Characterization of the 19q12 amplification including CCNE1 and URI in different epithelial ovarian cancer subtypes

Experimental and Molecular Pathology 98 (2015) 47–54 Contents lists available at ScienceDirect Experimental and Molecular Pathology j ourna l homepage: www.e lsev ie r .com/ locate /yexmp Characterization of the 19q12 amplification including CCNE1 and URI in different epithelial ovarian cancer subtypes Aurelia Noske a,⁎, Leigh A. Henricksen b, Bonnie LaFleur b, Anne-Katrin Zimmermann a, Alisa Tubbs b, Shalini Singh b, Martina Storz a, Daniel Fink c, Holger Moch a a Institute of Surgical Pathology, University Hospital Zurich, Schmelzbergstr. 12, CH-8091 Zurich, Switzerland b Ventana Medical Systems, Inc., 1910 East Innovation Park Drive, Tucson, AZ 85755, USA c Department of Gynecology, University Hospital Zurich, Frauenklinikstr. 10, 8091 Zurich, Switzerland ⁎ Corresponding author. E-mail address: [email protected] (A. Noske). http://dx.doi.org/10.1016/j.yexmp.2014.12.004 0014-4800/© 2014 Elsevier Inc. All rights reserved. a b s t r a c t a r t i c l e i n f o Article history: Received 9 December 2014 Accepted 15 December 2014 Available online 16 December 2014 Keywords: Epithelial ovarian cancer ISH 19q12 CCNE1 Cyclin E1 URI Background: CCNE1 is frequently amplified in high grade serous ovarian cancer andmay serve as a target for ovar- ian cancer treatment. URI is closely related to CCNE1 at the 19q12 amplicon andmay also contribute to the onco- genic effect. Our objective was to investigate the relevance of CCNE1 and URI gene amplification and protein expression in different histological subtypes of epithelial ovarian cancer (EOC). Methods: A novel dual-color 19q12 in situ hybridization (ISH), covering CCNE1 and URI, and chromosome 19 as a surrogate using Ventana BenchMark XT platformwas developed and applied to 148 EOCs. URI and CCNE1 ampli- fications were separately assessed by fluorescence in situ hybridization (FISH). Immunohistochemistry using a Cyclin E1 and a novel URI monoclonal antibody was performed. Results:Amplification of 19q12was found in 36.6%, CCNE1 in 21.7%,URI in 9.9%, and both genes simultaneously in 9% of EOC cases. High Cyclin E1 and URI protein expression were observed in 52.2% and 26.1%, respectively. Am- plification of 19q12 occurred in all EOC subtypes andwas associatedwith amplification and expression of CCNE1/ Cyclin E1, URI, TP53 mutation, and advanced stage. Conclusion: The novel 19q12 ISH probe reliably detects both CCNE1 and URI amplifications as confirmed by FISH. The combination of 19q12 amplification with Cyclin E1 and URI protein expression may help to select patients more likely to benefit from CDK2 targeted therapies. © 2014 Elsevier Inc. All rights reserved. 1. Introduction Epithelial ovarian cancer (EOC) represents a variety of histologic subtypes with different biological behaviors. High-grade serous ovarian cancer (HGSOC) is the most common subtype of EOC, which is charac- terized by poor prognosis due to late diagnosis in advanced tumor stages, frequent tumor relapse anddevelopment of chemotherapy resis- tance. At the molecular basis, HGSOC is dominated by a high frequency of TP53 mutations as well as high level of DNA amplifications and deletions. Genomic amplification at 19q12 has been observed in different human carcinomas, including EOC (Lin et al., 2000; Leung et al., 2006; Natrajan et al., 2012). CCNE1 has been attributed as the driver of tumor- igenesiswithin the 19q12 amplicon (Lin et al., 2000; Etemadmoghadam et al., 2010), and several signaling pathways (including TP53) are thought to control Cyclin E1 activity (Siu et al., 2012). Cyclin E1 in asso- ciation with the cyclin-dependent kinase 2 (CDK2) activates the cell cycle by promoting the G1–S phase transition. The CCNE1 gene is known to be amplified and/or over-expressed in EOC (Schraml et al., 2003). In HGSOC, amplification of CCNE1 has been observed in up to 20% of cases and the overall survival tended to be lower for those pa- tients (Cancer Genome Atlas Research N, 2011). It was further reported that CCNE1 amplification is related to primary platinum treatment failure in ovarian cancer patients (Etemadmoghadam et al., 2009). This is in contrast to BRCA1/BRCA2-mutant ovarian carcinomas,which ini- tially respond to platinum-based chemotherapy. Interestingly, CCNE1 am- plifications are mutually exclusive with BRCA1/BRCA2 mutations in HGSOC (Etemadmoghadam et al., 2013a). Other genes within the 19q12 amplicon may also contribute to ovarian cancer progression including URI (C19Orf2), C19Orf12, POP4, and PLEKHF1 (Etemadmoghadam et al., 2010; Theurillat et al., 2011). Recently, we and others identified URI am- plification in about 10% of human ovarian carcinomas and ovarian cancer cell lines (Theurillat et al., 2011; Davis et al., 2013). Therefore, the protein expression of the pro-survival protein URI may also contribute to the on- cogenic effect of 19q12 amplification. Cyclin E1 may be targeted therapeutically by its partner kinase CDK2. Currently, various clinical trials are ongoingwhich target individ- ual or multiple CDKs (Lapenna and Giordano, 2009; Cicenas and Valius, 2011). CDK inhibitors either block CDK activity or prevent their http://crossmark.crossref.org/dialog/?doi=10.1016/j.yexmp.2014.12.004&domain=pdf http://dx.doi.org/10.1016/j.yexmp.2014.12.004 mailto:[email protected] http://dx.doi.org/10.1016/j.yexmp.2014.12.004 http://www.sciencedirect.com/science/journal/00144800 