already mentioned, these include 1. Tumor multiplicity and assessment of total number of tumors at autopsy. 2. Time to primary tumor development. 3. Time to secondary tumor development. 4. Tumor growth rates of primary and/or secondary tumors. 5. Mitotic index of the primary tumor. 6. Cellular apoptosis within the primary tumor. 7. Angiogenesis and lymphangiogenesis with the primary tumor. 8. T-cell, macrophage, and DC infiltration of tumors.
4. Notes 1. Tumor and mouse models need to meet specific biologic criteria including heterogeneity, appropriate histology and metastatic propensity, appropriate genetic criteria depending on the targeted drug mechanism, limited immunogenicity and potentially etiology. 2. Models should have the potential to provide a correlation between model outcome and clinical activity, optimally with previous documentation of relevance between mice and humans (26, 40, 41).
Models of Metastasis in Drug Discovery
231
3. Prior to undertaking efficacy studies, baseline pharmacokinetic (PK) and toxicity data are needed, including an initial analysis of cellular/organ toxicity. It should be noted that the toxicity and therapeutic profile can differ significantly between push and continuous infusion. Thus, PK studies can help focus the initial dose finding and toxicological studies. 4. Prolonged and unnecessary enzymatic treatment, i.e., trypsinization of tumor cells, can also alter their survival and metastatic behavior in vivo. Moreover, viability tests (trypan blue exclusion) and even plating efficiency in vitro do not predict or correlate with the in vivo biological behavior of trypsinized cells. 5. The wash step is to remove traces of serum, which inhibits the action of trypsin. If necessary the wash step can be repeated, especially if using the cells for injection into animals. 6. It is common to purchase trypsin as 10× and to aliquot this in 50 ml tubes which are removed from the freezer as needed. It is important to limit the storage of the trypsin at 4◦ C temperatures to no longer than a week as the trypsin will auto-digest. 7. The intravenous injection of air into a mouse can be a lethal event. 8. No more than 107 cells should be injected i.v. as this results in pulmonary embolization. Typically, we inject 50,000– 150,000 murine tumor cells i.v.; however, with human cells, especially collogenase dissociated primary human tumor cells, higher numbers of cells are required. 9. It should be noted that this technique can also be successfully applied to hepatic metastasis. 10. Weight should also be monitored and if a cohort looses >30% weight, this dose can be identified as the maximum tolerated dose (MTD). References 1. Jemal, A., Siegel, R., Ward, E., Murray, T., Xu, J., Smigal, C. and Thun, M. J. (2006) Cancer statistics, 2006. CA Cancer J Clin 56, 106–130. 2. Talmadge, J. E., Wolman, S. R. and Fidler, I. J. (1982) Evidence for the clonal origin of spontaneous metastases. Science 217, 361–363. 3. Talmadge, J. E. and Fidler, I. J. (1982) Cancer metastasis is selective or random depend-
ing on the parent tumour population. Nature 297, 593–594. 4. Talmadge, J. E., Benedict, K., Madsen, J. and Fidler, I. J. (1984) Development of biological diversity and susceptibility to chemotherapy in murine cancer metastases. Cancer Res 44, 3801–3805. 5. Fidler, I. J. (1991) Orthotopic implantation of human colon carcinomas into nude mice provides a valuable model for the biology and
232
6.
7.
8.
9.
10.
11.
12. 13.
14. 15.
16.
17.
18.
Talmadge therapy of metastasis. Cancer Metastasis Rev 10, 229–243. Hoffman, R. M. (1999) Orthotopic metastatic mouse models for anticancer drug discovery and evaluation: a bridge to the clinic. Invest New Drugs 17, 343–359. Hoffman, R. M. (1997) Fertile seed and rich soil: development of patient-like models of human cancer by surgical orthotopic implantation of intact tissue, in update series: comprehensive textbook of oncology, Vol. 4 Schimpff, S. C. et al. (eds.), Williams & Wilkins, Baltimore, pp 127–144. Killion, J. J, Radinsky, R. and Fidler, I. J. (1999) Orthotopic models are necessary to predict therapy of transplantable tumors in mice. Cancer Metastasis Rev 17, 279–284. Dexter, D. L., Diamond, M., Creveling, J. and Chen, S. F. (1993) Chemotherapy of mammary carcinomas arising in ras transgenic mice. Invest New Drugs 11, 161–168. Moser, A. R., Hegge, L. F. and Cardiff, R. D. (2001) Genetic background affects susceptibility to mammary hyperplasias and carcinomas in Apc(min)/+ mice. Cancer Res 61, 3480–3485. Zeidman, I., McCutcheon, M. and Coman, D. R. (1950) Factors affecting the number of tumor metastases; experiments with a transplantable mouse tumor. Cancer Res 10, 357– 359. Fisher, B. and Fisher, E. R. (1968) Role of host and tumor calcium in metastases. Cancer Res 28, 1753–1758. Liotta, L. A., Saidel, M. G. and Kleinerman, J. (1976) The significance of hematogenous tumor cell clumps in the metastatic process. Cancer Res 36, 889–894. Fidler, I. J. and Talmadge, J. E. (1978) Methods in cancer research, in preclinical screeing, Vol. 2, Academic Press. Hart, I. R., Talmadge, J. E. and Fidler, I. J. (1983) Comparative studies on the quantitative analysis of experimental metastatic capacity. Cancer Res 43, 400–402. Fidler, I. J. (1980) Therapy of spontaneous metastases by intravenous injection of liposomes containing lymphokines. Science 208, 1469–1471. Donawho, C. K., Pride, M. W. and Kripke, M. L. (2001) Persistence of immunogenic pulmonary metastases in the presence of protective anti-melanoma immunity. Cancer Res 61, 215–221. Rebhun, R. B., Lazar, A. J., Fidler, I. J. and Gershenwald, J. E. (2008) Impact of sentinel lymphadenectomy on survival in a murine
19.
20.
21.
22.
23.
24. 25.
26. 27. 28.
29.
30.
31. 32.
model of melanoma. Clin Exp Metastasis 25, 191–199. Brinster, R. L., Chen, H. Y., Messing, A., Van, D. T., Levine, A. J. and Palmiter, R. D. (1984) Transgenic mice harboring SV40 T-antigen genes develop characteristic brain tumors. Cell 37, 367–379. Heineke, J. and Molkentin, J. D. (2006) Regulation of cardiac hypertrophy by intracellular signalling pathways. Nat Rev Mol Cell Biol 7, 589–600. Gossen, M. and Bujard, H. (1992) Tight control of gene expression in mammalian cells by tetracycline-responsive promoters. Proc Natl Acad Sci U S A 89, 5547–5551. Sharpless, N. E. and Depinho, R. A. (2006) The mighty mouse: genetically engineered mouse models in cancer drug development. Nat Rev Drug Discov. 9, 741–54. Weiss, B. and Shannon, K. (2003) Mouse cancer models as a platform for performing preclinical therapeutic trials. Curr Opin Genet Dev 13, 84–89. Jacks, T. (1996) Tumor suppressor gene mutations in mice. Annu Rev Genet 30, 603–636. Stringer, J. R., Larson, J. S., Fischer, J. M., Medvedovic, M., Hersh, M. N., Boivin, G. P. and Stringer, S. L. (2005) Modeling variation in tumors in vivo. Proc Natl Acad Sci U S A 102, 2408–2413. Hann, B. and Balmain, A. (2001) Building ‘validated’ mouse models of human cancer. Curr Opin Cell Biol 13, 778–784. Van, D. T. and Jacks, T. (2002) Cancer modeling in the modern era: progress and challenges. Cell 108, 135–144. Berger, M. R. (1999) Is there relevance for anticancer drug development, in relevance of tumor models for anticancer drug development, Chapter 54, Fiebig, H. H., Burger, B. A. (eds.) Karger Publishing, Basel, pp 15–27. Talmadge, J. E., Lenz, B. F., Klabansky, R., Simon, R., Riggs, C., Guo, S., Oldham, R. K. and Fidler, I. J. (1986) Therapy of autochthonous skin cancers in mice with intravenously injected liposomes containing muramyltripeptide. Cancer Res 46, 1160–1163. Schwartz, B., Birk, Y., Raz, A. and Madar, Z. (2004) Nutritional-pharmacological combinations–a novel approach to reducing colon cancer incidence. Eur J Nutr 43, 221–229. Schipper, H., Goh, C. R. and Wang, T. L. (1995) Shifting the cancer paradigm: must we kill to cure? J Clin Oncol 13, 801–807. Schuh, J. C. (2004) Trials, tribulations, and
Models of Metastasis in Drug Discovery
33.
34.
35.
36.
37.
trends in tumor modeling in mice. Toxicol Pathol 32 Suppl 1, 53–66. Seemann, M. D., Beck, R. and Ziegler, S. (2006) In vivo tumor imaging in mice using a state-of-the-art clinical PET/CT in comparison with a small animal PET and a small animal CT. Technol Cancer Res Treat 5, 537–542. Donelli, M. G., Rosso, R. and Garattini, S. (1967) Selective chemotherapy in relation to the site of tumor transplantation. Int J Cancer 2, 421–424. Willmanns, C., Fan, D., O’Brian, C., Radinsky, R., Bucana, C., Tsan, R. and Fidler, I. (1993) Modulation of doxorubicin sensitivity and level of P-glycoprotein expression in human colon carcinoma cells by ectopic and orthotopic environments in nude mice. Int J Oncol 3, 413–422. Bibby, M. C. (2004) Orthotopic models of cancer for preclinical drug evaluation: advantages and disadvantages. Eur J Cancer 40, 852–857. Kerbel, R. S. (2003) Human tumor xenografts as predictive preclinical mod-
38.
39.
40.
41.
233
els for anticancer drug activity in humans: better than commonly perceived-but they can be improved. Cancer Biol Ther 2, S134–S139. Fidler, I. J. (1986) Rationale and methods for the use of nude mice to study the biology and therapy of human cancer metastasis. Cancer Metastasis Rev 5, 29–49. Tan, M. H., Holyoke, E. D. and Goldrosen, M. H. (1977) Murine colon adenocarcinoma: syngeneic orthotopic transplantation and subsequent hepatic metastases. J Natl Cancer Inst 59, 1537–1544. Balmain, A. (2002) Cancer as a complex genetic trait: tumor susceptibility in humans and mouse models. Cell 108, 145–152. Siemann, D. W. (1987) Satisfactory and unsatisfactory tumor model factors influencing the selection of a tumor model for experimental evaluation, in rodent tumor models in experimental cancer therapy, Kallman, R. F. (ed), Pergamon Press, New York, pp 12–15.
Chapter 14 Methods for Evaluating Effects of an Irinotecan + 5-Fluorouracil/Leucovorin (IFL) Regimen in an Orthotopic Metastatic Colorectal Cancer Model Utilizing In Vivo Bioluminescence Imaging David Surguladze, Philipp Steiner, Marie Prewett, and James R. Tonra Abstract In testing novel anticancer therapies, researchers strive to utilize models that reflect the human disease as much as feasible. In this regard, orthotopic models are frequently developed because cancer cells in these models form tumors in, and metastasize from, a tissue environment similar to the tissue of origin of the cancer cells. Here we adapted an orthotopic colorectal cancer model, in which HT-29 colorectal cancer cells form tumors in the rectal lining and metastasize to the para-aortic lymph nodes with high frequency. Firefly luciferase-expressing HT-29 cells were used in this model to realize the benefits of bioluminescence imaging (BLI). A combination of irinotecan, 5-fluorouracil (5-FU), and leucovorin (LV) (IFL) was used as a standard chemotherapeutic regimen positive control. BLI allowed for the demonstration of the effects of IFL on tumor growth in the rectal lining, with tumor weight measurements at the end of the study reflecting total tumor burden. BLI also allowed relatively easy demonstration of reduced tissue metastasis with IFL treatment, compared to more time-consuming histological techniques. It is concluded that the orthotopic colorectal cancer model approach described represents a valuable tool for validating treatment strategies in this indication. Key words: HT-29, metastatic colorectal cancer, irinotecan, 5-FU, leucovorin, bioluminescence imaging, lymph node metastasis.
1. Introduction The preclinical evaluation of novel therapies for the treatment of cancer is dominated by studies tracking the growth of subcutaneous xenograft tumors with calipers in immunodeficient mice. G. Proetzel, M.V. Wiles (eds.), Mouse Models for Drug Discovery, Methods in Molecular Biology 602, DOI 10.1007/978-1-60761-058-8 14, © Humana Press, a part of Springer Science+Business Media, LLC 2010
235
236
Surguladze et al.
In these models, tumor cells are typically injected into the flank region where they generally do not form spontaneous metastasis to other tissues (1). In contrast, injecting tumor cells (2) or tumor fragments (3) into orthotopic sites provides a tissue environment more similar to the tissue of origin of the cancer cells and increases the frequency of spontaneous metastasis. Orthotopic models may, therefore, offer a more clinically relevant model to study treatment effects on metastasis, as well as primary tumor growth (4–6). Formation of regional lymph node metastasis can be an important step in dissemination of cancer cells. In colorectal cancer, lymph node metastasis frequently occurs in patients (7, 8) and is an important factor in staging the disease. In particular, the metastatic lymph node ratio (LNR; number of metastatic lymph nodes/number of examined lymph nodes) is predictive of overall survival (OS) and disease-free survival (DFS) in colorectal cancer patients (9, 10). Hence, an animal model of colorectal cancer with measurable lymphatic metastasis that allows for rapid evaluation of the effects of candidate treatment regimens on primary tumor growth and lymph node metastasis would be of great value. An orthotopic model of mCRC with a high rate of metastasis to the para-aortic lymph nodes has been developed utilizing implantation of tumor cells into the submucosal layer of the rectum (11, 12). In these models, primary tumor burden is evaluated by measuring tumor volume with a caliper, while lymph node metastasis is evaluated histologically. To achieve rapid and sensitive detection of tumor cells in tissues, several new imaging modalities have recently been introduced, including Bioluminescence Imaging (BLI). BLI technology is based on the detection of firefly luciferase protein expression by cells. The emitted signal is captured by a highly sensitive cooled CCD camera mounted in a light-tight specimen box. The advantages of BLI compared to conventional methods have been demonstrated successfully in animal models of various tumor types (13, 14). BLI allows detection of tumor cell burden even when the tumor is not visible or palpable. Moreover, tumor cell burden can be measured noninvasively over time with BLI in the same animal, at an internal site. BLI is also sensitive enough to detect micro-metastasis in tissues, and in some cases obviates the need for the more time-consuming and labor-intensive histological determination of the incidence of metastasis or metastatic burden. We have developed a firefly luciferase-expressing HT-29 colorectal cancer cell line, HT-29LP, to realize the benefits of BLI with an orthotopic colorectal cancer model. Primary tumors were established by implantation of HT-29LP cells into the submucosal layer of the rectum of mice, and tumor growth and lymph node metastasis were evaluated. A positive control treatment
Methods for Evaluating Effects of an (IFL) Regimen
237
was sought to demonstrate the ability of this model to establish the potential clinical relevance of novel treatment strategies. For metastatic colorectal cancer, 5-fluorouracil (5-FU) and leucovorin (LV) combination remains the backbone of standard chemotherapeutic regimens. Since adding irinotecan (CPT-11) to 5-FU/LV (IFL) increased patient response rate and survival compared to 5-FU/LV alone (14), IFL is considered to be among the current standard treatment options for colorectal cancer. Thus, an IFL regimen was evaluated as a positive control in the HT-29LP orthotopic model of colorectal cancer.
2. Materials 2.1. Generating LuciferaseExpressing HT-29 Cells
1. pGL3-Basic vector containing (Promega, Madison, WI).
the
luciferase
gene
2. Restriction enzymes SmaI, XbaI, and HpaI and modifying enzymes Klenow Fragment, calf intestinal alkaline phosphatase, and T4 DNA ligase (New England Biolabs, Ipswich, MA). 3. pLXSN retroviral plasmid (Clontech, Mountain View, CA). 4. RetroPack PT67 amphotropic packaging cell line (Clontech). 5. Dulbecco’s Modified Eagle Medium (DMEM), fetal bovine serum (FBS), and TrypLE Express Stable TrypsinLike Enzyme with Phenol Red (Invitrogen, Carlsbad, CA). 6. OptiMEM and Lipofectamine 2000 (Invitrogen). 7. Human HT-29 colorectal adenocarcinoma cell line (ATCC, Manassas, VA). 8. McCoy’s 5A medium (Invitrogen). 9. Polybrene sold as hexadimethrine bromide (Sigma, St. Louis, MO). 10. Geneticin (G418, Neomycin), 50 mg/ml solution in distilled water (Invitrogen, Carlsbad, CA). 11. Scienceware sterile cloning discs, 3.0 mm diameter (VWR, West Chester, PA). 12. Luciferase Assay System, 100 assays, E1500 (Promega, Madison, WI). 13. NEBuffer 4 (New England Biolabs, Inc., Ipswich, MA). R 14. Sterile 0.45 m Acrodisk syringe filters (Pall Life Sciences, Ann Arbor, MI).
238
Surguladze et al.
2.2. Cell Culture
1. Cell growth medium: One liter McCoy’s 5A Medium (Invitrogen) supplemented with 11 ml GlutaMAX (Invitrogen) and 110 ml FBS (HyClone, Logan, UT). Geneticin (Invitrogen) is also included in the medium at 0.8 mg/ml final concentration. 2. Dimethyl sulfoxide meeting USP testing specifications (Sigma, St. Louis, MO). 3. 0.9% sodium chloride for Irrigation, USP saline (B. Braun Medical Inc., Irvine, CA). 4. Corning CellSTACK Culture Chambers, growth area 6360 cm2 (Corning Life Sciences, Lowell, MA; Cat. #3270). 5. Trypan blue stain (Invitrogen). 6. Hemocytometer (VWR). 7. Polypropylene 15 and 50 ml tubes (BD Falcon, Franklin Lakes, NJ). 8. Cryogenic freezing tubes (Sarstedt, Newton, NC). 9. Erlenmeyer flasks 1000 ml (Corning Life Sciences). 10. Corning sterile polypropylene 250 ml tubes (Corning Life Sciences). 11. Clorox bleach (Office Depot).
2.3. In Vivo Efficacy Evaluation
1. Nu/Nu (Crl:NU/NU-nuBR) athymic female mice, 6 to 7-week-old (Charles River Laboratories, Wilmington, MA) (see Note 1). 2. Matrigel Basement Membrane Matrix (BD Biosciences, San Jose, CA). 3. Dulbecco’s phosphate-buffered saline (DPBS) (Invitrogen). 4. Individually packaged sterile syringes 1 cc (Becton Dickinson, Franklin Lakes, NJ). 5. Polypropylene hub stainless steel needles 27 G (Becton Dickinson). 6. IVIS 200 in vivo imaging system (Xenogen, Alameda, CA). 7. V-10 Mobile Cart System (VetEquip, Pleasanton, CA). 8. Micro Dissecting Scissors (Roboz Surgical Instrument Co., Inc., Gaithersburg, MD). 9. Hemostats (Roboz Surgical Instrument Co., Inc.). 10. Isoflurane 250 ml (Burns Veterinary Supply Inc., Westbury, NY).
Methods for Evaluating Effects of an (IFL) Regimen
11.
239
D -luciferin Firefly, potassium salt, 1.0 g/vial (Xenogen/ Caliper Life Sciences, Hopkinton, MA).
R syringe filters (Pall Life 12. Sterile 0.20 m Acrodisk Sciences, Ann Arbor, MI).
13. 96-well polypropylene plate (BD Falcon). 14. 5-fluorouracil (5-FU; Sigma). 15. Leucovorin (LV; folinic acid, calcium salt) (Sigma). 16. Irinotecan (CPT-11; LKT Laboratories, St. Paul, MN). 17. Living Image 2.50.1 Software (Xenogen/Caliper Life Sciences, Hopkinton, MA). 18. JMP7 Statistical Discovery Software (SAS Institute, Cary, NC).
3. Methods 3.1. Generating LuciferaseExpressing HT-29 Cells 3.1.1. Cloning of the Luciferase Gene into a Retroviral Plasmid
1. Digest 5 g of the pGL3-Basic vector with 1 l of the restriction enzymes SmaI and XbaI in 10 l of 1× NEBuffer 4 in an Eppendorf tube at 25◦ C for 4 h. Raise the temperature to 37◦ C and incubate overnight. This releases a 1714 bp fragment containing the luciferase gene. Add 1 l of the enzyme Klenow Fragment to fill-in the 5 overhang created by XbaI. This results in a DNA fragment with 5 and 3 blunt ends. 2. Digest 1 g of the pLXSN retroviral plasmid with 1 l of the restriction enzyme Hpal in 10 l of 1× NEBuffer 4 in a 1.5 ml Eppendorf tube at 37◦ C overnight. This produces a linear 5.9 kb backbone fragment with blunt ends. Add 1 l of calf intestinal alkaline phosphatase to the tube which removes 5 and 3 phosphoryl groups. This prevents the plasmid from self ligation. 3. Incubate 5 l of the purified luciferase insert (Smalblunt/Xbal-blunt) with 1 l of the linear pLXSN plasmid (Hpal-blunt) in the presence of 1 l T4 DNA ligase and 1 mM ATP and 3 l of water at 16◦ C overnight to create the retroviral plasmid pLXSN-luc. Measure DNA concentration with a spectrophotometer at 260 nm excitation, and check purity of DNA using 260/280 ratio. Verify correct orientation of the luciferase gene by enzyme restriction digest. Luciferase is expressed from the viral 5 long terminal repeat (LTR) sequence and the neomycin resistance marker gene is expressed from the SV40 protomer.
240
Surguladze et al.
3.1.2. Production of a Replication-Incompetent Retrovirus Containing the Luciferase Gene
1. Grow 1 × 107 cells of the RetroPack PT67 packaging cell line in T75 flasks in DMEM (10% FBS). PT67 cells are harvested with 2 ml TrypLE. Express from a confluent culture and resuspended in 3 ml DMEM (10% FBS) to inactivate the TrypLE Express. Prepare two new T75 flasks with DMEM (10% FBS) and seed 1.5 × 106 cells per flask. 2. The next day, transfect one flask containing PT67 cells with the retroviral plasmid pLXSN-luc. The second flask is used for a mock transfection which is a negative control transfection in the absence of a retroviral plasmid. Mix 3.0 ml serumfree OptiMEM with 120 l Lipofectamine 2000 under aseptic conditions in a biosafety cabinet and incubate for 5 min at room temperature. In parallel, prepare two vials with 1.5 ml serum-free OptiMEM. Add 30 g of pLXSN-luc to the first vial but no DNA to the second vial (mock). Transfer 1.5 ml from the tube with the OptiMEM/Lipofectamine 2000 mix to both tubes and incubate for 20 min. Add 3.0 ml from the vials with the pLXSN-luc or mock transfection solutions to the flasks with PT67 cells. Mix gently and incubate the flasks in a tissue culture laboratory designated for biosafety level 2 (BSL-2) work (see Note 2). 3. The next day, replace the medium with 8.0 ml of fresh DMEM (10% FBS). 4. After incubating for 3 days, transfer the culture medium containing the soluble retroviral particles (luciferase and mock) from the flasks to new 15 ml conical tubes. Since the medium might be contaminated with floating retrovirusproducing PT67 cells, it is important to remove these through filtration. The medium is pulled from the tubes into 10 ml syringes without needles. Attach a 0.45 m sterile filter to the syringe and force the medium with the retroviral particles through the filter into new tubes. The solutions containing the luciferase and mock retroviral particles can immediately be used for infections of the HT-29 cells or stored in 3.0 ml aliquots at −80◦ C. Repeated freezing and thawing reduces the titer of the retrovirus and should be avoided.
3.1.3. Infection of HT-29 Cells with a Luciferase-Expressing Retrovirus
1. Grow the HT-29 cell line in T75 flasks in McCoy’s 5A medium (10% FBS). When cells are confluent, harvest HT29 cells by adding 2.0 ml of TrypLE Express. After the cells detach from the flasks, resuspend the cells in McCoy’s medium (10% FBS) to inactivate the TrypLE Express. Count the cell concentration as described below. Prepare two small T25 flasks with McCoy’s medium (10% FBS) and seed 1 × 106 cells per flask. Keep the HT-29 cells in culture as they will be needed again later.
Methods for Evaluating Effects of an (IFL) Regimen
241
2. The next day, infect the HT-29 cells with luciferase or mock retroviral particles. A 2.5 ml solution containing retroviral particles is either prepared fresh or obtained from a frozen aliquot (see above). Prepare polybrene as 100× stock solution (0.8 mg/ml) in 1 × PBS (can be stored at −20◦ C). Add polybrene to the tubes containing the retroviral particles at a final concentration of 8 g/ml. Aspirate the medium from the HT-29 cells in the two T25 flasks. Initiate the infection by adding 2.5 ml of the solutions containing the luciferase or mock retroviral particles with polybrene to both flasks. 3. The next day, the infected HT-29 cells are passaged onto new tissue culture plates. Aspirate and discard the retroviruscontaining medium and wash the cells once with 1 × PBS. Harvest the cells with 0.75 ml TrypLE Express and resuspended in 2 ml McCoy’s medium (10% FBS). Plate 0.5 ml of the HT-29 cells infected with luciferase-expressing retrovirus on a 10 cm plates (low density). Plate 2.0 ml of the same cells on another 10 cm plate (high density). Repeat with the HT-29 cells that were infected with mock retrovirus. Plating the cells at a low and high density will help later when selecting neomycin-resistant colonies. 3.1.4. Selection of an HT-29 Cell Line Stably Transduced with the Luciferase Gene
1. One day after passaging the infected HT-29 cells onto new plates, start the selection process by adding geneticin directly to the medium on all four plates at a final concentration of 800 g/ml. 2. Replace the medium every 3 days with fresh McCoy’s medium (10% FBS) supplemented with 800 g/ml of geneticin for the next 2 weeks. 3. Luciferase-infected HT-29 cells start to form small colonies after 10 days of selection with geneticin. Mock-infected HT29 cells (negative control) start to detach from the plates after about 4 days and completely disappear after 10 days in the presence of geneticin. 4. After 3 weeks, the HT-29-luciferase colonies contain about 200–500 cells and are ready to be cloned. Pick one of the two plates where the colonies are not yet touching each other. Incubate 20 sterile cloning disks in 1 ml of TrypLE Express in a 10 cm tissue culture plate for 2 min. Aspirate the medium from the plate with the HT-29 luciferase colonies and carefully wash once with 1 × PBS. Using sterile tweezers put a moist cloning disk on top of a colony. Repeat with at least 20 colonies since not all colonies will grow after selection. After 8 min, lift cloning disks with HT-29 luciferase cells attached to them from the plate and transfer to separate wells of a 24-well tissue culture plate with 500 l of McCoy’s medium (10% FBS) containing 800 g/ml of geneticin.
242
Surguladze et al.
Passage cells to a 6-well plate, then to a T75 flask once confluent. 5. Once at least three stable HT-29-luciferase lines are established, determine luciferase enzymatic activity in cell extracts obtained from each line using the Luciferase Assay System from Promega by following the manufacturer’s instructions. Using this assay, a specific luciferase signal in the range of 107 –108 luciferase counts/mg protein can be expected. 6. Before moving the HT-29-luciferase cells from the BSL-2 lab back to a regular tissue culture lab, verify that the cells do not secrete infectious retroviral particles: a) Collect the culture medium from HT-29-luciferase cells grown in the absence of geneticin and filter through a 0.45 m filter as described above. Transfer 10 ml of the filtered medium onto a plate with parental HT-29 cells that were seeded at low density (see Section 3.1.3 Step 3). b) After 3 days, add geneticin at 800 g/ml to start the selection. Replace medium with new geneticin every 3 days for 14 days. c) If none of the parental HT-29 cells survive the selection process, the HT-29-luciferase cultures can be moved out of a BSL-2 lab. If the parental HT-29 cells form colonies in the presence of geneticin, the HT-29-luciferase cells most likely produced infectious retroviral particles. These cells should not be used in further studies and should be discarded. 3.2. Cell Culture 3.2.1. Freezing HT-29LP cells in aliquots for establishing a consistent supply
1. Aspirate HT-29LP medium (McCoy’s 5A, supplemented with 1% glutaMAX, 10% FBS, and 0.8 mg/ml geneticin), from adherent HT-29LP cells growing in a polystyrene flask or dish and add 2 ml TrypLE Express. Move the solution around the plate and quickly aspirate. 2. Add 1–1.5 ml TrypLE Express for T150 (150 cm2 ) tissue culture flask (or equivalent 150 × 25 mm tissue culture dish) or 0.5 ml for T75 tissue culture dish and make sure solution covers the cells. 3. Incubate plates at 37◦ C until cells slough off (about 5– 10 min). 4. Tap plate to get cells to one side and add 3 ml medium. Pipette cells up and down and wash plate with cell suspension. 5. Collect cells into 50 ml polypropylene conical tube and centrifuge cells at 1000 rpm (180 RCF) for 7 min.
Methods for Evaluating Effects of an (IFL) Regimen
243
6. Carefully aspirate supernatant and suspend cells in freezing buffer = 95% FBS, 5% DMSO (freezing buffer stored at 4◦ C). Add 4 ml freezing buffer for T150 flask or 2 ml for T75 flask. If cell number is known, make to about 5 × 106 cells/ml. 7. Make 1 ml aliquots into 2 ml cryogenic freezing tubes labeled with a lot #. 8. Put tubes in –80◦ C freezer for 24 h. 9. Place frozen tubes into liquid nitrogen for storage. 3.2.2. Thawing and Growing Cells
1. Record lot number of thawed tumor cells. 2. Put cell tubes into 37◦ C water bath for as brief a time as possible to defrost. 3. Resuspend cells by flicking tube and transfer into 5 ml of HT-29LP medium with geneticin in a 50 ml polypropylene Falcon tube. 4. Centrifuge cells at 1000 rpm for 7 min at 180 RCF. 5. Aspirate supernatant and resuspend in 20 ml medium in a 100 mm diameter Petri dish (culture dishes or plates are used when extra ventilation may improve the health of the cells, but for long-term culture, flasks should be used to protect from infection). 6. Incubate overnight at 37◦ C, 5% CO2 . 7. Verify cells have attached to the bottom of flask by tilting the flask to one side. If cells have attached, aspirate the supernatant. 8. Add 20 ml of fresh room temperature medium. 9. Culture at 37◦ C, 5% CO2 .
3.2.3. Feeding and Splitting Cells
1. Replace tissue culture medium by aspirating old medium and adding new 32–37◦ C medium (15–20 ml T75, 25–35 ml T150) every 2–3 days until cells reach 80% confluence. 2. When cells reach 80% confluence, aspirate medium. 3. Add 2 ml TrypLE Express, move it around the plate and quickly aspirate. 4. Add 1–1.5 ml TrypLE Express for T150 size and 0.5 ml for T75 and make sure solution covers the cells. 5. Incubate plates at 37◦ C until cells slough off (about 5 min). 6. Tap plate to get cells to one side and add medium to a total volume of 4–10 ml for T75 or T150 depending on cell concentration and growth rate (and desired time to next trypsinization). Pipette cells up and down and wash plate with cell suspension.
244
Surguladze et al.
7. Make dilutions (e.g., 1:4 of 10 ml total = 2.5 ml) into 25–35 ml fresh medium into a new T150, or 15–22 ml into a new T75. 8. Insure that cells are distributed evenly in plate, by rocking back and forth. 9. Incubate plates at 37◦ C, 5% CO2 for cells to grow. 3.2.4. Seeding cells into Corning 10-Tray CellSTACK
1. Prepare approximately 1.2 l of HT-29LP medium at room temperature in the medium bottle. 2. Collect cells from four 150 mm plates as outlined in Steps 1–6 for splitting cells above. Pipette collected cell suspension into one of the medium bottles (prepared in Step 1) to be used in the CellSTACK. 3. Place CellSTACK unit in the cell culture hood and remove one vent cap. Slowly pour the cell suspension and medium into the unit. Replace vent cap. 4. Lay unit on its side. Tilt unit back and forth at approximate 45◦ angles three times to mix the cell suspension. Place unit on its side again to let the suspension settle evenly among the 10 trays. Turn upright. 5. Incubate at 37◦ C, 5% CO2 for cells to grow.
3.2.5. Harvesting Cells from a Corning CellSTACK for Injection
1. Prepare a 4 l beaker containing ∼200 ml chlorine bleach. Prepare approximately 500 ml USP saline, an entire 100 ml bottle of TrypLE Express, and a sterile 500 ml Erlenmeyer collection flask containing 25 ml FBS all at room temperature. 2. Place CellSTACK in the hood and remove one cap. Pour off spent medium into the 4 l beaker. 3. Add ∼500 ml sterile USP saline. Close opening with cover cap. 4. Tilt unit back and forth at approximate 45◦ angles three times to rinse the cell surface. Remove caps and pour off into the 4 l beaker. 5. Pour in an entire 100 ml bottle of TrypLE Express at room temperature. Replace cap and place unit on its side to evenly distribute TrypLE Express solution. Remove unit from hood and place at 37◦ C to detach cells. 6. Check for cell detachment by holding unit up to the light. Return unit to hood, remove cap, and pour off cell suspension into the collection flask containing FBS. 7. Rinse unit twice by adding 50–100 ml medium for HT29LP cells without geneticin, swishing to rinse surfaces, and pouring into the collection flask.
Methods for Evaluating Effects of an (IFL) Regimen
245
8. Place the cell suspension into 250-ml Corning sterile polypropylene tubes. 9. Centrifuge at 1000 rpm (200 RCF) for 10 min. Aspirate off the supernatant. Resuspend cells by flicking bottom of tube. Add ∼50 ml of medium and put on ice. 3.3. In Vivo Efficacy 3.3.1. Counting Cells for Implantation
1. Add 0.5 ml cell suspension to 0.5 ml 0.4% trypan blue in a 15 ml polypropylene tube, invert, and incubate for 1–2 min. Add 4 ml of saline or PBS. 2. Count cell concentration using a hemocytometer. Count 4, 16 square areas (2 on both sample chambers) and calculate cell concentration in suspension = (total #cells/4) × 105 cells/ml. If necessary dilute cells another 1:10 in saline or PBS and count cell concentration. 3. Add medium (McCoy’s 5A, supplemented with 1% GlutaMax and 10% FBS without geneticin) to bring HT-29LP cells to the final desired concentration. For the utilized dose of 1 × 106 cells/mouse, the cells should be brought to 20 × 106 cells/ml. Put the cells on ice to bring to approximately 4◦ C. 4. Add an equi-volume (1:1) of cold Matrigel to the cell suspension. Invert the tube a few times. It is very important to keep cell tubes on ice from this point onward.
3.3.2. Inhalation Anesthesia method
1. When using an anesthesia machine, such as V-10 Mobile Cart System described here, ensure all carrier gas lines are securely attached and all chambers, breathing circuits, and nose cones are firmly in place (see Note 3). 2. Fill the vaporizer with isoflurane while the dial is in the “OFF” position. 3. Open the stopcock in the chamber tubing line and close the stopcock in the nose cone line. 4. Turn on oxygen tank and adjust oxygen flow rate to 1 liter per minute (1 lpm). 5. Place animal(s) in the induction chamber and secure the lid. 6. Turn on the isoflurane flow to the chamber to anesthetize mice and open flow to the chamber by turning the stopcock. Set the vaporizer dial to 3%, and animal(s) should begin to fall asleep within 1–2 min. 7. Breathing rate usually shows an initial increase. When breathing rate becomes near normal, and there is no reaction to rocking the chamber, push the “FLUSH” button on the machine for approximately 10 s.
246
Surguladze et al.
8. Immediately open the chamber and remove animal(s). Close the stopcock for the induction chamber and open the stopcock in the nose cone line. 9. Place animal(s) nose in the nose cone and turn the vaporizer dial to 1.5%. 10. When the procedure is completed remove animal(s) from anesthesia and mice should begin to wake within 1 min. 11. Repeat Steps 5–10 for each animal (see Note 4). 12. When all surgery is complete, turn off the vaporizer dial, the oxygen flowmeter, and oxygen tank. 3.3.3. Establishing Orthotopic HT-29LP Tumors in Nu/Nu Athymic Mice
1. If cells are clumped, disrupt the suspension via aspiration or gentle vortexing. 2. At least two 1 cc disposable polyethylene syringes should be used and kept on ice throughout the injection process. 3. Lay the anesthetized mouse on its back without restraint and with its tail toward the investigator. 4. Do not begin the procedure until mouse is areflexic (does not respond to pinching toes or touching an eye). 5. To prevent bleeding, double clipping of the anterior wall in the anorectal region is performed using hemostats (Fig. 14.1A).
Fig. 14.1. (A) Implantation of HT-29LP tumor cells into the posterior wall of the rectum. The anterior wall of the anorectal area is cut 7 mm in length between two hemostats to prevent colonic obstruction, resulting from tumor progression. Tumor cells are then injected submucosally using a 27 G needle. (B) At the end of the study period, the abdominal cavity is exposed through a midline incision and para-aortic lymph nodes (arrow), located around the abdominal aorta, are removed and imaged ex vivo.
Methods for Evaluating Effects of an (IFL) Regimen
247
6. The anorectal wall is cut 7 mm in length between the two hemostats to prevent colonic obstruction, resulting from rectal tumor progression (Fig. 14.1A). 7. Fill one syringe with 0.1 ml cell suspension, inject, place empty syringe on ice, remix cell suspension, and fill a second syringe for injection. Alternate syringes to keep them cold and avoid having the Matrigel solidify inside. 8. Use a 27 G needle, 1–1/2 in. needle, bevel parallel to the rectal wall, and insert into the submucosal layer of the rectum (Fig. 14.1A). 9. To prevent cell leakage, allow the needle to remain in the injection site for approximately 5–10 s after cell implantation, and then slowly remove the needle. This should prevent cell leakage, allow the Matrigel to harden, and minimize inconsistency in tumor growth. 3.3.4. Tumor Growth Measurement
1. Prepare a stock solution of D-luciferin at 15 mg/ml concentration in DPBS. Filter sterilize through a 0.2 m filter. Prepare enough to inject 10 l/g of body weight. Each mouse should receive 150 mg D-luciferin/kg body weight. For example, for a 30 g mouse, inject 300 l of 15 mg/ml stock to deliver 4.5 mg of luciferin (see Note 5). 2. Inject animals by an intraperitoneal i.p. route with Dluciferin solution (150 mg/kg, in DPBS). Subsequently, place animals into the plexiglass anesthesia box (XGI-8, Gas Anesthesia system connected to the IVIS system) and set the vaporizer dial to 2.5–3.5% isoflurane. 3. After the mice are fully anesthetized, transfer the mice from the induction chamber to the nose cones attached to the manifold and place them on black paper inside the imaging chamber of the IVIS system. 4. At 12 min (see Note 6) after D-luciferin injection, image animals in the dorsal position (laying on back) for 2–5 s at field of view (FOV) 23 (E), with medium binning selected in the IVIS menu (see Note 7). 5. Apply regions of interest (ROI) from displayed images around the tumor sites and quantify emitted signals as total photons/s using Living Image Software.
3.3.5. Randomize into Treatment Groups
1. When animals develop visible tumors at the site of cell implantation, approximately 12 days after cell implantation, obtain bioluminescence images of the animals in the ventral position using the IVIS system (Fig. 14.2A). Remove outlier animals with tumors larger or smaller compared to the majority of mice and randomize the remaining mice by bioluminescence intensity into one of the following two i.p.
248
Surguladze et al.
Fig. 14.2. In vivo monitoring of HT-29LP tumor growth and ex vivo bioluminescence imaging of lymph nodes. (A) Ventral images of representative anesthetized mice are shown on day 1 and day 35. Animals in the saline control group showed progressive tumor development over time following tumor cell implantation, indicated by increasing bioluminescence signal at the tumor site. BLI signal in the IFL treated animals were lower compared to the saline control group at the end of the study (day 35), indicating tumor growth inhibition. (B) Ex vivo images of para-aortic lymph nodes indicated that 11 of 13 animals (85%) developed lymph node metastasis in the saline control group. Treatment with IFL significantly reduced the frequency of metastasis (27%) versus the control (P = 0.002, chi-squared test).
treatment groups (n = 13–15 mice per group): USP saline at 20 l/gram body weight or IFL regimen: Irinotecan (CPT11) at 100 mg/kg in combination with 5-fluorouracil (5FU) at 60 mg/kg and leucovorin (LV) at 30 mg/kg (see Note 8). 2. Inject USP saline or IFL regimen once a week intraperitoneally (IP) for 35 days. 3. Record BLI signal intensity, quantified by Living Image 2.50.1 software, before and at the end of treatment (Fig. 14.3A).
Methods for Evaluating Effects of an (IFL) Regimen
249
Fig. 14.3. Treatment effect on tumor growth. (A) Treatment with IFL significantly (P = 0.0009, RM ANOVA) inhibited the growth of HT-29LP tumors as measured by bioluminescence imaging. (B) Tumor weight measurements confirmed antitumor effects of IFL. Mean tumor weight in the IFL group was significantly (P < 0.0001, one-way ANOVA) lower compared to the control.
4. At the end of the study period following lymph node dissection (see below), surgically remove primary tumors from the animals, isolate from surrounding tissues, and weigh (Figs. 14.1B and 14.3B). 5. Perform statistical analysis utilizing repeated measures ANOVA (JMP software, Version 5 from SAS Institute, Cary, NC) to evaluate the treatment effects on tumor growth. Compare tumor weight measurements between treatment groups utilizing one-way ANOVA. 6. Utilizing the methods described above, in the HT-29LP orthotopic model, animals in the control group showed progressive tumor development through the 34th day of treatment, indicated by increasing bioluminescence signal (BLI) at the tumor site (Fig. 14.2A). Treatment with IFL significantly (P = 0.0009, RM ANOVA) inhibited the growth of HT-29LP tumors as measured by bioluminescence imaging (Fig. 14.3A). Tumor weight measurements at day 35 of treatment confirmed the antitumor effect of IFL (Fig. 14.3B). Mean tumor weight in the IFL group was significantly lower than in the control group (P < 0.0001, oneway ANOVA). 3.3.6. Verifying the Presence of Metastasis Ex Vivo at Study Termination
1. Twenty-four hours after the last in vivo bioluminescence imaging, inject the animals i.p. with 150 mg/kg D-luciferin and euthanize them 2 min later. 2. Expose the abdominal cavity through a midline incision and locate para-aortic lymph nodes (Fig. 14.1B). Subsequently, surgically remove the lymph nodes and place them in DPBS containing 300 g/ml D-luciferin in 96-well polypropylene plates.
250
Surguladze et al.
3. Place plates into the IVIS imaging chamber, and image tissues for 1–2 min, at a platform height setting of level C in the imaging chamber (Fig. 14.2B). 4. Quantify emitted BLI signals from lymph node metastasis as total photons/sec using Living Image software. 5. Record the frequency of para-aortic lymph node metastasis in treatment groups. Evaluate treatment effect on the frequency of para-aortic lymph node metastasis using a chisquared test. 6. Utilizing the methods described above, ex vivo images of para-aortic lymph nodes indicated that 11 of 13 animals (85%) developed lymph node metastasis in the saline control group. Treatment with IFL significantly reduced the frequency of metastasis (27%) versus control (P = 0.002, chisquared test) (Fig. 14.2B).
4. Notes 1. This immunodeficient nude mouse was originally thought to be a BALB/c congenic. It was later determined that it was not inbred and is maintained as an outbred. It is expected that other strains of Nu/Nu mice could be used for this model (12). 2. When working with infectious retroviral particles, consider the following safety points: a. Always wear lab gowns, safety goggles and disposable gloves. b. Keep two small containers in the biosafety cabinet: one with 10% bleach to flush used pipettes and one with a small biohazard bag for solid waste. Close the bag in the biosafety cabinet and discard with the regular biohazard waste. Empty the 10% bleach with the used pipettes after each session. c. Turn on the UV light for 30 min after completion of work in the biosafety cabinet. 3. Animal anesthesia can be achieved with isoflurane without the use of specialized anesthesia equipment. Place animal(s) in a bell jar or screw top glass jar containing a cotton wool pad soaked with 1–2 ml of isoflurane until areflexic. After removal of the animal(s) from the jar, place the nose in a 50 ml Eppendorf tube with an isoflurane soaked cotton pad at the bottom. This method requires frequent evaluation of the depth of anesthesia.
Methods for Evaluating Effects of an (IFL) Regimen
251
4. Alternative methods of anesthesia include injectable anesthesia that is achieved by Avertin (200 mg/kg) or ketamine/xylazine (100/10 mg/kg) i.p. injections. However, it is difficult to control the depth of anesthesia when using injectable anesthetics. Additionally, it takes longer to induce or recover from these anesthetics compared to inhalation methods. 5. The stock solution (15 mg/ml) of luciferin can be transferred to smaller vials, each vial containing enough for a single day of in vivo measurements. These vials can be kept at −20◦ C for up to 6 months. 6. For luciferase-expressing models, a luciferin kinetic study needs to be performed prior to efficacy studies. For the HT29LP orthotopic model, animals were imaged in sequence mode with a 2-min delay between the images utilizing the IVIS system. The maximum BLI signal was detected at 12 min after D-luciferin injection and this time point was therefore selected for evaluating tumor burden in this model. 7. IVIS 200 imaging system allows for the imaging of up to five animals at a time in one field of view (FOV) (23 cm). The set FOV should remain constant during the study period to obtain consistent images. 8. IFL regimen is given i.p. in the following manner: CPT-11 is dosed first at 100 mg/kg, followed 1 h later by 5-FU/LV dosed in a single dosing solution at 60/30 mg/kg.
Acknowledgment The authors would like to thank Jessica Kearney for her role in determining the Maximum Tolerated Dose (MTD) of IFL regimen in Nu/Nu mice. References 1. Fogh J., Orfeo T., Tiso J., Sharkey F.E., Fogh J.M., Daniels W.P. (1980) Twentythree new human tumor lines established in nude mice. Exp Cell Biol 48, 229– 239. 2. Morikawa K., Walker S.M., Nakajima M., Pathak S., Jessup J.M., Fidler I.J. (1988) Influence of organ environment on the growth, selection, and metastasis of human colon carcinoma cells in nude mice. Cancer Res 48, 6863–6871.
3. Hoffman R.M. (1999) Orthotopic metastatic mouse models for anticancer drug discovery and evaluation: a bridge to the clinic. Invest New Drugs 17, 343–359. 4. Fidler I.J. (1995) Critical factors in the biology of human cancer metastasis. Am Surg 61, 1065–1066. 5. Hoffman R.M. (1994) Orthotopic is orthodox: why are orthotopic-transplant metastatic models different from all other models? J Cell Biochem 56, 1–3.
252
Surguladze et al.
6. Kubota T. (1994) Metastatic models of human cancer xenografted in the nude mouse: the importance of orthotopic transplantation. J Cell Biochem 56, 4–8. 7. Lo D.S., Pollett A., Siu L.L., Gallinger S., Burkes R.L. (2008) Prognostic significance of mesenteric nodules in patients with stage III colorectal cancer. Cancer 112, 50–4. 8. Wu Z.Y., Wan J., Li J.H., Zhao G., Yao Y, Du J.L., et al. (2007) Prognostic value of lateral lymph node metastasis for advanced low rectal cancer. World J Gastroenterol 13, 6048–6052. 9. Berger A.C., Sigurdson E.R., LeVoyer T., Hanlon A., Mayer R.J., Macdonald J.S., Catalano P.J., Haller D.G. (2005) Colon cancer survival associated with decreasing ratio of metastatic to examined lymph nodes. J Clin Oncol 23, 8706–8712. 10. Schumacher P., Dineen S., Barnett C., Fleming J., Anthony T. (2007) The metastatic lymph node ratio predicts survival in colon cancer. Am J Surg 194, 827–832. 11. Kashtan H., Rabau M., Mullen J.B.M., Wong A.H.C., Roder J.C., Shpitz B. (1992) Intra-rectal injection of tumor cells: a novel
12.
13.
14.
15.
animal model of rectal cancer. Surg Oncol 1, 251–256. Tsutsumi S., Kuwano H., Morinaga N., Shimura T., Aso T. (2001) Animal model of para-aortic lymph node metastasis. Cancer Lett 169, 77–85. Jenkins D.E., Hornig Y.S., Oei Y., Dusich J., Purchio T. (2005) Bioluminescent human breast cancer cell lines that permit rapid and sensitive in vivo detection of mammary tumors and multiple metastasis in immune deficient mice. Breast Cancer Res 7, R444–R454. Jenkins D.E., Yu S.-F., Hornig Y.S., Purchio T., Contag P.R. (2003) In vivo monitoring of tumor relapse and metastasis using bioluminescent PC-3 M-luc-C6 cells in murine models of human prostate cancer. Clin Exp Metastasis 20, 745–756. Saltz L.B., Cox J.V., Blanke C., Rosen L.S., Fehrenbacher L., Moore M.J., Maroun J.A., Ackland S.P., Locker P.K., Pirotta N., Elfring G.L., Miller L.L. (2000) Irinotecan plus fluorouracil and leucovorin for metastatic colorectal cancer. Irinotecan Study Group. N Engl J Med 343, 905–914.
Chapter 15 CML Mouse Model in Translational Research Cong Peng and Shaoguang Li Abstract Chronic myeloid leukemia (CML) is a myeloproliferative disorder characterized by increased proliferation of granulocytic cells without the loss of their capability to differentiate. CML is derived from the hematopoietic stem cells (1) with the Philadelphia chromosome resulting from of a reciprocal translocation between the chromosomes 9 and 22 t(9;22)-(q34;q11). This translocation produces a fusion gene known as BCR-ABL which acquires uncontrolled tyrosine kinase activity, constantly turning on its downstream signaling molecules/pathways, and promoting proliferation of leukemia cell through antiapoptosis and acquisition of additional mutations. To evaluate the role of each critical downstream signaling molecule of BCR-ABL and test therapeutic drugs in vivo, it is important to use physiological mouse disease models. In this chapter, we describe a mouse model of CML induced by BCR-ABL retrovirus (MSCV-BCR-ABL-GFP; MIG-BCR-ABL) and how to use this model in translational research. Key words: Chronic myeloid leukemia (CML), Philadelphia chromosome (Ph+ chromosome), BCR-ABL, hematopoietic stem cell, retroviral mouse model, translational research.
1. Introduction Chronic myeloid leukemia (CML) has been extensively studied and used as a model disease to investigate the molecular basis of leukemia, shedding light on the understanding of other human cancers as well. CML validates the concept that cancer is a genetic disease. CML is derived from the hematopoietic stem cells which harbor the BCR-ABL oncogene and acquire the selective growth advantages over the normal hematopoietic stem cells. These conditions are drawn based on the transplantation experiments in which peripheral blood (PB) cells, total bone marrow (BM) cells, and primitive cells (CD34+) from the CML patients have been G. Proetzel, M.V. Wiles (eds.), Mouse Models for Drug Discovery, Methods in Molecular Biology 602, DOI 10.1007/978-1-60761-058-8 15, © Humana Press, a part of Springer Science+Business Media, LLC 2010
253
254
Peng and Li
transplanted into irradiated nonobese severe combined immunodeficient (NOD/SCID) mice (2). Recipient mice transplanted with these PB or BM cells showed engraftment of the human leukemia cells in BM for up to 7 months and those transplanted with CD34+ cells showed a greater engraftment of leukemia cells. Although this xenograft model allows evaluating the capability of transplanted human leukemia cells to initiate and maintain CML disease in recipient mice, the mice with engrafted human leukemia cells did not develop lethal leukemia after 7 months. This calls for further improvement of this xenograft model as it is important to establish a faithful CML mouse model for evaluating promising therapeutic compounds and developing new therapeutic strategies. During the last 50 years, several milestones in CML research have been reached. In the early 1960s, Peter Nowell and David Hungerford discovered a small abnormal chromosome that was found in almost all human CML samples; later this abnormal chromosome was named as Philadelphia chromosome (Ph+ chromosome) (3). Ten years later, Janet Rowley proved that this Ph+ chromosome was a product of a reciprocal translocation between chromosome 9 and 22 (4). In 1984, Groffen et al. found that the t(9;22) translocation resulted in a previously known oncogene on chromosome 9, ABL, becoming fused with a previously unknown gene, BCR, on chromosome 22 (5). The molecular basis of Ph+ chromosome is the formation of the BCR-ABL oncogene. Many signaling molecules have been found to be downstream of BCRABL protein (Fig. 15.1). These downstream molecules include adapter proteins such as CRKL and CBL, protein kinase such as MAPK and PI3K, cytoskeleton protein such as paxillin and talin, and transcriptional factors such as ICSBP (1, 6, 7). BCRABL can promote cell proliferation, increase anti-apoptosis activity, and block differentiation of specific cell linage through activating key signal molecules/pathways these serve as promising targets in CML therapies.
Fig. 15.1. Signal pathways activated or inhibited by BCR-ABL oncoprotein. The figure shows some of all the downstream signalling pathways of BCR-ABL; note, more molecules which interact with BCR-ABL have been reported. GRB2, growth factor receptor-bound protein 2; PI3K, phosphatidylinositol 3-kinase; ICSBP, interferon consensus sequence binding protein.
CML Mouse Model
255
Much effort has been made to generate mouse models of Ph+ leukemia. BCR-ABL transgenic models have been made to express BCR-ABL transgene in mice. Different promoters that drive BCR-ABL expression have been tested in these models to express BCR-ABL in different target cells. These promoters include E (8), MPSV-LTR (8), metallothionein (8), BCR (10), and MSCV-LTR (11). Although all of these models show the expression of BCR-ABL in mice, there are at least two obvious defects in using these models: (1) not all mice harboring BCR-ABL develop myeloproliferative disorder, with some mice only developing lymphoid leukemia; and (2) disease latency is long, restricting the use of these models in developing therapeutic strategies for CML. In contrast, the retroviral transduction/transplantation model is a more faithful model of BCR-ABL induced CML. In 1990, Doley et al. co-cultured mouse bone marrow cells with the retroviral producer cells that produced BCR-ABL expressing retrovirus, and they transplanted these infected bone marrow cells into lethally irradiated recipient mice. Three different types of diseases were found in the recipients at up to 5 months post bone marrow transplantation. These diseases included CMLlike myeloproliferative syndrome, acute lymphoblastic leukemia, and a type of tumor involving macrophages (12). In the meantime, Kelliher et al. also established a retroviral system in which a JW-RX retrovirus expressing BCR-ABL was used to infect 5-FU pretreated donor mice bone marrow cells. After bone marrow transplantation, more than 90% of recipients developed tumors, with 50% of them developing a myeloproliferative syndrome that shares several features with the chronic phase of chronic myelogenous leukemia (13). Both of these studies proved that BCRABL is the primary cause of myeloproliferative syndromes in mice. However, there was more than one type of disease in the recipients, further as not 100% of mice developed CML with similar disease latency; hence it was still difficult to conduct drug testing experiments using these models. To overcome these deficiencies, improvements on the model system have been made including modified construct, transient retroviral packaging system, and changes of virus infection conditions. At present, we can induce CML in mice with high efficiency, shown by 100% induction of CML in mice (14). The same CML disease could be induced in most of the inbred mouse strain including C57BL/6, BALB/c, and viable gene knockout mice strains (15). Because all recipients develop CML with a short latency (about 3 weeks), this provides an excellent model for evaluating therapeutic agents for CML treatment (15). As CML is derived from the hematopoietic stem cells which harbor BCRABL oncogene, CML leukemia stem cells can also be studied in this model (15). In conclusion, this retroviral model system pro-
256
Peng and Li
vides a powerful tool for studying the CML disease mechanism and performing translational research.
2. Materials 2.1. Mice
1. C57BL/6J or BALB/cJ mice (The Jackson Laboratory, Bar Harbor, Maine, USA), 4–10-week-old (see Note 1). The donor and recipient have to be the same mouse strain. 2. Gamma irradiator: JL Shepherd 2000Ci Mark I Animal Irradiator.
