Upload
others
View
4
Download
0
Embed Size (px)
Citation preview
Intrinsic and Extrinsic Controls for Formalin Fixed, Paraffin Embedded Tissue
David L. Rimm M.D., Ph.DDirector, Yale Pathology Tissue Services
Professor, Dept. of PathologyYale University School of Medicine
Disclosure/Disclaimer• I am a consultant to, stockholder in, and scientific co-founder of
HistoRx Inc. the exclusive licensee of the AQUA® technology
• I am an author on the Yale held patent on the AQUA technology and receive royalties.
• This project has been funded in whole or in part with the federal funds from the National Cancer Institute, National Institutes of Health, under Contract No. HHSN261200800001E. The content of this publication does not necessarily reflect the views or policies of the Department of Health and Human Services, nor does mention of trade names, commercial products, or organizations imply endorsement by the U.S. Government.
The central problem: Standardization of Protein
Assessment of Formalin-Fixed Paraffin-Embedded tissue
• Definition of Extrinsic vs Intrinsic Control• Solution of the Extrinsic control problem• Progress toward an Intrinsic control or
Tissue Quality Index (previously designated TIC for tissue immuno-competence index)
Extrinsic vs Intrinsic Controls
• Extrinsic controls; control for and standardize all the processes from the stainer through the analysis
• Intrinsic controls; control for and standardize all the processes from the patient to the stainer (pre-analytic varibles)
Our solution to the Extrinsic Control Standardization Problem: The AQUA method of
Quantitative ImmunofluorescenceAQUA: Think like a moleculeSelection of regions only as a function of colocalization of molecular interactions
Other Software: Think like a human Assign significance to morphologically defined entities and use feature extraction to emulate human assignments
Example: a nuclear protein is measured by colocalization with DAPI in a cytokeratin positive region
Example: a nuclear protein emulates the human definition of nucleus and finds round or roundish entities, then counts signal within the roundish entities
http://www.tissuestudio.com/
Understanding the Difference between AQUA and other tissue analysis software
AQUA: Think like a moleculeSelection of regions only as a function of colocalization of molecular interactions
Other Software: Think like a human Assign significance to morphologically defined entities and use feature extraction to emulate human assignments
Solution: No pathologist to “agree” with since result is strictly derived from co-localization
Problem: Feature extraction software does not “agree” with the pathologist since tumors (and pathologists) are very different
http://www.tissuestudio.com/
0
2
4
6
8
10
12
CookieM Animal Beaker CookieM CookieM CookieM CookieM
CW JM JM JM MC MC MC
Day 1 Day 1 Day 1 Day 1 Day 1 Day 2 Day 3
AQ
UA
(R)
Sco
re (
log2
)
PM2PM1 PM1PM3 PM1
PM1 PM1
CV=4.3%
TMA
WTS
TMA-Tissue MicroarrayWTS-Whole Tissue Section
Cytokeratin Tumor Mask
Combine DAPI image and cytokeratin image then cluster to assign each pixel to a subcellular compartment
Estrogen Receptor
Σ compartmentpixel area
Σ target intensityin compartment pixels
= AQUAscore
Generating the AQUA®
score
40 patient controls(range of ER)
formalin-fixing ¶ffin-embedding
lysates
Western Blot alongside
recombinant ER
calculate ER as AQUA scores
Convert AQUA scores to pg/ug or other units
Control TMA
cell line panel(known range of ER)
calculate ER in pg/µg
Alley Welsh
Standardized Index Array ER antibody used is 1D5
Lowest positive vs. highest negative
ER
ER Alley Welsh
Expanded “levels” to visualize thresholdcontracted dynamic range of grayscale (max RGB input level 25516)
ER ER
ER ER
HC
C22
79H
2882
H19
3H
441
H12
99H
1666
H13
55S
KB
R3
MC
F7 10 n
g25
ng
50 n
g12
5 ng
EGFR
Rec EGFR
Beta tubulin MCF7
H1299 H1355
H193
H2279
H1666H2282H441
EG
FR le
vels
(ng/µg
)
Aqua score
SKBR3
0.85 ng/µg EG
FR le
vels
(ng/µg
)
EGFR negative EGFR positive
0.85 ng/µg
A.B.
