The Uterotrophic Assay

1
The Uterotrophic Assay
The OECD Validation Program and More A Case
Study in Validation
J. William Willie Owens The Procter Gamble Co.
2
Overview of Presentation
Illustrate All the Parts of a Validation
Program Illustrate Standardization of
Protocol Illustrate Conduct and Logistics of
Study Illustrate Variety of Issues that
Arise Illustrate that Nothing is Perfect
3
The Scientific Rationale

• Estrogen naturally regulates the growth of
• several female tissues uterus, vagina, and
• mammary gland
• Estrogen stimulated growth of uterus is cyclic,
• rapid, and dramatic in laboratory rodents
• 4-4.5 days total cycle, estrogen phase 2 days
• uterus does not reach min or max in cycle,
  dynamic range increases to 5-6 fold without
  cycle and minimal uterine weight

4
The Scientific Rationale

• Rodent uterus presents opportunity for rapid
• assay for estrogen agonists and antagonists
• The conditions and data
• Low endogenous levels of estrogens
  (immature or ovariectomized female)
• 3 days administration, sacrifice on 4th day
  (twice the time of natural estrogen growth)
• Quantitative data continuous gravimetric
  weight (not qualitative score or quantal
  categories, e.g., vaginal cornification
  assay)

5
The Uterotrophic Assay
The History early 1930s
Need to identify estrogen follow biological
activity through chemical fractionation Origina
l pharmaceutical needs assay possible
natural and synthetic estrogens Vaginal assay
subjective and quantal but castrated male
assay, used weights of responsive tissues in
the reproductive tract
6
The Uterotrophic Assay
The History early 1930s
At least 6 laboratories published basic
protocols Rat and mouse Immature intact,
immature OVX, adult OVX Consecutive
administration for 3-4 days Sometimes with
imbibed fluid, sometimes blotted weights Other
variations Primary focus was POTENT estrogens
7
The Uterotrophic Assay
The History 1970s
Estrogen receptor discovered Antagonists
discovered Compound discovery uses receptor
assay first Promising agents (still potent)
proceed to the uterotrophic assay (conserve
animals/time) Antagonist assay coadminister
test agent with potent reference, look for
decrease in weight versus reference only
group Research on complementary molecular
/cellular markers
8
The Uterotrophic Assay
The History 1990s
Estrogen mimics emerge as issue Typically, weak
agonists 2-5 orders of magnitude lower in
binding affinity than 17-beta estradiol Developmen
t of in vitro assays (reporter genes) Development
of new molecular and cellular markers Reapplicati
on of pharmaceutical paradigm Addition of in
vitro techniques to tiered scheme Attempts to
relate markers to uterine weight increase Begin
application of assay to very weak agonists
9
Types of Data
Mechanistic Data The endpoint or measurement is
linked mechanistically through a mode of
action to an effect or condition Empirical
Data An association is demonstrated between the
endpoint to an effect or condition Adverse
Effect Endpoint used for risk assessment
10
The Uterotrophic Assay What is its model and
prediction?
The growth of the uterus is a culmination of an
estrogen mode of action A cascade of molecular
and cellular events that begins with a ligand
binding the estrogen receptor (Mechanistic
data) Good in principle concordance with
binding data Antiestrogens block
response Response absent in alpha-estrogen
receptor knock out mice
11
The Uterotrophic Assay What is its model and
prediction?
Basic prediction
Positive uterine growth indicates that substance
may trigger estrogen mediated effects Testing
for adverse affects may be warranted If a
relevant route of administration is used, the
LOEL may provide helpful information on the
expected potency
12
Scientific Rationale Regulatory Use
In principle, we have a match up where
the uterotrophic assay can deliver the
information that is needed by the regulators.
13
The OECD Program
Phase I Standardize common protocols Demonstrate
it works with reference estrogen Refine as
necessary Note Validation works with a given
protocol With several versions OECD accepted
the need to work with several protocols to
demonstrate their characteristics
14
The OECD Program

• Phase I
• Rat as species of choice (mouse later?)
• EE (ethinyl estradiol not 17beta-estradiol as
  it binds avidly to alpha-fetoprotein)
• Two basic versions intact, immature ? OVX ?
• Two routes oral gavage and sc injection
• Also test extension of dosing to 7 days (from 3)
• Both wet (imbibed fluid) and blotted (no fluid)
  uterine weights
• Record body weights (toxicity, statistics)
• Voluntary expts. (evaluate histology, BrdU)