2.2. Generation of Retroviral Stocks
1. Cell lines: 293T cells (ATCC, Cat# CRL-11268) and NIH3T3 cells (ATCC, Cat# CRL-1658). 2. 293T culture medium: DMEM (Gibco, Bethesda, MD) supplemented with 10% fetal bovine serum (Gibco, Cat# 26140-079), 1% penicillin–streptomycin solution (Gibco, Cat# 15140-122), 1% L-glutamine solution (Gibco, Cat# 25030-081), and 1% MEM non-essential amino acid solution (Sigma, Cat# M7145). 3. NIH3T3 culture medium: DMEM (Gibco) supplemented with 10% new born calf serum (Lonza, Cat# 14-416F), 1% penicillin–streptomycin solution (Gibco), 1% L-glutamine solution (Gibco). 4. Calcium chloride (CaCl2 ) solution: prepare 2 M solution in sterile water, then filter in hood, and store at room temperature. 5. 2XHBS solution: weigh out 8 g NaCl, 0.37 g KCl, 106.5 mg Na2 HPO4 , 1 g dextrose (Gibco, Cat# 15023-021), 5 g HEPES (American Bioanalytical, Cat# AB00892-00500), and add sterile water to 500 mL and adjust pH to 7.05 with 10 N sodium hydroxide (see Note 2). 6. Polybrene (hexadimethrine bromide): (Sigma, Cat# 028K3730) make 8 mg/mL stock solution in H2 O and keep at −20◦ C. 7. 1X Trypsin–EDTA solution (Cellgro, Cat# 25-052-Cl).
2.3. Bone Marrow Cell Collection and Medium
1. Micro-dissecting scissor (Roboz, Cat# RS-5925) and microdissecting forceps (Roboz, Cat# RS-51900) 2. Precision glide needle 27 G 1/2 (BD, Product Number: 301230) and 10 mL BD Luer-LokTM tip syringe (BD, Product Number: 309604) 3. Acrodisc Syringe Filters with Super Membrane (Pall, Cat# 4184)
CML Mouse Model
257
4. DMEM (Gibco, Bethesda, MD) supplemented with 10% fetal bovine serum (Gibco), 1% penicillin–streptomycin solution (Gibco), 1% L-glutamine solution (Gibco), and 1% MEM non-essential amino acid solution (Sigma) 2.4. Bone Marrow Cell Culture and Transduction Medium
1. Bone marrow cell first time stimulation medium: 77% (v/v) DMEM, 15% (v/v) heat inactivated FBS, 5% (v/v) WEHI-3B conditioned medium, penicillin–streptomycin, 1.0 m g/mL ciprofloxacin, 200 mM L-glutamine, 6 ng/mL recombinant murine IL-3 (Peprotech, Cat# 213-13), 10 ng/mL recombinant murine IL-6 (Peprotech, Cat# 21616), and 50–100 ng/mL recombinant murine stem cell factor (SCF; Peprotech, Cat# 250-03). The total volume is 10 mL for each sample. For the second round stimulating medium, the volume is 4 mL for each sample. 2. First time transduction medium: 50% retroviral supernatant, 27% (v/v) DMEM, 15% (v/v) heat inactivated FBS, 5% (v/v) WEHI-3B conditioned medium, penicillin–streptomycin, 1.0 g/mL ciprofloxacin, 200 mM L -glutamine, 6 ng/mL recombinant murine IL-3 (Peprotech, Cat# 213-13), 10 ng/mL recombinant murine IL6 (Peprotech, Cat# 216-16), 50–100 ng/mL recombinant murine stem cell factor (SCF; Peprotech, Cat# 250-03), 1% (v/v) HEPES, and 20 g/mL polybrene. The total volume is 4 mL for each sample. 3. Second time transduction medium: 2 mL retroviral supernatant, 20 g/mL polybrene, and 1% (v/v) HEPES. 4. Red blood cell (RBC) lysis buffer: 150 mM NH4 Cl, 10 mM KHCO3 , and 0.1 mM EDTA (pH to 7.4). 5. Flow cytometry buffer (FACS buffer): PBS supplied with 1% BSA 6. 5-Fluorouracil (5-Fu, Sigma, 10 mg/mL in sterile PBS.
Cat#
6627)
solution:
7. 1XHBSS buffer (Fisher, Cat# 21-022-CV)
3. Methods 3.1. Generation of MIG-BCR-ABL (MSCV-BCR-ABLIRES-GFP) Virus Supernatant
1. Culture 293T cells in 15 cm tissue culture dish (there are about 1 × 108 cells in confluent plates). 2. When the 293T cells reach 90% confluence in the 15 cm dish, remove the medium and wash cells once with 1XPBS. Remove PBS, add 3 mL of trypsin–EDTA solution, and stop the reaction by adding 20 mL 293T medium. Col-
258
Peng and Li
lect cells carefully in 50 mL centrifuge tube and spin at 1500 rpm, 10 mins at room temperature. The 293T cells are passaged to 6 cm dish at 4 × 106 cells/dish at the day before transfection (see Note 3). 3. Change 4 mL fresh 293T medium to each dish before transfection. 4. In a 15 mL tube, add 10 g MIG-BCR-ABL plasmid, 5 g Ecopack plasmid (14), 62 l 2 M CaCl2 and sterile water to 500 l total volume. Briefly vortex. 5. Add 500 l 2×HBS to the tube and mix by vortexing for 10 s (see Note 4) 6. Gently and quickly drop the DNA/HBS solution onto 293T cells. 7. Rock the dishes forward and backward a few times to achieve even distribution of DNA/Ca3 (PO4 )2 particles. 8. After 24 h, remove the old medium and add 4 mL fresh 293T medium 9. After 48 h post-transfection of 293T cells, collect the supernatant by 10 mL BD syringe and filter the supernatant through 0.45 m syringe filter. 10. Aliquot virus supernatant in 4 mL/tube and store at −80◦ ºC (see Note 5). 3.2. Testing Viral Titer by Flow Cytometry
1. Culture NIH3T3 cells with NIH3T3 cell medium in 10 cm dish. 2. When the NIH3T3 cells reach 90% confluence, remove the medium and wash cells once with 1XPBS. Remove PBS, add 1 mL trypsin–EDTA solution, and stop the enzymatic reaction by adding 10 mL NIH3T3 medium. Collect cells carefully in 10 mL centrifuge tube and spin at 1500 rpm, 10 min at room temperature. The NIH3T3 cells are passaged to 10 cm dish at 0.6 × 105 cells/dish at the day before infection. 3. At the day of infection, remove the NIH3T3 cell medium and add virus supernatant serially diluted in 293T medium as 1:2, 1:8, and 1:16. Polybrene is added as 80 g/mL into retroviral supernatant. 4. After 3 h infection at 37 ºC, remove the virus supernatant and change to 10 mL NIH3T3 medium. 5. After 48 h culture, collect cells into 4 mL FACS buffer. Take 300 l cells to do the flow cytometry analysis for percentage of GFP-expressing cells. Normally, the good retroviral supernatant means the GFP% can reach to 90–95% at the 1:2 dilution, 75–85% at 1:8 dilution, and 60–70% at 1:16 dilution.
CML Mouse Model
3.3. Bone Marrow Cells Transduction and Transplantation (see Fig. 15.2) 3.3.1. Priming of Donor Mice at Day 0
259
1. Have donor mice ready (e.g., C57BL/6J or BALB/cJ) 2. Freshly suspend the 5-FU powder in PBS 3. Incubate in 37ºC water bath for 10–30 min, vortex the solution to help to dissolve the powder 4. Inject 5-FU to donor mice via tail vein (200 mg/kg, 0.6 mL for 30 g mouse; intravenous injection, i.v.) (see Note 6)
Fig. 15.2. Retroviral transduction/transplantation model of BCR-ABL induced CML. Donor mice are pretreated with 5FU, and bone marrow cells are stimulated with cytokines in vitro. After mice were infected twice with MIG-BCR-ABL retrovirus, donor bone marrow cells are transplanted into lethally irradiated recipients for induction of CML.
3.3.2. Bone Marrow Cell Collection of Donor Mice at Day 4
1. Kill primed donor mice with CO2 . 2. Sterilize the skin of the mice with 70% ethanol. 3. Collect femurs and tibias and place them in cold PBS, clip of the end of the bone, and briefly clean the muscle. 4. Flush out bone marrow cells with 293T medium. 5. Blow the cells with 10 mL pipette up-down to suspend cells. Normally, 2–3×107 total bone marrow cells can be harvested from 10 donors. 6. Spin down cells at 1500 rpm for 10 min. 7. Resuspend the cell pellet with the first time stimulation medium (less than 3 × 107 cells/dish in 10 mL medium). 8. Incubate cells at 37ºC for 24 h.
260
Peng and Li
3.3.3. First Time Transduction at Day 5
1. Collect the cells and spin cell at 1500 rpm for 10 min at room temperature. 2. Generate the first time transduction medium containing MIG-BCR-ABL retrovirus: 2 mL MIG-BCR-ABL virus is mixed with 2 mL of the first time transduction medium. 3. Add 4 mL of the transduction medium to suspend the cells and then transfer the cells to a 6-well plate. 4. Spin the cells in the 6-well plate at 2300 rpm at room temperature for 90 min. 5. Incubate the cells at 37ºC, 5% CO2 for 3–4 h. 6. Remove the supernatant from each well and add 4 mL second time stimulation medium, then incubate the cells at 37ºC overnight.
3.3.4. Second Time Transduction at Day 6
1. Remove 2 mL of supernatant from each well carefully and then add 2 mL of the second time transduction medium. 2. Spin the cells at 2300 rpm for 90 min. 3. Incubate the cells at 37ºC, 5% CO2 for 3 h. 4. Collect the cells and spin at 1500 rpm for 10 min. 5. Suspend the cell with 5 mL of HBSS solution.
3.3.5. Injection of BCR-ABL Transduced Donor Bone Marrow Cells into Lethally Irradiated Recipient Mice at Day 6
3.4. Monitoring Leukemia Development
3.4.1. Monitoring White Blood Cell Count
1. Recipient mice are treated by two doses of 550-cGy gamma (C57BL/6 J) or 450-cGy gamma (BALB/cJ) separated by 3 h. 2. Adjust the cell concentration to 1.25 × 106 /mL in HBSS solution then inject 0.5×106 cells (0.4 mL) to each mouse via tail vein (intravenous injection, i.v.). After transplantation, recipient mice are evaluated daily for signs of morbidity, weight loss, failure to thrive, and splenomegaly. Depending on the individual animal, hematopoietic tissues and cells are used for several applications, including histopathology, in vitro culture, FACS analysis (see Note 7), secondary transplantation, genomic DNA preparation, protein lysate preparation, or cell lineage analysis. 1. Eye bleed the mice to collect about 100 l peripheral blood. 2. Add 2 l peripheral blood to 18 l RBC buffer, mix, and lyse red blood cells on ice for 10 min. 3. Add 20 l trypan blue, mix well, and drop 15–20 l cells under the cover slip on a hemacytometer and count the live cells using a microscope (dead cells will take up trypan blue, so they are darker than the live cells).
CML Mouse Model
3.4.2. Preparing Peripheral Blood Cells for FACS
261
1. Eye bleed the mice to collect about 100 l peripheral blood into a 0.5 mL Eppendorf tube. 2. Add 1 mL of RBC buffer to the tube, mix well, and lyse the red blood cell on ice 10 min. 3. Spin cells at 6000 rpm at 4◦ C for 5 min. 4. Remove the supernatant and resuspend the cell pellet in 1 mL of PBS to wash the cells. 5. Spin cells at 6000 rpm for 5 min again and remove the supernatant. 6. Resuspend the cell pellet in FACS buffer and adjust cell concentration to 1×107 cells/mL. 7. Add phycoerythrin (PE) or allophycocyanin (APC) labeled or other fluorescent-labeled cell surface marker antibody to 50 l of cells. Mix well and stain the cells at 4ºC for 30 min. 8. Add 1 mL of FACS buffer to the cells and wash cells once by spinning at 6000 rpm, 4ºC for 5 min. 9. Resuspend the cell pellet in 300 l of FACS buffer and then perform flow cytometry analysis.
3.5. Example of Translation Research Using Ph+ CML Mouse Model
The MIG-BCR-ABL retroviral transduction/transplantation model of CML provides an excellent system not only for identifying critical therapeutic targets downstream of BCR-ABL but also for testing the efficiency of therapeutic agents. Previous studies have shown the effectiveness of imatinib in treating CML in our mouse disease model; i.e., imatinib (Gleevec, Novartis) can significantly prolong the survival of CML mice (15, 15). Another example of using this mouse model in translational research is to test the heat shock protein 90 (HSP90) as a critical therapeutic target in CML treatment (15). Below we briefly introduce how to use this model to test the effectiveness of HSP90 inhibitor IPI-504 (retaspimycin hydrochloride; Infinity Pharmaceuticals) in treating CML in mice.
3.5.1. CML Disease Induction in Mice
Induce CML as described in Section 3.1 through Section 3.3.
3.5.2. Drug Treatment of Mice
1. IPI-504 was dissolved in a solution containing 50 mM citrate, 50 mM ascorbate, 2.44 mM EDTA, pH 3.3. 2. Imatinib was dissolved in water. 3. The drugs were given orally in a volume of less than 0.5 mL by gavage (50 or 100 mg/kg, every other day for IPI-504, and 100 mg/kg, twice a day for imatinib) beginning at 8 days after bone marrow transplantation and continuing until the morbidity or death of the leukemic mice.
262
Peng and Li
4. Placebo is a solution containing 50 mM citrate, 50 mM ascorbate, 2.44 mM EDTA, pH 3.3. 3.5.3. Flow Cytometry Analysis
1. Hematopoietic cells were collected from peripheral blood and bone marrow of the CML mice 10 days after bone marrow transplantation. 2. Red blood cells were lysed with NH4Cl red blood cell lysis buffer (pH 7.4). 3. The cells were washed with PBS and stained with B220-PE for B cells, Gr1-APC for neutrophils, and Sca1-APC/c-kitPE for hematopoietic stem cells. 4. After staining, the cells were washed once with PBS and subjected to FACS analysis. Treatment with HSP90 inhibitor (IPI-504) resulted in the reduced peripheral leukemia cell count, spleen weight, and prolonged survival of mice with induced CML for wild type or T315I BCR-ABL. Mice with wild type (WT) or T315I-BCR-ABL transduced bone marrow cells from 5-FU-treated BALB/cJ donor mice were treated with a placebo, the Hsp90 inhibitor IPI-504, or imatinib alone, or the two agents in combination. All placebotreated mice developed and died of CML within 3 weeks after bone marrow transplant (BMT) (Fig. 15.3A). As expected, imatinib treatment was effective in treating WT-induced CML but not CML induced by T315I (Fig. 15.3A). In a dose-dependent manner, treatment with IPI-504 alone significantly prolonged survival of mice with WT CML, but even more markedly prolonged survival of mice with T315I-induced CML (Fig. 15.3A, P < 0.001). Treatment of mice with WT CML with both IPI-504 and imatinib was slightly more effective (but statistically insignificant) than with imatinib alone in prolonging survival of the mice (Fig. 15.3A), while treatment of mice with BCR-ABL-T315Iinduced CML with these two drugs did not further prolong survival of the mice compared with the mice treated with IPI504 alone (Fig. 15.3A). Prolonged survival of IPI-504-treated CML mice correlated with decreased peripheral blood BCRABL-expressing (GFP-positive) leukemia cells during therapy (Fig. 15.3B, P < 0.001) and less splenomegaly at necropsy (Fig. 15.3C). As lung hemorrhage caused by infiltration of mature myeloid leukemia cells is a major cause of death of CML mice (14), we further evaluated the therapeutic effect of IPI-504 on CML by examining the severity of lung hemorrhages at day 15 after BMT. Compared with placebo-treated mice, fewer hemorrhages were observed in the lungs of IPI-504-treated mice with BCR-ABLT315I-induced CML (Fig. 15.3D). Western blot analysis of spleen-cell lysates from the treated CML mice showed that IPI-504 reduced the levels of BCR-ABL protein in CML mice (Fig. 15.3E).
CML Mouse Model
263
Fig. 15.3. Hsp90 is a therapeutic target for CML induced by either BCR-ABL-WT or BCR-ABL-T315I. (A) Treatment with the Hsp90 inhibitor IPI-504 prolonged survival of CML mice. Mice with BCR-ABL-WT (left panel)- or BCR-ABL-T315I (right panel)- induced CML were treated with placebo (n = 15 for BCR-ABL-WT; n = 13 for BCR-ABL-T315I), imatinib (100 mg/kg, twice a day by gavage) (n = 8 for both BCR-ABL-WT and -T315I), IPI-504 (50 mg/kg, once every 2 days by gavage) (n = 20 for both BCR-ABL-WT and BCR-ABL-T315I), IPI-504 (100 mg/kg, once every 2 days by gavage) (n = 8 for both BCR-ABL-WT; n = 7 for BCR-ABL-T315I), and imatinib + IPI-504 (n = 12 for both BCR-ABL-WT and -T315I), respectively, beginning at day 8 after transplantation. The IPI-504-treated mice with BCR-ABL-T315I-induced CML lived longer than those with BCR-ABL-WT-induced CML (comparing between left and right panels). (B) Flow cytometric evaluation of the leukemic process in IPI-504- or imatinib-treated CML mice. The number of circulating leukemic cells (calculated as percentage of Gr-1+GFP+ cells × white blood cell count) in mice with BCR-ABL-WT (left panel)- or BCRABL-T315I (right panel)-induced CML treated with placebo, imatinib, IPI-504, or the combination of imatinib and IPI-504 was determined on day 14 after transplantation. (C) Spleen weights of CML mice treated with placebo, imatinib, IPI-504, and combination of imatinib and IPI-504. (Left panel) BCR-ABL-WT. (Right panel) BCR-ABL-T315I. (D) Photomicrographs of hematoxylin- and eosin-stained lung sections from drug-treated mice at day 14 after transplantation. (E) Western blot analysis of spleen-cell lysates for degradation of BCR-ABL in IPI-504-treated CML mice. IB indicates immunoblot. Adapted from (15) .
In the BMT CML model, imatinib prolongs survival of mice with BCR-ABL-induced CML, but does not stop progression of the disease, partially due to the inability of imatinib to completely eradicate leukemia stem cells (17). In our CML model, BCR-ABL-expressing hematopoietic stem cells (HSCs) have been identified as the leukemia stem cells (17). They are resistant to imatinib treatment and are capable of transferring the disease when sorted by FACS and transplanted to recipient mice. To
264
Peng and Li
Fig. 15.4. Targeting Hsp90 by IPI-504 inhibits survival of leukemia stem cells. (A) Mice with BCR-ABL-T315I-induced CML were treated with a placebo (n=5), imatinib (100 mg/kg, twice a day by gavage) (n=5), and IPI-504 (50 mg/kg, once every 2 days by gavage) (n=5), respectively, for 6 days beginning at day 8 after transplantation. Bone marrow cells were isolated from the treated CML mice, and leukemia stem cells were analyzed by FACS. The numbers of cells represent total leukemia stem cells in average from femur and tibia of each treated CML mouse. (B) IPI-504 had no inhibitory effect on survival of normal HSCs in mice. C57BL/6 J mice were treated with a placebo (n=5), imatinib (100 mg/kg, twice a day by gavage) (n=5), and IPI-504 (50 mg/kg, once every 2 days by gavage) (n=5), respectively, for 2 weeks. Bone marrow cells were isolated from the treated mice and were analyzed by FACS. Adapted from (15).
investigate whether HSP90 is also a major therapeutic target in leukemia stem cells, CML mice are treated by IPI-504 and leukemia stem cells monitored by FACS in vivo. Mice with BCRABLT315I-induced CML were treated with a placebo, imatinib, or IPI-504 for 6 days, and bone marrow cells were analyzed by FACS for GFP+Lin-c-Kit+Sca-1+cells. Imatinib treatment did not lower the percentage and number of leukemia stem cells, compared with the untreated group, whereas IPI-504 treatment had a dramatic inhibitory effect on the stem cells (Fig. 15.4A). To determine whether IPI-504 had an effect on normal HSCs in mice, WT mice were treated with IPI-504 or placebo for 2 weeks. Analysis of bone marrow from these mice showed that there was no change in levels of Lin- c-Kit+ Sca-1+ cells from any treatment group (Fig. 15.3B), indicating that IPI-504 treatment did not inhibit survival of normal HSCs.
4. Notes 1. In this retroviral transduction/transplantation CML model system, inbred mouse strains mice are highly recommended, as these mice have identical genetic background. This avoids modifier effects which would occur when using mixed, undefined, or outbred mouse strains. This genetic uniformity also assists in the development of consistent disease type with and similar disease latency. Most of human caners are a type of genetic diseases. Somatic mutations, chromosome deletions, or translocations could cause different type cancers. Inbred mice with stable genetic backgrounds are the best choice to be used as donors and recipients as secondary genetic events can be avoided.
CML Mouse Model
265
2. Making high-titer BCR-ABL retrovirus is critical. The 2XHBS is the most important reagent, and its pH needs to be adjusted exactly to 7.05, which can be kept at room temperature up to couple of months. 3. For making retrovirus using 293T cells, we recommend to make sure that the confluence of the cells is about 90%, this degree of cell confluence is critical to making high-titer virus. 4. During mixing the 2XHBS with DNA and CaCl2 , gently vortex and drip the solution evenly onto the cultured cells. 5. After collecting the virus supernatant, store them in−80◦ C immediately. This will help to retain the effectiveness of the virus for up to 1 year. 6. 5-FU has low solubility, so a 37ºC water bath is recommended with gently vortexing to improve the solubility. 7. After 8–10 days post bone marrow transplantation, FACS analysis can be performed to monitor the development of CML in mice. Normally, at this time point, the leukemia cells (GFP+GR1+) in peripheral blood of the mice could reach to 3–5% and after 2 weeks could reach to 40–60%. References 1. Ren, R. (2005) Mechanisms of BCR-ABL in the pathogenesis of chronic myelogenous leukaemia. Nat Rev Cancer 5, 172–183. 2. Wang, J. C., Lapidot, T., Cashman, J. D., Doedens, M., Addy, L., Sutherland, D. R., et al. (1998) High level engraftment of NOD/SCID mice by primitive normal and leukemic hematopoietic cells from patients with chronic myeloid leukemia in chronic phase. Blood 91, 2406–2414. 3. Nowell, P. C. and Hungerford, D. A. (1960) Chromosome studies on normal and leukemic human leukocytes. J Natl Cancer Inst 25, 85–109. 4. Rowley, J. D. (1973) Letter: a new consistent chromosomal abnormality in chronic myelogenous leukaemia identified by quinacrine fluorescence and Giemsa staining. Nature 243, 290–293. 5. Groffen, J., Stephenson, J. R., Heisterkamp, N., de Klein, A., Bartram, C. R. and Grosveld, G. (1984) Philadelphia chromosomal breakpoints are clustered within a limited region, bcr, on chromosome 22. Cell 36, 93–99. 6. Wong, S. and Witte, O. N. (2004) The BCRABL story: bench to bedside and back. Annu Rev Immunol 22, 247–306. 7. Deininger, M. W., Goldman, J. M. and Melo, J. V. (2000) The molecular biology of
8.
9.
10.
11.
12.
chronic myeloid leukemia. Blood 96, 3343– 3356. Hariharan, I. K., Harris, A. W., Crawford, M., Abud, H., Webb, E., Cory, S., et al. (1989) A bcr-v-abl oncogene induces lymphomas in transgenic mice. Mol Cell Biol 9, 2798–2805. Heisterkamp, N., Jenster, G., Kioussis, D., Pattengale, P. K. and Groffen, J. (1991) Human bcr-abl gene has a lethal effect on embryogenesis. Transgenic Res 1, 45–53. Castellanos, A., Pintado, B., Weruaga, E., Arevalo, R., Lopez, A., Orfao, A., et al. (1997) A BCR-ABL(p190) fusion gene made by homologous recombination causes B-cell acute lymphoblastic leukemias in chimeric mice with independence of the endogenous bcr product. Blood 90, 2168–2174. Inokuchi, K., Dan, K., Takatori, M., Takahuji, H., Uchida, N., Inami, M., et al. (2003) Myeloproliferative disease in transgenic mice expressing P230 Bcr/Abl: longer disease latency, thrombocytosis, and mild leukocytosis. Blood 102, 320–323. Daley, G. Q., Van Etten, R. A. and Baltimore, D. (1990) Induction of chronic myelogenous leukemia in mice by the P210bcr/abl gene of the Philadelphia chromosome. Science 247, 824–830.
266
Peng and Li
13. Kelliher, M. A., McLaughlin, J., Witte, O. N. and Rosenberg, N. (1990) Induction of a chronic myelogenous leukemia-like syndrome in mice with v-abl and BCR/ABL. Proc Natl Acad Sci U S A 87, 6649–6653. 14. Li, S., Ilaria, R. L., Jr., Million, R. P., Daley, G. Q. and Van Etten, R. A. (1999) The P190, P210, and P230 forms of the BCR/ABL oncogene induce a similar chronic myeloid leukemia-like syndrome in mice but have different lymphoid leukemogenic activity. J Exp Med 189, 1399– 1412. 15. Hu, Y., Liu, Y., Pelletier, S., Buchdunger, E., Warmuth, M., Fabbro, D., et al. (2004) Requirement of Src kinases Lyn, Hck and
Fgr for BCR-ABL1-induced B-lymphoblastic leukemia but not chronic myeloid leukemia. Nat Genet 36, 453–461. 16. Peng, C., Brain, J., Hu, Y., Goodrich, A., Kong, L., Grayzel, D., et al. (2007) Inhibition of heat shock protein 90 prolongs survival of mice with BCR-ABL-T315I-induced leukemia and suppresses leukemic stem cells. Blood 110, 678–685. 17. Hu, Y., Swerdlow, S., Duffy, T. M., Weinmann, R., Lee, F. Y. and Li, S. (2006) Targeting multiple kinase pathways in leukemic progenitors and stem cells is essential for improved treatment of Ph+ leukemia in mice. Proc Natl Acad Sci U S A 103, 16870–16875.
Chapter 16 Mouse Models for Studying Depression-Like States and Antidepressant Drugs Carisa L. Bergner, Amanda N. Smolinsky, Peter C. Hart, Brett D. Dufour, Rupert J. Egan, Justin L. LaPorte, and Allan V. Kalueff Abstract Depression is a common psychiatric disorder, with diverse symptoms and high comorbidity with other brain dysfunctions. Due to this complexity, little is known about the neural and genetic mechanisms involved in depression pathogenesis. In a large proportion of patients, current antidepressant treatments are often ineffective and/or have undesirable side effects, fueling the search for more effective drugs. Animal models mimicking various symptoms of depression are indispensable in studying the biological mechanisms of this disease. Here, we summarize several popular methods for assessing depression-like symptoms in mice and their utility in screening antidepressant drugs. Key words: Depression, animal models, antidepressant drug screening, despair, anhedonia, chronic stress.
1. Introduction The underlying pathophysiology of depression remains unclear despite the seriousness and prevalence of this disorder (1). Clinical symptoms of depression manifest at psychological, physiological, and behavioral levels and include changes in appetite and sleeping patterns, sad or irritable mood, psychomotor agitation, fatigue, anhedonia, poor concentration, feelings of guilt, and recurrent thoughts of suicide or death (2–5). While the introduction of monoamine-based antidepressants has promoted various neurotransmitter system-based models of depression (1), little is known about their mechanisms of therapeutic action. Additionally, up G. Proetzel, M.V. Wiles (eds.), Mouse Models for Drug Discovery, Methods in Molecular Biology 602, DOI 10.1007/978-1-60761-058-8 16, © Humana Press, a part of Springer Science+Business Media, LLC 2010
267
268
Bergner et al.
to 46% of depressed patients do not fully respond to initial monotherapy antidepressant treatments (6), collectively emphasizing the need for newer and more effective drugs (1, 2, 5, 7). Animal models are widely used to study the neurobiological mechanisms of depression (8–10). Ideal animal depression models must be reasonably analogous to the human symptoms, be able to be monitored objectively, be reversed by the same treatment modalities as humans, and be reproducible between laboratories (3, 5, 11). Although selected depression symptoms may be irreproducible in animals (e.g., thoughts of suicide), a number of models exhibit considerable construct validity when targeting other clinical endophenotypes of depression (4, 5, 12). Antidepressant treatment has been shown to affect the behavioral responses in these models (see further), indicating that certain depression paradigms are pharmacologically sensitive, and therefore, can be used in the testing of antidepressant drugs in mice. A clear distinction must be made between animal models of depression and animal tests (or screens) of antidepressant drugs. Examples of both types of animal paradigms, equally important for further progress in biological psychiatry and drug discovery, will be discussed here in detail, based on their common use in behavioral pharmacology research. Finally, automated versions of some of these tests are currently available (13, 14), enabling consistent behavioral measurement, standardization of experimental protocols, and increased throughput and testing.
2. Materials 2.1. Animals
1. Various inbred, selectively bred, and genetically modified (mutant or transgenic (15)) mice (see more details in chapter 18 on mouse models of anxiety in this volume). We recommend using most of the inbred strains listed in the T1priority list of the Mouse Phenome Project database (www.jax.org/phenome), especially C57BL/6J, 129S1/SvImJ, and BALB/c mice. We also recommend browsing the Mouse Genome Informatics (http://www.informatics.jax.org/) database by the depression phenotype to find appropriate transgenic or mutant strains (e.g., Disc1Rgsc1393 /Disc1Rgsc1393 , Tg(Syn1ADCY7)11004Btab/0 mice). 2. In general, avoid strains with overt motor or sensory deficits (e.g., vestibular, cardiovascular, visual) when using tests that may be confounded by these factors. In addition, males and females may exhibit different behavioral reactions to
Mouse Models for Studying Depression-Like States and Antidepressant Drugs
269
the experimental stimuli (16). Therefore, the sex is also an important factor to consider when choosing an appropriate animal for experimentation (see Note 1).
2.2. Housing
1. If mice are obtained from a commercial vendor or another laboratory, allow at least 1 week acclimation from shipping stress. In most cases, a much longer time will be required. Young mice recover more quickly from shipping stress (i.e., 1 week) than adult mice, which may require several weeks to acclimate. Food and water should be freely available, unless the intake is being controlled for experimental purposes. 2. Utilize plastic, solid-floored cages with sufficient space (e.g., <5 animals per cage). The mouse holding room should be kept at approximately 21◦ C, on a 12/12 h light cycle. As mice are nocturnal, the light cycle may be inverted if spontaneous activity measures are needed (16). 3. All experimental procedures (including handling, housing, husbandry, and drug treatment) must be conducted in accordance with national and institutional guidelines for the care and use of laboratory animals.
2.3. Requirements for Experimental Models
1. Sucrose consumption test: 4–10% sucrose (see Note 2) Home cage Two drinking bottles, one with pure water and the other with a sucrose solution (add sucrose to the animals’ standard drinking water, as a change in water type may dissuade animals from drinking). 2. Forced swim test (FST): Clean glass cylinder (e.g., height 25 cm, diameter 10–15 cm) Water maintained at 23–25◦ C Towels to dry animals after swimming Stop watch to calculate the duration of immobility. Optional: video camera for subsequent video tracking and data analysis (e.g., (13). For more information on behavioral tracking software, please see Table 16.1. Videotracking software requires highly developed algorithmic analysis of input; however, recording typically may be done with a standard video camera. Alternative methods of behavioral tracking include vibration-based (e.g., Bioseb, Vitrolles, France) FST activity monitoring.
270
Bergner et al.
Table 16.1 Automated video-tracking system manufacturers Name
City
Country
Web address
Any-Maze
Wood Dale, IL
United States
www.anymaze.com
Bioseb
Vitrolles
France
www.bioseb.com
CleverSys Inc.
Reston, VA
United States
www.cleversysinc.com
Harvard Apparatus
Holliston, MA
United States
www.harvardapparatus.com
Linton Instrumentation
Diss, Norfolk
England
www.lintoninst.co.uk
Medi Analytika Pvt. Ltd
Adyar, Chennai
India
www.medianalytika.com
Noldus
Leesburg, VA, or Wageningen
United States or Netherlands
www.noldus.com
Qubit Systems
Kingston, ON
Canada
www.quibitsystems.com
San Diego Instruments
San Diego, CA
United States
www.sandiegoinstruments.com
TSE Systems
Midland, MI
United States
www.tse-systems.com
India
3. Tail suspension test (TST): A shelf or tail suspension apparatus to suspend mice. The apparatuses may be wooden or plastic boxes (e.g., 680 × 365 × 280 mm), painted to contrast with mice. The design of the TST apparatus is usually negotiable, given that the animal is securely attached to a solid suspension apparatus and that this apparatus is at least 35 cm above the nearest surface (18, 19). Several companies provide behavioral tracking software that is flexible with the variations in experimental design and would yield reliable data that would translate between designs (Table 16.1). Additionally, there are also prefabricated apparatuses that can be purchased (see Table 16.2 for details). Tape measure to determine the height of suspension.
Table 16.2 Selected commercial suppliers of behavioral equipment for depression research Test apparatus
Manufacturer
Company web site
Tail suspension test
Panlab, Barcelona, Spain Columbus Instruments, Columbus OH, United States Bioseb, Vitrolles, France
www.panlab.com www.colinst.com www.bioseb.com
Forced swim test
Panlab, Barcelona, Spain San Diego Instruments, San Diego CA, United States Bioseb, Vitrolles, France
www.panlab.com www.sandiegoinstruments.com www.bioseb.com
Mouse Models for Studying Depression-Like States and Antidepressant Drugs
271
Adhesive tape to secure mice to suspension apparatus (see Notes 13 and 14). Optional: automated electromechanical strain gauge device, video-tracking system. In the tail suspension test (TST), mice initially engage in vigorous escape behaviors, but eventually succumb to immobility. Like the FST, longer durations of TST immobility infer a heightened degree of behavioral despair. As such, TST is a commonly used screening method for antidepressant properties of drugs and is highly sensitive to pharmacological manipulations. Antidepressant drugs generally decrease the duration of TST immobility in mice (14, 20–22). 4. Chronic mild stress (CMS): Supplementary cages for application of stressors. Various stressors, e.g., soiled rat bedding, confinement tube, or predator sounds (see Note 18).
3. Methods 3.1. Observations and General Procedures
1. Observers must refrain from making noise or movement, as their presence may alter animal behavior. Assess intra- and inter-rater reliability for consistency. See details in the chapter on animal models of anxiety. Note that strong scents (e.g., perfume) and loud or sudden noises should be avoided in the experimental room. 2. Allow at least 1 h acclimation of mice after their transfer from the animal holding room to the experimental room. 3. After each testing session, clean the equipment (e.g., with a 30% ethanol solution) to eliminate olfactory cues.
3.2. Drug Administration
1. All experimental protocols described here are compatible with testing various antidepressants, administered with a vehicle (e.g., saline). A typical experiment may include one or several drug-treated groups (e.g., several doses or several pre-treatment times) compared to a vehicle-treated group of mice. Usually (unless stated otherwise), 10 animals per experimental group will be needed, also providing adequate statistical power (see further). However, if the effects of the drugs are particularly robust, a smaller n (e.g., n = 7–8) may suffice. For mild effects, a larger number of animals (n =15– 16) may be required. 2. Common routes of injection include systemic (intraperitoneal [i.p.], intramuscular [i.m.], intravenous [i.v.], per oral [p.o.], subcutaneous [s.c.]) and local (intracerebral [i.c.] or
272
Bergner et al.
intracerebroventricular [i.c.v] or intranasal [i.n.]). Route of administration, dose, and pre-treatment time vary depending on strain sensitivity and the drug being used. Continuous drug infusion (using osmotic pumps, such as Alzet pumps) at a constant rate may be used to improve the availability of the drug, and implantable depots can be used for s.c. drug administration to achieve lasting therapeutic effect. 3.3. Data Analysis
1. Behavioral data may be analyzed with the Mann–Whitney U test for comparing two groups (parametric Student’s t test may be used only if data are normally distributed), or analysis of variance (ANOVA) for multiple groups, followed by an appropriate post hoc test. 2. Some experiments may require one-way ANOVA with repeated measures, or n-way ANOVA depending upon the number of groups tested (see more details in the chapter on mouse models of anxiety in this volume).
3.4. Sucrose Consumption Test
A core symptom of depression is anhedonia – a decreased interest in pleasurable activities (2). There are several commonly used tests to assess hedonic deficits in mice. The sucrose consumption test examines anhedonia in a relatively short period of time without the need for expensive equipment or extensive training of the test animals. In this model, a mouse is given free choice between water and a sucrose solution to drink. Usually, healthy mice show a clear preference for the sweetened water, while depressed animals demonstrate markedly less interest. A pure chance would result in animals drinking equally (50%) from each bottle, and a preference for sucrose of less than 65% is considered to be an indication of hedonic deficit (23). Since various antidepressant drugs reverse the anhedonia-like reduction in preference for sucrose, e.g., (24–26), this test is widely used in the screening of antidepressant drugs. As in most experiments, between 8 and 12 mice may be used per group in this test. However, as few as 6 mice may yield good results, if depression-like phenotypes are robust. C57BL/6J and 129S1/SvImJ mice respond well in this assay (see Note 2). We recommend finding suitable mouse strains through Mouse Genome Informatics or Mouse Phenome Database based on each laboratory’s individual scientific needs. 1. For a set period of time (e.g., 1, 3, or 7 days), allow experimental mice (housed in their standard home cages) access to two freely available water bottles – one containing tap water and the other containing a solution of up to 35% sucrose. To preclude side preference in drinking, switch the positions of the bottles halfway through the procedure (see Notes 2 and 3).
Mouse Models for Studying Depression-Like States and Antidepressant Drugs
273
2. Measure the volumes of sucrose solution and water consumed. Calculate the preference for the sucrose solution as a percentage of total liquid consumed and total sucrose intake in mg/g body weight. In addition, commercially available automated lick-counters (lickometers) may be used (e.g., by Lafayette Instrument Co, Lafayette IN, United States or Columbus Instruments, Columbus OH, United States). Assess the number of licks at each bottle for the duration of the test (i.e., 24–72 h) per 100 mg of body weight, and the preference for sucrose as a percentage of total licks (23, 27, 28) (see Notes 4–6). 3.5. Coat State Assessment
The coat state assessment is a fast and simple qualitative method of assessing mouse depression-like states through observation of the condition of an animal’s fur. In rodents, coat state tends to decline with increased depression, similar to depressed patients who frequently exhibit poor hygiene (29–31). Antidepressants have been shown to improve the coat condition of mice while reducing depression-like symptoms (29–31). For example, the reduction of corticotropin-releasing factor (CRF) has been associated with improved coat state (and is implicated in depression) (32). Of importance here, antidepressants (e.g., imipramine) and anxiolytics (e.g., chlordiazepoxide) have been shown to interact with corticotropin-releasing factor (33) (see Note 7). 1. After removing the animal from the home cage, assess the coat state in each of eight regions: head, neck, forepaws, dorsal coat, ventral coat, hindlegs, tail, and genital region. A coat with a healthy appearance (i.e., unchanged throughout the course of the experiment, normal coat state) should receive a score of 0. Conversely, a coat state that appears damaged or dirty (i.e., noticeably different from a normal coat state) should receive a score of 1. The average of the eight scores for each animal can then be compared among individuals or groups (30, 31). Due to the subjectivity of this assessment, it is beneficial to have more than one observer score each animal (these results should also be compared for inter-rater reliability). One way to minimize bias is to take the animal in question (i.e., the one with a seemingly dirty coat state) and compare it with another animal with an apparently normal coat state. Another way to observe an abnormal coat is to search for mild to severe piloerection, either generally or on specific body parts (32). This can either be in addition to, or independent from, a dirty coat appearance. Taken together, these symptoms signify the animal is not grooming normally and has declining hygiene, implicating overt depressive-like symptoms, (see Notes 8 and 9).
274
Bergner et al.
3.6. Forced Swim Test (FST)
1. This test can be performed manually or with automated video/software systems. Manual labor is not as high throughput as automated behavioral tracking software, and the latency of the observer to react can reduce the accuracy or data acquisition. Regardless of manual or automatic observation, the number of mice that can be tested depends on the duration of the test (6 min is enough to obtain reliable data and to determine significance in this test; therefore, 10 mice could be done per hour). Although it does not induce experimental depression in mice, the FST is one of the most commonly utilized ethological models of fast high-throughput antidepressant screening. The FST places mice in an inescapable aversive situation and measures their “despair,” (learned helplessness) by a measure of increased duration of immobility in the water. Animal FST immobility is markedly reduced by antidepressant drugs. The FST has good predictive validity and is widely used in research investigating acute and chronic effects of antidepressant drugs (20–23, 34) (see Note 10). 1. Place mice individually into a glass cylinder filled with 10 cm of water for 6 min. 2. As a measure of depression-like behavior, the total duration of immobility and the number of immobility episodes should be recorded. Immobility is defined as the absence of movement, unless they are necessary for the animal to stay afloat (head above water) (see Notes 11 and 12). 3. After testing, dry mice thoroughly with towels and return to their home cages.
3.7. Tail Suspension Test (TST)
1. Mice may be suspended by the tail on the edge of a shelf or in a special apparatus, at least 35 cm above the floor (from the beginning of the tail). 2. The mice should be secured by adhesive tape approximately 1 cm from the tip of the tail for 6 min (see Notes 13 and 14). 3. Researchers may choose to manually record data through direct observation or automatically collect data using a strain gauge device to detect movements. 4. Mice are considered immobile only when hanging passively and completely motionless (see Note 15).
3.8. Chronic Mild Stress (CMS)
5. Chronic mild stress (CMS) presents mice with an unpredictable barrage of stressors to induce (rather than simply measure) a depressed state. CMS reduces sucrose or saccharin intake in mice, a symptom of anhedonia (see above). CMS may also be responsible for decreases in sexual and aggressive
Mouse Models for Studying Depression-Like States and Antidepressant Drugs
275
behavior, changes in sleeping habits, loss of body weight, pituitary–adrenal hyperactivity, an increased threshold for brain stimulation reward, and an abolishment of place conditioning, making it a valid ethological model of depression. These behavioral deficits can be reduced through chronic treatments of antidepressants, accentuating the pharmacological sensitivity of CMS procedure (12, 35–37) (see Note 16). 1. Following a random schedule, expose mice to two or more stressors each day for 4–7 weeks (see Note 17). 2. Typical stressors may include cage tilting (e.g., 45◦ ), predator sounds, placement in an empty cage, placement in an empty cage with water on the bottom, damp sawdust, inversion of light/dark cycle, lights on during dark cycle, switching cages, food or water deprivation, short-term confinement in a tube, soiled cages with rat odors, and an inescapable footshock (35). Full experimental design must be published (e.g., degree and duration of cage tilt, dimensions of confinement tube and duration isolated in this tube, quality and duration of predator sounds, size of empty cage and duration of isolation), so that it can be compared across studies and between laboratories. Time of day during administration of stressors, as well as for assessing depression (sucrose intake, bouts of fighting or aggressive behavior, loss of body weight) must be recorded and standardized when possible. Additionally, duration of exposure to sucrose to measure intake should be provided by each study utilizing this model. The schedule of stressors should be random; however, they should be recorded and published citing the order in which they occurred to isolate this pattern as potentially manipulating experimental results (see Note 18). 3. To prevent habituation and enhance the unpredictable nature of the model, stressors should be applied at varying time intervals. 4. Following the period of stress, mice can be tested with behavioral models of depression such as coat state assessment, sucrose consumption test, FST, or TST (see these protocols above) to determine the effectiveness of the test.
4. Notes 1. Occasionally, mice may have altered cognitive domains that may be easily misinterpreted in models of depression (9). For example, mice with elevated learning and memory abilities may display active initial locomotion that decreases significantly over time. While this reduction in
276
Bergner et al.
locomotion may be attributable to heightened learning and habituation, it is often incorrectly assessed as behavioral despair. Likewise, mice with particularly low levels of memory and learning may be misinterpreted as persistently hyperlocomotive. The lack of habituation and decreased sensitivity to repeated stressors may be a result of a reduced learning phenotype, not hyperlocomotion. Similarly, mice displaying hypoactivity and increased sensitivity to repeated stressors (incorrectly categorized as anxious) may be associated with an increased level of depression and enhanced memory. Additionally, sustained hypoactivity coinciding with a decrease in habituation and sensitivity to repeated stressors may not be the result of increased anxiety or decreased despair. Rather, these behaviors may indicate reduced learning and memory, but heightened depression. Overall, cognitive functions may strongly modulate animal performance in ethological models of depression. To diminish the likelihood of incorrect interpretation of behavioral data as depression, it is recommended that mice are carefully tested in memory and learning specific tests (9). 2. The use of a 4–10% sucrose solution will usually generate good results for most mouse strains (e.g., C57BL/6J, 129S1/SvImJ mice). However, some strains (especially mutant or transgenic mice) may have abnormally reduced taste sensitivity, which would make assessment of their hedonic responses in this test difficult. Review Mouse Genome Informatics for mice with abnormal taste sensitivity (e.g., Gnat3tm1Rfm/Gnat3tm1Rfm). Most other mice will respond accurately to this test. However, always check taste sensitivity prior to performing a sucrose consumption test by using a standard taste sensitivity test (see specific mouse phenotyping literature (8–10) for details). Consider using a different strain if the problem persists. Alternatively, higher concentrations of sucrose (e.g., 20–35%) may be required. 3. To avoid confounds of metabolic factors and acute stress, allow food and water ad libitum prior to performing the sucrose consumption test. However, some strains may have altered water consumption (e.g., polydypsia), and the sucrose consumption test may not always be suitable for such strains. For example, this test may be unsuitable for diabetic (e.g, Hk2tm1Laak /Hk2+ ) or obese strains with altered water consumption. 4. Although this test can also be performed over a short time period (e.g., 2 h), mice consume so little over such a brief time and errors in measurement can result. Consider
Mouse Models for Studying Depression-Like States and Antidepressant Drugs
277
lengthening the period of the test to at least 24 h (a 3-day test will be more appropriate in most cases). Note that this protocol must have flexibility and adaptability to be useful as general guidance across laboratories and countries. Rigidly stated specifics can deter novice experimenters from implementing their own research ideas and techniques. 5. Neophobia to the presence of multiple water bottles and to the taste of sucrose may also confound behavioral results in this model. To avoid this problem, acclimate mice by giving them two bottles, each with the sucrose solution, for 72 h before the test, or with one water and one sucrose bottle for 1 h per day for 1 week. Also, consider lengthening the period of the test to at least 24 h. Researchers may choose to utilize video recording to document all water intake, however, end point analysis of overall sucrose consumption (as described earlier) is acceptable. 6. Depending on the length of time over which this test is conducted, mice may alternate between active and inactive phases, which demonstrate marked differences in the animals’ liquid consumption. When switching the positions of the bottles to avoid side preference, be sure to take shifting activity levels into account so that each bottle is in each position for the same amount of each activity phase. Some mouse strains may develop a metabolic syndromelike phenotype or have pathologically high rewardrelated phenotype (see Mouse Phenome Project or Mouse Genome Informatix database for these phenotypes of interest: e.g., metabolic syndrome-like phenotypes in Neil1tm1Rsld /Neil1tm1Rsld mice). Thus, their sucrose consumption may be abnormally affected and alternative methods of depression testing or other mouse strains may be required. 7. Various mouse strains may have different sensitivity in this test. For example, C57BL/6J mice can be somewhat resistant to the deleterious effects of chronic stress on the coat state (38). Strain differences may result in differing levels of grooming activity. For example, some inbred strains may be inherently poor (e.g., BALB/cJ) or excellent (e.g., A/J) groomers, regardless of stress levels. Some genetically modified mouse strains also display “compulsive” grooming behavior (39) that may mask any alterations in the animal’s coat state. Consider a more suitable strain if floor or ceiling effects occur. 8. In socially housed mice, hetero-grooming may confound self-grooming data. Single-housing mice may eliminate this confound, but this practice should be used with caution, as social isolation stress may induce aberrant behavioral
278
Bergner et al.
effects. Typically use 8–10 subjects per group in order to obtain reliable data. The nature of this specific test is to be used as complementary assessment of depression to supplement other models of depression. This test is relatively simple; however, it may be more useful when performed adjunct to another test, such as the tail suspension test (e.g., in the tail suspension test, before releasing the animal, simply score the coat state in addition to the other endpoints measured in this test). Similarly, coat state can be assessed while animals are still in the home cage, or with little manipulation outside of the home cage. 9. Some mouse strains may display pronounced balding patches due to alopecia (40) or increased auto- or heterobarbering behavior (41–43), which will make the coat state data less valid. Therefore, this model may not be used in high barbering strains. Likewise, stress per se may promote barbering in mice (44), thereby further confounding the coat assessment protocol. 10. Consider strain and individual differences in baseline immobility duration. C57BL/6J, BALB/cJ, and 129/SvEmJ strains have all been shown to provide reliable data and good sensitivity to pharmacological manipulations in this test (45). There is a growing number of mouse models with metabolic syndrome-like phenotypes, as well as with altered bone physiology (15). Mutant strains with calcium or bone deficiency (search Mouse Genome Informatics for specific examples) can potentially confound data in this respect, so this test would not be reliable in an experiment utilizing such mice. Similarly, obese mice may also confound data, as they could either be too buoyant or simply become exhausted during the test. Consequently, mutant animals with such phenotypes may have affected swimming abilities/buoyancy, and therefore may not be adequately compared in the FST with their wild type littermates. 11. Motor or vestibular deficits may result in poor (abnormal) swimming, including aberrant spinning, turning, and sinking, that may confound FST data. Examine such mice in specific motor or vestibular ability tests. Mice with poor swimming should be excluded from the FST. Also note that some popular inbred mouse strains (e.g., most 129 mouse substrains) are poor swimmers and develop spastic behaviors in FST situations that complicate their swimming. 12. Some mice exhibiting increased levels of FST immobility may be suffering from fatigue rather than depression per se. Evaluate the fatigability of animals in separate tests. If
Mouse Models for Studying Depression-Like States and Antidepressant Drugs
279
mice display high fatigability phenotypes, consider shortening the length of the test. 13. Some mice may fall from the apparatus due to poor fixation by the adhesive tape. Use a cushioned floor for the TST to prevent any damage to the animal and exclude such mice from the experiment. 14. Note that a moderately adhesive tape will be required (preferably, use vinyl tape or medical tape). Since most mice weigh about 25–35 g, duct tape is too strong and would not be required; also, tape of this grade would likely tear hair, skin, and possible part of the tail off the animal. 15. Some strains (e.g., C57BL/6J mice) display specific tail climbing behaviors and may not be an appropriate mouse model for this test (46). In contrast, BALB/cJ, DBA/2J, and BTBR strains were all shown to be reliable in this test and simultaneously responsive to drug effects (e.g., citalopram) (47). The growing number of mice with vestibular deficits (e.g., MRL/MpJ, Ce/J, and SJL/J inbred strains) (48); BDNF knockout mice (49) require further consideration, since strains with vestibular deficits may show an abnormal “spinning” phenotype in the TST, thereby confounding behavioral data in this model. Consider using other models of depression for testing these mice. Some mutant mice display other specific neurological abnormalities relevant to their TST performance. For example, mutation or deletion of the Pafah1b1 gene of mice on a mixed 129SvEv-NIH Black Swiss background showed a marked increase in “hind leg clutching” behavior (50) whereas hind leg clasping behavior is common in including serotonin transporter knockout mice on 129S1/SvImJ background (own observations). Such phenotypes may result in abnormally high immobility in this test (which can incorrectly be interpreted as low depression). In contrast, spontaneous mild seizures in some mice (see Mouse Genome Informatics database for examples) may lead to reduced TST immobility, again confounding depressionrelated data. 16. While CMS is a valid model of depression in mice, it is labor intensive, long in duration, and demanding of space. A practical recommendation for this model is thorough planning of all experiments and consistent completion of the entire CMS battery. 17. Consider strain differences in this paradigm. Some stressors may not affect all strains homogenously, and similarly, some models of depression may not accurately reflect depression in specific strains. For example, C57BL/6J mice are not sensitive to CMS affects on coat state (38).
280
Bergner et al.
18. Some suppliers (e.g., Keystone Country Store, www. keystonecountrystore.com) provide electronic devices that emit predator sounds, although rat vocalizations can also be recorded from live specimen. To standardize this, “predator sounds” should be published according to decibel, frequency, pitch, and length of recording. Unfortunately, soiled rat bedding cannot be obtained through a vendor. However, to standardize this stressor it would be possible to use a metabolic cage, collect, and measure the amount of urine and defecation of the rat, and then combine this with fresh rat bedding. A confinement tube is admittedly not descriptive. To isolate the mouse and cause a mild level of anxiety, insert the mouse into a restraint tube. These are typically similar in size and shape, although the exact design and specifications (dimensions) should be published in each manuscript utilizing this stressor. Some manufacturers include ITP (www.intoxproducts.com), AD Research (www.adinstruments.com), and ONARES (onares.com).
Acknowledgments This work was supported by NARSAD YI Award to AVK and by Stress Physiology and Research Center (SPaRC) of Georgetown University Medical School. References 1. Wong, M. L. and Licinio, J. (2004) From monoamines to genomic targets: a paradigm shift for drug discovery in depression. Nat Rev Drug Discov 3, 136–151. 2. Cryan, J. F. and Holmes, A. (2005) The ascent of mouse: advances in modelling human depression and anxiety. Nat Rev Drug Discov 4, 775–790. 3. Cryan, J. F., Markou, A. and Lucki, I. (2002) Assessing antidepressant activity in rodents: recent developments and future needs. Trends Pharmacol Sci 23, 238–245. 4. Cryan, J. F. and Mombereau, C. (2004) In search of a depressed mouse: utility of models for studying depression-related behavior in genetically modified mice. Mol Psychiatry 9, 326–357. 5. Cryan, J. F. and Slattery, D. A. (2007) Animal models of mood disorders: recent developments. Curr Opin Psychiatry 20, 1–7. 6. Fava, M. and Davidson, K. G. (1996) Definition and epidemiology of treatment-resistant depression. Psychiatr Clin North Am 19, 179–200.
7. Malatynska, E., Rapp, R., Harrawood, D. and Tunnicliff, G. (2005) Submissive behavior in mice as a test for antidepressant drug activity. Pharmacol Biochem Behav 82, 306–313. 8. Kalueff, A. V., Laporte, J. L., Murphy, D. L. and Sufka, K. (2008) Hybridizing behavioral models: a possible solution to some problems in neurophenotyping research? Prog Neuropsychopharmacol Biol Psychiatry 32, 1172–1178. 9. Kalueff, A. V.and Murphy, D. L. (2007) The Importance of cognitive phenotypes in experimental modeling of animal anxiety and depression. Neural Plasticity 2007, 1–7. 10. Geyer, M. A. and Markou, A. (1995) Animal models of psychiatric disorders, in Psychopharmacology: The fourth generation of progress (Bloom, F. and Kupfer, D. J., eds.). Raven Press, New York, pp. 787–798. 11. Frazer, A. and Morilak, D. A. (2005) What should animal models of depression model? Neurosci Biobehav Rev 29, 515–523.
