Main

Ovarian cancer is the major cause of gynecologic cancer mortality. In 2008, it is estimated that about 21 650 women will be diagnosed with ovarian carcinoma in the United States; and about 15 520 women will die of the disease, ranking as the eighth most frequent malignancy and fifth leading cause of female cancer deaths. The 5-year survival rate is only 45%.1

Owing to its location and the lack of early clinical symptoms and non-availability of effective screening tests, ovarian cancer is usually diagnosed at an advanced stage. Primary radical surgery with the aim of maximal cytoreduction followed by adjuvant chemotherapy with carboplatin and paclitaxel is the cornerstone in the clinical management of ovarian cancer. Although such therapeutic intervention results in a disease remission, most patients will relapse and die because of tumor progression.2

Prognosis and selection of therapy is influenced mainly by disease stage, in addition to patient age, histological type and grade of the tumor. Identification of novel molecular targets relevant to diagnosis, prognosis and prediction of response to therapy is essential. The new therapeutic strategies focus on the combination of molecular targeted therapies with cytotoxic agents, which might lead to an improvement in treatment response and patient outcome.

Topoisomerase IIα (Top IIα) plays a central role in DNA metabolism, including DNA replication, transcription, DNA recombination, chromosome condensation and organization of chromosome structure,3, 4, 5 and is a marker of cell proliferation.6, 7, 8 Top IIα is a molecular target for a variety of antineoplastic agents, including anthracyclines (eg, pegylated doxorubicin, daunorubicin), epipodophyllotoxins (eg, etoposide (VP16)), teniposide (VM-26), aminocridines (m-AMSA), anthracenediones (eg, mitoxantrone) and actinomycin D. They are referred as topoisomerase II inhibitors. These agents bind the DNA-Top IIα complex and inhibit the religation of DNA. As a consequence, DNA breaks accumulate in the cell and trigger cell death.9, 10, 11, 12 Topoisomerase IIα gene expression has been suggested to predict tumor sensitivity to chemotherapy.13, 14

Top IIα expression, genetic alteration and enzyme activity have been studied in malignancies of different types.15, 16, 17, 18, 19, 20 However, earlier studies investigated Top IIα mRNA expression using frozen tissues as it had been expected that formalin fixation and paraffin embedding result in fragmentation of nucleic acids and chemical modification by protein–protein and protein–nucleic acid cross-links, making RNA extracted from such processed tissues unsuitable for molecular analysis techniques. Nevertheless, recent studies have proved the reliable gene expression measurement of RNA isolated from formalin-fixed paraffin-embedded (FFPE) tissues.21, 22 As the detection of biomarkers on the mRNA level in FFPE tissue is becoming increasingly important in pathological diagnosis, we have evaluated a novel technique for RNA extraction from FFPE tissue, which is based on magnetic particles for separation and purification of nucleic acids. We attempted to assess the feasibility of measuring Top IIα gene expression in RNA isolated from archival FFPE tissue specimens, which are used routinely in pathology laboratories.

The cellular functions of topoisomerase IIα as a marker of proliferation, a molecular target for chemotherapeutic drugs and a predictive marker for response to topoisomerase II inhibitors raise the hypothesis that Top IIα might be involved in the development and progression of ovarian carcinoma. We further hypothesized that the nuclear export protein chromosomal region maintenance/exportin 1 (CRM1), which mediates nuclear export of proteins and mRNAs,23 is involved in the regulation of the subcellular localization of Top IIα.

In this study, we investigated these hypotheses in a large cohort of primary ovarian carcinomas. The aim of this study was to investigate topoisomerase IIα expression at RNA and protein level, and to evaluate the relationship between Top IIα expression and clincopathological factors as well as outcome variables in patients with invasive ovarian cancer on the one hand and CRM1 expression on the other. We focussed on the hypothesis that a correlation of Top IIα expression with the tumor proliferation can be detected at mRNA level in an FFPE tissue-based approach. We carried out measurement of mRNA expression by quantitative real-time reverse transcription polymerase chain reaction (qRT-PCR) using RNA isolated from FFPE tissues in addition to immunohistochemistry staining to evaluate protein expression.

