EPA issues policy on use of modern non-animal approaches for assessing endocrine disruptors

The EPA Endocrine Disruptor Screening Program (EDSP) is full of good news this week! Hopefully you saw my earlier post about acceptance of ToxCast data to satisfy the estrogen binding data requirement for certain chemicals. You may want to go read that first, if you have not done so already.

In tomorrow’s Federal Register, EPA will announce a new policy regarding “how EPA is planning to incorporate an alternative scientific approach to screen chemicals for their ability to interact with the endocrine system.”  This Federal Register announcement is to accompany the scientific support published last week by EPA in the Browne et al (2015) publication in Environmental Science & Technology

In essence, the information is the same … EPA has developed an approach to more rapidly screen the thousands of chemicals it needs to screen under EDSP (see the “EDSP Universe” here) in more cost-effective manner that will allow for mass prioritization of chemicals for subsequent testing that will reduce use of animals in toxicological testing.  The approach involves use of high-throughput toxicology data (generally, in vitro and in chemico/biochemical assays) generated through the ToxCast effort and integrating the relevant results into a computational “ER Model” (an in silico approach) to allow for individual chemicals to be scored for overall “bioactivity.”  The first such effort in this regard is described in detail in the Browne et al. paper.  Of over 1800 chemicals tested in the ER Model, over 1700 had “very low or no detectable ER bioactivity.” (As a complete aside, given that this entire 19-year very-slow-moving novel regulatory program was essentially founded on the notion that environmental exposure to xenobiotics was perturbing the estrogen pathway, its fabulous news that the vast majority of chemicals screened to date do not show appreciable estrogen bioactivity.)  It should be noted that the approach was vetted by the EPA’s Scientific Advisory Panel devoted to FIFRA, the federal pesticide law, in December 2014.  Based on this review and the Browne et al (2015) publication, EPA has concluded that “ER Model data are sufficient to satisfy the Tier 1 ER binding, ERTA and uterotrophic assay requirements.”  Even better, in future test orders issued under EDSP, the recipient will be allowed to cite the ER Model data as “Other Scientifically Relevant Inforamtion” (OSRI) or generate data using the 18 assays from ToxCast that were integrated into the ER Model or generate the ER binding, ERTA and uterotrophic assays per existing guidance.

Further advancements of in vitro, in chemico and in silico approaches will allow for similar efforts to be undertaken for the androgen and thyroid pathways as well.  In addition, this approach will hopefully establish proof-of-principle before wider application, such as to industrial chemicals under TSCA (which of course, is likely to undergo transformation as well in near future … see here and here for examples … but that is a different story). Expansion of this work will continue in several notable ways:

1.       EPA is still researching an androgen analogue to the ER Model (“AR Model”). When done and validated, this can be expected to fully replace the AR binding assay and Hershberger assay and partially replace the male rate pubertal and fish short-term reproduction study.

2.       Steroidogenesis and thyroid bioactivity models are also underway, which would fully replace the aromatase/steroidogenesis assays and amphibian metamorphosis assay, respectively.  They would also help partially replace the pubertal assays and the fish short-term reproductive study.

The eventual goal is to develop a set of non-animal high-throughput methods and models to fully replace all 11 Tier 1 assays.  It has been estimated that each compound tested in the full Tier 1 battery uses at least 600 animals and costs at least $500,000 overall, no including time, costs and resources that companies must devote to preparation on the regulatory submittals, review of the data prior to submission, development and submittal of OSRI, etc.  

EPA is soliciting public comments on the policy for at least 60 days from tomorrow (I can foresee the deadline getting extended for another 30 days, upon request from the regulated community and other interested parties).  Specifically, they are looking for comment on use of ToxCast/ER Model to replace the ER binding, ERTA and uterotrophic assays as well as comment on use of the screening results for the 1800 chemicals in a similar fashion.