Vickie S. Wilson

EDVMS Meeting

December 10-12-2003

Vickie S. Wilson

EDVMS Meeting

December 10-12-2003

Androgen Receptor Binding Assay Update

Androgen Receptor Binding Assay Update

Overview

• General introduction to binding assays• NICEATM/ICCVAM and Expert Panel• Summary of work completed

Training and Protocol Refinement Comparison of RPC and PV Scatchard analyses R1881 comparison 16 chemicals

• Future Direction

HRE

R RR DNA GENE A

mRNA

S

SHBG

mRNA

Blood 2

3

4

6

78

Cytoplasm

Nucleus

HSP HSP

tRNA

mRNA

Protein A

LEGRT '98

tf

r

Target Cell

1 S c

9

Gonad

S

S

S

S

SHBG

SS

SS S

R

S

3

10

5

aa

Luciferase

S

Two basic types of receptor binding Two basic types of receptor binding experiments experiments

• Saturation

Affinity of radioactive ligand for the receptor- Kd - Affinity of radioligand- Bmax - Binding sites

• Competition

Affinity of unlabeled ligand in competition with high affinity radioligand

- IC50, RBA- Ki – affinity of unlabeled ligand

Basic Steps in Receptor Binding AssaysBasic Steps in Receptor Binding Assays

Receptor (R) + [3H]Ligand (Free) Receptor:Ligand Complex (Bound)k1

k2

Incubate RR

3H

3H

R3H

3H

3H

3HR

3H

R

3H

R

Separate Boundfrom FreeMeasure

Radioactivity Bound

Analyze Results

Radiolabeled ligand

Test chemical

R

3H

Receptor

[ T ][ T ]11

[ T ][ T ]22

[ T ][ T ]77……

1E-11 1E-10 1E-9 1E-8 1E-7

0

20

40

60

80

100%

Bou

nd

Concentration, M

Competitive Binding Curve Competitive Binding Curve Quality DataQuality Data

1E-12 1E-11 1E-10 1E-9 1E-8 1E-7 1E-6 1E-5 1E-4 1E-3 0.01

0

20

40

60

80

100

% B

ound

Concentration, M

Standard Unknown 1

1E-11 1E-10 1E-9 1E-8 1E-7 1E-6 1E-5 1E-4 1E-3 0.01

0

20

40

60

80

100

% B

oun

d

Concentration, M

Standard Unknown 2

Example Binding Curves:Example Binding Curves:Examine data carefully for problemsExamine data carefully for problems

0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0

0.00

0.05

0.10

0.15

0.20

1/B

ound

1/Free R1881

0 M

50 M

100 M

200 M

Experimental Determination of Experimental Determination of Competitive Inhibition and KCompetitive Inhibition and Kii

0 50 100 150 2002

3

4

5

6

7

8

9

Slop

e X

100

[Competitor] (M)

Double reciprocal plotDouble reciprocal plot Slope replotSlope replot

EDC Expert Panel Report• Acknowledged the lack of a standardized in vitro AR

binding assay protocol

• Identified need for establishing comparative performance criteria

• Agreed on minimum procedural standards

• Acknowledged that RPC is “Gold Standard” for comparison purposes Most frequently used - Particularly useful as a

reference Has several disadvantages

• Recommended as high priority the development of an assay using purified, recombinant full-length AR

• Patent issues with hAR so an assay using an AR sequence from a species closely related to human may be necessary

Overview

• General introduction to binding assays• NICEATM/ICCVAM and Expert Panel• Summary of work completed

Training and Protocol Refinement Comparison of RPC and PV Scatchard analyses R1881 comparison 16 chemicals

• Future Direction

Comparison of RPC and PanVera Assays

2 ProtocolsRat Ventral Prostate Cytosol (RPC) - from EPA, RTDPanVera - from NCTR

Design: • 3 Technicians• Each tech ran every chemical in both protocols• 2 Duplicate tubes per run (3 runs in dup) • Positives were repeated by all 3 techs (6 runs)

19 Chemicals over a range of potenciesIdentified by number only

Test chemical concentrations as specified in each protocol

3.84.5

3.9

10.71 0.34

6.56.6

8.5

4.8

2.1 0.68

-11 -10 -9 -8 -7 -6 -11 -10 -9 -8 -7 -60

20

40

60

80

100

120

% r

ad

iola

bell

ed

1881 b

ou

nd

rAR STD rAR Mean PV STD PV Mean

Comparison of RPC and PV binding assays for R1881. The interassay CV for the PV assay is 13% versus 6% for the RPC

assay. Hence the PV assay is 2 fold more variable, which will require more replicates.

