National Capital Area Chapter Society of Toxicology

Predictive Teratogenicity Assessment: Current Applications and Future Directions

Karen Augustine, Ph.D. Discovery Toxicology, Bristol Myers Squibb, Hopewell, N.J.

Unexpected teratogenicity is a prevalent cause for toxicity-related attrition of drug candidates. It is a costly liability and tends to be identified relatively late in the pre clinical phase of drug development, following significant engagement in pre clinical and clinical studies. The origins of teratogenicity are related to the intended pharmacology having an essential role in embryo-fetal development and/or to an undesired “off-target” effect. Teratogenicity identified with endothelin receptor antagonists provides a classic example where proactive assessment of target-based liabilities could have been undertaken earlier, specifically in the discovery phase, potentially enabling better alignment of indication in context to the developmental toxicity risk. Such proactive approaches include thorough assessment of the literature to obtain an understanding of potential liabilities, additional follow up work conducted on the level of target expression and functional characterization using molecular biology and developmental model systems and applying in vitro teratogenicity screens. This presentation will describe the case study of unexpected teratogenicity associated with endothelin receptor antagonists under development in the 1990s and the types of proactive approaches that can now be utilized to proactively problem solve and potentially reduce later phase teratogenicity-based attrition.

Link to Slides:  http://www.toxicology.org/isot/rc/ncac/symposia_files/1_Augustine_NCAC-SOT_5'08.pdf

The Rodent Chronic Assay Database for Pesticides: Using Mode of Action Data to Assess the Human Relevance of Animal Tumors.

Vicki Dellarco, Ph.D., Senior Science Advisor, U.S. EPA, Office of Pesticide Programs, Washington DC.

The rodent cancer bioassay has formed the basis for health risk assessment and regulatory decisions for over 40 years. Several extrapolations or inference methods are necessary when using the results of this animal model to predict human health consequences. Unless there is evidence to the contrary, it is assumed that rodent data predict human effects, findings at high experimental doses predict effects at environmental exposure levels, and adult animals provide adequate coverage for predicting responses at different life stages, including children. Irrespective of the data base, the rodent liver is the most target organ most frequently showing a tumor response. An examination of the pesticide data base revealed that the mouse lung is also a common target which is consistent with the data showing that some frequently used mouse strains spontaneously develop pulmonary tumors. In addition to liver in the rat and mouse and lung in the mouse, thyroid follicular cell and Leydig cell tumors are frequently diagnosed in the rat following pesticide treatment. The hormone sensitive tumors in the rat thyroid and testicle typically represent species sensitivity. Unlike the rodent, the incidence of liver cancer in humans occurs at a much lower frequency compared to the much higher frequency of prostate and breast cancer. Prostate tumors rarely are diagnosed in rodents. Because of significant differences between rodent and human tumor incidences, data from the rodent cancer bioassay should be interpreted in combination with all available data to predict human hazard. The consideration of mode of action is extremely valuable to assess species differences and to inform dose-response extrapolation and modeling approaches. An approach that provides rigor and transparency to the weight of evidence evaluation of mode of action has developed through collaborative efforts of the US EPA, Health Canada, International Programme for Chemical Safety (IPCS), and the International Life Sciences Institute (ILSI). This approach uses a conceptual framework for assessing the weight of evidence for a postulated mode of action for a given toxic effect (including carcinogenicity). The IPCS/ILSI Framework is based on the Bradford Hill criteria. These criteria are used to evaluate the causal relationship between quantifiable key events in the animal that are necessary elements in the mode of action and tumor development. Once an understanding of the animal mode of action is established then the human relevancy can be assessed. This nest step involves a qualitative and quantitative concordance analysis of the key events between animal and human (e.g., consideration of comparative biology, kinetics/metabolism, anatomical variations, and relevant human disease states). The USEPA Office of Pesticide Programs has used this framework approach for several years with great success. In the case of pesticides that are rodent liver carcinogens and are not direct acting mutagens, two general of modes of action are observed, a receptor mediated pathway (e.g., PPARα agonism or CAR activation) or a non-receptor pathway (e.g., chronic cell injury/death and regenerative proliferation). Thiamethoxam, a broad spectrum insecticide that induces mouse liver tumors will be used as an example of the application of the IPCS/ILSI human relevance framework. This case study demonstrates how the framework is utilized, that it promotes the use of all relevant data (chemical specific and generic) and provides a structure to clearly present the evidence in a transparent manner. (This abstract represents the view of the author and does not necessarily represent the decisions or stated policies of the USEPA. Mention of trade names does not imply endorsement.)

