A Synopsis of the “Influence of Epigenetics,Genetics,and Immunology” Session Part A at the 35th Annual Society of Toxicologic Pathology Symposium

Genetic Variation in Nonhuman Primates (NHPs) and Impact of Toxicology Programs

In the first presentation, Dr. Jonathan Moggs of Novartis (Basel, Switzerland) spoke about the potential impact of genetic variation in nonclinical toxicology species, with a particular emphasis on NHPs. Although it has long been recognized that there are genetic differences that need to be considered when selecting appropriate species for use in nonclinical efficacy and toxicity studies, it is only recently that genetic differences between different subpopulations and strains within the same animal species have been determined. For example, whole genome sequences have recently been elucidated for cynomolgus monkeys of Mauritius (Ebeling et al. 2011), Vietnamese (Yan et al. 2011; Huh et al. 2012), and Malaysian origins (Higashino et al. 2012). Importantly, genetic sequence variation has been shown to occur between cynomolgus monkeys from distinct geographic origins; for example, there is variation in major histocompatibility (MHC) loci between Chinese and Indian populations. Furthermore, variation in microscopic findings has been seen in the same tissue in cynomolgus monkeys from different regions. As an example, there is a higher incidence of chronic diffuse gastritis in cynomolgus monkeys derived from the Philippines and Vietnam, as compared with Mauritius population, despite the presence of spiral-shaped bacteria in the stomach of monkeys from all 3 regions (Drevon-Gaillot et al. 2006). However, one limitation of these data sets is that they were based on only one or a few individuals, and there is currently a lack of genetic sequence and phenotypic information from multiple individuals across the various populations to more thoroughly explore the effects of genetic and environmental variation between populations.

Dr. Moggs emphasized that the genetic characterization of commonly utilized cynomolgus monkey populations can improve nonclinical pharmacology and toxicology study designs during drug development via selection of appropriate monkey populations, and he presented results from a recent initiative to more extensively characterize genetic variation within 3 distinct populations of cynomolgus monkeys using exome sequencing. Single nucleotide polymorphisms (SNPs) were identified that were in common or unique for cynomolgus monkeys originating from the Philippines, Vietnam, and Mauritius. Applications of the resulting genetic database include (1) providing a platform to rapidly assess extent of genetic variation in drug targets and molecular pathways, (2) supporting toxicology species selection, and (3) optimizing the design of assays and biomarkers supporting toxicity studies. Key areas that remain to be explored include the potential relevance of NHP genetic variants for humans and the extent to which genetic variation within NHP populations may affect phenotypic outcomes in toxicology studies. Dr. Moggs concluded the presentation by providing forward-thinking implications of enhanced nonclinical genetic databases, including guiding the selection of appropriate nonclinical models and supporting genome safety assessments for therapeutic genome modification.

Genetics: A Factor to Consider in Drug Safety Assessment Studies Using Cynomolgus Monkeys

In the second presentation of the session, Dr. Karissa Adkins of Pfizer (Groton, CT) presented her experiences with exploring genetic sequence variation in cynomolgus monkeys in the context of toxicity studies supporting drug development. Dr. Adkins cautioned that outlier findings in NHP studies that are often considered incidental may instead be a reflection of important interindividual genetic variability, particularly because environmental variability is kept to a minimum by study design and laboratory conditions. Importantly, because the genetic sequence homology between humans and cynomolgus monkeys is greater than 90%, outlier responses have the potential to reflect interindividual responses that may occur in human populations.

An example of this concept was provided in the context of delayed type IV hypersensitivity reactions of the skin and mucous membranes that occurred in only a subset of test article–dosed cynomolgus monkeys (9 of 62). Owing to observations that MHC haplotype variants have been associated with human hypersensitivity reactions to certain drugs (Hughes et al. 2004; Odueyungbo, Sheehan, and Haggstrom, 2010), Dr. Adkins and colleagues investigated whether MHC haplotype variation could explain why only certain cynomolgus monkeys experienced a reaction. In their studies, the MHC region was genotyped for the 62 monkeys, and microsatellite profiles were used to infer haplotypes within the MHC class I and class II genetic regions. This analysis enabled determination of a specific haplotype that was associated with the hypersensitivity reactions in the skin. This finding has the potential to prospectively inform selection of cynomolgus monkeys for studies and to address the potential risk for human populations. Details of this work are presented in the next article in this journal (Wu et al. in press).