C. D.
50 KD
91 KD
170 KD
EG
FR le
vels
(ng/µg
)
R=0.782
25
20
15
10
5
0
HC
C2279
H193
H1666
H2882
H441
H1299
H1355
SK
BR
3
MC
F7
12
10
8
6
4
2
0
109876543210
20 40 60 80 100 120 140 160
Tassos Dimou
Development and Commercialization Of a Quantitative Protein Measurement Technology
(AQUA) from the lab to the patient
Extrinsic vs Intrinsic Controls
• Extrinsic controls control for and standardize all the processes from the stainer through the analysis
• Intrinsic controls control for and standardize all the processes from the patient to the stainer (Pre-analytic variables)
Goals of our Project• Development of a Tissue Quality Index
(TQI): • obtained by developing a quantitative
intrinsic control that can measure the degree of degradation of any FFPE sample.
Pre-analytical Variables (incomplete list)
• Variable warm ischemic time• Variable cold ischemic time• Variable manipulation during gross cutting and prepping• Variable temperature during fixation• Variable total fixation time• Variable thickness of tissue blocks• Variable half life of fixative• Variable types/brands/ components of fixatives• Variable types of tissue processors• Variable solutions in the processor• Variable temperatures of different processor components• Variable types of embedding paraffin• Variable slide drying times• Variable slide oven temperatures
Pre-Analytic Variables; Can we treat them as a black box?
If we cannot control pre-analytical variables can we
quantify the damage or tissue degradation caused by them?
Can we disqualify specimens for companion dx testing?
Goal 1: To generate two “discovery” tissue sets to assess “pre-analytical” variability.Goal 2: Assessment of markers of cold ischemia (“housekeeping markers”) on discovery cohortsGoal 3: Assessment of markers of hypoxia on discovery cohortsGoal 4: Generation of a Multiplexed “Tissue Immunologic Competence” (TIC) Model (now Tissue Quality Index (TQI)) for normalization of tissue handling that measures tissue integrity for immunological assessmentGoal 5: Validation testing of the TIC Model in two core vs. resection specimen studies
Intrinsic Controls for FFPE tissue
ApproachGenerate
Intrinsic Control cohorts
Select and validate potential
antibodies/reagents
Test each reagent individually on
Intrinsic Control Cohorts
Generate simplest Multi-variable model
that can assess tissue quality (TQI)
Validate TQI
Perform Western Blotting on cell line positive and negative controls
Is there a band at the expected molecular weight?
Titer antibody on TMA containing control tissues and cell lines
Is staining localized and specific consistent with protein’s biological description?
Does expression level by IHC correlate with
WB?
Is the antibody reproducible between runs?
No Aband
on Antibo
dyYes
Yes
Yes
Yes
No
No
Yes
No
Antibody Validated
Histone 4 - Nuclear R =-0.398p = <0.0001
N = 77
02000400060008000
100001200014000160001800020000
0 100 200 300 400 500Time to Fixation (Mins)
His
tone
4 A
vera
ge N
ucle
ar A
qua
Scor
es
0
60
120
180
240
300
360
420
480
TN85
TN74
TN28
TN35
TN65
TN66
TN84
TN32
TN11
TN55
TN15
TN95
TN98
TN61
TN41
TN81
TN93
TN80
TN73
TN37
TN106
TN52
TN101
TN71
TN88
TN30
TN16
TN43
TN86
TN100
TN50
TN48
TN29
TN25
TN102
TN36
TN103
TN77
TN17
TN1
TN9
TN5
TN21
TN13
TN10
TN23
TN2
Tim
e to
Fix
atio
n (m
inut
es)
Construction of the Rochester Tissue Microarray (2x redundancy)
Two fold redundancyN=125 , tumor=93, normal=2, cell lines=10 control breast tumor=10 ,control lung tumor = 10Collected by Dr. David Hicks and colleague, University of Rochester Medical Center
Bx-
TRPa
ir 1
Bx
Bx-
TRPa
ir 1
TR
Tissue sample
Area of concern
Field of View (FOV)
Tumor
Core Bx – Resection Pair Cohorts
Example of CNB – Resection Cohort (from other studies)
pTyr1248HER2
1-N 2-N 3-N 4-N14-N
5-2+10
-2+13
-2+15
-2+16
-2+17
-2+18
-2+19
-2+20
-2+21
-2+22
-2+23
-2+24
-2+25
-2+9-3+11
-3+27
-3+28
-3+29
-3+30
-3+31
-3+32
-3+33
-3+34
-3+35
-3+36
-3+37
-3+38
-3+39
-3+40
-3+41
-3+0
1000
2000
3000
4000
5000
6000
CNBTRP<0.0001
Pairs
AQ
UA
Antibody Validation (Overview)
Perform Western Blotting on cell line positive and negative controls
Is there a band at the expected molecular weight?
Titer antibody on TMA containing control tissues and cell lines
Is staining localized and specific consistent with protein’s biological
description?
Does expression level by IHC correlate with WB?
