Abstracts

Day 1: Thursday, September 15, 2011

Joint Session I: Keynote Lecture

Have Animal Models of Disease Helped or Hindered the Drug Discovery Process?
Jackie Hunter PhD, CBE, OI Pharma Partners Ltd.

In the past decades the amount of information on pathways and mechanisms of disease has increased exponentially. Although more and more of this information is derived from human studies, animal models clearly have an important role to play in confirming and expanding on these findings. However most animal models of disease only model certain aspects of disease and do not fully recapitulate the pathophysiology of the human disease state. Therefore, if a mechanism is deemed of potential therapeutic importance on strong physiological and clinical evidence, is it sufficient to show that a compound is active in an animal model of the mechanism for its progression to man? Have we ruled out potentially effective compounds on the basis of their activities in animal models of disease? Conversely has an over-reliance on animal models in other areas led to high attrition rates in the clinic?

Joint Session II: Breat out into 3 Concurrent Workshops

Workshop 1: Highlights and Hurdles of Animal Models of Human Metabolic Disease

Using Mouse Knockouts to Guide Drug Discovery: LX4211 A Novel Diabetes Drug Candidate
Brian Zambrowicz, PhD, Lexicon Pharmaceuticals, Inc, The Woodlands, Texas

We have utilized a large scale mouse knockout and phenotypic analysis approach to identify novel and first-in-class mechanisms for drug discovery. Knockouts can be considered as models of potential antagonist drug action. They model an antagonist with 100% potency and 100% specificity. However, judgment must be used in the interpretation of knockout data as it pertains to both efficacy and safety for a given drug target. We have knocked out and analyzed the phenotypes of 4,650 genes using a broad screen of physiology and behavior relevant to disease. I will describe the general strategy and illustrate the utility of this strategy by discussing the preclinical and clinical data surrounding one diabetes drug candidate currently in phase 2b clinical trials. LX4211 is a first-in-class dual SGLT1 and SGLT2 inhibitor. In patients with type 2 diabetes, LX4211 produced large and rapid improvements in glycemic control that resulted in a 0.76% reduction in HbA1c over only four weeks of dosing. In addition, multiple metabolic and cardiovascular improvements were also observed. This dual mechanism of action results in improvements over SGLT2 selective compounds including the triggering of GLP-1 secretion by the gastrointestinal tract. This drug discovery and development work was guided by the knockout data for the two targets and supports the utility of the mouse for guiding the discovery and development of new drugs.

The Obesity Pipeline: Current Models and Strategies in the Development of Anti-Obesity Drugs
Kevin G. Murphy, PhD, Imperial College London

Obesity is now a global pandemic. The World Health Organization estimates that more than half a billion people are obese worldwide. Obesity is associated with increased morbidity and mortality due to associated diseases including type 2 diabetes mellitus and cardiovascular disease. Dietary and exercise advice, and current pharmacotherapy, have proven relatively ineffective. Bariatric surgery effectively treats obesity, but is an impractical option to cope with the increasing prevalence of obesity. New anti-obesity drugs are thus urgently needed. Energy balance is a homeostatic system. Body weight is regulated by a complex neuroendocrine signaling system between the central nervous system, the gastrointestinal tract and adipose depots. It is difficult to model neuronal circuitry and the endocrine system in vitro, and in vivo models have proved critical to our understanding of the systems controlling energy homeostasis. The signals regulating appetite and body weight appear similar across mammals, and animal models of obesity are used to determine the efficacy of potential anti-obesity agents. Such models have limitations; energy expenditure is differently regulated in rodents, redundancy and developmental compensation can make knockout models unreliable, and side effects, a major concern for obesity drugs, may not be detected. Nevertheless, animal models have identified a number of agents with potential as treatments for obesity which are now at various stages of drug development. The future may see combinations of anti-obesity agents being used to prevent weight gain and treat obesity.

Workshop 2: Highlights and Hurdles of Animal Models of Human Cardiovascular Disease

Angiotensin II Promote Diverse Forms of Aortic Aneurysms
Alan Daugherty, PhD, DSc, University of Kentucky, Lexington

Many laboratories have demonstrated that angiotensin II infusion into normo- and hyperlipidemic mice promotes the development of abdominal aortic aneurysms (AAAs). Angiotensin II-induced AAAs are characterized by a focal medial rupture that leads to a thrombus with adventitial dissection. Thrombotic material gradually resolves and is replaced by fibrous material, while there is progressive lumen dilation. Continued infusion for 3 months leads to progressive expansion that can be monitored by high frequency ultrasound. Continued aortic expansion is associated with progressive macrophages accumulation at the site of medial rupture. Interruption of macrophage function in mice with deficiency of the adaptor protein, MyD88, abates the development of angiotensin II-induced AAAs. More recently, it has been observed that angiotensin II infusion leads to aneurysms that are highly localized to the ascending aorta. Within 5 days of initiating angiotensin II infusion, there is dissection of the outer layers of the aortic media, with blood accumulating between elastin layers. Aneurysms in these regions are characterized by progressive concentric lumen expansion with increased medial thickening and pronounced elastin fragmentation. The development of angiotensin II-induced aneurysms is ablated in mice lacking AT1a receptors. Surprisingly, the smooth muscle cell-specific deletion of AT1a receptors had no effect on aneurysms in either region. In contrast, depletion of AT1a receptors in the endothelium attenuated the development of ascending aortic aneurysms. Mechanisms of this protective effect have not been defined.
 
