Discovery Strategies 2005

Improving the Predictive Value of Mouse Models in Drug Discovery and Development

May 16-18, 2005

Bringing together some of the best minds from both academic and industrial sectors, this year's Discovery Strategies meeting examines how the promise of select technologies and resource initiatives may be applied to expediting the delivery of safe, effective drugs to the clinic.

Session I Emerging Technologies and Strategies - focuses on how new approaches are being used to leverage the enormous power of the mouse in discovery, research, and development.  In particular, advances in key modeling areas will be addressed including new knockout strategies, humanization of the laboratory mouse and the application of siRNA in modeling potential new therapies.

Session II Phenotyping - emphasizes new strategies and resources that are being developed to provide comprehensive phenotype data on diverse mouse strains used in biomedical research, and how these resources are being used in pharmaceutical development.

Session III Mapping Functional Networks - examines the use of mouse models in pharmaceutical development from a systems biology perspective.  New approaches to the identification of functional networks underlying complex disease will also be presented.

Session IV Intellectual Property Rights - the organizers will convene a panel of experts to discuss intellectual property rights and how to promote sharing of intellectual resources across diverse sectors. Session Chair: Brian Stanton, Ph.D., DHHS, NIH

Session V Predictive Modeling of Drug Efficacy and Safety - explores issues in modeling human responses to drugs and how existing methods of predicting efficacy and safety might be improved.

Hosted by: The Jackson Laboratory Research Affiliates Program
Location: Bar Harbor, Maine, USA
Date: May 16-18, 2005
Phone: 207.288.6326
Fax: 207.288.6080
e-mail: discovery@jax.org
 

Speaker Abstracts Session I:

Session 1 Emerging Technologies and Strategies

Gene traps, targeting and the mouse knockout project
Mark Moore, Ph.D., National Human Genome Research Institute, abstract pending

In Vivo siRNA In Animal Disease Models: Improved Drug Discovery and Pathway
To siRNA Therapeutics
Martin Woodle, Ph.D., Intradigm Corporation

We have developed in vivo delivery methods for nucleic acids including nanoplex targeted siRNA.  These methods are used in murine models of disease as part of an efficacy-based discovery and validation method we call Efficacy-FirstÔ Discovery.  This method relies on gene perturbation of disease pathways with a readout of induced phenotype and efficacy for two key steps: 1) identification of genes and proteins associated with that phenotype and efficacy – the most relevant characteristic of candidate therapeutic targets – and 2) validation of genes as inhibitor drug targets.  Results using the system will be presented and evaluated for capabilities of the system using xenograft tumor models for cancer drug target discovery for novel monoclonal antibody candidates, considered in terms of capability to improve target quality and reduce discovery time.  The in vivo application of siRNA will also be reviewed for prospects as a novel therapeutic modality, in terms of efficient and clinically feasible means for systemic siRNA administration.  Intradigm is leading development of systemic targeting of siRNA with a uniquely effective ligand-targeted nanoparticle.  The combination of tissue selective delivery by the nanoparticle with gene selective inhibition by the siRNA opens the door to dual-targeted therapeutics.  Intradigm is developing a first product in this class, ICS-283, for inhibition of neovascularization and angiogenesis in cancer, ocular diseases, and other critical diseases.  In addition, Intradigm has worked with its partners to establish proof of concept for use of airway delivery of siRNA for treatment of SARS Coronavirus and other respiratory viral infections.

SCID Mouse Models for Human Biomedical Research
Leonard D. Shultz, The Jackson Laboratory