www.elsevier.com/locate/yexmp 48 A. Noske et al. / Experimental and Molecular Pathology 98 (2015) 47–54 interaction with cyclins, resulting in reduced cell proliferation. HGSOC with high-level CCNE1 amplification may represent a tumor type with potential treatment response to CDK inhibitors, butmay require specific patient selection to observe a therapeutic benefit (Etemadmoghadam et al., 2013b). Treatment decisions could potentially depend on the 19q12 (CCNE1/URI) copy number status, in combination with Cyclin E1 and/or URI protein expression. It is important to characterize potential therapeutic targets aswell as mechanisms of drug resistance in EOC with different molecular back- grounds. The objective of our study was to evaluate the relevance of the 19q12 amplicon including both CCNE1 and URI, in early and late tumor stages, aswell as different EOC subtypes.We investigated the im- portance of genomic amplification using a novel chromogenic ISH assay for automateddetection of the 19q12 amplicon and investigated CCNE1/ Cyclin E1 andURI amplification and expression by FISH and immunohis- tochemistry, respectively. 2. Methods 2.1. EOC cohort These studies utilize a tissuemicroarray (TMA) consisting of primary epithelial ovarian carcinomas (n = 148) diagnosed at the Institute of Surgical Pathology, University Hospital Zurich (Switzerland) between 1995 and 2005. Tissue samples were fixed in 4% neutral buffered form- aldehyde and embedded in paraffin. Routine hematoxylin and eosin sections were processed for histopathological evaluation. Histological subtype and grade were re-assessed according to the WHO classifica- tion 2014. Low grade serous ovarian carcinomas and borderline tumors were not included in this study. The carcinomaswere categorized in low grade (low grade endometrioid and mucinous) and high grade (high grade serous, high grade endometrioid, and clear cell). Tumor stage was classified according to the International Federation of Gynaecology and Obstetrics (FIGO) staging system. The median patient age at the time of surgery was 60 yrs, ranging between 25 and 87 yrs of age. The mean follow-up time was 39 months (range, 0.1–161 months). Data on adjuvant chemotherapy was known for all patients. In the majority of the cases (68%), a platinum-based combination therapywas adminis- tered. There were 29% of the patients who did not receive chemothera- py. The studywas approved by the Cantonal Ethics Committee of Zurich (StV 27–2009). 2.2. Detection of the 19q12 amplicon using a novel dual-color ISH automated assay The amplification status of the 19q12 region was identified using a DNA probe set (Ventana Medical Systems, Tucson, Arizona) bymeasur- ing the copy number ratio of the 19q12 amplicon to the chromosome19 (Fig. 1). The 19q12 DNP ISH probe (Ventana) is a DNA probe free of repetitive DNA sequences that cover a span of approximately 560 kb of chromosome 19q12 containing the coding sequences of CCNE1 and URI. Due to homology within the alpha-satellite sequences of chromosomes Fig. 1. Schematic diagram illustrating the positions of the probes u 1, 5 and19, it is difficult to specifically identify chromosome19. As a result of this homology, a second repeat-freeDNAprobewas developed to allow enumeration of chromosome 19 copy numbers. The INSR DIG probe hy- bridizes to a span of approximately 600 kb within 19p13.2 and 19p13.3 which includes the coding sequences for insulin receptor, INSR. The dual-color ISH assay was established and automated using the VENTANA BenchMark XT platform. Formalin-fixed paraffin embedded specimens (4 μm) were prepared by cell conditioning, protease treat- ment, and denaturation. Protocols used were specifically optimized for the intended cohort of this study. Briefly, after deparaffinization, sam- ples were treated with cell conditioning 2 (CC2) for four 12 min cycles followed by ISH protease 2 for 8 or 12 min. After co-denaturation, 19q12 DNP and INSR DIG probes were hybridized at 51 °C for 4 h and washed 3 cycles at 68 °C for 8 min. 19q12 DNP and INSR DIG signals were detected using VENTANA ultraView SISH DNP and VENTANA ultraView RED ISH DIG detection kits, respectively, and counterstained with hematoxylin II and bluing reagent. After optimization, the feasibil- ity of this assay was tested on a training set of 40 large sections of 4 μm thickness of FFPE ovarian cancer samples with known URI amplification status from a previous FISH study (Theurillat et al., 2011) in a blinded manner. First, the sample was evaluated for acceptable staining including the presence of appropriate signals in normal cells, adequate signal strength, and low background. Samples failing to meet these criteria were not included within the study data. After identifying regions for analysis, the reader recorded the number of 19q12 and INSR signals within 50 representative cells. In the case of TMAs containing small tis- sue sections, the total number of 50 nuclei was obtained by combining the enumerated nuclei from multiple cores. Amplification was defined as a function of the ratio of the average number of copies of the 19q12 amplicon to the average number of copies of chromosome 19 per cell. Both parameters are detectable on one slide and appear as black (19q12) and red (chr19) dots. A sample was considered amplified if the ratio of 19q12/INSR was found being equal to or greater 2.0 (Kuhn et al., 2014). Examples are shown in Fig. 2. 