Coauthor: Lisa Cassis, PhD, University of Kentucky, Lexington

Sex Differences in Cardiovascular Function in Genetically Manipulated Mice
Amrita Ahluwalia, BSc, PhD, Barts & The London School of Medicine, Queen Mary University of London

Globally cardiovascular disease (CVD) is the main cause of death accounting for 29% of all deaths in 2004. However, it is well-established that the incidence of CVD is significantly lower in females compared to age-matched males. Statistics show that this reduced susceptibility to CVD in females is lost following menopause when the rates of CVD are similar in both sexes (www.BHF.org). This sex-difference has been attributed, at least in part, to the protective effects of oestrogen. In particular, oestrogen-dependent upregulation of endothelial cell function has been implicated. The exact endothelial mediator(s) involved in this phenomenon is uncertain. I will discuss both the evidence in pre-clinical models implicating different endothelial pathways in sex-dependent protection and how these data correlate, or not, with human disease.

Modeling NSAID Activity in the Cardiovascular System
Jane A. Mitchell, PhD, Imperial College London, London, United Kingdom and  Barts & the London School of Medicine & Dentistry, London, United Kingdom

Non-steroidal anti-inflammatory drugs (NSAIDs), used for the treatment of inflammation and pain, are the most commonly used over-the-counter medication worldwide. These drugs share the common property of reducing prostanoid biosynthesis by inhibition of cyclo-oxygenase (COX)-1 and/or COX-2. In man, this produces well-documented effects in the cardiovascular system. On one hand, NSAIDs, particularly aspirin, can produce anti-thrombotic and cardiovascular protective effects. On the other, both non-selective COX-1/COX-2 inhibitors (e,g. diclofenac), and selective COX-2 inhibitors (e.g.celecoxib) are associated with detrimental cardiovascular effects including an increased risk of thrombotic events, particularly myocardial infarction. The anti-thrombotic effects of NSAIDs are relatively straightforward to study in animal models, where aspirin or COX-1 gene-knock out reduces platelet reactivity. Hypertension and gastric injury can also be demonstrated and modeled in laboratory animals treated with NSAIDs. Data from knock out mice suggests that both COX-1 and COX-2 are important in mediating these effects. The pro-thrombotic effects of NSAIDs, which manifest in man as increased rates of myocardial infarction, are serious, but relatively rare and unpredictable. An improved understanding of the relative roles of COX-1 and COX-2 throughout the circulation, and how they interact with underlying contributing factors, such as vascular inflammation or hypertension, to influence thrombotic risk is required.

Coauthors: Nicholas S. Kirkby, PhD, Imperial College London & Barts & the London School of Medicine & Dentistry and Timothy D. Warner, PhD, Imperial College London & Barts & the London School of Medicine & Dentistry, London, United Kingdom

Animals and Models: Interpretative Challenges and Drug Development
Garret A. FitzGerald, MD, University of Pennsylvania School of Medicine

A pivotal stage in drug development is the attempt to gain proof of concept in cells or more commonly animal models. The value of this exercise depends on both the strengths and limitations of the model and of the human beings who design the experiment and attempt to interpret its relevance. No model recapitulates the human condition, nor can any investigator interpret the human relevance of discrepancies between results in 3 monkeys versus 300 mice. Sometimes the end point of interest simply does not occur—spontaneous plaque rupture in hyperlipidemic mice. Other times, simulations of the clinical paradigm are artificial in the extreme—mice on a hot plate predicting the efficacy of analgesics in middle aged ladies with osteoarthritis of the knees. Commonly, there is a failure to appreciate the dynamic range of experimental paradigms as in studies of the thrombogenic vs antithrombotic effects of drugs. Finally, there is the general constraint of immunosuppression in assessing anticancer drugs in mice implanted with tumors. Here, the high spontaneous incidence of tumors in dogs offers and alternative approach to assessing chemoprevention strategies, albeit that frequently, tumor markers evident in humans may be incompletely recapitulated. Besides genomic differences relevant to drug action—for example in editing enzymes in dyslipidemia—gene deletion in mice may evoke counter-regulatory gene expression that may be absent with incomplete target blockade achieved by tolerated drug doses in humans. Finally, sufficient attention is rarely paid to measurements of drug exposure and their extrapolation to humans: what is a "low dose" of aspirin in a mouse? For these and other reasons, one seeks congruent information from distinct models, ideally crossing species and from both genetic and pharmacological manipulation of the target. In this setting, titration of drug exposure or the quantitative measure of drug effect (e.g. enzyme inhibition) can be related to functional impact in the model in a way that informs dose selection in humans. Knock in of gene variants may provide useful information on their potential functional importance in humans. Given all these potential limitations, it is unsurprising that some are totally dismissive of the value of animal models in drug development. However, when experiments are designed rationally and diversified evidence is obtained; animal models can be most useful in the prediction of drug response in humans and in the mechanistic elucidation of drug action.