Dr. Adkins has extended this work by better characterizing the genetic sequence information across cynomolgus monkeys of Mauritius origin. In this effort, whole exome sequencing data have been generated for 101 naive monkeys. Data analysis has shown that there is an average of approximately 8 sequence variants per 1,000 base pairs within the population, with the vast majority as SNPs. Dr. Adkins concluded that cynomolgus monkey populations are highly polymorphic; furthermore, the density and frequency of variants between the Mauritius subspecies of cynomolgus monkey and human populations are similar. These findings have important implications for translating functional effects of sequence variants between cynomolgus monkeys and humans. Dr. Adkins further stressed that understanding genetic variability has important implications for toxicology because these data can add important perspective to incidental findings in toxicity studies and can improve translatability of nonclinical results to clinical studies.

Low Frequency Clinical Adverse Drug Reactions Can Be Predicted and Studied by Using Genetically Diverse Mouse Populations

The third presentation of the session shifted the focus from monkeys to rodent models. Dr. Alison Harrill of the University of Arkansas for Medical Sciences (Little Rock, AR) advocated the use of mouse populations as a tool for exploring population diversity in adverse responses to drugs and chemicals. Classical toxicity studies are often limited to a single inbred strain, such as C57BL/6J mice—which contain no genetic diversity or outbred stocks such as CD-1 mice—which harbor very limited genetic diversity (Festing, 2010). Thus, important genetically mediated effects that are relevant to sensitive subpopulations of humans may be missed when using these models for toxicity studies. In the presentation, new mouse population models were discussed: (1) the Mouse Diversity Panel, comprised of dozens of inbred strains, (2) the Collaborative Cross, comprised of hundreds of recombinant inbred lines, and (3) the Diversity Outbred, a large outbred stock derived from the same 8 genetic founder strains as the Collaborative Cross. Dr. Harrill presented data showing that toxicity testing using a mouse population approach could retrospectively reveal human-relevant risks of adverse drug reactions, such as kidney injury that occurred in clinical trials of drug DB289 (Harrill et al. 2012) and liver enzyme elevations that occurred in clinical trials of PF-04287881 (Mosedale et al. 2014). Extending the model further, Dr. Harrill presented data demonstrating that Diversity Outbred mice can provide a useful model for clinically relevant idiosyncratic liver reactions, as demonstrated by studies using zileuton and green tea extract (GTE; Church et al. 2015).

In addition to modeling the diversity of human adverse reactions, Dr. Harrill presented data demonstrating that genetic variation within mouse populations can be exploited to identify genetic risk factors of adverse drug reactions as well as insights into toxicity modes of action. In the case of herbal supplements containing GTE, genome-wide association mapping in the Diversity Outbred mice identified candidate pharmacogenetic risk genes that were then investigated in human clinical cohorts. This translational pharmacogenetic approach identified 3 genes containing variation that distinguished human GTE liver injury cases from an ethnically matched control population (PER3, MFN2, and VPS13D; Church et al. 2015). In the case of PF-04287881, Dr. Harrill, along with Dr. Adkins and colleagues, utilized genetic sequence and liver gene expression information collected from sensitive and resistant mouse strains to conclude that variation in genes relating to assembly of phospholipids and energy metabolism contributed to an increase in susceptibility to liver injury induced by PF-04287881 (Mosedale et al. 2014). Dr. Harrill and her team have extended these studies to investigate the utility of Diversity Outbred mice for identification of genetic risk factors that may enable precision prescribing of otherwise toxic drugs.