There is a growing need for effective animal models to carry out experimental studies on the human hematopoietic and immune systems without putting individuals at risk. Progress in development of small animal models for human hematopoiesis and immunity has seen three major breakthroughs. (1) C.B17-scid mice: The initial discovery in 1988 demonstrating that human lymphohematopoietic cells could engraft in C.B17-scid mice. Limitations of this model included high levels of host NK cell activity, the development of murine T and B cells upon aging ("leakiness"), and their ability to support only very low levels of human hematolymphoid engraftment. (2) NOD-scid mice: The development of NOD-scid mice in 1995 by our laboratory. NOD-scid mice have reduced levels of NK cell activity, additional deficits in innate immunity, and support heightened levels of human hematolymphoid cell engraftment. The NOD-scid model has seen incremental improvements over the last 10 years with the development of NOD-scid mice homozygous for a targeted mutation at the b2 microglobulin (b2m) locus, the development of NOD mice with a null mutation in the recombination activating gene 1 (Rag1) locus, the null mutation in the perforin (Prf1) locus, and transgenic expression of human HLA molecules. Although improved as hosts of human lymphohematopoietic cells due to reduced innate immune function, these models remain limited by their short lifespan and the lack of development of a fully functional human immune system. (3)  Immunodeficient IL2rgnull mice: An alternative genetic strategy utilizing a targeted mutation at the interleukin 2 receptor common gamma (IL2rg) chain resulted in a major increase in support of engraftment of functional human hematolymphoid cells in immunodeficient mice. Deficiency of IL2rg  chain causes X-linked SCID in humans. This molecule is utilized by a number of cytokine receptors and is indispensable for IL-2, IL-4, IL-7, IL-9, IL-15, and IL-21 high-affinity ligand binding and signaling. Thus, IL2rg  chain deficiency causes severe impairment in innate and adaptive immunity. IL2rg  targeted mutations have been produced independently by three different groups and each of the IL2rgnull genetic stocks bred to scid or Rag2null mice have been used to develop immunodeficient models for human hematolymphoid engraftment. Our laboratory has developed a genetic stock of NOD-scid IL2rgnull mice, carried out phenotypic analyses, and evaluated support of engraftment of adult mice with human mobilized HSC and newborn mice with cord blood HSC. NOD-scid IL2rgnull mice lack mature lymphocytes and have multiple deficits in innate immunity that include an absence of NK cells. However, these mice survive beyond 16 months of age and resist lymphoma development. NOD-scid IL2rgnull mice support high levels of engraftment with human mobilized and cord blood CD34+ HSC. Engraftment results in multilineage human lymphoid and myeloid differentiation in the bone marrow.  De novo T cell development in the engrafted mice is validated by (1) high levels of T cell receptor excision circles, (2) complex TCRb repertoire diversity, and (3) proliferative responses to T cell mitogens. The development of a functional human immune system is validated by the finding of human humoral immune responses to ovalbumin and the development of functional mature human CD4+ as well as CD8+ T cells. NOD-scid IL2rgnull mice and further improvements in this model will facilitate research in many areas, including human hematopoiesis, immunity, defense against bioterrorism, regenerative medicine, diabetes, and cancer.

Imaging in CNS Drug Discovery
Richard Hargreaves, Ph.D., Merck Research Laboratories

Imaging is a key area of ‘translational research’ that provides a unique bridge from the laboratory to the clinic in many therapeutic areas. Imaging is especially important in neuroscience drug discovery and development since quantitative biomarkers as surrogate efficacy measures are often lacking and clinical trial endpoints can be confounded by high placebo response and take a long time to collect. Neuroimaging can be used pre-clinically to select candidate drug molecules during drug discovery and clinically to facilitate proof of concept testing and optimization of resources through prioritization of decision making during the development of new therapeutics. Conceptually, neuroimaging in drug discovery and development can be divided into four categories that are clearly inter-related.  1) Neuroreceptor mapping to examine the involvement of specific neurotransmitter systems in CNS diseases, drug occupancy characteristics and perhaps examine mechanisms of action; 2) Structural and spectroscopic imaging to examine morphological changes and their consequences; 3) Metabolic mapping to provide evidence of central activity and “CNS fingerprinting” the neuroanatomy of drug effects; 4) Functional mapping to examine disease drug interactions.  Positron Emission Tomography (PET) and Magnetic Resonance (MR) currently dominate the methodologies that are used for neuroimaging. Each technique, whilst powerful in its own right, has optimal value for understanding the pathophysiological characteristics of CNS diseases, their diagnosis and potential treatment outcomes when combined together due to the complimentary nature of the information provided. Neuroimaging is now central to research and drug development in the neurosciences and has begun to allow detection of the pharmacological and physiological consequences of drug action within the living brain.

