David L. Rimm M.D., Ph.D David L. Rimm M.D., Ph.D Professor, Dept. of PathologyProfessor, Dept. of Pathology

Yale University School of MedicineYale University School of Medicine

Intrinsic and Extrinsic controls for FFPE tissue

DisclosureDisclosure

•• I am a consultant, stockholder and I am a consultant, stockholder and scientific founder of HistoRxscientific founder of HistoRx

•• I am an author on the Yale held patent I am an author on the Yale held patent on the AQUA technology.on the AQUA technology.

Use of analyte measurement to Use of analyte measurement to determine patient managementdetermine patient management

Clinician suspects possible diabetes

Obtain tissue sample (blood)

Measure Blood glucose levels in mg/dl (objective)

Treat with appropriate therapy

Clinician suspects possible breast cancer

Obtain tissue sample (core biopsy)

Make Histologic Dx then measure estrogen receptor levels (subjective judgment)

Treat with appropriate therapy

AQUAAQUA®® method of analyte (estrogen receptor) method of analyte (estrogen receptor) measurement on a tissue slidemeasurement on a tissue slide

Step 1: Mask (define region of interest, exclude stroma, blank space, etc)

Step 2: Define the numerator and denominator

Concentration =Numerator

Denominator Σ compartmentpixel area

Σ target intensityin compartment pixels

Step 3: Calculate the AQUA score

Step 4: Convert to absolute concentration or normalize to set of uniform standards

= AQUAscore

TMA

WTS

TMA-Tissue MicroarrayWTS-Whole Tissue Section

CytokeratiinTumor 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

Quantification of western

MB

231

H 1

666

MB

436

CH

OSK

BR

3M

B 4

35

HT

29SU

M 1

59B

T 20

ZR 7

51B

T 47

4T4

7DM

CF7

Puro

0Pu

ro .0

1Pu

ro .1

Puro

.5Pu

ro 1

Puro

5

ERα

β-tubulin

pg ER / μg total protein

0

100

200

300

400

500

600

MB 231

H1666

MB 436

CHOSKBR3MB 43

5HT 29

SUM 159

BT 20ZR 75

1BT 474

T47D

MCF 7puro

0pu

ro .01

puro

.1pu

ro .5

puro

1puro

5

ER in Cell Lines by AQUA (no Envision)

0

100

200

300

400

500

600

MB 231

BT 20

HT 29SUM 15

9H16

66MB 43

6

CHOSKBR3BT 474

T47DPuro

.1MCF-7

(Puro

0)Puro

.5Puro

1Puro

5

AQ

UA

sco

re (n

o En

visi

on)

ER in Cell Lines by Western

0

100

200

300

400

500

600

SUM 159

CHOMB 23

1

H1666

MB 436

SKBR3

HT 29

BT 20BT 47

4

T47D

MCF-7 (P

uro 0)

Puro .1

Puro .5

Puro 1

Puro 5

pg E

R / ◊g

tota

l pro

tein

Converting ER AQUA scores to a concentration (pg/ g)

y = 1.101x - 26.513R2 = 0.865

0

100

200

300

400

500

600

700

0 100 200 300 400 500 600

AQUA score (no Envision)

pg E

R /

g to

tal p

rote

in

Use this equation to assign pg/μg value based on the AQUA scores for ER “negative” cell lines

Regression without ER “negatives”

Cell Line ER expression in pg/Πg

0

100

200

300

400

500

600

MB 231

BT 20

HT 29SUM 159

H1666

MB 436

CHO

BT 474

T47DMCF-7 (P

uro 0)

Puro .1

Puro .5

Puro 1

Puro 5

pg E

R / Πg

tota

l pro

tein

Use these cell lines (present on TMAs) to convert patient ER AQUA scores from YTMA 49 to an ER concentration (pg/ug)

Limit of detection

is 50 pg/μg

AQUA® analysis: Normalization/Standardization Example: HER2

CV=4.3% Cumulative

Operator 1

Operator 2

Stain 1

Mach.1

Stain 1

Mach. 3

Mach. 2

Mach. 1

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) S

core

(lo

g2)

Operator 3

Mach. 1

Stain 2 Stain 3Stain 1Stain 1 Stain 1

Mark Gustavson and Jason Christiansen

The HistoRx AQUAThe HistoRx AQUA®® platformplatformComplete solution Complete solution -- Hardware, Software, Hardware, Software,

ReagentsReagents

Automated Fluorescence Microscopy

– Expanded dynamic range of measurement

– Multi-parametricCommercially available with about 16 current placements worldwideAQUA analysis software compatible with .tiff imagesIn use by more than a dozen Pharma companies for drug developmentFDA clearance pending on AQUA software using ER as the example analyteUS Patent 7,219,016