Workshop 3: The 'Animalgesic' Effect in Modeling Pain

Modeling Inflammatory Hyperalgesia
Julie Keeble, PhD, King's College London

Inflammatory hyperalgesia is characterized by increased sensitivity to a noxious stimulus, caused by inflammation. Such a state is clearly demonstrated by both humans, for example in osteo- and rheumatoid arthritis, and the respective animal models. Modeling inflammatory hyperalgesia in animals can be achieved by targeting specific receptors on sensory nerves, e.g. TRPV1 and TRPA1 agonists, which is very useful in the search for antagonists for such targets. However, it is far more difficult to model inflammatory hyperalgesia in animals where the aim is to reproduce human disease states where a variety of mediators are involved. The stimulus for inflammatory hyperalgesia in animal models is rarely the same as compared with human disease. The assessment of the role of specific mediators in animal models of hyperalgesia is therein stimulus led. Albeit, pro-inflammatory stimuli such as carrageenan and complete Freund's adjuvant produce highly reliable and reproducible models of inflammation in rodents that is associated with significant thermal and mechanical hyperalgesia. Useful information can be gained from these models with respect to the potential role for different mediators/receptors in causing hyperalgesia in general, but not necessarily in relation to a specific disease state. Thermal and mechanical hyperalgesia are standard measures of pain in animal models, using techniques such as the Hargreaves test and Von Frey hairs. These have historically been very useful in identifying mediators/receptors that are involved in inflammatory hyperalgesia although new techniques are emerging that may be better tests of 'natural' pain behavior.

Use of Novel, Non-Reflex End Points for Detecting Analgesic Action in Rodents at Clinically Relevant Concentrations
Nick Andrews, PhD, Pfizer Global Research and Development, Sandwich, United Kingdom

The methods historically used in laboratory animals for preclinical prediction of the efficacy of novel analgesics have recently been critically evaluated by several authors and behavioural assessment of animals is called for (Mogil, (2009); Rice et al., (2008); Rice, (2010)). Reflex withdrawal-based paradigms do not measure the global impact of pain per se and furthermore do not address ethological validity. Rather than trying to replicate clinical features related to human pain in the animal laboratory, predictive validity may be improved by looking to re-instate specific, innate behaviours suppressed by pain e.g. rearing (Matson et al., 2007).Furthermore, a major advantage of looking to re-instate suppressed behaviours, (rather than dampening enhanced reflexes), is that compounds that impair motor function do not appear as false positives. We have taken this philosophy forward in developing a burrowing assay and also a novel operant assay that measures rearing activity (without suffering from habituation as happens with the method of Matson et al) to measure the effect of chronic pain in rats. We have found that gabapentin reverses burrowing deficits (30 mg/kg sc) and rearing deficits (10 mg/kg sc) induced by peripheral nerve injury and ibuprofen reverses burrowing deficits (30 mg/kg sc) induced by CFA. These doses produce plasma concentrations in the rat more in line with effective clinical plasma concentrations than doses generally used in rodent reflex withdrawal assays. Though further work is needed, it is tentatively suggested that re-instating suppressed behaviours, relevant to the rat, may give improved clinical dose prediction.

Coaurthors: Sinead Harper and Yasmin Issop MSc, Pfizer Global Research and Development, Sandwich, United Kingdom

Does Fibromyalgia Have Four Legs and a Tail?
Jon Levine, MD, PhD, University of California, San Francisco School of Medicine, San Francisco, California

The pathophysiology of chronic widespread pain remains poorly understood, and available therapies are only partially, and mostly unpredictably, effective. Progress has been seriously hindered by the lack of adequate experimental models. Given the strong clinical association between stress and diverse chronic widespread pain syndromes we have developed animal models based on this association and used them to evaluate pathophysiology. Depending on the nature of the stress, animals develop mechanical hyperalgesia and neuroplasticity in the primary afferent nociceptor. Our model, developed to study diffuse musculoskeletal pain (fibromyalgia syndrome), also demonstrates increased anxiety, and visceral hyperalgesia. The neuroplasticity in the nociceptor allows a response to levels of cytokines found after exposure to bacterial endotoxin. In addition, lesion of both neuroendocrine stress axes – the hypothalamic-pituitary-adrenal and sympathoadrenal – markedly attenuates pain in these models. Importantly, electrophysiological recordings from muscle nociceptors suggest neuropathic changes capable of contributing to enhanced nociceptor function. Thus, animal models of chronic widespread pain, which reproduce many aspects of the clinical syndrome, validate a role of neuroendocrine stress axes and cytokines, as well as suggest a role for neuropathic changes in the primary afferent nociceptor in chronic widespread pain.

