Discovery Strategies 2002
Integrated Approaches to Unraveling Complex Disease:
Phenotype to Genotype and Back Again
Conference Overview:
Availability of genomic sequence for human and model organisms coupled with the broad application of tools for manipulating the genome of model systems has provided a wealth of potential drug targets. The challenge now is in understanding the role each of these targets plays in complex disease processes. This meeting provides an opportunity to explore integrated, in vivo and in silico strategies for modeling and characterizing disease phenotypes in support of genomics-based drug discovery and development.
Hosted by: The Research Affiliates Program of The Jackson Laboratory
Location: The Portland Regency Hotel, Historic Old Port District, Portland, Maine
Speaker List:
|
Alan D. Attie |
Genetics and Gene Expression Studies of Obesity and Diabetes in Mice |
|
Yaacov Barak |
Customized Genetic Approaches to Pleiotropic Functions of Disease Genes - PPARs as a Model |
|
David Bergstrom |
Radiation-Induced Chromosomal Deletions as Genome Analysis Tools |
|
Molly Bogue |
Inbred Mouse Strains Revitalized: Sharpening a Classical Genetics Tool to Dissect Complex Traits |
|
Sydney Brenner |
Cell Map: A Project to Understand Function |
|
Steve Brown |
Mutagenisis and Genomics in the Mouse: Towards Systematic Alalysis of Mammalian Gene Function |
|
Gary Churchill |
Complex Trait Analysis: Prospects for a Predictive Genetics |
|
Ron Korstanje |
From QTL to Gene: Atherosclerosis and HDL cholesterol |
|
Jeremy M. Levin |
Integrating modeling and experimentation to understand biological complexity |
|
A. Jake Lusis |
Genome-wide Congenic Strains for Analysis of Complex Traits |
|
Joe Nadeau |
The Genetics of Health: Modifier Genes, Sensitized Crosses, and Complex Traits and Systems |
|
Russell Phillips |
From the Genome to Drug Targets via High Throughput Gene Knockouts and Systematic Phenotyping |
|
Kevin Seburn |
High-throughput Phenotyping: Screening Mice for Neurological Mutations in the JAX-Neuroscience Mutagenesis Facility |
|
Amanda Ewart Toland |
Cross-species Approaches to the Discovery of Cancer Susceptibility Genes |
Genetics and Gene Expression Studies of Obesity and Diabetes in Mice
Alan D. Attie
University of Wisconsin-Madison
Approximately 80% of people with type 2 diabetes mellitus (DM2) are obese. Yet, most obese people do not develop diabetes. The heritability of DM2 has been estimated to be about 50%. Yet, we do not know what genes interact with obesity to produce DM2. In addition, we do not have molecular markers to aid in predicting who among obese individuals is most likely to develop DM2.
We have exploited strain differences in diabetes susceptibility between C57BL/6 and BTBR mice to map several diabetes QTL. Several of these are at an advanced stage in positional cloning through the interval-specific congenic strain approach. In addition, we have carried out expression profile experiments that have revealed differences in gene expression that correlate with diabetes susceptibility.
We have combined our expression profiling work with genetics to treat mRNA abundance as a mappable QTL. In a small-scale study, we mapped the mRNA abundance of a group of genes that emerged from our gene expression study. We carried out a cluster analysis and computed principal components from this group of genes. Both of these data reduction approaches yielded a composite trait that yielded highly significant genetic linkages. This experiment can theoretically be scaled to a much larger collection of mRNAs.
Customized Genetic Approaches to Pleiotropic Functions of Disease Genes - PPARs as a Model
Yaacov Barak
The Jackson Laboratory
The primary challenges in the post-genome era are the analysis of individual gene functions and charting gene networks. These challenges can be complex, as demonstrated by the nuclear hormone receptors PPAR-gamma and PPAR-delta (peroxisome proliferator-activated receptors gamma and delta). PPARs are fatty acid-activated transcription factors with multiple physiological functions and prime pharmaceutical importance. PPAR-gamma is the receptor for thiazolidinediones (TZDs), which are insulin sensitizers prescribed to millions of type II diabetics worldwide. PPAR-gamma and TZDs also induce adipogenesis, increase HDL-cholesterol levels, ameliorate atherosclerosis, and promote tumor cell differentiation in vitro and in vivo. Recently developed PPAR-delta ligands raise HDL levels and ameliorate type II diabetes in obese-diabetic macaques. Deficiencies for both PPAR-gamma and PPAR-delta are embryonic lethal due to placental defects. We turned this potential setback into an advantage by utilizing the placenta and trophoblast for modeling PPAR function and for target gene screening. A proof of principle will be presented, where the whole process from identification and validation of a novel PPAR-gamma target to phenotypic analysis of its contribution to PPAR-gamma function was completed successfully. Early embryonic lethality nevertheless raises the need for alternative strategies to study PPAR functions later in life. We study the adipogenic functions of PPAR-gamma using embryonic stem cell chimeras. These studies teach us that PPAR-gamma is essential for the expansion and differentiation, but not commitment of pre-adipocytes; that a novel developmental feedback mechanism ensures repopulation of PPAR-gamma null fat pads by wild type adipocytes; and that adipogenesis involves PPAR-gamma-regulated paracrine signaling between neighboring adipocytes. Tissue- specific knockouts, using CRE-lox methodology, enable us to analyze PPAR-gamma function in mature adipocytes, as well as to address its tissue-specific contributions to insulin sensitization. The unique insights provided by each of these approaches emphasize the indispensability of customized, data-driven research strategies for addressing specific methodological challenges posed by individual drug targets.