Speaker Abstracts Session II:

The EUMORPHIA project: Phenotyping the mouse – from gene function to drug discovery
Steve D.M. Brown, Ph.D., and the Eumorphia Consortium
MRC Mammalian Genetics Unit, Harwell, OX11 ORD, UK

With the completion of the mouse genome sequence, one of the key goals for functional genomics is the creation of a series of mutant alleles for every mammalian gene. Large-scale mutagenesis efforts are underway using both gene-driven and phenotype-driven approaches to generate this resource of mouse mutants. An even greater challenge will be the determination of the phenotypic outcomes of each mutation. A vital element of this endeavour will be to develop standardised phenotyping platforms that allow for reproducibility of test outcome over both time and place. The EUMORPHIA programme, funded by the European Commission, is a consortium of 18 research institutes from across Europe working on establishing new approaches to phenotyping with a focus on improving and standardising phenotyping platforms for the mouse. A major achievement has been the development of a new robust primary screening strategy, EMPReSS (European Mouse Phenotyping Resource for Standardised Screens). This primary screen incorporates over 100 SOPs, each validated on a cohort of inbred strains across a number of laboratories. EMPReSS covers all of the major body systems, as well as generic approaches in pathology and gene expression. We have developed a database and web-based resource for the visualisation, searching and downloading of SOPs and other documents that constitute EMPReSS. In addition, we have established an associated ontological structure for describing phenotypes that importantly incorporates the phenotype assay (the SOP) as an integral component of the ontology. We are currently developing a phenome database to incorporate EUMORPHIA phenome data that will link phenotype data, SOPs and ontologies providing an integrated approach to the capture, analysis and dissemination of phenotype data. EUMORPHIA is continuing to build and enlarge EMPReSS with a particular focus now on the development of secondary and tertiary screens.

The availability of standardised screens and associated informatics structures and tools will be a vital underpinning for a systematic and rational functional annotation of the mouse genome and a necessary step in improving the mouse as a tool in the drug discovery process. It will be important to build upon programmes such as EUMORPHIA and others and to set international standards for phenotyping and phenotype analysis.

Mouse Phenome Project: Identifying Mouse Models
Molly Bogue, Ph.D., The Jackson Laboratory, abstract pending

Comprehensive Phenotyping of Genetically Modified Mice to Identify New Target Disease Indications
Rosalba Sacca, Ph.D., Pfizer, Inc.

The use of  genetically modified mice to establish gene function has been well documented  in the literature and their role in drug discovery  has been proven to be extremely valuable.  This presentation will discuss the generation of the Phenotype Pfinder platform by which KO mice for specific genes of interest are characterized, using a panel of assays to identify new functions for that gene as well as potential detrimental effects.  A number of examples will be highlighted to show the ability of this platform to identify new and unexpected phenotypes.

In Vivo Comprehensive Compound Profiling
David S. Grass, Ph.D., Xenogen Biosciences

abstract co-authors Beverly K. Jones, Heather S. Hain, Sylvie Ramboz, Arman Saparov, Michelle M. Bunzel, Aaron Corona, Yu Rao, Chris Shen, Bryan Strenkowski, and Olesia Buiakova.  Xenogen Biosciences, Cranbury, NJ  08512