Is the antibody reproducible between runs?
No Abandon Antibody
Yes
Yes
Yes
Yes
No
No
Yes
No
Antibody Validated
These markers show either an increase or a decrease of expression with time to fixation on the time to fixation TMA
GAPDH - Tumor Mask
R=0.076p = 0.48N = 87
0
5000
10000
15000
20000
25000
0 100 200 300 400 500
Time to Fixation (Mins)
GA
PDH
ave
rage
Tum
or M
ask
Aqu
a sc
ores
Histone 3 - Nuclear
R = 0.137p = 0.22N = 80
0
5000
10000
15000
20000
0 100 200 300 400 500
Time to Fixation (Mins)
Hist
one
3 av
erag
e Nu
clea
r
Aqua
Sco
res
Beta Tubulin - Cytoplasm
R = -0.029p = 0.79 N = 85
0
5000
10000
15000
20000
0 100 200 300 400 500
Time to Fixation (Mins)
Beta
Tub
ulin
Ave
rage
Cy
topl
asm
Aqu
a sc
ores
Cyclin D1-Nuclear
0
4000
8000
12000
16000
0 100 200 300 400 500Time to Fixation
Cyc
lin D
1 av
erag
e nu
clea
r A
QU
A s
core
R=-0.05p=0.66N=78
Correlation between markers of cold ischemia and hypoxia with time to fixation
Histone 4 - Nuclear R =-0.398p = <0.0001
N = 77
02000400060008000
100001200014000160001800020000
0 100 200 300 400 500Time to Fixation (Mins)
His
tone
4 A
vera
ge N
ucle
ar A
qua
Scor
es
YTMA 173-1-22 phospho-p44/42 MAPK Time to Fixation (Mins) one fold
R2 = 0.0233
0
2000
4000
6000
8000
10000
12000
14000
0 100 200 300 400 500Time to Fixation (Mins)
Aqu
a sc
ore
of M
APK
und
er T
M
(one
fold
)
While some show a downward or upward trend, heterogeneity is a concern
AKAP and Time to fixation (Average TM scores)
R2 = 0.0572
0
2000
4000
6000
8000
10000
12000
14000
0 50 100 150 200 250 300 350 400 450Time to Fixation (Mins)
AK
AP
av
era
ge
aq
ua
sc
ore
s
(TM
)
0
2000
4000
6000
8000
10000
12000
14000
16000
18000
20000
0 50 100 150 200 250 300 350 400 450
Time to Fixation (Mins)
His
tone
4 A
vera
ge N
ucle
ar A
qua
Scor
es
Time to Fixation<1hr Aqua - 8155
Time to Fixation 1-2hrAqua - 6671
Time to Fixation> 2hr Aqua – 5102.62
Spearman Rank Correlation:
P=<0.0001
R=-0.398
N=77
Histone 4 on Time to Fixation Array
Biopsy Histone 4 nuclear
20/25
10/18
21/20
12/9
11/9
5/21
5/19
18/18
7/38
14/29
10/24 13/24
9/10
4/27
10/7
13/28
5/24
17/23
21/21
16/14
0
2000
4000
6000
8000
10000
12000
14000
sample1
sample2
sample3
sample4
sample5
sample6
sample7
sample8
sample9
sample1
0
sample1
1
sample1
2
sample1
3
sample1
4
sample1
5
sample1
6
sample1
7
sample1
8
sample1
9
sample2
0
sample2
1
sample2
2
sample2
3
sample2
4
sample2
5
BX Average H4 NucRX Average H4 Nuc
Paired t-test: p=0.88
Resection Histone 4 nuclear
Histone 4 on 25 matched pairs of biopsies and resections
Heterogeneity washes out the effect of time to fixation
phospho-MAPK test to test reproducibility
R2 = 0.9645
0
5000
10000
15000
20000
0 2000 4000 6000 8000 10000 12000 14000YTMA 179-6-116 Aqua score of MAPK under TM
YT
MA
17
9-6
-6 A
qu
a
sc
ore
of
MA
PK
un
de
r T
M
YTMA 165- Tic Reproducibility
R2 = 0.9063
02000400060008000
10000120001400016000
0 5000 10000 15000 20000 25000YTMA 165-1-25 Aqua score of MAPK under TM
YT
MA
165
-1-3
7 A
qu
a sc
ore
of
MA
PK
un
der
TM
AntibodyValidated
Assay Reproducibility
YTMA 173-1-22 phospho-p44/42 MAPK Time to Fixation (Mins) one fold
R2 = 0.0233
0
2000
4000
6000
8000
10000
12000
14000
0 100 200 300 400 500Time to Fixation (Mins)
Aqu
a sc
ore
of M
APK
und
er T
M
(one
fold
) P = .3983
The Spearman Rank Correlation shows a trend (p-value not significant) towards negative correlation.