Inbred Mouse Strains Revitalized: Sharpening a Classical Genetics Tool to Dissect Complex Traits
Molly Bogue
The Jackson Laboratory
Multiple research approaches and new resources are required to reveal the full spectrum of genomic components and environmental factors that define a complex trait. Human genome-wide association mapping of complex diseases will be feasible when comprehensive sequence information is available and dense genome-wide SNPs have been identified. However, there are obstacles in human studies that cannot be overcome easily, including controlling heterozygosity and environment. These key factors can be controlled when using the laboratory mouse, a highly successful model organism for many human diseases and conditions.
In conjunction with contemporary and new technologies in mouse biology, enhanced classical approaches are essential to understanding complex traits. Inbred mouse strains have been used for decades in genetics research where hundreds of strains have been generated and described. Although inbred strains are obligate homozygotes with invariant genotypes, they are quite diverse phenotypically due to the assortment of genetic components inherited from the fortuitous admixture of multiple ancestral Mus species. Inbred strains are a powerful, enduring resource because they impart the unique opportunity to repeatedly access a genomically fixed population, bestow the confidence to reliably compare data across laboratories, provide the genetic background for many mutants, and they are commonly used in selective breeding strategies. Encompassing inbreds, mutants, and derivative strains gives extraordinary depth to the genetic resources available for complex trait analysis.
Haplotype association studies are pivotal for complex trait analysis. Unfortunately, most strains have not been adequately genotyped or systematically characterized for many phenotypic domains, making this a virtually untapped resource. To fill this gap and exploit the benefits of a completely sequenced mouse genome, the Mouse Phenome Project takes advantage of the enormous potential and remarkable utility of inbred strains and their derivatives. This project is an international collaborative effort to promote the quantitative phenotypic characterization of a defined set of strains under standardized conditions for a wide range of phenotypes of biomedical relevance. Phenotypic and SNPs data are stored in a central web-accessible database, the Mouse Phenome Database (MPD). The MPD serves as a repository for raw data and detailed protocols, and the website provides tools for data retrieval, query, and analysis. The MPD helps investigators identify genetic pathways and compensatory mechanisms and provides critical information for modeling disease processes. Together with other tools in the complex trait toolbox-genomics, genotype-driven technologies, and robust hybrid approaches-the MPD is an integral tool for identifying possible phenotype-genotype associations, exploring the influence of environment on phenotype, and detecting pleiotropism and epistasis.
Radiation-Induced Chromosomal Deletions as Genome Analysis Tools
David Bergstrom
The Jackson Laboratory
Geneticists have long used chromosomal deletions as powerful tools for studying the physical structure and functional content of complex genomes. In mouse, the availability of deletion complexes has traditionally been limited to those centered around a few visible loci as the result of cumbersome whole animal irradiation programs. Comprehensive sets of deletions collectively spanning entire genomes have been limited to simpler model organisms (e.g. Drosophila, C. elegans). The advent of modern embryonic stem (ES) cell techniques has now opened the door to the creation of radiation-induced chromosomal deletions at loci of interest throughout the mouse genome. Deletions created in culture can subsequently be used to create deletant mouse strains. These specialized strains can serve many purposes such as: (1) mapping tools for simple or complex traits, (2) substrates for mouse mutagenesis screens, (3) models of human contiguous gene deletion syndromes, (4) mapping reagents for imprinted or haploinsufficient loci, and (5) specialized strains for array analyses. Through the development of a resource known as DELBank, approximately 100 new deletion focal points have been randomly placed at an average spacing of every 10-15 cM throughout the mouse genome. Thus, investigators can begin to develop deletions around loci of interest immediately, without having to perform the laborious first step of targeting the locus. The generation and analysis of deletion complexes provides a powerful means to elucidate the functional content of complex mammalian genomes.