Mouse Models for Studying Depression-Like States and Antidepressant Drugs 12. Willner, P. (1997) Validity, reliability and utility of the chronic mild stress model of depression: a 10-year review and evaluation. Psychopharmacology (Berl) 134, 319–329. 13. Crowley, J. J., Jones, M. D., O Leary, O. F. and Lucki, I. (2004) Automated tests for measuring the effects of antidepressants in mice. Pharmacol Biochem Behav 78, 269– 274. 14. Juszczak, G. R., Sliwa, A. T., Wolak, P., Tymosiak-Zielinska, A., Lisowski, P., et al. (2006) The usage of video analysis system for detection of immobility in the tail suspension test in mice. Pharmacol Biochem Behav 85, 332–338. 15. Jackson-Laboratory. (2008) Mouse Genome Informatics. 16. Palanza, P. (2001) Animal models of anxiety and depression: how are females different? Neurosci Biobehav Rev 25, 219–233. 17. Deacon, R. M. (2006) Housing, husbandry and handling of rodents for behavioral experiments. Nat Protoc 1, 936–946. 18. Steru, L., Chermat, R., Thierry, B. and Simon, P. (1985) The tail suspension test: a new method for screening antidepressants in mice. Psychopharmacology (Berl) 85, 367–370. 19. Kos, T., Legutko, B., Danysz, W., Samoriski, G., and Popik, P. (2006) Enhancement of antidepressant-like effects but not brainderived neurotrophic factor mRNA expression by the novel N-methyl-D-aspartate receptor antagonist neramexane in mice Behav Pharmacol 318, 1128–1136. 20. Bai, F., Li, X., Clay, M., Lindstrom, T. and Skolnick, P. (2001) Intra- and interstrain differences in models of ”behavioral despair”. Pharmacol Biochem Behav 70, 187–192. 21. Bourin, M., Chenu, F., Ripoll, N. and David, D. J. (2005) A proposal of decision tree to screen putative antidepressants using forced swim and tail suspension tests. Behav Brain Res 164, 266–269. 22. Hunsberger, J. G., Newton, S. S., Bennett, A. H., Duman, C. H., Russell, D. S., et al. (2007) Antidepressant actions of the exercise-regulated gene VGF. Nat Med 13, 1476–1482. 23. Strekalova, T., Spanagel, R., Bartsch, D., Henn, F. A. and Gass, P. (2004) Stressinduced anhedonia in mice is associated with deficits in forced swimming and exploration. Neuropsychopharmacology 29, 2007–2017. 24. Jayatissa, M. N., Bisgaard, C. F., West, M. J. and Wiborg, O. (2008) The number of granule cells in rat hippocampus is reduced after chronic mild stress and re-established after
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
281
chronic escitalopram treatment. Neuropharmacology 54, 530–541. Xu, Q., Yi, L. T., Pan, Y., Wang, X., Li, Y. C., et al. (2008) Antidepressant-like effects of the mixture of honokiol and magnolol from the barks of Magnolia officinalis in stressed rodents. Prog Neuropsychopharmacol Biol Psychiatry 32, 715–725. Zhao, Z., Wang, W., Guo, H. and Zhou, D. (2008) Antidepressant-like effect of liquiritin from Glycyrrhiza uralensis in chronic variable stress induced depression model rats. Behav Brain Res 194, 108–113. Perona, M. T., Waters, S., Hall, F. S., Sora, I., Lesch, K. P., et al. (2008) Animal models of depression in dopamine, serotonin, and norepinephrine transporter knockout mice: prominent effects of dopamine transporter deletions. Behav Pharmacol 19, 566–574. Luo, D. D., An, S. C. and Xhang, X. (2008) Involvement of hippocampal serotonin and neuropeptide Y in depression induced by chronic unpredicted mild stress. Brain Res Bull 77, 8–12. Piato, A. L., Detanico, B. C., Jesus, J. F., Lhullier, F. L., Nunes, D. S., et al. (2008) Effects of Marapuama in the chronic mild stress model: further indication of antidepressant properties. J Ethnopharmacol 118, 300– 304. Yalcin, I., Aksu, F. and Belzung, C. (2005) Effects of desipramine and tramadol in a chronic mild stress model in mice are altered by yohimbine but not by pindolol. Eur J Pharmacol 514, 165–174. Yalcin, I., Aksu, F., Bodard, S., Chalon, S. and Belzung, C. (2007) Antidepressantlike effect of tramadol in the unpredictable chronic mild stress procedure: possible involvement of the noradrenergic system. Behav Pharmacol 18, 623–631. Ducottet, C., Griebel, G., and Belzung, C. (2003) Effects of the selective nonpeptide corticotropin-releasing factor receptor 1 antagonist antalarmin in the chronic mild stress model of depression in mice. Progress Neuro-Psychopharmacol Biol Psychiatry 27, 6. Zhang, L. and Barrett, J. E. (1990) Interactions of corticotropin-releasing factor with antidepressant and anxiolytic drugs: behavioral studies with pigeons. Biol Psychiatry 27, 953–967. Burne, T. H., Johnston, A. N., McGrath, J. J. and Mackay-Sim, A. (2006) Swimming behaviour and post-swimming activity in Vitamin D receptor knockout mice. Brain Res Bull 69, 74–78. Harkin, A., Houlihan, D. D. and Kelly, J. P. (2002) Reduction in preference for saccharin
282
36.
37.
38.
39. 40.
41.
42.
43.
Bergner et al. by repeated unpredictable stress in mice and its prevention by imipramine. J Psychopharmacol 16, 115–123. Willner, P., Moreau, J. L., Nielsen, C. K., Papp, M. and Sluzewska, A. (1996) Decreased hedonic responsiveness following chronic mild stress is not secondary to loss of body weight. Physiol Behav 60, 129–134. Pothion, S., Bizot, J. C., Trovero, F. and Belzung, C. (2004) Strain differences in sucrose preference and in the consequences of unpredictable chronic mild stress. Behav Brain Res 155, 135–146. Mineur, Y. S., Prasol, D. J., Belzung, C. and Crusio, W. E. (2003) Agonistic behavior and unpredictable chronic mild stress in mice. Behav Genet 33, 513–519. Greer, J. M. and Capecchi, M. R. (2002) Hoxb8 is required for normal grooming behavior in mice. Neuron 33, 23–34. Kalueff, A. V., Keisala, T., Minasyan, A., Kuuslahti, M., Miettinen, S., et al. (2006) Behavioural anomalies in mice evoked by ”Tokyo” disruption of the Vitamin D receptor gene. Neurosci Res 54, 254–260. Garner, J. P., Weisker, S. M., Dufour, B. and Mench, J. A. (2004) Barbering (fur and whisker trimming) by laboratory mice as a model of human trichotillomania and obsessive-compulsive spectrum disorders. Comp Med 54, 216–224. Kalueff, A. V., Minasyan, A., Keisala, T., Shah, Z. H. and Tuohimaa, P. (2006) Hair barbering in mice: implications for neurobehavioural research. Behav Processes 71, 8–15. Sarna, J. R., Dyck, R. H. and Whishaw, I. Q. (2000) The Dalila effect: C57BL/6 mice
44.
45.
46.
47.
48.
49.
50.
barber whiskers by plucking. Behav Brain Res 108, 39–45. Garner, J. P., Dufour, B., Gregg, L. E., Weisker, S. M. and Mench, J. A. (2004) Social and husbandry factors affecting the prevalence and severity of barbering (”whisker-trimming”) by laboratory mice. Appl Anim Lab Sci 89, 263–282. Lucki, I., Dalvi, A. and Mayorga, A. J. (2001) Sensitivity to the effects of pharmacologically selective antidepressants in different strains of mice. Psychopharmacology (Berl) 155, 315– 322. Mayorga, A. J. and Lucki, I. (2001) Limitations on the use of the C57BL/6 mouse in the tail suspension test. Psychopharmacology (Berl) 155, 110–112. Crowley, J. J., Blendy, J. A. and Lucki, I. (2005) Strain-dependent antidepressant-like effects of citalopram in the mouse tail suspension test. Psychopharmacology (Berl) 183, 257–264. Jones, S. M., Jones, T. A., Johnson, K. R., Yu, H., Erway, L. C., et al. (2006) A comparison of vestibular and auditory phenotypes in inbred mouse strains. Brain Res 1091, 40– 46. Rauskolb, S. (2008) Brain-derived neurotrophic factor: generation and characterization of adult mice lacking BDNF in the adult brain. University of Basel, Basel, Germany, p. 91. Paylor, R., Hirotsune, S., Gambello, M. J., Yuva-Paylor, L., Crawley, J. N., et al. (1999) Impaired learning and motor behavior in heterozygous Pafah1b1 (Lis1) mutant mice. Learn Mem 6, 521–537.
Chapter 17 Virus-Delivered RNA Interference in Mouse Brain to Study Addiction-Related Behaviors Amy W. Lasek and Nourredine Azouaou Abstract The use of viral vectors for gene transfer to specific brain regions is a powerful tool for determining gene function in mouse behavioral models. We have employed a lentiviral vector to deliver small-hairpin RNAs to areas of mouse brain implicated in behaviors relevant to drug addiction, such as the nucleus accumbens and ventral tegmental area. Delivery of virus expressing small-hairpin RNAs results in sustained target gene knockdown by RNA interference. Mice can subsequently be tested for behavioral responses to various drugs of abuse over the course of several weeks or months. Here we describe a method for stereotaxic delivery of lentivirus to mouse brain. This method is widely applicable to any behavioral experiment in which the role of a specific gene in a particular brain region is to be elucidated. Key words: Lentivirus, RNA interference, stereotaxic surgery, mouse models, drug addiction. Abbreviations: RNAi RNA interference shRNA small-hairpin RNA GFP green fluorescent protein DMA digital manipulator arm SAS stereotaxic alignment system AP anterior/posterior ML medial/lateral DV dorsal/ventral i.p. intraperitoneally s.c. subcutaneously VTA Ventral tegmental area
1. Introduction RNA interference (RNAi) has emerged as a useful method for determining gene function in mammals. Double-stranded RNA expression in cells results in sequence-specific knockdown of a G. Proetzel, M.V. Wiles (eds.), Mouse Models for Drug Discovery, Methods in Molecular Biology 602, DOI 10.1007/978-1-60761-058-8 17, © Humana Press, a part of Springer Science+Business Media, LLC 2010
283
284
Lasek and Azouaou
particular gene of interest (1). Delivery of double-stranded synthetic small-interfering RNAs (siRNAs) or virally expressed smallhairpin RNAs (shRNAs) are complementary approaches to making transgenic or knockout mice in order to determine the role of a gene in a biological process (2). The advantage of using RNAi in specific tissues of the mouse is that it can be done rapidly in comparison to making transgenic or gene knockout mice. In addition, RNAi can be used in the adult animal, thus overcoming issues associated with the developmental role of a gene or compensatory changes that occur when a gene is knocked out during embryogenesis. Determining the adult-specific role of a target gene may also reflect more accurately the response of the animal to treatment with small molecule inhibitors developed for disease therapy. Indeed, therapeutic approaches using RNAi are currently in development (1). A number of laboratories have used RNAi in the central nervous system (CNS) of mice to determine the therapeutic potential of gene knockdown or to delineate the role of a gene in neuropsychiatric conditions such as drug addiction (3–6). Lentiviral-expressed shRNAs are especially useful for maintaining sustained target gene knockdown when studying mice in complex behavioral models, such as those employed for drug addiction. Here we describe a method for injecting lentivirus directly into specific regions of the mouse brain. This method is applicable to any viral-based approach used in studies of CNS function in the mouse. In addition, this method can be used in any mouse strain required for the specific behavioral end point being studied. For instance, in studies of cocaine or alcohol addiction, an investigator may choose to use the C57BL/6J strain because of its robust response to cocaine and propensity to drink large quantities of ethanol, whereas for other alcohol-related behaviors, the DBA/2J mouse might be chosen (7). Lentiviral infection in the brain is generally monitored through co-expression of a fluorescent protein, such as green fluorescent protein (GFP), from the integrated lentivirus genome. For those investigators without the resources to clone shRNA sequences into the viral plasmid and produce the virus in-house, high-titer lentiviruses expressing shRNAs targeting nearly every gene in the mouse genome are commercially available (for example, query Open Biosystems at www.openbiosystems.com and Dharmacon at www.dharmacon.com, both part of Thermo-Fisher Scientific).
2. Materials 2.1. Cannulae for Virus Infusion
1. Stainless steel, 33 G hypodermic tubing (type 304), pre-cut to 34 mm (Small Parts, Inc., Miramar, FL).
Viral RNAi in Mouse Brain
285
2. Stainless steel, 26 G hypodermic tubing (type 304), pre-cut to 27 mm (Small Parts). 3. Polyethylene tubing, PE-20, 0.015” (inner diameter) × 0.043” (outer diameter) (Braintree Scientific, Braintree, MA). R Glue). 4. Cyanoacrylate glue (Super Glue or Krazy
5. Polystyrene weigh dishes (Fisher Scientific, Santa Clara, CA). 6. Crile hemostat, straight, serrated, 14 cm (Fine Science Tools, Foster City, CA). 2.2. Stereotaxic Alignment System
1. Stereotaxic alignment system base with mouse adaptor (David Kopf Instruments, Tujunga, CA, Model 1900). 2. Alignment indicator (David Kopf, Model 1905). 3. Micro-manipulator with three axis, 1 m resolution and digital display readout (David Kopf, Model 1940). 4. Centering microscope with 20X magnification (David Kopf, Model 1915). 5. Stereotaxic drill unit (David Kopf, Model 1911). 6. Dial test indicator (David Kopf, Model 1910). This tool is used to level the micro-manipulator arm with respect to the base unit during setup of the stereotaxic alignment system, according to the manufacturer’s instructions. It is not required during actual surgery. 7. Dual cannula insertion tool (David Kopf, Model 1973). 8. Infusion pump PHD 22/2000 (Harvard Apparatus, Holliston, MA).
2.3. Surgery
1. 70% Ethanol (v/v) in water, stored at room temperature. 2. Sterile water, stored at room temperature. 3. Hair trimmer (Oster). 4. Electric heating pad. 5. Adson forceps, straight, 12 cm (Fine Science Tools). 6. Drill bit for stereotaxic drill, #85, 0.011 diameter (David Kopf). 7. Tuberculin syringes with 26 G needle (Becton Dickinson, Franklin Lakes, NJ). 8. Sodium chloride for injection, 0.9% (Henry Schein, Melville, NY). 9. Disposable scalpel, No. 10 (Fisher Scientific). 10. Povidone–iodine swab sticks (Henry Schein). 11. Lidocaine HCl jelly, 2% (Henry Schein).
286
Lasek and Azouaou
12. Puralube veterinary ophthalmic ointment (Henry Schein). 13. Cyanoacrylate tissue glue (e.g., Vetbond tissue adhesive, Henry Schein). 14. Cotton Tip applicators, 6” (Fisher Scientific). 15. Hamilton gastight syringe, 2 L volume, 25 G (Model 7002, Fisher Scientific). 16. Ketamine, 100 mg/mL (Henry Schein). Dilute to 10 mg/mL in sterile 0.9% sodium chloride (saline). Store at room temperature. Ketamine is a controlled substance (Schedule III) and should be stored in a secure location when not in use and otherwise handled as per institutional requirements. 17. Xylazine, 20 mg/mL (Henry Schein). Dilute to 2 mg/mL in sterile saline. Store at room temperature. 18. Stainless steel utility scissors (Fine Science Tools). 19. High-titer (e.g., ≥107 pg p24 antigen/mL by ELISA or 108 transduction units (TU)/mL) lentivirus, stored at −80◦ C in 1.5 mL centrifuge tubes in 10–20 L aliquots. Thaw and store on ice during surgery procedure. Do not freeze–thaw more than twice. Handle lentivirus using Biosafety Level 2 precautions. 20. Mouse cage containing 1 deep alpha cellulose animal bedding (Alpha Dri, Dean’s Animal Feed Inc., Redwood City, CA). 21. Mouse water bottle filled with Children’s Tylenol (2 mg/mL) in drinking water.
3. Methods Prior to beginning surgeries, the stereotaxic alignment system (SAS) must be assembled and leveled according to manufacturer’s instructions. This process is important to increase the overall accuracy of the instrument. Although leveling only needs to be done once while setting up the equipment, it can take a few hours, so time must be allocated accordingly. In addition, users of the SAS should familiarize themselves with each piece of equipment and read the manual before attempting surgery. All animal protocols must be approved by the applicable institutional animal care and use committee (IACUC) prior to embarking on any surgery. Safety-modified lentivirus and other commonly used viruses (adenovirus, adeno-associated virus) should be prepared and handled according to Biosafety Level 2 guidelines.
Viral RNAi in Mouse Brain
287
The digital SAS from David Kopf Instruments allows for accurate targeting of specific brain regions in the mouse. Because the mouse brain extends up to 13 mm along the anterior/posterior (AP) axis, 9 mm along the medial/lateral (ML) axis, and 5 mm along the dorsal/ventral (DV) axis, a 1 mm error in viral infusion placement may lead to infection in an entirely different nucleus of the brain. The combination of a 1-micron resolution digital readout controlled by a digital manipulator arm (DMA) and an alignment indicator used in leveling the mouse head allows for increased targeting accuracy. Three-dimensional movement of the DMA and the mouse head are controlled separately. Individual knobs control movement in the X- (ML), Y(AP), and Z- (DV) planes. The components of the digital SAS are pictured in Fig. 17.1. During surgery, the DMA is first set to 0 at a focal point representing bregma (the intersection of the sagittal and coronal sutures on the mouse skull; Fig. 17.3). The mouse head is then moved to 0 such that bregma is centered and in focus. The mouse head is leveled in the sagittal and coronal planes. Holes are drilled bilaterally at the precise X and Y coordinates for the targeted brain region, and cannulae filled with virus are lowered into the holes to the Z coordinate. We have successfully targeted the nucleus accumbens (Lasek et al, in preparation), caudate putamen (6), ventral tegmental area (6), and amygdala (8), brain regions that are important for addiction-related behaviors. Coordinates for these brain regions are listed in Table 17.1 (see Note 1). It is critical to minimize damage to the mouse brain during virus infusion, so a very thin (33 G) stainless steel hypodermic tube is used as a cannula for virus delivery. A 26 G stainless steel tube provides rigidity to the 33 G tube and allows for a tight connection to polyethylene tubing and the cannula insertion tool (Fig. 17.2). Cannulae for virus infusion are prepared and tested as described below prior to surgery. Each cannula consists of one 33 G, 34 mm tube inserted into a 26 G, 27 mm tube. The tube ends are asymmetrically aligned such that one end of the 33 G tube extends 5–6 mm past the 26 G tube and the other end extends 1–2 mm (Fig. 17.2). The 6 mm extension is lowered into the mouse brain for virus infusion during stereotaxic surgery, and the 1 mm-extended end is attached to polyethylene (PE-20) tubing for connection to a gastight syringe and infusion pump. Damage is also minimized by performing a slow infusion of virus using the infusion pump. After surgery, mice are allowed to recover for 2 weeks prior to performing any behavioral experiments. This 2 week period allows for virus infection, expression of the shRNA targeting the gene of interest, knockdown of target gene RNA, and degradation of previously translated protein. The proper incubation time after surgery for maximum target gene knockdown must be
288
Lasek and Azouaou
Fig. 17.1. Digital stereotaxic alignment system. (A) The complete system, showing the mouse adaptor mounted on the base (1), the DMA (2) with the centering microscope attached, the digital monitor (3), and infusion pump with gastight syringes attached (4).
Viral RNAi in Mouse Brain
289
Table 17.1 Representative stereotaxic coordinates for virus infusion AP (Y) coordinatea
ML (X) coordinateb
DV (Z) coordinatec
Nucleus accumbens
+1.7
± 0.9
−4.6
Caudate putamen
+1.5
± 1.5
−3.2
Ventral tegmental area
−3.2
± 0.5
−4.7
Amygdala(central nucleus)
−0.85
± 3.1
−4.9
Mouse brain region
a Number
refers to mm from bregma. A positive value is anterior to bregma, while negative is posterior to bregma. The letter in parenthesis refers to the axis on the digital monitor. b Number refers to mm from midline. Positive values are to the right of midline, negative to the left. Both coordinates are used for bilateral injections. c Number refers to mm ventral from the top of the skull.
determined experimentally using available reagents to measure RNA and protein levels of the target gene. In addition, it is essential that mice are fully recovered from surgery for interpretation of subsequent behavioral experiments. 1. Cannulae should be prepared the day before surgery to allow time for glue to dry and to test for unobstructed liquid flow through tubing and cannulae.
3.1. Cannula Preparation
2. Insert one 34 mm, 33 G stainless steel hypodermic tube into one 27 mm, 26 G stainless steel hypodermic tube. Position the 34 mm tube so that 5–6 mm extends from one end of the 26 G tube, and 1–2 mm extends from the opposite end. Prepare 10 tubes this way and place into a plastic weigh dish.
Fig. 17.1. (Continued) X- and Z-axis knobs that control the DMA are labeled. The Y-axis knob is not visible in the picture, since it is located on the back side of the DMA. (B) A closer view of the mouse adaptor, showing the centering scope attached to the DMA (1), the centering height gauge (2), the palate bar and nose clamp (3), and the ear bars (4). X- and Z-axes, sagittal and coronal tilt, and the sagittal alignment knobs that control the movement of the mouse head are labeled. The Y-axis knob controlling movement of mouse head is not visible, since it is located at the back of the mouse adaptor. (C, D) Components of the SAS used during surgery. (C) Alignment indicator (1), centering height gauge (2), and centering microscope (3). The arrowhead shows the dials and the arrow the touch probes on the alignment indicator. (D) Dual cannula insertion tool, with Allen wrench (1) and stereotaxic drill unit, with wrenches and attached drill bit (2). The arrow points to the drill bit. The alignment indicator, microscope, dual cannula insertion tool, and drill are designed to easily attach to and remove from the DMA during surgery.
290
Lasek and Azouaou
Fig. 17.2. Virus infusion cannula. (A) Each cannula consists of a 33 G stainless steel hypodermic tube inserted into a 26 G tube. The 33 G tube extends from the 26 G tube 6 mm on one end and 1 mm on the other. The end with the 1 mm extension is bent at a right angle 10 mm from the end and inserted into PE-20 tubing. The PE-20 tubing will be attached to a gastight syringe on an infusion pump. (B) Placement of the two cannulae into the dual cannula insertion tool. The 33 G ends of the cannulae extend 6 mm from the end of the tool and are inserted bilaterally into the mouse brain during virus infusion.
3. Put a large drop of glue into a separate weigh dish and drag the end of a 20 or 200 L pipette tip through the glue so that a small droplet of glue is formed at the end of the tip. 4. Seal the two joints between each of the 33 and 26 G tubes prepared in Step 2 with the glue droplet, being careful not to seal the ends of the 33 G tube. The goal is to have unobstructed flow through the 33 G tube, without leakage through the 26 G tube. 5. Allow glue to dry for several hours to overnight. 6. Remove the cannulae from the weigh dish and use a Crile hemostat to bend each cannula at a right angle, 10 mm from the end of the tube with the 1 mm–33 G extension (Fig. 17.2, see Note 2). 7. Insert the 10 mm bent cannula end 2 mm into a piece of PE-20 tubing cut to approximately 75 cm (2.5 feet). 8. Test liquid flow through the cannula by filling a 1 mL 26 G tuberculin syringe with sterile water and attaching the syringe to the opposite end of the PE-20 tubing. Push water through the syringe and examine the stream that emerges from the cannula on the other side. The liquid should flow freely in order for the cannula to properly infuse virus into the mouse brain. Discard any clogged cannulae. Cannulae that leak from the joints can be re-glued and tested again. 3.2. Stereotaxic Surgery
1. For bilateral infusion of virus, attach two cannulae connected to PE-20 tubing to the dual cannula insertion tool
Viral RNAi in Mouse Brain
291
using the 3/32 . Allen wrench supplied with the cannula insertion tool. Be careful not to over-tighten the plastic pieces holding the cannulae in place, as they can crack with too much applied force. About 5–6 mm of the stainless steel 33 G tube should extend past the cannula insertion tool so that it can be lowered into the mouse brain (Fig. 17.2). 2. Flush each cannula with 0.5 mL of 70% ethanol using a 26 G tuberculin syringe as described in Section 3.1, Step 8. Flush again with 0.5 mL of sterile water and set aside (see Note 3). Fill two 2 L Hamilton gastight syringes with sterile water and attach to the infusion pump. 3. Weigh the mouse and inject 120 mg/kg ketamine and 8 mg/kg xylazine i.p. (for example, inject 0.3 mL of 10 mg/mL ketamine and 0.1 mL of 2 mg/mL xylazine into a male C57BL/6J mouse weighing 25 g). Place mouse back into cage. Generally, mouse will be fully anesthetized after 4 min (see Note 4). 4. Meanwhile, prepare the SAS. Attach the microscope to the digital manipulator arm (DMA) and the centering height gauge to the mouse adaptor. Using the X-, Y-, and Z-axis knobs that control the DMA, focus the microscope on the orange cross on the centering height gauge and align the black crosshair in the microscope with the orange cross. Reset the digital monitor to 0 in all three (X, Y, and Z) coordinates. 5. Calibrate the alignment indicator. Remove the microscope, raise the DMA up approximately 1 mm using the Z-axis knob, and attach the alignment indicator to the DMA. Lower the alignment indicator using the Z-axis knob until it touches the top of the centering height gauge and the indicator needles on the dials point to 0. If one of the indicator needles is not at 0, adjust the dial on the alignment indicator so that both needles are at 0. This procedure only needs to be done once at the beginning of a group of surgeries. 6. Remove the alignment indicator and the centering height gauge, attach the microscope and lower the DMA back to 0, such that the X, Y, and Z coordinates on the digital monitor read 0. This reference point represents the position of bregma on the mouse skull. The DMA should not be moved until after the mouse is placed onto the SAS and bregma is in focus. 7. Place a bed of paper towels folded in half on the mouse adaptor. This will be used as a cushion for the mouse during the surgery.
292
Lasek and Azouaou
8. Confirm that the mouse is fully anesthetized by gentle paw pinch. The mouse should not react to the pinch. The mouse must be monitored carefully during the entire procedure to confirm sedation (see Note 4). Use the electric trimmer to shave the top of the mouse’s head, from between the ears anterior to just between the eyes (approximately 1.5 cm). 9. Swab the exposed scalp with a povidone–iodine swabstick to disinfect. Apply lidocaine jelly to the scalp and ophthalmic ointment to the eyes. 10. Make a 1 cm incision with a scalpel through the midline of the scalp, so that bregma and lambda (the intersection of the sagittal and lambdoid sutures) are exposed on the skull (Fig. 17.3).
Fig. 17.3. The position of bregma on the mouse skull. (A) The mouse skull, viewed from above. Bregma is visible as the intersection of the sagittal and coronal sutures, and lambda as the intersection of the sagittal and lambdoid sutures. (B) Cartoon showing the position of the sutures (curved lines) as seen through the microscope mounted on the DMA. The microscopic view is a mirror image to that seen with the naked eye. The black cross represents the crosshair of the microscope, which will be the 0 point on the digital monitor. The crosshair is not directly aligned with bregma, but instead is positioned as the best fit of lines through the coronal and sagittal sutures.
11. Using the wooden end of a cotton-tipped applicator, push the mouse’s tongue out of and to the left of the mouth. Hold the mouth open with the wooden end of the cotton applicator and insert the palate bar into the mouth, so that the top incisors fit into the hole on the palate bar. Slide the nose clamp forward and gently tighten the screw to hold the nose in place. Do not apply too much pressure on the nose. 12. Move the ear bars so that the tips are anterior to the ear canals and touching the zygomatic arch to hold the head in place (see Note 5). Tighten the ear bar screws. Clean the skull with a cotton applicator soaked in 100% ethanol.
Viral RNAi in Mouse Brain
293
13. View the mouse head through the microscope and use the X, Y, and Z knobs on the mouse adaptor (not the DMA) to move the mouse head and bring bregma into focus. Note that the skull is viewed as a mirror image through the microscope. Align the black crosshair in the microscope with bregma (Fig. 17.3, see Note 6) by moving the mouse head. The digital monitor should still read 0 in all coordinates. 14. Move the DMA in the Y-axis to lambda. The digital monitor will reflect the distance in mm from bregma to lambda (see Note 7). If the crosshair on the microscope does not align with the center of lambda, the mouse head can be rotated using the sagittal alignment knob on the right front of the mouse adaptor plate. 15. Move the DMA in the Y-axis exactly half of the distance between bregma and lambda. Remove the microscope from the DMA, raise the DMA up 1 mm, and attach the alignment indicator. The alignment indicator in this step is used to level the skull in the sagittal and coronal planes. Adjust the number on the vernier scale on the alignment indicator to the distance between bregma and lambda. Lower the alignment indicator in the Z-axis using the DMA so that the touch probes are on the top of the skull, to each side of midline in the sagittal plane, and the indicator needles read 0. If one needle reads 0 before the other, then the skull is tilted and must be adjusted. Loosen the screw holding the sagittal (ML) tilt knob in place and turn the knob to adjust the tilt of the skull so that the indicator needles both read 0 when the probes are touching the skull. 16. Raise the alignment indicator in the Z-axis 1 mm using the DMA, turn the indicator 90o so that one touch probe is on bregma and the other on lambda. Lower the DMA in the Z-axis until the needles read 0. Loosen the screw holding the coronal (AP) tilt knob in place and turn the knob to adjust the tilt as in Step 15. Raise the alignment indicator 1 mm using the DMA, turn the indicator back to the original position, and lower the indicator again and read the needles to confirm that the head did not shift in the sagittal plane during the coronal tilt adjustment. 17. Remove the alignment indicator from the DMA and attach the microscope. Move the DMA in the Y-axis back to 0 and confirm that bregma is appropriately positioned and the digital display reads 0. Do not move the mouse head using the mouse adaptor knobs after this point. Instead move the
294
Lasek and Azouaou
DMA in the X- and Y-axes to the appropriate stereotaxic coordinates which will be displayed on the digital monitor. 18. Unscrew the collet nut from the drill and insert the drill bit into the collet. Screw the collet nut with bit onto the drill and tighten with the wrenches supplied with the drill. Raise the DMA in the Z-axis 5 mm and attach the drill to the DMA. 19. Apply a drop of sterile saline to the mouse skull right below the drill. Turn on the drill and lower the DMA in the Z-axis to drill a hole in the skull approximately 1 mm deep (see Note 8). 20. If injecting virus bilaterally, a second drill hole is made in the appropriate coordinate on the opposite side of midline. 21. Remove the drill from the DMA. 3.3. Virus Infusion
1. Flush one cannula attached to the dual cannula insertion tool (set aside in Section 3.2, Step 2) with 0.2 mL of sterile water as described in Section 3.1, Step 8. 2. Snip 1–2 mm from the end of the PE-20 tubing using stainless steel utility scissors. 3. Push 1 L of water out of the Hamilton gastight syringe on the infusion pump and attach the PE-20 tubing to the syringe. 4. Repeat steps 1–3 for the second cannula and syringe. 5. Touch the ends of the cannulae to a paper towel to remove droplets hanging from the ends. Push the remaining 1 L of water out of the syringes and observe the droplets that appear at the ends of the cannulae, indicating that the syringe and PE-20 tubing are properly attached. 6. Touch the ends of the cannulae to the paper towel to dry. Pull 0.2 L of air into the cannulae using each attached syringe. 7. Centrifuge the virus in a 1.5 mL Eppendorf tube in a microcentrifuge at top speed for 10 s. With a pipette, transfer 10 L of virus to the inside edge of the cap of the Eppendorf tube. 8. Hold the dual cannula insertion tool such that 1 mm of the cannulae tips are submerged in the virus in the tube cap. Pull each syringe back 1.1 L, so that the final reading on the syringe is 1.3 L. The 0.2 L air bubble and virus will be visible in the PE-20 tubing. Mark the front edge of the air bubble with a marker on the PE-20 tubing. 9. Raise the DMA in the Z-axis 5 mm. Attach the dual cannula insertion tool to the DMA. Using the X- and Y-axis knobs
Viral RNAi in Mouse Brain
295
on the DMA, bring the cannulae back to 0, such that they are centered on bregma. Lower the DMA so that the cannulae touch the top of the skull. Reset the Z-axis reading on the digital monitor to 0. This calibrates the cannulae in the Z-axis to the top of the skull. 10. Move the DMA in the X- and Y-axes back to the proper coordinates for injection. The dual cannula insertion tool is configured so that the left cannula is set at the negative X (ML) coordinate on the digital monitor, and the right cannulae is then manually adjusted using the knob on the dual cannula insertion tool to the opposite side. Continue to adjust the cannulae so that they fit directly over the holes drilled in the skull in Section 3.2, Steps 19 and 20. 11. Slowly lower the cannulae using the Z-axis knob on the DMA directly into the holes to the proper Z (DV) coordinate, visible on the digital monitor. 12. Set the infusion pump to infuse 1 L of virus at 0.2 L per minute (see Note 9). After the 5 min infusion, wait an additional 5 min. Withdraw the cannulae slowly from the brain using the Z-axis knob on the DMA. 13. Remove the dual cannula insertion tool from the DMA and immediately flush with 0.5 mL 70% ethanol. Set the tool aside. 14. Close the wound. Using Adson forceps to hold the wound closed, apply a small drop of tissue adhesive. Hold for 3–5 s and quickly remove the forceps (see Note 10). 15. Remove the animal from the mouse adaptor and inject 0.5 mL saline s.c. Place the animal in a cage filled with alpha cellulose bedding and partially positioned on an electric heating pad set to medium. 16. Monitor the animal until it recovers from the anesthetic and is moving freely around the cage. Place a water bottle filled with children’s Tylenol on the cage and allow the mouse free access to the solution for the first 2 days of recovery. If the scalp wound becomes infected, apply a topical antibiotic ointment.
4. Notes 1. Stereotaxic coordinates for injection into brain regions other than those listed in Table 17.1 should be calculated from a mouse brain atlas (9). Prior to surgery, use the atlas to determine optimal coordinates for injection in
296
Lasek and Azouaou
reference to bregma in the AP, DV, and ML planes. Test these coordinates on a small number of mice and modify as needed prior to beginning viral infusion surgeries. For the purpose of refining surgical coordinates, rather than injecting virus, 1 L of India ink diluted 1:10 in PBS may be infused, the mouse immediately euthanized, and the brain sectioned and examined under a light microscope to determine injection accuracy. Note that coordinates for a specific brain region may vary for different strains, ages, and sexes of mice. However, within a particular inbred strain of mice, coordinates for a specific region tend to be consistent. 2. When bending a cannula with the Crile hemostat, bend once at a 90o angle. Over-bending can break or occlude the cannula. 3. If flow through the cannula during the flushing step is blocked or very slow (i.e., dripping instead of a stream of fluid), the cannula should be removed from the PE-20 tubing and replaced with a fresh cannula. Several mm of the PE-20 tubing should be trimmed before inserting the new cannula to ensure a tight seal. The flow through the cannula should be tested before each surgery. 4. The amount of anesthetic to be used has been determined for 9–10 week-old male C57BL/6J mice weighing 25– 30 g. The amount of anesthetic for different strains and ages of mice will need to be determined empirically. In addition, any mouse not fully anesthetized within 15 min after the initial injection must be boosted with another injection, generally with 0.1 mL of 10 mg/mL ketamine. If the mouse is mounted on the mouse adaptor and begins to awaken during surgery, the injection can be given to the mouse while it is still mounted in the head holder. Lift up the back of the mouse by the tail and inject i.p. Wait 3–4 min before resuming surgery. If the mouse does not respond to the anesthetic, remove the mouse from the head holder and wait another 15–20 min before determining whether an additional booster injection is needed. Mice can easily overdose on anesthetic, so it is better to determine the most effective initial dose for all strains used rather than boosting multiple times. 5. The ear bars can also be inserted into the ear canals to hold the head in place. However, it is very easy to puncture the tympanum or to apply too much pressure and cause breathing difficulties. The head must be securely fixed in place using the nose clamp and ear bars. If the head moves during leveling or drilling, it will affect stereotaxic accuracy.
Viral RNAi in Mouse Brain
297
6. Bregma can vary among mice in the angle and point of intersection of each side of the coronal suture with the sagittal suture. As a result, it can be difficult to position the crosshair on the microscope directly on bregma. A solution to this problem is to find the best fit of lines through the coronal and sagittal sutures (Fig. 17.3). The crosshair on the microscope is placed at the intersection point of these lines; this is defined as the 0 point, represented on the digital monitor, instead of bregma per se. One can choose his or her own method of defining the 0 point, as long as the 0 point is consistently employed and the stereotaxic coordinates for a particular brain region with respect to this point are experimentally determined. 7. The average distance between bregma and lambda for an adult male C57BL/6J mouse is 4.2 mm (9). We have found that this distance can vary from 3.8 to 4.6 mm. For injections to coordinates located between bregma and lambda, an adjustment can be made to the AP coordinate based on this distance. For example, if the distance from bregma to lambda is 3.8 mm, divide 4.2 by 3.8 to obtain 0.90. This factor is multiplied by the original AP coordinate of −3.2 (for VTA injection) to give −2.88. The new AP coordinate for VTA injection would be rounded to −2.9. This scaling of the AP coordinate is particularly important for smaller regions that are harder to target, such as the VTA. 8. Drill holes deep enough to puncture the skull, but not so deep that the cortex is damaged. Raising and lowering the drill using the Z-axis knob on the DMA helps monitoring the progress of the hole. It is important that the hole is drilled completely through the skull, as otherwise it is difficult to lower the cannulae into the holes. We use a #85 drill bit, which generates a 0.28 mm diameter hole. The outer diameter of the 33 G cannula is 0.20 mm and can just fit into the hole. It can be difficult to align the cannula with a hole this size. A solution to this problem is to use a slightly larger drill bit (#75), which generates a 0.53 mm diameter hole. However, the smaller the hole, the less damage to the mouse skull during surgery and the more accurate insertion of the cannula into the brain region of interest. 9. During virus infusion, it is important to monitor the movement of the air bubble and virus through the PE-20 tubing, thus the need to mark the tubing at the front edge of the air bubble before starting the infusion. If the virus does not appear to be entering the brain (i.e., the bubble is not moving), then check the seal between the PE-20 tubing and the syringe. The tubing can also be “flicked” with a finger near
298
Lasek and Azouaou
the air bubble to promote movement. However, do not force the syringe manually. This causes excessive damage to the infusion site. The virus infusion steps described in Section 3.3 should be followed for each individual mouse. Loading a larger amount of virus into the tubing and trying to infuse multiple mice will not work, primarily because the 33 G cannula is easily clogged after lowering into the brain. 10. The tissue adhesive acts very quickly, so it is important to hold the wound closed with the forceps for only a few seconds. Otherwise, the scalp sticks to the forceps.
Acknowledgments The authors would like to thank Ulrike Heberlein for her support of this project, Heidi Lesscher for amygdala injection coordinates, Viktor Kharazia and Rajani Maiya for photographic assistance, and David Kapfhamer and Karen Berger for critical reading of the manuscript. This work was supported by funds awarded to Ulrike Heberlein by the State of California for medical research on alcohol and substance abuse through the University of California at San Francisco and the Alcohol Center for Translational Genetics at the Ernest Gallo Clinic and Research Center. References 1. Kim, D. and Rossi, J. (2008) RNAi mechanisms and applications. Biotechniques 44, 613–616. 2. Salahpour, A., Medvedev, I. O., Beaulieu, J. M., Gainetdinov, R. R. and Caron, M. G. (2007) Local knockdown of genes in the brain using small interfering RNA: a phenotypic comparison with knockout animals. Biol Psychiatry 61, 65–69. 3. Hommel, J. D., Sears, R. M., Georgescu, D., Simmons, D. L. and DiLeone, R. J. (2003) Local gene knockdown in the brain using viral-mediated RNA interference. Nat Med 9, 1539–1544. 4. Pulipparacharuvil, S., Renthal, W., Hale, C. F., Taniguchi, M., Xiao, G., et al. (2008) Cocaine regulates MEF2 to control synaptic and behavioral plasticity. Neuron 59, 621– 633. 5. Maskos, U., Molles, B. E., Pons, S., Besson, M., Guiard, B. P., et al. (2005) Nicotine reinforcement and cognition restored by targeted
6.
7.
8.
9.
expression of nicotinic receptors. Nature 436, 103–107. Lasek, A. W., Janak, P. H., He, L., Whistler, J. L. and Heberlein, U. (2007) Downregulation of mu opioid receptor by RNA interference in the ventral tegmental area reduces ethanol consumption in mice. Genes Brain Behav 6, 728–735. Crawley, J. N., Belknap, J. K., Collins, A., Crabbe, J. C., Frankel, W., et al. (1997) Behavioral phenotypes of inbred mouse strains: implications and recommendations for molecular studies. Psychopharmacology (Berl) 132, 107–124. Lesscher, H. M., McMahon, T., Lasek, A. W., Chou, W. H., Connolly, J., et al. (2008) Amygdala protein kinase C epsilon regulates corticotropin-releasing factor and anxietylike behavior. Genes Brain Behav 7, 323–333. Paxinos, G. and Franklin, K. B. J. (2001) The Mouse Brain in Stereotaxic Coordinates, 2nd ed. Academic Press, San Diego.
Chapter 18 Experimental Models of Anxiety for Drug Discovery and Brain Research Peter C. Hart, Carisa L. Bergner, Amanda N. Smolinsky, Brett D. Dufour, Rupert J. Egan, Justin L. LaPorte, and Allan V. Kalueff Abstract Animal models have been vital to recent advances in experimental neuroscience, including the modeling of common human brain disorders such as anxiety, depression, and schizophrenia. As mice express robust anxiety-like behaviors when exposed to stressors (e.g., novelty, bright light, or social confrontation), these phenotypes have clear utility in testing the effects of psychotropic drugs. Of specific interest is the extent to which mouse models can be used for the screening of new anxiolytic drugs and verification of their possible applications in humans. To address this problem, the present chapter will review different experimental models of mouse anxiety and discuss their utility for testing anxiolytic and anxiogenic drugs. Detailed protocols will be provided for these paradigms, and possible confounds will be addressed accordingly. Key words: Anxiety, experimental animal models, anxiolytic drugs, anxiogenic drugs, biological psychiatry, exploration.
1. Introduction Animal models are widely used for simulating human brain disorders and for providing insight into their neurobiological mechanisms (1–4). The latter is of great interest in the current neuroscientific community, given the increasing use of laboratory animals for screening various classes of psychotropic drugs (5, 6). The use of mice has been particularly beneficial, since fine-tuned manipulations of selected genes have led to new animal models relevant to drug discovery (3, 4, 7, 8). G. Proetzel, M.V. Wiles (eds.), Mouse Models for Drug Discovery, Methods in Molecular Biology 602, DOI 10.1007/978-1-60761-058-8 18, © Humana Press, a part of Springer Science+Business Media, LLC 2010
299
300
Hart et al.
It is important to understand, however, that any animal experiment in the laboratory is an artificial situation, and it may be biologically different from the natural behavior of the animal. Thus, it is crucial to correctly interpret the animal behavior observed in an experiment in order to identify parallels with specific human brain disorders. Although there are many other conceptual and methodological limitations of working with mice, this species shows much promise for future psychopharmacological research. In order for animal models to be useful, researchers must follow certain practices and methods which will optimize the translatability of data from animal models to human affective disorders. Here, we will present a broad review of some reliable methods of analyzing mouse anxiety, and their utility for screening for anxiolytic therapeutic agents. All these tests are of a complex nature and we would suggest that the reader explore each system to better understand the variables and subtleties of each test. We will also discuss how these protocols can be applied correctly, in order to avoid confounding experimental data.
2. Materials 2.1. Animals
1. Various inbred, selectively bred (for specific behavioral/physiological phenotypes), and genetically modified (mutant or transgenic) mice may be used, and some searchable online databases, such as Mouse Phenome Project (www.jax.org/phenome) or Mouse Genome Informatics (www.informatics.jax.org/), may provide appropriate strains for studying mouse anxiety. We recommend using most of the inbred strains listed in the Tier 1 list of the Mouse Phenome Project database, especially C57BL/6J, A/J, and 129S1/SvImJ mice (see http://phenome.jax.org/pub-cgi/phenome/mpdcgi?rtn= docs/pristrains for details) (see Note 1). 2. Generally, several different models of anxiety that target different domains (e.g., locomotion/exploration, risk assessment, defensive responses) are necessary in order to more fully characterize drug effects or a mutant mouse phenotype. The use of a single model, or only models targeting one particular behavioral domain, may not be sufficient. 3. Researchers should also take other factors like age, weight, sex, stage of estrous cycle, diet, and housing situation into account when designing experiments (see Note 2).
Experimental Models of Anxiety
2.2. Housing
301
1. If mice are obtained from a commercial vendor or another laboratory, allow at least 1 week acclimation from shipping stress. In most cases, a much longer time (e.g., 1 month) may be required for a better acclimation. 2. Housing animals in groups will help avoid social isolation stress/anxiety, but keeping groups small enough (e.g., not more than five animals per cage) will be necessary to avoid overcrowding stress. While overcrowding of mice may cause significant levels of stress, single housing is equally detrimental to experimental models. For example, social isolation may lead to altered neurobiology, increased basal anxiety, reduced exploration, and a profound vulnerability to depression-like behaviors (9, 10). 3. The room in which mice are housed should be kept at approximately 21◦ C, on a 12/12-h light cycle. As mice are nocturnal, the light cycle may be inverted if spontaneous activity measures are needed. 4. Food and water should be freely available, unless the intake is being controlled for experimental purposes. 5. Utilize plastic, solid-floored cages with sufficient space for mice to exercise and fully rear up. Note that enrichment items, such as cardboard tunnels, can improve general welfare but may also affect experimental outcomes or increase territorial aggression.
2.3. Drugs
1. All experimental protocols described here are compatible with drug testing. Researchers may choose from various antidepressants, anxiogenics, anxiolytics, or other psychotropic drugs, administered with a vehicle (e.g., saline). A typical experiment may include one or several drug-treated groups (e.g., several doses or several pre-treatment times) compared to a vehicle-treated group of mice. Usually (unless stated otherwise), 10 animals per experimental group will be needed, also providing adequate statistical power (see further). However, if the effects of the drugs are particularly robust, a smaller n (e.g., n = 7–8) may suffice. For mild effects, a larger number of animals (n = 15–16) may be required.
2.4. Observations, Video Recording, and General Procedures
1. Video-tracking software: Ethovision, Noldus, Nijmegen, Netherlands Videotrack system, Viewpoint, Lyon, France Loco-, Maze-, and Top Scan, Clever Sys Inc, Reston VA, USA
302
Hart et al.
2. Photobeam-based activity monitoring: Columbus Instruments, Columbus OH, USA Coulbourn Instruments, Whitehall PA, USA 3. Vibration-based activity monitoring: Laboras/Metris, Hoofddorp, Netherlands Bioseb, Vitrolles, France 2.5. Requirements for Experimental Models
1. Elevated-plus maze (EPM): Elevated maze with two open and two closed arms in the shape of a plus, made of steel, fiberboard, or Plexiglas (either transparent of painted matte black), see Table 18.1. Arms are typically 30 cm long and 5 cm wide. The apparatus is usually elevated 40–60 cm on sturdy legs (1, 9, 11).
Table 18.1 Selected commercial suppliers of behavioral equipment for anxiety research Test apparatus
Manufacturer
Company web site
Elevated-plus maze
Panlab, Barcelona, Spain Columbus Instruments, Columbus OH, USA
www.panlab.com www.colinst.com
Open field
ANY-Maze by Stoelting, Wood Dale IL, USA Noldus, Wageningen, Netherlands Panlab, Barcelona, Spain San Diego Instruments, San Diego CA, USA
www.anymaze.com www.noldus.com www.panlab.com www.sandiegoinstruments.com
Startle response
San Diego Instruments, San Diego CA, USA Columbus Instruments, Columbus OH, USA
www.sandiegoinstruments.com www.colinst.com
Hole board
Columbus Instruments, Columbus OH, USA myNeuroLab, St Louis MO, USA
www.colinst.com www.myneurolab.com
Light-dark box
Panlab, Barcelona, Spain Stoelting, Wood Dale IL, USA
www.panlab.com www.anymaze.com
2. Open field: Enclosed 50 × 30 cm wood, plastic, or Plexiglas arena, marked into 10-cm squares (Table 18.1). Gray or black arenas are typically used. If an arena is not available, a large animal cage marked into squares with indelible ink may be used (10, 12).
Experimental Models of Anxiety
303
3. Marble burying test: Woodchip bedding (e.g., aspen chips), up to 20 marbles (15 mm in diameter). Animal cages (e.g., large cage 30 × 20 cm for 20 marbles and smaller cages for 6–8 marbles) (13–16). 4. Defensive shock-prod burying test: Familiar test cage or home cage with plentiful bedding and a hole in the wall 2 cm above bedding (6, 17). Electrical probe connected to a shock source. Ruler for measuring depth to which prod is buried. Optional: large (e.g., 10 cm) object associated with shock. 5. Grooming analysis algorithm: Small (e.g., 20 × 20 × 30 cm) transparent observation box. Stressors to induce grooming: e.g., novel environment, predator exposure, bright light, or other means of artificially inducing grooming (e.g., water mist). Optional: video camera for subsequent frame-by-frame analysis (8). 6. Startle response: Observation box (similar to the open field test). Conditioned stimulus: e.g., a light, paired with a footshock, Table 18.1. Startle stimulus, such as an air puff or loud noise (1, 7, 18). 7. Social interaction test: Low-anxiety version: test apparatus (similar to the open field test) familiar to the animals, with low illumination. Mid–low anxiety version: familiar test apparatus with high illumination. Mid–high anxiety version: unfamiliar test apparatus with low illumination. High anxiety version: unfamiliar test apparatus with high illumination (1). 8. Suok test: Test apparatus: 2.6-m aluminum tube, 2 cm in diameter, marked into 10-cm segments with indelible ink) with fixed 50 × 50 × 1 cm Plexiglas side walls to prevent escape, elevated 20 cm from a cushioned floor. Optional (the light-dark version of the test): several 60-W light bulbs suspended 40 cm above one half of the test apparatus (18, 20). 9. Light-Dark Box Test: Test apparatus: a 2-compartment box, 30 × 30 × 30 cm each; with one black, and one transparent brightly illuminated boxes, separated by a sliding door (5).
304
Hart et al.
10. Stress-induced hyperthermia (SIH): Oiled rectal thermometer with rounded tip, up to 3 mm thick: Surgilube sterile surgical lubricant by Fougera & Co. (Melville, NY, USA). K-Y lubricant by Johnson & Johnson (Waltham, MA, USA). MLT1404 rectal probe for mice by Adinstruments, Inc (Colorado Springs, CO, USA). Cage or box (as in the open field test) to which mice can be transferred (21). 11. Hole-Board Test: Test apparatus (similar to the open field test) with holeboard insert (Table 18.1). The floor has four or more identical holes approximately 3 cm in diameter (22). 12. Rat Exposure Test: Medium (e.g., 40 × 30 × 30 cm) transparent observation box, with a wire mesh separating the two halves of the box. Small (e.g., 8 × 8 × 12 cm) black Plexiglas box, serving as the starting placement (home chamber for the mouse). Transparent Plexiglas tube (e.g., 4.5 cm in diameter, 13 cm in length) connecting the small black box to the medium transparent box (23). 13. Novel Object Test: Test apparatus similar to the open field test (see above). Novel objects: e.g., Mega Bloks structures (24).
3. Methods 3.1. Drug Administration
1. Common routes of injection include systemic [(intraperitoneal (i.p.), intramuscular (i.m.), intravenous (i.v.), per oral (p.o.), subcutaneous (s.c.)] and local [intracerebral (i.c.), intranasal (i.n.) or intracerebroventricular (i.c.v)). Continuous drug infusion (using osmotic pumps, such as Alzet pumps) at a constant rate may be used to improve the availability of the drug, and implantable depots can be used for s.c. drug administration to achieve a lasting therapeutic effect. Though not as commonly used in anxiety research, i.n. drug administration is a rapid, non-invasive method for drug delivery (25). By this method, drugs can be administered either by pipetting small (6 l) drops of a drug solution into each nostril once per minute (25) or by placing a 10-l drop of a drug to be inhaled on the end of the snout
Experimental Models of Anxiety
305
(26). As i.n. administration delivers the drug to the brain directly via the olfactory nerve and olfactory epithelium, this method may be favored for administration of drugs that are rapidly metabolized when given systemically or have difficulty crossing the blood–brain barrier (27). For example, in rats, i.n. dopamine has been shown to decrease grooming activity and increase locomotor activity in the open field at one tenth of the dose needed to observe these effects when the drug is administered i.p. (27); also see data on its antidepressant effects (28). 2. Route of administration, dose, and pre-treatment time vary depending on strain sensitivity and the drug being used. 3.2. Observations, Video Recording, and General Procedures
1. A computer, digital camera mounted above the test apparatus and video-tracking software will aid researchers in the collection of accurate behavioral data. 2. Alternative methods of behavioral tracking include photobeam-based or vibration-based activity monitoring. 3. In addition to automated tracking, an observer with a timer and data sheet to tally behaviors will allow comparison of data if video tracking is unreliable due to poor detectability (from poor angle or bad contrast; e.g., if fur color matches the background). 4. Observers must refrain from making noise or movement, as their presence may alter animal behavior. Assess intra- and inter-rater reliability for consistency. 5. Allow the animals at least 1 h acclimation after their transfer from the animal holding room to the experimental room. 6. Mice should be introduced to the testing environment during their normal waking cycle to prevent possible confounds. When performing ethological analysis as part of a battery of tests, consider how effects of these tests (such as habituation) may confound the mouse performance and drug sensitivity. 7. After each testing session, clean the equipment (e.g., with a 30% ethanol solution) to eliminate olfactory cues.
3.3. Data Analysis
1. Behavioral data may be analyzed with the Mann–Whitney U test for comparing two groups (parametric Student’s t test may be used only if data are normally distributed), or analysis of variance (ANOVA) for multiple groups, followed by an appropriate post hoc test. 2. Some experiments may require one-way ANOVA with repeated measures, while for more complex studies
306
Hart et al.
(e.g., those including treatment, genotype, sex, and/or stress) n-way ANOVA may be used. 3. Analysis of statistical power is becoming particularly important in animal research. Effect size (the difference between means of the two groups) in neurobehavioral research can be small (0.2), moderate (0.5), or large (0.8). Statistical power is the probability of finding statistical significance for a true hypothesis, and its common value used in behavioral research is 0.8. Factors that can affect statistical power include the experimental design (independent/dependent groups), one- or two-tailed hypotheses, statistical test chosen, effect size, and sample size. In order to determine an effect size, the researcher may rely on effect size reported in similar studies or can decide on it based on the goals of the study (e.g., use large effect size for robust phenotypes and small effect size for less profound differences). The researcher can then use statistical power-analyzing software to determine the sample size required for the level of power decided upon. The choice of sample size, based on power calculations, is increasingly important for institutional research ethics committees, to prove that neither too few nor too many subjects are used in the proposed research. 3.4. Elevated-Plus Maze (EPM) Test
Possessing good face-, construct-, and predictive validity, the EPM is a reliable and pharmacologically sensitive paradigm based on the conflict between innate rodent desire to explore and the fear of open, elevated areas (3, 5). Anxious mice generally have a lower ratio of open arm entries to total arm entries and display fewer explorative measures such as rears, wall leans, or head dips (3, 5). Anxiety also increases EPM freezing and stretch-attend postures. After administration of anxiolytic drugs (e.g., diazepam, chlordiazepoxide, ethanol), mice generally display more exploratory behaviors, a greater number of open arm entries, and an increased duration of time spent on open arms (9, 11). Anxiogenic drugs (e.g., pentylenetetrazole, picrotoxin) produce the opposite behavioral effects in this model. 1. Place rodent on the central platform of the EPM facing either an open arm or a closed arm consistently at the start of each trial (see Notes 1–3). 2. The open arm, closed arm, and total (open + closed arm) activity can be recorded for 5–10 min using a video-tracking system, while the researchers simultaneously document the number of arm entries (all four paws are on the arm) and time spent on each open arm (see Notes 4 and 5).