2.3. Analysis of CCNE1 amplification by Fluorescent In Situ Hybridization (FISH) For evaluation of CCNE1 copy numbers, we established a two-color FISH technique using a specific probe for CCNE1 (150 kb) and a refer- ence probe (570 kb) to the 19p13.11 region of chromosome 19 (FG0013 Abnova Corporation, Taipei, Taiwan). The TMA slide was de- paraffinized, hydrated and placed in deionized water. Afterwards, the section was pretreated with citrate buffer (pH 6) for 20 min (180 W), Pronase solution for 6min (37 °C), fixedwith formalin, and dehydrated. During the procedure, the slide was several times rinsed. After adding the FISH probe, denaturation for 5 min (75 °C) and hybridization for 16–24 h (37 °C) were achieved by using the ThermoBrite system (Ab- bott Laboratories, Abbott Park, Illinois, USA). After stringent washing (2× SSC/0.3% NP-40), the slide was mounted with diamidino-2- phenylindole (DAPI) solution. The FISH signals were analyzed with a sed for chromogenic detection of amplification of the 19q12. Fig. 2. Dual-color 19q12 ISH in EOC. A: Normal gene copy numbers of 19q12 (black signals) and chromosome 19 (red signals) and negative for 19q12 amplification. B: Increased copy numbers of 19q12 region but not of Chr19 equivalent to an amplification of 19q12. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.) 49A. Noske et al. / Experimental and Molecular Pathology 98 (2015) 47–54 Zeiss fluorescence microscope, AXIO Imager Z1. CCNE1 amplification was defined as the ratio of CCNE1 to CEN19p ≥2 or high polysomy with ≥4 copies in N40% cells (Kuhn et al., 2014). The URI amplification status was previously assessed by FISH (Theurillat et al., 2011). 2.4. Cyclin E1 and URI immunohistochemistry The above-mentioned TMA was incubated with a monoclonal anti- body against Cyclin E1 (clone HE12, Santa Cruz Biotechnology, CA) andwas scored based on nuclear expression, the intensity of the immu- noreaction (0— no staining, 1—weak, 2—moderate, and 3— strong) as well as the percentage of positive tumor cells. Afterwards, the H-score was calculated. The median H-score was used to determine the cut- point to define low and high expression of Cyclin E1. A novel rabbit monoclonal antibody specific for URI (1–21, VENTANA)was generated and the optimal immunohistochemical stain- ing procedure was tested on a multi-tissue array with different staining procedures. After evaluation by two labs and two pathologists, a con- sensus protocol was chosen for optimal staining. The URI antibody was then applied to the TMA. We observed a cytoplasmic URI expres- sion, which was scored according to the staining intensity (no, weak, moderate, and strong staining). 2.5. TP53 mutation analysis The EOC cohortwas previously investigated for TP53 genemutations in exons 5–8 by pyro-sequencing using the GS Junior 454 platform (Rechsteiner et al., 2013). 2.6. Statistics The statistical analysis was carried out with IBM SSPS software (ver- sion 21). Overall survival data were plotted as Kaplan–Meier curves and the statistical significancewas calculated using the log-rank test. The as- sociation between 19q12, CCNE1, and URI amplification status and clin- icopathological parameters, expression of Cyclin E1 and URI as well as TP53 gene mutation status were assessed either applying the Pearson's chi square or Fisher's exact test (for categorical variables) and two- sample t-test (for continuous variables). McNemar Test for correlated proportions was used to assess differences in paired proportions for assay comparisons. p-Values of 0.05 or lesswere considered statistically significant. 3. Results 3.1. 19q12 amplification occurs in all epithelial ovarian cancer subtypes The 19q12 ISH assay was applied on a TMA with different EOC subtypes. Due to the lack of tumor tissue or weak ISH signals not all tis- sue cores were available for the analysis. The assay interpretation was possible on 134 out of 148 tissue samples and was independently per- formed by two pathologists (A.N., S.S.). Discrepant cases were discussed until consensus was reached. Amplification of 19q12 was found in 49 (36.6%) of the carcinomas and significantly associated with advanced FIGO stage and older patient age. 19q12 amplification occurred in all histologic subtypes, but most commonly in high grade serous carci- nomas. The findings are summarized in Table 1. In the univariate Kaplan–Meier analysis, no differences in patient overall survival for Table 1 Association of 19q12, CCNE1, and URI amplification with clinicopathological features. 19q12 amplification n (%) p-Value CCNE1 amplification n (%) p-Value URI amplification n (%) p-Value Total 49/134 (36.6) 30/138 (21.7) 13/131 (9.9) Stage 0.006⁎ 0.014⁎ 0.753⁎ Early (FIGO I/II) 8/40 (20) 4/42 (9.5) 3/39 (7.7) Late (FIGO (III/IV) 40/87 (46) 26/90 (28.9) 10/86 (11.6) Grade 0.100⁎ 0.014⁎ 0.296⁎ Low grade 8/33 (24.2) 2/32 (6.3) 1/30 (3.3) High grade 41/101 (40.6) 28/106 (26.4) 12/101 (11.9) Histologic type 0.089⁎⁎ 0.042⁎⁎ 0.077⁎⁎ High grade serous 30/67 (44.8) 21/71 (29.6) 11/69 (15.9) Endometrioid 7/29 (24.1) 3/28 (10.7) 2/27 (7.4) Clear cell 10/25 (40) 6/26 (23.1) 0/23 (0) Mucinous 2/13 (15.4) 0/13 (0) 0/12 (0) Residual disease 0.628⁎ 0.258⁎ 0.99⁎ b1 cm 8/25 (32) 3/26 (11.5) 3/25 (12) ≥1 cm 26/65 (40) 16/68 (23.5) 8/66 (12.1) Age 0.049⁎ 0.001⁎ 0.037⁎ b60 yrs 17/62 (27.4) 6/65 (9.2) 2/60 (3.3) ≥60 yrs 32/72 (44.4) 24/73 (32.8) 11/71 (15.5) ⁎ Fisher's exact test. ⁎⁎ Pearson's chi square test. 50 A. Noske et al. / Experimental and Molecular Pathology 98 (2015) 47–54 19q12 amplified or non-amplified carcinomas was seen (log-rank test, p = 0.841). 