Copy Number Aberrations in Benign Serous Ovarian Tumors: a Case for Reclassification?
Total Page:16
File Type:pdf, Size:1020Kb
Published OnlineFirst October 5, 2011; DOI: 10.1158/1078-0432.CCR-11-2080 Clinical Cancer Human Cancer Biology Research Copy Number Aberrations in Benign Serous Ovarian Tumors: A Case for Reclassification? Sally M. Hunter1, Michael S. Anglesio6, Raghwa Sharma3, C. Blake Gilks6,7, Nataliya Melnyk6, Yoke-Eng Chiew4,5, Anna deFazio4,5 for the Australian Ovarian Cancer Study Group, Teri A. Longacre8, David G. Huntsman6,7, Kylie L. Gorringe1,2, and Ian G. Campbell1,2 Abstract Purpose: Serous ovarian carcinomas are the predominant epithelial ovarian cancer subtype and it has been widely believed that some or all of these may arise from precursors derived from the ovarian surface epithelium or fimbriae, although direct molecular evidence for this is limited. This study aimed to conduct copy number (CN) analysis using a series of benign and borderline serous ovarian tumors to identify underlying genomic changes that may be indicative of early events in tumorigenesis. Experimental Design: High resolution CN analysis was conducted on DNA from the epithelial and fibroblast components of a cohort of benign (N ¼ 39) and borderline (N ¼ 24) serous tumors using the Affymetrix OncoScan assay and SNP6.0 arrays. Results: CN aberrations were detected in the epithelium of only 2.9% (1 of 35) of serous cystadenomas and cystadenofibromas. In contrast, CN aberrations were detected in the epithelium of 67% (16 of 24) of the serous borderline tumors (SBT). Unexpectedly, CN aberrations were detected in the fibroblasts of 33% (13 of 39) of the benign serous tumors and in 15% (3 of 20) of the SBTs. Of the 16 cases with CN aberrations in the fibroblasts, 12 of these carried a gain of chromosome 12. Conclusions: Chromosome 12 trisomy has been previously identified in pure fibromas, supporting the concept that a significant proportion of benign serous tumors are in fact primary fibromas with an associated cystic mass. This is the first high resolution genomic analysis of benign serous ovarian tumors and has shown not only that the majority of benign serous tumors have no genetic evidence of epithelial neoplasia but that a significant proportion may be more accurately classified as primary fibromas. Clin Cancer Res; 17(23); 7273–82. Ó2011 AACR. Introduction At the time of diagnosis, women with epithelial ovarian cancer usually have advanced disease and as a consequence Ovarian cancer is a very significant health burden and the their prognosis is extremely poor (5-year survival rate for 7th leading cause of cancer death in women worldwide (1). stage III and IV disease is only 25%–30%; refs. 2, 3). For such a clinically significant disease, remarkably little is known about the molecular events that initiate the disease. Authors' Affiliations: 1Centre for Cancer Genomics and Predictive Med- Although the paradigm that malignancies arise through a icine, Peter MacCallum Cancer Centre, Melbourne; 2Department of Pathol- ogy, University of Melbourne, Parkville, Victoria; 3Anatomical Pathology, stepwise progression from benign precursors has been University of Sydney and University of Western Sydney at Westmead established for many malignancies, the archetypical exam- 4 5 Hospital; Department of Gynaecological Oncology, Westmead Institute ple being colorectal carcinogenesis (4), it remains unclear if for Cancer Research,University of Sydney at Westmead Millennium Institute, Westmead Hospital, Sydney, New South Wales, Australia; 6The Department this holds true for ovarian cancer. There is still considerable of Pathology and Laboratory Medicine, University of British Columbia; controversy as to what constitutes a true ovarian cancer 7 Genetic Pathology Evaluation Centre of the Prostate Research Centre and precursor; an important definition that needs to be made to Department of Pathology, Vancouver General Hospital and University of British Columbia, Vancouver, British Columbia, Canada; 8Department of understand the origins and identify new clinical interven- Pathology, Stanford University School of Medicine, Stanford, California tions for this lethal disease. Note: Supplementary data for this article are available at Clinical Cancer Serous ovarian carcinomas are the predominant clinically Research Online (http://clincancerres.aacrjournals.org/). important subtype but at present there is little experimental K.L. Gorringe and I.G. Campbell are co-senior authors. evidence from which to draw convincing conclusions about what constitutes the precursor(s) to this subtype. It has been Corresponding Author: Ian G. Campbell, VBCRC Cancer Genetics Research Laboratory, Peter MacCallum Cancer Centre, Locked Bag 1, widely believed that some or all of these arise from pre- A'Beckett Street, Melbourne, Victoria 