3.5. Open Field Test (OFT)
This test is based on the balance between the animal’s natural drive to explore novelty and its aversion to open illuminated areas. Measuring exploratory behaviors and generalized motor activity,
Experimental Models of Anxiety
307
the open field test is simple and the most frequently used model of mouse anxiety (2). In general, anxious mice exhibit more freezing, less time spent and a lower percentage of ambulation in the center of the arena (thigmotaxis), and fewer exploratory behaviors (see Note 6). Anxiolytic drugs generally increase exploration and reduce freezing and thigmotaxis (12). 1. After the apparatus is divided into central and peripheral zones, mice are placed consistently in a corner, or the center of the open field arena, and allowed to explore for 5–10 min. 2. Behavioral measurements can be recorded automatically with appropriate software and include time spent in the central area, distance traveled in the center as a ratio of total distance traveled, ambulation duration, time spent immobile (freezing), defecation score, and vertical activity such as rearing and wall leans (12) (see Notes 7 and 8).
3.6. Marble Burying Test
While not a direct model of anxiety per se, this simple test represents a pharmacologically sensitive method for assessing digging activity – a species-typical response to anxiogenic stimuli (13, 15, 29). Digging behavior is attenuated by low (nonsedative) doses of anxiolytic benzodiazepines and other ligands (14, 16). Control mice can be expected to bury roughly 75% of marbles, whereas drug-treated mice show a marked decrease in digging activity (30, 13). Mouse strains with high basal anxiety levels (e.g., 129S1/SvImJ) should be avoided when testing anxiogenic drugs to avoid a ceiling effect. Likewise, mice with very low basal anxiety may show poor results when examining the effects of anxiolytic drugs (see Note 9). 1. Cages should be filled with wood chip bedding approximately 5 cm deep. The bedding must be flattened to create an even surface. Use the same volume (e.g., 300 ml per 26 × 16 cm cage) of bedding in each cage. Although wood chips and shavings are most commonly used, sawdust has also been used with similar effectiveness in this model (32, 33). There are many suppliers of these types of bedding, including Shavings-Direct (www.shavings-direct.com), Doctors Foster and Smith (www.drsfostersmith.com), and local pet stores (e.g., Petsmart, www.petsmart.com). The important thing is to be consistent within experiments. 2. Mouse cage dimensions vary, but this test has been used effectively with 26 × 16 cm (for 6–8 marbles) or 30 × 20 cm (for 20 marbles) (14). Marbles should be placed on the surface of the bedding in a regular pattern, roughly 4 cm apart. 3. Place one mouse in each cage. After 30 min, count the number of buried marbles. Any marble covered 2/3 of its depth with bedding is considered “buried” (14).
308
Hart et al.
Alternatively, count fully covered (1/1) and partially covered (2/3) marbles separately, also calculating the sum of the latter two categories (see Notes 9–11). 4. Use a new clean cage with fresh bedding for each animal. Wash the marbles with ethanol after each test. 3.7. Defensive (Shock-Prod) Burying Test
Similar to the marble burying test, this paradigm is another pharmacologically sensitive method to assess rodent anxiety. Mice usually bury noxious stimuli posing an immediate threat (e.g., electrified shock-prod). The test has pharmacological validity, as benzodiazepines and the serotonergic anxiolytics potently suppress shock-prod burying in a dose-dependent manner, whereas anxiogenic drugs have been proven to increase this behavior (6). 1. In a standard-sized mouse cage (see above) with bedding 5 cm deep, insert a wire-wrapped prod (6–7 cm long) through a hole 2 cm above the bedding surface. 2. After the initial contact with the bare wires and the subsequent shock, record the behavior of the animal for 10–15 min. Behavioral measures of activity may include prod-directed burying, burying latency, height of pile at prod base, prod contacts (number, duration), prod contact latency, and stretch-attend postures directed at the prod (6) (see Note 12).
3.8. Grooming Analysis Algorithm
Anxious mice tend to display a disorganized behavioral sequencing of grooming (higher percentage of incorrect transitions, more interrupted bouts) and a longer duration of this behavior. In contrast, anxiolytic benzodiazepines normalize mouse grooming sequencing by significantly reducing interrupted bouts and incorrect transitions (8) (see Notes 13 and 14). 1. Induce grooming through exposure to novelty or a stressor. Alternatively, mist the animal with water using a spray bottle. Place the animal in a small transparent observation box for 5 or 10 min. 2. If using a video camera, begin recording. With a stopwatch, record cumulative measures of grooming activity, such as latency to onset, time spent grooming, and total number of bouts. A new bout takes place after an interruption of greater than 6 s; bouts containing pauses of less than 6 s are deemed “interrupted.” Additionally, record the patterning of each bout using the following scale: 0 – no grooming, 1 – paw licking, 2 – nose/face/head wash, 3 – body grooming, 4 – leg grooming, and 5 – tail/genitals grooming. 3. There are several types of incorrect transitions, including skipped (e.g., 1–4, 3–5), reversed (e.g., 3–2, 5–3), prematurely terminated (e.g., 3–0, 4–0), and incorrectly
Experimental Models of Anxiety
309
initiated (e.g., 0–3, 0–5) transitions. Calculate the percentage of interrupted bouts and the percentage of incorrect transitions; see (8) for details (see Notes 15 and 16). 3.9. Startle Response
The startle response test pairs a conditioned stimulus (sound, light) with a footshock to induce an anxiogenic “startle” response in mice. While the sensitivity of this test to many drugs is yet to be established, benzodiazepine and serotonergic anxiolytics have been effective in reducing the startle response (1). Since this model seems to be unaffected by motor phenotypes, activity levels, or neurological deficits, this test (unlike many other anxiety models discussed here) allows researchers to study mouse anxiety without these confounding factors (see Note 17). 1. In a conditioning trial, a conditioned stimulus (usually a light: e.g., 15 W) is paired with a footshock. The timing of the conditioned stimulus and footshock can be controlled by the data acquisition software for consistency (7). 2. In a separate trial 24 h later, the animals are presented with a startle stimulus (e.g., loud noise 70–80 db, or air puff) and their activity is recorded as a baseline. The startle stimulus can be presented in four blocks of five startles each, with 30–35 s between each startle stimulus (7). Peak and amplitude of the startle response can be recorded (e.g., using a piezoelectric accelerometer) and digitized (18). 3. 24 h later in testing trials, the conditioned stimulus is displayed immediately prior to the startle stimulus, and the observed response is compared to the baseline startle response. Stimuli should be presented when the animal is quiet and inactive (7) (see Notes 18 and 19).
3.10. Social Interaction Test
The social interaction test is a useful drug-sensitive approach to assess anxiety in mice, subjected to the apparatus similar to the OFT. There are four testing conditions which introduce varying levels of stress: (1) familiar test apparatus and low illumination; (2) familiar test apparatus and high illumination; (3) unfamiliar test apparatus and low illumination; and (4) unfamiliar test apparatus and high illumination. The level of anxiety across these conditions ranges from low to high, respectively. Overall, the duration and frequency of social interactions negatively correlate with anxiety. Because this test successfully isolates levels of anxiety (high vs. low) in the subjects, it has been used for pharmacological screening of both anxiolytic and anxiogenic drugs in their effectiveness for increasing or decreasing social interaction, respectively (1). 1. The test environment should be the same in all conditions, except for the test apparatus (familiar or unfamiliar) and the lighting (low or high illumination).
310
Hart et al.
2. Introduce the two mice into the test environment for 5 or 10 min, recording the duration and frequency of all social interactions (e.g., sniffing, following, chasing, touching, and biting). All behavioral endpoints mentioned should be recorded, including the frequency (number of bouts) and duration (cumulative of all bouts per trial). The best way would be to use a video-tracking system with either automatic or manual scoring, although it is possible for the observers to record these behaviors manually in real time. 3. After obtaining baseline data for each condition, administer an anxiolytic or anxiogenic drug to the mice. The same test (Step 2) can be conducted and analyzed relative to baseline data. Alternatively, compare drug-treated with saline-treated groups (only one animal in the interacting pair receives the drug) (see Notes 20–22). 3.11. Suok Test
The Suok test simultaneously examines anxiety, vestibular, and neuromuscular deficits by combining an unstable rod with novelty. To analyze anxiety, the threats of height, loss of balance, and novelty are presented and animal exploration is recorded. Anxiolytic or anxiogenic drugs will increase or decrease animal exploration, respectively. Risk assessment and vegetative behaviors are generally higher in anxious mice. The model is also sensitive to anxiety-evoked balancing deficits, since administration of anxiogenic drugs increases the number of falls and missteps, while anxiolytics generally improve balancing (18, 20). A light-dark modification of the test may also be employed, as the illuminated environment will represent an additional stressor. We recommend reviewing the Mouse Phenome and Mouse Genome Informatics Databases, for example, of anxious mouse strains, mice displaying low motor or vertical activity, hyperactive mice, and mice with sensory deficits or mouse strains with vestibular difficulties. Researchers should easily find mice that fit their specific experimental needs. 1. Place individual mice in the center of the apparatus facing either end (or, in the light-dark modification, orient the animal facing the dark end) (see Note 23). 2. Standing approximately 2 m away from the apparatus, record the following behavioral measures (for 5–10 min per animal): horizontal exploration activity (latency to leave central zone, number of segments visited with four paws, distance traveled, number of stops, time spent immobile, average inter-stop distance, number of stops near border separating light-dark areas of the apparatus), vertical exploration (number of vertical rears or wall leans), directed exploration (head dips, side looks), risk assessment behavior
Experimental Models of Anxiety
311
(stretch-attend postures), vegetative responses (number of defecation boli and urination spots), and vestibular/motor indices (number and latency of hind-leg slips and falls from rod). If the animal falls, replace it in the same position from where it fell (see Notes 24 and 25). 3.12. Light-Dark Box Test
This ethological model of anxiety measures the activity and time spent in brightly lit vs. dark compartments of the apparatus and is based on the animal’s innate desire to explore novel areas (5, 12). Anxious mice exhibit a profound preference for the dark area and display fewer exploratory behaviors (e.g., horizontal activity, vertical rears or wall leans) in the light. Increased duration of time spent in the light area and more exploratory behaviors can be seen following anxiolytic drug administration. 1. Place one animal into the dark compartment of the box for 5 min for acclimation 2. Lift the shutter to allow the mouse to move freely between the dark and the light compartments for 5 min. 3. Measure the latency to initial transition into the light box. Record the duration of time spent in each compartment, the number of transitions between them, and the distance traveled in each box. Additional indices may be vertical rearing, wall leans (in the light compartment), and the number of defecation boli (see Note 26).
3.13. Stress-Induced Hyperthermia (SIH)
This test relies on the evolutionarily important role of hyperthermia, a rise in body temperature upon encountering stressful stimuli which occurs across many species, including humans. In mouse SIH test, the insertion of a rectal thermometer records a 0.5–1.5◦ C increase in body temperature (see Note 27). SIH is reduced or prevented by different anxiolytic drugs, however, it seems to be unable to detect anxiogenic and antidepressant effects (21). 1. Animals should be put in individual cages the day before testing to avoid effects of acute isolation stress (see Note 28). 2. Baseline body temperature should be recorded. To test mouse rectal temperature, carefully insert a probe with a rounded tip (up to 3 mm thick) after dipping it in any kind of oil for lubrication. For example, use surgilube sterile surgical lubricant by Fougera & Co. (Melville, NY, USA), K-Y lubricant by Johnson & Johnson (Waltham, MA, USA), and MLT1404 rectal probe for mice by Adinstruments, Inc (Colorado Springs, CO, USA). The probe should be inserted consistently, approximately 2–2.5 cm for 10 s (see Note 29).
312
Hart et al.
3. Present the mouse with a stressor, such as a novel cage, and document the change in internal temperature (see Note 30). 4. After testing, mice may be re-socialized in grouped housing. They may be retested after in 1-week intervals. Typically, 10–15 mice per group are sufficient to observe significant effects (34) (see Note 31). 5. It is also possible to use an implanted temperature sensor to monitor temperature remotely and without using this type of invasive measurement. For example, microchip transponders (ELAMS, BioMedic Data Systems, Inc., Seaford DE, USA) have been shown to reliably monitor temperature without significant difference from rectal temperature measurements (35). 3.14. Hole-Board Test
Conceptually similar to the open field test (OFT), the hole-board test focuses on specific head dipping behaviors. Head dipping, an indication of directed exploration, can be vigorously affected by various drug classes, including anxiolytic and anxiogenic drugs. Due to its short duration and quantifiable behavioral measures, this test is a readily available method for the testing of classic or novel drugs (22). Place mice individually in hole-board apparatus and record behavior for 5–15 min, documenting traditional exploratory behaviors (as in the OFT, see above) and the number of head dips (see Notes 7, 8, and 32).
3.15. Rat Exposure Test
This test utilizes the natural defensive “avoidance” behavioral response of mice to signs of potential danger, such as a natural predator (e.g., rat). Defensive behaviors include stretchattend posture, stretch approach, freezing, burying, and hiding, and are measured as a function of risk assessment. This test has proven useful to determine strain differences in defensive behaviors and relative levels of anxiety in response to predators (see Note 33). Additionally, the defensive behaviors measured are sensitive to anxiolytics, making this paradigm useful in pharmacological screening (23). 1. Introduce the mouse into the small black box, which will serve as a “home chamber” (safe environment). The Plexiglas tunnel should allow free movement between the home chamber and the observation box. 2. On the first 3 days of testing, allow the mouse to explore the observation box for 10 min to become familiar with the environment. In these sessions, there should be no rat present. 3. On the fourth day, insert the mouse into the home chamber and the rat into the observation box for 10 min. The rat should be placed in the opposite side of the cage isolated from the mouse by the wire mesh.
Experimental Models of Anxiety
313
4. In every testing session, record the number of stretch-attend postures, stretch approaches, freezes, and number of times the mouse retreats to the home chamber. Also, measure the amount of time spent in the home chamber and observation box, as well as time in contact with the wire mesh. 3.16. Novel Object Test
This model investigates the approach-avoidance behaviors of mice in response to novel stimuli. Typically, mice tend to explore a novel object longer than a familiar one, and prior exposure to a stimulus increases consecutive approach behavior and decreases avoidance behavior. This robust behavior, as well as the simplicity of this model, makes this test particularly useful for measuring anxiety in a battery of tests (24). Anxiolytics have been shown to increase exploratory behavior of mice in novel environments (36), suggesting that the use of anxiolytics would similarly increase this behavior with novel objects. 1. On day 1, introduce the mouse into the test apparatus, allowing it to explore the environment for 30–60 min. 2. On testing day, insert the novel object into the center of the testing apparatus prior to introducing the mouse. Record the frequency and duration of exploratory behavior, such as approaches, sniffing, physical contacts (e.g., touching, licking, biting, etc.), wall leans, vertical rears, head dipping, and time spent near the novel object, for 10–30 min. Also record amount of avoidance behavior as time spent in the perimeter. 3. Any video-tracking system (see Table 18.2 for details) may be useful for measuring amount of movement and position within the test apparatus. Conversely, the apparatus may be sectioned off, and duration in each section can be recorded, comparing the perimeter sections to the novel object section (24) (see Note 34).
Table 18.2 Examples of video-tracking manufacturers Name
City
Country
Company web site
Clever Sys Inc.
Reston, VA
USA
www.cleversysinc.com
Noldus
Leesburg, VA, or Wageningen
USA, or Netherlands
www.noldus.com
San Diego Instruments
San Diego, CA
USA
www.sandiegoinstruments.com
TSE Systems
Midland, MI
USA
www.tse-systems.com
Qubit Systems
Kingston, ON
Canada
www.quibitsystems.com
Biobserve GMBH
Bohn
Germany
www.biobserve.com
314
Hart et al.
4. Notes 1. As genetic background greatly influences behavioral phenotypes in mice, comparative studies must take this into account. The use of inbred mice substantially decreases within-subject variation and also provides valuable insight into the neurobiological mechanisms that modulate specific phenotypes. The most updated detailed nomenclature for mouse strains should be used, and mice should be obtained from certified vendors or other reliable sources to ensure comparability of results between laboratories. In contrast, outbred mice do not seem to present similar benefits, and therefore, may yield more confounded results. 2. Age, gender, and strain differences affect elevated-plus maze (EPM) performance. Young females generally spend less time on the open arms than males, although this varies with the estrous cycle. Pro-estrus rodents spend significantly more time on open arms than di-estrus females (or male mice) (9). 3. All experimental procedures (including handling, housing, husbandry, and drug treatment) must be conducted in accordance with national and institutional guidelines for the care and use of laboratory animals. 4. Since lighting can affect behavior in the maze, make sure it is consistent on all arms. Red light in a darkened room is preferable, as mice cannot see red light. To avoid excessive freezing, testing environment should be kept quiet without disruptions. If the mouse freezes for more than 30% of the total test time, researchers should note of this abnormality, but continue testing. In case of unexpected or loud noises or other disruptions, the data should be discarded from analyses. 5. As some mice may fall off an open arm of the EPM, the data from these animals must be excluded from further analyses. 6. In some cases, reduced anxiety can be mistaken for hyperactivity. A minute-by-minute analysis of exploration and activity may aid in distinguishing these two different domains (37). 7. It is important not to misinterpret reduced locomotion due to high habituation as an anxiogenic response (see Table 18.3 for details). The mouse learning/memory phenotypes should be assessed in separate tests (37). If mice have poor habituation, this may result in increased “exploration” that should not be misinterpreted as decreased
Experimental Models of Anxiety
315
anxiety (37). Consider testing mouse cognitive functions in a separate study.
Table 18.3 Potential combinations of emotional and cognitive phenotypes (37) Phenotype
Elevated anxiety
Reduced anxiety
Elevated memory
Increased initial anxiety, increased habituation
Low initial anxiety, increased habituation
Reduced memory
Increased initial anxiety, decreased habituation
Low initial anxiety, decreased habituation
8. Mice with altered motor, vestibular, neurological, memory, or depression domains may need additional screening before use in the OFT. Variable aged ranges may be used, but all mice should be tested at the same age in a comparative experiment. 9. Some strain differences are apparent in digging behaviors. Slow or inactive mouse strains (e.g., 129S1/SvImJ) may be replaced with more active strains (e.g., C57BL/6 J) to achieve recordable amounts of burying data. It has been observed that younger mice (2–4 months old) tend to show enhanced digging behaviors as opposed to mice over 1-year-old (14). 10. If mice continue to display low burying activity, it may be useful to assess the environment for confounding factors. Unnecessary noise or stress should be eliminated and mice should be undisturbed throughout the experiment. Testing on cage-cleaning days may also cause mice to be less responsive to the new bedding (14). 11. Some strains with low burying/digging activity may require a longer (e.g., 45–60 min) testing time that may help reveal their phenotype. 12. Troubleshooting is the same as in the marble burying test. 13. In choosing a proper mouse strain for testing, it is important to consider possible motor and sensory deficits. For example, C57BL/6J strain is widely used as it has relatively no deficits in these domains and is sensitive to drug and behavioral testing. 14. Abnormally high grooming activity may be due to a strainspecific compulsive-like phenotype (consider using a more appropriate strain) or due to unintended stress in the animal facility (which may be assuaged by improved husbandry or enrichment).
316
Hart et al.
15. High baseline or transfer anxiety may lead to unusually low-grooming activity. This may be alleviated by using smaller observation boxes and dimmer lighting, as well as by improving handling techniques and lengthening acclimation time. Reduced grooming activity may also be due to a strain-specific low-grooming phenotype (for example, due to abnormal neurological/vestibular/motor phenotype) or overall inactivity of the stain being tested. 16. Detection of different stages of grooming behavior may sometimes be difficult. If using a video camera, replaying in slow motion will make the detection of transitions and interruptions much easier. 17. If the startle stimulus is auditory, some mouse strains may be insensitive to this test because of hearing deficits which can also be age related (e.g., C57BL/6J mice have a progressive hearing with onset after 10 months of age). To rule out this possibility, mice should be tested for hearing problems. If the mouse strain shows abnormally poor hearing, consider using a physical startle stimulus (e.g., air puff or bright light) or a different strain. 18. If the mouse does not show a heightened response to the startle stimulus in the testing trials, it may have cognitive deficits. Memory should be examined in separate, specific tests to ensure accurate data interpretation. 19. C57BL/10J and FVB/NJ strains have high startle responses and 129S1/SvImJ mice have low startle responses, whereas BALB/cJ and C57BL/6J strains have more moderate responses (38). Some animals may show an abnormally high startle response as a result of brain pathological over-excitation, and this abnormality should be investigated further. Also, consider baseline brain activity as well when administering drugs. For example, due to the floor/ceiling effect, anxiogenic drugs can be tested on mice with a low baseline startle response, whereas anxiolytics would yield clearer results if tested on mice with high baseline responses. Review literature for drug efficacy and concentrations. 20. Some mice display particularly high levels of social interaction, including FVB and C57BL/6J (39). Certain strains may be more likely to engage in social interaction because of their high sociability phenotype (which may be unrelated to their anxiety or emotionality profile per se). In this case, consider using other strains for this test. Low levels of social interaction may occur with the spontaneous deletion of the Dtnbp1 gene, leading to social withdrawal (40), or in
Experimental Models of Anxiety
317
some inbred mouse strains, for example, A/J, BALBcByJ, and BTBR T+ tf/J mice (39). 21. Thus, the use of some strains should be avoided as their autism-like behavior may prevent the relevance of this test as a model of anxiety. In performing the social interaction test for screening anxiolytic and anxiogenic drugs, it is suggested that the same two mice are not re-introduced into the same environment together, as this may eliminate the social novelty of the condition, and will affect their test performance. 22. Mice with abnormally poor or abnormally good cognitive abilities may produce aberrant behavior in this test (e.g., increased or decreased social interaction, respectively). To rule out this possibility, consider testing mice in some additional memory paradigms. Memory tests, such as the Morris Water Maze and OFT habituation, may be performed to assess cognitive functions in any abnormally behaving mouse. 23. Low motor or vertical activity may be a strain-specific phenotype. Inactive strains will produce less activity overall and may not be suitable for this model. Likewise, hyperactive strains generally display less non-horizontal exploration and may have difficulties with balance. A narrower apparatus will encourage the animal to show less horizontal activity, enabling it to focus on other behavioral responses. Differences in mouse size should also be addressed. Use animals of similar size, age, and weight to accurately compare between groups. 24. If the mouse displays abnormally high transfer anxiety, gently support it for approximately 5 s to facilitate a solid grip. If the animal continues to display high transfer anxiety, exclude it from the experiment. A dimly lit experimental room may help reduce anxiety. 25. Some strains have difficulties balancing on an aluminum rod, and a more textured surface (e.g., wood) may help stabilize the animal. Increasing the diameter of the rod is another possible solution. If mice continue to struggle with balance or motor abilities, assess motor and vestibular functions separately as these behaviors may be due to a neuromuscular or motor coordination problem unrelated to vestibular deficits or anxiety. 26. Certain strains of mice may be less inclined to explore the test environment, such as mice with anxiety- or depressionlike phenotypes (see Mouse Phenome and Mouse Genome Informatics Databases for details). Allow a longer acclimation and/or test time (e.g., 10 or 15 min) to reduce this
318
Hart et al.
factor. Some mouse strains (e.g., many albino mice) display visual deficits and may not be a suitable model for this test. Consider other mouse strains for the light-dark box testing. 27. While most strains respond consistently to this paradigm due to its independence from motor activity, specific mouse strains (e.g., FVB/N) have considerably higher baseline body temperatures and should be avoided in this model (21). However, this procedure has been shown to effectively induce hyperthermia to varying degrees in all inbred and outbred strains tested (41) and is also an effective indicator of stress in genetically modified animals (34). 28. In the group-housed mice, the last mice to be tested show an increase in body temperature (compared with the first mice) due to anticipatory anxiety. Therefore, animals should be tested individually, with at least 10 mice in each experimental group (21). 29. Temperature measurements should be performed at the same time due to circadian rhythm. Baseline body temperature is significantly higher during the night. If testing occurs during the dark phase, there may be an interference with the amplitude of hyperthermia when the stressor is presented (21). 30. Mice should be kept undisturbed before the experiment with proper handling and opening of cages to ensure accurate results. 31. The above protocol may be modified for the testing of anxiolytic drugs. Sixty minutes prior to the first temperature measurement, inject the mouse with the desired drug. The first temperature measurement serves as an acute stressor and is followed after 10 min with a second temperature measurement (34). 32. Certain drugs (e.g., ethanol) are known to be strain dependent in their effects and may not produce consistent results in the hole-board test. Many commonly used drugs (e.g., fluoxetine) have pronounced dose-dependent effects on head dipping behavior, and therefore, dosing should be carefully considered; see review in (22). 33. Certain mouse strains (e.g., 129S1/SvImJ or BALB/cJ) may not be useful in this test due to their hypoactivity and/or high anxiety phenotypes. If the mouse tested is very inactive and anxious, it may not even leave the home chamber, and this test will not work. In this case, use a milder stressor, such as an anesthetized rat, a toy rat, or rat odor. However, it may also be recommended to use a different mouse strain. Although this test is very useful for comparing defensive behaviors between mouse strains,
Experimental Models of Anxiety
319
some strains are not suitable for this test. For example, mice with sensory deficits (e.g., poor vision or olfaction) or with particular cognitive problems (e.g., poor working memory) will not provide reliable data in this paradigm. As mentioned above, it may help to check online mouse databases for selecting an appropriate mouse strain. 34. Similar to some other previously described tests, mouse strains with sensory or cognitive deficits may not provide reliable data in this model. In addition, some mice can exhibit strong neophobia, which would also confound behavioral data. Test mice prior to this experiment to screen for such defects and consider using alternate strains and/or extending the observation time.
Acknowledgments This work was supported by the NARSAD YI Award to AVK and by Stress Physiology and Research Center (SPaRC) of Georgetown University Medical School. References 1. Warnick, J. E. and Sufka, K. J. (2008) Animal models of anxiety: examining their validity, utility, and ethical characteristics, in Behavioral models in stress research (Kalueff, A. V. and LaPorte, J. L., eds.). Nova Biomedical Books, New York, pp. 55–71. 2. Flint, J. (2003) Animal models of anxiety and their molecular dissection. Semin Cell Dev Biol 14, 37–42. 3. Ohl, F. (2005) Animal models of anxiety. Handb Exp Pharmacol 169, 35–69. 4. Sousa, N., Almeida, O. F. and Wotjak, C. T. (2006) A hitchhiker s guide to behavioral analysis in laboratory rodents. Genes Brain Behav 5 Suppl 2, 5–24. 5. Borsini, F., Podhorna, J. and Marazziti, D. (2002) Do animal models of anxiety predict anxiolytic-like effects of antidepressants? Psychopharmacology (Berl) 163, 121–141. 6. De Boer, S. F. and Koolhaas, J. M. (2003) Defensive burying in rodents: ethology, neurobiology and psychopharmacology. Eur J Pharmacol 463, 145–161. 7. Falls, W. A., Carlson, S., Turner, J. G. and Willott, J. F. (1997) Fear-potentiated startle in two strains of inbred mice. Behav Neurosci 111, 855–861.
8. Kalueff, A. V., Aldridge, J. W., LaPorte, J. L., Murphy, D. L. and Tuohimaa, P. (2007) Analyzing grooming microstructure in neurobehavioral experiments. Nat Protoc 2, 2538– 2544. 9. Walf, A. A. and Frye, C. A. (2007) The use of the elevated plus maze as an assay of anxietyrelated behavior in rodents. Nat Protoc 2, 322–328. 10. Deacon, R. M. (2006) Housing, husbandry and handling of rodents for behavioral experiments. Nat Protoc 1, 936–946. 11. Hagenbuch, N., Feldon, J. and Yee, B. K. (2006) Use of the elevated plus-maze test with opaque or transparent walls in the detection of mouse strain differences and the anxiolytic effects of diazepam. Behav Pharmacol 17, 31–41. 12. Karl, T., Duffy, L. and Herzog, H. (2008) Behavioural profile of a new mouse model for NPY deficiency. Eur J Neurosci 28, 173–180. 13. Archer, T., Fredriksson, A., Lewander, T. and Soderberg, U. (1987) Marble burying and spontaneous motor activity in mice: interactions over days and the effect of diazepam. Scand J Psychol 28, 242–249.
320
Hart et al.
14. Deacon, R. M. (2006) Digging and marble burying in mice: simple methods for in vivo identification of biological impacts. Nat Protoc 1, 122–124. 15. Nicolas, L. B., Kolb, Y. and Prinssen, E. P. (2006) A combined marble buryinglocomotor activity test in mice: a practical screening test with sensitivity to different classes of anxiolytics and antidepressants. Eur J Pharmacol 547, 106–115. 16. Njung e, K. and Handley, S. L. (1991) Evaluation of marble-burying behavior as a model of anxiety. Pharmacol Biochem Behav 38, 63–67. 17. Mikics, E., Baranyi, J. and Haller, J. (2008) Rats exposed to traumatic stress bury unfamiliar objects–a novel measure of hypervigilance in PTSD models? Physiol Behav 94, 341–348. 18. Halberstadt, A. L. and Geyer, M. A. (2008) Habituation and sensitization of acoustic startle: opposite influences of dopamine D(1) and D(2)-family receptors. Neurobiol Learn Mem. 19. Kalueff, A. V., Keisala, T., Minasyan, A., Kumar, S. R., LaPorte, J. L., et al. (2008) The regular and light-dark Suok tests of anxiety and sensorimotor integration: utility for behavioral characterization in laboratory rodents. Nat Protoc 3, 129–136. 20. Kalueff, A. V. and Tuohimaa, P. (2005) The Suok (“ropewalking”) murine test of anxiety. Brain Res Brain Res Protoc 14, 87–99. 21. Bouwknecht, J. A., Olivier, B. and Paylor, R. E. (2007) The stress-induced hyperthermia paradigm as a physiological animal model for anxiety: a review of pharmacological and genetic studies in the mouse. Neurosci Biobehav Rev 31, 41–59. 22. Kliethermes, C. L. and Crabbe, J. C. (2006) Pharmacological and genetic influences on hole-board behaviors in mice. Pharmacol Biochem Behav 85, 57–65. 23. Yang, M., Augustsson, H., Markham, C. M., Hubbard, D. T., Webster, D., et al. (2004) The rat exposure test: a model of mouse defensive behaviors. Physiol Behav 81, 465–473. 24. Powell, S. B., Geyer, M. A., Gallagher, D. and Paulus, M. P. (2004) The balance between approach and avoidance behaviors in a novel object exploration paradigm in mice. Behav Brain Res 152, 341–349. 25. Martinez, J. A., Francis, G., Liu, W., Pradzinsky, N., Fine, J., et al. (2008) Intranasal delivery of insulin and a nitric oxide synthase inhibitor in an experimental model of amyotrophic lateral sclerosis. Neuroscience.
26. Ito, N., Nagai, T., Oikawa, T., Yamada, H. and Hanawa, T. (2008) Antidepressant-like effect of l-perillaldehyde in stress-induced depression-like model mice through regulation of the olfactory nervous system. Evid Based Complement Alternat Med. 27. De Souza Silva, M., Topic, B., Huston, J. and Mattern, C. (2008) Intranasal dopamine application increases dopaminergic activity in the neostriatum and nucleus accumbens and enhances motor activity in the open field. Synapse 62, 176–184. 28. Buddenberg, T. E., Topic, B., Mahlberg, E. D., de Souza Silva, M. A., Huston, J. P., et al. (2008) Behavioral actions of intranasal application of dopamine: effects on forced swimming, elevated plus-maze and open field parameters. Neuropsychobiology 57, 70–79. 29. Broekkamp, C. L., Rijk, H. W., Joly-Gelouin, D. and Lloyd, K. L. (1986) Major tranquillizers can be distinguished from minor tranquillizers on the basis of effects on marble burying and swim-induced grooming in mice. Eur J Pharmacol 126, 223–229. 30. Bruins Slot, L. A., Bardin, L., Auclair, A. L., Depoortere, R. and Newman-Tancredi, A. (2008) Effects of antipsychotics and reference monoaminergic ligands on marble burying behavior in mice. Behav Pharmacol 19, 145–152. 31. Bespalov, A. Y., van Gaalen, M. M., Sukhotina, I. A., Wicke, K., Mezler, M., et al. (2008) Behavioral characterization of the mGlu group II/III receptor antagonist, LY341495, in animal models of anxiety and depression. Eur J Pharmacol. 32. Gordon, C. J. (2004) Effect of cage bedding on temperature regulation and metabolism of group-housed female mice. Comp Med 54, 63–68. 33. Li, X., Morrow, D. and Witkin, J. M. (2006) Decreases in nestle shredding of mice by serotonin uptake inhibitors: comparison with marble burying. Life Sci 78, 1933–1939. 34. Olivier, B., Zethof, T., Pattij, T., van Boogaert, M., van Oorschot, R., et al. (2003) Stress-induced hyperthermia and anxiety: pharmacological validation. Eur J Pharmacol 463, 117–132. 35. Kort, W. J., Hekking-Weijma, J. M., TenKate, M. T., Sorm, V., and VanStrik, R. (1998) A microchip implant system as a method to determine body temperature of terminally ill rats and mice. Lab Anim 32, 9. 36. Klebaur, J. E. and Bardo, M. T. (1999) The effects of anxiolytic drugs on noveltyinduced place preference. Behav Brain Res 101, 51–57.
Experimental Models of Anxiety 37. Kalueff, A. V. and Murphy, D. L. (2007) The importance of cognitive phenotypes in experimental modeling of animal anxiety and depression. Neural Plast 2007, 1–7. 38. Paylor, R. and Crawley, J. N. (1997) Inbred strain differences in prepulse inhibition of the mouse startle response. Psychopharmacology (Berl) 132, 169–180. 39. Bolivar, V. J., Walters, S. R. and Phoenix, J. L. (2007) Assessing autism-like behavior in mice: variations in social interactions
321
among inbred strains. Behav Brain Res 176, 21–26. 40. zeng, Y. Q., Zhou, Z. Y., He, X., Wang, H., Guo, X. L., et al. (2008) Dysbindin deficiency in sandy mice causes reduction of snapin and displays behaviors related to schizophrenia. Schizophr Res. 41. Bouwknecht, J. A. and Paylor, R. (2002) Behavioral and physiological mouse assays for anxiety: a survey in nine mouse strains. Behav Brain Res 136, 489–501.
Chapter 19 Mouse Models of Neurodegenerative Diseases: Criteria and General Methodology Christopher Janus and Hans Welzl A man is but what he knoweth , Sir Francis Bacon (1).
Abstract The major symptom of Alzheimer’s disease is rapidly progressing dementia, coinciding with the formation of amyloid and tau deposits in the central nervous system, and neuronal death. At present familial cases of dementias provide the most promising foundation for modelling neurodegeneration. We describe the mnemonic and other major behavioral symptoms of tauopathies, briefly outline the genetics underlying familiar cases and discuss the arising implications for modelling the disease in mostly transgenic mouse lines. We then depict to what degree the most recent mouse models replicate pathological and cognitive characteristics observed in patients. There is no universally valid behavioral test battery to evaluate mouse models. The selection of individual tests depends on the behavioral and/or memory system in focus, the type of a model and how well it replicates the pathology of a disease and the amount of control over the genetic background of the mouse model. However it is possible to provide guidelines and criteria for modelling the neurodegeneration, setting up the experiments and choosing relevant tests. One should not adopt a “one (trans)gene, one disease” interpretation, but should try to understand how the mouse genome copes with the protein expression of the transgene in question. Further, it is not possible to recommend some mouse models over others since each model is valuable within its own constraints, and the way experiments are performed often reflects the idiosyncratic reality of specific laboratories. Our purpose is to improve bridging molecular and behavioural approaches in translational research. Key words: Tauopathy, Alzheimer’s disease, transgenic models, phenotype, behavioural tests. Abbreviations: AD A FAD FTD
Alzheimer’s disease -amyloid peptide amilial Alzheimer’s disease fronto-temporal dementia
G. Proetzel, M.V. Wiles (eds.), Mouse Models for Drug Discovery, Methods in Molecular Biology 602, DOI 10.1007/978-1-60761-058-8 19, © Humana Press, a part of Springer Science+Business Media, LLC 2010
323
324
Janus and Welzl FTDP-17 GLM MAPT MCI NFT
fronto-temporal dementia with parkinsonism linked to chromosome 17 general linear model microtubule-associated protein Tau mild cognitive impairment neurofibrillary tangles
1. Introduction One hundred years ago Alois Alzheimer described in a seminal paper the behavioural symptoms of his patient, Auguste D., who suffered from a mental illness (2), (see (3) for English translation of the original paper). He observed that “[S]he developed a rapid loss of memory. She was disoriented in her home, [. . .] She is completely disoriented in time and space. Her memory is seriously impaired. If objects are shown to her, she names them correctly, but almost immediately afterwards she has forgotten everything.” After the patient died, an autopsy revealed dense deposits outside and around nerve cells and twisted strands of fibre inside dead neurons in her brain. Today these two pathological hallmarks of Alzheimer’s disease (AD) are known to be extra-cellular plaques made largely of -amyloid peptide (A) and intracellular neurofibrillary tangles composed of hyper-phosphorylated microtubule-associated protein tau (4, 5). The current view is that almost all neurodegenerative disorders can be broadly classified as disorders of protein folding (6). The accumulation of misfolded proteins may be, as in the case of AD, intra- and extra-cellular, or only intracellular, with abnormally phosphorylated protein tau aggregates being most common. The most common is tau protein which aggregates within neurons in hyper-phosphorylated form (7, 8) and causes profound loss of neurons and atrophy of the brain (9–11). Neurodegenerative diseases characterised by the presence of hyper-phosphorylated tau are collectively termed tauopathies (7, 12). The degeneration of neurons in tauopathies leads to dementia, i.e., a progressive and accelerating decline in mental function. AD is one of the most devastating tauopathies in which a patient’s memory and ability to learn is initially compromised and eventually completely destroyed. Although the behavioural pathologies of tauopathies resemble each other, there is also much variation in the clinical picture due to specific combination of neuropathological changes, variations in the form of hyper-phosphorylated tau and the
Mouse Models of Neurodegenerative Diseases
325
individual spatio-temporal expression of neurodegeneration (reviewed by 7).
2. Tauopathy – Hallmarks and Characteristics 2.1. Compromised Behaviour
Tauopathies are diseases characterised by a progressive and severe decline in cognitive abilities that cannot be attributed to normal aging (13). Abnormalities in other behavioural systems often precede and accompany tauopathies (14). In AD, signs of mild cognitive impairment (MCI) precede overt dementia (15). In practice, individuals will only be classified as cognitively impaired when their ability to perform everyday tasks is compromised to the point that they are no longer able to function at home and in their community (16). MCI diagnostic does not always correctly predict the development of AD-type dementia. Data derived from neuroimaging, screening for genetic risk factors (17) or detection of increased tau protein levels in cerebrospinal fluid (18, 19) might provide false-positive indications of beginning dementia. Even significant hippocampal atrophy is not always a reliable marker and might be due to depression, Parkinson’s disease, or vascular dementia (17). Diagnosis of AD is also complicated since amyloid plaques and neurofibrillary tangles (NFTs) of hyperphosphorylated tau are present during normal aging. Further, depending on the type of tauopathy, patients may present a variety of other cognitive complaints including delusions, amnesia, executive dysfunction, apathy, agitation, and aggressive behaviour (8, 19–22).
2.2. Neuropathology
Tauopathies are characterised by neuronal dysfunction and loss that display a varying but overlapping spatio-temporal distribution (reviewed in 8). The neuropathological variability is the reason for the complex and also varying clinical phenotypes. In AD, progressive neuronal damage and death appear in brain regions critical for learning and memory (neocortex, hippocampus, amygdala, anterior thalamus, basal forebrain, and subcortical nuclei including the nucleus basalis of Meynert (23–28)). The highest atrophy is seen in the entorhinal cortex and hippocampus (29) which positively correlates with the degree of dementia (30). The development of neuropathology is paralleled by a decreased functionality of forebrain and brainstem monoaminergic and cholinergic systems (4, 31–33).
2.3. Intra- and Extra-cellular Protein Inclusions
The main characteristic of tauopathies is an age-progressing hyper-phosphorylation of the tau protein which accumulates in tangles with paired helical filaments, twisted ribbons, and/or
326
Janus and Welzl
straight filaments (7, 34). Excessive intra-neuronal deposition of tau protein is also a key feature of dying neurons during normal aging (35, 36). Kidd, studying brains of AD patients, was the first to show that both tangles and neuritic plaques contain dense accumulation of pathological paired helical filaments (37). In the mid-1980s, tau was identified to be the principal component of neurofibrillary lesions in the AD (34, 38, 39). Similar lesions were also found in other neurodegenerative diseases including corticobasal degeneration (40), amyotrophic lateral sclerosis/parkinsonism dementia complex of Guam (41, 42), Down syndrome (43), progressive supranuclear palsy, Pick’s disease, argyrophilic degeneration (reviewed by 7), myotonic dystrophy (44), and the family of fronto-temporal dementias (FTD) (45). AD is a special form of tauopathy additionally characterised by extra-cellular deposits of A and amyloid deposits in cerebral blood vessels (46). Plaques consist mainly of a 40–42 residue  amyloid peptide (A40 /A42 ), cleaved from the amyloid precursor protein (APP) and are surrounded by dystrophic nerve cells. The longer A42 species is normally present in small, soluble fractions in biological fluids (47), but it is elevated and early deposited in cases of familial Alzheimer’s disease (FAD) (48, 49). The increase in levels of both A40 and A42 correlates with progress in cognitive decline (50), and the increase in A40 peptide was detected in 40% of AD patients before overt amyloid plaques could be detected. Moreover, the increase in A preceded the formation of NFT, suggesting that at least in the frontal cortex of some cases, soluble species of A may precipitate the formation of NFTs. However, since tau pathology can occur in the absence of A other causes of NFT formation cannot be excluded (50). 2.4. Genetics
The first gene mutations related to familial neurodegeneration were identified in AD in the gene encoding amyloid b precursor protein (APP) (51–53), followed by mutations in presenilin 1 (PSEN1) (54, 55) and presenilin 2 (PSEN2) (56, 57). No mutations directly associated with NFTs have been identified in AD. However, several groups of patients with FTD inherited autosomal dominant mutations in MAPT in FTDP-17 were reported (58–60). Approximately 100 families with FTDP-17 have now been identified, with a total of 32 unique tau gene mutations (61, 62). Moreover, in addition to MAPT, mutations in progranulin, which encodes a growth factor involved in the regulation of multiple processes including development, wound repair, and inflammation, were recently shown to be implicated in FTDP-17 (63). Finally, mutations in PSEN1 gene, implicated in AD, can produce FTD-like phenotypes with the AD neuropathology (reviewed by 64).
Mouse Models of Neurodegenerative Diseases
2.5. Conclusions
3. Mouse Models of Neurodegeneration
3.1. Criteria of Mouse Models of Tauopathies
327
The identification of gene mutations implicated in tauopathies opened a way to model the diseases in mice. At translational research level, the work is focusing on the production of experimental animal models that reproduce the essential pathology and phenotype of human tauopathies, including the formation of abundant, specific to a disease intra- and/or extra-cellular protein inclusions, neuronal degeneration, and cognitive impairment. Therefore such animal models should have the following characteristics: (1) specific to a given disorder, intra- and/or extracellular deposition of misfolded proteins, (2) age-progressing dementia, (3) disturbance in behavioural systems not related directly to cognitive function but which are observed at specific stages of a disease, and (4) coinciding neuronal loss in diseasespecific brain regions and cytoarchitecture.
The first successful mouse model replicating major hallmarks of neurodegenerative disease was an AD mouse model created more than a decade ago by Games and colleagues (65). Other AD mouse models followed soon and were proven extremely informative, which was chronicled in number of scholarly reviews (66–77). Mouse models expressing genes implicated in other tauopathies followed closely (reviewed by 7, 70, 78, 79). At the juncture of over a decade long history of using mouse models of neurodegeneration, it seems that a robust and good model should meet the following guidelines: 1. Replication of clinical phenotypes. Since the cognitive decline and region-specific neuronal loss are central to neurodegenerative diseases, the model of a disease should recapitulate accurately these facets of the clinical phenotype. 2. Age-progressing phenotype. A credible model should exhibit age-progressive neuropathology and cognitive deficits which could be evident, to various extent, in paradigms addressing different memory systems. The extent of age-progressing behavioural impairment may eventually encompass non-cognitive systems due to a significantly increasing brain pathology. Although, this may raise operational complexities associated with the interpretation of the cognitive impairment being confounded by the emergence of impairment in non-cognitive behavioural systems (80), the use of such model during drugs screens may reveal which behavioural deficiencies can be ameliorated by a treatment at given stage of pathology.
328
Janus and Welzl
3. Control for the effect of genetic mutation. In the models employing genetic mutations (transgenic models), phenotyping changes should be robust and correlated with the presence of familial mutations, but should be absent or less overt in the age-matched mice expressing wildtype (wt, non-mutated) gene expressed at equal (or greater) steady-state levels. These mouse models require minimum variation in their genetic background in order to be sensitive to pick up subtle changes in a phenotype. Therefore, whenever possible inbred strains, having homozygous genomes, are used almost exclusively in biomedical research. Transgenic lines, however, are often created in one strain and later backcrossed to more suitable genetic background, which complicates the control over genetic variability. Wild-derived mouse strains or recombinant strains are usually avoided in transgenic research due to unwanted genetic diversity. Similarly, outbred stocks, used predominantly in genetics, toxicology, and pharmacology, are not recommended for transgenic research due to their genetic variability caused by genetic drift, directional selection, and genetic contamination during breeding (81–83). 4. Validation of a model. The independent confirmation and replication of the key facets of the phenotype in independent transgenic lines harbouring the same construct should be carried out in independent laboratories (84), including standardisation of expertise of technical personnel and differences in handling methods (85). 3.2. Caveats and Pitfalls of Using Mouse Strains
How close the mouse model replicates the pathology observed in a disease may often depend on the design of the model and behavioural systems in focus. The choice of a mouse strain is crucial, since many strains suitable for genetic manipulation are not particularly suitable for behavioural studies (86, 87). Genetic background of a mouse strain can be used as a tool in the analysis of a mutation (88), and the use of mapping and cloning strategies allows the identification of modifier genes existing in different mouse strains (89, 90). Noteworthy is that gene targeted or transgenic mice are usually initially created on 129 or FVB strain backgrounds. The reason for the latter is that the oocytes of these strains are large, thus increasing the probability of successful injection of the transgene construct. These strains, however, are not particularly suitable for behavioural testing (91) and it is often necessary to transfer the mutation to a more suitable background, usually C57BL/6 strain, by backcrossing for at least 10 generations. While the strategy largely results in the replacement of the donor background with the recipient background,
Mouse Models of Neurodegenerative Diseases
329
the region flanking the selected gene remains likely of donor origin. Thus, the genes in this flanking region travel with the selected transgene, and when comparing transgenic (or knockout) mice with non-transgenic littermates, allelic differences in the flanking genes can conceivably influence the traits of interest. The problem of genetic background was partially remedied in the case of knockout models in which genes are targeted directly in C57BL/6-derived embryonic stem (ES) cells (see Knockout Mouse Project (KOMP) http://www.komp.org/for further information). Many neurodegenerative mouse models, however, are still maintained on mixed and segregating genetic backgrounds, thus even inbred littermates are not genetically identical. Also, one has to bear in mind that many strains, like C3H/HeJ, SJL/J, FVB/NJ, MOLF/E, PL/J, SWR/J. BUB/BnJ, CBA/J, or NON/LtJ, are not suitable for behavioural testing when visual cues are of importance due to the presence of retinal degeneration (rd) http://www.jax.org Retinal degeneration (rd) is caused by an autosomal recessive mutation resulting in a rapid ageprogressing degeneration of rods and cones (92, 93). About 20% of all inbred mouse strains carry the rd causing PDE6B gene (94), and above listed strains are particularly prone to degeneration since they are homozygous for the PDE6B gene. Other strains, like A/J, BALB/cByJ, AKR/J, KK/H1J, to mention a few, are albino and can be expected to have mild defects in their vision (95, 96). Other deficits include age-progressing hearing loss in A/J, BALB/cByJ, C57BLKS/J, C57L/J, and C57BR/cdJ (but not C57BL/10 J) or other strains (129S1/SvlmJ, BALB/cByJ, or I/LnJ) may have partially developed corpus callosum (source, http://www.jax.org). In conclusion, taking into account the genetic strain background effect on behaviour, the breeding scheme of a transgenic line or the generation of multiple transgenic lines, the presence of retinal degeneration or other possible mutations expressed in homozygous state, the design of the transgenic mouse model has to be carefully planned to avoid serious confounding variables in most learning tasks which depend on visual acuity of animals. The observed pathology in a transgenic mouse model may also depend on the choice of the promoter used to drive a transgene expression. The most common promoters include APP promoter (AD mouse models) (97), brainenriched prion protein promoter (98–100), the platelet-derived growth factor b-chain (PDGFb) promoter (65) (both PrP and PDGF promoters resulting in a transgene expression also outside of the CNS), and neuronal specific Thy-1 promoter (101). Another problem with transgenic models relates to spontaneous genetic changes which may affect the phenotype of a model. Mice engineered to overexpress a transgene can potentially with time change the number of disease-causing transgene copies, leading to possible loss of a phenotype. Without routine checks of a trans-
330
Janus and Welzl
gene expression within and between the laboratories, the differences in the transgene copy number can prevent replication of the results between the laboratories. Awareness of this issue should prompt researchers to check periodically the genetic constitution of their transgenic stocks. 3.3. Modelling Human Dementia
Given the disparities between species, it can be challenging to draw definitive conclusions about the association of cognitive function between humans and transgenic mouse models. In order to make appropriate comparisons, tests of memory in rodent models of neurodegeneration should target cognitive systems that are found and conserved across species, including humans, and have a clearly delineated function and a defined neuroanatomy. Assessment of spatial navigation and its dependence on the hippocampus fulfils the above assumption, since this memory system is highly conserved in mammals (102). The neuroanatomical structure of hippocampus, together with changes in its synaptic plasticity during memory formation (103–110), serves as a welldefined model of memory that has been frequently employed in studies using rodent species (106, 109–112). Humans with temporal lobe damage also have severe impairments in learning and memory, including the recall of spatial locations and solving spatial maze tasks (113–115), confirming the involvement of the hippocampus in spatial memory in humans. Similar findings are seen in AD patients, who have significantly increased atrophy of the hippocampus (116, 117) and impaired performance in spatial navigation tests (118–123).
4. Methods and Experimental Design 4.1. Evaluation of Phenotype
The evaluation of the phenotype of a new mouse model of neurodegeneration should be based on a battery of tests characterising the physical and motor development of mice, their response to the array of basic stimuli, as well as the characterisation of the targeted by the model behavioural system(s). The detailed description of each test is beyond the scope of this chapter. The reader should consult existing sources, from textbooks (124, 125), general articles related to behavioural phenotypying (87, 126), specific articles related to analysis of strategies (127), memory (128, 129) in a water maze test, or methodological procedures related to spatial orientation tests (130). More specialised articles describe experimental approaches which can enhance learning in strains of mice known as poor performers in a specific test (131), or articles comparing performance of different species (e.g., rats and mice) in a test
Mouse Models of Neurodegenerative Diseases
331
(132) and also Chapters 17 and 19 of this volume. Above all, the existence of the recently launched Nature Protocols journal http://www.nature.com/nprot/index.html which publishes detailed experimental protocols, defies any attempt here to provide complete guidance to the plethora of behavioural paradigms available. Nature Protocols articles not only present theoretical background underlying each testing paradigm, detailed procedures, examples of collected data, but also suggest type of equipment, analyses, including trouble-shooting section. Instead, we have provided a list of most frequently used tests accompanied with appropriate information which should enable behavioural and molecular researchers to use this as a compendium facilitating the interpretation of the results found in scientific literature or as a starting point in establishing relevant testing paradigms of their models. 4.2. Evaluation of Mnemonic Function
The prevalence of spatial memory tasks in the characterisation of mouse models of neurodegeneration (Table 19.1) is justified by high evolutionary conservation of spatial memory across mammalian species. The main question asked during such characterisation relates to the abilities of a mouse genome to cope with the presence of an expressed transgene protein. The existence of possible effects due to strain genetic background, modifier genes, compensatory effects, and/or subtle differences in the experimental paradigms, including the strains’ different response to handling (133), can yield different, often contradictory results. Therefore, a broad characterization of mouse behaviour, including both hippocampus-independent memory systems (134–136) and other non-cognitive behavioural systems, such as changes in agitation and aggression levels (137), locomotor, exploratory, or stereotypic activity (138) can be very useful in evaluating transgenic mice. The results of such studies not only extend our understanding of the effect of these transgene on behaviour but also allow us to identify potential confounds in memory tests (139). Moreover, studying hippocampus-dependent memory in different testing paradigms may sometimes provide interesting additional information regarding a particular mouse model. For example, the APP Tg2576 mice were tested in T-maze alternation and contextual fear conditioning tasks (134). Investigators reported a significant impairment in T-maze alternation but, surprisingly, they found that the animals were unimpaired in both contextual (hippocampus-dependent) and auditory fear conditioning tests (hippocampus-independent task). The mice showed attenuated contextual discrimination only with a decrease in the salience of the context and without changes in tone conditioning discrimination. Such detailed validation of the existing mouse models
332
Janus and Welzl
Table 19.1 Evaluation of cognitive phenotypes of mouse models of neurodegeneration. These testing paradigms represent some of the most commonly used tests employed in behavioral evaluation of mouse models of neurodegenerative disease (Reprinted from (70) with permission) SHIRPA protocol (147) In most cognitive tests, learning rate and memory strength is inferred from measures of locomotor behavior, therefore any possible effect of a transgene on motor and/or perceptual systems can yield false-positive (impaired learning) results. To this end, a general phenotypic assessment of transgenic mice along with non-trangenic littermates must precede specific cognitive tests in order to eliminate these possible confounds. SHIRPA protocol provides a comprehensive evaluation of mice behavior ranging from the assessment of exploration and activity levels to thermal nociception. This battery of simple tests begins with procedures most sensitive to physical manipulation, like anxiety tests performed in the open-field or elevated plus or zero mazes. Other screens focus on gross phenotyping abnormalities, assessment of sensorimotor deficits (rota-rod), holeboard exploratory activity, and thermal analgesia. Although application of the full battery can be time consuming and requires a well-equipped lab, a subset of simple tests can be carried out and is highly recommended for initial characterization (126). MORRIS WATERMAZE (MWM) (148, 149) The MWM test has been the most widely used testing paradigm to study hippocampus-dependent spatial memory in rodent species. Reference memory or place discrimination version of MWM requires mice, trained with repeated trials over several days, to use external visual cues around the testing room to search for the hidden (barely submerged) escape platform in the water maze. Spatial navigation encompasses the development of different search strategies with spatial strategy (reflected by a direct swim to a platform) taking place at the end of this complex learning process (127, 150). The main dependent variable reflecting learning acquisition is escape latency – the time it takes a mouse to find a platform, or search path, which is less biased by the differences in swim speed. Memory bias is evaluated in trials where a mouse searches a pool where the hidden platform has been removed. Spatial learning is reflected by decreased escape latency or search path, while spatial memory by increased search in areas or quadrants of the pool containing a platform during training. An annulus-crossing index (a number of swimming over former platform location adjusted for swims in other 3 quadrants of a pool (127, 151, 152)) represents an alternative, more stringent measure of memory bias. In cases when more than one probe trial is carried out during training, a mean probe score (the mean percent of time spent in target quadrant during all probes (153)) can be used as a reliable memory evaluation index. Correspondingly, learning impairment is reflected by longer escape latency or search path during training, and memory impairment by displaced or random search, which is reflected by about 25% of time, spent in each of quadrant of the pool during a probe trial. To address episodic-like memory in mice, a more complex version of the MWM test was developed (154). In this test, numerous locations of the platform were used and the number of new locations learned during the whole training reflects learning capacity of an individual mouse. In a cued or visible platform version of the MWM test a platform location is marked by a visible cue that mice associate with an escape from water. This version of the test is often implemented as a control for normal visual acuity, an unimpaired learning of simple association between a proximal cue and an escape platform, or as demonstration of a comparable swim speed between studied genotypes. These controls should be used with caution, however, because in contrast to rats, some strains of mice with hippocampal lesions often show also partial impairment in the cue navigation task (155).
