Advertisement
Imaging, Diagnosis, Prognosis

Genomic Analysis Reveals the Molecular Heterogeneity of Ovarian Clear Cell Carcinomas

David S.P. Tan, Marjan Iravani, W. Glenn McCluggage, Maryou B.K. Lambros, Fernanda Milanezi, Alan Mackay, Charles Gourley, Felipe C. Geyer, Radost Vatcheva, Joanne Millar, Karen Thomas, Rachael Natrajan, Kay Savage, Kerry Fenwick, Alistair Williams, Charles Jameson, Mona El-Bahrawy, Martin E. Gore, Hani Gabra, Stanley B. Kaye, Alan Ashworth and Jorge S. Reis-Filho
DOI: 10.1158/1078-0432.CCR-10-1688 Published March 2011

Abstract

Purpose: Ovarian clear cell carcinomas (OCCC) are a drug-resistant and aggressive type of epithelial ovarian cancer. We analyzed the molecular genetic profiles of OCCCs to determine whether distinct genomic subgroups of OCCCs exist.

Experimental design: Fifty pure primary OCCCs were subjected to high-resolution microarray-based comparative genomic hybridization (aCGH). Unsupervised hierarchical clustering using Ward's linkage analysis was performed to identify genomic subgroups of OCCCs. Survival analysis was performed using Kaplan–Meier method and log-rank test. Cox-regression analysis was used to identify independent predictors of outcome. Differentially amplified regions between genomic subgroups of OCCCs were identified using a multi-Fisher's exact test.

Results: Hierarchical cluster analysis revealed two distinct clusters of OCCCs with different clinical outcomes. Patients from cluster-1 had a significantly shorter median progression-free survival (PFS) than those from cluster-2 (11 vs. 65 months, P = 0.009), although estimates for ovarian cancer–specific survival (OCS) did not reach statistical significance (P = 0.065). In multivariate analysis, suboptimal debulking surgery and genomic cluster were independently prognostic for PFS. Recurrently amplified genomic regions with a significantly higher prevalence in cluster-1 than cluster-2 OCCCs were identified and validated. HER2 gene amplification and protein overexpression was observed in 14% of OCCCs, suggesting that this may constitute a potential therapeutic target for a subgroup of these tumors.

Conclusions: OCCCs constitute a heterogeneous disease at the genomic level despite having similar histological features. The pattern of genomic aberrations in subgroups of OCCCs is of clinical significance. We have identified recurrently amplified regions that may harbor potential therapeutic targets for subgroups of OCCCs. Clin Cancer Res; 17(6); 1521–34. ©2011 AACR.

This article is featured in Highlights of This Issue, p. 1213

Translational/Clinical Relevance

Ovarian clear cell carcinomas (OCCC) are aggressive and chemoresistant tumors comprising approximately 13% of all epithelial ovarian cancers. OCCCs are considered high-grade tumors and histological features have not been shown to predict outcome. We performed high-resolution array comparative genomic hybridization (aCGH) analysis of OCCCs aiming to determine whether distinct genomic subgroups of these tumors exist. Unsupervised analysis based on the genomic copy aberrations revealed two distinct genomic subgroups of OCCCs (cluster-1 and cluster-2). Tumors from each cluster displayed distinct patterns of genetic aberrations; however, they did not significantly differ in terms of their clinicopathological and histological features. Cluster-1 OCCCs had a significantly shorter median progression-free survival (PFS) and subsequent multivariate analysis revealed that genomic cluster was an independent prognostic factor for PFS. Our data suggest that OCCCs are genomically heterogeneous and the pattern and complexity of genome-wide copy number aberrations may determine clinical outcome of OCCCs.

Introduction

Ovarian clear cell carcinoma (OCCC) is an uncommon subtype of epithelial ovarian carcinoma (EOC) that is associated with a reduced response rate following platinum-based chemotherapy (1–3) and a poorer prognosis, especially in advanced stage, in comparison with other histological subtypes of EOC (4–6).

The only established independent prognostic indicators of outcome for OCCCs are International Federation of Gynaecology and Obstetrics (FIGO) stage and the extent of residual disease following debulking surgery (3, 6, 7). Histopathological parameters have not been shown to be prognostically informative and OCCCs are considered to be high-grade tumors (8). Given that the current standard of care in EOC consists of optimal debulking surgery followed by platinum-based chemotherapy, the observed platinum-resistance of OCCCs, particularly in advanced FIGO stage, poses a considerable clinical challenge. Furthermore, there is a paucity of validated targeted agents or effective alternative systemic therapies for patients with this disease (4). The inability to predict outcomes or platinum/ taxane sensitivity in OCCC patients from intrinsic characteristics of the primary tumor also results in considerable difficulty for both patients and gynecological oncologists when making decisions about chemotherapy in both early and advanced-FIGO stage OCCC.

Microarray-based comparative genomic hybridization (aCGH) provides a genome-wide survey of copy number gains, losses, and amplification in solid tumors. This technique has proven useful in characterizing the genomic profiles of tumors, and identifying prognostic and predictive DNA copy number aberrations (9). Hitherto, only a few studies have focused on the analysis of DNA copy number changes associated with OCCC (10–13). Based on an analysis of OCCC cell lines, we previously identified PPM1D as a likely amplicon driver and potential therapeutic target in OCCCs harboring 17q23.2 amplification (11). More recently, Kuo and colleagues used a 250k single-nucleotide polymorphism array to analyze the genome-wide copy number changes in 12 affinity-purified OCCCs and identified a recurrent amplification of the putative oncogene ZNF217 (12). Only one previous aCGH study (using a cDNA platform comprised of 10,618 probes) was performed on OCCCs to identify prognostic markers (13), and this found that upregulation of ABCF2 was associated with poor prognosis in OCCCs.

Based on previous genomic analyses (11–13), we hypothesized that OCCCs would comprise a heterogeneous group of tumors in terms of their DNA copy number aberrations. Using a high resolution aCGH platform, we performed a detailed molecular characterization of 50 OCCCs with the following aims: (i) to characterize the patterns of copy number gains, losses and amplifications in these tumors and (ii) to assess whether genomic subgroups of OCCCs are associated with distinct outcomes.

Materials and Methods

Selection of ovarian clear cell carcinomas

Local ethical approval for this study was granted by the Royal Marsden Hospital Research and Ethics Committee on 11th May 2006 (RMH REC Committee Ref: 06/Q801/23). Archival formalin-fixed, paraffin-embedded (FFPE) OCCCs were obtained from the pathology departments of The Edinburgh Royal Infirmary (AW), The Belfast Health and Social Care Trust (WGM), Royal Marsden Hospital, London (CJ), and Hammersmith Hospital, London (ME). All OCCCs were reviewed and selected by pathologists with an interest in ovarian cancer pathology at respective hospitals before undergoing a second central review and selection process by 2 gynecological pathologists with experience and track record in ovarian cancer pathology (CJ, WGM).