Pharmacokinetic Pharmacodynamic Correlations in the Translation of Efficacy from Animal Pain Models
Garth Whiteside, PhD, Purdue Pharma, Cranbury, New Jersey

Consideration of pharmacokinetics in translating efficacy across species is essential. If the PK does not translate then the efficacy will also not translate. While clinical compounds are frequently used to validate preclinical models, as positive controls and to infer translatability from preclinical studies to clinical studies, consideration of the translatability of the pharmacokinetic parameters is all too often lacking. In addition, highly complex datasets, both clinical and preclinical, are frequently reduced to a one work summary—"worked", "didn't work", "failed" etc. To more fully understand preclinical and clinical datasets these one word summaries cannot reveal the full picture. In order to determine translatability and certainly to conclude a lack of translatability, one needs to more closely examine the data. This presentation focuses on the pharmacokinetic and pharmacodynamic relationship between a number of compounds that are active in both clinical and preclinical pain states. In addition, and to exemplify how continuous discourse between clinical and preclinical can inform a mechanistic understanding of drug action, the relationship between efficacy of a bisphosphonate, pain and extent of joint damage in an osteoarthritis model will be presented.

Joint Session III: Regulations and Best Practices in Disease Modeling in Animals

Animal Welfare and the 3Rs in European Biomedical Research
Dominic Wells, VetMB PhD, Royal Veterinary College, University of London

Historically the UK has had strict legislation governing the use of animals in research since 1876 and the 1986 Animals (scientific procedures) Act was enacted to ensure a strong animal welfare element was applied to the use of experimental animals. The main elements of this national licensing scheme are approval of the animal facility, approval for a specific set of experiments justified by a cost/benefit analysis and the licensing of individuals to perform specific techniques after approved training. It is a requirement of the Act that animal welfare is optimized wherever possible by consideration of the 3Rs (Replacement, Reduction and Refinement) and the establishment of humane end points to procedures. The recently approved EU Directive 2010/63 aims to extend broadly similar protection across Europe although the success of this will depend on the effectiveness of implementation by the member states. There are significant differences between the current UK system and the EU Directive and a UK consultation exercise will be held in summer 2011 to consider whether additional parts of the UK system will be retained. An important element of the EU Directive is the requirement to spread best practice in the 3Rs and a number of existing bodies could take up this role. It is gratifying to note that several organizations are rising to this challenge and that there is every indication that implementation of the EU Directive will significantly improve the welfare of experimental animals across Europe, especially in biomedical research.

US Public Policy and the Oversight of the Use of Animal Models in Biomedical Research and Testing
B. Taylor Bennett, DVM, PhD, National Association for Biomedical Research, Washington, DC

The oversight of animal use in biomedical research and safety assessment testing is provided by three agencies, the United States Department of Agriculture (USDA), the National Institutes of Health's Office of Laboratory Animal Welfare (OLAW) and the Association for the Assessment and Accreditation of Laboratory Animal Care International. Each of these agencies relies upon the Institutional Animal Care and Use Committee (IACUC) to provide assessment and oversight of an institution's Animal Care and Use Program (ACUP) through regular review of that program, inspection of the facilities that compose the program, review of concerns involving the care and use of animals, and review and approval of proposed activities involving animals and subsequent modifications to that usage. The IACUC must include as a voting member the veterinarian with institutional authority and responsibility for the health and well-being of the animals cared for and used in the program—the Attending Veterinarian (AV). The IACUC and the AV usually report directly to the Institutional Official (IO) who has ultimate responsibility for the ACUP. This presentation will review how this three pronged approach provides effective oversight while facilitating research and highlight the need for sound public policy to support this system.

Best Practices for the Use of Animals in Toxicological Research and Testing
William S. Stokes, DVM, National Institute of Environmental Health Sciences, Research Triangle Park, North Carolina

Laboratory animal models of human disease and injury are used in translational research to assess efficacy and pre-clinical safety of new medicines and vaccines. Animal models are also used to assess the safety of new chemicals and products like pesticides, industrial chemicals, consumer products, and food additives. However, disease or toxic injury can often result in significant pain and distress, and such studies can involve large numbers of animals. Welfare concerns have led to national policies and laws in the U.S. and other countries to ensure humane care and use of laboratory animals, and to require the consideration of ways to reduce, refine, and replace animal use before studies are approved. Continued advances in science and technology provide new opportunities to develop improved animal models and integrated safety assessment strategies that can reduce uncertainties in extrapolating from animals to humans, and can further reduce, refine, and in some cases replace animal use. Increased understanding of cellular and molecular mechanisms and perturbed pathways leading to disease and injury are leading to earlier in vivo and in vitro safety biomarkers. Such biomarkers can often serve as the basis for earlier humane endpoints to avoid or reduce the duration and severity of pain and distress in animal studies, and have also reduced and replaced animal use for some studies. Continued development, validation, and use of scientifically sound models is expected to improve animal welfare and further reduce and replace animal use while ensuring and advancing the health of people, animals, and the environment.

