Abstract

Over 80% of drug candidates fail in clinical trials, incurring high costs to the pharmaceutical industry. The primary reason is failure of an animal model to mimic human drug responses, due to limitations of currently used animal models in representing the genetic diversity within the species. To address this problem, PhysioGenix has developed a novel combinatorial breeding strategy, the PharmGenix™ panel, to capture greater genetic diversity within the rat genome and allow for preclinical drug screening for the identification of a strain that mimics the human “drug” response. Tacrine, which causes hepatotoxicity in a percentage of the human population, was tested to demonstrate the utility of the PharmGenixTM panel. A single dose of tacrine (35 mg/kg) was administered to each of six hybrid PharmGenix™ strains, CD-IGS and CDF strains. Rats were euthanized twenty-four hours later and serum analyzed for alanine transaminase (ALT) and aspartate transaminase (AST) levels, indicators of hepatotoxicity. Tacrine did not significantly elevate ALT levels in CD-IGS or CDF, however the PharmGenix™ panel showed significant ALT elevations in five of the six strains, ranging from 111% to 142% of control values. CD-IGS and CDF exhibited increases in AST levels of 82% and 174%, respectively, while the PharmGenix ™ rats showed much higher AST levels, ranging between 621% and 1069% of control values in four of the six strains. The ability to detect significant AST and ALT elevations in the PharmGenix™ panel, but not in CD-IGS or CDF, suggests a genetic component underlies the development of tacrine toxicity. Furthermore, the results demonstrate the possibility of obtaining a false positive by selecting the CD-IGS or CDF results, both of which are currently used extensively by industry. It is essential to select the right animal model for improving the predictive power of preclinical animal testing. By maximizing strain diversity there is a greater likelihood of finding a model that mimics the human drug response exhibited within the target class of drugs. The PharmGenixTM model system offers an opportunity to identify the most predictive rat strain to allow for only the most promising drug candidates to progress forward, thereby decreasing failure rates

Concepts

Genetically Diverse Rat Models

• Use of a single inbred strain enables reproducibility, but lacks genetic diversity.
• Outbred strains have limited genetic diversity and results may not be reproducible unless a higher number of animals are used.

Heterozygous Rat Models

Human populations are heterozygous exhibiting two forms of heterogeneity:

• Allelic (different alleles, same gene)
• Locus (different genes)
• The therapeutic and adverse effects to any drug is likely polygenic.

New solutions for drug candidate screening

• Screen drug candidate in PharmGenixTM panel that failed in clinical trials due to toxicity.
• Select strain that mimics human adverse drug response.
• Use selected animal strain to screen analogs to improve success rate of test compound.

Breeding the PharmGenix™ Panel

Four of the most genetically diverse inbred rat strains were selected using the Strain Calculator™, a proprietary genomics informatics tool, to create a panel of six F1 offspring. These six strains capture 57% of the allelic diversity in strains of rats. This process was repeated to develop a second panel of PharmGenix ™ animals (for a total of 12 F1 offspring) which together capture over 75% of the genetic diversity found in the rat genome

Advantages of the PharmGenix™ panel:

1. Twelve different genome backgrounds maximize heterogeneity (as in outbred rats).
2. Isogenic individuals lead to more phenotypic consistency (as in inbred rats).
3. Rats from each PharmGenixTM strain are genetically equivalent (as in inbred rats) and the genotype of the each strain can be reproduced, reducing the possibility of idiosyncratic responses.
4. All rats in the PharmGenixTM panel are heterozygous for the majority of alleles (as in humans).

Genetic Assessment of Sprague-Dawley (SD)

• Column A: genotypes obtained for a group of SD rats purchased from a leading animal supplier (n = 9).  
• Column B: genotypes obtained from a second batch of SD rats from the same supplier (n = 10).
• Column C: the average number of alleles from the 2 panels.  
• Column D: the maximum number of alleles found within the 48 most-commonly-used inbred strains of rats.  In an outbred strain that reflects the “general” population, we would expect the number of alleles to approach that found for the 48 strains; the number of alleles actually observed in the SD are far fewer than expected.

Effects of Genetic Diversity on Safety Assessment

Tacrine, a cholinesterase inhibitor that can cause hepatotoxicity in some humans, was administered (35 mg/kg for one day) to six hybrid PharmGenix™ strains, CD-IGS, and CDF strains (n=6 per strain). ALT levels were not significantly elevated in either the CD- IGS or CDF. In contrast, the administration of tacrine to the PharmGenix™ panel resulted in significant ALT elevations (111% to 142% of control) in PG1, PG2, PG4, PG5, and PG6.

CD-IGS and CDF exhibited tacrine-induced increases in AST of 82% and 174%, respectively.  The PharmGenix™ panel of rats showed much higher responses with PG1, PG2, PG4, and PG6 having elevated AST levels ranging between 621% and 1069% of the control.

Associated liver damage (hepatocyte swelling and vacuolization) was observed in the PharmGenix™ panel as illustrated in representative histology from PG4.

 PharmGenix panel conclusions

1. The PharmGenix™ panel is more effective at capturing the range of responses expected in a diverse human patient population. Maximizing diversity increases the likelihood of finding a rat model that mimics the human drug response exhibited within the target class of drugs.
2. The PharmGenix™ panel identifies strains sensitive to the liver toxicity of tacrine, found in 30-50% of human patients.
3. PharmGenix™ model system offers an opportunity to identify the most predictive rat strain to allow for only the most promising drug candidates to progress forward, thereby decreasing failure rates.

This work was supported in part by NIH grants 5 R44 ES011432-03 and 3R44 ES011432-03S1.