Link to Slides:  http://www.toxicology.org/isot/rc/ncac/symposia_files/2_Dellarco_NCAC-SOT%205'08.pdf

Use of Transgenics in Carcinogenicity Testing

Richard Storer, Ph.D., Senior Scientific Director, Dept. of Laboratory Sciences and Investigative Toxicology, Safety Assessment, Merck Research Laboratories.

International guidelines allow for use of a short-term cancer bioassay (26 weeks) in transgenic mice as a substitute for one of the two required long-term rodent bioassays in the preclinical safety evaluation of pharmaceuticals. The substitution may be allowed, or encouraged, when preclinical safety and/or pharmacology data suggest that an alternative assay could provide additional information not likely to be obtained from a second long-term bioassay. Utilization of these models in preclinical safety evaluation can significantly reduce animal use, time and manpower. The two models which have gained general acceptance by both sponsors and regulators are the CB6 1-RasH2 mouse hemizygous for the human H-ras transgene and the B6.129N5-Trp53 mouse heterozygous for a p53 null allele. These models have shown potential to provide, in addition to a short-term tumorigenicity endpoint, insight into mechanisms involved in tumor induction. Utilization of the p53^+/-model is of particular value for compounds with residual concern that genotoxic activity may contribute to tumorigenesis; the rasH2 model however may be accepted as an alternative without regard to evidence of genotoxic potential. In considering the utility of the p53 model, many genetic toxicologists view the ICH test battery and available adjunct assays as sufficient for evaluating the potential risk of genotoxic carcinogenesis and for making decisions with respect to initiation of clinical trials. However, results from a p53 assay can make an important contribution to the weight of evidence assessment of mode of action (genotoxic vs non-genotoxic) from any tumor findings in the long-term bioassay. In addition, since results from a short-term bioassay can be obtained relatively early in drug development, there is the potential for more timely assessment of cancer risk for individuals in clinical trials. For the rasH2 model, responses to non-genotoxic carcinogens and the rationale for use of this model will be discussed in the context of the debate as to which classes of compounds in this category are most important to detect.

Link to Slides:  http://www.toxicology.org/isot/rc/ncac/symposia_files/3_Storer_NCAC-SOT_5'08.pdf

 
Hepatotoxicity Testing; Predictive Strengths and Weaknesses

Rich Miller, D.V.M, Ph.D, VP, Safety Assessment, GlaxoSmithKline

Hepatotoxicity is a significant cause of discontinuation of drugs in development, as well as withdrawal of marketed medicines. Hepatic structure and function are generally similar across species used for preclinical testing and humans; specifically, hepatic cell make up and major roles (e.g., metabolism, detoxification, energy storage, bile formation, endocrine and immune function) are conserved across species. These similarities underpin the rationale for use of preclinical species in identifying potential human liver hazards. However, there is clearly a need to improve our ability to predict hepatotoxicity outcomes early in drug development due to varying hepatotoxic sensitivities in diverse patient populations and the multitude of mechanisms that can lead to hepatotoxicity. Increasingly sophisticated in vitro, imaging and molecular biology approaches offer promise in this regard. Hepatic transcriptome alterations representing well annotated toxicity pathways interpreted in the context of clinical and anatomic pathology endpoints can bolster sensitivity and predictivity of preclinical studies. In vitro assays including cellular, enzymatic and binding assays such as those using primary hepatocytes or hepatic cell lines, CYP inhibition assays, and reactive metabolite signal detection assays (i.e., GSH trapping), may also aid in identifying potential hepatotoxic properties early and can supplement in vivo studies. In summary, improvements in hepatotoxicity detection and prediction can be made by complementing sound preclinical testing paradigms with additional relevant endpoints in standard in vivo studies, coupled with detection of undesirable biologic properties using a variety of in vitro assays.

