From: Owens, Willie [mailto:owens.jw@pg.com]

Sent: Wed 3/26/2008 11:42 AM

To: Charles Eldridge

Cc: Timm, Gary; Smith, Jane

Subject: Follow up on antiestrogen Question

Dear Charles -

Two relevant points to try and further address your question yesterday
about the antiestrogen.... 

Pls inform the EPA officers to ensure this communication is within their
rules and bounds.  I'm also cc'ing both Gary Timm and Jane Smith to
ensure this is transparent and, hopefully, above board.

1.		 The uterotrophic and Hershberger approaches to antagonists are
almost identical. 

		 a.		 Step 1 or Phase1 as to run in multiple labs the dose response
curves for the reference positives, EE and TP, respectively. 

		 b.		 Step 2 was to identify an approximate ED70 on those dose
response curves.  The rationale was basically in two parts: 

   i.      Due to the significant magnitude of the changes in many
tissues such as the uterus, ventral prostate, seminal vesicles, and
Cowper's glands, the EC70 would provide adequate power for detection -
and with the rather low CVs of the solid tissues like the LABC and GP,
even with their lower response, adequate power would also be available.

   ii.      If we went higher than an ED70, due to the inherent
variability, some laboratories were likely to have reached a maximum on
the plateau with some risk that they would then be so "max'd" that some
insensitivity could result.

		 c.		 Step 3 was to test out the "hypothesis" - a pure antiestrogen,
ZM 189,154, and antiandrogen, Flutamide, respectively.  Two reference
doses and a dose response series for the antagonist.  Both worked
nicely.  This confirmed two key points in both cases: a) a positive
reference agonist dose adequate to test the laboratory and its
conditions was agreed and b) the dose of the antagonist to use for
testing the viability and performance of the antagonist assay was
agreed.

		 		 Willie's conclusion - at this point, all evidence is good as I
noted.  But, one can also do a very reasonable plausibility extension -
that is, the Hershberger antagonist assay worked like a charm (e.g., the
linuron data reproduction slide that I showed).  Given our knowledge of
the uterotrophic, there is simply nothing that argues against it not
working as well or even better. 

    I mentioned in closing the very extensive Japanese data set that had
been built up for the uterotrophic.  What I forgot - and want to
communicate - the Japanese have performed BOTH the agonist and
antagonist assays - so THERE IS AN EXTENSIVE ANTAGONIST DATA SET.  AND
they have been rather careful to explore the relationships of the OECD
proposed assays in concert ... something the EPA proposal will effective
do for about 70 compounds.  I am attaching one of the publications where
the utero, H, and modified 407 are run in unison.  METI has released
their utero and H data sets within the OECD, but I am not sure that I
can release those to you. 

	 a.	See Table 1. 

	 b.	My observations are the 300+ sets of data that I have seen are: 

       i.      If the chemical is negative on the agonist, the bulk of
the data are negative on the antagonist (as one would expect).

       ii.     If the chemical is positive on the agonist, there is
almost always a partial agonist-antagonist finding - that is, at the
dose where the agonist increase is observed - there is a parallel
decrease (absolute and statistically significant) at the same dose in
the antagonist assay.  

       iii.    Together, these data would argue fairly persuasively that
the antagonist assay is valid.

       iv.     HOWEVER, as with the example and thus my deliberate
purpose - when the body weights are dropping 4-5% WITH THREE DAYS OF
ADMINISTRATION, SO VERY RAPID GIVEN THE SHORT DURATION - then one often
sees an absolute decline in the stimulated uterus.  So we have the very
difficult issue here as with the other assays - when do the overt
toxicities cause us to see confounding responses that are difficult to
interpret.

Willie's key communication and conclusion - there is a clear data set to
support the antiestrogen segment of the uterotrophic as valid with
indications of key limitations - there have been 3 labs as I understand
it involved in generating the Japanese data sets - and the tamoxifen
standard response would be available as with this Table to assess
reproducibility - but I DO NOT know if other test compounds have been
actually done in several labs to full support interlaboratory assessment
in a strict sense.

So I hope this helps - and hope that it is within EPA bounds.  No bad
intents here.

Best regards, Willie

