Part two endometriosis associated cancer David G. Huntsman BC - - PowerPoint PPT Presentation

Rare tumours: some recent data and ideas Part two – endometriosis associated cancer

David G. Huntsman

BC Cancer Agency Vancouver General Hospital University of British Columbia Canada Research Chair in Molecular and Genomic Pathology

Clear cell and endometrioid ovarian carcinoma represent around 20% of cases in North America

Treated as one disease despite different clinical presentation and survival

Clinical disease heterogeneity

MP10 In retrospective cohorts expect up to a 20% misclassification in chart based pathology

F Kommoss 2002 << F Kommoss 2014 = B Gilks 2014

<<< =

• 300 cases centrally reviewed in 2002
• Reviewed again by same pathologist using

2014 WHO criteria : 54% concordance

• New histotypes showed 98% concordance with

second reviewer and stronger associations with outcome and biomarkers

Impact of histotype changes

Clear cell carcinoma of the ovary

• 2nd most common ovarian

carcinoma subtype in NA (12%) and more frequent in Asia

• Do not respond to standard
• varian chemotherapy
• No other treatments available
• May respond to radiotherapy
• Molecular basis little

understood

• Weird cousins of renal CCC

ARID1a mutations

• Common in OCCC, Endometrioid of ovary and uterus and MSI

positive gastric cancers

• Found in cancer types without tp53 mutations
• Occur in precancerous lesions but may not be initiating events
• Not prognostic
• Apart from association with PIK3Ca mutations no reproducible

evidence that ARID1a mutant ovarian cancers are different from non mutant cases of the same type

• Specific targeting of ARID1a mutant cancers has been

challenging

ARID1A Clear cell ovarian carcinoma

2010 NEJM Wiegand K, et al.; 2010 Science Jones S, et al.

Genomic perspective Clear cell ovarian carcinoma

ERBB2

• verexpressed and amplified

Pro-oncogenic/transforming growth factor receptor MET

• verexpressed and amplified

Pro-oncogenic/transforming growth factor receptor …and more recently highlighted…

2010 GynOnc Anglesio M, et al.

Activated pathways in OCCC

Anglesio et al 2011. Clin Can Res

IL6  STAT3  HIF1A & HIF2A(EPAS1) (activation of hypoxia-related survival pathways)

Elevated levels: IL6 (Activated) STAT3 (Nuclear) HIF1A HIF2A (EPAS1) Nuclear HIF1a in OCCC

Genomic disruptions Clear cell ovarian carcinoma

Anglesio 2011 Tan 2011

MET ERBB2 HNF1B

Endometrioid carcinomas

• Almost all are low grade yet some progress to higher

grade cancers

• Stage 1 low grade endometrioid carcinomas of ovary

have a very good prognosis

• Higher grade endometriod carcinomas and recurrent

low grade need new treatment approaches

• POLE mutations in 5% of cases, MSI in >20%
• Beta catenin mutations in 50% of cases
• Often present with synchronous uterine carcinoma

Synchronous uterine and ovarian carcinomas

• Up to 50% of low grade endometrioid carcinomas
• Most are low grade and T1a
• Due to excellent prognosis are considered to be separate primaries
• Genomic and data molecular studies low resolution and interpreted as

supporting separate primaries

• Data to be shown non-validated comparisons of somatic mutations

SEO_ VAN_ 65 SEO_ TBG_ 22 150 125 SEO_ VAN_ 43 SEO_ VAN_ 22 SEO_ VAN_ 27 SEO_ VAN_ 29 SEO_ VAN_ 14 SEO_ VAN_ 54 SEO_ TBG_ 15 SEO_ VAN_ 58 SEO_ VAN_ 56 SEO_ VAN_ 60 SEO_ TBG_ 31 SEO_ VAN_ 04 SEO_ VAN_ 07 SEO_ VAN_ 08 SEO_ VAN_ 33 SEO_ VAN_ 40

Endometriosis? Ovarian Endometrial

In almost all cases the uterine and

• varian cancers share somatic

mutations Anglesio JNCI 2016

Copy number plots showing clonality

Data and Conclusions

• Clonal relationships between the endometrial and
• varian cancers seen in but one of 20 cases

studies so far

• Analysis of endometriosis and normal

endometrium should inform whether these are metastatic cancers or whether a mutant field defect leads to both uterine cancer and ovarian cancer through endometriosis

• Are these true metastasis?

Endometriosis: the main risk factor for CCC and ENOC

• First described by Sampson in 1925.

Pearce et al Lancet Oncology 2012

• 13,226 controls, 7,919 cases including 674 CCC

and 1220 endometrioid

Are these uterine cancers in the wrong place?

Gounaris et al, J Pathology 2011

MET (HGFR) amplification and overexpression in OCCC?