Link to Slides:  http://www.toxicology.org/isot/rc/ncac/symposia_files/4_Miller_NCAC-SOT_5'08.pdf

 
The Generation and Use of Human Data in the Safety Evaluation of Pesticides, Personal Care Products, Food Ingredients and Pharmaceuticals

Gary Burin, MPH, Ph.D., D.A.B.T., Director, Toxicology and Risk Assessment, Technology Sciences Group Inc., Washington DC

Human data have played an important role in the safety assessment of chemicals found in a variety of products including pesticides, pharmaceuticals, personal care products and food ingredients. The types of intentional human studies that have been used in risk assessment include safety, efficacy and exposure studies. There is a well-established process for ensuring ethical conduct of these studies. In the absence of human data, risk to humans is assessed from animal models through the use of default values for interspecies and intraspecies sensitivity. Chemical Specific Adjustment Factors which are based upon human data have the potential to improve the accuracy and scientific basis of risk assessment by replacing the default values for interspecies and intraspecies uncertainty.

 
ToxCast: Developing predictive signatures of chemically induced toxicity.

Kavlock, R. J., Dix, D., Houck K, Judson, R. T. Knudsen, Martin, M. and Richard, A., National Center for Computational Toxicology, ORD, U.S. EPA, RTP, NC 27711.

ToxCast, the United States Environmental Protection Agency’s chemical prioritization research program, is developing methods for utilizing computational chemistry, bioactivity profiling and toxicogenomic data to predict potential for toxicity and prioritize limited testing resources. In the proof-of-concept phase, we are focused upon evaluating chemicals with an existing, rich toxicological database in order to provide an interpretive context for the high through put screening data. This set of 320 reference chemicals are largely derived from the active ingredients in food use pesticides and represent numerous structural classes and phenotypic outcomes, including tumorigens, developmental and reproductive toxicants, neurotoxicants and immunotoxicants. The goal of the program is to develop signatures based on the combined use of physico-chemical properties (the traditional independent variables in structure activity models) and the bioactivity data (derived from a broad spectrum of more than 400 readouts from biochemical assays, cell-based phenotypic assays, and genomic analyses of cells) that are predictive of responses in animal bioassays. The signatures derived for chemicals with toxicity data gaps could then be compared with those of the well characterized chemicals, and those with significant signatures would become priority candidates for testing in traditional animal bioassays. These data are being generated through a series of external contracts, and through collaborations within EPA and with the National Institutes of Health Chemical Genomics Center. Results of the proof of concept phase and supporting chemo-informatic infrastructure will be presented. This is an abstract of a proposed presentation.

Name:

______________________________________________________

Affiliation:

______________________________________________________

Address

_____________________________________________________

______________________________________________________

City:

__________________________________________

State:

_____ Zip Code: ____________

Area Code:

______ Phone: _________________FAX:____________________

E-mail:

______________________________________________________

Membership Type

______  Full Member  ($20) ______  Student  ($10)

SOT Member

_____  Yes

_____  No

Highest Degree Attained

_____  A.S.                 _____  M.P.H.

_____  B.A.                _____  M.S.

_____  B.S.                 _____  M.A.

_____  D.V.M.                        _____  Ph.D.

_____  D.V.M./Ph.D. _____  Sc.D.

_____  M.D.                _____  V.M.D.

_____  M.D./Ph.D.     _____  V.M.D./Ph.D.

Type of Affiliation

_____  Academia

_____  Consulting

_____  Contract Lab

_____  Government

_____  Industry- Chemical/Petroleum

_____  Industry- Pharmaceutical

_____  Industry- Other

_____  Other-  _____________