1. 2 n McKim Workshop on Reducing Data Redundancy in Cancer Assessment | 8 – 10 May 2012 | Baltimore, MD Highlighting the Need for AOPs in Streamlining Hazard Assessment Methods (for cancer assessment) Catherine Willett, PhD Director, Regulatory Toxicology Humane Society of the United States

2. 2 nd McKim Workshop on Reducing Data Redundancy in Cancer Assessment | 8 – 10 May 2012 | Baltimore, MD 1.Rodent Cancer Bioassay: cost/benefit? 2.Advantages of mechanism based approaches 3.AOP approaches and activities at OECD Outline

3. 2 nd McKim Workshop on Reducing Data Redundancy in Cancer Assessment | 8 – 10 May 2012 | Baltimore, MD Minimally 400 animals, 2 – 4 million USD and 3 years* Often performed in two species (drugs, food additives) Or rat bioassay plus shorter-term transgenic mouse assays 1 –ras H2 + p 53 +/- : use half the number of animals and for 6 months –Reduced animal use mitigated by need for creation and maintenance of multiple lines –Only improves identification if both used (geno and non-genotoxic)  no animal savings, minimal cost savings *Thomas RS, Pluta L, Yang L, Halsey TA. Toxicol Sci 2007. 97(1):55-64. 1 Alden et al. Veterinary Pathology. 2011. 48(3):772-784. 1. Rodent Cancer Bioassay: Cost

4. 2 nd McKim Workshop on Reducing Data Redundancy in Cancer Assessment | 8 – 10 May 2012 | Baltimore, MD Drugs (review of 533) –False positives: 80% –False negatives: 27% –Sensitivity: 73% 12 out of 44 chemicals with human concern not seen in either rodent species “When tested repeatedly, most molecules….had conflicting test results.” Alden et al. Veterinary Pathology. 2011. 48(3):772-784 1. Rodent Cancer Bioassay: Benefit?

5. 2 nd McKim Workshop on Reducing Data Redundancy in Cancer Assessment | 8 – 10 May 2012 | Baltimore, MD Other chemicals (e.g. NTP studies) –82% of 500 studies resulted in noncommittal classifications* –Only 57% of chemicals tested multiple time give consistent results –22% of all chemicals tested test positive 1 Largely due to findings at MTD and irrespective of biological plausibility Nevertheless results in classification as “possible human carcinogen” –Positive correlation? 9/10 known human carcinogens have tested positive either rats or mice in NTP studies All known human carcinogens have tested positive in some rodent assay, with the possible exception of arsenic “This result says more about the persistence of toxicologists than about the ability of a standard (rodent) protocol to predict human carcinogenicity.” 1 * Wasted Money Wasted Lives. People for the Ethical treatment of Animals. 2006. http://www.mediapeta.com/peta/pdf/Wasted-money-PDF.pdf http://www.mediapeta.com/peta/pdf/Wasted-money-PDF.pdf 1 Ennerver and Lave. Reg. Toxicol. Pharma. 2003. 38: 52-57 1. Rodent Cancer Bioassay: Benefit?

6. 2 nd McKim Workshop on Reducing Data Redundancy in Cancer Assessment | 8 – 10 May 2012 | Baltimore, MD 400 animals 2 – 4 million USD 3 years 1. Rodent Cancer Bioassay: Cost / benefit 2 False Positives: 80% False Negatives: 27% or 82% noncommittal classifications Too high

7. 2 nd McKim Workshop on Reducing Data Redundancy in Cancer Assessment | 8 – 10 May 2012 | Baltimore, MD 2. Advantages of mechanistic approaches 1.Refine design and interpretation of animal studies 2.Demonstrate human relevance (or not) 3.Improve predictivity for both human health and other target species 4.Design assessment strategies that are not dependent on animal testing

8. 2 nd McKim Workshop on Reducing Data Redundancy in Cancer Assessment | 8 – 10 May 2012 | Baltimore, MD 2. Early adoption of mechanistic approaches DNA reactive vs non-DNA reactive –Mouse transgenic lines –3Rs value highly debatable –Chemical alerts for DNA reactivity –Used to flag chemicals –Genotoxicity battery –Currently used to exclude chemicals –In vitro tests have high false positive rates –in vivo tests are insensitive –Battery misses non-DNA reactive chemicals –Cell transformation assays

9. 2 nd McKim Workshop on Reducing Data Redundancy in Cancer Assessment | 8 – 10 May 2012 | Baltimore, MD 2. Early adoption of mechanistic approaches Human Relevance Frameworks -Characterize MoA of each class of carcinogens -Determine which rodent MoA is possible relevant to humans -Built using case studies Timeline stolen from V. Dellarco: EPA & IPCS 1999-2001. Conceptual Framework for Evaluating a Mode of Action for Chemical Carcinogenesis. ILSI 2003. Framework for human relevance analysis of information on carcinogenic modes of action. ILSI 2005. Extends Framework to non-cancer outcomes & life stage information. IPCS 2006 & 2008 Adopts Human Relevance Framework: Boobis, et al. IPCS framework for analyzing the relevance of a non-cancer mode of action for humans. Crit Rev Toxicol. 2008;38(2):87-96.