3.2. CCNE1 and URI amplifications in EOC To evaluate the prevalence of pure CCNE1 amplification, we used FISH analysis and observed amplification in 30 cases (21.7%) which was significantly related to high grade EOC, advanced FIGO stage, high grade serous histology, and older patient age (Table 1). URI amplification, as previously assessed by FISH, was observed in 9.9% of the EOCs and associated with older patient age, but not with other clinicopathological parameters. However, nearly all URI amplifica- tions were identified in high grade serous carcinomas (Table 1). No dif- ferences in overall survival were observed for both genes (log-rank test for CCNE1 p = 0.337; for URI p = 0.605). We further found a co-amplification of CCNE1 andURI in 9 carcinomas (75%) while 19 CCNE1 amplified EOCs did not show URI amplification. 3.3. Cyclin E1 and URI protein expression in EOC subtypes Analysis of Cyclin E1 immunohistochemistry was possible in 138 carcinomas. Representative images of the different levels of the nuclear expression are shown in Fig. 3. The median H-score was 140 (range 0–300) and used to define the cut-point between low and high Cyclin E1 expression. Carcinomas with high expression levels were observed in 52.2% which were significantly more frequent in high grade EOC and in the clear cell and serous subtypes (Table 2). Evaluation of URI protein expression by immunohistochemistry was possible in 142 carcinomas. URI expression was exclusively detected in the cytoplasm with different intensity levels as shown in Fig. 4. Moder- ate and strong cytoplasmic staining was considered as URI positive and was observed in 26.1% of the carcinomas. URI positivity was significant- ly associated with the high grade serous subtype (Table 2). No differences in patient overall survival were observed for both proteins (log-rank test for Cyclin E1, p = 0.99 and URI, p = 0.605). 3.4. Association of 19q12 with CCNE1 and URI amplifications and expression The 19q12 ISH assay covers both CCNE1 and URI. Therefore we per- formed statistical analyses to compare the 19q12 amplification with the amplification status of both genes individually (Table 3). We found an association between 19q12 and URI amplifications (p = 0.0001, McNemar Test) as well as CCNE1 amplification (p = 0.0001, McNemar Test). 19q12 amplification was also significantly associated with high Cyclin E1 and URI expression levels (p = 0.011 and p = 0.032, respectively, McNemar Test) (Table 3). However, there were cases with high protein expression levels without any amplification and vice versa. Examples of the different protein expression patterns in 19q12 amplified EOC are shown in Fig. 5. 3.5. Gene amplification is associated with protein expression CCNE1 amplificationwas associatedwith increased Cyclin E1 expres- sion levels (p b 0.0001, McNemar Test). The majority of the CCNE1 am- plified EOC (27 out of 30) showed high Cyclin E1 expression levels. However, therewere 44 EOCswithout CCNE1 amplification but high Cy- clin E1 protein expression levels. URI amplification was significantly associated with high URI protein expression (p = 0.0001, McNemar Test) (Table 3). Our analysis re- vealed a high URI expression in 8 of 13URI amplified EOCs (61.5%). Con- versely, only 8 out of 37 cases with high URI expression showed URI amplification (21.6%). All URI amplified EOCs displayed also high Cyclin E1 protein expression. 3.6. TP53 mutations are frequent in 19q12 amplified EOC Since both TP53mutations and DNA amplifications are a hallmark of especially high grade serous EOC, we compared the TP53 mutational gene status previously evaluated in this EOC cohort (Rechsteiner et al., 2013) with the 19q12, CCNE1, and URI amplifications as well as Cyclin E1 and URI protein expression. TP53 mutations were identified in 71 of 131 EOCs (54.1%). We observed that 19q12 amplified EOC had signif- icantlymore TP53mutations (30 out of 46; 65%) compared to EOCwith- out 19q12 amplification (p = 0.001, McNemar Test) as shown in Table 3. In EOC without 19q12 amplification (n = 82), there were only 41 tumors with TP53mutations (50%). Similarly, EOC with CCNE1 and URI amplifications harbored significantly more TP53 mutations (22 out of 28; 78%; as well as 8 out of 13; 61%, each p = 0.0001, McNemar Test). However, there were 50 TP53 mutated EOCs without CCNE1 amplification and 59 TP53-mutated EOCs without URI amplifica- tion. Interestingly, high URI protein expression levels were significantly related to TP53 mutations (p = 0.0001, McNemar Test). Twenty-three out of 36 EOCs with high URI expression had TP53 mutation (63.8%) and 13 (36.1%) were wild type. In contrast, Cyclin E1 expression was not associated with TP53 mutation status. Fig. 3. Immunohistochemistry of Cyclin E1 in EOC. Examples of thedifferent expression levels are shown (magnification 20×). A:Mucinous EOCwith only lowexpression. B: Endometrioid EOC with weak to moderate expression. C–D: Clear cell ovarian carcinoma (left) and high grade serous ovarian cancer with high expression levels. 