Mouse Models of Neurodegenerative Diseases
333
Table 19.1 (Continued) OBJECT RECOGNITION (OR) (156, 157) This test exploits a natural tendency of rodents to explore novel objects and to show an exploratory preference for replaced or displaced objects. The dependence of object recognition memory on the hippocampus is related to the protocol of a test. Short delays between initial exploration phase and a memory test make OR test independent from the hippocampus (136), however, when longer delays (hours) are implemented, OR memory depends on hippocampus function (158, 159). Object memory impairment is demonstrated when an animal shows no preference in exploration (close proximity, nose contact) of a new or displaced object. FEAR CONDITIONING (FC) (160, 161) The FC paradigm which is an example of classical Pavlovian associative learning and involves an association of a neutral tone (conditioned stimulus, CS) paired with a brief electric foot-shock (unconditional stimulus, US) delivered in a novel context. Mice trained in that manner develop a fear response (conditioned response, CR), expressed as defensive or anti-predatory behavior in a form of freezing (complete cessation of movement) which coincides with autonomic and endocrine response (increased heart beat rate and blood pressure), and sensory alteration (analgesia, potentiated startle). The paradigm may involve two types of conditioning that can be performed simultaneously or independently during a training phase: contextual (CFC), when an animal develops an association between shock and training context (conditioning chamber), and tone fear conditioning when shock (US) is associated with a neutral tone (CS). The tone conditioning is performed either as delay conditioning paradigm when there is a temporal overlap between CS and US (a foot-shock is delivered within the last 1–2 s of tone duration), or more demanding trace fear conditioning which requires the association of a CS with an US across an interval of time known as trace interval (a foot-shock [US] is delivered after the tone [CS] is turned off). The time between CS and US can vary and an additional temporal processing is required because CS and US are separated therefore an animal has to retain a trace of CS across this time interval in order to associate it with the US. While delay tone conditioning is hippocampus independent but requires intact amygdala (162, 163), the trace and contextual fear conditioning are sensitive to hippocampal lesions (161, 164). The sensitivity of the mice to foot-shock can be established empirically recording the current thresholds that elicit specific response like flinch, jump (165). The lowest current eliciting learning (for mice a current of 0.35–0.4 mA is appropriate) should be used. Impairment in FC is evaluated during test phase on the following day after training, and is reflected by reduced freezing time when an animal is placed in familiar chamber (context conditioning) or when the animal is exposed to a conditioned tone in a new environment. CONDITIONED TASTE AVERSION (CTA) (166–168) CTA is a special form of classical Pavlovian conditioning, representing an adaptive specialization which defends an organism against repeated ingestion of toxic foods (166–169). CTA is well conserved in many different species including humans (169, 170). When acquiring a CTA response, an animal learns to associate the specific taste of a novel food, usually a saccharine solution (conditioned stimulus, CS) with experimentally induced through i.p. injection of lithium chloride after saccharine intake (unconditioned stimulus, US), nausea. Because of one trial pairing between CS and US, a long-lasting avoidance of food with this specific taste develops. The brain areas implicated in the CTA include the agranular insular cortex, the parvicellular thalamic ventral posteromedial nucleus, and the parabriachial nucleus of the pons, which are part of the gustatory pathway (171, 172) and the amydgala (173, 174). Impairment in CTA is reflected by increased saccharine intake as compared to control mice in choice tests (usually two bottles test, one containing water, one saccharine).
is necessary in order to provide a more powerful experimental framework for behavioural characterisation of future mod-
334
Janus and Welzl
els and to increase the effectiveness of screening of potential therapeutics. A pragmatic approach would dictate that robust phenotypes obtained in less labile tests (in which data collection is based on motor or strong sensory inputs) would be replicable within tolerable margins, while more labile phenotypes based on emotional or social behaviours may be strongly affected by differences in laboratory practice (140), especially in poorly managed animal colonies. 4.3. Experimental Design
The purpose of doing experiments is to distinguish between alternative hypotheses or explanations. However, even a perfectly designed experiment might lack sufficient power, if sample sizes of animals in experimental groups are small. Since it is often not known what effect on behaviour a mutation exerts, it is advisable to properly characterize a non-transgenic or wild-type control mice, thus establishing a yardstick for robust characterization of the phenotype in question, avoiding floor or ceiling effects in the data recording which can cause a skewed data distribution. Larger sample sizes are desirable (n = 8–12, or more (141) depending how robust a focal behaviour is), but attention has to be paid to ensure that the mice are not tested too long during the day, which may result in their fatigue or may span over different phases of the circadian cycle. Also, one has to be aware that a change in behaviour of mice may sometimes be caused not by the experimental treatment, but merely by the handling or attention paid to them by the experimenter. The effect, known in psychology as the Hawthorne effect (142), is often a cause of differences between the obtained results in various laboratories (85). A common error, which occurs less often in behavioural research but crops up frequently in physiological experiments, refers to treating repeated data points coming from the same subjects as independent from each other measures. This approach, called pooling fallacy (143), leads to an inappropriate increase in the sample size of mixed, dependent, and independent data points, thus violating many assumptions of experimental design and parametric data analysis. Problems with independence of data may arise in less obvious situations, when the obtained data correlate closely between mice coming from the same litter or between mice housed in the same cage. These litter- or cage-effects can be the result of, for example, differential maternal care, highly variable housing conditions (mice in cages placed at the bottom of a rack in a densely populated with racks rooms are kept in constant semi-darkness), singly housed animals are known to perform worse in learning and memory tasks, etc., and can introduce confounding factors often impossible to overcome due to the small number available and often difficult to derive mutant mice. Awareness of these issues, however, may help during the inspection and first steps in the interpretation of the raw data. If the data generated by mice from
Mouse Models of Neurodegenerative Diseases
335
the same litter or cage are on one spectrum of the distribution, one should consider the replication of the experiment using more careful and balanced assignment of mice to experimental groups. 4.4. Data Analysis
Most experiments, especially those evaluating learning and memory involve repeated tests or training sessions. Results are usually presented in blocks of training trials or days. Plotting data in that fashion usually reflects adequately the learning process, but blocking data over repeated trials reduces variance and may result in some unusual patterns of behaviour escaping attention. It is advisable, therefore to inspect the raw data, especially the data generated from first training trials to check if mutant mice are free from subtle motor or sensory deficits. It can be assumed that control and mutant mice, which are well habituated to handling and lab conditions but na¨ıve to a particular behavioural test, should show comparable performance during first training trial(s). Any cognitive impairment, if not confounded by compromised locomotor or sensory deficiency, should become apparent as training progresses, but should not be present at the beginning of training, unless of course the severity of cognitive decline impairs the interaction of an animal with the surrounding environment. Data are generally analyzed by analysis of variance (ANOVA), with genotype and/or treatment as between subject factors and training days and/or trials as within-subject factor(s). One issue to remember is that the data must meet the criteria of parametric statistics and in the case of repeated measure or within-subject factor, an assumption of compound symmetry must be met in order to avoid bias in the interpretation of the results of a test involving within-subject factor. The assumption of compound symmetry refers to a pattern of constant variances on the diagonal and constant covariances off the diagonal in the variance–covariance matrix. In practice this means that the correlations within the matrix of the repeated factor (days or trials) have to be the same at all distances between measurements. This assumption, however, is hardly met in the analysis of learning data, since as animals learn over time and improve their performance in a task, thus the variance decreases as learning progresses. A departure from the assumption of compound symmetry is usually evaluated by slightly the more stringent sphericity test (Mauchly sphericity test, SPSS GLM (Statistical Package for Social Sciences, SPSS Inc. Chicago)), and in cases of severe departures, degrees of freedom should be adjusted either by Greenhouse–Geisser ε-correction (tends to underestimate, especially when ε is close to 1) or by Huynh–Feldt estimator (which tends to overestimate ε) to avoid false-positive results (144).
4.5. Animal Facility and Behavioural Tests
Replication of the results are at the core of the falsification process of hypotheses or theories (145). However, even the best
336
Janus and Welzl
guidance or step-by-step description of procedures and methods pertaining to a specific testing paradigm may often yield unpredictable and different results between laboratories, despite very careful execution by experienced in the field researchers (84). Therefore, in this section we would like to alert readers to some potential problems and issues which may affect the execution of behavioural experiments and resulting data. The list of presented issues, by no means exhaustive, includes problems which are often not formally documented, or issues which are considered trivial or too obvious to mention in many the Material and Methods sections but are highly relevant. The characterisation of behavioural phenotypes starts with housing conditions in the animal facility. Animal rooms located in the vicinity of noisy, heavy traffic areas, like cage washing areas, are less desirable and in some cases may even lower the breeding rate, induce cannibalism, and increase hyper-reactivity in animals. The housing conditions, including the number of mice in a cage (not more than four mice over extended period), having minimum and consistent level of enrichment in a form of pressed cotton nesting material (nestlets), mouse huts or igloos, or pup tents (source http://www.bio-serv.com/) can significantly improve breeding and reduce anxiety of animals. Housing mice singly in cages is not recommended due to heightened rate of developing stereotyping behaviour, obesity, and decreased performance in learning ability. When breeding transgenic lines, it is not uncommon that the newly born transgenic pups tend to be smaller then their non-transgenic littermates (for example in the case of APP TgCRND8 mice, personal observation). Supplementing lactating females and their pups, especially at the pups age when they start to consume solid food (14–15 days of age), with easily available and more palatable moister powdered mouse Purina chow in a Petri dish, facilitates pups’ growth and can reduce the weight difference between transgenic TgCRND8 and their non-Tg littermates (unpublished data). The distance of the housing room to the behavioural testing room(s) is relevant and the transportation of mice between the two locations can be stressful, therefore adding appropriate time for acclimation to new testing conditions should be done. Last but not least, husbandry practices including care and feeding of animals, cleaning of equipment, physical surroundings, and routine checks of the stock health by experienced, well-trained, and well-managed facility staff guarantee good health, growth, reproduction, survival of mice. Personnel with poor management and/or inexperience in mouse handling and husbandry may adversely affect animal stress level and behaviour. Excessive noise produced during cage changing (for example changing cages under a hood and putting or stacking metal cage lids on the metal surface produces extremely noisy conditions, including high levels of ultrasounds which mice are sensitive to), undetected leaking water bottles or
Mouse Models of Neurodegenerative Diseases
337
wet cleaning equipment (mops and buckets) left in animal rooms produce impossible to pin point confounding variables all of which increase stress and anxiety levels of mice, consequently negatively and variably affecting their performance in the behavioural tests. As an example, in Fig. 19.1 we provide the results of training two cohorts of same-age C3B6 mice (mixed genetic background of C3 (C3H) and B6 (C57BL/6)) in the spatial reference memory version of a water maze test, in two different animal facilities. Both cohorts were trained by the same laboratory assistant, highly experienced with the behavioural procedure, mouse husbandry, and certified in laboratory animal medicine. Mice in animal facility A were maintained in a quiet room and highly qualified and well-managed personnel provided high quality husbandry care. The environment in colony B was more stressful and mice were exposed to noisy conditions. The comparison of mice performance between colonies (main between-subject factor) and the analysis of their learning (days as repeated measure or withinsubject factor) revealed no significant difference in the average performance between the colonies (F(1,23) = 1.2, NS), but it also revealed a significant interaction between the colony conditions and learning rate of mice (F(4,92) = 4.2, p < 0.01, colony by training days interaction). Mice in colony A showed a significant learning through a rapid improvement in their search for a hidden escape platform over days (Fig. 19.1, p < 0.001 – simple effects ANOVA with days as a repeated measure), whereas the mice in colony B, however, did not show any signs of improvement over training period (Fig. 19.1).
Fig. 19.1. Learning acquisition in the spatial reference memory version of a water maze test of mixed background C3B6 mice in two different animal colonies. Mice in the colony A were kept in a quiet room with appropriate for behavioral experiments husbandry practices, while the colony B had increased noise level and sub-optimal for behavioral studies conditions. The mice trained in the colony A showed a significant improvement in finding a hidden platform location during training (their search path was on average about 5 m shorter at the end of training as compared to at the beginning of training). On the other hand, the mice in colony B showed no improvement (the rate of improvement between day 1 and day 5 was about 1 m). See text for further details.
338
Janus and Welzl
5. Conclusions
The main goal of the generation of animal models of human diseases is to better understand their underlying pathology, which should lead to the discovery and tests of potential therapeutics. In our review we focused on neurodegenerative diseases which present a complex, age-progressing dementia with rapidly progressing neuronal death. It might be unrealistic to think that the full complexity of human brain disorder can be modelled in a mouse using a crude, single genetic modification. However, as we tried to outline in this chapter, the interpretation of the results coming from mouse models of neurodegeneration should not follow “one gene – one disease” paradigm. At present, none of the existing mouse models of tauopathy fully replicate the characteristics of the modelled disease. Using these mouse models to study well-conserved signalling pathways in vivo may be better warranted than replicating fully the complexity of dementia. Genetically modified mouse models are integral part of modern drug discovery, but the interpretation of the obtained results must be careful and carried out within constraints of a model and mouse biology. The intensive screens of many compounds would require systematic, well-controlled, standardized phenotyping approach which is presented by Sacca and colleagues in Chapter 3 of this volume. The initial characterisation of new models or detailed characterisation of specific aspects of existing models should be based on careful experimental design, including larger number of mice in completely randomized experimental designs in animal facilities which promote maximal expression of mouse natural behaviour. Rigor of the experimental design will ensure replicability of the results across the labs and between different models. Even seemingly good and robust models of a human disease can yield many false-positive results due to differences in methodology or less rigorously carried out experiments (146). Our intention is to highlight important aspects of experimental design which are not always identified a priori and which may often generate confounding factors seriously biasing obtained data. We argue that our success in the endeavour of modelling human cognitive impairment may often depend on how well we understand the behaviour of a mouse. Detailed analysis of the potential and limitation of a model and the interpretation of the results within the framework of mouse biology should improve considerably detailed evaluation of potential therapeutics. Testing a specific hypothesis, negative results should be as valuable as positive ones and should be made available to the scientific community.
Mouse Models of Neurodegenerative Diseases
339
Acknowledgements We would like to thank Amanda Hanna, Heather Melrose, Douglas Wahlsten, and Alexander Gaukhman for valuable comments on earlier drafts of the manuscript.
References 1. Dubos, R. (1968) So human an animal. Charles Scribner’s Sons, New York. ¨ 2. Alzheimer, A. (1907) Uber eine eigenartige Erkankung der Hirnrinde. Allg. Z. Psychiatrie Psychisch-Gerlichtlich Med 64, 146–148. 3. Stelzmann, R. A., Schnitzlein, H. N. and Murtagh, F. R. (1995) An English translation of Alzheimer’s 1907 paper, “Uber eine eigenartige Erkankung der Hirnrinde”. Clin Anat 8, 429–431. 4. Braak, H. and Braak, E. (1994) Pathology of Alzheimer’s disease, in Neurodegenerative diseases (Calne, D. B., ed.). Saunders, Philadelphia, pp. 585–613. 5. Braak, H. and Braak, E. (1997) Diagnostic criteria for neuropathologic assessment of Alzheimer’s disease. Neurobiol Aging 18, S85–88. 6. Kosik, K. S. and Shimura, H. (2005) Phosphorylated tau and the neurodegenerative foldopathies. Biochim Biophys Acta 1739, 298–310. 7. Lee, V. M., Goedert, M. and Trojanowski, J. Q. (2001) Neurodegenerative tauopathies. Annu Rev Neurosci 24, 1121–1159. 8. Dickson, D. W. (2003) Neurodegeneration: The molecular pathology of dementia and movement disorders. ISN Neuropath Press, Basel. 9. Cotman, C. W. and Su, J. H. (1996) Mechanisms of neuronal death in Alzheimer’s disease. Brain Pathol 6, 493–506. 10. Terry, R. D. (2006) Alzheimer’s disease and the aging brain. J Geriatr Psychiatry Neurol 19, 125–128. 11. Davies, R. R., Hodges, J. R., Kril, J. J., Patterson, K., Halliday, G. M., et al. (2005) The pathological basis of semantic dementia. Brain 128, 1984–1995. 12. Iqbal, K., Alonso Adel, C., Chen, S., Chohan, M. O., El-Akkad, E., et al. (2005) Tau pathology in Alzheimer disease and other tauopathies. Biochim Biophys Acta 1739, 198–210.
13. Albert, M. S. (1996) Cognitive and neurobiologic markers of early Alzheimer’s disease. Proc Natl Acad Sci USA 93, 13547–13551. 14. Morgan, J. and Sethi, K. D. (2007) Differential diagnosis, in Handbook of Parkinson’s disease (Pahwa, R. and Lyons, K. E., eds.). Informa Healthcare USA, Inc., New York, pp. 29–47. 15. Petersen, R. C., Smith, G. E., Waring, S. C., Ivnik, R. J., Tangalos, E. G., et al. (1999) Mild cognitive impairment: clinical characterization and outcome. Arch Neurol 56, 303–308. 16. Knopman, D. S., DeKosky, S. T., Cummings, J. L., Chui, H., Corey-Bloom, J., et al. (2001) Practice parameter: diagnosis of dementia (an evidence-based review). Report of the Quality Standards Subcommittee of the American Academy of Neurology. Neurology 56, 1143–1153. 17. Luis, C. A., Loewenstein, D. A., Acevedo, A., Barker, W. W. and Duara, R. (2003) Mild cognitive impairment: directions for future research. Neurology 61, 438–444. 18. Maruyama, M., Arai, H., Sugita, M., Tanji, H., Higuchi, M., et al. (2001) Cerebrospinal fluid amyloid beta(1-42) levels in the mild cognitive impairment stage of Alzheimer’s disease. Exp Neurol 172, 433–436. 19. Riemenschneider, M., Lautenschlager, N., Wagenpfeil, S., Diehl, J., Drzezga, A., et al. (2002) Cerebrospinal fluid tau and beta-amyloid 42 proteins identify Alzheimer disease in subjects with mild cognitive impairment. Arch Neurol 59, 1729–1734. 20. Cummings, J. L. (2004) Dementia with lewy bodies: molecular pathogenesis and implications for classification. J Geriatr Psychiatry Neurol 17, 112–119. 21. Victoroff, J., Zarow, C., Mack, W. J., Hsu, E. and Chui, H. C. (1996) Physical aggression is associated with preservation of substantia nigra pars compacta in Alzheimer disease. Arch Neurol 53, 428–434.
340
Janus and Welzl
22. Pahwa, R. and Lyons, K. E. (2007) Handbook of Parkinson’s disease. Informa Healthcare USA, Inc., New York. 23. Whitehouse, P. J., Price, D. L., Struble, R. G., Clark, A. W., Coyle, J. T., et al. (1982) Alzheimer’s disease and senile dementia: loss of neurons in the basal forebrain. Science 215, 1237–1239. 24. Morrison, J. H. and Hof, P. R. (1997) Life and death of neurons in the aging brain. Science 278, 412–419. 25. Arnold, S. E., Hyman, B. T., Flory, J., Damasio, A. R. and Van Hoesen, G. W. (1991) The topographical and neuroanatomical distribution of neurofibrillary tangles and neuritic plaques in the cerebral cortex of patients with Alzheimer’s disease. Cereb Cortex 1, 103– 116. 26. Hyman, B. T., Van Hoesen, G. W., Damasio, A. R. and Barnes, C. L. (1984) Alzheimer’s disease: cell-specific pathology isolates the hippocampus formation. Science 225, 1168– 1170. 27. Horn, R., Ostertun, B., Fric, M., Solymosi, L., Steudel, A., et al. (1996) Atrophy of hippocampus in patients with Alzheimer’s disease and other diseases with memory impairment. Dementia 7, 182–186. 28. Samuel, W., Terry, R. D., Deteresa, R., Butters, N. and Masliah, E. (1994) Clinical correlates of cortical and nucleus basalis pathology in Alzheimer dementia. Arch Neurol 51, 772–778. 29. Karas, G. B., Burton, E. J., Rombouts, S. A., van Schijndel, R. A., O’Brien, J. T., et al. (2003) A comprehensive study of gray matter loss in patients with Alzheimer’s disease using optimized voxel-based morphometry. Neuroimage 18, 895–907. 30. Jack, C. R., Jr., Shiung, M. M., Gunter, J. L., O’Brien, P. C., Weigand, S. D., et al. (2004) Comparison of different MRI brain atrophy rate measures with clinical disease progression in AD. Neurology 62, 591–600. 31. Jope, R. S., Song, L. and Powers, R. E. (1997) Cholinergic activation of phosphoinositide signaling is impaired in Alzheimer’s disease brain. Neurobiol Aging 18, 111–120. 32. Tong, X. K. and Hamel, E. (1999) Regional cholinergic denervation of cortical microvessels and nitric oxide synthase-containing neurons in Alzheimer’s disease. Neuroscience 92, 163–175. 33. Mattson, M. P. and Pedersen, W. A. (1998) Effects of amyloid precursor protein derivatives and oxidative stress on basal forebrain cholinergic systems in Alzheimer’s disease. Int J Develop Neurosci 16, 737–753.
34. Grundke-Iqbal, I., Iqbal, K., Quinlan, M., Tung, Y. C., Zaidi, M. S., et al. (1986) Microtubule-associated protein tau. A component of Alzheimer paired helical filaments. J Biol Chem 261, 6084–6089. 35. Braak, H. and Braak, E. (1997) Frequency of stages of Alzheimer-related lesions in different age categories. Neurobiol Aging 18, 351–357. 36. Tomlinson, B. E., Blessed, G. and Roth, M. (1968) rvations on the brains of nondemented old people. J Neurol Sci 7, 331–356. 37. Kidd, M. (1964) Alzheimer’s disease–an electron microscopical study. Brain 87, 307– 320. 38. Goedert, M., Wischik, C. M., Crowther, R. A., Walker, J. E. and Klug, A. (1988) Cloning and sequencing of the cDNA encoding a core protein of the paired helical filament of Alzheimer disease: identification as the microtubule-associated protein tau. Proc Natl Acad Sci U S A 85, 4051–4055. 39. Kosik, K. S., Joachim, C. L. and Selkoe, D. J. (1986) Microtubule-associated protein tau (tau) is a major antigenic component of paired helical filaments in Alzheimer disease. Proc Natl Acad Sci U S A 83, 4044–4048. 40. Paulus, W. and Selim, M. (1990) Corticonigral degeneration with neuronal achromasia and basal neurofibrillary tangles. Acta Neuropathol (Berl) 81, 89–94. 41. Elizan, T. S., Hirano, A., Abrams, B. M., Need, R. L., Van Nuis, C., et al. (1966) Amyotrophic lateral sclerosis and parkinsonismdementia complex of Guam. Neurological reevaluation. Arch Neurol 14, 356–368. 42. Hirano, A., Malamud, N., Elizan, T. S. and Kurland, L. T. (1966) Amyotrophic lateral sclerosis and Parkinsonism-dementia complex on Guam. Further pathologic studies. Arch Neurol 15, 35–51. 43. Hof, P. R., Bouras, C., Perl, D. P., Sparks, D. L., Mehta, N., et al. (1995) Age-related distribution of neuropathologic changes in the cerebral cortex of patients with Down’s syndrome. Quantitative regional analysis and comparison with Alzheimer’s disease. Arch Neurol 52, 379–391. 44. Kiuchi, A., Otsuka, N., Namba, Y., Nakano, I. and Tomonaga, M. (1991) Presenile appearance of abundant Alzheimer’s neurofibrillary tangles without senile plaques in the brain in myotonic dystrophy. Acta Neuropathol (Berl) 82, 1–5. 45. Mott, R. T., Dickson, D. W., Trojanowski, J. Q., Zhukareva, V., Lee, V. M., et al. (2005) Neuropathologic, biochemical, and molecular characterization of the frontotemporal
Mouse Models of Neurodegenerative Diseases
46.
47.
48.
49.
50.
51.
52.
53.
54.
55.
56.
dementias. J Neuropathol Exp Neurol 64, 420–428. Glenner, G. G. and Wong, C. W. (1984) Alzheimer’s disease: initial report of the purification and characterization of a novel cerebrovascular amyloid protein. Biochem Biophys Res Commun 120, 885–890. Vigo-Pelfrey, C., Lee, D., Keim, P., Lieberburg, I. and Schenk, D. B. (1993) Characterization of beta-amyloid peptide from human cerebrospinal fluid. J Neurochem 61, 1965–1968. Roher, A., Wolfe, D., Palutke, M. and KuKuruga, D. (1986) Purification, ultrastructure, and chemical analysis of Alzheimer disease amyloid plaque core protein. Proc Natl Acad Sci U S A 83, 2662–2666. Iwatsubo, T., Odaka, A., Suzuki, N., Mizusawa, H., Nukina, N., et al. (1994) Visualization of A beta 42(43) and A beta 40 in senile plaques with end-specific A beta monoclonals: evidence that an initially deposited species is A beta 42(43). Neuron 13, 45–53. Naslund, J., Haroutunian, V., Mohs, R., Davis, K. L., Davies, P., et al. (2000) Correlation between elevated levels of amyloid betapeptide in the brain and cognitive decline. Jama 283, 1571–1577. Goate, A., Chartier-Harlin, M. C., Mullan, M., Brown, J., Crawford, F., et al. (1991) Segregation of a missense mutation in the amyloid precursor protein gene with familial Alzheimer’s disease. Nature 349, 704–706. Chartier-Harlin, M. C., Crawford, F., Hamandi, K., Mullan, M., Goate, A., et al. (1991) Screening for the beta-amyloid precursor protein mutation (APP717: Val—Ile) in extended pedigrees with early onset Alzheimer’s disease. Neurosci Lett 129, 134–135. Chartier-Harlin, M. C., Crawford, F., Houlden, H., Warren, A., Hughes, D., et al. (1991) Early-onset Alzheimer’s disease caused by mutations at codon 717 of the beta-amyloid precursor protein gene. Nature 353, 844–846. Sherrington, R., Rogaev, E. I., Liang, Y., Rogaeva, E. A., Levesque, G., et al. (1995) Cloning of a gene bearing missense mutations in early-onset familial Alzheimer’s disease. Nature 375, 754–760. Campion, D., Flaman, J. M., Brice, A., Hannequin, D., Dubois, B., et al. (1995) Mutations of the presenilin I gene in families with early-onset Alzheimer’s disease. Hum Mol Genet 4, 2373–2377. Levy-Lahad, E., Wasco, W., Poorkaj, P., Romano, D. M., Oshima, J., et al. (1995) Candidate gene for the chromosome 1 famil-
57.
58.
59.
60.
61.
62. 63.
64.
65.
66.
67.
341
ial Alzheimer’s disease locus. Science 269, 973–977. Rogaev, E. I., Sherrington, R., Rogaeva, E. A., Levesque, G., Ikeda, M., et al. (1995) Familial Alzheimer’s disease in kindreds with missense mutations in a gene on chromosome 1 related to the Alzheimer’s disease type 3 gene. Nature 376, 775–778. Clark, L. N., Poorkaj, P., Wszolek, Z., Geschwind, D. H., Nasreddine, Z. S., et al. (1998) Pathogenic implications of mutations in the tau gene in pallido-ponto-nigral degeneration and related neurodegenerative disorders linked to chromosome 17. Proc Natl Acad Sci U S A 95, 13103–13107. Hutton, M., Lendon, C. L., Rizzu, P., Baker, M., Froelich, S., et al. (1998) Association of missense and 5’-splice-site mutations in tau with the inherited dementia FTDP-17. Nature 393, 702–705. Spillantini, M. G., Murrell, J. R., Goedert, M., Farlow, M. R., Klug, A., et al. (1998) Mutation in the tau gene in familial multiple system tauopathy with presenile dementia. Proc Natl Acad Sci U S Am 95, 7737–7741. Ghetti, B., Hutton, M. and Wszolek, Z. (2003) Frontotemporal Dementia and Parkinsonism Linked to chromosome 17 associated with Tau gene mutations (FTDP17 T), in Neurodegeneration: The molecular pathology of dementia and movement disorders (Dickson, D., ed.). ISN Neuropath Press, Basel, pp. 81–155. Goedert, M. and Jakes, R. (2005) Mutations causing neurodegenerative tauopathies. Biochim Biophys Acta 1739, 240–250. Baker, M., Mackenzie, I. R., PickeringBrown, S. M., Gass, J., Rademakers, R., et al. (2006) Mutations in progranulin cause taunegative frontotemporal dementia linked to chromosome 17. Nature 442, 916–919. Mendez, M. F. and McMurtray, A. (2006) Frontotemporal dementia-like phenotypes associated with presenilin-1 mutations. Am J Alzheimers Dis Other Demen 21, 281–286. Games, D., Adams, D., Alessandrini, R., Barbour, R., Berthelette, P., et al. (1995) Alzheimer-type neuropathology in transgenic mice overexpressing V717F beta-amyloid precursor protein. Nature 373, 523–527. Greenberg, B. D., Savage, M. J., Howland, D. S., Ali, S. M., Siedlak, S. L., et al. (1996) APP transgenesis: approaches toward the development of animal models for Alzheimer disease neuropathology. Neurobiol Aging 17, 153–171. Ashe, K. (2001) Learning and memory in transgenic mice modelling Alzhiemer’s disease. Learn Mem 8, 301–308.
342
Janus and Welzl
68. Ashe, K. H. (2005) Mechanisms of memory loss in Abeta and tau mouse models. Biochem Soc Trans 33, 591–594. 69. Dodart, J. C., Mathis, C., Bales, K. R. and Paul, S. M. (2002) Does my mouse have Alzheimer’s disease? Genes, Brain Behav 1, 142–155. 70. Eriksen, J. L. and Janus, C. G. (2006) Plaques, tangles, and memory loss in mouse models of neurodegeneration. Behav Genet 37, 79–100. 71. Higgins, G. A. and Jacobsen, H. (2003) Transgenic mouse models of Alzheimer’s disease: phenotype and application. Behav Pharmacol 14, 419–438. 72. Janus, C. and Westaway, D. (2001) Transgenic mouse models of Alzheimer’s disease. Physiol Behav 73, 873–886. 73. Price, D. L. and Sisodia, S. S. (1998) Mutant genes in familial Alzheimer’s disease and transgenic models. Annu Rev Neurosci 21, 479–505. 74. Seabrook, G. R. and Rosahl, T. W. (1999) Transgenic animals relevant to Alzheimer’s disease. Neuropharmacology 38, 1–17. 75. van Leuven, F. (2000) Single and multiple transgenic mice as models for Alzheimer’s disease. Prog Neurobiol 61, 305–312. 76. Spires, T. L. and Hyman, B. T. (2005) Transgenic models of Alzheimer’s disease: learning from animals. NeuroRx 2, 423–437. 77. Wong, P. C., Cai, H., Borchelt, D. R. and Price, D. L. (2002) Genetically engineered mouse models of neurodegenerative diseases. Nat Neurosci 5, 633–639. 78. Hall, G. F. and Yao, J. (2005) Modeling tauopathy: a range of complementary approaches. Biochim Biophys Acta 1739, 224– 239. 79. Melrose, H. L., Lincoln, S. J., Tyndall, G. M. and Farrer, M. J. (2006) Parkinson’s disease: a rethink of rodent models. Exp Brain Res 173, 196–204. 80. Le Cudennec, C., Faure, A., Ly, M. and Delatour, B. (2008) One-year longitudinal evaluation of sensori-motor functions in APP751SL transgenic mice. Genes Brain Behav 7, 83–91. 81. Cui, S., Chesson, C. and Hope, R. (1993) Genetic variation within and between strains of outbred Swiss mice. Lab Anim 27, 116– 123. 82. Festing, M. F. (1974) Genetic reliability of commercially-bred laboratory mice. Lab Anim 8, 265–270. 83. Festing, M. F. (1974) Genetic monitoring of laboratory mouse colonies in the Medical Research Council Accreditation Scheme
84.
85.
86.
87.
88.
89.
90.
91.
92.
93.
94.
95. 96.
for the suppliers of laboratory animals. Lab Anim 8, 291–299. Crabbe, J. C., Wahlsten, D. and Dudek, B. C. (1999) Genetics of mouse behavior: interactions with laboratory environment. Science 284, 1670–1672. Wahlsten, D., Metten, P., Phillips, T. J., Boehm, S. L., 2nd, Burkhart-Kasch, S., et al. (2003) Different data from different labs: lessons from studies of gene-environment interaction. J Neurobiol 54, 283–311. Banbury Conference on Genetic Background in Mice (1997) Mutant mice and neuroscience: recommendations. Neuron 19, 755–759. Crawley, J. N., Belknap, J. K., Collins, A., Crabbe, J. C., Frankel, W., et al. (1997) Behavioral phenotypes of inbred mouse strains: implications and recommendations for molecular studies. Psychopharmacology (Berl) 132, 107–124. Takahashi, J. S., Pinto, L. H. and Vitaterna, M. H. (1994) Forward and reverse genetic approaches to behavior in the mouse. Science 264, 1724–1733. Dietrich, W. F., Lander, E. S., Smith, J. S., Moser, A. R., Gould, K. A., et al. (1993) Genetic identification of Mom-1, a major modifier locus affecting Min-induced intestinal neoplasia in the mouse. Cell 75, 631–639. Gould, K. A., Luongo, C., Moser, A. R., McNeley, M. K., Borenstein, N., et al. (1996) Genetic evaluation of candidate genes for the Mom1 modifier of intestinal neoplasia in mice. Genetics 144, 1777–1785. Wahlsten, D., Cooper, S. F. and Crabbe, J. C. (2005) Different rankings of inbred mouse strains on the Morris maze and a refined 4arm water escape task. Behav Brain Res 165, 36–51. Jimenez, A. J., Garcia-Fernandez, J. M., Gonzalez, B. and Foster, R. G. (1996) The spatio-temporal pattern of photoreceptor degeneration in the aged rd/rd mouse retina. Cell Tissue Res 284, 193–202. Ogilvie, J. M. and Speck, J. D. (2002) Dopamine has a critical role in photoreceptor degeneration in the rd mouse. Neurobiol Dis 10, 33–40. Sidman, R. L. and Green, M. C. (1965) Retinal Degeneration in the Mouse: Location of the Rd Locus in Linkage Group Xvii. J Hered 56, 23–29. Guillery, R. W. (1974) Visual pathways in albinos. Sci Am 230, 44–54. Rice, D. S., Williams, R. W. and Goldowitz, D. (1995) Genetic control of retinal
Mouse Models of Neurodegenerative Diseases
97.
98.
99.
100.
101.
102.
103.
104.
105.
106.
107.
projections in inbred strains of albino mice. J Comp Neurol 354, 459–469. Lamb, B. T., Sisodia, S. S., Lawler, A. M., Slunt, H. H., Kitt, C. A., et al. (1993) Introduction and expression of the 400 kilobase amyloid precursor protein gene in transgenic mice. Nat Genet 5, 22–30. Chishti, M. A., Yang, D. S., Janus, C., Phinney, A. L., Horne, P., et al. (2001) Early-onset amyloid deposition and cognitive deficits in transgenic mice expressing a double mutant form of amyloid precursor protein 695. Biol Chem 276, 21562–21570. Hsiao, K., Chapman, P., Nilsen, S., Eckman, C., Harigaya, Y., et al. (1996) Correlative memory deficits, a-beta elevation, and amyloid plaques in transgenic mice. Science 274, 99–102. Hsiao, K. K., Borchelt, D. R., Olson, K., Johannsdottir, R., Kitt, C., et al. (1995) Age-related CNS disorder and early death in transgenic FVB/N mice overexpressing Alzheimer amyloid precursor proteins. Neuron 15, 1203–1218. Sturchler-Pierrat, C., Abramowski, D., Duke, M., Wiederhold, K.-H., Mistl, C., et al. (1997) Two amyloid precursor protein transgenic mouse models with Alzheimer diseaselike pathology. Proc Natl Acad Sci U S A 94, 13287–13292. Squire, L. R. (1992) Memory and the hippocampus: a synthesis from findings with rats, monkeys, and humans. Psychol Rev 99, 195–231. Barnes, C. A., Rao, G. and McNaughton, B. L. (1996) Functional integrity of NMDAdependent LTP induction mechanisms across the lifespan of F-344 rats. Learn Mem 3, 124–137. Bliss, T. V. P. and Collingridge, G. L. (1993) A synaptic model of memory: long-term potentiation in the hippocampus. Nature 361, 31–39. Collingridge, G. L., Kehl, S. J. and McLennan, H. (1983) Excitatory amino acids in synaptic transmission in the Schaffer collateral-commissural pathway of the rat hippocampus. J Physiol (Lond) 334, 33–46. Eichenbaum, H. (1996) Learning from LTP: a comment on recent attempts to identify cellular and molecular mechanisms of memory. Learn Mem 3, 61–73. Fazeli, M. S., Errington, M. L., Dolphin, A. C. and Bliss, T. V. P. (1988) Long– term potentiation in the dentate gyrus of the anaesthetized rat is accompanied by an increase in protein efflux into push–pull cannula perfusates. Brain Res 473, 51–59.
343
108. Malenka, R. C. and Nicoll, R. A. (1993) NMDA-receptor-dependent synaptic plasticity: multiple forms and mechanisms. Trends Neurosci 16, 521–527. 109. Morris, R. G. M. (1989) Synaptic plasticity and learning: Selective impairment of learning in rats and blockade of long-term potentiation in vivo by the N-methyl-D-aspartate receptor antagonist AP5. J Neurosci 9, 3040– 3057. 110. Morris, R. G. M. (1990) Toward a representational hypothesis of the role of hippocampal synaptic plasticity in spatial and other forms of learning. Cold Spring Harb Symp Quant Biol 55, 161–173. 111. O’Keefe, J. and Nadel, L. (1978) The hippocampus as a cognitive map. Oxford University Press, Oxford. 112. Olton, D. S., Becker, J. T. and Handelman, G. E. (1979) Hippocampus space and memory. Behav Brain Sci 2, 313–365. 113. Milner, B. S., W.B. (1957) Loss of recent memory after bilateral hippocampal lesions. J Neurol Neurosurg Psychiat 20, 11–21. 114. Smith, M. L. and Milner, B. (1981) The role of the right hippocampus in the recall of spatial location. Neuropsychologia 19, 781–793. 115. Milner, B. (1965) Visually-guided mazelearning in man: effects of bilateral hippocampal, bilateral frontal hippocampal lesions. Neuropsychologia 3, 317–338. 116. Elgh, E., Lindqvist Astot, A., Fagerlund, M., Eriksson, S., Olsson, T., et al. (2006) Cognitive dysfunction, hippocampal atrophy and glucocorticoid feedback in Alzheimer’s disease. Biol Psychiatry 59, 155–161. 117. Rodriguez, G., Vitali, P., Calvini, P., Bordoni, C., Girtler, N., et al. (2000) Hippocampal perfusion in mild Alzheimer’s disease. Psychiatry Res 100, 65–74. 118. Carlesimo, G. A., Mauri, M., Graceffa, A. M., Fadda, L., Loasses, A., et al. (1998) Memory performances in young, elderly, and very old healthy individuals versus patients with Alzheimer’s disease: evidence for discontinuity between normal and pathological aging. J Clin Exp Neuropsychol 20, 14–29. 119. Ghilardi, M. F., Alberoni, M., Marelli, S., Rossi, M., Franceschi, M., et al. (1999) Impaired movement control in Alzheimer’s disease. Neurosci Lett 260, 45–48. 120. Kavcic, V. and Duffy, C. J. (2003) Attentional dynamics and visual perception: mechanisms of spatial disorientation in Alzheimer’s disease. Brain 126, 1173–1181. 121. Monacelli, A. M., Cushman, L. A., Kavcic, V. and Duffy, C. J. (2003) Spatial disorientation in Alzheimer’s disease: the remembrance of things passed. Neurology 61, 1491–1497.
344
Janus and Welzl
122. Pai, M. C. and Jacobs, W. J. (2004) Topographical disorientation in communityresiding patients with Alzheimer’s disease. Int J Geriatr Psychiatry 19, 250–255. 123. Rizzo, M., Anderson, S. W., Dawson, J. and Nawrot, M. (2000) Vision and cognition in Alzheimer’s disease. Neuropsychologia 38, 1157–1169. 124. Crawley, J. N. (2007) What’s wrong with my mouse?: Behavioural phenotypying of transgenic and knockout mice, 2nd ed. John Wiley & Sons, Inc., New Jersey. 125. Whishaw, I. Q. and Kolb, B. (2005) The behavior of the laboratory rat: a handbook with tests. Oxford University Press, Inc., Oxford. 126. Crawley, J. N. and Paylor, R. (1997) A proposed test battery and constellation of specific behavioral paradigms to investigate the behavioural phenotypes of transgenic and knockout mice. Horm Behav 31, 197–211. 127. Janus, C. (2004) Search strategies used by APP transgenic mice during spatial navigation in the Morris water maze. Learn Mem 11, 337–346. 128. Markowska, A. L., Long, J. M., Johnson, C. T. and Olton, D. S. (1993) Variable-interval probe test as a tool for repeated measurements of spatial memory in the water maze. Behav Neurosci 107, 627–632. 129. Spooner, R. I. W., Thomson, A., Morris, R. G. M. and Salter, S. H. (1994) The Atlantis platform: a new design and further developments of Buresova’s on-demand platform for the water maze. Learn Mem 1, 203–211. 130. Dudchenko, P. A., Goodridge, J. P., Seiterle, D. A. and Taube, J. S. (1997) Effects of repeated disorientation on the acquisition of spatial tasks in rats: dissociation between the appetetive radial arm maze and aversive water maze. J Exp Psychol 23, 194–210. 131. Chapillon, P. and Debouzie, A. (2000) BALB/c mice are not so bad in the Morris water maze. Behav Brain Res 117, 115–118. 132. Whishaw, I. Q. and Tomie, J. A. (1996) Of mice and mazes: similarities between mice and rats on dry land but not water mazes. Physiol Behav 60, 1191–1197. 133. Wahlsten, D., Metten, P. and Crabbe, J. C. (2003) A rating scale for wildness and ease of handling laboratory mice: results for 21 inbred strains tested in two laboratories. Genes Brain Behav 2, 71–79. 134. Corcoran, K. A., Lu, Y., Turner, R. S. and Maren, S. (2002) Overexpression of hAPPswe impairs rewarded alternation and contextual fear conditioning in a transgenic mouse model of Alzheimer’s disease. Learn Mem 9, 243–252.
135. Janus, C., Welzl, H., Hanna, A., Lovasic, L., Lane, N., et al. (2004) Impaired conditioned taste aversion learning in APP transgenic mice. Neurobiol Aging 25, 1213–1219. 136. Mumby, D. G. (2001) Perspectives on object-recognition memory following hippocampal damage: lessons from studies in rats. Behav Brain Res 127, 159–181. 137. Kumar-Singh, S., Dewachter, I., Moechars, D., Lubke, U., De Jonghe, C., et al. (2000) Behavioral disturbances without amyloid deposits in mice overexpressing human amyloid precursor protein with Flemish (A692G) or Dutch (E693Q) mutation. Neurobiol Dis 7, 9–22. 138. Lalonde, R., Dumont, M., Staufenbiel, M., Sturchler-Pierrat, C. and Strazielle, C. (2002) Spatial learning, exploration, anxiety, and motor coordination in female APP23 transgenic mice with the Swedish mutation. Brain Res 956, 36–44. 139. Gerlai, R., Fitch, T., Bales, K. R. and Gitter, B. D. (2002) Behavioral impairment of APP(V717F) mice in fear conditioning: is it only cognition? Behav Brain Res 136, 503–509. 140. Wahlsten, D., Bachmanov, A., Finn, D. A. and Crabbe, J. C. (2006) Stability of inbred mouse strain differences in behavior and brain size between laboratories and across decades. Proc Natl Acad Sci U S A 103, 16364–16369. 141. Scott, S., Kranz, J. E., Cole, J., Lincecum, J. M., Thompson, K., et al. (2008) Design, power, and interpretation of studies in the standard murine model of ALS. Amyotroph Lateral Scler 9, 4–15. 142. Sprinthall, R. C. (1987) Basic statistical analysis, 2nd ed. Addison-Wesley, Reading MA. 143. Machlis, L., Dodd, F. W. D. and Fentress, J. C. (1985) The pooling fallacy: problems arising when individuals contribute more than one observation to the data set. Zeitschrifte fur Tierpsychologie 68, 201–214. 144. Stevens, J. (1990) Intermediate statistics: a modern approach. Lawrence Erlbaum Associates, Inc., Publishers, Hillsdale, New Jersey. 145. Popper, K. (1963) Conjectures and refutations. Routledge and Keagan paul, London. 146. Benatar, M. (2007) Lost in translation: treatment trials in the SOD1 mouse and in human ALS. Neurobiol Dis 26, 1–13. 147. Rogers, D. C., Fisher, E. M., Brown, S. D., Peters, J., Hunter, A. J., et al. (1997) Behavioral and functional analysis of mouse phenotype: SHIRPA, a proposed protocol for comprehensive phenotype assessment. Mamm Genome 8, 711–713.
Mouse Models of Neurodegenerative Diseases 148. Morris, R. (1984) Developments of a watermaze procedure for studying spatal learning in the rat. J Neurosci Methods 11, 47–60. 149. Morris, R. G. M. (1981) Spatial localization does not require the presence of local cues. Learn Motiv 12, 239–260. 150. Wolfer, D. P. and Lipp, H. P. (2000) Dissecting the behaviour of transgenic mice: is it the mutation, the genetic background, or the environment. Exp Physiol 85, 627–634. 151. Gass, P., Wolfer, D. P., Balschun, D., Rudolph, D., Frey, U., et al. (1998) Deficits in memory tasks of mice with CREB mutations depend on gene dosage. Learn Mem 5, 274–288. 152. Wehner, J. M., Sleight, S. and Upchurch, M. (1990) Hippocampal protein kinase C activity is reduced in poor spatial learners. Brain Res 523, 181–187. 153. Westerman, M. A., Cooper-Blacketer, D., Mariash, A., Kotilinek, L., Kawarabayashi, T., et al. (2002) The relationship between Abeta and memory in the Tg2576 mouse model of Alzheimer’s disease. J Neurosci 22, 1858–1867. 154. Chen, G. Q., Chen, K. S., Knox, J., Inglis, J., Bernard, A., et al. (2000) A learning deficit related to age and beta-amyloid plaques in a mouse model of Alzheimer’s disease. Nature 408, 975–979. 155. Logue, S. F., Paylor, R. and Wehner, J. M. (1997) Hippocampal lesions cause learning deficits in inbred mice in the Morris water maze and conditioned-fear task. Behav Neurosci 111, 104–113. 156. Bohut, M. C., Soffi´e, M. and Poucet, B. (1989) Scopolamine affects the cognitive processes involved in selective object exploration more than locomotor activity. Psychobiology 17, 409–417. 157. Save, E., Poucet, B., Foreman, N. and M-C., B. (1992) Object exploration and reactions to spatial and nonspatial changes in hooded rats following damage to parietal cortex or hippocampal formation. Behav Neurosci 106, 447–456. 158. Hammond, R. S., Tull, L. E. and Stackman, R. W. (2004) On the delay-dependent involvement of the hippocampus in object recognition memory. Neurobiol Learn Mem 82, 26–34. 159. Vnek, N. and Rothblat, L. A. (1996) The hippocampus and long-term object memory in the rat. J Neurosci 16, 2780–2787. 160. LeDoux, J. E. (1993) Emotional memory systems in the brain. Behav Brain Res 58, 69-79. 161. Phillips, R. G. and Ledoux, J. E. (1992) Differential contribution of amygdala and
162. 163.
164.
165.
166. 167. 168.
169.
170.
171.
172.
173.
174.
345
hippocampus to cued and contextual fear conditioning. Behav Neurosci 106, 274–285. LeDoux, J. E. (2000) Emotion circuits in the brain. Annu Rev Neurosci 23, 155–184. Repa, J. C., Muller, J., Apergis, J., Desrochers, T. M., Zhou, Y., et al. (2001) Two different lateral amygdala cell populations contribute to the initiation and storage of memory. Nat Neurosci 4, 724–731. McEchron, M. D., Bouwmeester, H., Tseng, W., Weiss, C. and Disterhoft, J. F. (1998) Hippocampectomy disrupts auditory trace fear conditioning and contextual fear conditioning in the rat. Hippocampus 8, 1 638–646. Bourtchuladze, R., Frenguelli, B., Blendy, J., Cioffi, D., Schutz, G., et al. (1994) Deficient long-term memory in mice with a targeted mutation of the cAMP-responsive elementbinding protein. Cell 79, 59–68. Garcia, J., Hankins, W. G. and Rusinak, K. W. (1976) Flavor aversion studies. Science 192, 265–266. Revusky, S. H. and Bedarf, E. W. (1967) Association of illness with prior ingestion of novel foods. Science 155, 212–214. Rozin, P. and Kalat, J. W. (1971) Specific hungers and poison avoidance as adaptive specializations of learning. Psychol Rev 78, 459–486. Garcia, J., Kimeldorf, D. J. and Koeling, R. A. (1955) Conditioned aversion to saccharin resulting from exposure to gamma radiation. Science 122. Bures, J., Bermudez-Rattoni, F. and Yanamoto, T. (1998) Conditioned taste aversion: memory of a special kind. Oxford University Press, Oxford. Rosenblum, K., Meiri, N. and Dudai, Y. (1993) Taste memory: the role of protein synthesis in gustatory cortex. Behav Neural Biol 59, 49–56. Kruger, L. and Mantyh, P. W. (1989) Gustatory and related chemosensory systems, in Integrated systems of the CNS, Part II (Bj¨orklund, A., H¨okfelt, T. and Swanson, L. W., eds.). Elsevier Science Publishers, Amsterdam, pp. 323–411. Lamprecht, R. and Dudai, Y. (1996) Transient expression of c-Fos in rat amygdala during training is required for encoding conditioned taste aversion memory. Learn Mem 3, 31–41. Lamprecht, R., Hazvi, S. and Dudai, Y. (1997) cAMP response element-binding protein in the amygdala is required for longbut not short-term conditioned taste aversion memory. J Neurosci 17, 8443–8450.
Chapter 20 Neuromuscular Disease Models and Analysis Robert W. Burgess, Gregory A. Cox, and Kevin L. Seburn Abstract Neuromuscular diseases can affect the survival of peripheral neurons, their axons extending to peripheral targets, their synaptic connections onto those targets, or the targets themselves. Examples include motor neuron diseases such as amyotrophic lateral sclerosis, peripheral neuropathies, such as CharcotMarie-Tooth diseases, myasthenias, and muscular dystrophies. Characterizing these phenotypes in mouse models requires an integrated approach, examining both the nerve and the muscle histologically, anatomically, and functionally by electrophysiology. Defects observed at these levels can be related back to onset, severity, and progression, as assessed by “quality-of-life measures” including tests of gross motor performance such as gait or grip strength. This chapter describes methods for assessing neuromuscular disease models in mice, and how interpretation of these tests can be complicated by the inter-relatedness of the phenotypes. Key words: Motor neuron diseases, amyotrophic lateral sclerosis (ALS), spinal muscular atrophy (SMA), peripheral neuropathies, Charcot-Marie-Tooth diseases, hereditary motor and/or sensory neuropathies (HSMNs), congenital myasthenic syndromes, neuromuscular junction, muscular dystrophies, Duchenne’s disease.
1. Introduction In this chapter, we will describe genetic models of neuromuscular diseases in mice, and the methods of analysis used to understand the underlying pathophysiology. In discussing these diseases, we are specifically considering those that have their primary pathology in either the lower motor neurons (those with cell bodies in the ventral horn of the spinal cord, but including their peripheral axons and synaptic terminals) or the muscle fibers themselves. Other neurological conditions with exclusively upper motor G. Proetzel, M.V. Wiles (eds.), Mouse Models for Drug Discovery, Methods in Molecular Biology 602, DOI 10.1007/978-1-60761-058-8 20, © Humana Press, a part of Springer Science+Business Media, LLC 2010
347
348
Burgess, Cox, and Seburn
neuron, basal ganglia, or cerebellar involvement are not discussed. We are also emphasizing diseases that affect the peripheral motor system more than those affecting peripheral sensory systems. The disease models can be grouped into four primary categories (Fig. 20.1). (1) Motor neuron diseases, in which the death of motor neuron somata in the spinal cord results in denervation of the muscles, progressive flaccid paralysis, and usually premature death. In humans, examples of such diseases would include amyotrophic lateral sclerosis (ALS) and spinal muscular atrophy (SMA). (2) Peripheral neuropathies, in which axonal integrity or conduction is not maintained, resulting in axon degeneration and impaired connectivity of the nervous system and the musculature.
Fig. 20.1. Anatomy of the neuromuscular system. The neuromuscular system includes the spinal cord, the skeletal muscles, and the motor and sensory nerves that connect them. In cross-section, the spinal cord contains motor neuron cell bodies in the ventral horn and sensory neurons in the dorsal horn. Motor axons exit the spinal cord through the ventral roots and join sensory axons in the spinal and peripheral nerves. The cell bodies of the sensory neurons, which are bipolar neurons, are in the dorsal root ganglia, and enter the spinal cord through the dorsal roots. In the periphery, motor and sensory fibers are almost always comingled. Major sensory endings relevant to the neuromuscular system include muscle spindles, which sense muscle tension and proprioceptive endings such as Golgi tendon organs. Together with motor neurons these neurons innverate skeletal muscle and circuitry within the spinal cord balances force between flexor and extensor muscles, although this is crucial for proper neuromuscular function, it is not discussed in detail here. The diseases discussed in the chapter can generally be grouped into those that affect motor neurons in the spinal cord, those that affect the peripheral axons or synapses (neuromuscular junctions), and those that affect the muscles themselves. (Color figure is available online).
Neuromuscular Disease Models and Analysis
349
These diseases are often less severe than the motor neuron diseases, and although potentially debilitating, they do not necessarily result in shortened lifespan. In humans, these diseases include the Charcot-Marie-Tooth diseases, or hereditary motor and/or sensory neuropathies (HSMNs). These conditions may be the result of demyelination or defects intrinsic to the motor axon itself. (3) myasthenias, which are diseases of the neuromuscular junction (NMJ). These diseases impair synaptic transmission and can be caused by presynaptic or postsynaptic defects, or by defects in the extracellular matrix of the synaptic cleft, particularly by affecting the localization or activity or acetylcholine esterase (AChE) at the NMJ. In humans, these diseases are the Congenital Myasthenic Syndromes (CMS), which are rare but typically very severe. Adult onset myasthenias, such as Myasthenia Gravis, tend to be autoimmune and are not considered in this chapter. (4) Muscular dystrophies, which are diseases that affect the muscle fibers themselves, often leading to a loss of sarcolemmal integrity and degeneration of the muscle fibers, although some diseases may progress through other mechanisms. In humans, these diseases include congenital muscular dystrophies, such as Duchenne’s, and also limb-girdle muscular dystrophies. In reality, the diseases in humans and the phenotypes in mice cannot be so cleanly grouped. For instance, ALS results in lower motor neuron degeneration, but also has upper motor neuron involvement, or congenital muscular dystrophy type 1A, which has profound muscular dystrophy, but also severe hypomyelination in the central and peripheral nervous system. In some cases, mouse models fully recapitulate the human disease pathologies, and in other cases there are differences, which are both caveats for disease-oriented research and interesting opportunities for examining disease mechanisms through comparative physiology. However, recognizing the full phenotypic spectrum of these diseases is important for designing effective therapies. Overtly, the phenotypes of mice with any of these conditions may be quite similar. They are generally smaller, shaky either from weakness or from myelination defects, have difficulty fully supporting their weight, and atrophy or wasting, most often in the hind limbs. The methods we will outline below could be applied toward an existing model, for instance to determine whether a drug treatment or genetic manipulation changes the course of the disease, or to a novel model to determine the basis of an overt neuromuscular phenotype. It should be noted that the overt phenotype is a very sensitive predictor of neuromuscular dysfunction, although it may be subtle, such as in the mdx mouse, a model of Duchenne’s muscular dystrophy resulting from mutation in the dystrophin gene. We will first mention useful mouse models in each of the disease categories listed above, and then describe methods for
350
Burgess, Cox, and Seburn
assessing each pathology, as well as general methods, including basic neuromuscular physiology and behavioral tests of gross motor performance.