Discussion: Can We Model Human Disease in Animals?
Margaret S. Landi, VMD, GlaxoSmithKline, King of Prussia, Pennsylvania

This section of the seminar will set the stage for the debate, from the negative side, on the question if animals can be models of diseases in humans. Areas to be covered include, but are limited to, a brief history of animals in biomedical and behavioral research. Published reports on the challenges of efficacy models and toxicity testing in animals in drug discovery and development will also be reviewed.

Joint Session IV: Recent Advances in Generating Genetically Modified Animals for Therapeutic Discovery

Diverse Types of Pluripotent Stem Cells and Their Relevance for Animal Modeling
Roger A. Pedersen, PhD, University of Cambridge

Recent studies have revealed a distinction between the pluripotent stem cells that can be derived from the inner cell mass of mouse blastocysts and from the epiblast layer of later, post-implantation stage embryos. Blastocyst-derived mouse embryonic stem cells (mESCs) can contribute to all body tissues in chimeras, but do not form the trophoblast cells of the placenta. Mouse epiblast-derived stem cells (EpiSCs) scarcely contribute to chimeras but seemingly can form trophoblast cells. Intriguingly, the properties of human embryonic stem cells (hESCs) derived from blastocysts mirror those of mouse EpiSCs derived from later stages. Thus, the properties of mESCs, EpiSCs and hESCs appear to reflect their developmental stage, rather than their species of origin. We studied the differentiation of hESCs and EpiSCs into mesoderm and found that they respond similarly to the epiblast layer of mouse embryos. Rather than forming trophoblast, however, they appear to form a sub-population of placental mesenchyme that expresses marker genes shared with placental trophoblast. Thus, ability for chimera formation is a defining feature of diverse types of pluripotent stem cells. Advances in methods for deriving and maintaining pluripotency of embryonic stem cells from other species have enabled researchers to establish rat ESCs. These are capable of contributing extensively to chimeras with rat or mouse embryos, thus resembling mESCs. This raises the possibility that human ESCs with properties similar to mESCs and rESCs could be derived, either from embryos or from hESCs. Asking whether such ESCs would participate in chimeras has both experimental and ethical challenges.

Targeted Gene Disruption in Rats via Zinc Finger Nucleases
Aron M. Geurts, PhD, Medical College of Wisconsin

Genetic engineering in laboratory model systems is the benchmark for investigation of functional mechanisms of the roles of genes in biological and disease traits. In practice, routine genetic engineering in vertebrate model systems has been generally limited to mice, due to the availability and protocols to manipulate embryonic stem cells. We recently demonstrated the use of Zinc Finger Nucleases (ZFNs) for rapid genetic modification of the rat embryo genome by pronuclear microinjection as an alternative to stem cell engineering. Using ZFNs, we can generate both knock-out and knock-in modifications to the rat genome in a timeframe of a few months. Thus, direct manipulation of the embryo genome obviates the need for stem cells for several engineering approaches and reduces the time and effort necessary to produce a genetically modified animal. We have now applied this approach to generate variants in 100 rat genes in a span of two years and are beginning to uncover new and interesting insights into complex human disease. Improvements in the art of genetic engineering in the rat, partnered with its reputation as an excellent physiological and behavioral research model, offer great promise to impact discovery and intervention into human biology and disease.
 
Coauthor: Howard Jacob, Medical College of Wisconsin.

Porcine Induced Pluripotent Stem Cells For Generating Swine Biomedical Models
Steve Stice, PhD, University of Georgia, Regenerative Bioscience Center, Athens, Georgia

Porcine induced pluripotent cells (piPSC) will be used to generate potentially complex animal models for human diseases and modeling allogenic pluripotent stem cell and tissue transplants in large animal models. Generated piPSC chimeric offspring demonstrated that piPSC can contribute to the in vivo development of multiple tissues. However transplantation of undifferentiated piPSC into a disease model is unlikely, given the potential for teratomas formation in the recipient animals. Conditions for guided differentiation of pluripotent stem cells, including induced pluripotent stem cells, into progenitors that do not generate teratomas has been developed using human pluripotent stem cells in our laboratory. Shortly after we isolated three of the initial of NIH registered human embryonic stem cells lines we generated neural progenitor lines. We used these to produce more uniformly differentiated neuronal cells ranging from motor neurons to dopaminergic neurons and now these have been successfully used in stroke and spinal cord injury rodent animal models. Traditionally, in vitro differentiation processes for embryonic stem cells often utilize embryoid body formation because this is a relatively easy procedure, but rarely results in a uniform population of germ layer specific progenitor cells. We have developed more defined conditions required to produce uniform human neural, and now mesenchymal and germ cell specified progenitor cells in adherent human embryonic and now induced pluripotent ESC cultures without embryoid body formation. The development of equivalent piPSC derived progenitor cells will allow us to test allogenic stem cell based therapies in large animal models.
 