15
The OECD Program

• Immature, Intact Female
• - Start pnd 17-20 (birth pnd 0)
• Once per day, 3 consecutive days, same time
  each day
• Sacrifice, 24 hours after last dose
• Both oral gavage and sc injection expts
• n 6

16
The OECD Program
Uterine Weight Long Evans Rats Showing Window
for Stable Uterine Weights
100
Entering puberty E2 levels rising
Low E2 levels Stable weight low SDs
80
60
Mean Uterine Weight (mg)
40
20
0
PND 20 21 22 23 24 25
26 27 28 29 30 31
32 n 10 9 10 10 9 13
10 9 10 10 10 10
13
Postnatal Day
Example from Lichtensteigers lab Univ of Zurich
17
The OECD Program

• OVX Female
• - OVX 6-8 weeks of age (42-56 days)- Allow 7
  days to regress
• Once per day, 3 consecutive days, same time
  each day
• Satellite protocol using 7 day dosing
• Sacrifice, 24 hours after last dose
• Sc injection only
• n 6

18
Uterine Weight (mg)
19
The OECD Program

• Phase 1 Conclusions
• Basically, success!
• No evident difference in versions (none
  superior)
• Expected difference for EE in po vs sc
• Blotted weight slightly more sensitive
  (less variable)
• Some animal husbandry issues with immature
• Extend OVX regression to 14 days
• No impact from strain, diet, etc.
• Overall consistency
• Adequate power with n 6

20
The OECD Program
OVX Control Blotted Weights
21
The OECD Program

• Phase 1 Observations
• Lab 19 possibly less sensitive, no evident
  reason
• Antagonists (ZM 189,154) demonstration was
  successful
• Slightly greater induction of blotted weight
  with immature animals 5-fold vs 3.5 fold
• Imbibed fluid possibly greater in OVX _at_ 3 days
  (but far less imbibed fluid _at_ 7 days)
• Model dose curve with Hill equation ED10
  supports statistical analysis

22
The OECD Program

• Phase II
• Two segments Dose response Coded
  (blind) Multichemical
• Demonstrate capability to detect weak agonists
• Demonstrate reproducibility (both EE and weak)
• Continue testing of protocol capabilities
• Continue wet and blotted uterine weights
• Continue body weights (toxicity, statistics)
• Voluntary experiments (histology, BrdU)

23
The OECD Program Chemical Selection

• Weak Agonists
• Bisphenol A
• Genistein
• Methoxychlor
• Nonylphenol
• o,p-DDTNegative Chemical
• Dibutylphthalate

24
Chemical Selection ER Binding Data
25
The OECD Program Bisphenol A
26
The OECD Program Genistein
27
The OECD Program Methoxychlor
28
The OECD Program Nonylphenol
29
The OECD Program o,p-DDT
30
The OECD Program Dose Response Summary
All weak agonists were detected Protocols were
qualitatively similar Includes one po study
with OVX Route of administration influence
substance-specific for lowest dose that is
statistically significant BPA GN / NP
/ MX o,p-DDT Blotted weights again slightly
less variable Magnitude of response varied
(partial agonists)
31
The OECD Program Dose Response Summary

• However, all was not perfection
• - One lab did not record terminal body weights
  Therefore, not included in analysis
• Several instances statistical significance was
  not recorded
• One discordant data set

32
The OECD Program Dose Response Summary
5 out of 83, no statistical significance BPA,
NP, o,p-DDT the weakest agonists 4 cases
labs did only 3 intermediate doses - highest
dose missing - uterine weights were rising
33
The OECD Program Dose Response Summary
34
The OECD Program Dose Response Summary
mg/kg/d
35
The OECD Program Immature Vehicle Control Weights
36
The OECD Program Coded Multichemical - Design
All Samples are Coded and Prepared at Repository
so Dilutions at Laboratories are Same Two EE
Doses Included - Compare to same doses in Phase
1 (Reproducibility over time) - Same EE doses
used in Dose Response (Reproducibility over
time not simultaneous) One Dose to One from
Dose Response for Each Weak Agonist - 3rd
or 4th dose from dose response (Reproducibility
over time) One Dose of Negative Chemical
37
(No Transcript)
38
The OECD Program Coded Multichemical
39
The OECD Program Coded Multichemical
Laboratories replicated their EE results -
Between Phase 1 and Phase 2 (where
laboratories were in both phases) - Within
Phase 2 where duplicates were done
40
The OECD Program Coded Multichemical