Materials and methods

Study Population

The study group included 133 patients who underwent surgery for primary ovarian cancer at the Department of Gynecology and Obstetrics of the Charité University Hospital, Berlin, Germany between 1991 and 2004. Specimens of all patients were examined and diagnosed at the Institute of Pathology, Charité University Hospital, Berlin, Germany. Selection of cases was on the basis of availability of tumor tissue together with follow-up data. Histology and grade were re-evaluated by an experienced gynecopathologist (CD). Hematoxylin and eosin (H&E)-stained sections were cut from paraffin-embedded blocks and used for histopathological evaluation; grading of tumors was assessed using Silverberg grading system;24 and staging of tumors was carried out according to the International Federation of Gynecology and Obstetrics (FIGO) staging system.

Patients underwent primary surgery with the goal of maximal tumor resection followed by standard combination chemotherapy with carboplatin and paclitaxel. None of the patients had received chemotherapy before surgery and collection of the tumor specimens. The patient characteristics are presented in Table 1. Data regarding postoperative chemotherapy were available for 121 patients (91%); of these patients, 112 (92.6%) received a platinum-based chemotherapy, 3 (2.5%) received other regimens and 6 (5.0%)—who were FIGO stage I—did not receive any chemotherapy.

Table 1 Characteristics of ovarian carcinoma patients
Full size table

Data on intraoperative residual tumor were available for 93 patients; of these patients, 81 (87.0%) had a postoperative residual tumor of less than 2 cm (Table 1).

Immunohistochemical Staining and Evaluation

Immunohistochemical staining was performed on tissue microarrays (TMAs), which were constructed by selection of representative tumor areas that were marked on the routine H&E-stained histological sections. For each case, four tissue cores (1.5 mm diameter) from different tumor areas of the sample tissue blocks were punched using a tissue micro-arrayer (Beecher Instruments; Woodland, CA, USA), and embedded into a new paraffin array block. Immunohistochemical staining was performed according to standard procedures. Briefly, slides were deparaffinized in xylene, and gradually rehydrated in graded solutions of ethanol. For antigen retrieval, slides were boiled in 0.01 M citrate buffer (pH 6.0) in a pressure cooker for 5 min, and they were incubated in protein block reagent (Dako, Glostrup, Denmark) for 10 min at room temperature, and subsequently incubated overnight at 4°C with a polyclonal rabbit antibody directed against human topoisomerase IIα (Bio-trend chemicals, Cologne, Germany), diluted 1:400 in an antibody diluent solution (Zytomed, Berlin, Germany). This was followed by incubation with a biotinylated anti-Ig universal secondary antibody and the biotin–streptavidin-amplified detection system multilink-alkaline phosphatase (Biogenex, San Ramon, CA, USA). After the development of color with fast-red chromogen system (Sigma, Steinheim, Germany), the slides were counterstained with Mayer's hematoxylin, and cover slips were applied with Aquatex (Merck, Darmstadt, Germany).

Topoisomerase IIα immunostaining in tumor cells was evaluated without prior knowledge of clinicopathological parameters and patient outcome. Two staining patterns of Top IIα were observed (nuclear and cytoplasmic); they were assessed and scored independently. Top IIα immunoreactivity was evaluated according to the percentage of positive cells and the intensity of staining. The percentage of positive cells was scored as 0 (0%); 1 (1–10%); 2 (11–50%); 3 (51–80%); and 4 (> 80%). The staining intensity was scored as 0 (negative), 1 (weak), 2 (moderate) and 3 (strong). For the immunoreactive score (IRS), the percentage of positive cells and staining intensity was multiplied, resulting in a value between 0 and 12. To distinguish cases with a weak or a strong expression of Top IIα in tumor tissue, we combined cases with an IRS of 0–3 to one group with negative to weak Top IIα expression (Top IIα-low expression), whereas cases with an IRS of 4–12 were combined into Top IIα-high-expression group.

Immunohistochemical staining and evaluation of CRM1 were performed according to the standard procedures as described previously.25

RNA Extraction

A total of 100 cases, which were a subset of the total cohort were examined for Top IIα mRNA expression in FFPE tissue, as well. The clinicopathological data for this subgroup of patients are shown in Table 1, and is generally comparable to the whole cohort.