Epigenetics in Toxicology

In the fourth presentation, Dr. Jonathan Moggs took the stage again, this time to explore the potential impact of epigenetics in toxicity studies. Epigenetic modifications of the genome include DNA methylation and chromatin protein modifications that govern the ability to transcribe a gene in a given location, determining whether a gene exists in a silent, poised, or active transcriptional state. Consequently, while mammals have a single genome, the epigenetic status of each cell varies based on its state of differentiation and can be altered by environmental challenges (Lillycrop and Burdge, 2014). Epigenetic mechanisms have been shown to be involved in a number of disease states, including, but not limited to, cancer, neurologic disorders, type 2 diabetes, systemic lupus erythematosus, and infectious diseases. Dr. Moggs suggested that knowledge of epigenetic mechanisms can enhance translational safety sciences, with particular emphasis on providing tissue-based biomarkers of pharmacodynamic or toxicologic effects and on providing mechanistic insight for long-lasting drug-induced cellular perturbations (e.g., carcinogenicity, immune memory, and developmental and transgenerational effects).

Dr. Moggs presented case studies demonstrating potential applications for genome-wide epigenetic molecular profiling in tissues derived from nonclinical toxicity studies. The case studies had a common theme of derisking drug-induced carcinogenicity, particularly for nongenotoxic carcinogenesis (NGC), for which there are no sufficiently accurate or well-validated short-term assays that enable detection (Ellinger-Ziegelbauer et al. 2009). Dr. Moggs and colleagues from the Innovative Medicines Initiative biomarkers and molecular tumour classification for non-genotoxic carcinogenesis (MARCAR) consortium have utilized integrated epigenomic and transcriptomic profiling data in a well-characterized NGC model of phenobarbital-induced liver tumors in mice to identify novel early biomarkers for NGC. These biomarkers included progressive phenobarbital-mediated increases in the expression of the Dlk1-Dio3 imprinted gene cluster noncoding RNAs, including Meg3 (also called Gtl2). This finding also provides novel insight into the mechanism underlying phenobarbital-induced NGC because Meg3 has been shown to regulate genes in the transforming growth factor (TGF)-β pathway (Mondal et al. 2015). TGF-β has been shown previously to play a key role in promoting rodent liver tumors in the context of phenobarbital via regulation of the proliferative potential of hepatocytes. Dr. Moggs then went on to describe recent initiatives for evaluating heritable (transgenerational) epigenetic changes in the context of exposure to xenobiotics, including the molecular epigenomic profiling of sperm derived from a rodent model in which multigenerational epigenetic adaptation of the hepatic wound healing response has been observed (Zeybel et al. 2012). Together, these case studies are providing a foundation for the application of epigenetic end points in drug safety assessment.

Epigenetics and Angiogenesis

In the final presentation of the session, Dr. Hellmut Augustin of the German Cancer Research Center and Heidelberg University (Heidelberg, Germany) discussed implications for epigenetics in the regulation of vascular function, with particular emphasis on endothelial cells (ECs). ECs line the inner wall of the vasculature, regulating vascular tone, blood cell circulation, inflammation, and platelet activity. Cardiovascular disease remains a significant cause of mortality worldwide (Reddy and Yusuf, 1998). Owing to a central role in regulating essential hemostatic functions, dysfunction of ECs is an early predictor of atherosclerosis and future cardiovascular adverse events (Daiber et al. 2016).

Dr. Augustin and his team investigated epigenetic modifications and their potential contributions to EC dysfunction, hypothesizing that there are epigenetic changes that facilitate angiogenesis that supports tumor growth. Thus, identification of key epigenetic changes during adolescent vessel maturation has the potential to identify targets for novel therapeutic agents. Dr. Augustin described his approaches to answer this question, involving methylome and transcriptome analysis of capillary ECs in lungs of mice at different life stages (infants, young adults, and mature adults). The detailed results are the subject of forthcoming publications; briefly, genome-wide increases in DNA methylation were observed to occur during vascular maturation. Specific loss and gain of epigenetic modifications were observed to be life-stage dependent and associated with increases in gene pathways related to regulation of the immunologic responses. These data provide a foundation for future investigations to identify novel therapeutic targets for EC dysfunction and tumor-related angiogenesis.

Summary

In conclusion, the session provided insights into several areas where knowledge of genetic and epigenetic variation may be an excellent resource for drug development. Whether it is exploiting natural genetic variation within a species for prospective study design or precision medicine, retrospective analysis, identification of epigenetic biomarkers of adverse effects, or modulation of therapeutic targets, these technologies have transformative potential. A key future direction will be to identify ways for stakeholders to share genetic data to collectively advance the field.