Binding Binding CurvesCurves

Example of Example of Variability Variability

between runsbetween runs

6 runs of same chemical in both

protocols

RPC

PV

PV Binding Assay for 3039 (DEHP)

Comparison of RPC and PV for p,p’-DDE

Examples IllustratingConcerns with PV Assay

9080

34

15 11

27

128122

10494

8578

4.3E-10 4.3E-9 4.3E-8 4.3E-7 4.3E-6 4.3E-50

50

100

150

% r

ad

iola

be

lle

d 1

88

1 b

ou

nd PV hAR STD

High Interassay VariabilityBinding Greater than 100%

Comparison of RPC and PV for E2

1E-11 1E-10 1E-9 1E-8 1E-7 1E-6 1E-5 0.0001 0.0010

20

40

60

80

100

120

140

% r

ad

iola

bell

ed

1881

bou

nd PV hAR VP rAR-2003

U-Shaped Curve

1E-10 1E-9 1E-8 1E-7 1E-6 1E-5 0.0001 0.001 0.010

10

20

30

40

50

60

70

80

90

100

110

% r

ad

iola

belled

1881 b

ou

nd PV hAR VP rAR-2003

Curve Shifted

Analysis of Assay Comparison

High intra-assay variability in PV • 3.5% of duplicates rejected. Discrepancy of greater than 25%

High inter-assay CV in PV assay• Twice the rejection rate of the RPC

Several PV assays with extraordinarily high CVs

Other Issues• Some U-Shaped binding curves in PV• Binding greater than 100% in some PV assays

Different concentrations of unknowns used in RPC and PV assays complicates comparison of assays

Scatchard Display289-L 10/7/2002

0 2500 5000 7500 10000 12500 150000

2500

5000

7500

10000

12500

Specific Binding

Bo

un

d/F

ree

WA 2-22 Saturation StudiesRun 289-L 10/7/02

0 1 2 3 4 5 6 7 8 9 10 110

2500

5000

7500

10000

12500

15000

Equation 1Best-fit values BMAX KDStd. Error BMAX KD95% Confidence Intervals BMAX KD

289L

139490.8800

376.30.04999

13142 to 147560.7728 to 0.9873

R1881 (nM)

Sp

ecifi

c B

ind

ing

(D

PM

)

Run Kd, nM Bmax, fmol/mg

288J 0.9418 65.29289L 0.880 64.75290J 0.9615 66.0291L 0.8710 64.59

Mean 0.914 +/- 0.04 65.16 +/- 0.64

Saturation Binding Acceptable

• Two technicians• Two Runs per technician• Duplicates per run• Runs on two different days

Reference Chemical (R1881) Comparison

2 Technicians each ran twice with duplicates – 4 reps (Subtask 3.2)

Repeated – 2 technicians; 6 runs each – 12 reps (Subtask 3.5) - Sixteen total replicates

Analysis was a nested ANOVA with a 5 x 2 x 8 x 2 design (5 concentrations of R1881; 2 techs; 8 replicates per tech; 2 duplicate observations per replicate)