8006, Australia. Phone: 613-9656- cursors originating from the ovarian surface epithelium, 1803; Fax: 613-9656-1411; E-mail: [email protected] via inclusion cysts or serous benign and borderline tumors doi: 10.1158/1078-0432.CCR-11-2080 (5–8). Obvious candidate precursors are serous ovarian Ó2011 American Association for Cancer Research. cystadenomas and cystadenofibromas, which are benign www.aacrjournals.org 7273 Downloaded from clincancerres.aacrjournals.org on September 27, 2021. © 2011 American Association for Cancer Research. Published OnlineFirst October 5, 2011; DOI: 10.1158/1078-0432.CCR-11-2080 Hunter et al. cursors to some invasive serous ovarian carcinomas, we Translational Relevance have conducted high-resolution CN analysis on a series of benign and borderline serous ovarian tumors. This is the Ovarian cancer is a very significant health burden and first ultra-high-resolution CN analysis of benign serous the seventh leading cause of cancer death in women. At tumors of the ovary. the time of diagnosis, women often have advanced disease and as a consequence their prognosis is extremely Materials and Methods poor. Our understanding of the progression of ovarian cancer through precursor stages and the molecular genet- Tissue samples ic events underlying these changes is very limited. Only fresh frozen tissue samples were used in this for CN Although a number of candidate precursor lesions have and mutation analyses. All samples were collected with the been proposed, the true contribution of these precursor informed consent of patients and the study was approved by lesions to the onset of ovarian cancer is unresolved. the Human Research Ethics Committees at the Peter Mac- Identifying genuine ovarian cancer precursors and defin- Callum Cancer Centre, Queensland Institute of Medical ing key molecular genetic events initiating and promot- Research, University of Melbourne and all participating ing tumorigenesis have important implications in early hospitals. Patients with ovarian tumors were identified detection and treatment of ovarian cancers. through 2 primary sources: (i) 9 at hospitals in South- ampton (24), United Kingdom, (ii) 54 through the Austra- lian Ovarian Cancer Study (25). Pathology reviews were done independently by 2 gynecologic pathologists (R. lesions with a cystic mass of 1 cm or more in diameter, lined Sharma and C.B. Gilks). Pathology review was conducted with a single layer of cuboidal to columnar epithelium and on cryosections adjacent to the tissue from which DNA was commonly associated with a fibromatous stromal mass extracted (n ¼ 63). (Supplementary Fig. S1). The serous epithelial layer of these tumors typically displays minimal cellular proliferation and Microdissection and DNA extraction no nuclear atypia. These are relatively common tumors, A representative hematoxylin and eosin (H&E) stained accounting for 60% of all serous ovarian tumors, whereas section was assessed and needle microdissection was done serous borderline tumors (SBT) and low-grade serous car- using 10 mm sections to obtain high percentage tumor cinomas (LGSC) account for 10%–15% and 2%–9% of all epithelial cell and fibroblast cell components. DNA was serous ovarian tumors, respectively (9–12). Benign serous extracted using the Qiagen Blood and Tissue Kit (Qiagen). tumors are presumed by many to be precursors to SBTs Normal DNA extracted from blood lymphocytes was avail- based on similarities in the cystic structure of some SBTs and able for all 63 patients. the frequent detection of cases with a benign cyst coexisting with a SBT or cases comprising predominantly benign cysts CN arrays with regions of atypical proliferation (13). Despite the A subset of cases were processed by Affymetrix for the cooccurrence of benign, borderline, and low-grade carci- OncoScan (Molecular Inversion Probe) assay, which con- noma epithelial components, direct molecular evidence sists of a 330K probe set that allows the detection of supporting benign lesions as precursors is limited. Although genome-wide, allele-specific CN. OncoScan data normal- some studies have shown the existence of KRAS and BRAF ization was done by Affymetrix Inc. as previously mutations in ovarian serous cystadenomas and cystadeno- described (26, 27). Where sufficient material ( 250 ng fibromas co-existing with a region of atypical proliferation DNA) was available, the Affymetrix SNP6.0 (1.8M probe or adjacent SBT (14), mutations in these genes have not set) arrays were utilized for ultra-high-resolution allele- been detectable in solitary benign tumors. specific CN analysis, although prior to its release the SBTs have been firmly established as the likely precursor Affymetrix 500K array was used. For the SNP6.0 array lesions to LGSCs, sharing similar rates of KRAS and BRAF the input was reduced from the recommended 500 to 250 mutation