All of the patients underwent surgical debulking as initial management. None of the selected patients had received any systemic anticancer therapy prior to primary debulking surgery. Tumor stage was defined according to the FIGO staging system. In total, 88 archival surgical resection specimens of primary, pure (no component of EOC other than OCCC) OCCCs were identified, of which only 50 tumors had DNA of sufficient quality for aCGH analysis (see the following text). A review of the histopathological characteristics of these 50 tumors was performed by one of the authors (WGM), blinded to the results of aCGH analysis. Histopathological parameters assessed included: (i) predominant architectural pattern—glandular, solid, papillary or tubulocystic; (ii) mitotic count per 10 high-power fields from most mitotically active areas (50 high-power fields counted); (iii) degree of nuclear pleomorphism categorized as mild (round to oval nuclei with even distribution of chromatin and inconspicuous nucleoli); moderate (irregular nuclei with chromatin clumping and moderate sized nucleoli); and severe (large pleomorphic nuclei with coarse chromatin and large irregular nucleoli); (iv) presence or absence of necrosis; (v) presence or absence of an adenofibroma-like growth pattern—defined as dilated glands lined by epithelial cells within a fibrous stroma; (vi) presence or absence of inflammation.

DNA extraction

Five 8-μm-thick paraffin sections of each OCCC included in the study were stained with nuclear fast red and selected tumor areas were manually microdissected with a sterile needle (Terumo Corporations) under a stereomicroscope (Olympus SZ61s) to remove lymphocytic infiltrates and stromal components to achieve a minimum of a greater than 70% composition of tumor cells, as previously described (14–17). DNA from the FFPE samples was extracted using the DNeasy kit (Qiagen Ltd) according to the manufacturer's recommendations. DNA was quantified using the Quant-IT PicoGreen dsDNA Assay Kit (Molecular Probe; Invitrogen) according to manufacturer's instructions. A multiplex-PCR–based quality control assay was performed for DNA extracted from archival FFPE OCCC samples (15). Out of 88 OCCC samples, 50 samples yielded DNA of sufficient quantity and quality to be subjected to aCGH analysis.

aCGH analysis

The aCGH platform used for this study was constructed by the Breakthrough Microarray Laboratory and comprises a fluorescence in situ hybridization (FISH) mapped tiling-path rearray set of 32,000 bacterial artificial chromosome (BAC) clones from Children's Hospital Oakland Research Institute (http://bacpac.chori.org/) spaced at approximately 50 kb intervals throughout the genome. All 32,000 BAC clones were spotted onto single Corning GAPSII-coated glass slides (Corning). This platform has been shown to be as robust as and to have comparable resolution with high density oligonucleotide arrays (18, 19).

The aCGH protocol has previously been described (14–17). Briefly 400 to 500 ng of reference (normal DNA extracted from blood lymphocytes pooled from 24 females) and tumor DNA were labeled with Cy3- or Cy5-conjugated dCTP (Amersham Biosciences) using random primed BioPrime DNA labeling (Invitrogen Life Technologies) according to the manufacturer's protocol, modified to incorporate 1.0 mmol/L Cy dye, 0.6 mmol/L dCTP, and 1.2 mmol/L dATP, dGTP, and dTTP. After labeling at 37°C for 18 hours, nonincorporated reaction constituents were removed by MinElute Reaction Cleanup (Qiagen Ltd). Labeled reference and tumor DNA were coprecipitated with 100 μg human Cot-1 DNA (Invitrogen Life Technologies) at −20°C for 1 hour (DNA from cell line and fresh frozen tissue at −80°C for 1–2 hours) followed by resuspension in 50 μL hybridization buffer (50% deionized formamide, 10% w/v dextran sulphate, 2× SSC, 2% SDS, 20 μg yeast tRNA). Denaturation of labeled DNA was performed at 70°C for 15 minutes, followed by incubation at 37°C for 30 minutes to allow blocking of repetitive sequences by human Cot-1 DNA. Denatured DNA samples were hybridized to the microarray at 37°C overnight. Posthybridization, slides were washed in 2× SSC/0.1% SDS for 15 minutes at 45°C to remove the coverslips, followed by 50% deionized formamide/2× SSC wash for 15 minutes at 45°C. Finally, slides were washed in 2× SSC/1% SDS for 30 minutes at 45°C and twice in 0.2× SSC for 15 minutes at room temperature. After posthybridization washing, slides were dried by centrifugation at 2,000 rpm for 2 minutes.

Following hybridization and washes, slides were scanned using an Axon 4000B scanner (Axon Instruments) and images were processed using Genepix Pro 5.1 image analysis software (Axon Instruments). Log2 ratios were normalized for spatial and intensity-dependent biases using a 2-dimensional loess regression followed by a BAC-dependent bias correction as previously described (15). This left a final dataset of 31,213 clones with unambiguous mapping information according to the Feb 2009 ensembl assembly 55 build 37 (hg 19) of the human genome (http://www.ensembl.org). Data were smoothed using a local polynomial adaptive weights smoothing (aws) procedure for regression problems with additive errors (20).

Thresholds for copy number gains and losses were chosen to correspond to 3 standard deviations of the normal ratios obtained from the filtered clones mapping to chromosomes 1 to 22, assessed in multiple hybridization between DNA extracted from a pool of male and female blood donors as previously described (Log2 ratio of ±0.12; refs. 15, 17). A categorical analysis was applied to each clone on the array after classification as amplified/high level gain, gain, loss, or no-change according to their aws-smoothed Log2 ratios. Low level gain was defined as an aws-smoothed Log2 ratio between 0.12 to 0.4, corresponding to approximately 3 to 5 copies of the locus, whereas gene amplification/high level gain was defined as having an aws-smoothed Log2 ratio more than 0.4, corresponding to more than 5 copies. These figures were obtained by comparison with interphase FISH and chromogenic in situ hybridization (CISH) data for markers at different chromosomal locations as previously described (11, 16, 17). Data processing and analysis was carried out in R 2.9.0 (http://www.r-project.org/) and BioConductor 2.5 (http://www.bioconductor.org/), making extensive use of modified versions of the packages aCGH, marray, and aws in particular (15).

Differentiation of genomic patterns

We classified the genomic profiles of OCCCs into 3 subgroups following the system proposed for breast cancer by Hicks and colleagues (21): ‘simplex’, ‘sawtooth’, and ‘firestorm’. Cases were considered of ‘simplex’ pattern if their genomic profiles were characterized by broad segments of duplication and deletion, usually comprising entire chromosomes or chromosome arms, with occasional isolated narrow peaks of amplification. Complex patterns included ‘sawtooth’ and ‘firestorm’ profiles. Cases with a ‘sawtooth’ profile were characterized by many narrow segments of duplication and deletion, often alternating and affecting all chromosomes. In these cases, although multiple gains and losses occurred throughout the genome, amplifications were rarely found. Cases were considered of ‘firestorm’ pattern if they resembled the ‘simplex’ type except that the profiles contained at least 1 localized region of clustered, relatively narrow peaks of amplification, with each cluster confined to a single chromosome arm. Genome patterns (i.e., ‘simplex’, ‘sawtooth’, and ‘firestorm’) were determined through visual inspections of the genome plots by 3 observers independently (DT, RN, and ML) as previously described (11, 16). In all cases, a perfect agreement between the observers was achieved.