Summary:Summary:• Absolute values of Estrogen Receptor can be obtained by conversion from Western Blot to Cell Line Standards to Breast Cancer Tissue

• The minimal expression level detectable by this system is about 50pg/ug total protein

• Range of ER expression in patients is between 50pg/ug and about 1500 pg/ug

• Reproducibility for AQUA has shown coefficient of variation (CV) of around 5%

• Someday: CLIA guidelines for Estrogen Receptor: Threshold for detection = 100pg/ug total proteinAcceptable variation = +/- 10% (20pg/ug)

StandardizationStandardization

•• Extrinsic controlsExtrinsic controls–– Controls for the process of analyte detection Controls for the process of analyte detection

external to the FFPE sample (ie the antigen external to the FFPE sample (ie the antigen retrieval, staining, visualization and quantification)retrieval, staining, visualization and quantification)

•• Intrinsic controlsIntrinsic controls–– Controls for specimen integrity based on Controls for specimen integrity based on

properties of the tissue that have been subjected properties of the tissue that have been subjected to all of the preto all of the pre--analytical processes in parallel analytical processes in parallel with the analyte (ie controls for underwith the analyte (ie controls for under--fixation, cold fixation, cold ischemic time, etc.)ischemic time, etc.)

Key PreKey Pre--analytic Variables for analytic Variables for measuring analytes on slide in formalinmeasuring analytes on slide in formalin--

fixed paraffin embedded tissuefixed paraffin embedded tissue

1.1. Time to Fixation (cold ischemic time)Time to Fixation (cold ischemic time)2.2. Method of Tissue AcquisitionMethod of Tissue Acquisition3.3. Fixative and Method of FixationFixative and Method of Fixation

AQUA score of pAKT in Tumor Mask

0

50

100

150

200

250

300

350

400

450

1 2 3 4 5 6 7 8 9 10

Biopsy vs. Tumor Resection

who

le ti

ssue

sec

tion

AQUA

sco

re

Biopsy tissuefollow up tumor resection

8 1211 20 5 26

25 2310 26

22 19 17 29

11 11 6 6

9 10

Number of fields examined on each tumor

Whole section analysis of core biopsy vs. resection to assess affect of ischemic time on stability

Yalai Bai and Eirini Pectacides

pAKT Expression

050

100150200250300350

S17-S15

S10-S9

S22-S21

S14-S13

S8-S18

S7-S23

S5-S16

S20-S4

S6-S19

S28-S

29Pairs

AQ

UA

Sco

re

BxTR

10 26 22 19 25 23 9 1017 29 11 11

11 20 8 22

5 26 7 12

Lower Expression of pAKT in the resections than in CNBs

Wilcoxon Signed Ranks testp=0.004

AKT Expression

0

50

100

150

200

250

S17-S15

S10-S9

S22-S21

S14-S13

S8-S18

S7-S23

S5-S16

S20-S4

S6-S19

S28-S29

Pairs

AQ

UA

Sco

re

15 48 14 44

17 12

13 8 20 30

5 16 14 49 7 15 7 18 9 27

Wilcoxon Signed Ranks testp=0.065

Whole section analysis of core biopsy vs. resection to assess affect of ischemic time on stability for Total AKT

pERK Expression

0

200

400

600

800

1000

1200

S17-S

15S10

-S9

S22-S

21S14

-S13

S8-S18

S12-S

11S7-S

23S5-S

16S20

-S4

S28-S

29S30

-S31

Pairs

AQ

UA

Sco

re

9 28 15 40

15 9 8 6 12 20 7 6 6 30 9 6 6 2523 28

7 12

Lower Expression of pERK in the resections than in CNBs

Wilcoxon Signed Ranks testp=0.005

CNBsTumor Resection

ERK Expression

0

500

1000

1500

2000

2500

3000

3500

S17-S

15

S10-S

9

S22-S

21

S14-S

13

S8-S18

S7-S23

S20-S

4

S24-S

25

S26-S

27

S30-S

31

Pairs

AQ

UA

Sco

re

8 20

13 24 8 199 6

10 159 11 7 8

12 177 13

18 21

Wilcoxon Signed Ranks testp=0.38

Wilcoxon signed rank testP value 0.0098

Significant Difference in ER Levels Significant Difference in ER Levels between Core Bx and Resectionbetween Core Bx and Resection