2. Materials 2.1. Mouse Strains Used for Neuromuscular Disease Research
The tables below list mouse models used in neuromuscular disease research. Strains were included based on their popularity, but also on their uniqueness in cases where a mutation illustrates an apparently distinct disease mechanism. In many cases, the targeted mutations (usually nulls) result in non-neurological phenotypes, in some cases embryonic lethality, and in others, no phenotype. Neuropathy phenotypes in such models are often generated by transgenic expression of human disease alleles, either in wild type of knockout backgrounds. General caveats in using any of these strains include variation in phenotype in different genetic backgrounds, increased phenotypic variability in hybrid genetic backgrounds, and the existence of multiple alleles for some genes, which may have markedly different characteristics. 1. Motor neuron disease models Examples of mouse mutations resulting in the loss of ventral horn motor neurons are given (Table 20.1). In many instances, these models are created by the transgenic overexpression of human disease genes, such as SOD1 alleles that cause ALS. In such cases, the loss-of-function mouse mutation frequently does not cause the same motor neuron disease. The mouse models may also vary in phenotype, not only because of the different alleles introduced, but also because of transgene expression levels (copy number) and integration sites. In other cases, the human and mouse phenotypes simply differ, such as Tbce, where human mutations cause congenital hypoparathyroidism, mental retardation, facial dysmorphism, and extreme growth failure (40), whereas the mouse mutations cause motor axon degeneration. The genetics of Smn1 in the mouse is also complicated. Mice have only one Smn gene, while humans have two. To replicate the human SMA condition, mouse models frequently involve the transgenic expression of human SMN2 and/or variant forms of human SMN1 in an Smn1 knockout background. Also noteworthy is the wide array of cellular functions these genes normally serve. Some have a logical association with motor neurons, such as those involved in axonal transport, but others are ubiquitous “house keeping” genes that selectively affect motor neurons when mutated.
Cellular function
RNA/DNA Helicase
Assembly of spliceosomal snRNP complexes
Tubulin-specific chaperone
p150(glued), component of dynein/dynactin complex.
P50 (dynamitin) component of dynein/dynactin complex
Dominant gain of toxic function unrelated to dismutase activity
Gene
Ighmbp2
Smn1
Tbce
Dctn1
Dctn2
SOD1
Table 20.1 Motor neuron disease models
Tg
Tg
KO/KI, Tg
Sp
Tg + KO
Sp
Tg/KO/ knockin/ spontaneous
ALS
Unknown
dSBMA
Kenny–Caffey syndrome/ HRD
SMA
DSMA1
Human disease
Motor neuron degeneration, distal axonopathy, and NMJ denervation
Late onset (>10 months) slowly progressive motor neuron degeneration and denervation of muscle
Tg and KI hets, late-onset (>10 mo), slowly progressive weakness and altered NMJ integrity.
Caudiocranial degeneration of motor axons, ∼7 week lifespan
KO is embryonic lethal, addition of TgSMN2, SMN27 and/or SMN1A2G produce lifespans ranging from 2 to 227 d
Motor neuron degeneration, cardiomyopathy ∼8–10 week lifespan
Mouse phenotype 2J
(52, 53)
(48–51)
(39, 45–47)
Reference
(57)
(continued)
mSod1-G86R, (58–64) hSOD1G93A, G85R, G37R, H46R/H48Q
Tg-Dctn2
Dctn1tm1Cai ; (54–56) TgDCTN1(G59S)
pmn
Tg-SMN2, TgSMN27 , TgSMN1A2G ; Smn1tm1Msd
nmd , nmd
J
Alleles
Neuromuscular Disease Models and Analysis 351
Cellular function
ER localized, lipid synthesis, transport, or sensing
Axonal transport
Phosphoinositide phosphatase
Golgi-associated retrograde protein (GARP) complex of vesicle sorting proteins
Gene
Cln8
Dync1h1
Fig4
Vps54
Table 20.1 (continued)
Sp, KO
Sp
Sp
Sp
Tg/KO/ knockin/ spontaneous
Unknown
CMT4J, ALS, and PLS
CLN8; EPMR– epilepsy, progressive with mental retardation Unknown
Human disease
Vacuolization and degeneration of brainstem and ventral horn neurons; KO embryonic lethal
Pale coat color, axonal degeneration – motor and sensory, limited segmental demyelination
Hets: sensory neuropathy with muscle spindle deficiency. Hom: neonatal lethality
Neuronal ceroid lipofuscinoses and retinal photoreceptor degeneration
Mouse phenotype
(71–73)
(68–70)
(65–67)
Reference
wr, (74) Gt(RRI497) Byg
plt
Loa, Cra1, Swl
mnd
Alleles
352 Burgess, Cox, and Seburn
Neuromuscular Disease Models and Analysis
353
2. Peripheral neuropathy models Examples of demyelinating (type 1), X-linked, axonal (type 2), and recessive (type 4) Charcot-Marie-Tooth diseases are given (Table 20.2). The demyelinating and recessive models fairly accurately reproduce the human disease phenotypes caused by mutations in their orthologous genes. For genes such as Pmp22 and Egr2, there are many alleles in mice, with some phenotypic differences. This is particularly the case for Egr2, which is involved in numerous developmental pathways, and a demyelinating neuropathy is present only in a partial-loss-of-function allele (Egr2lo ), null mice have hypomyelination in the CNS and PNS, but do not survive long enough to study. Interestingly, the peripheral neuropathy phenotype is more difficult to reproduce in the axonal models. Motor neuron phenotypes are observed in transgenic mice overexpressing mutant forms of Mfn2 or Nefl, but these phenotypes are not evident in straight lossof-function alleles. The Hspb1 mutant mice do not have a peripheral neuropathy, which may again indicate a requirement for expression of a mutant form of the protein, or it may reflect differences in expression pattern and function between mice and humans. The GarsNmf249 mouse provides a phenotypically accurate axonal CMT model. The allele was identified in a phenotype-driven screen for neuromuscular disease mutations, and no dominant phenotype is observed in mice lacking Gars expression at the RNA level. This suggests the mutant form of the protein is required to cause the phenotype and is an argument in favor of phenotype-based approaches to generating accurate disease models. Note that some mutations, such as the aggregate-forming Neflpro transgenics and other alleles of FIG4 in humans, can cause motor neuron loss and not just axonopathy, possibly indicating that some axonal neuropathies may in fact be part of a spectrum of motor neuron disease manifestations. It should also be noted that there are several identified CMT-genes in humans that do not yet have mouse models, such as RAB7, GDAP1, HSPB8, and YARS. 3. Myasthenia models Examples of mutations causing myasthenia phenotypes are given (Table 20.3), including those with presynaptic, synaptic, trans-synaptic signaling, or postsynaptic functions. In several cases, mouse mutations with relevant phenotypes, or in relevant pathways, are listed in the absence of an established human disease caused by mutations in the orthologous gene (Vacht, Ache, Agrn, Lrp4). In addition, the mouse mutations are frequently a complete loss-offunction (knockout), whereas the human mutations are a
Cellular function
Peripheral myelin packing
Peripheral myelin component, transmembrane cell adhesion
Zinc finger transcription factor, immediate early gene
Connexin32, gap junction component
Pmp22, peripheral myelin protein 22 kDa
Mpz, myelin Protein Zero
Egr2, early growth response 2
Gjb1, gap junction protein beta1
Demyelinating, type 1 CMTs
Gene
KO, Tg
KO, KI
KO, Tg
Sp, KO, ENU
Tg/KO/ spontaneous
Table 20.2 Peripheral neuropathy disease models
X-linked CMT
Congenital hypomyelinating neuropathy, CMT1D
CMT1B
CMT1A
Human disease
Late onset demyelinating neuropathy
Embryonic letha (null)l, lack of CNS, PNS myelination
Demyelinating neuropathy, hypomyelination
Demyelinating neuropathy
Mouse phenotype
Null, TgWT and R142W
Several, including null and Egr2lo
Null, Tg– WT, S63C, S63del.
Trembler (3)
Alleles
(continued)
(81–83)
(79, 80)
(76–78)
(75)
Reference
354 Burgess, Cox, and Seburn
Cellular function
Fig4
Prx, periaxin
PDZ-scaffolding protein
Chaperone
Hspb1, heat shock protein 1, 27 kDa
Recessive CMTs
KO, Tg
Sp
KO CMT4J
CMT4F
CMT2F
CMT2E
Sp, KO
Glycine to tRNAgly aminoacylation
Gars, glycyl tRNA synthetase Nefl, neurofilament light chain KO
CMT2D SMA-DV
Tg
CMT2A
Human disease
Mitochondrial fusion/transport
Tg/KO/ spontaneous
Mfn2, mitofusin2
Axonal, type 2 CMTs
Gene
Table 20.2 (continued)
Demyelination, neuropathic pain, shortened internodal distance
No phenotype (expression pattern?)
Impaired regeneration (null), motor neuron loss (Tg)
Loss of large diameter sensory and motor axons in periphery
In Tg, motor neuropathy, hind limb weakness
Mouse phenotype
Pale tremor
Null
Null
Null, Tg.Neflpro
Nmf249 XM256
Null, TgT105M
Alleles
(72)
(4, 88)
(87)
(85, 86)
(3)
(84)
Reference
Neuromuscular Disease Models and Analysis 355
ACh degradation
Ache, acetylcholine esterase KO
KO
KO
KO
Agrin receptor (with Lrp4), tyrosine kinase receptor Agrin receptor
Musk, musclespecific kinase
Lrp4, LDLreceptorrelated protein 4
Nerve-derived differentiation factor
Agrn, agrin
ENU
KO
KO
Synaptic differentiation/structure, trans-synaptic signaling
Colagen like tail, AChE anchoring
ACh vesicular transport
ACh synthesis
Cellular function
Colq, collagenic tail of AChE
Synaptic
Vacht, vesicular acetylcholine transporter
Presynaptic Chat, choline acetyltransferase
Gene
Tg/KO/ sponta neous
Table 20.3 Myasthenic syndrome disease models
CMS with AChR deficiency
End plate AChE deficiency
CMS with episodic apnea
Human disease
Neonatal lethal, no postsynaptic differentiation
Neonatal lethal, no postsynaptic differentiation
Neonatal lethal, no postsynaptic differentiation
Severe myasthenia, weakness, reduced viability
Severe myasthenia, weakness, reduced viability
Myasthenia, behavioral changes
No synaptic transmission, neonatal lethal
Mouse phenotype
Null
Null (and autoimmune)
Deletion, isoformspecific, conditional
Null
Null
Partial loss of function
Conditional null
Alleles
(continued)
(102)
(101)
(96–100)
(94, 95)
(93)
(92)
(89–91)
Reference
356 Burgess, Cox, and Seburn
Intracellular scaffolding of AChRs
Intracellular adaptor protein
Extracellular matrix protein
Rapsn, rapsyn
Dok7, downstream of kinase-7
Lamb2, lamininbeta2
KO
KO
KO
Tg/KO/ sponta neous
AChR subunit
AChR subunit (adult)
AChR subunit (embryonic)
Chrnb1, acetylcholine receptor beta1
Chrne, acetylcholine receptor epsilon subunit
Chrng, acetylcholine receptor gamma subunit KO, KI
KO
KI, phosphoryation deficient
Acetylcholine receptor (AChR) defects (postsynaptic)
Cellular function
Gene
Table 20.3 (continued)
Escobar syndrome
CMS with AChR deficiency, fast, and slow channel
CMS with AChR deficiency and slow channel
Pierson syndrome (kidney)
CMS with AChR deficiency CMS1B, limb girdle
Human disease
Neonatal lethal (null). Broadened end plate band (KI)
Death at 2–3 months with progressive weakness
NMJ morphology changes
Failure of NMJ maturation, death at 4 weeks
Neonatal lethal, no postsynaptic differentiation
Neonatal lethal, no AChR clustering
Mouse phenotype
Null, epsilon knockin
Null
Three intracellular tyrosines
Null
Null
Null
Alleles
(109– 111)
(107, 108)
(106)
(105)
(104)
(103)
Reference
Neuromuscular Disease Models and Analysis 357
358
Burgess, Cox, and Seburn
partial-loss-of-function (Rapsn, Musk, Dok7). Such models are useful for understanding gene function, but accurate models of the associated human disease will require more specific alterations, such as generation of knockin alleles recreating the human genetic changes. In some cases, loxPconditional alleles offer compromise models, where the timing and extent of the loss of function can be controlled, even though the alleles created are complete loss-of-function following cre-mediated excision (Chat, Agrn). Elimination of Lamb2 has both pre- and postsynaptic effects in mice, but human mutations are characterized by kidney dysfunction and not myasthenia (Pierson Syndrome). Kidney defects are also observed in the Lamb2 mutant mice (41). In humans, mutations in nicotinic acetylcholine receptor subunits (CHRNA, CHRNB1, CHRND, CHRNG, CHRNE) cause a variety of myasthenic syndromes. The mouse models of these diseases are incomplete largely because the function of these genes is well understood (encoding ligand-gated ion channels), and the biophysical properties of variants can be effectively studied using in vitro and heterologous systems. 4. Muscular dystrophy models Examples of mouse mutations exhibiting a muscular dystrophy phenotype are presented Table 20.4. A prevalent mechanism for these mutations is a failure in linking the extracellular matrix to the intracellular cytoskeleton via components of the dystroglycan glycoprotein complex (DGC), resulting in compromised sarcolemmal integrity. Similar dystrophies are caused by defects in membrane repair, which is critical for maintaining mechanically contracting muscle fibers. A second class of genes causing muscular dystrophy includes those associated with the structure of the sarcomere and the cytoskeleton. Finally, a variety of other examples are given that arise through more disparate or poorly understood mechanisms ranging from phospholipid biosynthesis to RNA binding. In many cases, multiple alleles exist for any given gene, often with varying phenotypes that correlate with the severity of the loss of function. For example, Lama2 was first identified as dystrophia muscularis (dy) in 1955, an allele that is a severe hypomorph. Additional alleles including milder, partial loss of function (dy2J, dy7J), and a complete null (tm1Stk, also called dy3K) have been identified or engineered since. The appropriate allele for use in research should be chosen based on the experiment, for example, the severity of a hypomorphic mutation could be either enhanced or suppressed, in contrast, a null allele has no retained activity so any suppression has to come through other pathways or activation of downstream factors.
Cellular function
DGC-associated, sarcolemmal integrity
Glycosylation of ␣-dystroglycan ECM
Membrane repair
DGC associated
Type VI collagen
Dmd
Large
Lama2
Dysf
Dtna
Col6a1
Sarcolemmal/ ECM
Gene
KO
KO
Sp, KO
Sp, ENU, KO
Sp
Sp, ENU, KO
Tg/KO/ KI/ENU/ spontaneous
Table 20.4 Muscular dystrophy disease models
Bethlem myopathy and Ullrich congenital muscular dystrophy
Left ventricular noncompaction (LVNC)
LGMD2B, Miyoshi myopathy
MDC1A
MDC1D
Duchenne and Becker muscular dystrophy
Human disease
Dystrophic pathology as early as 3 d; mitochondrial dysfunction
Dystrophic pathology by 1 month and progressing to 6 months
Dystrophic pathology by 3 months, progressive necrosis and fatty infiltration
Progressive dystrophy and ocular abnormalities Variable severity among alleles, progressive dystrophy, and hypomyelination
Dystrophic pathology by 2 week, particularly progressive in diaphragm
Mouse phenotype
tm1Gmb
tm1Jrs
im (SJL/J), prmd (A/J); tm1Kcam, tm1Meho
dy, -2 J, -3 J, -6 J, -7 J; tm1Eeng, tm1Stk
myd, vls
mdx, -2Cv, -3Cv, -4Cv, -5Cv; tm1Mok, tm1Khan
Alleles
(continued)
(130, 131)
(129)
(126–128)
(121–125)
(119, 120)
(112–118)
Reference
Neuromuscular Disease Models and Analysis 359
Type XV collagen
Structural protein of caveolae
Col15a1
Cav3
Sarcomere assembly and passive tension
Alpha-B crystalline
Regulate Hsp70 family molecular chaperones Intermediate filament; Z discs
Z-disc protein
Ttn
Cryab
Bag3
Myot
Des
Skeletal actin
Acta1
Cytoskeletal/sarcomeric
Cellular function
Gene
Table 20.4 (continued)
Tg, KO
KO
KO
KO
Sp, ENU, KO
KO
KO
KO
Tg/KO/ KI/ENU/ spontaneous
LGMD1A
Desminrelated myopathy
Myofibrillar myopathy
Myofibrillar myopathy
LGMD2J, TMD
Nemaline myopathy 3
LGMD1C
Unknown
Human disease
Myofibrillar myopathy, KO no pathology
Tg-Myot(T57I); tm1Moza
tm1Cap, tm1Cba, tm1Ltho
Gt(OST16086)Lex
Myofibrillar degeneration ∼4 week lifespan Cardiac, skeletal and smooth muscle myofibrillar defects
tm1Wawr
mdm; shru; tm1Her
tm1Jll
tm1Mls
tm1Pih
Alleles
Muscle wasting and dystrophic pathology after 40 week
Progressive dystrophy, onset by 2 week, 8–10 week lifespan. KO embryonic lethal
Neonatal lethal ∼10 d
Mild myopathic changes; dilated and longitudinally oriented T tubules
Dystrophic pathology after 3 months; microvessel damage
Mouse phenotype
(continued)
(145, 146)
(142–144)
(140, 141)
(139)
(136–138)
(134, 135)
(133)
(132)
Reference
360 Burgess, Cox, and Seburn
Phosphatidyl choline synthesis
PI(3)P-phosphatase
Component of nuclear lamina
Mitochondrial electron transport
CUG repeat expansion in 3 UTR
CUG RNA binding protein
Sialic acid biosynthesis
Mtm1
Lmna
Cox10
Dmpk
Mbnl1
Gne
Cellular function
Chkb
Other
Gene
Table 20.4 (continued)
KO/Tg
Distal myopathy with rimmed vacuoles (DMRV)
Myotonic dystrophy
Myotonic dystrophy 1 (DM1)
Tg, KO
KO
COX deficiency
X-linked myotubular myopathy EDMD2 and EDMD3
Unknown
Human disease
Conditional KO
KI, KO
KO
Sp
Tg/KO/ KI/ENU/ spontaneous
(156)
Mbnl3/3 (tm1Sws) tm1Sngi, Tg-GNE (D176V)
Myotonia; myonuclear RNA foci Decreased motor performance >30 weeks, inclusion bodies
(157)
(152–155)
Myotonia; myonuclear RNA foci
Tg-HSALR , Tg-DM300, Tg-Dm960, tm1Rdd
(149, 150)
LmnaH222P/H222 , tm1Stw
(151)
(148)
(147)
Reference
tm1Jman
rmd
Alleles
tm1Ctm
Slowly progressive myopathy after 3 months
Knockin mutations develop later dystrophy, 9 months lifespan; KO alleles lifespan ˜8 week
Progressive myopathy by 4 week with a 6–14 week lifespan
Rostrocaudal progressive dystrophy with fatty infiltration
Mouse phenotype
Neuromuscular Disease Models and Analysis 361
362
Burgess, Cox, and Seburn
5. Research tools A variety of strains that are useful in neuromuscular disease research are listed Table 20.5. Cre transgenic strains mediate the deletion of sequences flanked by loxP sites (“floxed” sequences), the most common construct used in condition mutations in mice. Tissue specificity is achieved by selectively driving cre expression, and both neuron-specific and musclespecific strains are listed. In addition, temporal control can be achieved by using inducible cre systems such as the SLICK mice. A second class of transgenic mice that greatly simplify analysis is those expressing markers such as fluorescent proteins (largely GFP derivatives) in selected cell types. Many of these strains incorporate the same DNA construct and variable expression occurs as a result of genomic insertion site of the transgene (position effect variegation), for example, many constructs use the Thy1 promoter to drive expression in the nervous system, but the precise expression pattern of each of these strains is distinct. This also highlights a caveat for anyone considering transgenic studies, the expression level, and pattern of each transgene must be confirmed before results can be reliably interpreted. Finally, the Wlds allele is an interesting spontaneous genomic rearrangement that creates a fusion of the Ube4e and Nmnat genes. This mutation suppresses distal axonal degeneration in a number of acute (injury) and genetic models. It appears that the Nmnat portion of the fusion may be primarily responsible for conferring this activity, and it is proposed to function as a chaperone and to prevent mitochondrial dysfunction (42–44). This strain can therefore be useful for exploring pathogenic mechanisms in different neurodegenerative conditions. 2.2. Histology
1. Hematoxylin and eosin stain (H&E) 2. Cresyl violet/luxol fast blue stain (CV/LFB) 3. Bouin’s fixative 4. Decalcifying solution (Fisher Cal-Ex) 5. Paraffin 6. Xylene 7. Ethanol series (70% to 95% to 100%)
2.3. Assessing Motor Neuron Loss
1. Strong fixative: 2% paraformaldehyde, 2% glutaraldehyde in 0.1 M cacodylate buffer 2. No. 2 forceps 3. Toluidine Blue
Neuromuscular Disease Models and Analysis
363
Table 20.5 Strains useful for neuromuscular research Strain name
Jax #
Phenotype/utility
Reference
B6.Cg-Tg(Nes-
3771
Pan-neuronal expression of cre from embryonic ages
(158)
HB9-cre, B6.129S1Mnx1tm4(cre)Tmj /J HSA-cre
6600 NA
Motor neuron-specific expression of cre from embryonic ages Muscle-specific expression of cre from myoblast stage (including satellite cells in adult)
(159, 160) (161)
Mck-cre
NA
Muscle-specific expression of cre from early myotube stage
(162)
7845
Knockin allele, skeletal muscle, and dermis expression of cre
(163)
NA 7606, 7610
Schwann cell-specific expression of cre Inducible cre (ESR-cre) expression coupled to YFP expression, driven by Thy1 promoter
(164) (165)
Thy1-YFP-16 B6.Cg-Tg(Thy1-YFP)16Jrs/J
3709
YFP expression in all motor neurons embryonic ages
(166)
Thy1-YFP-H, B6.Cg-Tg(Thy1YFPH)2Jrs/J
3782
(166)
Thy1-CFP-23 B6.Cg-Tg(Thy1CFP)23Jrs/J
3710
YFP expression in sparse subset of motor neurons beginning at 3–4 weeks of age (useful for tracing motor units) CFP expression in all motor neurons
Thy1-Mito-CFP B6;CB-Tg(Thy1CFP/COX8A)Lich/J
7940, 6614, 6617
Mitochondrially localized CFP in all motor neurons, useful for transport/trafficking studies
(167)
S100-XFP, B6;D2-Tg(S100BEYFP)1Wjt/J
5620, 5621
YFP of GFP expression in Schwann cells
(168)
HB9-GFP, B6.Cg-Tg(Hlxb9GFP)1Tmj/J
5029
Motor neuron-specific expression of GFP (also useful differentiation/cell type marker in vitro)
(169)
Brainbow mice B6.Cg-Tg(Thy1Brainbow1.0)Lich/J B6.Cg-Tg(Thy1Brainbow2.1)Lich/J
7901, 7910, 7911, 7921
Multi-colored neuronal expression
(170)
NA
Slowed distal axonal degeneration in response to a variety of insults
(171)
Cre transgenic strains Nestin-cre, cre)1Kln/J
Myf5-cre B6;129S4-Myf5
tm3(cre)Sor
/J
P0-cre SLICK mice, B6.Cg-Tg(Thy1cre/ESR1,-EYFP)Gfng/J Fluorescent protein strains
(166)
Other Wlds Wallerian degeneration slow
364
Burgess, Cox, and Seburn
2.4. Internodal Distance Assessment
1. No. 5 forceps
2.5. Light Microscopy of the NMJ
1. 2% paraformaldehyde (electron microscopy grade)
2. 30-G needle
2. Anti-neurofilament marker, e.g., the monoclonal antibody 2H3 and anti-SV2 (Developmental Studies Hybridoma Bank, University of Iowa, see http://dshb.biology.uiowa. edu) 3. Postsynaptic receptor marker, e.g., fluorescent conjugates of alpha-bungarotoxin (Molecular Probes, Invitrogen) 4. Schwann cell marker: anti-S100 antibody
2.6. Electron Microscopy of the NMJ
1. Paraformaldehyde (electron microscopy grade)
2.7. Evan’s Blue staining
1. Sterile saline solution
2. Fixative: 2% paraformaldehyde, 2% glutaraldehyde in 0.1 Mm cacodylate buffer
2. Evan’s blue (Sigma), dissolved in sterile saline at 10 mg/ml
3. Methods The protocols below provide an intermediate level of detail. Specifics are provided when they are critical to the success of the experiment, but otherwise should be sought in the references. Our goal is to provide a list of experimental approaches for assessing neuromuscular function and to illustrate caveats and possible interpretations so that appropriate controls can be used and accurate conclusions can be drawn. 3.1. Motor Neuron Diseases 3.1.1. Histological Analysis of the Spinal Cord
1. Dissect the vertebral column. Note: the cell bodies of lumbar motor neurons are more rostral than their exit points, i.e., L4 cells bodies are in upper lumbar or even lower thoracic vertebrae, even though the spinal nerve exits at the fourth lumbar vertebrae. The more caudal vertebral column contains only the cauda equina, the dorsal, and ventral roots of lower lumbar and sacral motor neurons (Fig. 20.2). 2. Fix the spinal cord in the intact vertebral column using Bouin’s fixative, either by immersion after dissection or by transcardial perfusion prior to dissection. The bone of the vertebrae will eventually decalcify in Bouin’s (requiring 2–4 weeks), or after 24 h, the sample can be transferred to decalcifying solution overnight.
Neuromuscular Disease Models and Analysis
365
Fig. 20.2. The spinal cord and cauda equina. The caudal spinal cord does not completely fill the caudal vertebral column. As a result, the dorsal and ventral roots of lumbar and sacral neurons are very long, spanning from the more rostral cell bodies to the more caudal exit points. Therefore, the dissected spinal cord resembles a horse’s tail. The practical effect of this anatomy is that it becomes very hard to reliably determine the level of the spinal cord that is being studied in any given cross-section.
3. Rinse the sample thoroughly (>12 h) in water and process for paraffin embedding (dehydrate through at ethanol series and then to xylene, do not allow nervous system samples to sit longer than necessary in ethanol or white matter tracts will look like Swiss cheese). 4. Embed the samples for cross-section (horizontal section). 5. Stain 4–5 m sections using standard hematoxylin and eosin protocols (H&E) or with cresyl violet/luxol fast blue (CV/LFB). H&E stains protein rich cells (eosin) and counterstains nuclei (hemotoxylin). CV/LFB also stains protein,
366
Burgess, Cox, and Seburn
but also stains myelin. If available, an automated tissue processor is recommended for consistency of staining. 6. Interpretation: dying cells in the ventral horn of the spinal cord may be distinguished by loss of Nissl substance in the cell bodies, by pyknotic nuclei, and by pink cytoplasm in cells that are early in the degeneration process (see Note 1). 3.1.2. Assessing Motor Neuron Number (Loss)
In the spinal cord, motor neuron number can be assessed in cross-sections as described above. However, there are two primary challenges. First, motor neuron cell bodies need to be accurately identified in the ventral horn. Markers such as HB9-GFP, Thy1.2-YFP, or ChAT can be useful to positively identify motor neurons, but these are often incompatible with strong fixatives, paraffin embedding, or histological counterstains. A second challenge is being sure that comparisons are made at exactly the same level of the spinal cord, which is difficult considering the cord does not fill the full rostral/caudal extent of the vertebral column and the number of motor neurons per cross-section varies greatly with level. An easier approach that we prefer is to count axons in the ventral roots as they exit the spinal cord (Fig. 20.3). The ventral roots can be dissected, embedded, cross-sectioned, and myelinated axons can be counted after staining. A decrease in axon number should reflect a loss in motor neurons in the ventral horn. 1. Perfuse an animal transcardially with strong fixative (2% paraformaldehyde, 2% glutaraldehyde in 0.1 M cacodylate buffer, standard electron microscopy fixative). 2. Expose the sciatic nerve dorsally adjacent to the femur and follow it proximally to the sciatic notch where it disappears under the pelvis. 3. Using a pair fine blunt-point scissors carefully split pelvis in the direction followed by the nerve and gently separate the split bone. Continue to follow the nerve until branches into L6, the L5, and L4 spinal nerves are evident (NB: L6 is most distal and smallest). 4. Perform a hemilaminectomy (removing the dorsal bone of the vertebrae) in the lumbar region using a pair of No. 2 forceps to remove small pieces of bone until spinal cord is exposed from the dorsal midline laterally to the entry point of the spinal nerves. 5. Grasping each spinal nerve, follow it to its bifurcation into the ventral (motor) and dorsal (sensory) roots (see Note 2). 6. Cut the ventral root free at the bifurcation point and as close to the spinal cord as possible (NB: ventral roots enter cord proximal to spinal nerve entry point). 7. Process for plastic embedding (as for electron microscopy).
Neuromuscular Disease Models and Analysis
367
Fig. 20.3. Dorsal and ventral root dissection. The anatomy of the ventral and dorsal roots during dissection is shown. Mouse is prone (rostral is up) and the left sciatic nerve (Sci) has been exposed along the femur and to its origin at the lumbar spinal nerves. The spinal nerves are labeled (lumbar 3 through 6). The ischium (Isch) has been removed to follow the sciatic nerve. The inset shows the L4 root separated into its ventral and dorsal roots, note the DRG associated with the dorsal root. The scale bar is 6 mm. (Color figure is available online).
8. Cut 500 nm cross-sections near the middle of the sample to avoid dissection damage and stain with Toluidine Blue. The most straightforward interpretation is that decreased axon number will reflect death of motor neurons in the spinal cord. See Fig. 20.4 for an example of this analysis. This interpretation is strengthened if combined with spinal cord histology described above. Degenerating axon profiles may also be seen, which may indicate an axonopathy in the absence of evidence for dying cells in the ventral horn. An often-invoked complication to this analysis is that motor axons may branch and this would mask the loss in axon number. Motor axons typically branch only after they have entered the muscle. While a pathological state may cause them to branch at the level of the ventral root, we have not encountered any examples of this, although it is a formal possibility and a caveat to this analysis. Furthermore, if axon loss is seen,
368
Burgess, Cox, and Seburn
Fig. 20.4. An example of motor neuron loss reflected in ventral root axon number. The L4 ventral root stained for myelin is shown in cross-section in (A) a control mouse, (B) an nmd mouse (Ighmbp2 mutant), (C) an nmd mouse carrying a modifier locus from CAST, and (D) an nmd mouse with transgenic rescue of the Ighmbp2 gene driven in the nervous system. For details, see (39).
the worst-case scenario is that the number is an underestimate due to branching. Degenerating axons in the peripheral nerves can also be examined as described below, and electrophysiological estimates of motor unit number could be informative. Motor units are defined as a single motor neuron (axon) and the muscle fibers its terminals innervate. In motor neuron diseases, the number of motor units is anticipated to decrease as motor neurons die, but the size of motor units may increase with compensatory sprouting and reinnervation (e.g., (1)). Interpretation can be further confounded by factors such as a change in muscle fiber number or innervation of muscle fibers by multiple motor axons. Therefore, the best interpretation results from corroborating evidence from the spinal cord, nerve, synapse, and the muscle. 3.2. Peripheral Neuropathies
The femoral nerve provides an excellent system for evaluating peripheral neuropathy, provided there is hind limb involvement. In combination with counts of ventral roots, femoral axon counts can be used to distinguish peripheral neuropathy from motor neuron death. The nerve has a primarily motor branch that innervates the quadriceps, and a primarily sensory branch that becomes the saphenous nerve more distally (Fig. 20.5). Each branch can be easily dissected free (the animals can be transcardially perfused before, or the nerves can be fixed by immersion after dissection if care is taken to be sure they are extended at full-length when immersed). 1. Nerves should be plastic embedded and cross-sectioned as above for the ventral roots. 2. Axons can be counted from Toluidine Blue-stained sections. 3. The distribution of axon diameters, myelin thickness, and Gratios (inner/outer diameters, the inner being the axoplasm, the outer including the myelin) can also be determined.
Neuromuscular Disease Models and Analysis
369
Fig. 20.5. Femoral nerve dissection. The motor and sensory branches of the femoral nerve are exposed. The mouse is supine and the right hip is shown (forceps are retracting the abdominal wall, A, P, M, L are anterior, posterior, medial, and lateral respectively, H is hamstring muscles). Some adipose tissue has been removed for clarity. The motor branch of the femoral nerve innervates the quadriceps (Q). The sensory branch becomes the saphenous nerve, which runs adjacent to the saphenous vein (Saph) on the medial side of the thigh. Dissecting the nerve where the tick marks provides a reasonable length of nerve to work with. Note the sensory branch sometimes runs as two fascicles and both should be taken to get reproducible counts. The scale bar is 2 mm. (Color figure is available online).
This may be done most accurately by low magnification (4000–6000×) transmission electron microscopy. 4. The assessment of axon diameters may reveal general axonal atrophy, or a missing class, such as large diameter, fast motor axons. 5. The assessment of myelin layering and myelin thickness may reveal a demyelinating or hypomyelinating neuropathy. 6. The G-ratio may indicate abnormal reciprocal signaling between the axon and the myelinating Schwann cell or may highlight thin myelin or conversely, thin axons, since there is normally a rough correlation between axon diameter and myelin thickness.
370
Burgess, Cox, and Seburn
7. Other peripheral nerve pathologies such as onion bulbs (indicating rounds of demyelination/remyelination), Schmidt–Lanterman Incisures (indicating abnormal myelin packing), myelinating Schwann cells wrapping multiple axons, and bundles of regenerating axons can also be seen in these sections. For examples of these pathologies see Fig. 20.6 and ref. (2, 3).
Fig. 20.6. Peripheral neuropathy phenotypes. A cross-section of the motor branch of the femoral nerve in a control mouse (A) and a Gars Nmf249/+ model of Charcot-Marie-Tooth 2D (B) are shown. The insets highlight the varied axon diameters in a control nerve and the almost complete absence of large diameter axons in the mutant nerve. (C) The same mutation examined by transmission electron microscopy demonstrates degenerating axon profiles. Note the irregularities in the myelin are fixation artifact and not pathology, highlighting the need to always process control samples in parallel. (D) A myelinating Schwann cell that has ensheathed multiple axons, probably representing a failure in radial sorting during early postnatal development. Note, this is different from a Remak bundle of small sensory axons, in which a nonmyelinating Schwann cell enwraps a number of axons in a single basal lamina (see the bottom right corner of (C). (E) Nodes of Ranvier can be examined by light microscopy in teased nerve preparations. The nucleus of the Schwann cell (black arrow) is typically midway between the nodes (white arrows). (E) An example of hypomyelination in the ventral root of a Lama2 dy/dy mouse. Normally, 100% of the ventral root axons are myelinated. In this mutation, bundles of large but unmyelinated axons are evident.
3.2.1. Internodal Distance Assessment
In addition to axon loss/atrophy and defects in myelination, the internodal distance can also affect nerve conduction velocities. 1. To determine internodal distance, dissect a 1–2 cm segment of peripheral nerve such as the femoral or sciatic and fix as above (see Section 3.1.2). 2. Tease the nerve longitudinally to individual fibers using No. 5 forceps or a 30-G needle. Keeping the nerve immersed in a drop of PBS, tease the nerve directly on a microscope slide. 3. Coverslip the teased nerves and view using Nomarski-DIC optics.
Neuromuscular Disease Models and Analysis
371
4. Measure internodal distances and correlate them with axon diameters. 5. Again, there should be a rough correlation, with larger axons having longer internodal distances. This analysis requires software calibrated for digital image analysis to determine the distances ( Fig. 23.6 ). 3.2.2. Nerve Conduction Velocities
Nerve conduction velocities are used diagnostically to distinguish type 1 (demyelinating) and type 2 (axonal) neuropathies in humans. Type 1 neuropathies typically have pathologically reduced NCVs (below 30 m/s compared to normal values near 50 m/s in humans). Nerve conduction velocities can also be measured in the mouse as described below. However, significant decreases in axon diameter or internodal distance can also contribute to decreased NCV (3, 4). Furthermore, NCVs record the fastest (largest) axons present and may therefore miss axonal pathologies that do not cause a marked decrease in these neurons. Therefore, this functional measure should again be combined with an examination of axon diameters, myelination, and internodal distance to determine the underlying mechanism.
3.3. Myasthenias
Myasthenias are diseases of the NMJ. This synapse is highly accessible, highly stereotyped in its morphology, and easily visualized by light or electron microscopy. Evaluation of the morphology of the junctions is generally very well correlated with function, although electrophysiology may be required to fully assess synaptic transmission and to determine if defects are pre- or postsynaptic.
3.3.1. Staining NMJs for Light Microscopy
Neuromuscular junctions can be visualized by light microscopy following labeling of the presynaptic nerve terminal and the postsynaptic acetylcholine receptors. For best results, muscles should be prepared for longitudinal sections, and NMJs in an en face orientation can be imaged. In almost all muscles, the end plate band is near the middle of the muscle and this represents the region of interest. 1. Muscles should be lightly fixed in buffered 2% paraformaldehyde (2–4 h on ice) (see Note 3). 2. The muscles can be prepared for staining in a number of ways, but should be oriented for longitudinal sections. 3. Cutting thick, frozen sections using a cryostat (20–40 m thick sections) or using a vibratome to cut 50 m sections of unfrozen tissue (remove the tendons to facilitate sectioning) gives good results. Muscle fibers can also be teased directly on slides to obtain individual fibers.
372
Burgess, Cox, and Seburn
4. Samples are then stained using standard immunocytochemistry procedures. 5. Presynaptic antigens that work well include a cocktail of anti-neurofilament (for instance the monoclonal antibody 2H3) and anti-SV2 to fully label both the axon and the nerve terminal. 6. The postsynaptic receptors are brightly and specifically labeled using fluorescent conjugates of alphabungarotoxin. 7. Standard immuno staining techniques involving application of the cocktail of primary antibodies followed by washes, then application of a cocktail of the secondary antibody and ␣-bungarotoxin followed by washes, work well provide patience is used (for instance, primary antibodies should be applied overnight followed by several hours of washing the next day). 8. Standard antibody dilution buffers such as PBS with BSA or normal goat serum as blockering agents can be used, provided generous detergent (0.5–1% triton X-100) is also included. 9. Samples can be viewed on a standard fluorescence scope, but given the large size and three-dimensional nature of the samples, a confocal Z-series usually gives better results. 10. A number of defects can be readily observed, including partially innervated or completely denervated postsynaptic receptor sites, fragmented or shrunken postsynaptic receptors, atrophied axons or terminals, and swollen or dystrophic axons or terminals. More subtle defects include sprouting nerve terminals and multiple innervation or synaptic sites (single innervation is normally present after 2 weeks of age). Normal NMJs have a contiguous pretzellike morphology, the AChR staining has a zebra stripe appearance (the junctional folds), and the nerve terminals completely overlap the receptors. Examples are shown in Fig. 20.7. Caveats include minor differences in size and shape from muscle to muscle, and fixation artifacts that can eliminate staining, especially presynaptically. In general, defects in the presynaptic terminal, such as partial retraction, are reflected less-precise definition in the postsynaptic receptors. Additional analyses that can be informative include staining with anti-S100 to visualize Schwann cells (the terminal Schwann cells play an important role in guiding terminal sprouting and reinnervation (5, 6)) and histochemical stains to visualize acetylcholine esterase (7).
Neuromuscular Disease Models and Analysis
373
Fig. 20.7. Neuromuscular junctions, light microscopy. In a control neuromuscular junction (A), the nerve (stained green using anti-neurofilament plus anti-SV2, or using a Thy1-YFP transgenic strain) completely overlaps the postsynaptic AChRs (stained red using fluorescent -bungarotoxin conjugates). (B) An example of partial innervation, where an atrophied motor axon and terminal fails to completely cover the AChRs on the muscle. (C) An example of frank denervation, where a site of postsynaptic AChRs with no associated nerve (arrowhead) is observed near a site of partial innervation (double arrowhead). (B) and (C) are examples from the Gars Nmf249/+ mouse model of Charcot-Marie-Tooth 2D (3). Other NMJ pathologies are evident in an example from an unpublished spontaneous mutation. The axons and terminals have irregular diameters and varicosities (arrowheads) and postsynaptic sites are fragments (double arrowheads). Such changes are predictive of eventual denervation. These pathologies are also observed in very old mice (greater than 15 months), but are evident in this mutant by three months of age. (Color figure is available online).
3.3.2. Electron Microscopy of the NMJ
NMJs can also be visualized by transmission electron microscopy for a more detailed look at pre- and postsynaptic anatomies. 1. Muscles should be prepared for electron microscopy using standard techniques, including rapid fixation (preferably by perfusion) with glutaraldehyde-based fixatives. 2. The muscle should be dissected free and trimmed for crosssections at the point where the nerve enters the muscle. The end plate band is only a narrow region near the middle of the muscle (see Note 4).
374
Burgess, Cox, and Seburn
3. Samples should be postfixed, osmicated, and embedded in plastic, cross-sectioned, and mounted on EM grids using standard procedures (for example, (8)). The challenge to viewing sections is finding NMJs. They can be spotted by scanning the grids at 10–12 K magnification and concentrating on areas where axons, fat, or blood vessels are also present. NMJs are rarely found in areas where the muscle fibers tightly stacked. Detailed images can be obtained at 30–60 K magnification. Normal NMJs have a generally polarized nerve terminal with accumulations of 40–50 nm small clear vesicles near the presynaptic membrane and mitochondria located farther away (Fig. 20.8). In mice, the terminal Schwann cell capping the nerve terminal can be difficult to resolve. The postsynaptic membrane has a series of junctional folds invaginating into the muscle fiber. At the mouth (crest) of each fold, the membrane appears electron dense because of the accumulation of AChRs. The synaptic cleft is pronounced and contains a visible basal lamina.
Fig. 20.8. Neuromuscular junctions analyzed by transmission electron microscopy. (A) In wild-type mice, the motor nerve terminal (MN) is depressed into the muscle fiber surface. The terminal is polarized, with small clear vesicles near the presynaptic membrane and mitochondria in the more proximal portion of the terminal. The postsynaptic membrane has deep convolutions (junctional folds, JF) and the membrane near the tops of these folds is very electron dense because of the high density of acetylcholine receptors (arrowheads). (B) In some myasthenias where the nerve sprouts but remains in contact with the muscle, terminals with mitochondria and vesicles are observed in the absence of any postsynaptic specialization. Presumably these are sprouting terminals that have not established a functional connection. (C) Partial innervation of postsynaptic sites is evident as elaborate junctional folds in the muscle membrane with no overlying nerve terminal. In these examples, the interpretations were aided by light microscopy examination of other samples as described in Fig. 20.8 in parallel with electron microscopy. The mutation shown in (B, C) is an unpublished ENU-induced allele of agrin.
Pathological deviations include an absence of junctional folds, partial innervation (folds without an overlying nerve terminal), and vacuolated mitochondria. Assessing more subtle defects, such as changes in vesicle number, requires a statistical analysis on many junctions. 3.4. Muscular Dystrophies 3.4.1. Histological Analysis of Muscular Dystrophies
Muscle pathology can be accurately assessed by histology, focusing on similar hallmarks to those used in the diagnosis of human muscular dystrophies (Fig. 20.9). Muscle weight to body weight ratios can be used to indicate pathology in the muscle that is not
Neuromuscular Disease Models and Analysis
375
simply proportional to decreased body size. Mice with muscular dystrophy or atrophy will typically have lower body weights, but an even greater reduction in muscle weight:body weight. 1. Muscles can be fixed by perfusion or immersion in Bouin’s fixative, dehydrated, and embedded in paraffin (record body weights and muscle weights following dissection, prior to processing). 2. Cross-sections should be cut using a microtome and stained using H&E. All of these techniques are standard histological protocols. Care should be taken to collect sections from standardized regions of the muscle (such as the belly) to avoid differences in fiber number, fiber size, or composition that can vary along the longitudinal axis in limb muscles. 3. Samples should be evaluated for muscle fiber diameters (and the uniformity of fiber sizes), centrally located myonuclei (a sign of a regenerated muscle fiber), fibrosis, fatty infiltration, and atrophied muscle fibers. 4. Interpretation: muscular dystrophies can be distinguished from other neuromuscular conditions based on the loss of muscle fibers, fibrosis, and signs of degeneration/regeneration. Atrophy resulting from denervation may have a similar appearance, but the affected fibers are typically scattered throughout the muscle (reflecting the anatomy of motor units and the pattern of innervation by motor neurons), while dystrophies tend to affect most of the fibers in a particular region of the muscle. 3.4.2. Muscle Fiber Integrity, Evan’s Blue Staining
Many muscular dystrophies result in the loss of sarcolemmal integrity. This is true of dystrophies affecting the Dystrophin/Glycoprotein Complex (DGC), including defects in dystrophin, the sarcoglycans, and dystroglycan, or in dysferlin, which is involved in membrane repair. Other mutations, such as those in titin, do not result in a loss of muscle fiber integrity. Therefore, particularly for the mechanistic evaluation of new models, it is important to determine if membrane integrity is compromised. This is easily done using Evan’s blue, an Azo dye that binds albumin and is normally excluded from healthy cells, but infiltrates into the cytoplasm of compromised cells (Fig. 20.9). 1. Evan’s blue dye should be dissolved in sterile saline at 10 mg/ml. 2. This solution can then be injected IP using 0.1 ml per 10 g of body weight (100 mg/kg). 3. Within a few hours of injection, exposed skin, such as the feet, tail, and ears, should have a pronounced blue tinge. 4. After 12–18 h (overnight), dissect muscles of interest and view.
376
Burgess, Cox, and Seburn
Fig. 20.9. Dystrophy phenotypes. (A) In histological staining of wild-type muscle (H&E), the muscle fibers are closely packed, regular in size, and the nuclei are near the cortex of the muscle fiber. (B) In a dystrophic muscle (Lama2dy/dy is shown), the fibers are variable in diameter, some are atrophied, fibrotic cells are replacing muscle fibers, and nuclei in the center of fibers indicate regenerated fibers. (C, D) Mice lacking dysferlin, which is involved in membrane repair, have a progressive dystrophy. Muscle is histologically normal (C) until approximately 8 months of age, by 14 months (D), the muscle is severely dystrophic, with a great deal of fatty infiltration. (E) The rmd mouse mutation also has a severe dystrophy with fatty infiltration into the muscle, but the phenotype is much more severe in the hind limbs (shown) than the forelimbs. (F) Antibody staining for myosin isoforms (fast myosin is shown) can determine if certain fiber types are selectively sensitive. The myd mouse is shown, in which both fast and slow fibers show central nuclei and signs of dystrophy, indicated both fiber types are affected. (G, H) Fiber-type staining can also identify grouping, as seen with slow myosin stains of control (G) and GarsNmf249/+ muscle (H). Such grouping is indicative of denervation and reinnervation and not a dystrophy intrinsic to the muscle. (I, J) Muscle fiber integrity can be assessed by Evan’s blue dye infiltration. In control muscle (I), Evan’s blue administered intraperitoneally is excluded from the muscle fibers but stains the membranes and connective tissue. In dystrophic muscles in which the sarcolemmal membrane integrity is compromised (mdx is shown), the dye stains the entire fiber (J). (Color figure is available online).
5. All tissue will have a bluish cast, but dystrophic muscle will have marked streaks of blue due to the compromised fibers. 6. This can be most readily assessed by embedding the tissue and cutting cryostat cross-sections. 7. Under fluorescence (rhodamine filters), the Evan’s blue fluoresces red. Healthy muscle will have positively labeled membranes, as well as blood vessels, but muscles with compromised fiber integrity will contain fibers in which the entire cross-sectioned cytoplasm is strongly fluorescent. This technique is simple and very sensitive. Intraperitoneal injection is straightforward and as effective as intravenous injection. Positive muscle fibers can also be the result of necrosis or injury in the muscle, but this usually results in a few, widely scattered
Neuromuscular Disease Models and Analysis
377
fibers instead of a large number of fibers grouped in a specific region of the muscle. 3.4.3. Caveats
As with all neuromuscular diseases, there can be tremendous variation muscle to muscle, and even within a muscle (this is true for neuropathies as well, see, for example, (9)). For instance, in some dystrophy models the diaphragm is severely affected while in others, it is spared. Therefore, it is very important to evaluate the same muscle in each mouse and to evaluate more than one muscle. A picture of the anatomy of the lower hind limb of the mouse is shown in Fig. 20.10. The underlying basis of these differences in not clear and does not seem to be as obvious as differences in fiber type composition, activity levels, or force generated. Within a model, the pattern of pathology is usually fairly reproducible, but there can again be animal-to-animal variability that complicates quantification of results.
Fig. 20.10. The muscles of the lower hind limb in cross-section. In this image, anterior is down and medial is left. Abbreviations are as follows: MG, medial gastrocnemius; LG, lateral gastrocnemius; Plant, plantaris; PN, plantar nerve; Sol, soleus; Fib, fibula; EDL, extensor digitorum longii; TA, tibialis anterior; Tib, tibia. The mouse muscles are predominantly fast muscle fibers, but the soleus is valuable for its high percentage of slow fibers. Note, the darker mass on the posterior portion of the leg is a lymph node that provides a convenient landmark when sectioning to establish that reproducible sections are examined in the proximal/distal axis. Also, the peripheral muscles in the section are the hamstrings, which insert along the tibia in the lower leg in the mouse. (Color figure is available online).
Other techniques such as serum levels of muscle creatine kinase can also be used to assess dystrophy, but this requires great care in mouse handling. Creatine kinase is released from damaged
378
Burgess, Cox, and Seburn
muscle fibers. If the mouse struggles significantly while the sample is collected (more than a few seconds), values will be artificially high and variable. In addition, hemolysis in the sample will also reduce accuracy. Therefore, while this measurement is outwardly straightforward and valuable, obtaining consistent results can be challenging. 3.5. Techniques Generally Applicable to Neuromuscular Disease Models 3.5.1. Gross Motor Performance
3.5.2. Grip Strength
The disease models described above, motor neuron diseases, neuropathies, myasthenias, and muscular dystrophies, can all benefit from an analysis of gross motor performance. The rotarod is often used to assess neuromuscular function and other movement disorders and has been successfully used to describe defects in many mouse models (e.g., (10–12)). The widespread use of the device is due in part to the apparent simplicity of the testing procedures. However, recent work emphasizes that, as with any behavioral test, there are a number of potential confounding factors (13, 14). In our experience data collection is more efficient using gait analysis (treadmill). In addition raw video data is often useful on its own and the extensive and varied measurements that are possible offer significant flexibility. Grip strength measures also provide a more direct assessment of muscle force generation that can be combined with electrophysiological techniques described below, and are only minimally confounded by other parameters affecting motor performance such as coordination or balance. Therefore, our preferred methods are grip strength and gait analysis. Measure of grip strength gives a good indicator of muscle strength and is analogous to grip strength measures in humans, e.g. (15), which reveal weakness as a major presenting symptom of neuromuscular dysfunction. To measure grip strength, the mouse is prompted to grab a bar connected to a force transducer with either its hind or its fore paws (or less often all paws using a grid). Once the mouse achieves a grip the tester typically pulls the mouse horizontally away from the bar until the animal is no longer able to maintain its grip. The peak force registered by the transducer is recorded. This approach, recommended by equipment suppliers, requires careful attention to the rate and direction of force applied by the tester to “break” the animals grip. If this approach is used, reliability should be confirmed for each individual performing the test and between testers if multiple individuals are collecting data. However, to improve reproducibility for forelimb grip strength we recommend orienting the force transducer vertically (Columbus Instruments, Columbus, OH, USA) and modifying the test procedure as follows. A weight (100 g) is attached to the base of the mouse’s tail using a small plastic clip. The mouse is held by the scruff of the neck with the tail and weight in the
Neuromuscular Disease Models and Analysis
379
tester’s other palm. The mouse is advanced toward the bar until it instinctively grasps the bar with both paws. The mouse is slowly lowered to a vertical position and both animal and weight are released. The transducer records the peak force generated by the animal just prior to grip loss. This method insures consistent force and acceleration away from the grip bar and is not any more traumatic for the mouse. Typically measures from three repeat trials are averaged to represent the value for each animal. This approach can be used to track changes in longitudinal studies and has been used effectively for most standard inbred strains (http://phenome.jax.org). 3.5.3. Gait Analysis
Analysis of gait has a long history, and because it is also routinely used in humans, establishing clinical relevance for murine disease models is facilitated. The simplest method for quantification of gait in mice is to paint the feet of the mice and to motivate them to walk across a piece of paper. Manual measurements of footprints can then be made to derive stride length and stance width. This method is particularly useful for quantification in mice with overt, observable movement defects, but may also be sufficient for more subtle phenotypes (e.g., (16)). Practically, the greatest difficulty with this method is in motivating the mice to walk, and investigators have addressed this with a variety of strategies. For example, placing the mice in a lighted “corridor” facing toward a darkened space will both restrict wandering and improve motivation, as mice will seek to “escape” to the dark enclosure. Two additional considerations for this method are that only a limited number of parameters can be measured and that the speed of locomotion cannot be controlled. Thus, common time domain parameters that divide a stride (step) into component phases of stance and swing are not possible and measurements that are derived need to consider locomotory speed in the interpretation. We expanded the utility of the basic footprint analysis by employing video recordings of mice walking on a treadmill (Fig. 20.11). The mice are placed in an enclosure on a treadmill with a clear plastic tread (Columbus Instruments, Columbus, OH). The ventral surface of the mouse is reflected in a mirror placed at 45◦ under the tread and is recorded by an adjacent digital video camera (Basler, Inc). Typically a video clip of ∼10 s provides sufficient data for valid measurement. Interactive analysis software (Clever Sys Inc, Reston VA) is used to track body position and paw placement of each paw during locomotion at a fixed speed. The software derives the standard time domain gait parameters for each paw (e.g., stride and stance, swing phase times) as well as a variety of additional measures (body angle, foot placement angles, inter-limb phase ratios, etc.). The variety of measures that can be derived with such a system allow detailed description of gait in many different models and
380
Burgess, Cox, and Seburn
Fig. 20.11. Gait analysis device. (A) The gait analysis device is shown. The mouse is place on the treadmill in the green box. The mouse would be facing left and the loop at right in the chamber bumps the mouse’s tail if they lag on the treadmill. The white box in front contains the digital camera that videos the mouse using an angled mirror. (B) The automated data analysis is trained to recognize the mouse’s paws in the video and gait parameters are calculated, as well as body axis, toe spread, and other data.
offers the significant advantage of comparing animals at the same speed of locomotion and may be more sensitive to some subtle neuromuscular changes prior to overt movement deficits (17). However, the treadmill represents a novel context for the mice,
Neuromuscular Disease Models and Analysis
381
and we have noted changes in treadmill gait with repeated trials and age both of which can complicate characterization of progressive changes. Also, different mouse strains vary in their willingness to walk consistently on the treadmill, and compliance may be further reduced with repeated exposures (for details see (18)). These factors need to be considered carefully for each model in order to optimize experimental design. 3.5.4. Electrophysiology
We will describe basic electrophysiological analysis of neuromuscular function, similar to evaluation using electromyography by neurologists. More specialized techniques such as two electrode voltage clamp for monitoring synaptic transmission may also be useful, but are beyond the basic nature of this chapter. We describe briefly techniques for assessment of muscle contractile function including electromyography and nerve conduction velocity. These techniques provide measures of muscle force output and fatigability and insight into contractile dynamics related to energy supply, calcium handling, and excitation–contraction coupling. In addition, motor unit number can be estimated and significant deficits in synaptic function may be detectable.