All samples included in the study contained at least 60% tumor tissue as evaluated by H&E staining. A 10-μm thick section was cut from each paraffin block and set in a 1.5-ml Eppendorf tube. RNA was extracted using magnetic bead-based technique developed by Siemens Healthcare Diagnostics Research (Köln, Germany) according to a standard protocol provided by the manufacturer. In brief, the FFPE section was deparaffinized in xylol and rehydrated in graded solutions of ethanol. The pellet was let to dry at room temperature for 10 min, lysed in lysis buffer and sodium dodecyl sulfate (SDS) for 5 min at 95 °C; then proteinase K was added and 2-h incubation at 56 °C with shaking was carried out. A binding buffer and the magnetic beads were added, and nucleic acids were allowed to bind to the beads for 15 min at room temperature aided by shaking. On a magnetic rack, the supernatant was removed and beads were washed several times with washing buffers. After the addition of elution buffer and incubation for 10 min at 70 °C with strong shaking, the supernatant was transferred to a new tube without touching the beads on the magnetic rack. This was followed by DNAse I treatment for 30 min at 37 °C, and subsequently the inactivation of DNAse I, using TURBO DNA-free kit according to the procedure provided by the manufacturer (Ambion, Austin, TX, USA); after a short centrifugation, the supernatant was removed carefully, kept at −80 °C and used for real-time qRT-PCR).

Real-Time qRT-PCR

Topoisomerase IIα mRNA expression in a total of 100 FFPE samples was evaluated by real-time qRT-PCR that was performed using intron-spanning TaqMan primers and fluorogenic probes for Top IIα and RPL37A genes (Siemens Healthcare Diagnostics). All the reactions were carried out on a LightCycler instrument (Roche Diagnostics, Mannheim, Germany) using Superscript III Platinum One-Step qRT-PCR System (Invitrogen, Karlsruhe, Germany) in a 20-μl reaction volume. The following conditions were employed: 30 min at 50 °C for reverse transcription, 2 min at 95°C for Taq activation followed by 40 cycles of 95 °C for 15 s and 60°C for 30 s for amplification. No-template controls were included in every reaction run.

The comparative threshold cycle (Ct) method was used for the calculation of amplification fold as specified by the manufacturer. RPL37A was employed as a housekeeping gene at a cycle threshold (Ct) of 25. Top IIα and RPL37A mRNA quantities were analyzed in duplicate, and mean Ct levels were used for data analyses. A normalized score (40-ΔCt) that has a proportional correlation to relative RNA transcriptional level was calculated. In cases where Top IIα mRNA expression could not be measured despite adequate positive controls, the Ct value was estimated as 40, resulting in a normalized score of zero. Using ( 2 40 - Δ C t ) virtual copy numbers were calculated. To separate cases with low and high expressions, the median of virtual copy numbers was used as the cutoff point for Top IIα mRNA expression.

Statistical Evaluation

The statistical significance of the association between Top IIα expression and clinicopathological parameters was assessed by Fisher's exact test or the χ2-test. Mann–Whitney and Kruskal–Wallis tests were used for non-parametric comparisons. The probability of overall and progression-free survival as a function of time was determined by the Kaplan–Meier method, and differences in survival curves were compared by the log-rank test. Multivariate survival analysis was performed using the Cox regression model. Generally, P-values <0.05 were regarded as significant. For the statistical analyses, SPSS software version 15.0 (SPSS Inc., Chicago, IL, USA) was applied.

