Transcript HarvardCytology23ndCourse.Ross.2012

Gene Expression and Gene
Sequencing Using Cytology
Specimens
24st Annual Advances in Cytology Cytology
June 13, 2012
Jeffrey S. Ross, M.D.
Albany Medical College
Albany, NY
[email protected]
Gene Expression Profiling in Cytology
• Can be performed on fresh FNA material
• Difficult to perform on FFPE cell block
materials
• Clinical utility limited to Oncology
– Cancer of unknown primary site
– Prognosis assessment
– Prediction of response to targeted therapy
– Prediction of response to conventional therapy
Gene Expression Profiling Techniques
• RT-PCR
– For individual genes
– For genesets
• Oncotype DX
• Others
• Genomic Micorarrays
– For genesets
• Mammaprint
• Others
• Mathematic Models and Algorithms
– For results classification
– For disease outcome prediction
Advantages of Cytology Specimens
• Relatively easy to obtain new specimens
• Ease of repeated sampling for therapy response
information
• Lower cost of sampling
• FNA is Less invasive and better tolerated
• Improved patient compliance
• Tumor cell reduced cohesiveness enriches for
malignant cells vs stromal and inflammatory cells
Disadvantages of Cytology Specimens
• Small sample size
– Lower mRNA yield
– Reservoir of sample may be limited or absent
• May lack relevant histologic correlation(s)
• Tumor cell reduced cohesiveness enriches for
malignant cells vs stromal and inflammatory
cells
Expression Profiling for Disease
Classification Examples
• Lymphoma and Leukemia
• Solid Tumors (Breast, Prostate, Colorectal
and Lung Cancers)
• Carcinoma of Unknown Primary Site
• Microarray vs Multiplex RT-PCR
Microarray Classification of Non-Hodgkin’s Lymphoma
Staudt. Cancer Cell Vol. 2, No. 5, 11/02: 363 - 366
Site of Origin for Metastatic Adenocarcinoma
Dennis et al Cancer Res 2002;62:5999-6005
Expression Profiles of
61 Genes by SAGE
Established Tumor Markers by RTPCR in Common Adenocarcinomas
Expression Profiling in Breast Cancer
•
•
•
•
•
Molecular Portraits
Molecular Grading
ER/PR Testing
HER2 Testing
Oncotype DxTM Recurrence Score by RT-PCR
(Genomic Health)
• MammaprintTM 70 Gene Recurrence Predictor
(Agendia)
• Other Multigene Predictors
Comparison of Multigene Predictor Test Platforms (1)
Feature
IHC
Starting Material
FFPE
FISH
CISH
FFPE
RT-PCR
Microarray
FFPE or
Fresh/Frozen
Fresh/Frozen
Slide Based/
Morphology
Driven
Yes
Yes
No
No
Requires
Microdissection
No
No
Yes
Yes
Processing
Impact
Yes
Minimal
Minimal
No
Number of
Genes Tested
Small
Small
Intermediate
Large
Type of
Measurement
Semiquantitative
Semiquantitativ
e
Quantitative
Quantitative
Complexity of
Statistical
Algorithm
Straightforward
Straightforward
Complex
Highly
Complex
Comparison of Multigene Predictor Test Platforms (2)
Platform
Feature
IHC
FISH
CISH
RT-PCR
Microarray
Potential for False
Discovery
Low
Low
Intermediate
High
Ability to assess
multiple pathways
simultaneously
Low
Low
Intermediate
High
Stand Alone
Prognostic value
Established
Established
Established
Established
Prediction of
Response to
Hormonal
Therapies
Established
In current routine
practice
Not
Established
Established
Not in current
routine practice
Established
Not in current
routine practice
Prediction of
Response to HER2
Targeted Therapies
Established
In current routine
practice
Established
In current
routine
practice
Established
Not in current
routine practice
Established
Not in current
routine practice
Comparison of Multigene Predictor Test Platforms (3)
Platform/Feature
IHC
FISH
CISH
RT-PCR
Microarray
Prediction of
Response to
Cytotoxic Therapies
Not established
Not in current
routine clinical
practice
Not
established
Not in
routine
practice
Established
In current
routine clinical
practice
Established
Not in current
routine clinical
practice
Prediction of
Therapy Toxicity
Not established
Not in current
routine clinical
practice
Not
established
Not in
routine
practice
Not established
Not in current
routine clinical
practice
Not established
Not in current
routine clinical
practice
Cost
Comparatively
Low ($100-400)
Comparatively Low
($300-600)
Very High
($3,500 or
higher)
Very High
($3,500 or
higher)
Amenable to
standardization
Low
Low
Intermediate
High
Amenable to
Automated
Processing
Intermediate
Intermediat
e
High
High
Molecular Portrait of Breast Cancers
Basal–like
HER-2
“Normal
Luminal B
Luminal A
Sørlie et al. Proc Natl Acad Sci U S A. 2001 Sep 11;98(19):10869-10874.