Unsupervised analysis of aCGH data

For unsupervised analysis of aCGH data, copy number changes were categorized according to the aforementioned thresholds for each clone and converted to integer values (high-level gains/ amplified = 2, gains = 1, losses = −1 or no copy number change = 0) before a hierarchical clustering analysis was performed for all OCCCs profiled using Ward's linkage analysis (Euclidean distance; ref. 22) as previously described (16, 23).

For supervised analysis, copy number changes were categorized as gains, losses, or amplifications according to the aforementioned thresholds for each clone before using Fisher's exact test [with adjustment for multiple testing using the step-down permutation procedure maxT, providing strong control of the family-wise type I error rate (FEWER; refs. 16, 23)] to identify statistically significant (adjusted P value < 0.05) differences between the genomic profiles of cluster-1 and cluster-2 OCCCs. Data processing and statistical analysis was carried out in R 2.9.0 (http://www.r-project.org/) and BioConductor 2.5 (http://www.bioconductor.org/). The aCGH data (i.e., Log2 ratios and smoothed Log2 ratios) are available on the ROCK database (http://rock.icr.ac.uk/collaborations/TanDS).

Statistical analysis of clinical outcome data

Known prognostic factors in EOC were recorded for each patient as shown in Table 1. Characteristics of patients in cluster-1 and cluster-2 (age, FIGO stage, chemotherapy, and histopathological characteristics) were compared using Fisher's exact test (for categorical data) or a Mann–Whitney test (for continuous data; Table 1). P values less than 0.05 were considered to be significant. Ovarian cancer specific survival (OCS) and progression-free survival (PFS) for patients in cluster-1 versus cluster-2 were compared using Kaplan–Meier estimates and a log-rank test (Fig. 3B). The OCS and PFS of patients in the cohort were also calculated for other known prognostic factors (FIGO stage I vs. II/III/IV, <2 cm vs. >2 cm residual disease after debulking surgery and age) and the different histopathological characteristics using the same method (Table 2). A Cox regression model was fitted to predict for OCS and PFS by including prognostically significant characteristics for OCS and PFS in univariate analysis (Table 2). The resulting hazard ratios were calculated with a 95% CI. All statistical analyses and graphs were performed using SPSS version15 software.

Figure 1.
Figure 1.

Representative genome plots of ovarian clear cell carcinomas illustrating the “complex-firestorm”, “complex-sawtooth”, and “simplex” genomic patterns. The aws log2 ratio of each BAC clone is plotted on the y-axis and according to its genomic location on the x-axis. BAC clones that are gained (aws log2 ratio > 0.12) or amplified (aws log2 ratio > 0.4) are represented as green dots, clones with no change in copy number (aws log2 ratio = −0.12 to 0.12) are represented by black dots and clones which are lost (aws log2 ratio < −0.12) are represented as red dots. A, complex “firestorm” genomic patterns; B, complex “sawtooth” genomic patterns; and C, “simplex” genomic patterns (see text for details).

Figure 2.
Figure 2.

Frequency of copy number gains, losses, and amplifications in ovarian clear cell carcinomas. A, frequency of copy number changes observed in 50 OCCCs. The proportion of tumors in which each clone is gained (log2 ratio > 0.12, green bars) or lost (log2 ratio < −0.12, red bars) is plotted on y-axis for each BAC clone according to genomic location on x-axis. The vertical dotted line represents the chromosome centromere. B, frequency of amplifications (log2 ratio > 0.4, green bars) in OCCCs. Recurrent 17q12 amplification in OCCCs marked with asterix. C, CISH validation of 17q12 (HER2) amplification showing a HER2 nonamplified and HER2-negative OCCC, and a HER2-amplified OCCC with HER2 3+ overexpression (CISH images shown at ×630 magnification and immunohistochemistry images shown at ×200 magnification).

Figure 3.
Figure 3.

Unsupervised hierarchical clustering analysis based on copy number aberrations of ovarian clear cell carcinomas (n = 50) reveals distinct genomic subgroups associated with outcome. A, cluster dendrogram derived from unsupervised (Ward's cluster) analysis of DNA copy number changes (i.e., categorical data obtained from aCGH analysis) reveals 2 distinct OCCC subgroups with a heatmap showing copy number changes across all chromosomes in cluster-1 and cluster-2 OCCCs with corresponding clinical characteristics and genomic patterns for each OCCC included. B, Kaplan–Meier survival plots for OCCC specific survival and PFS of cluster-1 versus cluster-2 tumors. P values are calculated using the log-rank test.

View this table:
Table 1.

Clinicopathological features and genomic patterns in cluster 1 and 2 tumors

View this table:
Table 2.

Survival analysis for ovarian clear cell carcinoma patients (n = 50)

CISH, FISH, and immunohistochemical analysis

Genomic loci found to be recurrently amplified in OCCCs (17q12 encompassing the HER2 gene) and also to be significantly more frequently amplified in cluster-1 than cluster-2 tumors (17q21.2–21.32, 19q13.2, 20q13.13, and 20q13.31-q13.32) were analyzed on whole tissue sections from selected tumors from this cohort by CISH and FISH.

For HER2 copy number assessment, FDA-approved ready-to-use digoxigenin-labeled Spotlight amplification probe for HER2 (Invitrogen) was employed. In 1 of 7 HER2-amplified OCCCs there was insufficient tissue for CISH and immunohistochemical analysis. Therefore, CISH was only performed on 6 of 7 HER2-amplified OCCCs.

CISH probes using 4 BACs (RP11–812O05, CTD-2234O23, RP11–1365J18, RP11–769M09) mapping to a smallest region of amplification (SRA) on 17q21.2–21.32, and FISH probes using BACS mapping to SRAs on 19q13.2 (RP11–688J23 RP11–36B02, RP11–691F18), 20q13.13 (RP11–223M11, RP11–444K09), and 20q13.31-q13.32 (RP4–539E4, RP5–843L14) were generated as previously described (24). Due to limited tissue availability, we performed FISH to validate the presence of amplification in 4 OCCC samples for each amplicon and used 3 negative controls (i.e., samples reported not to have amplification of the region by microarray-based CGH analysis) for each probe. Analysis of CISH hybridizations was performed at ×400 and ×630 magnification on a multiheaded microscope by 2 of the authors (DT, FCG), blinded to the results of aCGH analysis. FISH probes were visualized using a Zeiss Axioplan 2 microscope equipped with a CCD camera, and analyzed with Cytovision software version 2.81 (Applied Imaging International) by 2 of the authors (FM, DT). Signals for CISH and FISH were counted in nonoverlapping nuclei of at least 60 morphologically unequivocal neoplastic cells. Amplification was defined as more than 5 signals per nucleus in more than 50% of cancer cells, or when more than 50% of cells harbored large gene copy clusters as previously described (15, 16, 25, 26).