3.5.5. Muscle Contractile Function
Contractile properties in rodents can be measured either in vitro in a dissected muscle or in vivo in an intact preparation with an anesthetized animal (e.g., (19–20)). Measurements made under isometric conditions are perhaps most common and use the most straightforward setup. The addition of servomotors for dynamic control of muscle length allows simulation of dynamic conditions (eccentric, isotonic, etc.) that may be modified by disease or other processes (22, 23). For in vitro studies, the muscle is anchored by ligating the tendon (origin) to a support, for in vivo studies, the bone (femur) is clamped to prevent movement. The other tendon (insertion) is then coupled to a force transducer. In both cases, a recording electrode is also placed in contact with the muscle to record the compound action potential, and a stimulating electrode is used to stimulate the nerve or the muscle, as described below. It should be noted that both in vitro and in vivo preparations present considerable technical challenges and are best developed/established in consultation with an experienced laboratory. In general, an in vitro preparation is somewhat less demanding because surgical preparation is less demanding and extended life support/maintenance of the animal during experimentation is not required. Muscle force can be elicited by nerve stimulation to test both muscle function and the integrity of the nerve–muscle connection or by direct muscle stimulation to evaluate only muscle contractile independent of the synapse. The latter measure reflects the total force that the muscle is able to generate, if it is significantly
382
Burgess, Cox, and Seburn
greater than that obtained by nerve stimulation, it implies denervation, or a failure in conduction or synaptic transmission in the nerves. Two general types of stimuli are typically used, a single brief (100–200 s) stimulus of intensity sufficient to activate all functional connections, referred to as maximal twitch force (Pt), and trains of stimuli delivered at a frequency (100–200 pps) which induces tetanic fusion of individual contractions to produce maximum tetanic force (P0). 3.5.6. Contractile Measures
Measurement of single twitches provides insight into contraction/relaxation dynamics of the muscle. For example, reduced AChE or altered Ca2+ handling in the muscle causes a slower relaxation phase, whereas shifts toward “faster” ATPase isoforms will reduce time to reach peak twitch force (Fig. 20.12).
Fig. 20.12. Muscle twitch force. The force generated by a control muscle in response to a single 200 s stimulus (black trace) compared to two mutants showing exaggerated force (red) or slower relaxation time (orange) is shown. This is an unpublished mutation that we interpret as having decreased AChE at the synapse, possible as a secondary compensation for impaired presynaptic neurotransmitter release. (Color figure is available online).
Stimulus trains of varying lengths (0.5–1.2 s) and frequencies of 10–200 pps can provide additional information. As stimulus frequency is increased, there is greater tetanic fusion and higher peak forces are produced. A force–frequency curve (F–F) can also be generated to evaluate factors that affect muscle contractile speed (time-to-peak, half-relaxation). If, for example, disease processes lead to slower contraction/relaxation time, more tetanization occurs at lower frequencies and the shape of the F–F curve will be shifted compared to normal muscle. A variety of protocols exist for measurement of muscle fatigue in mice. In general a series of stimulus trains are repeated at a set frequency for several minutes. The protocol selected depends on the muscle being tested and can vary depending on the model and the experimental objective. Different laboratories have established unique protocols that can be implemented (24–28). The basis for the difference in protocols is not always evident, different stim-
Neuromuscular Disease Models and Analysis
383
ulus paradigms may have been selected on a theoretical basis or determined empirically for the purposes of a given laboratory. Long stimulus trains will also induce “tetanic fade,” observed as a decrease in force during a single contraction. Fade is measured as the ratio of final force to initial peak force. Diseased muscle/nerve may show a lower ratio. Accompanying measurement of EMG can reveal if fade is due to failure of the muscle or the nerve. In purely muscle defects (muscle dystrophy) force will drop without any change in EMG whereas reduced force accompanied by reduced EMG suggests a defect in transmission (NMJ) (e.g., (3)) or excitation–contraction coupling. 3.5.7. Motor Unit Number Estimates (MUNE)
Delivering brief nerve stimuli with gradually increasing amplitudes will evoke an incremental increase in Pt as additional motor units are recruited. If step-wise increases in twitch force or electromyogram amplitudes are recorded the distinct increments can be counted to provide an estimate of motor unit number or the number functional motor axons innervating a given muscle (Fig. 20.13). There are a variety of MUNE techniques (29) but no universally accepted standard has emerged. Nonetheless, the method, originally used in humans (30), has been used successfully in other species including mice (e.g., (1)) and recently developed methods can facilitate implementation (31).
3.5.8. Nerve Conduction Velocities (NCV)
In contractile experiments the time from the stimulation of the nerve to the CMAP recorded in muscle provides an estimate of NCV. The length of the nerve from the stimulating electrode to the muscle can simply be measured and divided by the time. However, the time recorded in this way includes the delay for synaptic transmission, which may be increased in models with synaptic defects. If this is a concern or if the only parameter desired is NCV, then the measurement can be obtained non-invasively with a relatively simple setup (e.g., (3)). Using the sciatic nerve, NCV can be calculated by measuring the latency of compound motor action potentials recorded in the muscle of a rear paw. Action potentials are produced by subcutaneous stimulation at two separate sites: proximal stimulation at the sciatic notch and distally at the ankle. NCV is then calculated by using the two latencies and conduction distance. Decreases in nerve conduction velocity most often reflect defects in myelination, but may also be the result of changes in internodal distance, decreased axon diameters, or altered excitability.
384
Burgess, Cox, and Seburn
4. Conclusion The pathophysiology of neuromuscular diseases is highly integrated, reflecting the extensive reciprocal signaling and functional interdependency of the peripheral nervous system and the skeletal muscles. As a result, changes in the nerves result in changes in the muscles and vice versa, making it a challenge to distinguish primary and secondary effects. This is particularly true when new models are being examined and it is unknown where the causative gene normally functions. As a result, the functions of genes and the mutant phenotypes must be examined comprehensively, and the results must be taken as a whole in order to understand how the final phenotype arises. In this chapter, we have given examples of mouse models used in neuromuscular disease research,
Fig. 20.13. Motor unit estimation. Stimuli of gradually increasing intensity are applied to the nerve and the force generated by the muscle is recorded. The number of steps of force produced approximately reflects the number of motor units in the nerve.
Neuromuscular Disease Models and Analysis
385
and provided a battery of analyses that can be applied to these models to understand the underlying disease mechanisms. While many of these diseases are now well understood, many remain equally mysterious, and continued efforts to understand both the genetics and physiology are needed to eventually have a complete understanding of the neuromuscular system.
5. Notes 1. Purple condensed cells in H&E can be “Dark Neuron Artifact” caused by deforming the tissue before it was well fixed (32–34). 2. The dorsal root has an enlargement, the dorsal root ganglion containing the sensory cell bodies near the point of bifurcation. 3. The paraformaldehyde should be of high quality (such as that supplied by Electron Microscopy Services) and should be prepared fresh each day. Over fixation or low-grade fixatives will dramatically reduce or eliminate the antigenicity of presynaptic proteins. 4. It is better to lose some NMJs than to waste a lot of effort sectioning regions of the muscle where there are no synapses. 5. There are a numberof other useful online resources and reference texts that further explain the genetics, physiology, and clinical features of neuromuscular diseases in both humans and mice. The database “Online Mendelian Inheritance in Man (OMIM)” is particularly useful as a source of genetic information related to all forms of heritable human diseases (http://www.ncbi.nlm.nih.gov/sites/entrez?db=OMIM& itool=toolbar). For a resource of mouse genetics, including mapping data, alleles, molecular characterization, and expression information, the Mouse Genome Informatics web site provides a comprehensive and current database (http://www.informatics.jax.org/). Other online resources related to specific human diseases also exist, such as the Inherited Peripheral Neuropathies website, providing genetic and molecular information pertaining to inherited diseases of the peripheral nervous system (http://www.molgen.ua.ac.be/CMTMutations/). Useful reference texts include Myology (A. G. Engel, C FranziniArmstrong, Eds.), Peripheral Neuropathy (P. J. Dyck and P. K. Thomas, Eds.), Pathology of Peripheral Nerves (J. M. Schroder, Ed.), and Neuromuscular Disorders: Clinical and Molecular Genetics (A. E. H. Emery, Ed.). See ref. (35–38).
386
Burgess, Cox, and Seburn
References 1. Hegedus, J., Putman, C. T. and Gordon, T. (2007) Time course of preferential motor unit loss in the SOD1 G93A mouse model of amyotrophic lateral sclerosis. Neurobiol Dis 28, 154–164. 2. Grohmann, K., Schuelke, M., Diers, A., Hoffmann, K., Lucke, B., Adams, C., et al. (2001) Mutations in the gene encoding immunoglobulin mu-binding protein 2 cause spinal muscular atrophy with respiratory distress type 1. Nat Genet 29, 75–77. 3. Seburn, K. L., Nangle, L. A., Cox, G. A., Schimmel, P. and Burgess, R. W. (2006) An active dominant mutation of glycyl-tRNA synthetase causes neuropathy in a CharcotMarie-Tooth 2D mouse model. Neuron 51, 715–726. 4. Court, F. A., Sherman, D. L., Pratt, T., Garry, E. M., Ribchester, R. R., Cottrell, D. F., et al. (2004) Restricted growth of Schwann cells lacking Cajal bands slows conduction in myelinated nerves. Nature 431, 191–195. 5. Son, Y. J. and Thompson, W. J. (1995) Nerve sprouting in muscle is induced and guided by processes extended by Schwann cells. Neuron 14, 133–141. 6. Son, Y. J. and Thompson, W. J. (1995) Schwann cell processes guide regeneration of peripheral axons. Neuron 14, 125–132. 7. Enomoto, H., Araki, T., Jackman, A., Heuckeroth, R. O., Snider, W. D., Johnson, E. M., Jr., et al. (1998) GFR alpha1-deficient mice have deficits in the enteric nervous system and kidneys. Neuron 21, 317–324. 8. Patton, B. L., Cunningham, J. M., Thyboll, J., Kortesmaa, J., Westerblad, H., Edstrom, L., et al. (2001) Properly formed but improperly localized synaptic specializations in the absence of laminin alpha4. Nat Neurosci 4, 597–604. 9. Pun, S., Santos, A. F., Saxena, S., Xu, L. and Caroni, P. (2006) Selective vulnerability and pruning of phasic motoneuron axons in motoneuron disease alleviated by CNTF. Nat Neurosci 9, 408–419. 10. Morgan, D., Munireddy, S., Alamed, J., Deleon, J., Diamond, D. M., Bickford, P., et al. (2008) Apparent behavioral benefits of tau overexpression in P301L tau transgenic mice. J Alzheimers Dis 15, 605–614. 11. Stam, N. C., Nithianantharajah, J., Howard, M. L., Atkin, J. D., Cheema, S. S. and Hannan, A. J. (2008) Sex-specific behavioural effects of environmental enrichment in a transgenic mouse model of amyotrophic lat-
eral sclerosis. Eur J Neurosci 28, 717–723. 12. Turgeman, T., Hagai, Y., Huebner, K., Jassal, D. S., Anderson, J. E., Genin, O., et al. (2008) Prevention of muscle fibrosis and improvement in muscle performance in the mdx mouse by halofuginone. Neuromuscul Disord 18, 857–868. 13. Bohlen, M., Cameron, A., Metten, P., Crabbe, J. C. and Wahlsten, D. (2009) Calibration of rotational acceleration for the rotarod test of rodent motor coordination. J Neurosci Methods 178, 10–14. 14. Rustay, N. R., Wahlsten, D. and Crabbe, J. C. (2003) Influence of task parameters on rotarod performance and sensitivity to ethanol in mice. Behav Brain Res 141, 237–249. 15. Bautmans, I. and Mets, T. (2005) A fatigue resistance test for elderly persons based on grip strength: reliability and comparison with healthy young subjects. Aging Clin Exp Res 17, 217–222. 16. Fernagut, P. O., Diguet, E., Labattu, B. and Tison, F. (2002) A simple method to measure stride length as an index of nigrostriatal dysfunction in mice. J Neurosci Methods 113, 123–130. 17. Wooley, C. M., Sher, R. B., Kale, A., Frankel, W. N., Cox, G. A. and Seburn, K. L. (2005) Gait analysis detects early changes in transgenic SOD1(G93A) mice. Muscle Nerve 32, 43–50. 18. Wooley, C. M., Xing, S., Burgess, R. W., Cox, G. A. and Seburn, K. L. (2009) Age, experience and genetic background influence treadmill walking in mice. Physiol Behav 96, 350–361. 19. Brooks, S. V., Faulkner, J. A. and McCubbrey, D. A. (1990) Power outputs of slow and fast skeletal muscles of mice. J Appl Physiol 68, 1282–1285. 20. Seburn, K. L. and Gardiner, P. (1995) Adaptations of rat lateral gastrocnemius motor units in response to voluntary running. J Appl Physiol 78, 1673–1678. 21. Lynch, G. S., Hinkle, R. T., Chamberlain, J. S., Brooks, S. V. and Faulkner, J. A. (2001) Force and power output of fast and slow skeletal muscles from mdx mice 6-28 months old. J Physiol 535, 591–600. 22. Brooks, S. V. and Faulkner, J. A. (1988) Contractile properties of skeletal muscles from young, adult and aged mice. J Physiol 404, 71–82. 23. Grange, R. W., Gainer, T. G., Marschner, K. M., Talmadge, R. J. and Stull, J. T. (2002) Fast-twitch skeletal muscles of dys-
Neuromuscular Disease Models and Analysis
24.
25.
26.
27.
28.
29.
30.
31.
32.
33. 34.
35.
trophic mouse pups are resistant to injury from acute mechanical stress. Am J Physiol Cell Physiol 283, C1090–1101. Belluardo, N., Westerblad, H., Mudo, G., Casabona, A., Bruton, J., Caniglia, G., et al. (2001) Neuromuscular junction disassembly and muscle fatigue in mice lacking neurotrophin-4. Mol Cell Neurosci 18, 56–67. Danieli-Betto, D., Esposito, A., Germinario, E., Sandona, D., Martinello, T., Jakubiec-Puka, A., et al. (2005) Deficiency of alpha-sarcoglycan differently affects fastand slow-twitch skeletal muscles. Am J Physiol Regul Integr Comp Physiol 289, R1328–1337. Grange, R. W., Meeson, A., Chin, E., Lau, K. S., Stull, J. T., Shelton, J. M., et al. (2001) Functional and molecular adaptations in skeletal muscle of myoglobinmutant mice. Am J Physiol Cell Physiol 281, C1487–1494. Parry, D. J. and Desypris, G. (1985) Fatiguability and oxidative capacity of forelimb and hind limb muscles of dystrophic mice. Exp Neurol 87, 358–368. Swallow, J. G., Garland, T., Jr., Carter, P. A., Zhan, W. Z. and Sieck, G. C. (1998) Effects of voluntary activity and genetic selection on aerobic capacity in house mice (Mus domesticus). J Appl Physiol 84, 69–76. Shefner, J. M. (2001) Motor unit number estimation in human neurological diseases and animal models. Clin Neurophysiol 112, 955–964. McComas, A. J., Fawcett, P. R., Campbell, M. J. and Sica, R. E. (1971) Electrophysiological estimation of the number of motor units within a human muscle. J Neurol Neurosurg Psychiatry 34, 121–131. Major, L. A., Hegedus, J., Weber, D. J., Gordon, T. and Jones, K. E. (2007) Method for counting motor units in mice and validation using a mathematical model. J Neurophysiol 97, 1846–1856. Ebels, E. J. (1975) Dark neurons: a significant artifact: the influence of the maturational state of neurons on the occurrence of the phenomenon. Acta Neuropathol 33, 271–273. Garman, R. H. (1990) Artifacts in routinely immersion fixed nervous tissue. Toxicol Pathol 18, 149–153. Jortner, B. S. (2006) The return of the dark neuron. A histological artifact complicating contemporary neurotoxicologic evaluation. Neurotoxicology 27, 628–634. Schr`eoder, J. M. (2001) Pathology of peripheral nerves: an atlas of structural and
36. 37. 38.
39.
40.
41.
42.
43.
44.
45.
46.
387
molecular pathological changes, Springer, Berlin/New York. Emery, A. E. H. (1998) Neuromuscular disorders: clinical and molecular genetics, John Wiley, Chichester/New York. Dyck, P. J. and Thomas, P. K. (2005) Peripheral neuropathy, 4th ed. Elsevier Saunders, Philadelphia, PA. Engel, A. and Franzini-Armstrong, C. (2004) Myology: basic and clinical, 3rd ed. McGraw-Hill, Medical Pub. Division,INTbreak; New York. Maddatu, T. P., Garvey, S. M., Schroeder, D. G., Hampton, T. G. and Cox, G. A. (2004) Transgenic rescue of neurogenic atrophy in the nmd mouse reveals a role for Ighmbp2 in dilated cardiomyopathy. Hum Mol Genet 13, 1105–1115. Hershkovitz, E., Rozin, I., Limony, Y., Golan, H., Hadad, N., Gorodischer, R., et al. (2007) Hypoparathyroidism, retardation, and dysmorphism syndrome: impaired early growth and increased susceptibility to severe infections due to hyposplenism and impaired polymorphonuclear cell functions. Pediatr Res 62, 505–509. Noakes, P. G., Miner, J. H., Gautam, M., Cunningham, J. M., Sanes, J. R. and Merlie, J. P. (1995) The renal glomerulus of mice lacking s-laminin/laminin b2: nephrosis despite molecular compensation by laminin b1. Nat Genet 10, 400–406. Zhai, R. G., Cao, Y., Hiesinger, P. R., Zhou, Y., Mehta, S. Q., Schulze, K. L., et al. (2006) Drosophila NMNAT maintains neural integrity independent of its NAD synthesis activity. PLoS Biol 4, e416. Zhai, R. G., Zhang, F., Hiesinger, P. R., Cao, Y., Haueter, C. M. and Bellen, H. J. (2008) NAD synthase NMNAT acts as a chaperone to protect against neurodegeneration. Nature 452, 887–891. Press, C. and Milbrandt, J. (2008) Nmnat delays axonal degeneration caused by mitochondrial and oxidative stress. J Neurosci 28, 4861–4871. Maddatu, T. P., Garvey, S. M., Schroeder, D. G., Zhang, W., Kim, S. Y., et al. (2005) Dilated cardiomyopathy in the nmd mouse: transgenic rescue and QTLs that improve cardiac function and survival. Hum Mol Genet 14, 3179–3189. Grohmann, K., Schuelke, M., Diers, A., Hoffmann, K., Lucke, B., Adams, C., et al. (2001) Mutations in the gene encoding immunoglobulin -binding protein 2 cause spinal muscular atrophy with respiratory distress type 1. Nat Genet 29, 75–77.
388
Burgess, Cox, and Seburn
47. Cox, G. A., Mahaffey, C. L. and Frankel, W. N. (1998) Identification of the mouse neuromuscular degeneration gene and mapping of a second site suppressor allele. Neuron 21, 1327–1337. 48. Monani, U. R., Pastore, M. T., Gavrilina, T. O., Jablonka, S., Le, T. T., Andreassi, C., et al. (2003) A transgene carrying an A2G missense mutation in the SMN gene modulates phenotypic severity in mice with severe (type I) spinal muscular atrophy. J Cell Biol 160, 41–52. 49. Le, T. T., Pham, L. T., Butchbach, M. E., Zhang, H. L., Monani, U. R., Coovert, D. D., et al. (2005) SMN Delta 7, the major product of the centromeric survival motor neuron (SMN2) gene, extends survival in mice with spinal muscular atrophy and associates with full-length SMN. Hum Mol Genet 14, 845–857. 50. Monani, U. R., Sendtner, M., Coovert, D. D., Parsons, D. W., Andreassi, C., Le, T. T., et al. (2000) The human centromeric survival motor neuron gene (SMN2) rescues embryonic lethality in Smn(-/-) mice and results in a mouse with spinal muscular atrophy. Hum Mol Genet 9, 333–339. 51. Schrank, B., Gotz, R., Gunnersen, J. M., Ure, J. M., Toyka, K. V., Smith, A. G., et al. (1997) Inactivation of the survival motor neuron gene, a candidate gene for human spinal muscular atrophy, leads to massive cell death in early mouse embryos. Proc Natl Acad Sci U S A 94, 9920–9925. 52. Martin, N., Jaubert, J., Gounon, P., Salido, E., Haase, G., Szatanik, M., et al. (2002) A missense mutation in Tbce causes progressive motor neuronopathy in mice. Nat Genet 32, 443–447. 53. Bommel, H., Xie, G., Rossoll, W., Wiese, S., Jablonka, S., Boehm, T., et al. (2002) Missense mutation in the tubulin-specific chaperone E (Tbce) gene in the mouse mutant progressive motor neuronopathy, a model of human motoneuron disease. J Cell Biol 159, 563–569. 54. Puls, I., Jonnakuty, C., LaMonte, B. H., Holzbaur, E. L., Tokito, M., Mann, E., et al. (2003) Mutant dynactin in motor neuron disease. Nat Genet 33, 455–456. 55. Chevalier-Larsen, E. S., Wallace, K. E., Pennise, C. R. and Holzbaur, E. L. (2008) Lysosomal proliferation and distal degeneration in motor neurons expressing the G59S mutation in the p150Glued subunit of dynactin. Hum Mol Genet 17, 1946–1955. 56. Lai, C., Lin, X., Chandran, J., Shim, H., Yang, W. J. and Cai, H. (2007) The G59S mutation in p150(glued) causes dysfunc-
57.
58.
59.
60.
61.
62.
63.
64.
65.
66.
tion of dynactin in mice. J Neurosci 27, 13982–13990. LaMonte, B. H., Wallace, K. E., Holloway, B. A., Shelly, S. S., Ascano, J., Tokito, M., et al. (2002) Disruption of dynein/dynactin inhibits axonal transport in motor neurons causing late-onset progressive degeneration. Neuron 34, 715–727. Ripps, M. E., Huntley, G. W., Hof, P. R., Morrison, J. H. and Gordon, J. W. (1995) Transgenic mice expressing an altered murine superoxide dismutase gene provide an animal model of amyotrophic lateral sclerosis. Proc Natl Acad Sci U S A 92, 689–693. Kunst, C. B., Messer, L., Gordon, J., Haines, J. and Patterson, D. (2000) Genetic mapping of a mouse modifier gene that can prevent ALS onset. Genomics 70, 181–189. Bruijn, L. I., Becher, M. W., Lee, M. K., Anderson, K. L., Jenkins, N. A., Copeland, N. G., et al. (1997) ALS-linked SOD1 mutant G85R mediates damage to astrocytes and promotes rapidly progressive disease with SOD1-containing inclusions. Neuron 18, 327–338. Wang, J., Xu, G., Gonzales, V., Coonfield, M., Fromholt, D., Copeland, N. G., et al. (2002) Fibrillar inclusions and motor neuron degeneration in transgenic mice expressing superoxide dismutase 1 with a disrupted copper-binding site. Neurobiol Dis 10, 128–138. Wong, P. C., Pardo, C. A., Borchelt, D. R., Lee, M. K., Copeland, N. G., Jenkins, N. A., et al. (1995) An adverse property of a familial ALS-linked SOD1 mutation causes motor neuron disease characterized by vacuolar degeneration of mitochondria. Neuron 14, 1105–1116. Gurney, M. E., Pu, H., Chiu, A. Y., Dal Canto, M. C., Polchow, C. Y., Alexander, D. D., et al. (1994) Motor neuron degeneration in mice that express a human Cu,Zn superoxide dismutase mutation. Science 264, 1772–1775. Heiman-Patterson, T. D., Deitch, J. S., Blankenhorn, E. P., Erwin, K. L., Perreault, M. J., Alexander, B. K., et al. (2005) Background and gender effects on survival in the TgN(SOD1-G93A)1Gur mouse model of ALS. J Neurol Sci 236, 1–7. Messer, A. and Flaherty, L. (1986) Autosomal dominance in a late-onset motor neuron disease in the mouse. J Neurogenet 3, 345–355. Bronson, R. T., Lake, B. D., Cook, S., Taylor, S. and Davisson, M. T. (1993) Motor neuron degeneration of mice is a model of neuronal
Neuromuscular Disease Models and Analysis
67.
68.
69.
70.
71.
72.
73.
74.
75.
76.
77.
ceroid lipofuscinosis (Batten’s disease). Ann Neurol 33, 381–385. Ranta, S., Zhang, Y., Ross, B., Lonka, L., Takkunen, E., Messer, A., et al. (1999) The neuronal ceroid lipofuscinoses in human EPMR and mnd mutant mice are associated with mutations in CLN8. Nat Genet 23, 233–236. Hafezparast, M., Klocke, R., Ruhrberg, C., Marquardt, A., Ahmad-Annuar, A., Bowen, S., et al. (2003) Mutations in dynein link motor neuron degeneration to defects in retrograde transport. Science 300, 808–812. Chen, X. J., Levedakou, E. N., Millen, K. J., Wollmann, R. L., Soliven, B. and Popko, B. (2007) Proprioceptive sensory neuropathy in mice with a mutation in the cytoplasmic Dynein heavy chain 1 gene. J Neurosci 27, 14515–14524. Ilieva, H. S., Yamanaka, K., Malkmus, S., Kakinohana, O., Yaksh, T., Marsala, M., et al. (2008) Mutant dynein (Loa) triggers proprioceptive axon loss that extends survival only in the SOD1 ALS model with highest motor neuron death. Proc Natl Acad Sci U S A 105, 12599–12604. Chow, C. Y., Landers, J. E., Bergren, S. K., Sapp, P. C., Grant, A. E., Jones, J. M., et al. (2009) Deleterious variants of FIG4, a phosphoinositide phosphatase, in patients with ALS. Am J Hum Genet 84, 85–88. Chow, C. Y., Zhang, Y., Dowling, J. J., Jin, N., Adamska, M., Shiga, K., et al. (2007) Mutation of FIG4 causes neurodegeneration in the pale tremor mouse and patients with CMT4J. Nature 448, 68–72. Zhang, X., Chow, C. Y., Sahenk, Z., Shy, M. E., Meisler, M. H. and Li, J. (2008) Mutation of FIG4 causes a rapidly progressive, asymmetric neuronal degeneration. Brain 131, 1990–2001. Schmitt-John, T., Drepper, C., Mussmann, A., Hahn, P., Kuhlmann, M., Thiel, C., et al. (2005) Mutation of Vps54 causes motor neuron disease and defective spermiogenesis in the wobbler mouse. Nat Genet 37, 1213–1215. Suter, U., Welcher, A. A., Ozcelik, T., Snipes, G. J., Kosaras, B., Francke, U., et al. (1992) Trembler mouse carries a point mutation in a myelin gene. Nature 356, 241–244. Giese, K. P., Martini, R., Lemke, G., Soriano, P. and Schachner, M. (1992) Mouse P0 gene disruption leads to hypomyelination, abnormal expression of recognition molecules, and degeneration of myelin and axons. Cell 71, 565–576. Wrabetz, L., Feltri, M. L., Quattrini, A., Imperiale, D., Previtali, S., D’Antonio,
78.
79.
80.
81.
82.
83.
84.
85.
86.
87.
389
M., et al. (2000) P(0) glycoprotein overexpression causes congenital hypomyelination of peripheral nerves. J Cell Biol 148, 1021–1034. Wrabetz, L., D’Antonio, M., Pennuto, M., Dati, G., Tinelli, E., Fratta, P., et al. (2006) Different intracellular pathomechanisms produce diverse Myelin Protein Zero neuropathies in transgenic mice. J Neurosci 26, 2358–2368. Le, N., Nagarajan, R., Wang, J. Y., Araki, T., Schmidt, R. E. and Milbrandt, J. (2005) Analysis of congenital hypomyelinating Egr2Lo/Lo nerves identifies Sox2 as an inhibitor of Schwann cell differentiation and myelination. Proc Natl Acad Sci U S A 102, 2596–2601. Topilko, P., Schneider-Maunoury, S., Levi, G., Baron-Van Evercooren, A., Chennoufi, A. B., Seitanidou, T., et al. (1994) Krox-20 controls myelination in the peripheral nervous system. Nature 371, 796–799. Nelles, E., Butzler, C., Jung, D., Temme, A., Gabriel, H. D., Dahl, U., et al. (1996) Defective propagation of signals generated by sympathetic nerve stimulation in the liver of connexin32-deficient mice. Proc Natl Acad Sci U S A 93, 9565–9570. Anzini, P., Neuberg, D. H., Schachner, M., Nelles, E., Willecke, K., Zielasek, J., et al. (1997) Structural abnormalities and deficient maintenance of peripheral nerve myelin in mice lacking the gap junction protein connexin 32. J Neurosci 17, 4545–4551. Jeng, L. J., Balice-Gordon, R. J., Messing, A., Fischbeck, K. H. and Scherer, S. S. (2006) The effects of a dominant connexin 32 mutant in myelinating Schwann cells. Mol Cell Neurosci 32, 283–298. Detmer, S. A., Vande Velde, C., Cleveland, D. W. and Chan, D. C. (2008) Hindlimb gait defects due to motor axon loss and reduced distal muscles in a transgenic mouse model of Charcot-Marie-Tooth type 2A. Hum Mol Genet 17, 367–375. Lee, M. K., Marszalek, J. R. and Cleveland, D. W. (1994) A mutant neurofilament subunit causes massive, selective motor neuron death: implications for the pathogenesis of human motor neuron disease. Neuron 13, 975–988. Zhu, Q., Couillard-Despres, S. and Julien, J. P. (1997) Delayed maturation of regenerating myelinated axons in mice lacking neurofilaments. Exp Neurol 148, 299–316. Huang, L., Min, J. N., Masters, S., Mivechi, N. F. and Moskophidis, D. (2007) Insights
390
88.
89.
90.
91.
92.
93.
94.
95.
96.
97.
Burgess, Cox, and Seburn into function and regulation of small heat shock protein 25 (HSPB1) in a mouse model with targeted gene disruption. Genesis 45, 487–501. Gillespie, C. S., Sherman, D. L., FleetwoodWalker, S. M., Cottrell, D. F., Tait, S., Garry, E. M., et al. (2000) Peripheral demyelination and neuropathic pain behavior in periaxindeficient mice. Neuron 26, 523–531. Brandon, E. P., Lin, W., D’Amour, K. A., Pizzo, D. P., Dominguez, B., Sugiura, Y., et al. (2003) Aberrant patterning of neuromuscular synapses in choline acetyltransferase-deficient mice. J Neurosci 23, 539–549. Misgeld, T., Burgess, R. W., Lewis, R. M., Cunningham, J. M., Lichtman, J. W. and Sanes, J. R. (2002) Roles of neurotransmitter in synapse formation. Development of neuromuscular junctions lacking choline acetyltransferase. Neuron 36, 635–648. Buffelli, M., Burgess, R. W., Feng, G., Lobe, C. G., Lichtman, J. W. and Sanes, J. R. (2003) Genetic evidence that relative synaptic efficacy biases the outcome of synaptic competition. Nature 424, 430–434. Prado, V. F., Martins-Silva, C., de Castro, B. M., Lima, R. F., Barros, D. M., Amaral, E., et al. (2006) Mice deficient for the vesicular acetylcholine transporter are myasthenic and have deficits in object and social recognition. Neuron 51, 601–612. Feng, G., Krejci, E., Molgo, J., Cunningham, J. M., Massoulie, J. and Sanes, J. R. (1999) Genetic analysis of collagen Q: roles in acetylcholinesterase and butyrylcholinesterase assembly and in synaptic structure and function. J Cell Biol 144, 1349–1360. Adler, M., Manley, H. A., Purcell, A. L., Deshpande, S. S., Hamilton, T. A., Kan, R. K., et al. (2004) Reduced acetylcholine receptor density, morphological remodeling, and butyrylcholinesterase activity can sustain muscle function in acetylcholinesterase knockout mice. Muscle Nerve 30, 317–327. Mesulam, M. M., Guillozet, A., Shaw, P., Levey, A., Duysen, E. G. and Lockridge, O. (2002) Acetylcholinesterase knockouts establish central cholinergic pathways and can use butyrylcholinesterase to hydrolyze acetylcholine. Neuroscience 110, 627–639. Burgess, R. W., Nguyen, Q. T., Son, Y. J., Lichtman, J. W. and Sanes, J. R. (1999) Alternatively spliced isoforms of nerve- and muscle-derived agrin: their roles at the neuromuscular junction. Neuron 23, 33–44. Burgess, R. W., Skarnes, W. C. and Sanes, J. R. (2000) Agrin isoforms with distinct amino
98.
99.
100.
101.
102.
103.
104.
105.
106.
107.
108.
termini: differential expression, localization, and function. J Cell Biol 151, 41–52. Lin, W., Burgess, R. W., Dominguez, B., Pfaff, S. L., Sanes, J. R. and Lee, K. F. (2001) Distinct roles of nerve and muscle in postsynaptic differentiation of the neuromuscular synapse. Nature 410, 1057–1064. Harvey, S. J., Jarad, G., Cunningham, J., Rops, A. L., van der Vlag, J., Berden, J. H., et al. (2007) Disruption of glomerular basement membrane charge through podocytespecific mutation of agrin does not alter glomerular permselectivity. Am J Pathol 171, 139–152. Gautam, M., Noakes, P. G., Moscoso, L., Rupp, F., Scheller, R. H., Merlie, J. P., et al. (1996) Defective neuromuscular synaptogenesis in agrin-deficient mutant mice. Cell 85, 525–535. DeChiara, T. M., Bowen, D. C., Valenzuela, D. M., Simmons, M. V., Poueymirou, W. T., Thomas, S., et al. (1996) The receptor tyrosine kinase MuSK is required for neuromuscular junction formation in vivo. Cell 85, 501–512. Weatherbee, S. D., Anderson, K. V. and Niswander, L. A. (2006) LDL-receptorrelated protein 4 is crucial for formation of the neuromuscular junction. Development 133, 4993–5000. Gautam, M., Noakes, P. G., Mudd, J., Nichol, M., Chu, G. C., Sanes, J. R.,et al. (1995) Failure of postsynaptic specialization to develop at neuromuscular junctions of rapsyn-deficient mice. Nature 377, 232–236. Okada, K., Inoue, A., Okada, M., Murata, Y., Kakuta, S., Jigami, T., et al. (2006) The muscle protein Dok-7 is essential for neuromuscular synaptogenesis. Science 312, 1802–1805. Noakes, P. G., Gautam, M., Mudd, J., Sanes, J. R. and Merlie, J. P. (1995) Aberrant differentiation of neuromuscular junctions in mice lacking s-laminin/laminin beta 2. Nature 374, 258–262. Friese, M. B., Blagden, C. S. and Burden, S. J. (2007) Synaptic differentiation is defective in mice lacking acetylcholine receptor betasubunit tyrosine phosphorylation. Development 134, 4167–4176. Missias, A. C., Mudd, J., Cunningham, J. M., Steinbach, J. H., Merlie, J. P. and Sanes, J. R. (1997) Deficient development and maintenance of postsynaptic specializations in mutant mice lacking an ’adult’ acetylcholine receptor subunit. Development 124, 5075–5086. Witzemann, V., Schwarz, H., Koenen, M., Berberich, C., Villarroel, A., Wernig, A.,
Neuromuscular Disease Models and Analysis
109.
110.
111.
112.
113.
114.
115.
116.
117.
118.
et al. (1996) Acetylcholine receptor epsilonsubunit deletion causes muscle weakness and atrophy in juvenile and adult mice. Proc Natl Acad Sci USA 93, 13286–13291. Takahashi, M., Kubo, T., Mizoguchi, A., Carlson, C. G., Endo, K. and Ohnishi, K. (2002) Spontaneous muscle action potentials fail to develop without fetal-type acetylcholine receptors. EMBO Rep 3, 674–681. Koenen, M., Peter, C., Villarroel, A., Witzemann, V. and Sakmann, B. (2005) Acetylcholine receptor channel subtype directs the innervation pattern of skeletal muscle. EMBO Rep 6, 570–576. Liu, Y., Padgett, D., Takahashi, M., Li, H., Sayeed, A., Teichert, R. W., et al. (2008) Essential roles of the acetylcholine receptor gamma-subunit in neuromuscular synaptic patterning. Development 135, 1957–1967. Sicinski, P., Geng, Y., Ryder Cook, A. S., Barnard, E. A., Darlison, M. G. and Barnard, P. J. (1989) The molecular basis of muscular dystrophy in the mdx mouse: a point mutation. Science 244, 1578–1580. Chapman, V. M., Miller, D. R., Armstrong, D. and Caskey, C. T. (1989) Recovery of induced mutations for X chromosome-linked muscular dystrophy in mice. Proc Natl Acad Sci U S Am 86, 1292–1296. Danko, I., Chapman, V. and Wolff, J. A. (1992) The frequency of revertants in mdx mouse genetic models for Duchenne muscular dystrophy. Pediatr Res 32, 128–131. Cox, G. A., Phelps, S. F., Chapman, V. M. and Chamberlain, J. S. (1993) New mdx mutation disrupts expression of muscle and nonmuscle isoforms of dystrophin. Nature Genetics 4, 87–93. Im, W. B., Phelps, S. F., Copen, E. H., Adams, E. G., Slightom, J. L. and Chamberlain, J. S. (1996) Differential expression of dystrophin isoforms in strains of mdx mice with different mutations. Hum Mol Genet 5, 1149–1153. Araki, E., Nakamura, K., Nakao, K., Kameya, S., Kobayashi, O., Nonaka, I., et al. (1997) Targeted disruption of exon 52 in the mouse dystrophin gene induced muscle degeneration similar to that observed in Duchenne muscular dystrophy. Biochem Biophys Res Commun 238, 492–497. Kudoh, H., Ikeda, H., Kakitani, M., Ueda, A., Hayasaka, M., Tomizuka, K., et al. (2005) A new model mouse for Duchenne muscular dystrophy produced by 2.4 Mb deletion of dystrophin gene using Cre-loxP recombination system. Biochem Biophys Res Commun 328, 507–516.
391
119. Grewal, P. K., Holzfeind, P. J., Bittner, R. E. and Hewitt, J. E. (2001) Mutant glycosyltransferase and altered glycosylation of alphadystroglycan in the myodystrophy mouse. Nat Genet 28, 151–154. 120. Lee, Y., Kameya, S., Cox, G. A., Hsu, J., Hicks, W., Maddatu, T. P., et al. (2005) Ocular abnormalities in Large(myd) and Large(vls) mice, spontaneous models for muscle, eye, and brain diseases. Mol Cell Neurosci 30, 160–172. 121. Kuang, W., Xu, H., Vachon, P. H., Liu, L., Loechel, F., Wewer, U. M., et al. (1998) Merosin-deficient congenital muscular dystrophy. Partial genetic correction in two mouse models. J Clin Invest 102, 844–852. 122. Michelson, A. M., Russell, E. S. and Harman, P. J. (1955) Dystrophia muscularis: a hereditary primary myopathy in the house mouse. Proc Natl Acad Sci U S A 12, 1079–1084. 123. Miyagoe, Y., Hanaoka, K., Nonaka, I., Hayasaka, M., Nabeshima, Y., Arahata, K., et al. (1997) Laminin alpha2 chain-null mutant mice by targeted disruption of the Lama2 gene: a new model of merosin (laminin 2)-deficient congenital muscular dystrophy. FEBS Lett 415, 33–39. 124. Patton, B. L., Wang, B., Tarumi, Y. S., Seburn, K. L. and Burgess, R. W. (2008) A single point mutation in the LN domain of LAMA2 causes muscular dystrophy and peripheral amyelination. J Cell Sci 121, 1593–1604. 125. Sunada, Y., Bernier, S. M., Utani, A., Yamada, Y. and Campbell, K. P. (1995) Identification of a novel mutant transcript of laminin a2 chain gene responsible for muscular dystrophy and dysmyelination in dy2J mice. Hum Mol Genet 4, 1055–1061. 126. Bittner, R. E., Anderson, L. V., Burkhardt, E., Bashir, R., Vafiadaki, E., Ivanova, S., et al. (1999) Dysferlin deletion in SJL mice (SJLDysf) defines a natural model for limb girdle muscular dystrophy 2B. Nat Genet 23, 141–142. 127. Bansal, D., Miyake, K., Vogel, S. S., Groh, S., Chen, C. C., Williamson, R., et al. (2003) Defective membrane repair in dysferlindeficient muscular dystrophy. Nature 423, 168–172. 128. Ho, M., Post, C. M., Donahue, L. R., Lidov, H. G., Bronson, R. T., Goolsby, H., et al. (2004) Disruption of muscle membrane and phenotype divergence in two novel mouse models of dysferlin deficiency. Hum Mol Genet 13, 1999–2010. 129. Grady, R. M., Grange, R. W., Lau, K. S., Maimone, M. M., Nichol, M. C., Stull, J. T., et al. (1999) Role for alpha-dystrobrevin
392
130.
131.
132.
133.
134.
135.
136.
137.
138.
139.
Burgess, Cox, and Seburn in the pathogenesis of dystrophin-dependent muscular dystrophies. Nat Cell Biol 1, 215–220. Bonaldo, P., Braghetta, P., Zanetti, M., Piccolo, S., Volpin, D. and Bressan, G. M. (1998) Collagen VI deficiency induces early onset myopathy in the mouse: an animal model for Bethlem myopathy. Hum Mol Genet 7, 2135–2140. Irwin, W. A., Bergamin, N., Sabatelli, P., Reggiani, C., Megighian, A., Merlini, L., et al. (2003) Mitochondrial dysfunction and apoptosis in myopathic mice with collagen VI deficiency. Nat Genet 35, 367–371. Eklund, L., Piuhola, J., Komulainen, J., Sormunen, R., Ongvarrasopone, C., Fassler, R., et al. (2001) Lack of type XV collagen causes a skeletal myopathy and cardiovascular defects in mice. Proc Natl Acad Sci U S A 98, 1194–1199. Galbiati, F., Engelman, J. A., Volonte, D., Zhang, X. L., Minetti, C., Li, M., et al. (2001) Caveolin-3 null mice show a loss of caveolae, changes in the microdomain distribution of the dystrophin-glycoprotein complex, and t-tubule abnormalities. J Biol Chem 276, 21425–21433. Crawford, K., Flick, R., Close, L., Shelly, D., Paul, R., Bove, K., et al. (2002) Mice lacking skeletal muscle actin show reduced muscle strength and growth deficits and die during the neonatal period. Mol Cell Biol 22, 5887–5896. Ilkovski, B., Cooper, S. T., Nowak, K., Ryan, M. M., Yang, N., Schnell, C., et al. (2001) Nemaline myopathy caused by mutations in the muscle alpha-skeletal-actin gene. Am J Hum Genet 68, 1333–1343. Garvey, S. M., Rajan, C., Lerner, A. P., Frankel, W. N. and Cox, G. A. (2002) The muscular dystrophy with myositis (mdm) mouse mutation disrupts a skeletal musclespecific domain of titin. Genomics 79, 146–149. Gotthardt, M., Hammer, R. E., Hubner, N., Monti, J., Witt, C. C., McNabb, M., et al. (2003) Conditional expression of mutant Mline titins results in cardiomyopathy with altered sarcomere structure. J Biol Chem 278, 6059–6065. May, S. R., Stewart, N. J., Chang, W. and Peterson, A. S. (2004) A Titin mutation defines roles for circulation in endothelial morphogenesis. Dev Biol 270, 31–46. Brady, J. P., Garland, D. L., Green, D. E., Tamm, E. R., Giblin, F. J. and Wawrousek, E. F. (2001) AlphaB-crystallin in lens development and muscle integrity: a gene knock-
140.
141.
142.
143.
144.
145.
146.
147.
148.
149.
150.
out approach. Invest Ophthalmol Vis Sci 42, 2924–2934. Homma, S., Iwasaki, M., Shelton, G. D., Engvall, E., Reed, J. C. and Takayama, S. (2006) BAG3 deficiency results in fulminant myopathy and early lethality. Am J Pathol 169, 761–773. Selcen, D., Muntoni, F., Burton, B. K., Pegoraro, E., Sewry, C., Bite, A. V., et al. (2009) Mutation in BAG3 causes severe dominant childhood muscular dystrophy. Ann Neurol 65, 83–89. Li, Z., Colucci-Guyon, E., Pincon-Raymond, M., Mericskay, M., Pournin, S., Paulin, D., et al. (1996) Cardiovascular lesions and skeletal myopathy in mice lacking desmin. Dev Biol 175, 362–366. Milner, D. J., Weitzer, G., Tran, D., Bradley, A. and Capetanaki, Y. (1996) Disruption of muscle architecture and myocardial degeneration in mice lacking desmin. J Cell Biol 134, 1255–1270. Thornell, L., Carlsson, L., Li, Z., Mericskay, M. and Paulin, D. (1997) Null mutation in the desmin gene gives rise to a cardiomyopathy. J Mol Cell Cardiol 29, 2107–2124. Moza, M., Mologni, L., Trokovic, R., Faulkner, G., Partanen, J. and Carpen, O. (2007) Targeted deletion of the muscular dystrophy gene myotilin does not perturb muscle structure or function in mice. Mol Cell Biol 27, 244–252. Garvey, S. M., Miller, S. E., Claflin, D. R., Faulkner, J. A. and Hauser, M. A. (2006) Transgenic mice expressing the myotilin T57I mutation unite the pathology associated with LGMD1A and MFM. Hum Mol Genet 15, 2348–2362. Sher, R. B., Aoyama, C., Huebsch, K. A., Ji, S., Kerner, J., Yang, Y., et al. (2006) A rostrocaudal muscular dystrophy caused by a defect in choline kinase beta, the first enzyme in phosphatidylcholine biosynthesis. J Biol Chem 281, 4938–4948. Buj-Bello, A., Laugel, V., Messaddeq, N., Zahreddine, H., Laporte, J., Pellissier, J. F., et al. (2002) The lipid phosphatase myotubularin is essential for skeletal muscle maintenance but not for myogenesis in mice. Proc Natl Acad Sci U S A 99, 15060–15065. Sullivan, T., Escalante-Alcalde, D., Bhatt, H., Anver, M., Bhat, N., Nagashima, K., et al. (1999) Loss of A-type lamin expression compromises nuclear envelope integrity leading to muscular dystrophy. J Cell Biol 147, 913–920. Arimura, T., Helbling-Leclerc, A., Massart, C., Varnous, S., Niel, F., Lacene, E., et al. (2005) Mouse model carrying H222P-Lmna
Neuromuscular Disease Models and Analysis
151.
152.
153.
154.
155.
156.
157.
158.
159.
160.
mutation develops muscular dystrophy and dilated cardiomyopathy similar to human striated muscle laminopathies. Hum Mol Genet 14, 155–169. Diaz, F., Thomas, C. K., Garcia, S., Hernandez, D. and Moraes, C. T. (2005) Mice lacking COX10 in skeletal muscle recapitulate the phenotype of progressive mitochondrial myopathies associated with cytochrome c oxidase deficiency. Hum Mol Genet 14, 2737–2748. Mankodi, A., Logigian, E., Callahan, L., McClain, C., White, R., Henderson, D., et al. (2000) Myotonic dystrophy in transgenic mice expressing an expanded CUG repeat. Science 289, 1769–1773. Seznec, H., Agbulut, O., Sergeant, N., Savouret, C., Ghestem, A., Tabti, N., et al. (2001) Mice transgenic for the human myotonic dystrophy region with expanded CTG repeats display muscular and brain abnormalities. Hum Mol Genet 10, 2717–2726. Orengo, J. P., Chambon, P., Metzger, D., Mosier, D. R., Snipes, G. J. and Cooper, T. A. (2008) Expanded CTG repeats within the DMPK 3’ UTR causes severe skeletal muscle wasting in an inducible mouse model for myotonic dystrophy. Proc Natl Acad Sci U S A 105, 2646–2651. Reddy, S., Smith, D. B., Rich, M. M., Leferovich, J. M., Reilly, P., Davis, B. M., et al. (1996) Mice lacking the myotonic dystrophy protein kinase develop a late onset progressive myopathy. Nat Genet 13, 325–335. Kanadia, R. N., Johnstone, K. A., Mankodi, A., Lungu, C., Thornton, C. A., Esson, D., et al. (2003) A muscleblind knockout model for myotonic dystrophy. Science 302, 1978–1980. Malicdan, M. C., Noguchi, S., Nonaka, I., Hayashi, Y. K. and Nishino, I. (2007) A Gne knockout mouse expressing human GNE D176V mutation develops features similar to distal myopathy with rimmed vacuoles or hereditary inclusion body myopathy. Hum Mol Genet 16, 2669–2682. Tronche, F., Kellendonk, C., Kretz, O., Gass, P., Anlag, K., Orban, P. C., et al. (1999) Disruption of the glucocorticoid receptor gene in the nervous system results in reduced anxiety. Nat Genet 23, 99–103. Yang, X., Arber, S., William, C., Li, L., Tanabe, Y., Jessell, T. M., et al. (2001) Patterning of muscle acetylcholine receptor gene expression in the absence of motor innervation. Neuron 30, 399–410. Arber, S., Han, B., Mendelsohn, M., Smith, M., Jessell, T. M. and Sockanathan, S. (1999) Requirement for the homeobox gene Hb9 in
161.
162.
163.
164.
165.
166.
167.
168.
169.
170.
171.
393
the consolidation of motor neuron identity. Neuron 23, 659–674. Schwander, M., Leu, M., Stumm, M., Dorchies, O. M., Ruegg, U. T., Schittny, J., et al. (2003) Beta1 integrins regulate myoblast fusion and sarcomere assembly. Dev Cell 4, 673–685. Bruning, J. C., Michael, M. D., Winnay, J. N., Hayashi, T., Horsch, D., Accili, D., et al. (1998) A muscle-specific insulin receptor knockout exhibits features of the metabolic syndrome of NIDDM without altering glucose tolerance. Mol Cell 2, 559–569. Tallquist, M. D., Weismann, K. E., Hellstrom, M. and Soriano, P. (2000) Early myotome specification regulates PDGFA expression and axial skeleton development. Development 127, 5059–5070. Feltri, M. L., D’Antonio, M., Previtali, S., Fasolini, M., Messing, A. and Wrabetz, L. (1999) P0-Cre transgenic mice for inactivation of adhesion molecules in Schwann cells. Ann N Y Acad Sci 883, 116–123. Young, P., Qiu, L., Wang, D., Zhao, S., Gross, J. and Feng, G. (2008) Single-neuron labeling with inducible Cre-mediated knockout in transgenic mice. Nat Neurosci 11, 721–728. Feng, G., Mellor, R. H., Bernstein, M., Keller-Peck, C., Nguyen, Q. T., Wallace, M., et al. (2000) Imaging neuronal subsets in transgenic mice expressing multiple spectral variants of GFP. Neuron 28, 41–51. Misgeld, T., Kerschensteiner, M., Bareyre, F. M., Burgess, R. W. and Lichtman, J. W. (2007) Imaging axonal transport of mitochondria in vivo. Nat Methods 4, 559–561. Zuo, Y., Lubischer, J. L., Kang, H., Tian, L., Mikesh, M., Marks, A., et al. (2004) Fluorescent proteins expressed in mouse transgenic lines mark subsets of glia, neurons, macrophages, and dendritic cells for vital examination. J Neurosci 24, 10999–11009. Wichterle, H., Lieberam, I., Porter, J. A. and Jessell, T. M. (2002) Directed differentiation of embryonic stem cells into motor neurons. Cell 110, 385–397. Livet, J., Weissman, T. A., Kang, H., Draft, R. W., Lu, J., Bennis, R. A., et al. (2007) Transgenic strategies for combinatorial expression of fluorescent proteins in the nervous system. Nature 450, 56–62. Coleman, M. P., Conforti, L., Buckmaster, E. A., Tarlton, A., Ewing, R. M., Brown, M. C., et al. (1998) An 85-kb tandem triplication in the slow Wallerian degeneration (Wlds) mouse. Proc Natl Acad Sci U S A 95, 9985–9990.
Chapter 21 Murine Model of Cutaneous Infection with Streptococcus pyogenes Eva Medina Abstract Despite the medical advances achieved during the last century to fight against bacteria, viruses, fungi and parasites, infectious diseases are still a major cause of death, disability, and social and economic upheaval for millions around the world. Challenges remain in countering microorganisms even where antibiotics and vaccines are available. Much remains to be learned about basic aspects of the host–pathogen relationship and the complexity of the immune response to infection. Animal models represent a powerful tool to dissect the host response to infection, as well as the pathogenesis of the microbe. One of the advantages of using animal models is that both genetic and environmental factors that may influence the course of an infection can be controlled, allowing a precise cause–effect analysis of the host–pathogen interactions. In addition, there are no real alternatives to whole animal models in the study of integrative physiology and dynamic pathophysiologic alterations. The use of animal models has also proven invaluable for testing the efficacy of experimental antimicrobial agents and their therapeutic regimes. The mouse model is the most widely used for many reasons, including its cost effectiveness, the high number of immunological reagents available for this species, and the relative ease of biocontainment. Mouse strains with specific properties such as transgenic mouse strains with gene insertion or targeted mutation (knock-out) are very effective tools for studying the role of specific genes controlling the immune response to infectious pathogens. Murine models will remain the most appropriate tool for evaluating new therapeutic strategies for the treatment of various diseases. The closer the model is adapted to the human disease, the more reliable will be the results. In this chapter, the experimental procedures required to establish a mouse model of cutaneous and soft tissue infection are detailed. This model has provided invaluable insights into the pathogenicity of the agent for the human host. Key words: Subcutaneous inoculation, Streptococcus pyogenes, skin and soft tissue infection.
1. Introduction Skin and soft tissue infections are defined as infections of the epidermis, dermis, or subcutaneous tissue. They are among the most common human bacterial infections and are frequently observed G. Proetzel, M.V. Wiles (eds.), Mouse Models for Drug Discovery, Methods in Molecular Biology 602, DOI 10.1007/978-1-60761-058-8 21, © Humana Press, a part of Springer Science+Business Media, LLC 2010
395
396
Medina
in clinical practice (1). Most skin and soft tissue infections are caused by Gram-positive bacteria, primarily Staphylococcus aureus and Streptococcus pyogenes (2). Although these infections can be successfully treated using empirical antimicrobial therapy, the increasing prevalence of antibiotic resistance among some bacterial strains implies that new treatment options are required (3). Experimental mouse models of skin and soft tissue infections have played a critical role in providing detailed data about the efficacy of specific antimicrobial agents, pharmacokinetics, and disease pathogenesis. In this chapter, an experimental mouse model of skin and soft tissue infection induced by Streptococcus pyogenes is described. Rapidly progressing soft tissue infections due to group A beta-hemolytic streptococcus, or Streptococcus pyogenes can present with erysipelas, cellulitis, or with necrotizing fasciitis (4). Erysipelas is an acute inflammation of the superficial skin and cutaneous lymphatic vessels, whereas cellulitis is an infection of the lower dermis and subcutaneous tissue (4). The most severe soft tissue infection caused by S. pyogenes is necrotizing fasciitis, also known as the “Flesh-eating disease.” Necrotizing fasciitis is a deep-seated infection of the subcutaneous tissue that results in the rapidly progressive destruction of fascia and fat (4). Despite prompt antibiotic therapy and surgical debridement, streptococcal necrotizing fasciitis is associated with high death rates ranging from 20 to 60% (4). The mouse model described here mimics many of the features of severe S. pyogenes skin infection in humans and involves the subcutaneous inoculation of S. pyogenes resulting in severe local infection that lead to systemic bacterial dissemination, multiorgan failure, and death. This model of streptococcal skin infection provides a very useful tool for testing drugs with potential effect on bacterial dissemination. In this case, drug therapy should start soon (1–6 h) after bacterial inoculation. Also drugs that can stop progressive bacterial growth can be tested in this infection models and the therapy should start after the systemic dissemination of S. pyogenes (approximately at 24 h after bacterial inoculation).