51A. Noske et al. / Experimental and Molecular Pathology 98 (2015) 47–54 4. Discussion In this study, we investigated two oncogenic drivers, CCNE1 andURI, within the 19q12 amplicon in epithelial ovarian cancer. Novel therapeu- tic approaches require a detailed knowledge of oncogenic drivers to guide treatment decisions. Recent molecular studies uncovered the mo- lecular profile of high grade serous ovarian cancer (HGSOC), the most common and aggressive EOC subtype. The genomic landscape of HGSOC is characterized by ubiquitous somatic TP53 mutations and numerous DNA amplifications, including 19q12 amplification with CCNE1 and URI as possible oncogenic drivers (Theurillat et al., 2011; Davis et al., 2013; Table 2 Association of URI and Cyclin E1 protein expression with clinicopathological features. URI high expression n (%) p-Value Cyclin E1 high expression n (%) p-Value Total 37/142 (26.1) 72/138 (52.2) Stage 0.407⁎ 0.999⁎ Early (FIGO I/II) 9/43 (20.9) 22/41 (53.7) Late (FIGO (III/IV) 26/92 (28.3) 47/90 (52.2) Grade 0.078⁎ 0.002⁎ Low grade 5/35 (14.3) 9/32 (28.1) High grade 32/107 (29.9) 63/106 (59.4) Histologic type 0.007⁎⁎ 0.010⁎⁎ High grade serous 27/72 (37.5) 42/72 (58.3) Endometrioid 7/30 (23.3) 11/27 (40.7) Clear cell 2/26 (7.7) 17/26 (65.4) Mucinous 1/14 (7.1) 2/13 (15.4) Residual disease 0.311⁎ 0.999⁎ b1 cm 5/26 (19.2) 14/26 (53.8) ≥1 cm 22/70 (31.4) 36/68 (52.9) Age 0.004⁎ 0.864⁎ b60 yrs 9/65 (13.8) 32/63 (50.7) ≥60 yrs 28/77 (36.4) 40/75 (53.3) ⁎ Fisher's exact test. ⁎⁎ Pearson's chi square test. Bowtell, 2010). The prevalence of gene amplification and/or protein ex- pression of these genes in other ovarian carcinoma subtypes is unclear. To determine the frequency of somatic CCNE1 amplifications, we used a CCNE1-specific probe for FISH analysis. In total, CCNE1 amplifica- tionwas found in 21% of all EOC subtypes and in 30% of HGSOC,which is consistent with previous reports (Cancer Genome Atlas Research N, 2011; Karst et al., 2014; Nakayama et al., 2010; Pils et al., 2014). In our study, CCNE1 amplification was significantly associated with advanced tumor stage and poor tumor differentiation, suggesting that CCNE1 am- plification may be involved in ovarian cancer progression. Importantly, CCNE1 amplification was also detected in clear cell and endometrioid subtypes, but not inmucinous EOC. Similar data have been also reported by Nakayama et al. (2010). CCNE1 has been proposed as the key oncogenic driver within the 19q12 amplicon in various human carcinomas (Lin et al., 2000; Etemadmoghadam et al., 2010), but it has become evident that this region comprises other co-regulated genes as shown in gastric, ovar- ian, and breast cancers (Leung et al., 2006; Natrajan et al., 2012; Etemadmoghadam et al., 2010; Theurillat et al., 2011). Therefore, it might be important that these other genes are also identified by assays for the characterization of the 19q12 amplification. We and others have recently shown that URI (C19orf2), contiguous to CCNE1, is amplified in a subset of ovarian carcinomas and ovarian cancer cell lines (Etemadmoghadam et al., 2010; Theurillat et al., 2011; Davis et al., 2013), suggesting that CCNE1 may not be an exclusive driver at 19q12. So far, it is not clear whether CCNE1 or URI or both constitute the main drivers of this amplicon. URI gene amplifications were most common in HGSOC in 10–15% (Theurillat et al., 2011) and less prevalent in other EOC subtypes (b10%). Importantly, 75% of the URI-amplified EOC showed co- amplification of CCNE1, suggesting that solitary URI amplification oc- curs in 25% of EOC cases. Amplification of URI but not of CCNE1 was also detected in a recent high-throughput siRNA screening of ovarian cancer cell lines. In this assay, URI, but not CCNE1 was related to cell viability (Davis et al., 2013). Down-regulation of URI selectively Fig. 4. Immunohistochemistry of the novel URI antibody in EOC. Examples of the different expression levels are shown (magnification 20×). A:Negativity in a highgrade serous carcinoma. B: Weak staining. C: Moderate. D: Strong expression in a high grade serous carcinoma. 52 A. Noske et al. / Experimental and Molecular Pathology 98 (2015) 47–54 affects the viability of those cells that feature amplification of URI, which implies that it may function as an “addicting” oncogene (Theurillat et al., 2011). Recently, Theurillat et al. suggested that co-amplification of CCNE1 and URImay provide a selective advantage for carcinomas — CCNE1 as a promoter of cell cycle and URI for cell survival. In contrast, activation of the cell cycle may also stimulate tumor cell proliferation, which renders ovarian carcinoma cell lines more sensitive to platinum agents (Bedrosian et al., 2004). To better evaluate the 19q12 amplicon, we developed a novel dual- color 19q12 ISH assay, encompassing CCNE1 and URI. We observed an amplification of 19q12 in 36.6% of the EOCs. This is in line with our CCNE1 and URI FISH data and previous studies reporting an amplifica- tion of CCNE1 in up to 26% and of URI in 10% (Cancer Genome Atlas Research N, 2011; Theurillat et al., 2011; Karst et al., 2014; Nakayama et al., 2010; Pils et al., 2014). This 19q12 ISH assay is a fully automated Table 3 Association of 19q12 amplification with URI, CCNE1/Cyclin E1, and TP53 status. 