Immunohistochemistry for HER2 expression was performed using the HercepTest (code A0485; Dako) and scored according to ASCO/CAP guidelines (27).

Results

Genomic architecture patterns of ovarian clear cell carcinomas

The clinical details of all 50 patients are summarized in Table 1. All patients were diagnosed between October 1988 and March 2007 with a median follow-up period of 24 months (range 1.2 to 159.6 months).

High-resolution (approximately 50 kb) aCGH analysis of 50 primary OCCCs revealed the presence of 3 distinct genomic patterns (Fig. 1): one characterized by few chromosomal aberrations involving whole chromosomes or chromosomal arms (‘simplex’), one characterized by multiple gains and losses throughout the genome (‘sawtooth’), with very few high-level amplifications, and a third pattern characterized by multiple, clustered amplifications mapping to a single chromosome or chromosomal arm (‘firestorm’). These patterns were similar to those described by Hicks and colleagues (21) in breast cancer. Out of the 50 OCCCs profiled, 19 (38%), 11 (22%), and 20 (40%) tumors displayed simplex, complex-sawtooth, and complex-firestorm genomic profiles, respectively. These findings were similar to those derived from our previous aCGH analysis of OCCC cell lines (33% simplex, 17% complex-sawtooth, and 42% complex-firestorm; ref. 11).

Genomic copy number alterations in ovarian clear cell carcinomas

Having established the genomic profiles of the OCCCs, a detailed analysis of the aCGH-derived copy number changes was performed. Figure 2 shows the frequency plots of copy number gains, losses, and amplifications in the OCCCs analyzed. A summary of recurrent regions (at least 5/50 tumors, i.e., ≥10%) of gains (Log2 ratio > 0.12) and losses (Log2 ratio < −0.12) and recurrent regions (at least 2/50 tumors, i.e., ≥4%) of high-level gain/ amplification (Log2 ratios > 0.4) and known copy number variations (CNV) for all 50 OCCCs is included in Supplementary Table 1.

We (16, 28) and others (12) have previously demonstrated that genes mapping to recurrently amplified regions in OCCCs may function as amplicon drivers and hence potential therapeutic targets (Supplementary Table 1). Regions with a high frequency of recurrent amplification included 20q13.2 (harboring ZNF217, which has previously been reported as being amplified in 36% of OCCCs; ref. 12) in 11 of 50 (22%) OCCCs, and 17q12-q21.32 which was amplified in 9 of 50 (18%) OCCCs and harbors the HER2, TOP2A, GAST, JUP, and BRCA1 genes (Supplementary Tables 1 and 2). We previously showed that PPM1D is amplified and overexpressed in 10% of OCCCs and is a potential therapeutic target in OCCCs (11). In this study, PPM1D was amplified in 3 of 50 (6%) OCCCs. Genes within regions of recurrent amplification are listed in Supplementary Table 2.

To validate the results of our aCGH data, we performed CISH using probes mapping to the HER2 gene, which was amplified in 7 of 50 OCCCs. A perfect agreement between CISH and aCGH results was found (Fig. 2, Supplementary Table 3), validating the presence of amplification of these regions in a subset of OCCCs. Immunohistochemical analysis of HER2 expression also revealed a perfect correlation between HER2 amplification and overexpression in our OCCCs (Supplementary Table 3). No association between HER2 amplification and clinical outcome (PFS and OCS) was observed.

Hierarchical clustering analysis reveals 2 genomic subgroups of OCCCs that are associated with distinct clinical outcomes

Given the heterogeneity in terms of patterns of genomic aberrations observed in OCCCs, we carried out a hypothesis generating unsupervised hierarchical clustering class discovery analysis based on categorical aCGH data from the 50 OCCCs. This analysis revealed that OCCCs were subdivided into 2 distinct clusters (cluster-1 OCCCs and cluster-2 OCCCs) based on their copy number changes (Fig. 3).

We next determined the associations with clinico-pathological, histological, and molecular features of these 2 clusters defined by the patterns of copy number aberrations. There were no significant differences between tumors from cluster-1 and cluster-2 in terms of their histological characteristics, FIGO stage, or residual disease following primary debulking surgery (Table 1). Importantly, however, cluster-1 was significantly enriched for tumors with complex-firestorm and complex-sawtooth genomic patterns, whereas cluster-2 tumors displayed a significantly higher frequency of the tumors with a simplex genomic pattern (P < 0.001; Table 1 and Fig. 3).

To determine if the 2 observed clusters were associated with the outcome of patients with OCCCs, we compared the PFS and OCS outcomes across the cohort of 50 OCCC patients. The Kaplan–Meier survival estimates (Fig. 3) showed that cluster-1 patients (median survival 11 months, 95% CI: 3 to 19) had a significantly shorter PFS than cluster-2 patients (median survival 65 months, 95% CI 22 to 108, P = 0.009) following first-line treatment. The Kaplan–Meier survival estimates for median OCS (Fig. 3) also revealed a trend for a poorer prognosis for cluster-1 patients (19 months, 95% CI: 0 to 69) compared with cluster-2 patients (66 months, 95% CI: 45 to 87, P = 0.065).

Univariate Kaplan–Meier (log-rank test) analysis determined that out of all clinicopathological parameters, only FIGO stage (P = 0.001), extent of debulking surgery (P < 0.001) and genomic cluster (P = 0.009) were significantly associated with PFS and only stage (P = 0.001) and extent of debulking surgery (P < 0.001) were significantly associated with OCS (Table 2 and Table 3). The genomic architectural patterns (complex vs. simplex) of OCCCs were not associated with PFS or OCS (Table 2).

Multivariate analysis using a Cox regression model including all variables that were significant at univariate analysis (i.e., FIGO stage, residual disease, and cluster) was performed to identify independent prognostic factors for our cohort of OCCC patients. Genomic cluster and suboptimal debulking surgery (residual disease > 2 cm) were found to be independent predictors of PFS (Table 2). As expected, our analysis confirmed that advanced FIGO stage and suboptimal debulking surgery were independent prognostic factors for poorer OCS (Table 2).

Genomic differences between cluster-1 and cluster-2 OCCCs

Given that cluster-1 and cluster-2 tumors did not differ in terms of their clinicopathological and histological features (Table 1), we sought to define the genomic differences between these molecular subgroups of OCCCs (Fig. 4). Multi-Fisher's exact test corrected for false discovery rate using the step-down permutation procedure maxT (15, 16) revealed that 26 recurrent regions of high-level gain/ amplification as being significantly more prevalent (adjusted P < 0.05) in cluster-1 (i.e., the poor outcome) OCCCs. These included 8q22.1 (encompassing CCNE2), 17q21.2–q21.32 (encompassing GAST, JUP, BRCA1), 17q21.33 (encompassing ABCC3, TOB1), 19q13.2 (encompassing AKT2), 19q13.31 (encompassing BCL3, ERCC1, FOXA3), 20q11.23 (encompassing SRC), and 20q13.13 (encompassing SNAI1; Fig. 4 and Supplementary Table 4 for full list of regions and genes).

