Prevalidation
and
Validation
Study
Plan
for
Minced
Testes
Assay
Gary
Timm
Presented
to
EDMVS
June
4,
2003
Objectives
°
To
assess
relevance
of
minced
testes
assay
for
detecting
compounds
that
affect
steroidogenesis
 
Measure
change
in
testosterone
production
relative
to
controls
°
To
assess
reliability
 
Measure
variability
in
testosterone
measurements
of
participating
laboratories
 
Measure
variability
in
mean
response
among
participating
laboratories
Data
Interpretation
°
Assay
will
detect
interference
with
key
steps
in
the
steroidogenic
pathway:

 
Decreases
in
steroidogenic
signal
transduction
 
Interference
with
the
transport
of
cholesterol
from
the
cytoplasm
to
the
mitochondria
via
StAR
 
Inhibition
of
enzymes
involved
in
the
conversion
of
cholesterol
to
testosterone
C27
HO
Cholesterol
Pregnenolone
C21
HO
O
Dehydroepiandrosterone
C19
HO
O
Androstenedione
O
O
O
O
Progesterone
17 ­
Hydroxyprogesterone
O
O
OH
C21
HO
O
OH
17 ­
Hydroxypregnenolone
P450scc
3 HSD
P450c17
P450c17
O
HO
H
Dihydrotestosterone
C19
OH
HO
C
18
Estradiol
HO
O
Estrone
TESTOSTERONE
OH
O
C
19
17KSR
P450arom
P450arom
5 ­
RED
Testis
Ovary
Theca
Cells
(
Females)

Granulosa
Cells
Testis
and
Peripheral
Tissues
Leydig
Cells
Intracellular
Compartment
Corpus
Luteum
Mitochondria
Cytoplasm
Data
Interpretation
(
2)

°
Interference
with
steroidogenesis
will
result
in
a
decrease
or
increase
in
measured
testosterone
relative
to
controls
°
Assay
is
unlikely
to
detect
 
Inhibition
of
aromatase
 
5 ­
reductase
inhibitors
Inhibition
of
these
downstream
steps
would
result
in
an
increase
in
testosterone
production
Basic
Protocol
°
Optimization
study
will
determine
parameters
of
protocol
°
Replicate
Runs
 
3
with
hCG
stimulation
 
3
without
hCG
stiulation
°
Dosing
 
3
dose
levels
 
1
positive
control
 
Media
control
°
Sample
3­
4
time
points
Prevalidation
Studies
°
Purpose:

 
To
obtain
initial
information
on
protocol
transferability
 
Primary
test
of
relevance
°
Studies:

 
Protocol
optimization
study
 
Baseline
study
 
Pilot
study
 
Multichemical
study
Prevalidation
Studies
°
Optimization
of
minced
testes
protocol
has
been
completed
°
Baseline
study
 
Two
labs
to
run
optimized
protocol
 
3
runs
without
hCG
 
3
runs
with
hCG
challenge
 
Measure
testosterone
formation
and
LDH
 
No
test
chemical
 
3
replicates
Prevalidation
Studies
(
2)

°
Pilot
studies
 
Aminoglutethimide
(
positive
control)

 
Ethane
dimethanesulfonate
(
Leydig
cell
toxicant)

 
Two
labs
 
Three
replicates
°
Multichemical
studies
 
9
challenge
chemicals
 
Two
labs
 
Two
replicates
Selection
of
Reference
Chemicals
°
Selected
for
known
mode
of
action
°
Limited
by
availability
 
Pharmaceuticals
are
difficult
at
best
to
procure
because
the
require
material
transfer
agreements
 
Many
pharmaceuticals
are
not
available
from
the
manufacturer
°
Thus,
we
will
have
to
duplicate
many
of
the
chemicals
in
validation
that
were
used
in
prevalidation,
if
we
want
to
cover
modes
of
action
Reference
Chemicals
X
Aromatase
Fenarimol
X
X
Leydig
cell
toxicant
EDS
X
X
StAR
inhibitor
Dimethoate
X
Inhibits
C­
AMP
Bisphenol
A
+
Cont
+
Cont
Aromatase
P450scc
Aminogluteth
imide
Val
Preval
Mode
2
Mode
1
Chemical
Reference
Chemicals
X
X
Negative
chemical
Vinclozolin
?