We have developed a comprehensive phenotyping platform that allows for the efficient comprehensive evaluation of compounds in vivo.  To validate this platform, we have performed proof of principle studies using the generic anxiolytic compound, buspirone.  Buspirone is a partial agonist of the serotonin 5HT1A receptor and an agonist of the dopamine D2 receptor and the a1 adrenoreceptor. Its primary metabolite, 1-(2-Pyrimidinyl)-piperazine (1PP), also exhibits bioactivity and is an agonist of the 5HT1A receptor and the a2 adrenoreceptor.  C57BL/6 mice received a single dose of either vehicle (PBS), 1mg/kg, or 6 mg/kg of buspirone via i.p. injection.  Results from the elevated plus maze assay confirmed that buspirone reduced anxiety related behavior in mice, concomitant with a short term reduction in locomotor activity.  However, assays examining metabolism and immunology also showed dramatic effects in response to buspirone.  Glucose excursion was increased following an oral glucose tolerance test and insulin levels were significantly reduced in response to buspirone.  Levels of serum adiponectin and leptin were unaffected.  Food intake, measured over a three hour period, was also unchanged, but triglycerides were significantly elevated.  Heart rate was significantly depressed in buspirone-treated mice but systolic blood pressure was unaffected. Corticosterone levels were significantly increased.  Total white blood cells, neutrophils, lymphocytes and monocytes were significantly reduced.  In summary, behavioral differences attributable to the primary indication of buspirone were confirmed.  However, significant differences were found in glucose metabolism, heart rate, number and distribution of white blood cells, serum corticosterone and trigyceride levels. Studies utilizing a chronic dosing regimen via continuous infusion are currently in progress.  The results to date are consistent with what has been reported in the literature for the effects of this compound in rodents.  In addition, the results are consistent with not only the primary indication of buspirone in humans (anxiolytic), but also some of the other observed physiological effects seen in humans. 

We intend to apply this platform to confirm expected primary indications, identify new secondary indications, and/or identify unexpected side effects for lead compounds, optimized lead compounds, preclinical drug development candidates, and/or clinical drug candidates.  In addition, this platform could be utilized to reposition existing therapeutics, and thus expand the market potential for such therapeutics. 

Speaker Abstracts Session III:

Statistical Assessments of Time Series GeneChip® Data Using Multi-Factorial ANOVA Provide Novel Insights into Sleep Deprivation and the Function of PPAR-g2 in Mice
Presenter Keith Shockley, Ph.D., The Jackson Laboratory, co-authors M. Mackiewicz, J. E. Zimmerman, C. Ackert-Bicknell, O. P. Lazarenko, C. J. Rosen, B. Lecka-Czernik, A. I. Pack and G. A. Churchill. 

Technological advances leading to decreasing costs and wider access to high-throughput expression platforms are enabling more and more researchers to conduct time course expression studies. While promising, it remains to determine how significance measures resulting from these studies should be used to resolve underlying functional networks when considering dynamic changes in transcriptional response. Although there is currently no agreement for how to best conduct and analyze a time course microarray study, it is clear that certain key aspects should be carefully considered in each investigation. For instance, due to differences in fabrication technology precision, treatments of Affymetrix GeneChip® array data require different considerations than those that have been commonly employed in two-color platforms. Other important aspects include experimental design, assessment of raw data quality, implementation of normalization procedures used to generate gene expression measures, determination of a statistical framework to assess questions of primary interest and the application of pathway analysis procedures to characterize outcomes. We will address these concerns through multiple-way analysis of variance (ANOVA) incorporating shrinkage variance components in the context of two different time course studies described below. Examples from each of these studies demonstrate that specific contrasts and multiple linear regression trend tests of expression profiles result in greater power to elucidate expression changes.

The objective of the first study was to compare time-specific differences in gene expression in different brain regions of mice associated with sleep deprivation. In this experiment, animals kept awake for selected time increments were compared to time-matched animals allowed to sleep normally during their typical resting periods. After ANOVA-based statistical assessment of gene expression estimates, the overrepresentation of gene set members in the final gene lists were determined by the Fisher exact test. It was discovered that coping mechanisms for sleep deprivation were offset by temporal effects in the cerebral cortex and hypothalamus. Additionally, stress response genes were induced in sleep disturbed mice, while engaging in normal rest cycles lead to higher expression of genes involved in the synthesis of membrane components. The purpose of the second study was to examine temporal mechanisms by which the nuclear receptor PPAR-g2 exerts anti-osteoblastic and pro-adipocytic effects in the presence or absence of the biomedically relevant ligand rosiglitazone in UAMS-33 mouse cell cultures. The presence of multiple treatment factors in the latter experiment presented opportunities to fully exploit an ANOVA framework within the time series experimental design. A pathway-specific correlation analysis identified several gene clusters important in osteoblast/adipocyte cell differentiation and bone homeostasis, indicating that PPAR-g2 is a major regulator of marrow mesenchymal stem cell differentiation. 