Figure 4.
Figure 4.

Genomic differences between cluster 1 and cluster 2 ovarian clear cell carcinomas. A, frequency plots of chromosomal gains and losses in cluster-1 (n = 23) and cluster-2 (n = 27) OCCCs. B, frequency plots of chromosomal high-level gains/ amplifications in cluster-1 (n = 23) and cluster-2 (n = 27) OCCCs. The proportion of tumors in which each clone is gained/ amplified (green bars) or lost (red bars) is plotted on y-axis for each BAC clone according to genomic location on x-axis. Vertical dotted lines represent chromosome centromeres. Results of Fisher's exact test performed with the segmented values for each clone are shown below the frequency plots of both OCCC clusters in A and B. Clones with an adjusted P value of less than 0.05 are plotted on y-axis (inverse Log10 of P value) according to genomic location (x-axis). C, CISH validation of 17q21 amplification showing a cluster-2 OCCC without 17q21 amplification and a cluster-1 OCCC with 17q21 amplification (All CISH images shown at ×600 magnification).

We performed CISH (Fig. 4 and Supplementary Table 3) using probes mapping to the smallest region of amplification on 17q21.2–q21.32 which confirmed this region was amplified at a significantly higher prevalence (Multi-Fisher's adjusted P < 0.05) in the poor prognosis cluster-1 tumors (6/23) compared with cluster-2 OCCCs (0/27; Supplementary Table 3). Further validation of 3 additional amplicons, 19q13.2 (encompassing AKT2), 20q13.13 (encompassing SNAI1), and 20q13.31-q13.32 (encompassing TFAP2C) that were also amplified at a significantly higher prevalence (Multi-Fisher's adjusted P < 0.05) in the poor prognosis cluster-1 tumors, was also performed using FISH (Supplementary Figure 1 and Supplementary Table 3). aCGH and FISH results displayed 100% agreement in the cases tested.

Discussion

In this study, we demonstrate that OCCCs constitute a heterogeneous group of cancers in terms of their patterns of copy number aberrations and identified several regions of recurrent amplification in our OCCCs that have previously been described in OCCCs and OCCC cell lines (11, 12). In particular, recurrent amplification of the 20q13 locus harboring ZNF217 and the 17q23.2 amplicon encompassing PPM1D were found in 20% and 6% of the OCCCs studied, respectively (Supplementary Table 2). In addition, our aCGH analysis also revealed recurrent amplifications of the HER2 gene in 14% of cases, which were subsequently validated by CISH. Importantly, all OCCCs harboring HER2 gene amplification displayed HER2 protein overexpression as defined by immunohistochemical analysis.

In view of the previous evidence from experimental models demonstrating that genes consistently overexpressed when amplified may constitute potential therapeutic targets (11, 16, 28) and the positive results from clinical trials carried out in patients with breast and gastric cancers displaying HER2 gene amplification and protein expression (29–31), our data suggest that HER2 may also be a potential therapeutic target in OCCCs. In a previous phase II study of trastuzumab monotherapy (32) in 41 patients with recurrent EOCs (including 7 OCCCs), an overall response rate of only 7% was observed. However, patient selection was only based on HER2 2+ (27/41) or 3+ (14/41) immunohistochemical expression levels without assessment of HER2 copy number status (32). Therefore, additional studies to assess the predictive value of HER2 amplification and overexpression in OCCCs in response to trastuzumab are warranted.

In agreement with previous large retrospective studies on OCCCs (3, 6, 33), we confirmed that (i) histopathological features cannot identify prognostic subgroups of OCCCs, (ii) advanced FIGO stage and suboptimal debulking (>2 cm residual disease) are independent predictors of OCS, and (iii) suboptimal debulking is an independent predictor of PFS in OCCCs.

OCCCs are considered high grade (Grade III) tumors by definition, given their relatively poor outcome and the fact that histopathological features cannot distinguish prognostic subgroups in these cancers. Here we demonstrate that OCCCs are heterogeneous at the genomic level. Unsupervised clustering analysis resulted in the identification of 2 subgroups of OCCCs, which are characterized by distinct patterns of genomic aberrations and clinical outcomes. Importantly, cluster-1 (i.e., poor outcome) OCCCs have more complex genomes, including recurrent amplifications that may be predictive of poorer outcomes and possibly mediate tumor aggressiveness and treatment resistance.

A hypothesis generating analysis of the associations between the genomic groups identified by hierarchical clustering revealed that they are associated with PFS, independent of FIGO stage and residual disease following primary cytoreductive surgery. These results suggest that the genomic copy number changes in primary, pure OCCCs may have a direct influence on tumor behavior and clinical outcome, which may be independent of FIGO stage, residual disease. There were significant differences in the genomic architectural profiles of OCCCs between the genomic clusters, with a higher prevalence of complex-sawtooth and firestorm patterns being observed in cluster-1 tumors and simplex genomic patterns being predominantly observed in cluster-2 tumors (Table 1), however these genomic patterns were not predictive of PFS or OCS (Table 2).

Although the histopathological features of cluster-1 and cluster-2 tumors did not differ, we identified 26 genomic regions (Supplementary Table 4) with a significantly higher prevalence of high-level gain/ amplification in cluster-1 compared with cluster-2 OCCCs. Our aCGH results were supported by CISH and FISH, which confirmed the amplification of selected regions (Supplementary Table 3 and Supplementary Figure 1). These amplicons encompass a number of genes, including CCNE2 on 8q22.1 (34), JUP (35), GAST (36) and BRCA1 (37) on 17q21.2–q21.32, ERCC1 on 19q13.31 (38), AKT2 on 19q13.2 (39, 40), SRC on 20q11.23 (41); and SNAI1 on 20q13.13 (42) that, if overexpressed, may account for the observed prognostic differences between the 2 genomic OCCC clusters (Supplementary Table 4). In particular, the shorter PFS observed in cluster-1 suggests that these amplicons may harbor potential mediators of tumor aggressiveness and chemoresistance, and therefore novel therapeutic targets for further assessment in OCCCs.

We performed CISH or FISH using probes mapping to SRAs that were significantly more frequently amplified in cluster-1 than in cluster-2 OCCCs, including 19q13.2, 20q13.13, and 20q13.31–q13.32 and 17q21.2–q21.32. These analyses validated the presence of these amplifications in cases from cluster-1. Of particular interest is the 17q21.2–q21.32 region which contains a number of putative oncogenes including EIF1 (43), JUP (35), GAST (36), and FKBP10 (44) but, in the context of EOC, the tumor suppressor gene BRCA1, which encodes a nuclear phosphoprotein involved in transcriptional regulation and DNA repair (37), is of particular interest. Mutations and downregulation of this gene are associated with the hereditary breast and ovarian cancers (37) and increased sensitivity to platinum-based chemotherapy (45, 46). Conversely, higher BRCA1 mRNA expression has been associated with platinum-resistance and poorer outcomes in EOC (45), and transcript variants of BRCA1 have been shown to increase cell proliferation by upregulating cyclin-D1 expression (47). Hence, one could hypothesize that amplification and overexpression of genes within 17q21.2–q21.32 may confer a survival advantage to OCCC tumor cells and worse patient outcome.