19q12 amplification n (%) p-Value⁎ URI-FISH (n = 122) b0.0001 Not amplified 37/112 (33) Amplified 9/10 (90) URI-IHC (n = 134) 0.032 Low level 29/100 (29) High level 20/34 (58.8) CCNE1-FISH (n = 131) b0.0001 Not amplified 19/102 (18.6) Amplified 29/29 (100) Cyclin E1-IHC (n = 131) 0.011 Low level 14/64 (21.9) High level 35/67 (52.2) TP53 (n = 128) 0.001 Wild-type 16/57 (28.1) Mutation 30/71 (42.3) ⁎ McNemar Test. approach providing permanently stained slides and the possibility to in- terpret the ISH signals in the context of tissue morphology. We further attempted to evaluate the association between 19q12 amplification andprotein expression of Cyclin E1 andURI. The combina- tion of both immunohistochemistry and ISH assay, similar to the assess- ment of the Her2 status in breast cancer, would be an interesting tool to test the 19q12 amplicon as a potential predictive marker. For this pur- pose,we alsoused a novelmonoclonalURI antibody. However, although URI protein expressionwas associatedwithURI gene amplification, high URI expression was also seen in about 25% of EOCs without URI amplifi- cation, which is similar to the situation, observed for CCNE1/Cyclin E1 amplification versus overexpression status. The staining results with our novel monoclonal URI antibody differed somewhat from the previ- ously reported URI immunohistochemistry (Etemadmoghadam et al., 2009), because it showed amore precise discrimination of the cytoplas- mic staining intensity in low and high expressions. In this study, we demonstrate that 19q12 amplification is associated with high Cyclin E1 and URI protein expression in about 50% of EOCs. However, we observed high expression levels of both proteins, not only in amplified, but also in non-amplified carcinomas. The discor- dance between amplification and protein expression is consistent with findings from recent reports (Natrajan et al., 2012; Karst et al., 2014; Nakayama et al., 2010). Cyclin E1 overexpressing EOC lacking CCNE1 amplification may have disordered transcription or posttranscriptional modifications that contribute to CCNE1/Cyclin E1 up-regulation (Karst et al., 2014). The discordance between 19q12 amplification and Cyclin E1/URI expression has to be considered, when IHC or FISH is used as prognostic or even as a predictive biomarker, because we have identi- fied subgroups of Cyclin E1 positive EOC with and without URI protein expression. Previous studies have reported correlations among CCNE1 amplification andCyclin E1 expression, but several studies showcontro- versial results in regard to the prognostic value of CCNE1/Cyclin E1 am- plification and/or protein expression levels. CCNE1 amplification significantly correlated with shorter disease-free survival in ovarian cancer patients in a multivariate analysis (Nakayama et al., 2010). Fig. 5. Examples of 19q12 amplified EOC and their patterns of Cyclin E1 and URI expression. Case 7 shows an amplification of 19q12 and high Cyclin E1 expression but negativity for URI. Case 156 is characterized by 19q12 amplification and expression of both proteins. 53A. Noske et al. / Experimental and Molecular Pathology 98 (2015) 47–54 However, Nakayama et al. foundno relation to overall survival in an EOC cohort of different histological types. The TCGA analysis of high grade serous ovarian carcinomas showed a trend that patients with CCNE1 am- plification have a shorter overall survival (p = 0.072) (Cancer Genome Atlas Research N, 2011). CCNE1 copy number, assessed by qPCR, was sig- nificantly associated with shorter progression-free and overall survival in patients with advanced serous cancer (Etemadmoghadam et al., 2010). The study also showed that higher gene expression levels of CCNE1 bymi- croarray were related to shorter progression-free but not to overall sur- vival. The authors concluded that copy number status is much more informative than gene expression. In contrast to these findings, high CCNE1 gene expression, measured by qPCR, correlatedwith better overall survival in serous EOC of FIGO II–IV stages (Pils et al., 2014). That study further demonstrated that CCNE1 copy number significantly correlated with expression, but not with patient outcome. Similarly, the impact of Cyclin E1protein expression on patient outcome is controversially report- ed (Heeran et al., 2012; Mayr et al., 2006). Generally, there is a concordance between methods analyzing ex- pression of CCNE1 gene and protein. As demonstrated in our analysis, not all of the CCNE1 amplified tumors show high protein expression levels. Different approaches and cut-point definitions were used and may result in controversial outcome results. Therefore, the prognostic strength of the 19q12 amplicon in EOC has to be further investigated. Amplified CCNE1 was suggested as a predictive marker for primary treatment failure (platinum-based chemotherapy) in ovarian cancer patients (Etemadmoghadam et al., 2009). So far, it is not clear how CCNE1 contributes to resistance and/or poor response to standard che- motherapy. In our previous study,URI amplification and protein expres- sion were significantly related to poor survival, and URI amplification in ovarian cancer cells may contribute to cisplatin resistance (Theurillat et al., 2011). Further prospective studies are needed to evaluate whether either CCNE1 or URI, or both genes' amplification status are el- igible predictive markers for chemotherapy response. Finally, we observed a significant association between 19q12 and CCNE1/URI amplifications as well as URI protein expression levels with the presence of TP53 mutations. This is consistent with findings from high-grade breast carcinomas, which harbored TP53 gene (exons 2–11) mutations and CCNE1 amplification simultaneously (Natrajan et al., 2012). High grade serous ovarian carcinomas are mainly driven by TP53 mutations and widespread DNA copy number changes, espe- cially 19q12, which results in chromosomal instability and tumor progression (Cancer Genome Atlas Research N, 2011; Bowtell, 2010). Taken together, CCNE1 and URI may constitute