What Information Fulfills EDSP Screening Requirements?

1. What Information Fulfills EDSP Screening Requirements? Steve Levine, Ph.D. Science Fellow Ecotoxicology & Risk Assessment Monsanto Company ISRTP Meeting September 9, 2009

2. Topics Covered • Background on the EDSP • EDSTAC recommendations • Functional equivalence and Other Scientifically Relevant Information (OSRI) • International Regulatory Guideline Studies • TOX • ECOTOX • WoE approaches

3. Tier 1 Screening • The number of substances planned for Tier 1 screening is staggering • Pesticides • Commercial chemicals • Cosmetic ingredients • Food additives • Nutritional supplements • Certain mixtures • Therefore, screening must be efficient, robust, and predictive at identifying potential EACs

4. EDSTAC Framework • Priority setting • Based on potential human exposure • residues in food and drinking water • residential • occupational • Higher priority given to chemicals likely to have human exposure via multiple pathways • Tier 1 screening • In vitro and in vivo assays • Identify chemicals that can potentially interact with the endocrine system • Tier 2 testing • Will establish the relationship between dose of an EAC and the potential effects observed • Assess risks humans and wildlife

5. EDSTAC Recommendations • Recommended a minimal number of screens & tests to make sound decisions, thereby, reducing the time needed to make decisions • T1S screens should be MoA based to identify specific types of endocrine activity • Should be broadly predictive (sensitivity & specificity) • Should produce data that can be interpreted as either positive or negative (minimize Type I & II error rates) • Should be inexpensive and relatively quick and easy to perform

6. Tier 1 Screening Battery

7. Alternative Means to Meet Tier 1 Screening • EDSTAC recommended that it should be permissible to meet T1S requirements by submitting data that are “functionally equivalent” to the data generated by the T1S battery • Functionally equivalent information could be submitted for one or more of the recommended T1S assays or for the entire battery • Pointed out that functionally equivalent informationexists for chemicals that have reproductive and developmental toxicity testing

8. Functional Equivalence & OSRI • Functional equivalence = data of a suitable nature and quality even if different methods were used i.e., provides the same essential predictive information as T1S • OSRI = information that informs the determination as to whether the substance may have a similar effect produced by a substance that interacts with EAT systems • OSRI may either be (1) functionally equivalent to information obtained from the Tier 1 assays or (2) data from assays that perform the same procedure / function as EDSP Tier 1 Screens

9. EPA’s Approach for Considering OSRI • Overview paper @ www.regulations.gov • Anyone may submit OSRI not just Test Order Recipients • Submitters should provide the information or cite previously submitted information • Information will receive different weights in a WoE - the quality of the information will weigh heavily • Factors will include sensitivity, specificity, confidence in the conclusions and applicability across taxa • Factors will also include route & duration/frequency of administration, dose levels, age at exposure, species, number of test units, variability, and whether the data provides a basis for conclusions for potential interaction with the endocrine system in mammalian and non-mammalian systems • OSRI for Tier 1 or Tier 2: metabolism, carcinogenicity, reproductive and developmental, and toxicogenomic

10. OSRI from OPPTS and OECD Tests • TOXICOLOGY • Sub-chronic (OPPTS 870.3100) • Chronic (OPPTS 870.4100) • Developmental/Teratology (OPPTS 870.3700) • 2-gen reproduction (OPPTS 870.3800) • ECOTOXICOLOGY • Avian reproduction (OPPTS 850.2300) • Chronic invert reproduction (OPPTS 850.1300/1350) • Early Life-stage (OPPTS 850.1400) • Fish full life-cycle (OPPTS 850.1500)

11. OSRI Will Include • Toxicity tests following international regulatory guidelines where full histopathology is carried out on: • Endocrine tissues (pituitary, thyroid, parathyroid, pancreas adrenal, ovary and testis) • Hormone sensitive male tissues (prostate, epididymides, seminal vesicles) • Hormone sensitive female tissues (mammary, uterus and reproductive tract) • And toxicity protocols that focus on reproduction, fertility and development for mammalian species and other taxa (birds, fish) • These studies provide the strongest evidence of potential endocrine effects and should be weighted above less comprehensive data

12. Rat Multi-Gen • The 1996 USEPA guideline for a multiple generation reproductive and developmental toxicity assay was revised to include developmental benchmarks predictive of ED potential • A diverse set of female endpoints (ovaries, uterus, vagina, and mammary glands) • Day of vaginal opening and first estrus • Mating and fertility indices • number of implantation sites, estrous cyclicity • Male tissue weights and histopathology (testes, epididymis, prostate, and seminal vesicles) and other male parameters (sperm number and analyses)

13. Bridging OSRI to T1S Assays • High concordance between the results of rat multi-gen and the rat Uterotrophic assay • Consistency between MEDs for the Uterotrophic assay and estrogen-related LOEL/LOAEL in multi-generation testing • High concordance between the Uterotrophic assay and the ER binding assay & the stably transfected TA for a large number of chemicals • 2-gen data should meet the requirement for the pubertal assays: • Pubertal assays largely have redundant endpoints with the 2-gen study

14. OSRI to T1S Assays - ECOTOX • Avian reproduction studies • Gold standard or highest tier of avian testing • Bobwhite quail and mallard duck • Studies do not include exposure during all relevant stages of development or organ histopathology • However, provides valuable information in a WoE context to assess potential effects of endocrine active substances • Sensitivity of 1-gen vs 2-gen Japanese quail? • Reproduction study design • Dietary exposure for ~10 weeks prior egg-laying • Photoperiod change initiates egg-laying • Exposure continues for ≥8 weeks • Endpoints: eggs laid, eggs damaged, eggs set, egg shell thickness, viable embryos, live embryos, hatchlings, 14-day survivors, eggs laid/female, eggs laid/female/day, 14-day survivors/female

15. OSRI to T1S Assays - ECOTOX • Fish full life cycle (FFLC) study • Gold standard for aquatic vertebrate ecotox testing • Entire LC is exposed = most sensitive life-stage is tested • ~260 to 300 days for FTHD minnows = T1S spp. • FFLC study design • ELS → growth & repro → ELS • First gen: hatching success, time-to-hatch, survival, growth and development, reproductive success (# of spawns, # of eggs) • Second-gen: time-to-hatch, survival, growth and development

16. Principles for Evaluating Data in a WoE. • Consistent pattern of responses for a MoA (+ or -) • The nature and range of the biological effects observed • The shape of DRCs curves when available • The severity and magnitude of the effects • Interpretation made in the context of biological significance & biological plausibility • The presence or absence of responses in multiple taxa • Evaluate results in the context of natural variability using control data (historical and concurrent)

17. Principles for Evaluating Data in a WoE • Results from apical assays (MTD dose) need to be weighted appropriately when accompanied by decreases in BW or other potential confounders • In vivo results generally are considered to have more weight than in vitro results • Available in vitro assays should not be used as yes/no determinants to proceed to Tier 2 (ER binding example) • SARs = critical step in the WoE process • A WoE approach for OSRI must be developed along with the WoE framework for the T1S

18. Closing Thoughts • SARs are OSRI and weighted appropriately • There will be instances where OSRI will meet all or some of the T1S requirements, particularly for food use pesticides. • Development of a transparent WoE approach is desperately needed not only in the interpretation of OSRI but for the T1S data before testing is required.