Coauthor: Franklin West, University of Georgia, Regenerative Bioscience Center, Athens, Georgia

Session V: Break out into 2 Parallel Workshops

Workshop 4: Phenotype Screening and the Impact of Genetic Background

The Knockout Mouse Project (KOMP) Repository
K. C. Kent Lloyd, DVM, PhD, Mouse Biology Program, University of California, Davis, California

The NIH-funded KOMP Repository is the archive and distribution center for targeting vectors; ES cell clones, live mouse lines, and frozen embryos and sperm that have been produced for the 8500 knockout genes targeted by the KOMP Mutagenesis Program. The Repository conducts quality assurance procedures to ensure the viability, genotype, pathogen-free status, and genomic integrity of all KOMP products. In addition, the Repository provides investigators with value-added services, such as microinjection, germline transmission testing, cryopreservation, and colony management. The Repository currently holds 75,302 targeted ES cell clones for 6909 unique genes, and to date has received 3316 orders from investigators. The Repository itself has injected 659 ES cell clones to produce 446 knockout mouse lines (a 68% germline transmission rate), indicating that injection of 3 clones per gene will ensure a >96% chance of generating a genotypically-confirmed knockout mouse. Because the NIH intended the resource to become self-supporting, fees are charged to both academic and for-profit users to obtain KOMP products and services. The Repository maintains an easily navigable website (www.komp.org) where one can search, browse, order products and services, register interest in genes and receive automatic notices when products become available, access customer and technical services (1-888-KOMP-MICE, service@komp.org), receive news updates, follow the KOMP blog, view FAQs, and download relevant protocols. Besides ensuring the utility, longevity, and vitality of this unique resource, the KOMP Repository is key to the success of KOMP-Phase 2, which seeks to functionally annotate all protein coding genes in the mammalian genome.

Functional Genomics in the Rat and the Knockout Rat Consortium (KORC)
Eric M. Ostertag, University of Kentucky and Transposagen Biopharmaceuticals Inc., Lexington

The laboratory rat has been the desired animal model for human disease over the past century. Rat relevance to human physiology coupled with its size, ease of manipulation, and breeding characteristics are attributes of this animal model. The Knockout Rat Consortium (KORC) is a repository of genetically modified rat models that were created using multiple technologies. The KORC website is www.knockoutrat.org and contains a database, news items, a discussion form, and useful links. The KORC now has over 300 genetically modified rats in its database. Examples of knockout rats available include a SCID model, a p53 knockout, a model of pain (Trpc4), a model of hydrocephalus (Myo9a), a new model of obesity (Mc4r) and a Sod3 knockout for hypertension. The SCID rat contains an insertional mutation within the adenosine deaminase (Ada) gene, which is an underlying cause of many cases of severe combined immunodeficiency (SCID) in humans. A p53 knockout rat is also available with extensive phenotype information, including tumor spectrum data. The Trpc4 knockout rat is a novel model for pain research and demonstrates greatly reduced sensitivity to visceral pain. The Myo9a knockout enables researchers to study hydrocephalus in an animal model that allows facile surgical manipulation. Mc4r mutations are the single most common form of monogenic obesity in humans. The Sod3 knockout rat is hypertensive due renal failure from oxidative stress. On behalf of the KORC we invite all interested investigators to participate in our efforts to assemble a repository of knockout rats.

Coauthors: Aron M. Geurts, Medical College of Wisconsin, Milwaukee, and Edwin Cuppen, University Medical Center Utrecht, Netherlands

Genetic Resource Populations for the Laboratory Mouse
Gary A. Churchill, PhD, The Jackson Laboratory, Bar Harbor, Maine

Genome-wide association studies in human populations have raised the bar for genetic mapping and analysis of Mendelian and complex traits. However, the genetic structure of human populations, uncontrolled environmental variables, and limitations on experimental interventions and phenotyping present significant barriers to investigations of biological processes in humans. If we could design an ideal model system for genetic studies, what properties would it have? High genetic diversity is desirable to enable the broadest possible scope of discovery. High mapping resolution is needed to identify causal genes with confidence and precision. Absence of population structure and rare alleles combined with full genomic sequences of ancestral haplotypes will substantially improve power and reduce required sample sizes. A good model system does not need to have a natural population structure but the evolutionary origin and context in which segregating variants arose should be understood. The ability to reproduce genotypes leads us to consider inbred models, but natural heterozygosity is also desirable. In depth phenotyping tools, a high-density genotyping platform, and methods to work with transgenic constructs are essential. An experimental system should allow for both discovery and validation. I will describe how the Collaborative Cross and Diversity Outbred mouse populations together fulfill these criteria and provide an ideal system for genetic analysis in a mammalian model organism.