Molecular Grading of Breast Cancer
• Gene expression profiling data indicates
that there are 2 molecular grades of breast
cancer
• Histologic Grade 2 cases redistribute into
Molecular Grades 1 and 2
• Molecular grading has outperformed
histologic grading in multivariate analysis of
traditional prognostic factors including
ER/PR and HER2 status
Patterns of expression of grade-related genes and their
association with histologic grade (HG) and relapse-free
survival. GGI score of each tumor is plotted below the
corresponding column. Relapse-free survival times in years are
indicated below the GGI scores.
Sotiriou C, et al.. Gene expression profiling in breast cancer: understanding the molecular basis of histologic grade to improve
prognosis. J Natl Cancer Inst. 2006 Feb 15;98(4):262-72.
Microscopic Grading in the Molecular Era
• How can we reduce the number of grade II cases and
increase the number of grade I and grade III tumors?
• Grade II (3+3+1 = 7/9 “Moderately Differentiated”)
– Architecture 3
– Nuclear grade 3
– Mitosis count 1
• Use Ki-67 labeling index to control for artifact low
mitotic figure count in cases where tumors were left
at room temperature for too long prior to fixation
ER Testing: Concordance Between IHC Status
and mRNA Levels
Relative ER mRNA
Expression
25
20
15
10
5
0
+ + + + + + + + + + + + ++ + + + + + + + + + + + - - - - - - - - - - - - - - - - - - - + - + - -
ER Status by IHC on Core Needle
Biopsies
HER-2 Testing: Is CISH the “Kish of Death” for FISH
and IHC?
•
Will IHC continue to be the international method of choice for screening with
2+ cases triaged to FISH?
•
Will primary FISH testing become the standard?
•
Will mRNA detection gain in popularity?
•
Will the recently approved CISH (SISH) assay become the preferred method?
•
Will the ToGA Trial and FDA approval for trastuzumab in gastric/GEJ cancer
change how HER2 testing is done?
•
Will pertuzimab and trastuzumab-DM1 require HER2+ testing prior to use?
IHC/FISH/CISH Concordance
Overall FISH/CISH Concordance = 98%
Hanna WM, Kwok K. Chromogenic in-situ hybridization: a viable alternative to fluorescence
in-situ hybridization in the HER2 testing algorithm. Mod Pathol. 2006 Apr;19(4):481-7.
Multigene Classifiers and Predictors of Breast
Cancer Clinical Outcome*
*Ross, JS, Hatzis C, Symmans WF, Pusztai L, Hortobagyi GN. Commericalized multi-gene predictors of clinical outcome in breast
cancer. Oncologist. 2008;13:477-493.