Sigma Plots’s Ligand software

R1881 BindingAll runs converged and had R2 values greater than 99%

1.3

33

1.3

43

1.2

97

5 1.1

88

0.7

91

7

0.7

78

0.8

05

1

0.9

22

05

0.7

48

5

0.9

01

7

1.0

96

1.0

83

0.9

81

85

1.0

53

4

0.9

91

75

0.9

95

2

30

3

30

4

30

5

30

6

31

8

31

9

32

0

32

1

32

2

32

3

33

2

33

3

33

4

33

5

33

7

33

80.7

0.8

0.9

1

1.1

1.2

1.3

1.4

EC50s E-9

8.8

87

8.8

75

8.8

78

8.9

28

9.1

01

9.1

12

9.0

94

9.0

37

9.1

26 9.0

47 8

.96

8.9

68

9.0

16

8.9

79

9.0

05

9.0

05

303

304

305

306

318

319

320

321

322

323

332

333

334

335

337

3388.7

8.8

8.9

9

9.1

9.2

9.3

- LogEC50s

EC50 and log EC50 by RunEC50 and log EC50 by Run• Shows clustering of results over time• CV of reps (8) within batch = 4.6%• CV between batches = 22.5%• Note similarity of reps between 2 technicians

1.2

9

0.8

24

5

1.0

33

6

SOW 3.2SOW 3.5 A

SOW 3.5 B0.7

0.8

0.9

1

1.1

1.2

1.3

1.4

EC50s E-9

8.8

92

9.0

86

8.9

9

SOW 3.2 SOW 3.5 A SOW 3.5 B8

8.5

9

- LogEC50s

EC50 and log EC50 - Mean and SE “Batch” Clustering of Results Over Time

All Three groups differ significantly from each otherCV between batches = 22.5%

Summary and Conclusions R1881R1881 Comparison

• Binding assay with R1881 was run 16 times in three “batches” by 2 technicians

• CV for duplicates – about 5%

• Interassay CV – about 22%

• Each run provided an excellent fit - R-squared values greater than 99%

• In the worst case, the IC50 values varied by 2 fold (0.7 X10-9 to 1.3 X 10-9)

• Success

0.68

0.538

P 2 19 P2 220.5

0.55

0.6

0.65

0.7

0.75

4831 EC50 E-64-androstene-3,17-dione

1.28

1.047

P 2 19 P2 220.5

0.6

0.7

0.8

0.9

1

1.1

1.2

1.3

1.4

1.5

R1881 EC50 E-9

Task 3.3 AR Binding protocol comparison. Battelle concluded that there were only slight differences between the two protocols. However, we found that several of these were statistically significant. The "experiment" was unacceptable as designed, so such results should be ignored until the hypothesis is tested in a true side-by-side experiment.

p < 0.05

p < 0.012

2.5

2.21

P 2 19 P2 221.5

2

2.5

3

4833 EC50 E-95 alpha DHT

6.81

6.05

P 2 19 P2 225

5.5

6

6.5

7

7.5

4837 EC50 E-6Corticosterone

p < 0.053p < 0.2

Results of 16 ChemicalsResults of 16 Chemicals

• Original Report from Battelle classified • 14 Chemicals as Binders• 2 Chemicals as Non-Binders

• EPA Review reclassification• 10 Binders• 4 Equivocal • 2 Non-binders

• Equivocal binders - need additional experiments to define Ki

• Chemicals were each run 2-3 times but better experimental design needed before detailed statistical analysis

4-tert- OctylphenolMethoxychlorVinclozolinProcymidone

LinuronCyproterone Acetate17-EstradiolP,p’-DDEMedroxyprogesterone

AcetateMethyltrienoloneTestosteroneProgesteroneDexamethasoneSpironolactone

BINDERS

AtrazineDi(2-ethylhexyl)phthalate

(DEHP)

EQUIVOCAL

NON-BINDERS

Recombinant Androgen Receptor

Expert Panel recommended as high priority the development of an assay using purified, recombinant full-length AR

- Patent issues with human AR - Species closely related to human

Questions with truncated (chimeric) AR

Ongoing work at EPA, RTD - Chimpanzee cDNA library obtained- Screening for full length AR

Future Direction

• Supplement binding data of 16 chemicals with additional runs and conduct statistical analysis (intralaboratory)

• Work on recombinant system is being conducted but lags behind

• desirable but 2-3 years for development and standardization• no commercial or non-commercial source available

• Move forward with RPC assay• standard data set• comparative performance criteria• interlaboratory study