Biopsy patient 5 pMAPK Immunohistochemistry
Resection patient 5 pMAPK Immunohistochemistry
Average AQUQ Score 5500
Average AQUQ Score 407
0
2000
4000
6000
8000
10000
12000
14000
16000
18000
20000
sample2
sample3
sample4
sample5
sample6
sample7
sample9
sample1
0
sample1
1
sample1
2
sample1
6
sample1
7
sample1
9
sample2
0
sample2
1
sample2
2
sample2
3
sample2
4
biopsy pMAPKaverage resection pMAPKaverage
Assessment of pMAPK expression on 25 matched pairs
of biopsies and resections
14/1339/11
6/14
13/31
8/15
10/28
18/12
6/79/42
18/35
14/36
12/13
23/27
21/52
6/21
44/45
30/4331/30
Samples missing: 1,8,13,14,15,18,25 Results for 18 pairs
Paired t-test: p=0.0011
Mixed Effects Model for Estimating Number of Fields Required for Immunostaining
Juliana Tolles, Yalai Bai and Annette Molinaro
Estrogen Receptor: Estimated Prediction Error Criterion
Model = linear regressionInput=average of sampled fields (simulated from a normal distribution for each subject)“Truth”=average over all fields
Juliana Tolles, Yalai Bai and Annette Molinaro
The number of FOVs required are a function of the protein examined
Number of 20X fields of view (FOVs) to find stable minimum in mixed effects modeling
Juliana Tolles, Yalai Bai and Annette Molinaro
Measurement of Variability and Heterogeneity of Estrogen Receptor
Between PatientsMean difference: 149 pts (CI: 92-241)
Between Biopsy and ResectionMean decrease: 134 pts (CI: 62-204)
Between Regions within a SampleMean difference: 39 pts (CI: 29-53)
Juliana Tolles and Annette Molinaro
Problem!! Need to find markers that are both highly homogeneous and highly
sensitive to pre-analytic variables
• Go beyond antibodies? Eosin• Phospho-modification? pTyr antibodies• Your suggestions here_______________
Intra Array Reproducibility of Eosin on YTMA173-2-8, march2011
R2 = 0.3405
0
500
1000
1500
2000
2500
3000
3500
0 500 1000 1500 2000 2500 3000Eosin in TM, norm, first fold
Eosi
n in
TM
, nor
m, s
ec fo
ld
Average Eosin Expression in TM on YTMA173-2-8, March2011
R2 = 0.0035
0
500
1000
1500
2000
2500
3000
3500
4000
4500
5000
0 50 100 150 200 250 300 350 400 450time to fixation in min
norm
AQ
UA
Sco
re o
f Eos
in in
TM
Distribution of norm AQUA Scores in TM for Eosin at 1 to 50, 5 min at RT
0
500
1000
1500
2000
2500
3000
3500
4000
4500
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28
Staining with Eosin at 1 to 50
-to measure the amount of free amines and protein degradation
-Predict increase free amine with increased degradation
Reproducibility for pTyrosine on breast tests
R2 = 0.6294
0
2000
4000
6000
8000
10000
12000
0 2000 4000 6000 8000 10000 12000 14000 16000 18000
first fold, pTyrosine in TM, norm AQUA Scores
sec
fold
, pTy
rosi
ne in
TM
, nor
m A
QU
A
Scor
es
AntibodyValidated
Treatment of breast test with Lambda Phosphatase for 2 hours at 37degree celsius
0
2000
4000
6000
8000
10000
12000
1 2 3 4 5 6 7 8 9 10
norm
AQ
UA
Sco
re in
TM
of p
Tyro
sine
pTyrsone1:250 phosphatase treated slide, incubated with pTyrosine 1 to 250
Sample #
Average pTyrosine Exression on YTMA173-2 and Time to Fixation
R2 = 0.0762
0
2000
4000
6000
8000
10000
12000
14000
16000
0 50 100 150 200 250 300 350 400 450time to fixation
Ave
rage
pTy
rosi
ne in
TM
, nor
m A
QU
A S
core
P=0.0079
Intra Array Reproducibility for pTyrosine 4G10 on YTMA 173-2 at 1 to 250, March 2011
y = 0.5745x + 1072.6R2 = 0.5407
0
2000
4000
6000
8000
10000
12000
0 2000 4000 6000 8000 10000 12000 14000 16000first fold pTyrosine, norm TM AQUA Scores
sec
fold
pTy
rosi
ne, n
orm
TM
AQ
UA
Sco
res
SummaryGenerate
Intrinsic Control cohorts
Select and validate potential
antibodies/reagents
Test each reagent individually on
Intrinsic Control Cohorts
Generate simplest Multi-variable model
that can assess tissue quality (TQI)
Validate TQI
Perform Western Blotting on cell line positive and negative controls
Is there a band at the expected molecular weight?