The 50 tumors profiled in this study represent the largest high-resolution aCGH analysis of OCCCs to date. However, given the retrospective accrual of the cohort analyzed and its relatively small size, our results are best regarded as hypothesis generating and will require independent validation in a larger set of OCCCs. In fact, a trend for cluster-1 tumors to have a shorter OCS than cluster-2 OCCCs was observed, but this did not reach statistical significance (P = 0.065). It cannot be ruled out that the lack of association between genomic clusters and OCS may be due to the limited sample size. Nevertheless, our results do suggest that OCCCs are genetically heterogeneous and that the level of genetic complexity of OCCCs (i.e., the equivalent of a ‘genomic grade’) may predict the outcomes of patients with OCCCs. Besides, we have also identified 26 regions of recurrent high level gain/amplification which may harbor genes that are predictive of poorer outcome and potential therapeutic targets in OCCCs. Further analyses to investigate the ‘drivers’ of these recurrent amplicons are warranted.

Disclosure of Potential Conflict of Interests

No potential conflicts of interest were disclosed.

Acknowledgments

This study was funded by Cancer Research UK, Breakthrough Breast Cancer and National Health Service funding to the National Institute of Health Research Biomedical Research Centre. J.S. Reis-Filho is the recipient of the 2010 CRUK Future Leaders Prize.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Footnotes

  • Note: Supplementary data for this article are available at Clinical Cancer Research Online (http://clincancerres.aacrjournals.org/).

  • Received June 25, 2010.
  • Revision received November 17, 2010.
  • Accepted November 30, 2010.