5 
reductase
MK­
434
X
Inhibits
C­
AMP
Lindane
X
Aromatase
P450scc
Ketoconazole
X
X
3 ­
HSD
inhibitor
Genistein
X
P450c17
Flutamide
Val
Preval
Mode
2
Mode
1
Chemical
Selection
of
Laboratories
°
Laboratories
to
be
selected
by
the
Contractor
by
open
solicitation
°
Laboratories
must
be:

 
Independent
 
Experienced
in:

°
In
vitro
test
methods
°
Cell
and
tissue
culturing
°
Test
chemical
administration
°
Enzyme
kinetics
and
inhibition
studies
 
Knowledge
of
steroidogenesis
 
Compliance
with
GLP
Measurements
of
Reliability
1.
Coefficient
of
variation
across
studies
 
Study
standard
deviation/
mean
of
studies
 
Reflects
the
spread
among
study
means
in
relation
to
their
average
value
2.
Ratio
of
between­
to
within­
study
standard
deviation
 
Standard
deviation
across
studies/
average
standard
error
within
studies
 
Reflects
relative
contribution
to
total
variation
of
the
variability
among
study
means
as
compared
to
the
precision
within
studies
3.
Comparison
of
within­
lab
SD
to
Average
within­
lab
SD
 
Standard
deviation
of
lab
I/
geometric
mean
within­
lab
SD
 
Measures
the
homogeneity
of
within
study
variation
across
laboratories
 
Can
identify
poor
performing
labs
Determination
of
Number
of
Laboratories
°
Sensitivity
analysis
for
each
measure
of
reliability
was
prepared
using
literature
values.
(
Fail,
Gray
Laskey)

°
For
Criterion
1
(
CV
interval
factor):

 
95%
confidence
interval
factors
were
calculated
as
a
function
of
the
number
of
laboratories
and
the
number
of
replicate
determinations
per
laboratory.

 
A
95%
confidence
interval
on
the
characteristic
of
interest
is
calculated
by
multiplying
the
point
estimate
by
the
confidence
interval
factor.

 
Confidence
interval
factor
is
sensitive
to
the
number
of
labs
and
approaches
1
as
the
number
of
labs
increases,
flattens
after
~
8
labs
 
Criterion
1
is
not
sensitive
to
the
number
of
replicate
determinations
per
lab
Determination
of
Number
of
Laboratories
(
2)

°
Criterion
2
(
Lower
and
Upper
confidence
interval
factor
of
between:
within
SD)

 
95%
confidence
interval
factors
were
calculated
as
a
function
of
the
number
of
laboratories
and
the
number
of
replicate
determinations
per
laboratory
 
Lower
confidence
interval
is
sensitive
to
both
number
of
labs
and
number
of
replicates
 
Upper
confidence
level
is
sensitive
only
to
the
number
of
labs
 
Flattens
out
after
~
8
labs
Determination
of
Number
of
Laboratories
(
3)

°
Criterion
3
(
Lower
confidence
factor
of
withinlaboratory
standard
deviations
to
average
withinlaboratory
standard
deviation)

 
95%
confidence
interval
factors
were
calculated
as
a
function
of
the
number
of
laboratories
and
the
number
of
replicate
determinations
per
laboratory
 
Sensitive
to
number
of
replicates
 
Not
sensitive
to
number
of
labs
 
Flattens
after
~
8
replicates
Determination
of
Number
of
Laboratories
(
4)

°
Conclusions
 
Based
on
available
data,
6­
10
laboratories
are
needed
to
achieve
a
high
confidence
indication
of
assay
reliability
 
We
shall
select
6
laboratories
as
the
actual
variability
in
these
studies
should
be
less
than
in
the
literature
where
different
protocols
were
used
 
~
8
replicates
are
needed
to
obtain
a
high
confidence
estimate
of
within
laboratory
standard
deviations.