Titer antibody on TMA containing control tissues and cell lines
Is staining localized and specific consistent with protein’s biological description?
Does expression level by IHC correlate with
WB?
Is the antibody reproducible between runs?
No Aband
on Antibo
dyYes
Yes
Yes
Yes
No
No
Yes
No
Antibody Validated
Histone 4 - Nuclear R =-0.398p = <0.0001
N = 77
02000400060008000
100001200014000160001800020000
0 100 200 300 400 500Time to Fixation (Mins)
His
tone
4 A
vera
ge N
ucle
ar A
qua
Scor
es
Thanks to:Rimm Group:Valsamo (Elsa) AnagnostouBonnie Gould RothbergVeronique NeumeisterSeema AgarwalAnastasios DimouHuan ChengMaria BaqueroAlley WelshJason HannaJennifer BordeauxHalley WimberlySummar SiddiquiElisabeth RichardsonHollis VirayYalai BaiRobert Camp
Yale Pathology Tissue ServicesLori CharetteJoe Salame Aruna MadanSudha Kumar Peter Gershkovich
Outside Yale CollaboratorsKonstantinos Syrigos (Athens)Gerold Bepler (Moffitt-KCI)Daniel Hayes and SWOGElaine Alarid (UW)Bruce Haffty (CINJ)
Yale CollaboratorsAnnette Molinaro
Karen LostritoJuliana Tolles
Harriet KlugerRuth HalabanSteve AriyanDaniel Boffa
Catherine SullivanFrank DetterbeckLynn TanoueLyndsay Harris
Work supported by grants from the NCI, DOD, the Susan G Komen Foundation for the Cure and the NCI Office of Biospecimen and Biorepository Research (OBBR)
www.tissuearray.orgRimm Lab 2010
Same Genome – Different Proteome
Why we measure protein
Why we measure protein in situ
Auguste Renoir : The Luncheon of the Boating Party C.1881
Claude Monet: The Stroll, Camille Monet and Her Son Jean (Woman with a Parasol)C. 1875
Why we measure protein with a machine
47
Marker of cold ischemia: Histone H4
Western Blotting Expression Range Graph
Quantitative Immunofluorescence
0
2000
4000
6000
8000
10000
12000
14000
16000
1/50 1/100 1/500 1/1000 1/5000
bottom three Aqua scorestop three Aqua scoressubtraction
Dilution 1:50 AQUA=14307
Dilution 1:100 AQUA=8391
Dilution 1:500 AQUA=7962
Dilution 1:1000 AQUA=7450
Dilution 1:5000 AQUA=576
Tubulin
H4
H441
H135
5
HCC2
279
A431
SKBR
3BT
474
MCF
-7
H166
6H1
299
MB4
68BT
20
AntibodyValidated
Histone 4 Reproducibility (Tic Array)All spots
R = 0.763
0
500
1000
1500
2000
2500
0 2000 4000 6000 8000 10000 12000
H4 Nuclear Scores YTMA 165_1_ March
H4
Nuc
lear
Aqu
a sc
ores
YTM
A 1
65_1
_ Fe
b
Assay Reproducibility
0
2000
4000
6000
8000
10000
12000
14000
16000
TN85 TN47 TN67 TN89 TN63 TN46 TN17 TN8
25min 41min 45min 49min 68min 75min 95min 135min
average pMAPK on biopsies average pMAPK on TMAspots
Paired t-test: p=0.073
Assessment of pMAPK Expression on Biopsies and
matched TMA Spots on Time to Fixation Array
13
2927
15
34
30
34
20
pTyrosine Expression Range Graph
0
2000
4000
6000
8000
10000
12000
14000
16000
18000
1:25 1:100 1:250 1:500 1:1000 1:5000
Aqu
a S
core
of p
Tyro
sine
und
er T
M topbottomsubtraction
4G10 Platinum, Anti-Phosphotyrosine
Mouse monoclonal Antibody cocktail IgG2b
Millipore, Cat. # 05-1050
0
2000
4000
6000
8000
10000
12000
14000
16000
18000
1:25 1:100 1:250 1:500 1:1000 1:50000
1
2
3
4
5
6
7topbottomratio
pTyrosine Expression Range Graph