Results

Clinical and Pathological Characteristics of Patients with Ovarian Tumors

Samples from 133 patients with invasive ovarian carcinomas were investigated for Top IIα immunoreactivity. The distribution of clinicopathological parameters is shown in Table 1. The mean (median) age of patients at surgery was 57.3 (57) years (range: 32–81). A total of 82 (61.7%) specimens were serous carcinomas, 7 (5.2%) were mucinous carcinomas, 17 (12.9%) were endometroid carcinomas, 6 (4.5%) were clear cell carcinomas, 7 (5.2%) were transitional cell carcinomas and 14 (10.5%) were undifferentiated carcinomas. For the statistical analysis, endometrioid, mucinous and transitional cell tumors were grouped into one group (non-serous tumors). Tumors were well differentiated in 22 (16.5%) cases, moderately differentiated in 57 (42.9%) cases and poorly differentiated in 54 (40.6%) cases. The nuclear polymorphism was low in 21 (15.8%) cases, medium in 69 (51.9%) cases and high in 43 (32.3%) cases. Overall survival data were available for all patients; 34 (25.6%) patients had died during follow-up period. Overall survival time ranged from 2 to 83 months with a median (mean) survival time of 36 (35.9) months. Data on disease recurrence were available for 114 (86%) patients. Progression-free survival was defined as the time between diagnosis and the first clinical or pathological evidence of local or distant disease recurrence. Progression-free survival ranged from 4 to 69 months with a median (mean) survival time of 24.0 (26.5) months.

Top IIα Immunostaining in Ovarian Tumors

Nuclear immunoreactivity of Top IIα was observed in 94.7% of ovarian carcinomas (126 out of 133), and was mostly of moderate-to-strong intensity. In most cases, an additional cytoplasmic staining was detected (85.7%, 114 out of 133 cases). Cytoplasmic staining was generally weaker than nuclear signals, and was diffusely distributed among different tumor areas. Stromal cells were negative for Top IIα expression in most cases; however, in some tumors, a nuclear Top IIα expression was observed in stromal fibroblasts. Different staining patterns of Top IIα are shown in Figure 1a–d.

Figure 1
figure 1

Immunohistochemistry. Expression of topoisomerase IIα in ovarian tissues. (a) Weak nuclear staining of normal ovarian surface epithelium. (b) Very faint expression of Top IIα in sparse nuclei of a highly differentiated serous ovarian carcinoma (arrows). (c) Invasive complexes of a moderately differentiated carcinoma reveal moderate nuclear and weak cytoplasmic Top IIα expression. (d) A poorly differentiated ovarian serous carcinoma with moderate-to-strong expression of Top IIα in the majority of the nuclei (arrows). The tumor cell cytoplasm is diffusely stained in moderate intensity (note the nuclear Top IIα expression in stromal fibroblasts (arrowheads)).

Full size image

Top IIα mRNA Expression in Ovarian Carcinomas

For the determination of mRNA expression, 100 cases were studied. Expression of Top IIα mRNA was detectable in 76% of these specimens. The median normalized score was 5.8 (range: 0.1–8.9); the 25th and 75th percentiles were 2.74 and 7.06, respectively. There was no significant difference in Top IIα mRNA expression between low and high Top IIα protein-expressing tumors (data not shown).

Association of Top IIα Expression with Clinicopathological Features and CRM1 Expression

Differences in Top IIα expression (RNA and protein) between different groups of clinicopathological features were investigated by non-parametric tests (Figure 2). As shown in Figure 2a, high Top IIα mRNA levels were significantly associated with higher tumor grade (G1 vs G2–3: P=0.003, Mann–Whitney test). Additionally, grades 2 and 3 tumors showed a trend toward higher levels of Top IIα cytoplasmic protein expression (Figure 2b; P=0.053, Mann–Whitney test). As the Silverberg grading system used for the classification of ovarian carcinomas is on the basis of the three morphological parameters (nuclear polymorphism, mitotic rate and growth pattern), we evaluated the correlation between Top IIα expression and each of these parameters independently. We observed a significant association between Top IIα mRNA expression and mitotic rate (Figure 2c; P=0.001, Kruskal–Wallis test), but there was no correlation with nuclear polymorphism or growth pattern. Advanced stage tumors (II–IV) had higher Top IIα mRNA expression compared with stage I tumors (Figure 2d, P=0.011, Mann–Whitney test).

Figure 2
figure 2

Correlation of topoisomerase IIα expression with clinicopathological variables in ovarian carcinoma. (a) Top IIα mRNA is significantly highly expressed in carcinomas of higher Silverberg grade (P=0.003, Mann–Whitney test). (b) High Silverberg grade tumors show a trend toward higher cytoplasmic Top IIα protein expression (P=0.053, Mann–Whitney test). (c) Top IIα mRNA expression correlates significantly with mitotic rate (no of mitoses/10 high power field; P=0.001, Kruskal–Wallis test). (d) Top IIα mRNA expression correlates significantly with FIGO stage (P=0.011, Mann–Whitney test).