Mechanisms and Patterns of Genome Organization in the Mouse Genome
Joel Graber, Ph.D., The Jackson Laboratory

The physical and functional organizations of a genome are correlated outcomes of evolution. Inbred strains of mice provide a unique opportunity for exploring these relationships, representing as they do, diverse genomes originally separated by millions of generations that were then scrambled in the laboratory and subjected to intense selection during inbreeding to homozygosity. Here we show that the resulting pattern of chromosome organization includes regional domains of functionally related elements that promote the co-inheritance and survival of compatible sets of alleles.  There are also patterns of linkage disequilibrium between domains on separate chromosomes; these are distinctly non-random and form networks with scale-free architecture. The strong conservation of gene order among mammals suggests that the domains and networks we find likely characterize all mammals, and possibly beyond.

Integrating genetic and gene expression data: Application to cardiovascular and metabolic traits
Aldons J. Lusis, Ph.D., Department of Microbiology Immunology and Molecular Genetics and UCLA School of Medicine, University of California, Los Angeles

Integrating genetic and gene expression data: application to cardiovascular and metabolic traits.  Quantitative trait locus (QTL) analysis provides a straightforward method for mapping genes involved in complex traits in genetic crosses with mice or rats.  However, with the exception of genes that explain a large fraction of the variance of a trait, it has been difficult to go from a locus to a gene.  We and others are developing an approach that combines genetic segregation and gene expression profiling to help understand the molecular pathways involved in complex traits.  We applied the strategy to two intercrosses, one between mouse strains DBA/2 and C57BL/6 (Schadt et al. Nature, 2003, 422: 279) and the other between strains C3H and C57BL/6.  In the first cross, consisting of 114 females, we examined transcripts in liver, and in the second cross, consisting of 334 males and females, we examined transcripts in liver, fat, muscle, and brain. In this approach, the loci controlling transcript abundance, as detected by whole genome expression arrays, are mapped in the same way as other, physiologic, traits.  These loci, called expression QTL, or eQTL, are of two varieties.  Roughly one quarter of the loci with lod scores over four map to the structural gene encoding the  transcript, and thus, appear to act in cis (cis eQTL).  The others map to distal regions and are therefore trans eQTL. We have carried out a classic cis-trans test on a subset of the cis eQTL and found that the majority of the predicted QTL exhibited evidence of cis regulation (Doss, et al. Genome Res. 2005, April 18). Also, over 95% of the cis eQTL map to regions of high SNP density between the strains whereas the trans eQTL are found equally in high and low SNP density regions. We have used the data to construct co-expression networks based on correlations of the transcript levels in the genetic crosses.  The networks consist of “modules” of highly interconnected genes that are only loosely connected with other modules.  Some of the modules are significantly correlated with physiological traits examined in the crosses. We have also attempted to examine causal interactions relating DNA variation, gene expression, and physiologic traits.  Several genes predicted to be causal for the amount of visceral fat have been validated using transgenic mice. Finally, the coincidence of chromosomal loci for a clinical trait QTL and a cis-acting eQTL appears to be a useful method to nominate genes from within the disease susceptibility locus.

Expression-based identification of adipose tissue genes involved in metabolic disease
Yaacov Barak, Ph.D., Kathy Snow, Suyeon Kim.
The Jackson Laboratory, Bar Harbor ME 04609

Obesity and its associated Syndrome X, which includes type II diabetes, hypertension, and atherosclerosis, are a public health problem of epidemic proportions, and the leading cause of death in developed countries. Adipose tissue is a critical regulator of glucose and lipid homeostasis, and its central role in type II diabetes and Syndrome X is demonstrated in lipodystrophy, where acute fat degeneration leads to typical obesity-associated metabolic disorders. Phenotypic convergence of the diametrically opposed obesity and lipodystrophy suggests that the physiological status of adipose tissue in obesity is a functional equivalent of its physical degeneration in lipodystrophy. Molecular characterization of fat tissue in lipodystrophy should therefore provide important insights into its degenerative responses in obesity. Here, we will demonstrate a molecular correlate to the phenotypic convergence of lipodystrophic and obese adipose tissues, in the form of multiple gene expression changes that are common to both. Most of these differentially expressed genes have not been previously studied in relation to fat biology. Several patterns arise, including deregulated inflammation and evidence that a robust, adipocyte-produced pro-inflammatory chemokine is central to the etiology of lipodystrophy. In addition, a disproportionate number of imprinted genes and genes potentially associated with neuronal growth are deregulated. Additional genes, which do not cluster into obvious functional groups, but exhibit unique expression profiles, include novel secreted factors and a transcription factor that underlies a rare congenital diabetes syndrome in humans. Overall, this differential expression-based approach refreshes our understanding of adipose tissue biology and disease, and offers new strategies to identify prospective drug targets in the battle against Syndrome X.