the main drivers of the 19q12 amplicon. Understanding the amplification status of CCNE1/ URI may further help to identify ovarian cancer patients most likely to respond to standard treatment or may benefit from therapeutic ap- proaches targeting cell cycle checkpoints. To further evaluate the 19q12 ISH assay (in combination with protein expression analysis) and to clarify the potential role as a predictive test, an investigation of a larger cohort with treatment and outcome data is required. Conflict of interest statement Holger Moch and Aurelia Noske (Institute of Pathology, University Hospital Zurich) received research funds from Roche/Ventana. Acknowledgments We would like to thank André Fitsche, Jasmine Roth, and Katherine Smith for technical and study coordination assistance. The authors, Holger Moch and Aurelia Noske, received research funds from Roche/Ventana. 54 A. Noske et al. / Experimental and Molecular Pathology 98 (2015) 47–54 References Bedrosian, I., Lu, K.H., Verschraegen, C., Keyomarsi, K., 2004. Cyclin E deregulation alters the biologic properties of ovarian cancer cells. Oncogene 23, 2648–2657. Bowtell, D.D., 2010. The genesis and evolution of high-grade serous ovarian cancer. Nat. Rev. Cancer 10, 803–808. Cancer Genome Atlas Research N, 2011. Integrated genomic analyses of ovarian carcino- ma. Nature 474, 609–615. Cicenas, J., Valius, M., 2011. The CDK inhibitors in cancer research and therapy. J. Cancer Res. Clin. Oncol. 137, 1409–1418. Davis, S.J., Sheppard, K.E., Pearson, R.B., Campbell, I.G., Gorringe, K.L., Simpson, K.J., 2013. Functional analysis of genes in regions commonly amplified in high-grade serous and endometrioid ovarian cancer. Clin. Cancer Res. 19, 1411–1421. Etemadmoghadam, D., deFazio, A., Beroukhim, R., Mermel, C., George, J., Getz, G., Tothill, R., Okamoto, A., Raeder, M.B., Harnett, P., Lade, S., Akslen, L.A., Tinker, A.V., Locandro, B., Alsop, K., Chiew, Y.E., Traficante, N., Fereday, S., Johnson, D., Fox, S., Sellers, W., Urashima, M., Salvesen, H.B., Meyerson, M., Bowtell, D., 2009. Group AS: integrated genome-wide DNA copy number and expression analysis identifies dis- tinct mechanisms of primary chemoresistance in ovarian carcinomas. Clin. Cancer Res. 15, 1417–1427. Etemadmoghadam, D., George, J., Cowin, P.A., Cullinane, C., Kansara, M., Australian Ovar- ian Cancer Study G, Gorringe, K.L., Smyth, G.K., Bowtell, D.D., 2010. Amplicon- dependent CCNE1 expression is critical for clonogenic survival after cisplatin treat- ment and is correlated with 20q11 gain in ovarian cancer. PLoS One 5, e15498. Etemadmoghadam, D., Weir, B.A., Au-Yeung, G., Alsop, K., Mitchell, G., George, J., Australian Ovarian Cancer Study G, Davis, S., D'Andrea, A.D., Simpson, K., Hahn, W.C., Bowtell, D.D., 2013a. Synthetic lethality between CCNE1 amplification and loss of BRCA1. Proc. Natl. Acad. Sci. U. S. A. 110, 19489–19494. Etemadmoghadam, D., Au-Yeung, G., Wall, M., Mitchell, C., Kansara, M., Loehrer, E., Batzios, C., George, J., Ftouni, S., Weir, B.A., Carter, S., Gresshoff, I., Mileshkin, L., Rischin, D., Hahn, W.C., Waring, P.M., Getz, G., Cullinane, C., Campbell, L.J., Bowtell, D.D., 2013b. Resistance to CDK2 inhibitors is associated with selection of polyploid cells in CCNE1-amplified ovarian cancer. Clin. Cancer Res. 19, 5960–5971. Heeran,M.C., Hogdall, C.K., Kjaer, S.K., Christensen, L., Blaakaer, J., Christensen, I.J., Hogdall, E.V., 2012. Limited prognostic value of tissue protein expression levels of cyclin E in Danish ovarian cancer patients: from the Danish ‘MALOVA’ ovarian cancer study. APMIS 120, 846–854. Karst, A.M., Jones, P.M., Vena, N., Ligon, A.H., Liu, J.F., Hirsch, M.S., Etemadmoghadam, D., Bowtell, D.D., Drapkin, R., 2014. Cyclin e1 deregulation occurs early in secretory cell transformation to promote formation of fallopian tube-derived high-grade serous ovarian cancers. Cancer Res. 74, 1141–1152. Kuhn, E., Bahadirli-Talbott, A., Shih Ie, M., 2014. Frequent CCNE1 amplification in endo- metrial intraepithelial carcinoma and uterine serous carcinoma. Mod. Pathol. 27, 1014–1019. Lapenna, S., Giordano, A., 2009. Cell cycle kinases as therapeutic targets for cancer. Nat. Rev. Drug Discov. 8, 547–566. Leung, S.Y., Ho, C., Tu, I.P., Li, R., So, S., Chu, K.M., Yuen, S.T., Chen, X., 2006. Comprehensive analysis of 19q12 amplicon in human gastric cancers. Mod. Pathol. 19, 854–863. Lin, L., Prescott, M.S., Zhu, Z., Singh, P., Chun, S.Y., Kuick, R.D., Hanash, S.M., Orringer, M.B., Glover, T.W., Beer, D.G., 2000. Identification and characterization of a 19q12 amplicon in esophageal adenocarcinomas reveals cyclin E as the best candidate gene for this amplicon. Cancer Res. 60, 7021–7027. Mayr, D., Kanitz, V., Anderegg, B., Luthardt, B., Engel, J., Lohrs, U., Amann, G., Diebold, J., 2006. Analysis of gene amplification and prognostic markers in ovarian cancer using comparative genomic hybridization for microarrays and immunohistochemical analysis for tissue microarrays. Am. J. Clin. Pathol. 