10. 2 nd McKim Workshop on Reducing Data Redundancy in Cancer Assessment | 8 – 10 May 2012 | Baltimore, MD Short-term screening tests * –Step 1: 90-day “screen” E.g. hepatocarcinogenicity: hepatocellular necrosis, hypertrophy, cytomegaly, increased liver weight –Step 2: mechanistic screens (most 90-days) Histopathology Serum enzymes Acyl Co-A oxidase CYP induction CAR, PSR, AHR binding ER binding (or histologic evaluation of endocrine-sensitive organs) Iron staining Reversibility Metabolic activation *Cohen, S.M. Toxicol. Pathol. 2010. 38(3):487-501. Proposals to improve efficiency of rodent-based assessment

11. 2 nd McKim Workshop on Reducing Data Redundancy in Cancer Assessment | 8 – 10 May 2012 | Baltimore, MD Testing strategy to rule out carcinogens * –Negative genotoxicity in standard battery –Negative hormonal perturbation in chronic studies effects on endocrine-related tissues, hormone levels –Lack of histopathologic risk factors in 6 month rat study neoplasia in any tissue –Negative in 6 month transgenic mouse Both genotox and non-genotox models required? –Positives go into a standard rat bioassay * Sistare et al. Toxicol. Path. 2011.39: 716-744. Proposals to improve efficiency of rodent-based assessment

12. 2 nd McKim Workshop on Reducing Data Redundancy in Cancer Assessment | 8 – 10 May 2012 | Baltimore, MD 2. Molecular approaches Gene expression profiling following 13 week exposure* –Predict lung tumors in B6C3F1 female mice –Overall accuracy 77.5% in predicting mouse lung tumors –25 chemicals sufficient to develop predictive model (could create predictive models for 8 organs using 200 chemicals) –Gene profiles placed into biologic process categories 1 Used to calculate BMD values for liver or lung Some transcriptional and tumor incident BMD values correlate well  transcriptional BMD values could potentially be used as points of departure for cancer as well as non-cancer risk assessment * Thomas, R.S. et al. Toxicol. Sci. 2009. 112(2): 311-321. 1 Thomas, R.S. et al. Toxicol. Sci. 2011. 120 (1): 194-205. Could the predictive capacity of ‘omics be improved by AOPs?

13. 2 nd McKim Workshop on Reducing Data Redundancy in Cancer Assessment | 8 – 10 May 2012 | Baltimore, MD 3. Adverse Outcome Pathway Approaches Adverse Outcome Pathway: a chemical and biological description of what occurs when a substance interacts with a living organism and results in an adverse reaction – a biological map from the initiating event through the resulting adverse outcome that describes both mechanism and mode of action. From: Ankley et al. Environ.Toxicol.Chem. 2010. 29 (3): 730–741.

14. 2 nd McKim Workshop on Reducing Data Redundancy in Cancer Assessment | 8 – 10 May 2012 | Baltimore, MD 3. Adverse Outcome Pathway Approaches AOPs can be useful for: Near-term: –Developing chemical categories and structure activity relationships –Increasing certainty of interpretation of both existing and new information –Developing integrated testing strategies that maximize useful information gained from minimal testing Longer-term: –Identifying key events for which non-animal tests can be developed, thereby facilitating mechanism-based, non-animal chemical assessment –Creating predictive toxicological assessments with low uncertainty and high human relevance

15. 2 nd McKim Workshop on Reducing Data Redundancy in Cancer Assessment | 8 – 10 May 2012 | Baltimore, MD Adverse Outcome Pathway QSAR focus area Chemicals Receptor Binding ER Binding Liver Cells Altered Protein Expression Vitellogenin Liver Altered proteins Gonad Ova-testis; Sex- reversed; Fecundity Sex reversal; Altered behavior; Repro. ER-mediated Reproductive Impairment In vivo MOLECULAR Target CELLULAR Response TISSUE/ORGAN INDIVIDUAL Skewed Sex Ratios; Yr Class POPULATION In vitro Assay focus area Toxicity Pathway 3. AOP activities at OECD: expert consultation on the ER decision framework in Feb 2009 P. Schmieder, McKim conference 2008.

16. 2 nd McKim Workshop on Reducing Data Redundancy in Cancer Assessment | 8 – 10 May 2012 | Baltimore, MD 3. AOP activities at OECD 2010 Workshop on using mechanistic information in forming chemical categories near-term recommendations: 1) Develop AOPs for well-established effects (e.g., skin sensitization) as well as several longer term health and ecotoxicological endpoints. 3) Establish, populate and maintain an accessible, electronic repository e.g. Effectopedia. 4) Develop a strategic plan including : –an information template for developing and assessing AOPs –guiding principles for assessing completeness and acceptance of an AOP –a format for attaining mutual acceptance of an AOP. 5) Harmonize terminology associated with AOPs. 6) Integrate AOPs in the OECD QSAR Toolbox.

17. 2 nd McKim Workshop on Reducing Data Redundancy in Cancer Assessment | 8 – 10 May 2012 | Baltimore, MD 3. AOP activities at OECD: Sensitization draft AOP OECD 2011. (Draft) The Adverse Outcome Pathway for Skin Sensitisation Initiated by Covalent Binding to Proteins

18. 2 nd McKim Workshop on Reducing Data Redundancy in Cancer Assessment | 8 – 10 May 2012 | Baltimore, MD 3. AOP activities at OECD Agreed at last Joint Meeting in June 2011 that AOPs were to become the cornerstone for all future projects in the test guidelines programme. Under development: A process for development and maintenance of AOPs A Guidance Document for Developing and Assessing Completeness of AOPs

19. 2 nd McKim Workshop on Reducing Data Redundancy in Cancer Assessment | 8 – 10 May 2012 | Baltimore, MD Thank you. Catherine Willett, PhD Director, Regulatory Toxicology, Risk Assessment and Alternatives kwillett@humanesociety.org kwillett@humanesociety.org t +01.240.599.6785 The Humane Society of the United States 2100 L Street NW Washington, DC 20037 humanesociety.org/animalsinlaboratories humanesociety.org/animalsinlaboratories