References

  1. 1.↵
    1. du Bois A,
    2. Luck HJ,
    3. Meier W,
    4. Adams HP,
    5. Möbus V,
    6. Costa S,
    7. et al.
    A randomized clinical trial of cisplatin/paclitaxel versus carboplatin/paclitaxel as first-line treatment of ovarian cancer. J Natl Cancer Inst 2003;95:1320–9.
    OpenUrlAbstract/FREE Full Text
  2. 2.↵
    1. Piccart MJ,
    2. Bertelsen K,
    3. James K,
    4. Cassidy J,
    5. Mangioni C,
    6. Simonsen E,
    7. et al.
    Randomized intergroup trial of cisplatin-paclitaxel versus cisplatin-cyclophosphamide in women with advanced epithelial ovarian cancer: three-year results. J Natl Cancer Inst 2000;92:699–708.
    OpenUrlAbstract/FREE Full Text
  3. 3.↵
    1. Takano M,
    2. Kikuchi Y,
    3. Yaegashi N,
    4. Kuzuya K,
    5. Ueki M,
    6. Tsuda H,
    7. et al.
    Clear cell carcinoma of the ovary: a retrospective multicentre experience of 254 patients with complete surgical staging. Br J Cancer 2006;94:1369–74.
    OpenUrlCrossRefPubMed
  4. 4.↵
    1. Tan DS,
    2. Kaye S
    . Ovarian clear cell adenocarcinoma: a continuing enigma. J Clin Pathol 2007;60:355–60.
    OpenUrlAbstract/FREE Full Text
  5. 5.↵
    1. Winter WE 3rd,
    2. Maxwell GL,
    3. Tian C,
    4. Carlson JW,
    5. Ozols RF,
    6. Rose PG,
    7. et al.
    Prognostic factors for stage III epithelial ovarian cancer: a Gynecologic Oncology Group Study. J Clin Oncol 2007;25:3621–7.
    OpenUrlAbstract/FREE Full Text
  6. 6.↵
    1. Chan JK,
    2. Teoh D,
    3. Hu JM,
    4. Shin JY,
    5. Osann K,
    6. Kapp DS
    . Do clear cell ovarian carcinomas have poorer prognosis compared to other epithelial cell types? A study of 1411 clear cell ovarian cancers. Gynecol Oncol 2008;109:370–6.
    OpenUrlCrossRefPubMed
  7. 7.↵
    1. Takano M,
    2. Sugiyama T,
    3. Yaegashi N,
    4. Suzuki M,
    5. Tsuda H,
    6. Sagae S,
    7. et al.
    Progression-free survival and overall survival of patients with clear cell carcinoma of the ovary treated with paclitaxel-carboplatin or irinotecan-cisplatin: retrospective analysis. Int J Clin Oncol 2007;12:256–60.
    OpenUrlCrossRefPubMed
  8. 8.↵
    1. McCluggage WG
    . My approach to and thoughts on the typing of ovarian carcinomas. J Clin Pathol 2008;61:152–163.
    OpenUrlAbstract/FREE Full Text
  9. 9.↵
    1. Tan DS,
    2. Lambros MB,
    3. Natrajan R,
    4. Reis-Filho JS
    . Getting it right: designing microarray (and not ‘microawry’) comparative genomic hybridization studies for cancer research. Lab Invest 2007;87:737–54.
    OpenUrlCrossRefPubMed
  10. 10.↵
    1. Hirasawa A,
    2. Saito-Ohara F,
    3. Inoue J,
    4. Aoki D,
    5. Susumu N,
    6. Yokoyama T,
    7. et al.
    Association of 17q21-q24 gain in ovarian clear cell adenocarcinomas with poor prognosis and identification of PPM1D and APPBP2 as likely amplification targets. Clin Cancer Res 2003;9:1995–2004.
    OpenUrlAbstract/FREE Full Text
  11. 11.↵
    1. Tan DS,
    2. Lambros MB,
    3. Rayter S,
    4. Natrajan R,
    5. Vatcheva R,
    6. Gao Q,
    7. et al.
    PPM1D is a potential therapeutic target in ovarian clear cell carcinomas. Clin Cancer Res 2009;15:2269–80.
    OpenUrlAbstract/FREE Full Text
  12. 12.↵
    1. Kuo KT,
    2. Mao TL,
    3. Chen X,
    4. Feng Y,
    5. Nakayama K,
    6. Wang Y,
    7. et al.
    DNA Copy Numbers Profiles in Affinity-Purified Ovarian Clear Cell Carcinoma. Clin Cancer Res 2010;16:1997–2008.
    OpenUrlAbstract/FREE Full Text
  13. 13.↵
    1. Tsuda H,
    2. Ito YM,
    3. Ohashi Y,
    4. Wong KK,
    5. Hashiguchi Y,
    6. Welch WR,
    7. et al.
    Identification of overexpression and amplification of ABCF2 in clear cell ovarian adenocarcinomas by cDNA microarray analyses. Clin Cancer Res 2005;11:6880–8.
    OpenUrlAbstract/FREE Full Text
  14. 14.↵
    1. Geyer FC,
    2. Weigelt B,
    3. Natrajan R,
    4. Lambros MB,
    5. de Biase D,
    6. Vatcheva R,
    7. et al.
    Molecular analysis reveals a genetic basis for the phenotypic diversity of metaplastic breast carcinomas. J Pathol 2010;220:562–73.
    OpenUrlCrossRefPubMed
  15. 15.↵
    1. Marchio C,
    2. Natrajan R,
    3. Shiu KK,
    4. Lambros MB,
    5. Rodriguez-Pinilla SM,
    6. Tan DS,
    7. et al.
    The genomic profile of HER2-amplified breast cancers: the influence of ER status. J Pathol 2008;216:399–407.
    OpenUrlCrossRefPubMed
  16. 16.↵
    1. Natrajan R,
    2. Lambros MB,
    3. Rodriguez-Pinilla SM,
    4. Moreno-Bueno G,
    5. Tan DS,
    6. Marchió C,
    7. et al.
    Tiling path genomic profiling of grade 3 invasive ductal breast cancers. Clin Cancer Res 2009;15:2711–22.
    OpenUrlAbstract/FREE Full Text
  17. 17.↵
    1. Reis-Filho JS,
    2. Drury S,
    3. Lambros MB,
    4. Marchio C,
    5. Johnson N,
    6. Natrajan R,
    7. et al.
    ESR1 gene amplification in breast cancer: a common phenomenon? Nat Genet 2008;40:809–810; author reply810–802.
    OpenUrlPubMed
  18. 18.↵
    1. Coe BP,
    2. Ylstra B,
    3. Carvalho B,
    4. Meijer GA,
    5. Macaulay C,
    6. Lam WL
    . Resolving the resolution of array CGH. Genomics 2007;89:647–53.
    OpenUrlCrossRefPubMed
  19. 19.↵
    1. Gunnarsson R,
    2. Staaf J,
    3. Jansson M,
    4. Ottesen AM,
    5. Göransson H,
    6. Liljedahl U,
    7. et al.
    Screening for copy-number alterations and loss of heterozygosity in chronic lymphocytic leukemia–a comparative study of four differently designed, high resolution microarray platforms. Genes Chromosomes Cancer 2008;47:697–711.
    OpenUrlCrossRefPubMed
  20. 20.↵
    1. Hupe P,
    2. Stransky N,
    3. Thiery JP,
    4. Radvanyi F,
    5. Barillot E
    . Analysis of array CGH data: from signal ratio to gain and loss of DNA regions. Bioinformatics 2004;20:3413–22.
    OpenUrlAbstract/FREE Full Text
  21. 21.↵
    1. Hicks J,
    2. Krasnitz A,
    3. Lakshmi B,
    4. Navin NE,
    5. Riggs M,
    6. Leibu E,
    7. et al.
    Novel patterns of genome rearrangement and their association with survival in breast cancer. Genome Res 2006;16:1465–79.
    OpenUrlAbstract/FREE Full Text
  22. 22.↵
    1. Meunier B,
    2. Dumas E,
    3. Piec I,
    4. Bechet D,
    5. Hebraud M,
    6. Hocquette JF
    . Assessment of hierarchical clustering methodologies for proteomic data mining. J Proteome Res 2007;6:358–66.
    OpenUrlCrossRefPubMed
  23. 23.↵
    1. Marchio C,
    2. Iravani M,
    3. Natrajan R,
    4. Lambros MB,
    5. Geyer FC,
    6. Savage K,
    7. et al.
    Mixed micropapillary-ductal carcinomas of the breast: a genomic and immunohistochemical analysis of morphologically distinct components. J Pathol 2009;218:301–15.
    OpenUrlCrossRefPubMed
  24. 24.↵
    1. Lambros MB,
    2. Simpson PT,
    3. Jones C,
    4. Natrajan R,
    5. Westbury C,
    6. Steele D,
    7. et al.
    Unlocking pathology archives for molecular genetic studies: a reliable method to generate probes for chromogenic and fluorescent in situ hybridization. Lab Invest 2006;86:398–408.
    OpenUrlCrossRefPubMed
  25. 25.↵
    1. Reis-Filho JS,
    2. Pinheiro C,
    3. Lambros MB,
    4. Milanezi F,
    5. Carvalho S,
    6. Savage K,
    7. et al.
    EGFR amplification and lack of activating mutations in metaplastic breast carcinomas. J Pathol 2006;209:445–53.