• Two sets of EE problems
• Some ambiguity in directions
• High and low EE doses
• - The high EE concentration was the low
  dilution
• - The low EE concentration was the high
  dilution
• Apparently, some labs misreported the doses.
• Four discovered this when requested to review
  records
• One could not resolve the issue in reviewing
  records
• A second discovered a dilution error, where
  low EE
• concentration was several fold lower than
  intended

41
The OECD Program Coded Multichemical
2. One laboratory repeated its insensitivity
with EE in Phase 1 in Phase 2.
Protocol C adult OVX model, subcutaneous
injection
This one lab did not observe statistical
difference with coded EE in Phase 2
0.0
0.01
0.03
0.1
0.3
1.0
3.0
10.0
42
The OECD Program Coded Multichemical Results
Compound
43
The OECD Program Coded Multichemical
Multichemical doses were either 3 or 4
out of five from dose response expts Several in
lower portion of dose response curve Thus, in
both dose response and multichemical, several
laboratories did not achieve statistical
significance (but this pattern repeated in
both sets of experiments)
44
The OECD Program Coded Multichemical
Bisphenol A Protocol A 600 mg/kg/d Dose
Response 3 pos/1 neg
Multichemical 6 pos/4 neg
1 negative 3 animals
died n3 Nonylphenol Protocol B 80
mg/kg/d Dose Response 6 pos/3 neg
(4 pos _at_ 80) Multichemical
11 pos/4 neg 1
negative Lab 12 with high
control uterine wts
45
The OECD Program Coded Multichemical
Results with negative chemical
dibutylphthalate 36 total experiments in 4
protocols 3 events with statistically
significant increase 2 events with
statistically significant decrease All within
/- 40 weight of controls Implication there
will be possible instances of both false
positives and false negatives due to group to
group variability
46
The OECD Program Coded Multichemical
Other observations Dose preparation issues
Hydrophobic substances difficult to prepare
Heating and mechanical techniques were needed
to achieve dispersion (not complete
dissolution) Two laboratories reported
inability to prepare Toxicity observed in some
laboratories Deaths observed and euthanasia
required with some chemicals
47
The OECD Program Phytoestrogen Dietary Analysis
Concerns over phytoestrogens in lab
diets Rare instance of high control uterine wts
leading to insensitivity of bioassay Linked
to high phytoestrogen content of diet All
laboratories retained sample of diet Diets
analyzed (genistein, daidzein, coumestrol)
48
The OECD Program Phytoestrogen Dietary Analysis
Control Uterine Weights and Dietary Phytoestrogen
49
The OECD Program Phytoestrogen Dietary Analysis
Control Uterine Weights and Body Weight
50
The OECD Program Laboratory Strain
Concern over sensitivity of rodent strain In
this study, laboratories reported using Sprague
Dawley and Wistar rats (both supply houses
and in-house sources) No difference in uterine
weights of control animals attributable to
strain No difference in responses to EE or any
of the weak agonists attributable to strain
51
Willies Comments

• Laboratories should demonstrate capability
  withEE and at least one agonist (GN, NP, BPA)?
  one insensitive lab, no technical explanation
• Guidance is needed dose preparation/insolubility
  
• Data acceptance criteria are needed
• ExamplesControl uterine weights
  immature and OVXToxicity (MTD and animal welfare)

52
Conclusions

• Protocols have been standardized
• Protocols replicate and are consistent with high
  potency reference estrogen EE
• Protocols replicate and are consistent withfive
  low potency weak estrogen agonists
• Number and diverse competency of labs supportthe
  robustness (but not perfection) of the bioassay

53
Conclusions

• Strong scientific rationale for the bioassay
  - Uterus is a natural and responsive target -
  Substantial evidence that uterine growth is
  linked to estrogen mechanism
• Final Statistical Analysis is Still Underway!
• Peer Review Planning is in Progress
• OECD Test guideline intensive review comment