Minimizing Strain Influences in a Genetically Modified Mouse Phenotyping Platform
Michael D. Hayward, PhD, Taconic, Cranbury, New Jersey

The characterization of phenotypes in genetically modified mice are frequently performed as in-depth mechanistic studies but they can also be scaled up to serve as a discovery tool to identify novel physiological functions of genes or therapeutically relevant targets. The choice of ES cell and the choice of which strain to backcross to (congenic approach) should ideally yield data with the lowest possible standard deviation related to genetic heterogeneity. However, that approach can be complicated by the existence of substrains of the most popular strains used (e.g., 129 and C57BL/6) and genetic differences that exist among these substrains. Several differences between two of the more commonly used substrains, C57BL/6J and C57BL/6NTac have been identified, such as differences in nociception and insulin secretion. The characterization of cohorts on a mixed background (such as 129XC57BL/6) are frequently done, but less frequently recognized is that studies are also done on mixed substrains (such as C57BL/6J X C57BL/6NTac). In the case when a pure substrain congenic line is not practical, good statistical practices, such as power calculations, become critical. Since breeding of genetically modified mice is costly, attempts have been made to minimize that expense by balancing the need for statistical power, the costs of breeding and control of the background strain. An approach that we have used commercially (Xenogen Biosciences/Pfizer Agreement) and has also been used in some international phenotyping consortia (i.e., EMPREeSSslim), to a more limited extent, is to “multiplex” in vivo tests by conducting them serially. The serial design ensures that a majority of the resources go towards characterization rather than the breeding of the mice and allows for a larger group size in the assays, thus reaching a statistical power whereby even modest changes can be detected. Our approach, therefore, was to power studies to allow for detection of at least modest changes from a WT littermate control, include assays with overlapping physiological systems to provide cross-functional interpretive value and to employ challenge assays.
 
Coauthors:  Olesia Buiakova, David S. Grass, Taconic

Behavioral Differences Among C57bl/6 Substrains
Camron D. Bryant, PhD, Department of Human Genetics, University of Chicago

C57BL/6 ("B6") mice at the Jackson Laboratories ("J") and NIH ("N") represent the two core substrains of B6 mice, each of which have their own substrains. Investigators often assume that a B6 mouse is a B6 mouse but accumulating studies clearly demonstrate that this is a dangerous assumption. Specifically, there are large behavioral differences among B6 substrains, e.g., in motor coordination, pain sensitivity, fear learning, neuromuscular strength, anxiety-like behavior, depressive-like behavior, working memory, alcohol preference, psychostimulant sensitivity, aggression, startle, and prepulse inhibition. I will first highlight the main behavioral findings that have been replicated across laboratories. As an example, we and others have noted that the C57BL/6J strain shows a low level of fear learning relative to other B6 substrains and thus, it might be ill-suited to detect the effects of manipulations that would be expected to decrease this phenotype. Molecular genetic studies also document SSLPs, SNPs, and CNVs that distinguish the B6 substrains—I will briefly summarize this growing list of variants. Genetic variation among B6 substrains likely contributes to the behavioral variation and thus, it is crucial that investigators unambiguously document and carefully choose which B6 substrains to employ for forward and reverse genetic studies of behavior.
 
Reference:
Animal Models and Their Value in Predicting Efficacy and Toxicity. New York Academy of Sciences and Global Medical Excellence Cluster in collaboration with Imperial College London and King's College London.

Workshop 5: Vascular Inflammation and Pain — Understanding the Link

Sensory Hyperresponsiveness of the Lung
Maria G. Belvisi, PhD, Imperial College London

Asthma and chronic obstructive pulmonary disease (COPD) are inflammatory diseases of the airway characterized by airflow limitation. Asthmatics at the more severe end of the disease spectrum and COPD patients are not effectively managed with standard therapy (i.e., glucocorticoids and β2-adrenoceptor agonists). Furthermore, there are no treatments for COPD that can halt the progression of the disease. Cough is a common symptom of both diseases and also presents as a disease in its own right. Treatment options are limited and a recent meta-analysis concluded that OTC remedies are ineffective and there is increasing concern about their use in children. TRPA1 channels are non selective cation channels that are activated by a range of natural products (eg. allyl isothiocyanate), a multitude of environmental irritants (eg. acrolein which is present in air pollution, vehicle exhaust and cigarette smoke) and inflammatory mediators (eg. cyclocpentane prostaglandins). TRPA1 is primarily expressed in small diameter, nociceptive neurons where its activation probably contributes to the perception of noxious stimuli. Inhalational exposure to irritating gases, fumes, dusts, vapours, chemicals and endogenous mediators can lead to the development of cough. The respiratory tract is innervated by primary sensory afferent nerves which are activated by mechanical and chemical stimuli. Recent data suggests that activation of TRPA1 on these vagal sensory afferents by these irritant substances could lead to central reflexes including dyspnoea, changes in breathing pattern and cough which contribute to the symptomatology and pathophysiology of respiratory diseases.
 