Multi-Gene Prognostic Tests
• Pathways in common
– Proliferation
– Hormone Receptor
– HER2
• Challenges for Clinical Acceptance
– Associations are by chance only
• Overfitting the data
• Separate validation group must be used
– Bias
– Generalizability
Companion Diagnostics: Potential Uses in Cancer Drug
Development
• Biomarkers for drug response and drug-induced toxicity
• Comparison of human response to pre-clinical animal models
• Identify genes with variants that may define patient
populations
• Identify proteins as potential biomarkers
• Shorten duration and lower cost of clinical trials by selecting
patients more likely to respond and less likely to suffer toxicity
• Improve patient compliance with “custom-designed”
medicines
• Achieve “best in class” status and premium pricing to
overcome market segmentation
Important Anti-Cancer Drugs Requiring
Companion Diagnostic Testing for Use
•
•
•
•
•
•
•
Tamoxifen and AI’s: ER/PR
Trastuzumab and Lapatinib: HER2
Imatinib and Dasatinib: BCR:ABL, CKIT, PDGF
Cetuximab and Panitumumab: KRAS, (BRAF)
Gefitinib and Erlotinib: EGFR
Vemurafenib: BRAF
Crizotinib: EML4:ALK
Fine Needle Aspirate vs Core Biopsy
FNA
Less invasive
Cell suspension enriched for
tumor cells
Average 80-85% Tumor Cells
CBX
More invasive
Includes stromal, benign epithelial
and inflammatory cells
Average 55-60% Tumor Cells
Total RNA Yield From Single Pass FNA or Core
Biopsy
Count
10
FNA:
Usable RNA in 46 / 63
samples (73%)
mean = 3.6 µg
median = 2.2 µg
5
0
-5
0
5
10
15
20
25
RNA yield FNA
4
CBX:
Usable RNA in 15 / 20
samples (75%)
mean = 2.8 µg
median = 2.0 µg
Count
3
2
1
0
-5
0
5
10
15
RNA yield CBX
20
25
Symmans et al. Cancer
15;97(12):2960-2971
FNA vs Core Biopsy Conclusion
• FNAs and CBxs are similarly suitable for
transcriptional profiling
• FNAs have significant advantages
–
–
–
–
–
More acceptable to patients,
Higher consent rate
Faster to perform
Less expensive
Provide A Higher Percentage Of Cancer Cells In The Sample
• The Stromal Transcriptional Component Will
Probably NOT Influence The Predictive Power Of
Most Genomic Studies For:
–
–
–
Prognosis
Biology
Therapy Response (?)
DNA Sequencing in Cytology
• DNA yield from FNAs may not always be
satisfactory
• Needle core biopsies now preferred
• DNA can also be extracted from FFPE cellblocks
from fluids and FNAs
• Sequencing NSCLC FNAs is especially important
–
–
–
–
EGFR for gefitinib/erlotinib
EML4:ALK for crizotinib
KRAS for cetuximab (not approved in NSCLC)
EML4:ALK for crizotinib
Background (1)
• Next Generation DNA Sequencing (NGS) has recently
been applied to FFPE cancer biopsies and major
resections (Ross JS et al. J Clin Oncol 29: 2011)
• Current Hot-Spot Genotyping only detects:
– Mutations restricted to specific exons and codons
• NGS detects:
–
–
–
–
Whole exome mutations in numerous cancer related genes
Insertions and deletions
Translocations and fusions
Copy number alterations (amplifications)
Background (2)
• Recently, biomarker testing has emerged as a major driver of
the selection of therapy for non-small cell lung cancer
(NSCLC), colorectal cancer (CRC) and melanoma
• Currently, “hot-spot” DNA sequencing and FISH are used to
select therapies for solid tumors:
–
–
–
–
EGFR genotyping in NSCLC for tyrosine kinase inhibitor (erlotinib)
EML4:ALK translocation testing in NSCLC for crizotinib
KRAS mutation testing in CRC for cetuximab
BRAF mutation testing in melanoma for vemurafenib
• The emergence of comprehensive genomic profiling by NGS
has led investigators to question whether more thorough
gene sequencing techniques could discover potential targets
for the treatment of relapsed and metastatic NSCLC not
currently searched for in current routine practice
Cancer Genome Profiling Workflow
<14-21 days
EGFR Activating Mutation – NSCLC
•
•
•
•
Mutation: EGFR_c.2573T>G_p.L858R
Freq=32%, depth=53
79 year old white female
FNA of lung mass: NSCLC
FNA sample cytocentrifuged and
converted to an FFPE tissue block.
Very small numbers of viable tumor
cells. Extensively necrotic
.