Fig 1 from Yamamoto et al, 2012. Mod Path

In second study by Yamamoto et al. MET overexpression and copy number changes were also correlated with atypical endometriosis that was synchronous with OCCC Endometriosis Adjacent atypical endometriosis

Regions of endometriosis that are synchronous to OCCC

Features found in OCCC can be found in adjacent endometriosis

H&E IHC CISH

VOA1048

Adjacent Atypical Endometriosis vs. OCCC

21

DAH145 - VOA1048 (in some cases the adjacent atypical endometriosis is essentially cancer

22

V O A 1 4 8 . A 1 5 a t y p i c a l e n d

• m

e t r i

• s

i s − a d j a c e n t V O A 1 4 8 . A 6 L e f t O v a r y C C C V O A 1 4 8 . B 6 E n d

• m

e t r i a l P

• l

y p V O A 1 4 8 . T L e f t O v a r y C C C

CCC (3a) AT-E-osis (3b) E-osis (3e) E-osis (3f)

Fig 3 A B

Anglesio J Path 2015

DAH72 – VOA734

24 ARID1A

Conclusions

• Adjacent atypical endometriosis can have a near

complete complement of mutations -final transformation events may not be mutations

• So far no explanation for why endometriosis can lead

to two such distinct cancers

• Are there more sensitive clonal marks for tracking

relatedness

• Is there a screening window ?
• What about endometriosis not associated with

cancer?

Deep infiltrating endometriosis

• Will other clinically

relevant forms of endometriosis have somatic mutations as have been seen in endometriosis associated with cancer

? Is endometriosis a partially competent neoplasm

Deep Infiltrating Endometriosis “Case 2”

NTC CTRL G12V CTRL CASE2 (LCM) Normal Tissue CASE2

Endometriosis (LCM)

KRAS Double-mutation positive

G12V G12A WT

Endometriosis, CCOC and ENOCa

?

How do two such different cancers arise from the same precursor?

How do different cancers arise from the same precursor? Do distinct mutations drive distinct oncogenic pathways

?

ENOCA and CCC: commonly mutated genes

Summary of specific genomic findings

• No single feature exclusive to endometrioid or CCC

discovered

• No feature seen exclusively in ARID1a wild type

cancers seen

• KMT2B (MLL4) the most commonly mutated “new”

gene of interest

Landscapes: Can the genomic landscape inform our understanding of the pathogenesis of these cancers

• Higher level view of cancer genome enables

identification of signatures that point to mutational process

• ENOCa and CCC compared to GCT and HGSCa

Copy number changes

GCT<<ENOCa<CCC<<HGSCa

Signatures as well as specific mutations track with cancer types

Signatures of mutational processes in human cancers: Alexandrov et al Nature 2014

ENOCA and CCC: genomic landscapes

ENOCA and CCC: genomic landscapes

APOBEC MMR AGE

How do different cancers arise from the same precursor?

?

Although differences no mutation is exclusive to these cancer types some landscape features are enriched Cancer associated mutations may precede transformation process (Anglesio)

Cysteine Biosynthetic Pathway

Methionine Homocysteine Cystathionine Cysteine

CBS CTH

Glutathione

Higher in clear cell

CCOC ENOC HGS

CTH is Highly Expressed in Clear Cell Ovarian Cancer

CTH and CBS Expression in Cell Lines

CTH CBS a-Tubulin A2780 IGROV1 TOV112D 2008 JHOC5 JHOC7 JHOC9 OVISE OVMANA OVTOKO RMG2 TOV21G CaOV3 Hey Kuramochi CaOV3 Hey OVCAR 3 OVCAR 4 OVCAR 5 OVSAYO

ENOC CCC HGS

Homocysteine Cystathionine Cysteine

CBS CTH

The Origins of Endometriosis Associated Ovarian Cancer?

Cell of Origin for Endometrioid Ovarian Cancer? Cell of Origin for Clear Cell Ovarian Cancer?

CCOC and EndoCa and the ovary

• Both cancers are associated with endometriosis
• Although cancer associated mutations occur in

endometriosis at other sites, transformation occurs almost exclusively within ovarian endometriomas

• CCOC and EndoCa look similar to their endometrial

counterparts and have similar mutations – do these cancers arise from different cells of origin?

• The IL6 pathway is dominant in OCCC, whereas

ARID1A/PIK3CA mutation occurs in approximately 50%

• f cases
• Is OCCC more than on disease and if so what marks

each type (proteomics screen)

Thanks

• My lab:, Niki Boyd ,Michelle Woo, Leah Prentice, Melissa

McConechy, Winnie Yang, Sarah Mains-Bandiera, Clara Salamanca, Michael Anglesio, Alicia Tone, Hector Li Chang, Yemin Wang, Jay Chen, Tony Karnezis,, Madlen Maassen and Janine Senz

• Sohrab Shah -- Bioinformatics: Jairhu Ding, Yikan Wang, Ali

Bashashati, Gavin Ha, Andrew McPherson, Gavin Ha

• GSC: Marco Marra, Martin Hirst, Gregg Morin
• Collaborators: Stefan Kommoss, M Kobel, Blaise Clarke, J

Brenton, AM Mes-Masson, D Bowtell, B Vanderhyden, A Okamoto and Sam Aparicio

• OvCaRe BC: Blake Gilks, Dianne Miller, Ken Swenerton, Paul

Hoskins, YZ Wang, Nelly Auersperg, Brad Nelson, Cal Roskelly, Tom Ehlen, Anna Tinker, Jessica McAlpine