Coauthors: Megan Grace, Eric Dubuis, and Mark A. Birrell, Imperial College London.

Modeling Inflammation and Microvascular Dysfunction
Felicity N.E. Gavins, PhD, BSc, Imperial College London

A hallmark of inflammation is the recruitment of blood-borne leukocytes in the microvasculature, which involves a complex set of events that can occur both locally and systemically. Both in vivo and in vitro evidence have demonstrated molecular and cellular pathways involved in this multi-step cascade.
 
Inflammation and microvascular dysfunction can be visualized using intravital microscopy (IVM), a quantitative and qualitative in vivo molecular imaging technique that enables further understanding of the physiology and pathophysiology of different diseases, and the cellular and subcellular distribution of potential therapeutic compounds to be assessed. All of which results in a cost-effective approach to identify potential novel therapeutic drug targets for disease and accelerate preclinical drug development.
 
We have used the technique of IVM to investigate the cardinal signs of inflammation and microvascular injury associated with ischaemia-reperfusion (I/R) injury and sepsis. We, and others, have found increasing evidence suggesting that members of the for myl peptide receptor (FPR) family, in particular human FPR2/ALX play an important role in mediating the anti-inflammatory and pro-resolving effects of several peptides and non-peptidyl small-molecule compounds in a number of different disease models. This talk will review evidence and suggest that FPR ligands, particularly in the brain, could be novel and exciting anti-inflammatory therapeutics for the treatment of a variety of clinical conditions, including stroke and sepsis.

Role of TRPV1 and TRPA1 Channels in Inflammatory Pain
Susan D. Brain, PhD, King's College London

Pain is a poorly controlled component of arthritis. We have been working at a fundamental level to develop models in which we can learn more of the crosstalk of inflammatory mediator and signalling systems. Our central hypothesis is that these results will in turn inform mechanisms by which pharmacological targets may offer therapeutic benefit. Mice deleted of the TRPV1 gene are protected from secondary thermal hyperalgesia in a CFA joint inflammation model, but TNFα levels remain high in the TRPV1 knockout mice, suggesting that TNFα generation occurs before TRPV1 activation (Keeble et al., Arthritis Rheum 2005 52:3248-56). We considered how we could learn more about the ability of TNFα to interact with the TRPV1 system. We simplified our model and studied slocal (ipsilateral and contralateral) effects after intraplantar administration into the ipsilateral paw. This allowed us to use genetically modified mice and also to give blocking drugs either locally or systemically. We realised that a complex downstream cascade of events occurred that involved cyclo-oxygenase products (Russell et al., Pain 2009 142:264-74). Progressing, we were well placed to learn of the distinct role of TRPA1, (through use of TRPA1 wildtype and knockout mice). We have now confirmed these results, using TRPA1 antagonists (Fernandes et al., Arthritis Rheum 2011 63:819-29).
 
We thank the Arthritis Research Campaign and a BBSRC-led IMB capacity building award for support.

Developing TRP-Channel Blockers for Inflammatory Pain
O-G Berge, DDS, PhD, AstraZeneca R&D, Södertälje, Sweden

The algogenic and desensitizing actions of capsaicin intrigued scientists for decades and inspired industrial programs to develop novel analgesics. These efforts accelerated with the cloning of the "capsaicin receptor" TRPV1 published in 1997 and identification of other members of the TRP ion channel family with potential roles in sensory functions. In recent years, several TRPV1 antagonists have been tested in man but mechanism-related as well as off-target effects have hampered their clinical development. The clinical utility of blocking TRPV1 is still unclear. Analgesic efficacy in man has been reported in experimental and acute postoperative pain but negative data have been announced in pain due to osteoarthritis. The list of TRP targets for drug discovery and development related to inflammatory pain includes TRPA1, TRPM8, TRPV3 and TRPV4. Other channels like TRPM2 and TRPM4 may influence immunological functions of relevance for inflammatory pain. As drug targets, the TRP channels are challenging due to extensive expression patterns and complex biology with tetrameric and in some cases heteromeric assembly of units and signaling associated with several activating and regulating mechanisms. Endogenous ligands and physiological functions are not fully identified for several potentially important channels and exogenous compounds activating the channels, however useful for physiological characterization, frequently have properties that make them unsuitable as tools for drug discovery. Poor species crossover in pharmacology and biology has been shown for several targets, posing difficulties for animal modeling. In spite of the challenges, the TRP channel family remains a promising field for drug development.

* Additional abstracts to come.