Full size image

There was significant association between nodal stage (pN), and each of nuclear and cytoplasmic protein expression (P=0.030 and P=0.034, respectively; Mann–Whitney test). mRNA expression correlated significantly with tumor stage (pT; P=0.006, Mann–Whitney test). No significant association was observed between RNA or protein expression and patient age or metastasis stage (pM).

In addition to the established clinicopathological parameters, we investigated the link between the expression of Top IIα and CRM1. Cytoplasmic immunostaining of Top IIα was significantly linked to CRM1 expression (P=0.008, Figure 3).

Figure 3
figure 3

Association of protein expression of topoisomerase IIα and CRM1 in ovarian carcinomas. (a) Association of cytoplasmic Top IIα protein expression with CRM1 protein expression in ovarian carcinomas (P=0.026, Fisher's exact test). (b) Correspondence between cytoplasmic immunostaining of Top IIα and CRM1 in the same tumor specimen stained for Top IIα (left) and for CRM1 (right).

Full size image

Top IIα Expression and Patient Survival

Patients whose tumors revealed a high nuclear as well as cytoplasmic Top IIα protein expression showed a shorter overall survival as well as progression-free survival compared to patients with tumors of low expression (Figure 4). In univariate survival analysis, nuclear protein expression of Top IIα was found to be a significant prognostic factor for overall survival (P=0.045), and cytoplasmic protein expression showed a borderline significance for overall survival (P=0.056). A trend toward a reduced progression-free survival (P=0.084) was seen in patients with high cytoplasmic expression of Top IIα protein, whereas there was no significant correlation between nuclear expression of Top IIα protein and progression-free survival. mRNA expression of Top IIα had no significant association with neither overall nor progression-free survival (Figure 4).

Figure 4
figure 4

Univariate survival analysis according to nuclear or cytoplasmic protein expression of topoisomerase IIα in patients with ovarian cancer. Nuclear Top IIα expression is a prognostic factor for overall survival (a; P=0.045, log-rank test). A trend toward a reduced overall (b; P=0.056) as well as progression-free survival (c; P=0.084) in patients whose tumors express Top IIα in the cytoplasm is evident. Solid lines: cases with high expression; dotted lines: cases with low expression; +: censored.

Full size image

The known prognostic factors in ovarian carcinoma, such as lymph node stage, histological grade, and age at diagnosis, in addition to intraoperative residual tumor were prognostic factors for overall survival in our study group, as well (Table 2). For progression-free survival, FIGO stage, histological type, tumor stage, lymph node stage, tumor grade and residual tumor were significant prognostic factors (data not shown).

Table 2 Univariate survival analysis (Kaplan–Meier)
Full size table

Discussion

In this study, we investigated the expression of topoisomerase IIα in primary invasive ovarian carcinomas at the protein and mRNA levels. Our results show an increased expression of Top IIα in higher grade and advanced stage tumors as well as a correlation of high expression with worse survival. Furthermore, to the best of our knowledge, for the first time, we report a link between cytoplasmic expression of Top IIα and CRM1 in human ovarian tumors.

We have undertaken an approach that facilitates the evaluation of both mRNA and protein expression from the same routinely processed FFPE tissue blocks used in pathology laboratories for diagnostic purposes. We were thus able to evaluate gene expression in a large clinicopathologically characterized ovarian cancer cohort at the transcriptional and translational levels retrospectively without the requirement of special processing or preservation of specimens, as opposed to earlier studies on Top IIα in ovarian cancer surgical specimens, which required frozen tissue samples for the determination of Top IIα protein and RNA expression.13, 14, 26, 27, 28 We employed a new technique for RNA extraction from FFPE tissues, which is based on magnetic beads separation of nucleic acids, and used the sensitive real-time qRT-PCR method with specific TaqMan primer/probe to assess the transcriptional level of the Top IIα gene. Using this new technique, we showed that the determination of mRNA expression helps to identify aggressive tumors (high grade and high stage), and was linked to the mitotic activity of the tumor cells, yet does not add prognostic information compared with the evaluation of Top IIα protein expression. Interestingly, mRNA expression was not related to the protein expression levels, which could point to a post-translational regulation of the Top IIα. This post-translational regulation might be related to the interesting staining patterns (nuclear and cytoplasmic), which also suggest that other proteins might interact with Top IIα protein and modify its cellular localization and function.