This
information
will
be
generated
by
the
positive
and
negative
controls.
Validation
°
6
labs
°
Baseline
studies
 
3
runs
without
hCG
 
3
runs
with
hCG
challenge
 
Measure
testosterone
formation
and
LDH
 
No
test
chemical
°
Pilot
studies
 
Aminoglutethimide
(
positive
control)

 
Ethane
dimethanesulfonate
(
Leydig
cell
toxicant)
Validation
°
Coded
sample
studies
 
5
chemicals
 
2
replicates
per
laboratory
 
Aminogluthethimide
is
positive
control
 
Vinclozolin
(
AR
antagonist)
used
as
negative
chemical
°
Will
modify
validation
study
plan
based
on
results
of
prevalidation
work
Reference
Chemicals
by
Mode
of
Action
Fenarimol
Aromatase
Vinclozolin
Vinclozolin
Negative
MK­
434
5 
reductase
Genestein
Genestein
3 ­
HSD
Flutamide
P450c17
+
Cont
+
Cont
Ketoconazole
P450scc
Dimethoate
Dimethoate
StAR
inhibitor
Lindane
BPA
C­
AMP
inhibitor
Validation
Prevalidation
Mode
of
Action
Data
Analysis
°
Intra­
Laboratory
Analysis
Assess
chemically
related
testosterone
inhibition
within
laboratories
°
Inter­
Laboratory
Analysis
Assess
extent
of
heterogeneity
of
chemical
inhibition
effects
across
laboratories
Data
Analysis
Strategy
°
Large
numbers
of
comparisons
can
be
identified,
for
both
intra
and
inter
laboratory
analyses
 
Baseline
studies
°
Preval:
2
labs
x
3
replicates
=
6
°
Validation:
6
labs
x
2
replicates
=
12
 
Positive
Control
(
aminoglutethimide)

°
Preval
pilot:
2
labs
x
3
replicates
=
6
°
Validation
pilot:
6
labs
x
2
replicates
=
12
°
Preval
high
dose:
2
labs
x
9
chems
x
2
replicates
=
36
°
Validation
high
dose:
6
labs
x
5
chems
x
2
reps
=
60
Data
Analysis
Strategy
(
2)

°
Preval/
val
chemicals
2
labs
x
4
chems
x
2
replicates
=
16
6
labs
x
4
chems
x
2
replicates
=
48
°
Preval
chemicals:
2
labs
x
5
chems
x
2
replicates
=
20
°
Validation
chemicals:
6
labs
x
1
chem
x
2
replicates
=
12
Total
of
228
studies
°
In
validation
study
carry
out
each
analysis
considered
a
priori
to
be
possibly
toxicologically
relevant
°
Make
recommendations
in
final
report
concerning
which
analyses
are
most
informative
and
so
should
be
included
in
assay
standard
practice
Intra­
Laboratory
Analysis
°
Principal
Endpoints
 
Cumulative
Testosterone
Concentration
 
Cumulative
LDH
Concentration
(
3
doses
and
3­
4
time
points)

°
Similar
analyses
for
both
endpoints
°
Variation
in
Effects
 
Across
chemicals
 
Across
graded
chemical
doses
Components
of
Variation
°
Rat­
to­
rat
°
Testis­
to­
testis
within
rat
°
Fragment­
to­
fragment
within
testis
°
Block
of
assays
performed
simultaneously
(
e.
g.
day­

today

°
Variance
components
need
to
be
accounted
for
in
the
statistical
analysis
Inter­
Laboratory
Analysis
°
Focus
on
primary
and
secondary
responses
from
intra­
laboratory
analysis
°
Assess
extent
of
heterogeneity
of
responses
across
laboratories
°
Reference
 
American
Society
for
Testing
and
Materials
(
1988).
"
Standard
Practice
for
Conducting
an
Inter­
Laboratory
Study
to
Determine
the
Precision
of
a
Test
Method"
Measures
of
Variation
Among
Laboratories
°
Heterogeneity
of
within­
laboratory
means
across
laboratories
°
Heterogeneity
of
within­
laboratory
standard
deviations
across
laboratories
°
Ratio
of
laboratory­
to­
laboratory
standard
deviation
to
average
within­
laboratory
standard
deviation
°
Coefficient
of
variation
across
laboratories
Reporting
°
Each
laboratory
will
report:

 
That
protocol
was
followed
 
Difficulties
in
executing
the
studies
 
Summary
of
data
 
Raw
data
°
Validation
Study
Report
Graphical
Summary
Displays
°
Prepare
control
charts
and
associated
control
limits
°
Display
intra­
laboratory
statistics
side­
by­
side
across
 
Chemicals
 
Graded
doses
 
Laboratories
°
Identify
outlying
laboratories
and
nature
of
discrepancies