KRAS Mutation – CRC
•
•
•
•
•
Mutation: KRAS_c.35G>T_p.G12V
Freq=30%, depth=283
52 year old white male
KRASG12V mutation by “hot-spot” genotyping at
Commercial Laboratory
pT3 pN2 pMx CRC
Classic
CRC with
origin from
mucosal
surface at
lower right
BRAF V600K Mutation – Metastatic MM
•
•
•
•
•
Mutation: BRAF_c.1798_1799GT>AA_p.V600K
Freq=10%, depth=416
77 year old white male
Thick melanoma of back
Multiple posterior cervical lymph nodes positive for metastatic
melanoma
Metastatic Melanoma to
a cervical lymph node
EML4-ALK Translocation in NSCLC
•
PF-02341066 (PF-1066) “Critzotinib”
– Oral ALK4 receptor kinase inhibitor
– Phase I Trial on NSCLC patients with EML4-ALK
translocation
• “echinoderm microtubule-associated protein-like 4”
– “anaplastic lymphoma kinase”
– 10/19 (53%) had a partial response ASCO ‘09
– 76 patients detected by break-apart FISH ASCO ‘10
•
•
•
Seen in 5-13% of adenocarcinomas
–
–
•
•
ORR (overall response rate) 64%
DCR (disease control rate) 90%
Papillary, solid and signet ring cell features
Mostly never or non-smokers
All ELM4-ALK Positive NSCLC is Negative for EGFR
mutation
ELM4-ALK translocation can be detected by FISH
or PCR
–
IHC for ALK available, but no outcome data for crtizotinib
treated patients
Shaw et al. J Clin Oncol. 27;2009:4247-4253.
NSCLC: JAK2 Mutation Detected by NGS
•
•
•
•
•
•
Sample: SM86
Mutation: JAK2_c.1849G>T_p.V617F
Freq=4%, depth=205
77 year old white female
Lung adenocarcinoma diagnosed by pleural biopsy
Patient diagnosed with polycythemia vera
Low power of pleural biopsy positive for
adenocarcinoma
c.1849G>T p.V617F
High power view shows adenocarcinoma of the lung.
Rare capillaries not blood filled. No nucleated RBC or
blasts seen.
G T A T G T GT C T G T G G A
Val
Cys
Val
Cys
Gly
HER2 Gene Copy Number Alteration Validation
Increased HER2 CNA detected by NGS
HER2 FISH Positive Breast Invasive Duct
Carcinoma Demonstrating High HER2 CNA
CONFIDENTIAL
ERBB2 RARA
HER2 Protein 3+ Expression by IHC
35
cMET Copy Number Alteration in Clear Cell Ovarian
Carcinoma Validation
cMET average CNA by FISH at APS is 6.6 using
Abbott-Vysis Probes for cMET and CEP 7
cMET IHC at
APS shows H
Score = 300
with 100% 3+
Membrane
Staining
cMET gene CNA at FMI
estimated at 6 copies in
40% purity
CONFIDENTIAL
36
Novel ALK Fusion in CRC Detected by NGS
A 5,194,955-bp tandem duplication
generates an in-frame C2orf44-ALK gene
fusion
The RNA sequence of the C2orf44-ALK
gene fusion shows aberrant splicing
pT4pN1pM1 Mucinous Adenocarcinoma
associated with a serrated sessile polyp
RNA sequencing shows an 89.8-fold
increase in expression of ALK beginning
at exon 20 relative to exons 1–19.
Lipson et al. Nature Med, Feb, 2012
Novel RET:KIF 5B Rearrangement in NSCLC
(11.3Mb Pericentric Inversion)
KIF5B
RET
Break
Break
ATG
ATG
32,316,377 bps
43,611,118 bps
ATG
ATG
RET-KIF5B
ATG
KIF5B-RET
KIF5B-RET
Translation
Kinesin
Novel gene fusion joining exons 1-15 of KIF5B to exons
12-20
20) of RET in lung adenocarcinoma
Coiled coil Tyrosine kinase
KIF5B (exons 1—15)
RET (exons 12—
Lipson et al. Nature Med, Feb, 2012
Percentage Of Samples With Actionable Mutations Across
Major Tissue Types
71% cases carried ≥1 plausibly actionable mutation
32 % cases carried ≥2 plausibly actionable mutations
N=94
N=76
N=31
N=29
N=24
100%
90%
80%
Percent of Cases
70%
60%
50%
No Mutations Found
Unknown Actionability
40%
Actionable Mutations
30%
20%
10%
0%
Colorectal
Lung(NSCLC)
Prostate
Tumor Type
39
CONFIDENTIAL
Breast
Melanoma
Comparison of NGS with Traditional Hot-Spot Genotyping
in NSCLC, CRC, Breast Cancer and Melanoma
Also Detected by Hot-Spot Genotyping
N = 111
Targeted Therapies for Cancer
Molecular profiling is driving many new targeted cancer therapeutics
• ~500 compounds hitting ~140
targets in development
• Growing number of newly
identified potential targets
Subset of analyzed targets
listed; data from BioCentury
Online Intelligence
Database