Mutagenesis and Genomics in the Mouse: Towards Systematic Analysis of Mammalian Gene Function
Steve Brown
MRC Mammalian Genetics Unit and UK Mouse
Systematic approaches to mouse mutagenesis are vital for future studies of gene function and the identification of new drug targets. We have undertaken a major ENU mutagenesis programme incorporating a large genome-wide screen for dominant mutations (Nolan et al. Nature Genetics 25: 440-443). Nearly 30,000 mice have been produced and the majority screened employing a systematic and semi-quantitative screening protocol - SHIRPA (Rogers et al. Mammalian Genome 8: 711-713). SHIRPA is a hierarchical screening protocol employing a rapid and efficient primary screen for deficits in muscle and lower motor neuron function, spinocerebellar function, sensory function, neuropsychiatric function and autonomic function. Moreover, in the primary screen blood is collected from mice and subjected to a comprehensive clinical chemistry analysis, and in addition more sophisticated neurobehavioural tests have been applied to a large fraction of the mutagenised progeny. Subsequently, secondary and tertiary screens of increasing complexity can be employed on animals demonstrating deficits in the primary screen. Progeny testing of mice carrying abnormal phenotypes indicates that 2% of mice from the screen carry a new heritable dominant phenotype. Over 150 mutants have been confirmed as heritable and added to the mouse mutant catalogue and, overall, we can extrapolate that we have recovered around 500 mutants from the screening programme. For further information on the project and details of data derived from the screening see: http://www.mgu.har.mrc.ac.uk. We are currently using frozen sperm and IVF for the rapid generation of small backcrosses in order to map many of the newly catalogued mutations to the mouse genome. We have mapped over 70 mutants to date and confirmed that many of the novel phenotypes represent mutations at previously unidentified loci in the mouse genome. We have also begun to develop gene-driven mutagenesis approaches using ENU (Coghill et al. Nature Genetics 30: 255-256). We have established parallel archives of DNA and frozen sperm from a large cohort of male progeny of mutagenised animals. Screening of the DNA archive allows us to identify mice carrying point mutations in our gene of interest. Subsequently, the mutant can be recovered from the sperm archive and phenotyped. The approach promises a rapid methodology to recovering an allelic series of point mutations for any gene in the mouse genome enabling a more profound analysis of gene function. The use of both phenotype-driven and gene-driven ENU mutagenesis for the generation of a new mutant map of the mouse will be a powerful resource available to the mouse and human genetics communities at large for future gene function studies.
From QTL to Gene: Atherosclerosis and HDL cholesterol
Ron Korstanje
The Jackson Laboratory
Our lab concentrates on quantitative trait loci (QTL) associated with cholesterol related diseases such as gallstones, atherosclerosis and hypercholesterolemia. Atherosclerosis is a difficult phenotype in the mouse and can be induced only in a few transgenic strains or in a small number of inbred strains by feeding a high fat diet containing cholic acid. So far, eleven QTL for atherosclerosis in the mouse have been reported by several groups. We are currently close to identifying the genes for two of these. The level of HDL cholesterol is a much easier phenotype. Many studies have reported HDL cholesterol QTL and some of these QTL are found in multiple crosses. The mouse HDL QTL map will be shown along with the QTL found for HDL levels in humans. Some QTL have obvious candidate genes whose role in cholesterol metabolism has already been established, such as Abca1, Srb1, and Apoa1. Our lab is attempting to identify genes involved in HDL levels for some QTL, focussing on those with human QTL in homologous regions of the chromosome. We are using a variety of techniques to narrow the region, several of which will be illustrated, and real-time PCR and microarrays to test expression levels of candidate genes. Although the process of identifying genes from QTL has been slow, novel genetic resources, such as the near complete human and mouse genome sequences, microarray technology, single nucleotide polymorphisms (SNP's), have accelerated it considerably. Not all of the many QTL found to date are of equal interest. Looking at concordance between QTL studies in different species will give us some clues on which QTL are of major interest and should be focused on.
Genome-wide Congenic Strains for Analysis of Complex Traits
A. Jake Lusis
UCLA School of Medicine
While congenic strains are clearly useful for the analysis of complex traits, they are expensive and time consuming to construct. We have developed two "libraries" of congenic mice with a set of introgressed segments covering all 19 autosomes and the X chromosome. Each of these libraries is comprised of more than 60 individual congenic strains, each carrying an introgressed segment representing about 2% of the genome. Both libraries use C57BL/6J as the background strain while the introgressed segments come from either DBA/2J or CAST/Ei, strains chosen because of the high level of previous QTL mapping carried out in CAST/Ei X C57BL/6 and DBA/2 X C57BL/6 crosses. Most of the DBA congenic strains are now established as homozygous breeding stocks, and the CAST congenic strains should be established within a year. Due to the small size of the introgressed chromosomal segment, these congenic strains will greatly speed positional cloning of genes underlying QTLs, as well as helping to resolve multiple QTLs that may be present on a chromosome. We also anticipate that extensive phenotyping of the full set of congenic strains will allow the direct mapping of QTLs for these phenotypes. Initial phenotype data is being collected for a panel of traits related to atherosclerosis, obesity, diabetes, and insulin resistance. These include body weight, length, bone density and adiposity as well as measures of plasma lipids, insulin and glucose. We have also begun collaborative efforts to determine strain-specific differences in gene expression using micro-arrays to examine in a variety of tissues including, liver, skeletal muscle, heart, adipose tissue and brain. We anticipate that these combined data will be very powerful in analyzing the genetic variation underlying complex traits and in revealing the pathways responsible for the resulting phenotypic variation.
From the Genome to Drug Targets via High Thoughput Gene Knockouts and Systematic Phenotyping
Russell Phillips
Deltagen
Genome sequencing efforts have identified thousands of genes with no known function representing potential new drug targets for the pharmaceutical industry. The challenge is to identify which targets have therapeutic potential early in the drug discovery process. Using Deltagen's mouse knockout technology, it is now possible to knockout hundreds of genes per year and to assign in vivo mammalian function to novel members of gene families representing drugable targets, including GPCRs, ion channels, nuclear hormone receptors, kinases, and phosphatases. A comprehensive systems biology approach integrating extensive phenotypic analysis, microarray expression studies, disease challenge models and pathway analysis has enabled the identification of key novel targets for the treatment of disease. Examples illustrating the discovery of novel drug targets using this systems biology approach in the areas of Inflammation and Diabetes will be presented.
High-throughput Phenotyping: Screening Mice for Neurological Mutations in the JAX-Neuroscience Mutagenesis Facility
Kevin Seburn
The Jackson Laboratory
The Jackson Laboratory Neuroscience Mutagenesis Facility (JAX-NMF) (http://www.jax.org/nmf/) is producing and making available mouse models of human CNS diseases using large-scale mutagenesis. The mutagen ENU is being used in genome-wide screens of C57BL/6J mice. Currently heritability has been confirmed for 11 potentially novel mutant lines in several different focus areas. Four lines are fully established and mice are available for distribution and an additional 27 phenotypic deviants are at various stages of heritability testing. High-throughput phenotype screens are underway in several focus areas including motor function, epilepsy, neural obesity, learning and memory, affective disorders, sensorimotor gating and vision. Lower throughput screens for mice with mutations affecting ingestive behavior and smell are also underway and are being refined. Results vary from screen to screen and some screens are still in their "ramp-up" phase, but overall 25% of the families (equivalent to one mutagenized genome) tested to date had at least one phenotypic deviant in one of the screens. Heritability has been proven in 39% of the lines attempted. Presentation and discussion will focus on two novel technologies that are expected to improve overall yields over the course of the next year and that have wide applicability. An automated live-in cage system (CLAMS รค, Columbus Instruments, Columbus Ohio), conceived as a multi-domain screen, is providing promising early results in deviant detection and will be described in detail. A second device for use in the detection of motor domain deviants was recently completed. High-throughput capture of digital video images allows for automated analysis of gait patterns in mice. Validation studies and preliminary screening results will be discussed. Supported by NIH grant NS41215 to W.N. Frankel.
Cross-species Approaches to the Discovery of Cancer Susceptibility Genes
Amanda Ewart Toland
Cancer Research Insitute, UCSF
One in three individuals are diagnosed with cancer in an average lifetime, whereas two out of three are not. Cancer risk in these individuals is a combination of environmental factors and low penetrance cancer modifier genes. An individual's susceptibility to cancer is therefore partly determined by the combination of inherited cancer susceptibility and resistance genes. Initial identification of these cancer risk genes has been done in the mouse because human low penetrance genes are extremely difficult to find using traditional methods and the choice of candidate genes for association studies are based on educated guesswork which can miss the unknown or unexpected. In this presentation, I will review our cross-species strategy that utilizes linkage and allelic specific tumor loss/gain in the mouse in combination with association studies and preferential allelic loss/gain in the human to identify cancer risk genes. Allelic variants in both mouse and human can then be used to better understand cancer risk and as targets for intervention.