An interesting finding of our study was the detection of a distinct cytoplasmic localization of Top IIα in a subset of tumors. Several cell culture experiments have pointed out to a correlation of cytoplasmic trafficking of Top IIα with acquired drug resistance to Top II poisons in cell lines under drug selection, or in patient samples obtained after exposure to chemotherapy.29, 30

Recently, de Lucio et al31 has characterized a human lung cancer cell line with an intrinsic resistance to etoposide that was mediated by cytoplasmic localization of topoisomerase IIα. Cytoplasmic Top IIα expression was further found to be associated with increased resistance, and hence poor survival in leukemia in children treated with a regimen containing a Top II poison.32 Resistance to Top II inhibitors could thus be due to an ectopic cytoplasmic localization of Top IIα, leading to the decreased formation of enzyme/DNA complexes in the nucleus, which could be targeted by the Top IIα inhibitors, and by this mechanism reducing the sensitivity of the cell to the drug.

Interestingly, in our study, we observed cytoplasmic expression of Top IIα in samples from patients who have not received any chemotherapy before collection of tumor specimens.

We found a significant correlation between Top IIα and CRM1 cytoplasmic proteins. In earlier cell culture studies, it was reported that nuclear export of Top IIα is mediated by CRM1, a mechanism, which could be blocked by the CRM1 inhibitor lepotomycin B (LMB).33, 34, 35 Our data reveal an in vivo implication of nuclear export mechanisms in Top IIα cytoplasmic localization.

We observed that Top IIα expression has a significant association with stage; this finding is similar to the results reported by van der Zee et al.26 The finding of an association of increased expression of Top IIα with poor differentiation is in line with the findings in ovarian cancer,26 and in prostate tumors.15, 16

Similar to our finding of a significant inverse correlation of Top IIα expression and survival, other groups have detected an unfavorable prognostic impact of Top IIα expression in ovarian cancer as well as in other series of neoplasms, such as breast cancer, synovial sarcoma, lung cancer and endometrial carcinoma.20, 36, 37, 38, 39, 40 One explanation for the shorter survival rate associated with elevated Top IIα levels could be an enhancement of tumor cell proliferation, which results in an increased tumor aggressiveness.6, 7, 41 On the other hand, high Top IIα expression might indicate tumor sensitivity to Top IIα inhibitors, as it reflects target availability. Several authors had shown a direct correlation between the sensitivity of a cell to Top IIα targeting drugs and the level of Top IIα,42, 43, 44 and in vitro chemosensitivity testing of ovarian carcinoma specimens has provided evidence that the Top IIα expression could reflect chemosensitivity to Top II-targeting drugs.45 Top IIα expression could thus be used as a predictor of response to chemotherapy. Patients with increased Top IIα expression might benefit more from chemotherapeutic protocols involving Top II-targeted drugs; hence their outcome could be improved. Our finding is in accordance with other studies in ovarian cancer patients treated with platinum-based regimens.13, 40 Yet, due to its role as a marker for an active cell cycle, Top IIα expression might be an indicator for chemoresponse in general as the latter is strongly dependent on proliferation.

In conclusion, Top IIα expression showed prognostic importance as shorter survival was associated with increased Top IIα expression. On the basis of our results, we suggest Top IIα as a candidate prognostic factor for ovarian cancer, which should be evaluated in large-scale clinical trials. The determination of the immunoreactive pattern of Top IIα expression, together with other clinicopathological factors, might help to identify high-risk patients, and to plan an individualized therapy. In the future, it would further be interesting to characterize the potential role of Top IIα as a predictor of tumor response to chemotherapy in prospective studies on ovarian cancer patients, including samples obtained pre- and post-chemotherapeutic treatment, especially with Top IIα inhibiotors, such as pegylated doxorubicin and etoposide.