126, 101–109. Nakayama, N., Nakayama, K., Shamima, Y., Ishikawa, M., Katagiri, A., Iida, K., Miyazaki, K., 2010. Gene amplification CCNE1 is related to poor survival and potential therapeutic target in ovarian cancer. Cancer 116, 2621–2634. Natrajan, R., Mackay, A., Wilkerson, P.M., Lambros, M.B.,Wetterskog, D., Arnedos, M., Shiu, K.K., Geyer, F.C., Langerod, A., Kreike, B., Reyal, F., Horlings, H.M., van de Vijver, M.J., Palacios, J., Weigelt, B., Reis-Filho, J.S., 2012. Functional characterization of the 19q12 amplicon in grade III breast cancers. Breast Cancer Res. 14, R53. Pils, D., Bachmayr-Heyda, A., Auer, K., Svoboda, M., Auner, V., Hager, G., Obermayr, E., Reiner, A., Reinthaller, A., Speiser, P., Braicu, I., Sehouli, J., Lambrechts, S., Vergote, I., Mahner, S., Berger, A., Cacsire Castillo-Tong, D., Zeillinger, R., 2014. Cyclin E1 (CCNE1) as independent positive prognostic factor in advanced stage serous ovarian cancer patients — a study of the OVCAD consortium. Eur. J. Cancer 50, 99–110. Rechsteiner, M., Zimmermann, A.K., Wild, P.J., Caduff, R., von Teichman, A., Fink, D., Moch, H., Noske, A., 2013. TP53 mutations are common in all subtypes of epithelial ovarian cancer and occur concomitantly with KRAS mutations in the mucinous type. Exp. Mol. Pathol. 95, 235–241. Schraml, P., Bucher, C., Bissig, H., Nocito, A., Haas, P., Wilber, K., Seelig, S., Kononen, J., Mihatsch, M.J., Dirnhofer, S., Sauter, G., 2003. Cyclin E overexpression and amplifica- tion in human tumours. J. Pathol. 200, 375–382. Siu, K.T., Rosner, M.R., Minella, A.C., 2012. An integrated view of cyclin E function and reg- ulation. Cell Cycle 11, 57–64. Theurillat, J.P., Metzler, S.C., Henzi, N., Djouder, N., Helbling, M., Zimmermann, A.K., Jacob, F., Soltermann, A., Caduff, R., Heinzelmann-Schwarz, V., Moch, H., Krek, W., 2011. URI is an oncogene amplified in ovarian cancer cells and is required for their survival. Cancer Cell 19, 317–332. http://refhub.elsevier.com/S0014-4800(14)00201-9/rf0105 http://refhub.elsevier.com/S0014-4800(14)00201-9/rf0105 http://refhub.elsevier.com/S0014-4800(14)00201-9/rf0085 http://refhub.elsevier.com/S0014-4800(14)00201-9/rf0085 http://refhub.elsevier.com/S0014-4800(14)00201-9/rf0035 http://refhub.elsevier.com/S0014-4800(14)00201-9/rf0035 http://refhub.elsevier.com/S0014-4800(14)00201-9/rf0065 http://refhub.elsevier.com/S0014-4800(14)00201-9/rf0065 http://refhub.elsevier.com/S0014-4800(14)00201-9/rf0055 http://refhub.elsevier.com/S0014-4800(14)00201-9/rf0055 http://refhub.elsevier.com/S0014-4800(14)00201-9/rf0040 http://refhub.elsevier.com/S0014-4800(14)00201-9/rf0040 http://refhub.elsevier.com/S0014-4800(14)00201-9/rf0040 http://refhub.elsevier.com/S0014-4800(14)00201-9/rf0040 http://refhub.elsevier.com/S0014-4800(14)00201-9/rf0020 http://refhub.elsevier.com/S0014-4800(14)00201-9/rf0020 http://refhub.elsevier.com/S0014-4800(14)00201-9/rf0020 http://refhub.elsevier.com/S0014-4800(14)00201-9/rf0045 http://refhub.elsevier.com/S0014-4800(14)00201-9/rf0045 http://refhub.elsevier.com/S0014-4800(14)00201-9/rf0070 http://refhub.elsevier.com/S0014-4800(14)00201-9/rf0070 http://refhub.elsevier.com/S0014-4800(14)00201-9/rf0110 http://refhub.elsevier.com/S0014-4800(14)00201-9/rf0110 http://refhub.elsevier.com/S0014-4800(14)00201-9/rf0110 http://refhub.elsevier.com/S0014-4800(14)00201-9/rf0090 http://refhub.elsevier.com/S0014-4800(14)00201-9/rf0090 http://refhub.elsevier.com/S0014-4800(14)00201-9/rf0090 http://refhub.elsevier.com/S0014-4800(14)00201-9/rf0075 http://refhub.elsevier.com/S0014-4800(14)00201-9/rf0075 http://refhub.elsevier.com/S0014-4800(14)00201-9/rf0075 http://refhub.elsevier.com/S0014-4800(14)00201-9/rf0060 http://refhub.elsevier.com/S0014-4800(14)00201-9/rf0060 http://refhub.elsevier.com/S0014-4800(14)00201-9/rf0010 http://refhub.elsevier.com/S0014-4800(14)00201-9/rf0010 http://refhub.elsevier.com/S0014-4800(14)00201-9/rf0005 http://refhub.elsevier.com/S0014-4800(14)00201-9/rf0005 http://refhub.elsevier.com/S0014-4800(14)00201-9/rf0005 http://refhub.elsevier.com/S0014-4800(14)00201-9/rf0115 http://refhub.elsevier.com/S0014-4800(14)00201-9/rf0115 http://refhub.elsevier.com/S0014-4800(14)00201-9/rf0115 http://refhub.elsevier.com/S0014-4800(14)00201-9/rf0095 http://refhub.elsevier.com/S0014-4800(14)00201-9/rf0095 http://refhub.elsevier.com/S0014-4800(14)00201-9/rf0015 http://refhub.elsevier.com/S0014-4800(14)00201-9/rf0015 http://refhub.elsevier.com/S0014-4800(14)00201-9/rf0100 http://refhub.elsevier.com/S0014-4800(14)00201-9/rf0100 http://refhub.elsevier.com/S0014-4800(14)00201-9/rf0100 http://refhub.elsevier.com/S0014-4800(14)00201-9/rf0080 http://refhub.elsevier.com/S0014-4800(14)00201-9/rf0080 http://refhub.elsevier.com/S0014-4800(14)00201-9/rf0080 http://refhub.elsevier.com/S0014-4800(14)00201-9/rf0030 http://refhub.elsevier.com/S0014-4800(14)00201-9/rf0030 http://refhub.elsevier.com/S0014-4800(14)00201-9/rf0025 http://refhub.elsevier.com/S0014-4800(14)00201-9/rf0025 http://refhub.elsevier.com/S0014-4800(14)00201-9/rf0050 http://refhub.elsevier.com/S0014-4800(14)00201-9/rf0050 http://refhub.elsevier.com/S0014-4800(14)00201-9/rf0050 Characterization of the 19q12 amplification including CCNE1 and URI in different epithelial ovarian cancer subtypes 1. Introduction 2. Methods 2.1. EOC cohort 2.2. Detection of the 19q12 amplicon using a novel dual-color ISH automated assay 2.3. Analysis of CCNE1 amplification by Fluorescent In Situ Hybridization (FISH) 2.4. Cyclin E1 and URI immunohistochemistry 2.5. TP53 mutation analysis 2.6. Statistics 3. Results 3.1. 19q12 amplification occurs in all epithelial ovarian cancer subtypes 3.2. CCNE1 and URI amplifications in EOC 3.3. Cyclin E1 and URI protein expression in EOC subtypes 3.4. Association of 19q12 with CCNE1 and URI amplifications and expression 3.5. Gene amplification is associated with protein expression 3.6. TP53 mutations are frequent in 19q12 amplified EOC 4. Discussion Conflict of interest statement Acknowledgments References