    OpenUrlCrossRefPubMed
  26. 26.↵
    1. Natrajan R,
    2. Weigelt B,
    3. Mackay A,
    4. Geyer FC,
    5. Grigoriadis A,
    6. Tan DS,
    7. et al.
    An integrative genomic and transcriptomic analysis reveals molecular pathways and networks regulated by copy number aberrations in basal-like, HER2 and luminal cancers. Breast Cancer Res Treat 2010;121:575–89.
    OpenUrlCrossRefPubMed
  27. 27.↵
    1. Wolff AC,
    2. Hammond ME,
    3. Schwartz JN,
    4. Hagerty KL,
    5. Allred DC,
    6. Cote RJ,
    7. et al.
    American Society of Clinical Oncology/College of American Pathologists guideline recommendations for human epidermal growth factor receptor 2 testing in breast cancer. J Clin Oncol 2007;25:118–45.
    OpenUrlAbstract/FREE Full Text
  28. 28.↵
    1. Turner N,
    2. Pearson A,
    3. Sharpe R,
    4. Lambros M,
    5. Geyer F,
    6. Lopez-Garcia MA,
    7. et al.
    FGFR1 amplification drives endocrine therapy resistance and is a therapeutic target in breast cancer. Cancer Res 2010;70:2085–94.
    OpenUrlAbstract/FREE Full Text
  29. 29.↵
    1. Piccart-Gebhart MJ,
    2. Procter M,
    3. Leyland-Jones B,
    4. Goldhirsch A,
    5. Untch M,
    6. Smith I,
    7. et al.
    Trastuzumab after adjuvant chemotherapy in HER2-positive breast cancer. N Engl J Med 2005;353:1659–72.
    OpenUrlCrossRefPubMed
  30. 30.↵
    1. Romond EH,
    2. Perez EA,
    3. Bryant J,
    4. Suman VJ,
    5. Geyer CE Jr,
    6. Davidson NE,
    7. et al.
    Trastuzumab plus adjuvant chemotherapy for operable HER2-positive breast cancer. N Engl J Med 2005;353:1673–84.
    OpenUrlCrossRefPubMed
  31. 31.↵
    1. Okines AF,
    2. Cunningham D
    . Trastuzumab in gastric cancer. Eur J Cancer 2010;46:1949–59.
    OpenUrlCrossRefPubMed
  32. 32.↵
    1. Bookman MA,
    2. Darcy KM,
    3. Clarke-Pearson D,
    4. Boothby RA,
    5. Horowitz IR
    . Evaluation of monoclonal humanized anti-HER2 antibody, trastuzumab, in patients with recurrent or refractory ovarian or primary peritoneal carcinoma with overexpression of HER2: a phase II trial of the Gynecologic Oncology Group. J Clin Oncol 2003;21:283–90.
    OpenUrlAbstract/FREE Full Text
  33. 33.↵
    1. Takano M,
    2. Kita T,
    3. Kikuchi Y,
    4. Yaegashi N,
    5. Kuzuya K,
    6. Tsuda H,
    7. et al.
    Clinical characteristics of clear cell adenocarcinoma of the ovary -Japan Clear Cell Carcinoma Study Group. Journal of Clinical Oncology, 2005 ASCO Annual Meeting Proceedings. 2005;23 No. 16S, Part 1 of II (Jun1 supplement): 5123.
  34. 34.↵
    1. Caldon CE,
    2. Musgrove EA
    . Distinct and redundant functions of cyclin E1 and cyclin E2 in development and cancer. Cell Div 5:2.
  35. 35.↵
    1. Pan H,
    2. Gao F,
    3. Papageorgis P,
    4. Abdolmaleky HM,
    5. Faller DV,
    6. Thiagalingam S
    . Aberrant activation of gamma-catenin promotes genomic instability and oncogenic effects during tumor progression. Cancer Biol Ther 2007;6:1638–43.
    OpenUrlCrossRefPubMed
  36. 36.↵
    1. Ferrand A,
    2. Wang TC
    . Gastrin and cancer: a review. Cancer Lett 2006;238:15–29.
    OpenUrlCrossRefPubMed
  37. 37.↵
    1. Tan DS,
    2. Marchio C,
    3. Reis-Filho JS
    . Hereditary breast cancer: from molecular pathology to tailored therapies. J Clin Pathol 2008;61:1073–82.
    OpenUrlAbstract/FREE Full Text
  38. 38.↵
    1. Vilmar A,
    2. Sorensen JB
    . Excision repair cross-complementation group 1 (ERCC1) in platinum-based treatment of non-small cell lung cancer with special emphasis on carboplatin: a review of current literature. Lung Cancer 2009;64:131–9.
    OpenUrlCrossRefPubMed
  39. 39.↵
    1. Weng D,
    2. Song X,
    3. Xing H,
    4. Ma X,
    5. Xia X,
    6. Weng Y,
    7. et al.
    Implication of the Akt2/survivin pathway as a critical target in paclitaxel treatment in human ovarian cancer cells. Cancer Lett 2009;273:257–65.
    OpenUrlCrossRefPubMed
  40. 40.↵
    1. Xing H,
    2. Weng D,
    3. Chen G,
    4. Tao W,
    5. Zhu T,
    6. Yang X,
    7. et al.
    Activation of fibronectin/PI-3K/Akt2 leads to chemoresistance to docetaxel by regulating survivin protein expression in ovarian and breast cancer cells. Cancer Lett 2008;261:108–19.
    OpenUrlCrossRefPubMed
  41. 41.↵
    1. Pengetnze Y,
    2. Steed M,
    3. Roby KF,
    4. Terranova PF,
    5. Taylor CC
    . Src tyrosine kinase promotes survival and resistance to chemotherapeutics in a mouse ovarian cancer cell line. Biochem Biophys Res Commun 2003;309:377–83.
    OpenUrlCrossRefPubMed
  42. 42.↵
    1. Blechschmidt K,
    2. Sassen S,
    3. Schmalfeldt B,
    4. Schuster T,
    5. Hofler H,
    6. Becker KF
    . The E-cadherin repressor Snail is associated with lower overall survival of ovarian cancer patients. Br J Cancer 2008;98:489–95.
    OpenUrlCrossRefPubMed
  43. 43.↵
    1. van Agthoven T,
    2. Sieuwerts AM,
    3. Meijer D,
    4. Meijer-van Gelder ME,
    5. van Agthoven TL,
    6. Sarwari R,
    7. et al.
    Selective recruitment of breast cancer anti-estrogen resistance genes and relevance for breast cancer progression and tamoxifen therapy response. Endocr Relat Cancer 2010;17:215–30.
    OpenUrlAbstract/FREE Full Text
  44. 44.↵
    1. Coss MC,
    2. Stephens RM,
    3. Morrison DK,
    4. Winterstein D,
    5. Smith LM,
    6. Simek SL
    . The immunophilin FKBP65 forms an association with the serine/threonine kinase c-Raf-1. Cell Growth Differ 1998;9:41–8.
    OpenUrlAbstract
  45. 45.↵
    1. Quinn JE,
    2. James CR,
    3. Stewart GE,
    4. Mulligan JM,
    5. White P,
    6. Chang GK,
    7. et al.
    BRCA1 mRNA Expression Levels Predict for Overall Survival in Ovarian Cancer after Chemotherapy. Clin Cancer Res 2007;13:7413–20.
    OpenUrlAbstract/FREE Full Text
  46. 46.↵
    1. Tan DS,
    2. Rothermundt C,
    3. Thomas K,
    4. Bancroft E,
    5. Eeles R,
    6. Shanley S,
    7. et al.
    “BRCAness” syndrome in ovarian cancer: a case-control study describing the clinical features and outcome of patients with epithelial ovarian cancer associated with BRCA1 and BRCA2 mutations. J Clin Oncol 2008;26:5530–6.
    OpenUrlAbstract/FREE Full Text
  47. 47.↵
    1. ElShamy WM,
    2. Livingston DM
    . Identification of BRCA1-IRIS, a BRCA1 locus product. Nat Cell Biol 2004;6:954–67.
    OpenUrlCrossRefPubMed
Back to top
Clinical Cancer Research: 17 (6)
March 2011
Volume 17, Issue 6

Sign up for alerts

View this article with LENS

Open full page PDF
Article Alerts
Email Article
Citation Tools
Genomic Analysis Reveals the Molecular Heterogeneity of Ovarian Clear Cell Carcinomas
David S.P. Tan, Marjan Iravani, W. Glenn McCluggage, Maryou B.K. Lambros, Fernanda Milanezi, Alan Mackay, Charles Gourley, Felipe C. Geyer, Radost Vatcheva, Joanne Millar, Karen Thomas, Rachael Natrajan, Kay Savage, Kerry Fenwick, Alistair Williams, Charles Jameson, Mona El-Bahrawy, Martin E. Gore, Hani Gabra, Stanley B. Kaye, Alan Ashworth and Jorge S. Reis-Filho
Clin Cancer Res March 15 2011 (17) (6) 1521-1534; DOI: 10.1158/1078-0432.CCR-10-1688
del.icio.us logo Digg logo Reddit logo Twitter logo CiteULike logo Facebook logo Google logo Mendeley logo

Jump to section

Advertisement