Speaker Abstracts Session IV:

Intellectual property rights (IPR) are a cornerstone of the biotechnology
industry.  However, small biotechnology firms, large pharmaceutical
companies, and the academic community have different IPR goals and
practices.  There is a perceived tension between reserving access to
fundamental genetic information for basic research and preserving industrial
IPRs.   This issue has attracted the attention of governments throughout the
world.  In the USA, the US Patent and Trademark Office has promulgated
patentability standards to address some of these issues, while NIH has
promulgated best practices for licensing of genomic inventions.
Internationally, the OECD is developing best practices for the licensing of
genetic inventions.  On the legal side, patent pooling has been suggested as
a possible solution to the problem of access to biotechnology.  This session
will explore this perceived problem and discuss patent pooling and voluntary
compliance with best practice licensing practices as means to address these
issues.  Following formal presentation, open discussion will bring forth
various perspectives.

Speaker Abstracts Session V:

Opportunities and Challenges for Improving the Predictive Value of Mouse Models in Oncology Drug Discovery and Development
Robert Wild, Ph.D., Bristol-Myers Squibb Co.

Animal models have long played an important role in pharmaceutical drug discovery and development. In the oncology area, both transplantable syngeneic tumor models as well as human tumor xenografts implanted in nude mice have provided significant insights into the therapeutic potential of numerous anticancer agents. Nevertheless, data obtained from these studies often times failed to truly predict clinical activity, particularly which cancer indication or histology would be most suited for a given drug candidate. Recent advances in understanding the molecular pathogenesis of cancer have shifted the focus of oncology drug development from traditional cytotoxic approaches to molecular targeted strategies. This general change in philosophy as well as recent emerging technologies have opened the opportunity to improve our testing paradigms to more accurately mimic cancer disease in preclinical animal models. For instance, advances in gene manipulation and transgene technologies now allow the generation of target-selective and -driven disease models that provide excellent tools to assess the role of a new molecular target in cancer tumorigenesis and/or normal animal physiology. In addition, these models offer a unique opportunity to do initial proof of concept studies on the effects of new pharmacological entities in vivo. Moreover, the recent development of sophisticated imaging technologies now permit the non-invasive tracking of multiple parameters in vivo, which in some cases have direct translational value in the clinical setting. Likewise, advances in pharmacogenomics and proteomics allow us to identify specific molecular signatures, or biomarkers for our tumor models, which may predict responsiveness more accurately than traditional means. Finally, in the advent of molecular targeted therapies, it is becoming increasingly important to correlate conventional antitumor activity endpoints with molecular pharmacodynamic (PD) and surrogate biomarker readouts. More specifically, recent accomplishments suggest that successful modeling of pharmacokinetic (PK) and pharmacodynamic (PD) relationships in preclinical cancer models may prove of significant predictive value when applied to a molecularly well defined clinical setting. In summary, this presentation will give a broad overview of in vivo tumor model selection/development strategies and provide some detailed examples for its application to various cancer drug discovery projects.


Does the lack of genetic diversity in animal models currently used for safety testing put the public at risk?