2. Materials 1. Phosphate-buffered saline (PBS): 8 g NaCl, 0.2 g KCl, 1.44 g Na2 HPO4 , and 0.24 g KH2 PO4 in 800 ml of distilled H2 O. Adjust the pH to 7.4 with HCl. Add H2 O to 1 l. Sterilize by autoclave.
Murine Model of Cutaneous Infection
397
2. 70% (v /v) ethanol. R System PT 4000 (KINE3. Homogenator: POLYTRON MATICA AG, Littau/Luzern, Switzerland).
4. EasyClean Dispersing aggregates (KINEMATICA AG, Littau/Luzern, Switzerland). 5. 1–3 ml syringes and several sizes of needles. 6. Heat lamp. 7. Equipment for restraint of the animal. 8. Centrifuge (RCF: 3345 × g; capacity: 4 × 400 ml). 9. THY broth: Todd–Hewitt broth (Beckton Dickisnson, TM MD, USA) supplemented with 1% Bacto yeast extract (Beckton Dickisnson, MD, USA). 10. Blood agar plates (Beckton Dickisnson, MD, USA). 11. Glycerol (Roth, Karlsruhe, Germany Cat. Nr. 4043.3). 12. Isoflurane (Baxter, Germany). 13. Gauze or cotton. 14. Closed glass container with raised floor. 15. Caliper. 16. Electric shaver for small animals. 17. Bacteria (see Note 1): different strains of S. pyogenes have been used in mouse models of streptococcal skin infection including the commercially available S. pyogenes strains DSM 2071, which can be obtained from the German Culture Collection or from The American Type Culture Collection (ATCC 2105). 18. Mice (see Note 2): inbred 7- to 8-week-old female C57BL/6NHsd mice (Harlan-Winkelmann, Borchen, Germany) or C57BL/6 J mice (The Jackson Laboratory, Bar Harbor, USA) have been used in this infection model (see Note 3). Mice should be housed in microisolator cages. Other mouse strains such as outbred CD1 mice, inbred BALB/c, C3H/HeN as well as hairless crl:SKH1(hrhr) Br mice have also been used in models of streptococcal skin infection (see Note 4). Although outbred mice have been used in this model of infection, the use of inbred mouse provide a genetically better-defined model with the advantage of high reproducibility not available with standard outbred mouse strains. In addition, the phenotypic uniformity of an inbred mouse strain means that sample size can be reduced in comparison with the use of outbred stocks.
398
Medina
3. Methods 3.1. Inoculum Preparation
1. Stock bacteria cultures are maintained in THY broth plus 20% glycerol at –70◦ C. 2. Plate a small amount of stock onto blood agar the day before infection and incubate overnight at 37◦ C. 3. Inoculate a fresh colony from a blood agar plate into THY medium. 4. Incubate at 37◦ C until midlog phase (OD600 = 0.5–0.6). 5. Harvest streptococci by centrifugation and wash twice with sterile PBS. 6. Adjust to a concentration of 5×108 colony-forming units (CFU)/ml in sterile PBS (see Note 5). For standardization of streptococci inoculum, a regression curve should be constructed for the selected strain by plotting the log number of serial bacterial dilutions against the percentage transmittance of the suspensions read at 600 nm wavelength in a specR trophotometer (Novospec II, Pharmacia). Transmittance of uninoculated THY broth from the same batch should be used as blank. 7. The inoculum concentration should be verified by plating 10-fold serial dilutions onto blood agar and counting after incubating for 24 h at 37◦ C (see Note 6).
3.2. Subcutaneous Infection Procedure
1. Anesthetize mice by inhalation of isoflurane (see Note 7): 1.1 Pour isoflurane on gauze or cotton placed in the bottom of a closed glass container. 1.2 Place a raised floor over the soaked material to allow isoflurane to vaporize without impregnating the mouse fur. 1.3 Place the mouse in the container and close the lid. 1.4 Keep the animal inside until reaching slow and regular breathing. 1.5 Remove the animal from the container to perform infection. 2. Shave the back of mice using an electric shaver for small animals. 3. Inoculate streptococci on the shaved back of the mouse (alternative sites for the initiation of subcutaneous infections are the flanks and the scruff of the neck) by subcutaneous injection: 3.1. Fill 1 ml syringe carrying a 22 G needle with 100 l of a suspension containing 107 CFU of bacteria.
Murine Model of Cutaneous Infection
399
3.2. Place anesthetized animal on a clean paper towel. 3.3. Insert the needle under the skin of the intercapsular area and inject slowly with moderate pressure. 3.4. Inoculate control mice with sterile PBS. 3.3. Key Parameters to Monitor Infection
The main parameters commonly used for monitoring the course of an experimental skin and soft tissue infection are the death or survival of the animal, the development and progression of the local skin lesion, and the removal of organs for determination of bacterial loads, immune response analysis, or histological evaluation. Survival curves can be constructed by monitoring the time of death of individual mice after bacterial inoculation (see Note 8). 1. Lesion development: the development of a lesion at the site of infection can be monitored by measuring daily the area of dermonecrosis with a caliper and calculated by using the formula: (L × W ) A= 2 where L is length and W is width. 2. Quantification of bacterial loads: groups of mice are killed at selected intervals to count the numbers of bacteria at the local skin and systemic organs (see Note 9): 2.1 Euthanize the infected animal by CO2 asphyxiation. 2.2 Place animal on absorbent paper tissue over the dissection board and fix the upper and lower extremities with dissection pins. 2.3 Remove the infected lesion by using a sterile scalpel or scissors and place in a 10 ml tube containing 4.5 ml of sterile PBS. 2.4 Remove carefully the spleen, liver, kidneys, and lungs and place them in a 10 ml homogenization tubes containing 4.5 ml of sterile PBS. 2.5 Homogenize the skin and organs. 2.6 Perform 10-fold serial dilutions of tissue homogenate by mixing 100 l of tissue homogenate and 900 l of sterile PBS. 2.7 Apply 100 l of each dilution to a blood agar plate and gently swirl the plate. 2.8 Incubate plates at 37◦ C for 24 h. 2.9 Determine the bacterial load within the organ using the following formula: C × Vt CFU/organ = Vp × DF
400
Medina
where C is the colony counts, Vt is the total volume of the organ homogenate, VP is the volume plated, and DF is the dilution factor of the sample. 3.4. Blood Collection
Sampling of blood from mice can be performed using several methods depending upon the purpose and volume of the blood sample.
3.4.1. Blood Collection from the Tail Vein
From the tail vein, small amounts of blood for monitoring systemic bacterial dissemination from the local site of infection can be collected at different times of infection without the need to kill the animal (see Note 10): 1. Restrain the animal. 2. Warm the tail with a heat lamp to dilate the vessels. 3. Extend the tail and make a small incision with a sterile scalpel blade in the lateral vein at the distal one-third of the tail. 4. Collect the blood flowing from the incision with a pipette. 5. Plate blood immediately onto blood agar after making 10-fold serial dilutions using sterile PBS. 6. Incubate plates at 37◦ C for 24 h. 7. Count colonies and express as CFU/ml of blood: CFU/ml =
C Vp × DF
where C is the colony counts, VP is the volume plated, and DF is the dilution factor of the sample. 3.4.2. Blood Collection by Cardiac Puncture
By cardiac puncture, when animal survival is not required, exsanguinations by cardiac puncture yields the maximal volume of blood: 1. Anesthetize the mouse and place on its back over clean paper towels. 2. Insert a 20 G needle attached to a 1 or 3 ml syringe just below and slightly to the left of the xiphoid cartilage at the base of the sternum, at a 20◦ angle. 3. Apply very slight negative pressure on the barrel of the syringe and aspirate gently until blood flow comes to an end (see Notes 11 and 12). 4. Euthanize the animal by CO2 asphyxiation.
3.5. Quantification of Cytokines in Serum Samples
1. Allow blood to clot. 2. Separate serum by centrifugation at 4◦ C and 450×g for 10 min.
Murine Model of Cutaneous Infection
401
3. Determine cytokines levels by Enzyme-Linked ImmunoSorbent Assay (ELISA) (see Note 13). 4. Serum samples can also be stored at –70◦ C until required for analysis.
4. Notes 1. S. pyogenes is considered to be a potential hazard to personnel, and Biosafety Level 2 practices and facilities should be used when working with these pathogens. All work with this microorganism should have prior approval of the institutional Biological Safety Department and should strictly follow the guidelines and regulations for the handling of this pathogen. 2. All protocols using live animals must first be reviewed and approved by the corresponding Governmental and Institutional Animal Care and Use Committee and must conform to the regulations regarding the care and use of laboratory animals. 3. Female mice are recommended since male mice are extremely susceptible to S. pyogenes infection (5). 4. Different inbred mouse strains have been shown to strongly differ in their level of susceptibility to S. pyogenes infection (5). Thus, CBA and C3H/HeN have been reported to be very susceptible, whereas BALB/c and DBA/2 are much more resistant (5). The reader is encouraged to consult the relevant literature for details relating to the levels of susceptibility of different mouse strains to S. pyogenes infection. 5. Different strains of S. pyogenes vary strongly in their levels of virulence. It is usually a good idea to determine a 50% lethal dose (LD50 ) for the selected strain of S. pyogenes. For this purpose, groups of 10 mice are inoculated with increasing numbers (e.g., 104 , 106 , 107 , 108 CFU) of bacteria. The LD50 is determined by using the Reed–Muench formula (6): ⎤ Dilution ⎥ ⎢ % mortality > 50%−50 coefficient of − ⎥ anti Lg10 ⎢ ⎦ ⎣ % mortality > 50% − % mortatity < 50% % of mortality > 50% ⎡
The selection of a lethal (leading to 100% mortality) or sublethal (leading to >50% survival) dose
402
Medina
should be performed according to the purpose of the experiment. 6. The inoculum should be used immediately after preparation since the viability of S. pyogenes in PBS decreases over time. 7. All procedures using isoflurane should be conducted in a fume hood that continuously exhausts anesthetic gases away from personnel. 8. In some countries, for ethical reasons, the mice are not allowed to die from infection and they must be euthanized or processed for sampling when they develop signs of severe infection (e.g., 20% of weight loss, lethargy, piloerection, stop taking food, and water). 9. The number of mice per group should be adequate to allow statistical analysis of the data. Because of the phenotypic uniformity, experimental infections using inbred mouse strains will require smaller number of animals to reach statistical significance than those employing outbred strains. 10. No more than 10% of the blood volume should be removed at any one time and maximum once a week. The maximum volume of a single sample should not exceed 0.25 ml since the average blood volume of an adult mouse is 2.5 ml. 11. Advancing or retracting the needle may be necessary to obtain a maxilla volume. 12. To prevent hemolysis, remove needle from the syringe before transferring blood to collection tube. 13. The simplest method for determination of cytokines by ELISA is to use commercially available kits with matching antibody pairs and standards for the specific cytokine. The reader can consult web sites such as http://www.biocompare.com/ to select from a wide range of commercially available kits with the corresponding manufactures instructions for use.
References 1. Vinh, D. C. and Embil, J. M. (2005) Rapidly progressive soft tissue infections. Lancet Infect Dis. 5, 501–513. 2. Moet, G. J., Jones, R. N., Biedenbach, D. J., Stilwell, M. G., and Fritsche, T. R. (2007) Contemporary causes of skin and soft tissue infections in North America, Latin America, and Europe: Report from the SENTRY
Antimicrobial Surveillance Program (1998– 2004). Diagn Microbiol Infect Dis. 57, 7–13. 3. Segreti, J. (2005) Efficacy of current agents used in the treatment of Grampositive infections and the consequences of resistance. Clin Microbiol Infect. 11 (suppl 3), 29–35.
Murine Model of Cutaneous Infection 4. Bisno, A. L. and Stevens, D. L. (1996) Streptococcal infections of skin and soft tissues. N Engl J Med. 334, 240–245. 5. Medina, E., Goldmann, O., Rohde, M., Lengeling, A., and Chhatwal, G. S. (2001)
403
Genetic control of susceptibility to group A streptococcal infection in mice. J Infect Dis. 184, 846–852. 6. Reed, L. J. and Muench,H. A. (1938) A simple method of stimating fifty percent end points. Am J Hyg. 27, 493–497.
Chapter 22 Murine Model of Pneumococcal Pneumonia Eva Medina Abstract Respiratory tract infections remain among the most common clinical problems worldwide. Pneumonia or inflammation of the lungs can be caused by infection with bacteria, viruses, and other organisms. Pneumonia management has been challenged by the widespread distribution of antibiotic-resistant strains of Streptococcus pneumoniae, the commonest cause of community acquired pneumonia. Experimental models of pneumonia have played a crucial role for testing the efficacy of antimicrobial agents as well as for gaining a better understanding of the disease pathogenesis. These models have also received increased attention as tools for deriving pharmacodynamic data and for determining the clinical significance of drug resistance. Key words: Respiratory infection, Streptococcus pneumoniae , pneumonia, intranasal inoculation.
1. Introduction Streptococcus pneumoniae is a major human pathogen that colonizes the upper respiratory tract and causes both lifethreatening diseases such as pneumonia, sepsis, and meningitis and milder but common diseases, such as sinusitis and otitis media (1). S. pneumoniae is the leading cause of bacterial pneumonia. The burden of pneumococcal infections is particularly large among children and the elderly and is exacerbated by the rising numbers of isolates resistant to antibiotics (2). Mouse models of pneumonia have been used to characterize the course of infection by determining animal survival after infection, the bacterial loads in lungs and other organs, levels of inflammation, and histology of lung tissue (3). In addition, studies in the mouse model of streptococcal pneumoniae have provided critical information on antimicrobial efficacy that is directly relevant to the treatment G. Proetzel, M.V. Wiles (eds.), Mouse Models for Drug Discovery, Methods in Molecular Biology 602, DOI 10.1007/978-1-60761-058-8 22, © Humana Press, a part of Springer Science+Business Media, LLC 2010
405
406
Medina
of human infection (4, 5). Murine models of S. pneumoniae have been also essential for testing strategies of vaccination against this pathogen (6). In the mouse model described here, pneumonia is induced after the intranasal instillation of S. pneumoniae. When using this model for evaluation of antimicrobial agents, therapeutic treatment should start after the infection has already been established in the lungs (6–12 h after bacterial inoculation).
2. Materials 1. Phosphate-buffered saline (PBS): 8 g NaCl, 0.2 g KCl, 1.44 g Na2 HPO4 , and 0.24 g KH2 PO4 in 800 ml of distilled H2 O. Adjust the pH to 7.4 with HCl. Add H2 O to 1 l. Sterilize by autoclave. 2. 70% (v /v) ethanol. R 3. Homogenator: POLYTRON System PT 4000 (KINEMATICA AG, Littau/Luzern, Switzerland).
4. EasyClean Dispersing aggregates (KINEMATICA AG, Littau/Luzern, Switzerland). 5. 1 ml syringes and 25 G needles. 6. THY broth: Todd–Hewitt broth (Beckton Dickinson, MD, USA) supplemented with 1% BactoTM yeast extract (Beckton Dickinson, MD, USA). 7. Blood agar plates (Beckton Dickinson, MD, USA; Cat. Nr. 221165). 8. Glycerol (Roth, Karlsruhe, Germany). 9. Isoflurane (Baxter, Germany). 10. Gauze or cotton. 11. Closed glass container with raised floor. 12. 10% neutral buffered formalin: 10% Formaldehyde in PBS (see Note 1). R , Ravensburg, Germany). 13. Ketamine (Narketan
14. Xylene (J.T. Baker Chemical Co, Phillipsburg, USA). 15. Bacteria (see Note 2): the majority of the published studies have used the serotype 2 S. pneumoniae (NCTC 7466) that can be obtained from the National Collection of Type Cultures, London, United Kingdom. Other strains suitable for the mouse model are the serotype 3 S. pneumoniae (ATCC 6303) and serotype 1 S. pneumoniae (ATCC 6301), both can be obtained from the American Type Culture Collection.
Murine Model of Pneumococcal Pneumonia
407
16. Mice (see Note 3): several inbred strains of mice have been used in this infection model including BALB/c and CBA/Ca among others (see Note 4).
3. Methods 3.1. Inoculum Preparation
1. Stock bacteria cultures are maintained in THY broth plus 20% glycerol at –70◦ C. 2. Plate a small amount of stock onto blood agar the day before infection and incubate overnight at 37◦ C in 5% CO2 . 3. Inoculate a fresh colony from a blood agar plate into THY medium. 4. Incubate at 37◦ C until mid-log phase (OD600 = 0.3–0.4). 5. Harvest pneumococci by centrifugation and wash twice with sterile PBS. 6. Adjust the inoculum to the desired concentration in sterile PBS (see Note 5). For standardization of S. pneumoniae inoculum, a regression curve should be constructed for the selected strain by plotting the log number of serial bacterial dilutions against the percentage optical density of the suspensions read at 600 nm wavelength in a spectrophotomeR ter (Novospec II, Pharmacia). Optical density of uninoculated THY from the same batch is used as blank. 7. The inoculum concentration should be verified by plating 10-fold serial dilutions onto blood agar and counting after incubating for 24 h at 37◦ C (see Note 6).
3.2. Intranasal Infection Procedure
1. Anesthetize mice by inhalation of isoflurane in order to facilitate aspiration (see Note 7): 1.1 Pour isoflurane on gauze or cotton placed in the bottom of a closed glass container. 1.2 Place a raised floor over the soaked material to allow isoflurane to vaporize without impregnating the mouse fur. 1.3 Place the mouse in the container and close the lid. 1.4 Keep the animal inside until reaching slow and regular breathing. 1.5 Remove the animal from the container to perform infection.
408
Medina
2. Holding the mice vertically, deliver a 40 l volume of bacterial inoculum to the nostrils to induce aspiration pneumonia. Inoculate control mice with sterile PBS. 3.3. Bacterial Counts in the Lungs as Parameter to Monitor Infection
1. Euthanize mice at selected time intervals following infection by cervical dislocation. Remove the lungs aseptically and place them in a 10 ml homogenization tube containing 4.5 ml of sterile saline. 2. Homogenize the lungs and perform 10-fold serial dilutions of lung homogenate in sterile PBS. 3. Apply 100 l of each dilution to a blood agar plate and gently swirl the plate. 4. Incubate plates at 37◦ C for 24 h. 5. Determine the bacterial load within the lungs using the following formula:
CFU/lungs =
C × Vt V p × DF
where C is the colony counts, Vt is the total volume of the lung homogenate, VP is the volume plated, and DF is the dilution factor of the sample. 3.4. Histological Examination of Lung Tissue to Monitor Infection
1. At chosen intervals after bacterial inoculation, mice are euthanized by an overdose of ketamine (see Note 8). 2. Remove lungs and fix them with neutral buffered formalin for 48 h at 4◦ C. 3. Wash lungs with dH2 O and dehydrate the tissue by serial immersion in ethanol (50, 70, 95, and 100%) 1 h each, finishing with incubation for 1 h in 100% xylene. 4. Embed lung tissue in paraffin and cut 5 m thickness sections. 5. Dewax tissue section starting with 100% xylene for 1 h and continuing with serial immersion in ethanol (100, 95, 70, and 50%) 1 h each finishing with dH2 O. 6. Stain sections and mount (see Note 9). 7. Examine by light microscopy.
3.5. Leukocyte Recruitment into the Lungs as Parameter to Monitor Infection
The recruitment of leukocytes into the lungs following infection with S. pneumoniae is analyzed by broncho-alveolar lavage (BAL) fluid counts: 1. Euthanize mice by an overdose of ketamine (see Note 8). 2. Disinfect the neck region with 70% ethanol.
Murine Model of Pneumococcal Pneumonia
409
3. Make a middle incision in the neck area and retract the skin with forceps. 4. Cut away the muscles overlying the trachea carefully to not affect the aorta. 5. Make a small incision in the upper part of the trachea and insert a polyethylene catheter attached to a 25 G needle on a 1 ml tuberculin syringe (see Note 10). 6. Slowly inject 0.5 ml of sterile PBS through the catheter into the mouse lungs recovering the resultant BAL. 7. Determine total leukocyte counts under optical microscope using a Neubauer chamber. 8. Determine differential cell counts on cytospin smears of lung washes (3 min at 100×g) stained by the Wright–Giemsa method. 3.6. Quantification of Inflammatory Cytokines in Broncho-alveolar Lavage (BAL): As Parameter to Monitor Infection
1. Centrifuge lavage fluids for 10 min at 800×g. 2. Remove supernatant with a pipette and use for quantification of cytokines by ELISA (see Note 11).
4. Notes 1. Work with 10% buffered neutral formalin is recommended to be executed in a well-ventilated area, wearing goggles, gloves, and lab coat. Storage areas should have appropriate ventilation systems. 2. S. pneumoniae is considered to be a potential hazard to personnel, and Biosafety Level 2 practices and facilities should be used when working with this pathogen. All work with this microorganism should have prior approval of the institutional Biological Safety Department and should strictly follow the guidelines and regulations for the handling of this pathogen. 3. All protocols using live animals must first be reviewed and approved by the corresponding Governmental and Institutional Animal Care and Use Committee and must conform to the regulations regarding the care and use of laboratory animals. 4. Different mouse strains strongly differ in their degree of susceptibility to S. pneumoniae (7). While CBA/Ca and SJL are highly susceptible, BALB/c is highly resistant and C3H/He, FVB/n, NIH, AKR, C57BL/6, and DBA/2 exhibit intermediate resistance.
410
Medina
5. Concentrations ranging between 5×105 and 107 CFU can be used. 6. Due to propensity of S. pneumoniae to autolysis, the inoculum should never be vortexed or subjected to drastic shaking. 7. All procedures using isoflurane should be conducted in a fume hood that continuously exhausts anesthetic gases away from personnel. 8. CO2 asphyxiation is not recommended since it can cause perivascular edema in the lungs or alveolar hemorrhage. 9. The staining technique will depend on the type of cell of interest. Hematoxylin and eosin (H&E) stain is the most widely used method in histology. Using this technique, basophilic white blood cells stain dark blue, eosinophilic white blood cells stain bright red, and neutrophils stain a neutral pink. 10. The catheter should not be insert too far into the trachea since this will result in the dispensing of PBS into only one lobe of the lungs. 11. The simplest method for determination of cytokines by ELISA is to use commercially available kits with matching antibody pairs and standards for the specific cytokine. The reader can consult web sites such as http://www. biocompare.com/ to select form a wide range of commercially available kits with the corresponding manufactures instructions for use. References 1. Austrian, R. (1999) The pneumococcus at the millennium: not down, not out. J Infect Dis. 179 Suppl 2, S338–S341. 2. Ortqvist, A., Hedlund, J., and Kalin, M. (2005) Streptococcus pneumoniae: epidemiology, risk factors, and clinical features. Semin Respir Crit Care Med. 26, 563–574. 3. Chiavolini, D., Pozzi, G., and Ricci, S. 2008. Animal models of Streptococcus pneumoniaedisease. Clin Microbiol Rev. 21, 666–685. 4. Craig, W. A. (1998) Pharmacokinetic/ pharmacodynamic parameters: rationale for antibacterial dosing of mice and men. Clin Infect Dis. 26, 1–10. 5. Moine, P., Vall´ee, E., Azoulay-Dupuis, E., Bourget, P., B´edos, J. P., Bauchet, J.,
and Pocidalo, J. J. (1994) In vivo efficacy of a broad-spectrum cephalosporin, ceftriaxone, against penicillin-susceptible and -resistant strains of Streptococcus pneumoniae in a mouse pneumonia model. Antimicrob Agents Chemother. 38, 1953– 1958. 6. Steinhoff, M.C. (2007) Pneumococcal vaccine animal model consensus group. Animal models for protein pneumococcal vaccine evaluation: a summary, Vaccine 25, 2465– 2470. 7. Kadioglu, A. and Andrew, P. W. (2005) Susceptibility and resistance to pneumococcal disease in mice. Brief Funct Genomic Proteomic. 4, 241–247.
Chapter 23 Murine Model of Polymicrobial Septic Peritonitis Using Cecal Ligation and Puncture (CLP) Eva Medina Abstract Although a number of animal models such as endotoxic shock and bacteremia have been used to study the pathogenesis of sepsis, cecal ligation and puncture (CLP) represents a peritonitis model with clinical features of polymicrobial infection comparable with those of peritonitis in humans. The CLP consists in the surgical perforation of the legated cecum of mice that results in immediate and constant drainage of cecal bacteria into the peritoneal cavity. The severity of the diseases depends on the diameter of the needle used for the perforation as well as on the number of cecal punctures. The CLP model of sepsis in mice is the most commonly used for studying the process of septic peritonitis and can be used as a preclinical model to test the efficacy of pharmacological agents for the treatment of sepsis. Key words: Septic peritonitis, cecal ligation and puncture, polymicrobial sepsis.
1. Introduction Sepsis is a complex clinical syndrome characterized by a severe infection in the body and bloodstream and patients with septic peritonitis have a particular high mortality rate of 60–80% (1, 2). Bacterial invasion of the peritoneal cavity due to intestinal leakage after major abdominal surgery is the most frequent cause of septic peritonitis. Due to the large number of microorganisms in the bowel, this infection is by nature polymicrobial. Septic peritonitis is characterized by massive infiltration of neutrophils and macrophages into the peritoneum where these cells are the first line of defense for clearing invading microorganisms. However, once they fail to restrict microbes to the peritoneal cavity, microbes may reach the blood stream, resulting in an G. Proetzel, M.V. Wiles (eds.), Mouse Models for Drug Discovery, Methods in Molecular Biology 602, DOI 10.1007/978-1-60761-058-8 23, © Humana Press, a part of Springer Science+Business Media, LLC 2010
411
412
Medina
overwhelming systemic immune response via the production of proinflammatory mediators such as cytokines frequently leading to multi-organ failure, septic shock, and death (2). Although several mouse models of sepsis have been developed such as endotoxic shock or bacteremia (3, 4), the cecal and ligation puncture (CLP) model mirrors more closely the clinical course of human abdominal sepsis (5). The CLP surgery model takes into account the impact of pathogenic infection typically observed in sepsis since it involves trauma (perforation) to the bowel, thereby permitting the introduction of multiple bacterial strains into the peritoneal cavity. Furthermore, the magnitude of the septic challenge can be controlled by changing the size of the needle used to puncture the ligated cecum. The mouse peritonitis model has been extensively used for the evaluation of numerous antimicrobial compounds and it has been pivotal for investigating fundamental issues in the relationship between infecting pathogen and antibiotic therapy.
2. Materials 1. Columbia blood agar (Beckton Dickinson, MD, USA). 2. Surgical staples. 3. 3-0 silk suture. 4. 4-0 braided absorbable suture. R , Ravensburg, Germany). 5. Ketamine (Narketan R , Leverkusen, Germany). 6. Xylazine (Rompun
7. Mice (see Notes 1 and 2): any variety of mice can be used. Pathogen-free CD1 and NMRI mice from either sex have also been commonly used. As mentioned above, we recommend the use of inbred mouse strains since they are genetically well defined and uniform providing the advantage of high reproducibility not available with standard outbred mouse strains. In addition, due to the phenotypic uniformity of an inbred mouse strain, the sample size can be smaller than when using outbred stocks. 8. Wright–Giemsa stain. 9. Cytospin apparatus (Shandon, Runcorn, UK). 10. Neubauer chamber.
3. Methods 3.1. Surgical Procedure
1. Anesthetize the animal by intraperitoneal injection of a mixture of ketamine (80 mg/kg/body weight) and xylazine (16 mg/kg/body weight) in 0.2 ml of sterile PBS.
Murine Model of Polymicrobial Septic Peritonitis
413
2. Place anesthetized animal on a clean paper towel, shave the anterior abdominal wall, and disinfect with 70% ethanol. 3. Make an approximately 1 cm longitudinal incision in the left-lower quadrant of the abdomen using a scalpel to open the peritoneal cavity (see Note 3). 4. Bring out the cecum (see Note 4). 5. Ligate one third of the distal cecum with a 3-0 silk immediately below the ileocecal valve. 6. The ligated cecum is then punctured once or twice using a hypodermic needle (see Note 5). 7. Press gently on the tied segment to ensure that a small amount of feces is extruded on to the surface of the bowel. 8. Return the cecum to the peritoneal cavity. 9. Close the wound using 4-0 braided absorbable suture for the muscle layer and the skin with surgical staples. 10. In sham controls, the cecum is exposed but not ligated or punctured, then returned to the abdominal cavity. 11. Administer 1 ml of sterile PBS intraperitoneally as a fluid resuscitation measure, immediately following surgery and place the animals on a heating pad until they recover from the anesthetic. 3.2. Quantification of Bacterial Loads in the Peritoneal Cavity
1. Euthanize the mice by CO2 asphyxiation (see Note 6). 2. Cut open the skin of the abdomen in the midline after thorough disinfection and without injury to the muscle. 3. Using a sterile 5 ml syringe and 20 G needle, inject 5 ml of sterile PBS into, and aspirate out of, the peritoneal cavity twice. 4. Make 10-fold serial dilutions of peritoneal lavage samples and plate on Columbia blood agar. 5. Incubate plates for 18 h at 37◦ C. 6. Count colonies and express as CFU/ml of lavage fluid.
3.3. Quantification of Bacterial Loads in the Blood
1. Euthanize the mice by CO2 asphyxiation. 2. Collect blood by cardiac puncture. 3. Plate blood immediately onto Columbia blood agar after making 10-fold serial dilutions using sterile PBS. 4. Count colonies and express as CFU/ml of blood: CFU/ml =
C Vp × DF
where C is the colony counts, VP is the volume plated, and DF is the dilution factor of the sample.
414
Medina
3.4. Quantification of Bacterial Loads in Systemic Organs
1. Euthanize the mice by CO2 asphyxiation. 2. Remove carefully the spleen, liver, kidneys, and lungs and place them in a 10-ml homogenization tube containing 4.5 ml of sterile saline. 3. Homogenize organs and perform 10-fold serial dilutions of tissue homogenate by mixing 100 l of tissue homogenate and 900 l of sterile PBS. 4. Apply 100 l of each dilution to a Columbia blood agar plate and gently swirl the plate. 5. Incubate plates at 37◦ C for 24 h. 6. Determine the bacterial load within the organ using the following formula:
CFU/organ =
C × Vt Vp × DF
where C is the colony counts, Vt is the total volume of the organ homogenate, VP is the volume plated, and DF is the dilution factor of the sample. 3.5. Quantification of Cytokines and Chemokines in the Peritoneal Cavity
1. Perform peritoneal lavage as described for the quantification of bacterial loads in the peritoneal cavity. 2. Clarify peritoneal fluids by centrifugation for 10 min at 1000 rpm. 3. Remove the supernatant and determine the level of cytokines and chemokines by ELISA (see Note 7).
3.6. Determination of Peritoneal Cell Counts
1. Perform peritoneal lavage as described for the quantification of bacterial loads in the peritoneal cavity. 2. Determine total leukocyte counts under optical microscope using a Neubauer chamber. 3. Determine differential cell counts on cytospin smears of peritoneal fluids (100×g for 3 min) stained by the Wright– Giemsa method.
4. Notes 1. The animals should be fasted overnight prior to surgery. 2. All protocols using live animals must first be reviewed and approved by the corresponding Governmental and Institutional Animal Care and Use Committee and must conform
Murine Model of Polymicrobial Septic Peritonitis
415
to the regulations regarding the care and use of laboratory animals. 3. There are no major blood vessels in the region, but care should be taken to avoid any muscular damage, a source of subcutaneous hematoma. 4. The mouse cecum is located on the left lateral side of the abdomen, curved on itself, and filling the space between the stomach and the liver lobes. 5. The gauge of the needle can vary from 18 to 26 G, depending on the desired mortality to be induced. 6. The number of mice per group should be adequate to allow statistical analysis of the data. Because of the phenotypic uniformity, experimental infections using inbred mouse strains will require smaller number of animals to reach statistical significance than those employing outbread strains. 7. The simplest method for determination of cytokines by ELISA is to use commercially available kits with matching antibody pairs and standards for the specific cytokine. The reader can consult web sites such as http://www. biocompare.com/ to select from a wide range of commercially available kits with the corresponding manufactures instructions for use. References 1. Cohen, J. (2002) The immunopathogenesis of sepsis. Nature 420, 885–891. 2. Hotchkiss, R. S. and Karl, I. E. (2003) The pathophysiology and treatment of sepsis. N Engl J Med 348, 138–150. 3. Fink, M. P. and Heard S. O. (1990) Laboratory models of sepsis and septic shock. J Surg Res 49, 186–196.
4. Deitch, E. A. (1998) Animal models of sepsis and shock: a review and lessons learned. Skock 9, 1–11. 5. Wang, P. and Chaudry, I. H. (1998) A single model of polymicrobial sepsis: cecal ligation and puncture. Sepsis 2, 227– 233.
INDEX
A
serum, 94–95, 97, 99–100, 140, 142, 147, 154, 189, 257, 400 tail vein, 103, 142–144, 146, 217, 220–221, 259–260, 400 Body composition, 46–48, 140, 149–151 Bone marrow, 16–17, 50, 108–111, 128, 217, 253, 255–257, 259–260, 262, 264–265 Bowel, 411–413 Brain, 40–41, 48, 72, 81, 87–89, 275, 283–298, 299–319, 324–325, 327, 333 Brain imaging, 87–89 Burying test, 303, 307–308, 315
Acyl coenzyme A:cholesterol acyltransferase (ACAT), 158–159 Adipocytes, 136, 150 Albumin pharmacokinetics, 102 Alopecia areata, 194, 196, 208 Amyotropic Lateral Sclerosis (ALS), 348–352 Alzheimer’s disease, 40, 51, 324, 326 Amyloid plaques, 40, 325–326 Anesthesia, 81, 83, 87, 143, 149, 201–202, 245–247, 250–251 Angiogenesis, 218, 228–230 Anhedonia, 267, 272, 274 ANOVA, 14–15, 249, 272, 305–306, 335, 337 Antibody, 46, 50, 68, 93–94, 96–97, 102, 128, 184, 191, 197, 204, 261, 364, 402, 415 See also Therapeutic antibody, monoclonal antibodies Antidepressant, 267–279, 301, 305, 311 Anxiety, 268, 271–272, 276, 280, 299–319, 332, 336–337 Anxiogenic drugs, 306–310, 312, 316–317 Anxiolytic drugs, 306–307, 311, 318 Apolipoprotein E, 41–42 Apoptosis, 203–205, 230, 254 Arthritis, 41–42, 46, 68, 181–192 Atherosclerosis, 41–42, 158, 172 Atopic dermatitis, 195 Autoantigen, 125–127, 129 Autochthonous tumors, 226 Autoimmune, 9, 106, 119–130, 182, 194, 208, 349, 356 Autoimmunity, see Autoimmune
C C57BL/6, 5, 26, 98, 111, 143, 170, 182–183, 187, 191, 260, 315, 397, 409 Cancer BCR-ABL, 253, 264 chemotherapeutic, 228, 237 chronic myeloid leukemia, 253–265 colorectal cancer, 50, 235–251 5-fluorouracil, 235–251 irinotecan, 235–251 leucovorin, 235–251 leukemia, 253, 264 lymph node metastasis, 236 melanoma, 216 metastasis, 215–231, 236 metastatic colorectal cancer, 235–251 oncogene, 56, 225, 253 Philadelphia chromosome, 253 staging, 236 Trp53 knockout mouse, 9 tumor, 215, 216, 225–226, 228, 235–236 growth measurement, 247 model, 215, 225–228 xenograft, 226, 235–236 Candidate gene, 50 Carboxyl ester lipase (CEL), 158 Cardiac hypertrophy, 40 Cardiac puncture, 143, 400, 413 Cardiovascular disease (CVD), 157–159, 161, 172, 175 CD-1, 4, 12–17, 19 Cecal ligation and puncture, 411–415 CEL, see Carboxyl ester lipase (CEL) Cellulitis, 396 Charcot-Marie-Tooth diseases, 349, 353 Chimera, 50–51, 128 Chloramphenicol, 12–13, 15–16, 19 Cholesterol, 46–47, 141, 157–176 Chronic myeloid leukemia (CML), 253–265
B BALB/c, 12, 14, 16, 19, 107–108, 250, 255, 268, 397, 401, 407, 409 Barbering, 278 BCR-ABL, 253–255, 257–263, 265 Behavior, 46–47, 49, 152, 231, 271, 274–275, 277–279, 300, 305, 307, 316, 333 Behavioral response, 312 Behavioral test, 378 Beta-amyloid peptide, 324, 326 Beta-galactosidase, 69–70 Beta-2-microglobulin, 106 Bile acids, 164, 171, 173–175 Bioluminescence, 80, 86–87, 90, 228, 235–251 Biomarker, 10–11, 49 Blister, 199 Blood collection, 97, 103, 140–144, 147–150, 400 cardiac puncture, 143, 400, 413 pressure, 47, 49, 141, 333
G. Proetzel, M.V. Wiles (eds.), Mouse Models for Drug Discovery, Methods in Molecular Biology 602, c Humana Press, a part of Springer Science+Business Media, LLC 2010 DOI 10.1007/978-1-60761-058-8,
417
MOUSE MODELS FOR DRUG DISCOVERY
418 Index
Chronic proliferative dermatitis, 204–205, 208 Chronic stress, 277 CIA, see Collagen-induced arthritis (CIA) Clinical score, 188, 191 CML, see Chronic myeloid leukemia (CML) CNV, see Copy Number Variation (CNV) Collagen, 41–42, 46–47, 181–192, 359–360 Collagen-induced arthritis (CIA), 41–42, 181–191 Colorectal cancer, 50, 235–251 Conditional gene targeting, 61–63, 68, 72 Conditional knockout, 4 Congenital Myasthenic Syndromes, 349 Copy Number Variation (CNV), 25 Cre-ER, 43, 63 Cre-ERT2, 63 Cre/loxp, 40, 42, 61–63, 65, 67, 72, 362, 363 Cre recombinase, 40, 62 Cryopreservation embryos, 26–29 sperm, 26–29, 31–34 Cryoprotective, 29, 31 CVD, see Cardiovascular disease (CVD) Cytokine, 46, 50, 107, 111, 120, 182, 190, 208, 259, 400–402, 409–410, 412, 415
D Data analysis, 91, 169, 174, 269, 272, 305–306, 334–335, 380 DBA, 41–42, 45, 137, 182, 184–185, 187–188, 190–191, 279, 284, 401, 409 Deficiency, 109, 113, 278, 335, 352, 356–357, 361 Dementia, 324–327, 330, 338 Demyelination, 349, 352, 355, 369 Depression, 267–280, 301, 315, 325 Dermatitis, 195, 204–205, 208 Dermis, 199, 363, 395–396 Despair, 271, 274, 276 DEXA, 46, 149 Diabetes, 40, 46, 48, 115, 119–130, 135–154 See also Idd; Type 1 diabetes (T1D); Type II diabetes mellitus (T2DM) Diet chow, 165–166 high fat, 46–47, 49, 141, 148, 151, 153 powder, 166 repelleted, 166 semi-purified, 163, 165, 167, 170, 174, 176 western diet, 158 Diet-induced obesity (DIO), 137–138 Dose-response, 12, 14–15, 84–86, 200 Drug addiction, 284 administration, 198, 271–272, 304–305, 311 development, 1–19, 65, 159, 175, 181–182 discovery, 37–52, 55–72, 135–154, 162, 215–231, 268, 299–318, 338 Dual-label, 168, 170 Duchenne’s disease, 349, 359
E Electrophysiology, 371, 381 Elevated-Plus Maze (EPM), 302, 306, 314 ELISA, 97–101, 140, 144–145, 184, 190–192, 286, 401–402, 409, 415 Embryo, 6, 26–28, 30, 34, 59, 123
Embryonic stem (ES) cells, 51, 59, 107, 225 Epidermis, 199–200, 203–204, 395–396 Epilepsy, 48, 352
F Fluorescent-activated cell sorting (FACS), 46, 97, 99, 257–258, 260–265 Fasting, 136, 146, 148–149, 154, 166, 170 Fat, 46–47, 49, 137–138, 149, 151, 171, 200, 374, 396 Fat pad, 224–225 Fatty acids, 46, 141, 145–146, 171 Fc fusion proteins, 96 FcRn, see Neonatal Fc receptor (FcRn) FDA, 2, 19, 196 Fear, 306, 331, 333 Fecal sterols, 164 Fibrosis, 40, 375 Filaggrin, 195 Fixative, 197, 207–208, 227, 362, 364, 366, 373, 375, 385 Flow cytometry, 109–110, 227, 257–258, 261–264 Flp/Frt, 61–63, 65–66, 126, 353, 355 Flp recombinase, 62 See also Flp/Frt Fluorescence, 372, 376 Fluorescent-activated cell sorting (FACS), 46, 97, 99, 257–258, 260, 262, 264–265 Food intake, 46–47, 136, 146, 152–153 Footpad, 222–223 Forced Swim Test (FST), 269–271, 274–275, 278 5-FU (5-fluorouracil), 235–251, 257 Full thickness skin grafts, 196, 208
G Gait analysis, 378–381 Gene replacement, 43, 50 See also Gene targeting; Knockout; Knock-in (KI) Gene targeting, 59–71, 72, 98, 157 GEMMs/GEMs, 37–52, 55–73, 216, 225–226, 230 Genetically engineered mouse models, see GEMMs/GEMs Genetically modified mice, 38 Genetic background, 4, 25, 47, 51, 98, 162, 216, 226, 264, 314, 328, 337, 350 Genetic contamination, 25, 328 Genetic disorder, 72, 106, 255, 299–300, 327, 385 Genetic drift, 5, 6, 25–26, 328 Gene trap, 69–70 Genome, 3, 4, 24, 26–28, 38–39, 41, 51, 55–73, 107, 268, 276–277, 279, 310, 317, 331, 385 Genotyping, 39, 47, 70, 99 GFP, see Green fluorescent protein (GFP) Glucose, 30, 40, 46, 136, 139–140, 144, 146–147, 154 Glucose-stimulated insulin secretion (GSIS), 139–140, 147–148 Glucose tolerance test (GTT), 46, 136, 146–147, 153–154 Graft, 105–115 Graft-versus-host disease (GVHD), 105–115 Green fluorescent protein (GFP), 68, 257–258, 262–263, 284, 362–363, 366 Grip strength, 378–379 Grooming Analysis, 303, 308–309 GVHD, see Graft-versus-host disease (GVHD)
MOUSE MODELS FOR DRUG DISCOVERY 419 Index H Hair, 195, 197, 202, 206, 208, 224, 279, 285 Hair cycle, 195, 200–201, 203, 205, 207 Hawthorne Effect, 334 HDL, see High-density lipoprotein (HDL) Hearing loss, 329 Heart failure, 40 Heart rate, 47–48 Hematopoietic stem cell (HSC), 106, 108, 128–129 Hepatocytes, 50–51, 112, 136, 150 Hereditary Motor and/or Sensory Neuropathies (HSMNS), 349 High-density lipoprotein (HDL), 158, 161–162, 167, 175 Hippocampus, 88, 325, 330–333 Histology, 112, 140–141, 150–151, 154, 226, 230, 362, 367, 374, 410 Histopathology, 46, 56, 197, 202–207, 218, 228, 260 HLA transgenics, 119–130 Hole Board Test, 304, 312, 318 Homologous recombination, 38, 41, 51, 60–61, 65, 68–71, 120–121, 157 See also Gene targeting Housing, 103, 143, 153, 269, 277, 300–301, 312, 314, 334, 336 See also Husbandry HPRT1, see Hypoxanthine phosphoribosyltransferase 1 (HPRT1) HSC, see Hematopoietic stem cell (HSC) HSMNS, see Hereditary Motor and/or Sensory Neuropathies Humanized mice, 51, 68–69, 108 See also Humanized knock-in mice Husbandry, 103, 269, 314–315, 336–337 Hyperglycemia, 137–138 Hyperinsulinemia, 138, 151 Hyperleptinemia, 137 Hypoxanthine phosphoribosyltransferase 1 (HPRT1), 69
I Ichthyosis, 194–195 Idd, see Insulin-dependent diabetes (Idd) IL2 receptor gamma chain, 107 Imaging, 3, 68, 79–92, 149, 228, 235–251, 325 Immunization, 114 Immunocompromised, 50 Immunodeficient, 51, 102, 106–107, 113–114, 196, 235–236, 254 Implant, 27–28 Implantation, 227–228, 236, 245–248 Inbred mouse, 9–10, 24, 39, 41, 176, 255, 264, 278, 317, 329, 397, 402, 415 strain, 4–6, 8, 12–13, 19, 25, 296 Inducible, 40, 42–44, 59, 63–65, 70, 72, 363 Infection bacteria, 395 bacterial load, 399, 405, 408, 413–414 cutaneous, 395–402 lentivirus, 58–59, 71–72, 284, 286 See also Lentiviral local, 396 lung, 399, 406, 408 peritonitis, 411–415 pneumonia, 405–410 polymicrobial, 411–415
respiratory, 405 respiratory tract, 405 retrovirus, 58, 69, 240–241, 255 sepsis, 405, 411–412 skin, 396–397 soft-tissue, 395–396, 399 Staphylococcus aureus, 396 Streptococcus pyogenes, 395–402 viral, 70, 193, 284 Infectious disease, 193 Inflammation, 41, 46, 54, 141 Inflammatory disease, 181 Inhibitor, 47–49, 159–161, 261–263 Injection, 28, 33, 39, 56, 68, 70, 80, 83–84, 87, 147, 162, 199, 219–221, 227, 246, 251, 271, 295, 297, 333, 398, 412 Inoculation, 396, 399, 406, 408 Insulin, 29, 40, 120, 125, 127–128, 136, 141, 144–145, 147–148, 152, 154 Insulin-dependent diabetes (Idd), 120, 122 Insulin secretion, 136–137, 139–140, 147–148 Insulin sensitivity, 136–137, 148–149 Insulin tolerance test (ITT), 148, 154 IRES, see Internal ribosome entry site (IRES) Internal ribosome entry site (IRES), 66–68 Intestine, 48, 150, 159–160, 167, 172, 174–175 Intranasal inoculation, 406 In vitro fertilization (IVF), 26–27, 32–35 In vivo imaging, 68, 80–81, 83–86, 91, 238 Irinotecan, 235–251 Isogenic strain, 7–8, 18 IVF, see In vitro fertilization (IVF)
K Keratinocytes, 203–205 Knock-in (KI), 38, 43–44, 51, 65–68, 225, 351, 354, 357, 359, 361 Knock-down, 43, 52, 71–72 Knockout, 3, 9, 60–61, 63–65, 105–115, 122, 158–160, 182, 284, 329, 353–354
L LDL, 157–159, 161–162, 172, 175, 356 Learning, 275–276, 314, 329–330, 332–333, 335–337 Lentiviral, 57–59, 70–72, 284 Leptin, 46, 48, 137 Leucovorin, 235–251 Leukemia, 106, 114, 254, 260–261, 263–264 Light-Dark Box Test, 303, 311, 318 Lipopolysaccharides (LPS), 46 Lipoprotein, 41, 159, 162, 175–176 Liver, 50, 87, 89, 112, 145, 154, 158, 162, 167, 175, 229, 399, 415 Locomotor, 305, 332, 335 LPS, see Lipopolysaccharides (LPS) Luciferase, 79–82, 86–87, 90, 237, 239–242, 251 Lung, 218, 221–222, 262, 405, 408–409 Lymph node, 222, 224, 236, 249, 250, 377 Lymphoma, 111–112
M Mab, see Monoclonal antibodies (Mab) Magnetic Resonance Imaging, 149 Melanoma, 216, 221–222
MOUSE MODELS FOR DRUG DISCOVERY
420 Index
Memory, 275–276, 315, 317, 324, 330–331, 335, 337 Metastasis, 215–231, 236, 248, 250 Metastatic colorectal cancer, 235–251 MHC class I, 94, 108–114, 120–121 MHC class II, 108–109, 111–121 Microscopy, 197, 206–207, 364, 366, 370–374, 408 Monoclonal antibodies (Mab), 96–97, 101, 103 Motor neuron, 348–349, 351, 353, 362–363, 366–368 Motor neuron diseases, 348–349, 364–368, 378 Muscle, 46, 51, 199, 203, 259, 348–349, 351–352, 360, 363, 376, 381, 383, 413 Muscular dystrophies, 349, 374–378 Mutation, 4, 5, 60, 107, 129, 202, 208, 328, 334, 350, 370, 374, 382 Myasthenias, 349, 371–374, 378
N NEFA, see Nonesterified free fatty acids (NEFA) Neonatal Fc receptor ( FcRn), 93–103 Neuroanatomy, 330 Neurodegenerative disease, 327, 332 Neuromuscular disease, 347–385 Neuromuscular junction, 349, 373 Neuropathic pain, 48, 355 Neuropathology, 325–327 NOD, 105–115, 120, 122–130, 254 NOD-scid IL2rγ null , 107–108, 109–114, 128 NONcNZO10/LtJ, 138 Nonesterified free fatty acids (NEFA), 140–141, 144–146 Novel Object Test, 304, 313 Nu/Nu mice, 250
O Obesity, 46, 48, 137–138, 160, 172, 336 Oncogene, 56, 225, 253–256 Oocyte, 26–27, 33–35, 56–57, 123, 328 Open Field, 46, 302, 304, 306–307, 312, 332 Orthotopic tumors, 225 Osmotic pumps, 142, 196, 199–200, 272, 304 Outbred, 2, 4–8, 10, 15, 18, 250, 264, 314, 328, 397 Ovary, 27, 33
P Pancreas, 136, 150–151 Pathogenesis, 106, 108, 114, 122, 396 Pathology, 46, 56, 196, 202–207, 260, 325–327, 338, 350, 359–360, 374, 385 Paw swelling, 182, 187–188 PCR, see Polymerase chain reaction (PCR) Peripheral neuropathies, 348, 368–370, 385 Peritonitis, 411–415 Pharmacodynamics, 86, 159 Pharmacokinetics (PK), 91, 96, 101–103, 152, 159, 231, 396 Phenotype, 4–5, 38, 40, 42, 44, 47, 51, 62, 122, 153, 205, 277, 300, 314, 316, 325, 328, 349, 353, 358–360, 370, 379, 384 Phenotyping, 3, 38, 44–49, 276, 328, 332, 338 Philadelphia chromosome (Ph+chromosome), 254, 259–261, 267 Phytosterols, 16, 169, 172–173, 176 Pneumonia, 405–410 Polymerase chain reaction (PCR), 39, 61, 99 Pooling fallacy, 334
PPAR-γ , 136, 141 PPRE-Luc, 80, 84 Preclinical, 51, 56, 59, 65, 68, 89, 96, 98, 102, 194, 228–229, 235–236 Presenilin 1, 326 scid Prkdc , 106, 109 See also Scid Promoter, 40, 56–57, 66–67, 70–71, 329, 362–363 Psoriasis, 194, 208 Pulmonary tumor, 221–222
R Rat, 2–3, 6, 48, 96, 170, 184, 271, 280, 304, 312–313, 318 Rat Exposure Test, 304, 312–313 Regenerative medicine, 51 Replacement, 38, 43, 50–51, 96, 328–329 See also Gene replacement Reporter gene, 69, 79 See also Beta-galactosidase; Luciferase; Green fluorescent protein (GFP) Reporter mouse, 79–92 Respiratory infection, 405–406 See also Infection Retinal degeneration, 202, 329 Retroviral, 58–59, 228, 237, 239–240, 250, 255–256, 259, 264 Retrovirus, 240–241, 255, 260, 265 Reverse cholesterol transport, 161–162, 172–175 Rheumatoid arthritis (RA), 181 RNA interference (RNAi), 52, 71–73, 283–298 ROSA26, 67, 69–71 Rosiglitazone, 136, 138–139, 141–142, 152
S Scid, 103, 105–115, 128–129, 227, 254 Screening, 1–20, 38, 61, 80, 85–86, 92, 193–209, 271, 274, 300, 315, 334 Selection Cassette, 40, 60–63 See also Selection Marker Selection Marker antibiotic, 40 diphtheria toxin, 61–62 ganciclovir, 62 hygromycin, 62 negative, 61–62, 128 neomycin, 61, 241 positive, 61–62 puromycin, 62 thymidine kinase, 62 Sepsis, 405, 412 Septic peritonitis, 411–415 Serum half-life, 95, 100 ShRNA, 71–73, 284, 287 Signal/noise ratio, 10, 16–18 SiRNA, 71–72, 142 Skin, 9, 152, 186, 189, 193–209, 225, 279, 395, 397, 399 Skin disorder, 193–209 Skin infection, 396–397 Social Interaction Test, 303, 309–310, 317 Soft tissue infection, 395, 396, 399 Sperm, 26–29, 31–35 Spinal cord, 347–348, 364–366, 368 Spinal Muscular Atrophy (SMA), 348, 350–351, 355 Spontaneous metastasis, 216–218, 222–230, 236
MOUSE MODELS FOR DRUG DISCOVERY 421 Index Staphylococcus aureus, 396 Startle Response, 302–303, 309, 316 Statin, 157 Statistics, 19, 335 Stem cell, 26, 28, 51–52, 59–71, 106–108, 114, 225, 253, 255, 257, 262–263 See also Embryonic stem (ES) cells; Hematopoietic stem cell (HSC) Streptococcus pneumoniae, 405 Streptococcus pyogenes, 395–402 Streptozotocin (STZ), 138 Stress-Induced Hyperthermia (SIH), 304, 311–312 STZ, see Streptozotocin (STZ) Subcutaneous injection (s.c.), 199 Subcutaneous inoculation, 396 Subcutaneous tissue, 395–396 Sucrose, 152, 171, 269, 272–273, 275–277 Suok Test, 303, 310–311 Superovulation, 26, 29–30, 33, 35, 130 Surgery, 224–225, 246, 285–287, 289–296, 411, 413–414 Survival curves, 399
Therapeutic antibody, monoclonal antibodies, 68, 96 Thymus, 50, 112, 121 Tissue collection, 140–141, 149–151 Toxicity testing, 2, 6–12, 19 Transgene, 39–40, 57–58, 60, 66, 69–71, 98, 122, 124–227, 255, 328–330, 332, 363 Transgenic, 9, 39–40, 51, 56–59, 66, 70, 93–103, 119–130, 142, 226, 255, 268, 284, 334, 350 Translational research, 253–265, 327 Transplantation, 51, 106, 229, 253, 255, 259–260, 263, 265 Trp53 knockout mouse, 9 Tumor cells, 217–218, 223, 227–228, 231, 236, 246 Tumor Growth Measurement, 247 Tumor models, 225–230 Type 1 diabetes (T1D), 119–122, 124, 127, 129–130 Type II diabetes mellitus (T2DM), 136–138, 141, 147, 150–151
T
W
Tail Suspension Test (TST), 270–271, 274, 278 Tamoxifen, 43, 90 Tauopathy, 325–327, 338 T cells, 46, 63, 106, 108, 113, 120–121, 125–128, 182 T1D, see Type 1 diabetes (T1D) T2DM, see Type II diabetes mellitus (T2DM) Tetracycline, 40, 72, 226 Tet-ON system, 72 See also Tetracycline
Welfare, 152, 201, 301
V Virus, see Retrovirus
X Xenograft, 51, 226, 229, 235, 254
Y YAC, see Yeast artificial chromosomes (YAC) Yeast artificial chromosomes (YAC), 57