Yalai Bai

Cytokeratin AQUA in Tumor Compartment

S17-S

15S10

-S9

S22-S

21S14

-S13

S8-S18

S12-S

11S7-S

23S5-S

16S20

-S4

S6-S19

0

1000

2000

3000

4000

BiopsyTumor Resections

30 23

16 8

13 10

17 47

7 7

12 2013 47

7 5

11 18

21 53

Biopsy vs. Tumor Resections

AQ

UA

Sco

re o

f Who

le T

issu

e Se

ctio

ns

ER AQUA in Nuclear Compartment

S17-S

15S10

-S9

S22-S

21S14

-S13

S8-S18

S12-S

11S7-S

23S5-S

16S20

-S4

S6-S19

0100200

300400500600

700800900

5

46

18832

46 21

7

20

7

17

6

16

21

33

19

12

711

Biopsy vs. Tumor ResectionsA

QU

A S

core

of W

hole

Tis

sue

Sect

ion

53

ER Expression normalized by Cytokeratin (CK) of ER EXP

0

0.1

0.2

0.3

0.4

0.5

0.6

S17 S15

S10 S9

S22 S21

S14 S13

S8 S18

S12 S11

S7 S23

S5 S16

S20 S4

S6 S19

ER A

QU

A/ C

K A

QU

A

Biopsy

Tumor Resection

Normalization by CK expression results in loss Normalization by CK expression results in loss of statistically significant decrease in ERof statistically significant decrease in ER

CoPath Database (Yale Pathology) ER AQUA

+ +

- +

- + weak

+ -

- -

YTMA 94-2 ER AQUA (tissue)

0

50

100

150

200

250

300

350

400

450

500

s96-8

961

1s96

-1866

91s

96-10

746

1s96

-2091

2s9

6-334

42s

96-21

954

1s96

-6413

s96-3

158

s1s9

6-172

39s9

6-833

5s9

6-125

05s9

6-339

Normal

s96-4

629

Normal

Normal

96-33

02s9

6-158

941s

96-22

076

2s96

-1643

s96-7

903

s96-1

0541

1s96

-6624

S96-61

78s9

6-119

711s

96-13

586

s96-7

594

2s96

-8655

1s96

-9882

s96-7

229

s96-4

818

s96-1

7238

1s96

-2188

51s

96-15

415

s1s9

6-126

72s9

6-210

021s

96-40

74

Specimen Number

AQ

UA

Sco

re o

f ER

Thanks to:Thanks to:Rimm Group:Robert CampElsa AnagnostouBonnie Gould RothbergVeronique NeumeisterSeema AgarwalEirini PectasidesHuan ChengMaria BaqueroAlley WelshJason HannaJennifer BordeauxMatt McRaeBill BradleyCamil LiceagaSummar SiddiquiLiz Killiam

Tissue Microarray FacilityLori CharetteJoe SalameSudha KumarPeter Gershkovich

Outside YaleLisa Ryden (Lund)Karin Jirstrom (Malmo)David Huntsman (BCCA)Karen Gelmon (BCCA)Elaine Alarid (UW)

Yale CollaboratorsAnnette MolinaroGina ChungHarriet KlugerRuth HalabanLyndsay HarrisJohn McClaskey

Former Rimm Group:Marisa Dolled-FilhartMaciej Zerkowski Kyle DeVitoMelissa CreggerTony McCabeSatish AvvaruGreg TedeschiMark GustavsonGretchen GraffCarola ZallesAaron BergerSharon MoulisMalini HarigopalChris MoederJena Giltnane

www.tissuearray.orgRimm Lab Summer ‘08

•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 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 tissueIntrinsic Controls for FFPE tissue

Goal-specific Task: Milestone:1A. To collect tissue blocks from patients with matched pairs of core biopsies and resection specimens

1A. Establishment of a fully annotated set of 300 core needle biopsy (CNB) and resection tissue pairs from breast cancer patients

1B. To collect a cohort of patient tissues with recorded time to fixation

1B. Construction of a 2x redundancy tissue microarray (TMA) representing 100 specimens with a wide range of recorded times to fixation

2. To select, titer, validate and assess expression of a series of 15-25 proteins that could have value in monitoring cold ischemia “house-keeping genes”

2. To select and report 3-5 proteins whose change is proportional and reflective of degeneration as a function of cold ischemic time

3. To select, titer, validate and assess expression of a series of 15-25 proteins that could have value in monitoring hypoxia

3. To select and report 3-5 proteins whose change is proportional and reflective of degeneration as a function of hypoxia

4. To use Part-DSA or other methods to define a set of markers from the best results from milestones 2 and 3 to generate five “Tissue Immunologic Competence”(TIC) models

4. To report the 5 best TIC models in preparation for competitive testing to choose the best one

5A. To test the 5 TIC models to assess EGFR expression on whole sections from the Thoracic Surgery Core Series Study.

5. To report, patent and prepare to commercialize the best TIC model based on the results of Tasks 5A-B.

5B. To test the 5 TIC models on an independently collected series of breast cancer cases with matched core biopsies to test ER expression.

Objective Optimization of Objective Optimization of Antibody ConcentrationAntibody Concentration

1

2

3

4

5

6

7

8

9

10

10

100

1000

10000

100000

Rat

io

AQ

UA

(R) S

core

s

Mark Gustavson and Marisa Dolled-Filhart