H. Jacob, S. Nye ,N. Cozzi, J. Baye, D. Evans, Y. Evrard, S. Korb, H. Vernon, A. Wittenburg, M. Hessner, X. Wang, R. Roman, PhysioGenix, Inc. and the Medical College of Wisconsin, Milwaukee, WI 53226

Purpose:  Safety testing in Pharma is currently done by using inbred or outbred rat or mouse strains.  Inbred strains offer reproducibility but have limited predictive value to other rat strains let alone humans.  Outbred strains lack the perceived degree of genetic diversity and test populations cannot be reproduced in reasonable sample sizes.  To address this, PhysioGenix developed genetically diverse PharmGenix rat panels that capture 82% of the commercially available genetic diversity in the rat and are faithfully reproduced in a controlled fashion

Methods:  PharmGenix rats were tested for sensitivity to the toxic effects of the antibiotic gentamicin, the analgesic acetaminophen and the Alzheimer’s drug tacrine.  Biomarkers in urine and blood were analyzed, organ pathology was assessed and gene expression by microarray analysis was performed.

Results:  The differential response exhibited by PharmGenix rat strains revealed genetic components underlie toxicity to gentamicin.  PharmGenix rats also responded differentially to acetaminophen, but the pattern of toxicity was dependent on the target organs (liver, kidney).  Tacrine elevated serum transaminases in three PharmGenix strains, but hepatotoxicity was missed by the Sprague-Dawley and F344.  For mechanistic studies, knowing that PharmGenix rats respond differently to drugs enables 20-fold enrichment in finding relevant susceptibility and resistance genes by microarray and this number is further refined when combining haplotype information.

Conclusions:  Genetic background has a large impact on the response to known renal and liver toxicants and suggests that results of efficacy or safety testing performed in any single strain of rats will not predict the range of responses expected in humans.

Genetic Variation in Toxicity: Improving Prediction
Ken Paigen, Ph.D., The Jackson Laboratory

It is likely that genetics contributes to a substantial fraction of individual differences in susceptibility to adverse reactions, whether directly or by influencing responses to environmental factors.  We can model this variation in response by exploiting the genetic differences among inbred strains of mice.  Our confidence that what the mice teach us is relevant to human biology is enhanced by the discovery in recent years that it is largely the same set of genes that determine susceptibility of both mice and man to a variety of human pathologies.

The potential now exists to improve our ability to predict adverse reactions among genetic sub-sets of the human population, even though these may be infrequent.

Opportunities and limitations of computational methods in Safety Assessemnt
Scott Boyer, Ph.D., Astrazeneca

Experimental models used to assess the safety of drug candidates can be complimented by the judicious use of computational tools.  These tools can range from 'predictive' quantitative structure activity models to 'hypothesis generation' tools that have the potential to focus experimentation in safety studies.   Real world application of chemogenomics for linking chemical structure to a specific adverse effect via proteins and pathways will also be presented.  A balanced and pragmatic view on the real world application of these tools is necessary and a discussion of both successes and failures will be presented.


Use of transgenic mice in drug safety assessments
David Jacobson-Kram, Ph.D.,DBAT., Office of New Drugs, Center for Drug Evaluation and Research, U.S. Food and Drug Administration

Determining the carcinogenic potential of materials to which humans have significant exposures is an important, complex and imperfect exercise.  Not only are the methods for such determinations protracted, expensive and utilize large numbers of animals, extrapolation of data from such studies to human risk is imprecise.  Toxicologists have long recognized these shortcomings but the two-year chronic rodent study has remained the gold standard.  Recent developments in the field of molecular oncology and development of methods to insert or inactivate specific genes in animals have provided the tools with which to develop the next generation of carcinogenicity assays.  With improved understanding of oncogene activation and tumor suppressor gene inactivation a number of animal models have been developed to dramatically reduce latency for chemically-induced cancers.  This has led to the development of shorter carcinogenicity assays. Also, because the spontaneous tumor frequencies in these animals are low during the in-life portion of the study, and studies are terminated well before the health complications of advanced aging are observed, it has been possible to reduce the group sizes and reduce animal usage.  FDA’s adoption of ICH SIB in 1997, (ICH, 1997) “Testing for the Carcinogenicity of Pharmaceuticals,” opened the door for the use of such transgenic models in regulatory toxicology.  This presentation reviews the current state of the science and its application to regulatory issues.

--end of conference program--

Past Conferences: