T-LAK Cell-Originated Protein Kinase (TOPK) as a Prognostic Factor and a Potential Therapeutic Target in Ovarian Cancer

Ovarian cancer is the fifth leading cause of cancer-related death in women (1). Diagnosis of this malignancy tends to be delayed because its symptoms do not become overt until tumors get large and these symptoms often mimic other pathological conditions. In fact, more than 60% of patients with ovarian cancer are diagnosed at an advanced stage, and their 10-year survival rate was reported as low as 21% (1, 2). Peritoneal dissemination is the most common progression pattern of ovarian cancer. Because the dissemination spreads broadly in abdomen, complete surgical resection is not applicable for ovarian cancer patients with dissemination (3). Platinum agents are widely used to treat patients with advanced stages of ovarian cancer, and its response rate is known to be more than 70% (4). However, the majority of the patients subsequently and rapidly experiences relapse of the disease that is generally incurable (5).

TOPK (T-lymphokine–activated killer cell-originated protein kinase, also known as PBK or PDZ-binding kinase) is a Ser/Thr protein kinase and highly activated during cell mitosis (6). TOPK was suggested as a promising molecular target for the development of cancer therapy, because it was highly transactivated in various types of cancer while its expression was hardly detectable in normal tissues except the testis and fetal tissues (7). Its high expression was shown to be correlated with a more aggressive phenotype in several types of cancers, such as breast, lung, prostate, and colorectal cancers as well as Ewing sarcoma (8–13).

Through high-throughput compound library screening and subsequent extensive structure–activity relationship studies, OTS514 and its derivative OTS964 were developed as potent TOPK inhibitors (14). We recently reported that liposomal formulation of OTS964 could achieve complete regression of tumors in the human lung cancer xenograft model (15) and that treatment with OTS514 significantly elongated overall survival in a murine engraft model of acute myeloid leukemia (16).

In our best knowledge, no literature describing a role of TOPK in ovarian cancer has been reported. In this study, we describe the clinical significance of TOPK and explore therapeutic efficacy of TOPK inhibitors, OTS514 and OTS964, in ovarian cancer.

Tumor tissues for immunohistochemistry were obtained by surgical resection from 189 patients, including 26 cases of borderline tumor and 163 cases of epithelial ovarian cancer (EOC). For primary cell culture, a total of 34 samples were obtained from 28 patients with ovarian cancer. Among them, three samples were obtained from malignant ascites in 3 patients with recurrence. Of 31 remaining samples from nonrecurrence cases, 14 were obtained from ovarian tumors, 4 from omental tumors, and 13 from malignant ascites, including 6 cases which tumor cells were obtained from multiple sites (e.g., collected from ovarian and omental tumors from 1 patient). All the patients who participated in this study provided written informed consent for the collection and research use of their materials, and the use of these samples was approved by the appropriate institutional ethics committees. All the patients underwent surgical procedure and were diagnosed with the International Federation of Gynecology and Obstetrics (FIGO) staging criteria.

We investigated the TOPK and Ki-67 expression on formalin-fixed paraffin-embedded (FFPE) specimens of 163 EOCs and 26 borderline tumors using anti-TOPK antibody (BD Transduction Laboratories) and anti-Ki-67 antibody (Dako). Immunohistochemical staining was performed on full section slides using Histofine Simple Stain MAX PO (Nichirei Bioscience Inc.) and DAB⁺, Liquid (DAKO) according to the manufacturer's instructions. Tumor tissue sections were deparaffinized, hydrated, and treated with Target Retrieval Solution pH9 (Dako), followed by overnight incubation with the monoclonal anti-TOPK antibody (1:100 dilution) or anti-Ki67 antibody (1:400 dilution) at 4°C. Positive staining levels of TOPK were classified according the positivity in tumor cells as 0 with <1% positive tumor cells, 1+ with 1% to 25% positive tumor cells, 2+ with 25% to 50% positive tumor cells, and 3+ with ≥50% positive tumor cells. Similarly, Ki-67 staining was scored and more than 50% of staining was defined as strong expression.

Human ovarian cancer cell lines, ES-2, OV90, PA-1, SKOV3, SW626, A2780, and CaOV3 were purchased from American Type Culture Collection (Rockville, MD). OVSAHO, OVTOKO and RMG-I were purchased from Japan Collection of Research Bioresources Cell Bank (Ibaraki, Osaka). OVCAR3 was provided from Tohoku University. Cells were cultured under the recommendations of their respective depositors. All cell lines were authenticated by STR analysis.

Expression levels of TOPK protein and FOXM1 mRNA were examined by Western blot and RT-PCR, respectively, as described previously (17, 18). Anti-TOPK antibody (BD Biosciences) and anti-β-actin antibody (Sigma-Aldrich) were used as primary antibodies. Regarding RT-PCR, total RNAs were extracted from cell lines using RNeasy Mini Kit (Qiagen) according to the manufacturer's directions, then reversely transcribed using SuperScrive III First Strand Synthesis System (Invitrogen) following the manufacturer's instruction. RT-PCR was performed using ViiA 7 system (Life Technology). The primer sequences were 5′-CAACCGCTACTTGACATTGGA-3′ and 5′-TCACCGGGAACTGGATAGG-3′ for FOXM1, and 5′-CGACCACTTTGTCAAGCTCA-3′ and 5′-GGTTGAGCACAGGGTACTTTATT-3′ for GAPDH. mRNA levels of FOXM1 were normalized by those of GAPDH.

In vitro cell viability assay was performed as described previously (15). Briefly, ovarian cancer cells were plated in 96-well plates (100 μL each) at a density that is expected to show linear growth (ES-2, 1,000 cells; OVCAR3, 6,000 cells; OV90, 4,000 cells; OVSAHO, 6,000 cells; OVTOKO, 2,000 cells; PA-1, 2,000 cells; RMG-1, 6,000 cells; SKOV3, 2,000 cells; SW626, 4,000 cells; A2780, 3,000 cells; and CaOV3, 5,000 cells). The cells were allowed to adhere overnight before an exposure to compounds for 72 hours at 37°C. For methyl thiazolyl tetrazolium (MTT) assay, cell counting kit-8 (Dojindo Molecular Technologies, Inc.) was added in the 96-well plate which was read at 450 nm of wavelength using the iMark microplate absorbance reader (Bio-Rad) after reaction for 1 hour. All assays were carried out in triplicate.

ES-2 cells (2 × 10⁶ cells) were injected intraperitoneally into the BALB/cSLC-nu/nu mice (Japan SLC Inc.). After 5 days of injection, 75 mice were randomized into 3 groups (each 25 mice) for vehicle, 25 mg/kg of OTS514 administration, and 50 mg/kg of OTS514 administration. OTS514 compound was prepared in a vehicle of 0.5% methylcellulose and given by oral gavage for 14 days (25 mg/kg of OTS514) or 11 days (50 mg/kg of OTS514). The weight and abdominal circumference were monitored as an indicator of tolerability for OTS514 and tumor aggressiveness.

Surgical tissue samples were cut into pieces of less than 1 mm, and then gently pipetted with prewarmed 0.25% of trypsin-EDTA for 3 minutes. The tissues were resuspended with 10 mL of cold HBSS with 2% of FBS. After centrifugation for 5 minutes at 1,500 rpm, the supernatant was removed as much as possible. Subsequently, samples were treated with 2 mL of prewarmed dispase (STEMCELL Technologies) and 200 μL of DNase I (STEMCELL Technologies), followed by gentle pipetting for 2 minutes, resuspending with 10 mL of HBSS with 2% of FBS, and filtration through 100-μm, 70-μm, and 40-μm cell strainers (BD Falcon). Then, single suspension cells were centrifuged for 5 minutes at 1,500 rpm and counted. Finally, ovarian cancer cells were magnetically isolated from cells using the EasySep Human EpCAM Positive Selection Kit (STEMCELL Technologies) according to the manufacturer's instructions.

Freshly isolated primary ovarian cancer cells were seeded into 96-well collagen coated plates (5 × 10⁴ cells/well) and incubated with TOPK inhibitors at different concentrations (0, 1, 10, and 100 nmol/L) for 72 hours. The cell proliferation activity was measured using the CellTiter 96 AQueous One Solution Cell Proliferation Assay (Promega) according to the manufacturer's instructions. Similarly, the cells were measured for the cytotoxicity every 24 hours using the CellTox Green Cytotoxicity Assay (Promega) according to the manufacturer's instructions. All assays were done in triplicate.

The correlation of TOPK expression levels and clinicopathological characteristics were analyzed using the Fisher exact test, and prognostic values were analyzed by the Kaplan–Meier method and the log-rank test. The statistical analysis of the in vivo study was conducted using the Kaplan–Meier method and statistically analyzed with the log-rank test. Comparison of expression levels of TOPK or FOXM1 and growth-inhibitory effects/cytotoxicity of TOPK inhibitors on primary cancer cells were analyzed by the t-test (one-way). Statistical analyses were performed using JMP v11 (SAS) and GraphPad Prism 6. In all tests, differences were considered to be significant at P < 0.05.

We first examined TOPK expression levels in normal ovarian epithelial cells and cancer cells using The Cancer Genome Atlas database (19) and confirmed its high level of transactivation in ovarian cancer cells (Supplementary Fig. S1a). Another dataset (20) showed that TOPK expression was very low or absent in normal organs, including fallopian tubes, except for testis (Supplementary Fig. S1b). Then, we analyzed TOPK protein expression by immunohistochemical staining in 26 borderline tumors and 163 EOCs. The expression levels of TOPK were scored as 0, 1+, 2+, or 3+ according the definition described in the Materials and Methods section, and we classified the tumors into two groups, a low TOPK expression group (0 and 1+) and a high TOPK expression group (2+ and 3+; Fig. 1A). As shown in Table 1, high TOPK expression was observed in 84 (52%) of 163 EOCs examined, although its elevated expression was observed in only 2 (7%) of 26 borderline tumors (P < 0.001). We then examined correlation of TOPK expression levels with several clinicopathological variables, and found the significant association of its high expression with advanced stages (P = 0.010). We also found that significant correlation between Ki-67 and TOPK expression levels (P = 0.017). Expression of TOPK was relatively higher in serous and endometrioid adenocarcinomas than in the other subtypes, such as clear and mucinous adenocarcinomas (Supplementary Fig. S2a). We subsequently compared TOPK expression level with prognosis of ovarian cancer patients and identified that high TOPK expression was significantly associated with poor prognosis of the patients in progression-free survival (PFS; P = 0.017; Fig. 1B). High expression of TOPK was also tended to be associated with shorter overall survival (OS; P = 0.083). The association of TOPK expression and PFS/OS was more obvious when we analyzed patients at early stages (PFS, P = 0.008; OS, P = 0.006; Fig. 1C). On the other hand, no difference between TOPK high and low groups was observed when we analyzed patients at advanced stages (PFS, P = 0.44; OS, P = 0.16; Fig. 1D). mRNA expression level of TOPK using the Kaplan–Meier plotter (21) also showed the significantly shorter PFS in early-stage ovarian cancers with highly TOPK mRNA (P = 0.04), but not in advanced-stage cases (P = 0.97; Supplementary Fig. S2b). The reason is probably that many other progressive factors are also activated in advanced ovarian tumors.

To evaluate therapeutic potential of TOPK inhibitors, we examined expression levels of TOPK in 11 ovarian cancer cell lines. Our immunoblot results revealed that TOPK was very highly expressed in 9 of 11 ovarian cancer cell lines (Fig. 2A). We then tested in vitro growth-inhibitory effects of two TOPK inhibitors, OTS514 and OTS964, by measuring the IC₅₀ values of 11 ovarian cancer cell lines (Fig. 2B and C; Supplementary Figs. S3 and S4). IC₅₀ values of OTS514 for these cancer cell lines were very low as 3.0 (CaOV3) to 46 nmol/L (OVTOKO) and those of OTS964 were between 14 (CaOV3) to 110 nmol/L (RMG-I). Protein levels of TOPK in both ES-2 and SKOV3 cell lines were significantly reduced after 48-hour treatment of OTS514 or OTS964 at IC₅₀ concentrations (Fig. 2D). We also detected reduction of the FOXM1 mRNA level in both ES-2 and SKOV3 cell lines treated in the same condition (Fig. 2E).

We further examined the in vivo efficacy of OTS514 using an ovarian cancer peritoneal dissemination xenograft model which was established by the ES-2 ovarian cancer cell line. Treatment dose and duration were determined on the basis of pharmacokinetic data for orally administrated OTS514 in mice, as shown in Supplementary Fig. S5. Five days after intraperitoneal injection of ES-2 cells into abdominal cavity, 75 mice were randomized into 3 groups (each 25 mice) according to administration of vehicle, 25 mg/kg of OTS514, or 50 mg/kg of OTS514. Vehicle control or OTS514 (25 mg/kg or 50 mg/kg) was orally administered every day. The treatment duration was planned for 14 days (from day 5 to day 18) for both groups. However, because one mouse in the 50 mg/kg group died with >20% body weight loss and one showed the >20% body weight loss, we decided to terminate the administration at day 15 (although the average body weight in this group was 0.97, compared with that at the beginning of the treatment). We sacrificed 5 mice each from 3 groups at day 15 and confirmed peritoneal dissemination in all 15 mice, but ascitic fluid was not obvious in all of the 10 mice treated with OTS514 administrated groups (Fig. 3A). As shown in Figure 3B, overall survivals of the two groups treated with OTS514 were significantly improved compared with that of the vehicle group (25 mg/kg, P < 0.001; 50 mg/kg, P = 0.011). The ranges of relative median body weight in each group were 1.0 to 1.24 in vehicle, 1.0 to 1.15 in 25 mg/kg group, and 0.97 to 1.21 in the 50 mg/kg group (Fig. 3C). We sacrificed and examined 6 mice, which survived until day 32, among 40 mice treated with the OTS514 and, interestingly, found no evidence of tumor in their abdomen. As shown in Figure 3D, all the 18 dead mice in the vehicle treatment group showed visually enlarged abdomen and increased body weight. Among 34 dead mice with OTS514 treatment, 13 mice (treated with 25 mg/kg) and 12 mice (treated with 50 mg/kg) showed visible enlargement of abdomen (due to the increase of ascites) before death.

We further examined ex vivo growth-inhibitory and cytotoxic effects of TOPK inhibitors (OTS514 and OTS964) against ovarian cancer cells that were freshly isolated from ovarian cancer patients as described in Materials and Methods. We examined cell viability of 31 primary ovarian cancer cells after treatment with various concentrations of OTS514 or OTS964 for 3 days. As shown in Figure 4A, strong growth-inhibitory effects were observed after treatment with 10 nmol/L and 100 nmol/L of OTS514. Figure 4B and Supplementary Fig. S6A summarized the cytotoxic effects of OTS514 and OTS964, respectively. Cytotoxicity against primary cancer cells with treatment of 100 nmol/L of OTS514 (P < 0.001) or OTS964 (P < 0.001) was confirmed. We also compared the sensitivity of cancer cells with TOPK inhibitors according to origins (cancer cells in ovary, omentum, or malignant ascites), but found no difference in the sensitivity by the locations (Fig. 4C; Supplementary Fig. S6b). In addition, we compared the sensitivity to TOPK inhibitors according to histological types, because clear-cell carcinomas are usually resistant to the chemotherapeutic agents. We observed no significant differences in the sensitivity to these drugs among different histological types of ovarian cancer (Fig. 4D; Supplementary Fig. S6c), but observed strong cytotoxic effects of OTS514 on cells derived from malignant ascites samples from three intensively pretreated and recurrent ovarian cancer patients (10 nmol/L, P = 0.045; 100 nmol/L, P = 0.003; Fig. 4E). To further validate the cytotoxic effect of TOPK inhibitors, we monitored changes in membrane integrity that occurred as a result of cell death. With a fluorescence-based assay, significant ex vivo cytotoxicity was observed at day 3 in cells treated with 100 nmol/L of OTS514 (Supplementary Fig. S7).

In this study, we investigated a potential of TOPK as a promising molecular target for the development of drugs to treat ovarian cancer and demonstrated (i) significant association between high TOPK expression and poor prognosis of ovarian cancer patients; (ii) strong cytotoxic effects of TOPK inhibitors, OTS514 and OTS964, against ovarian cancer cell lines; (iii) in vivo efficacy of OTS514 using ovarian cancer peritoneal dissemination xenograft model; and (iv) efficacy of these two compounds to patient-derived primary ovarian cancer cells.

Cancer therapies such as radiation and chemotherapy are not only targeting cancer cells but also damage on rapidly proliferating normal cells and cause severe adverse reactions in cancer patients (22). Recently, molecular-targeted therapy or immunotherapy using antibodies has been developed as new treatment modalities. However, some of these approaches still cause on-target adverse reactions because targeted molecules do not have the high specificity to cancer cells and are often expressed in normal cells (23). Because it was very highly upregulated in ovarian cancer cells and its expression was hardly detectable in normal ovary and fallopian tube as well as other normal organs except the testis, TOPK is considered as a promising molecular target for the development of therapeutics for ovarian cancer. The TOPK locus was shown to have 59% heterozygous loss and 6.5% homozygous loss in ovarian cancer in the TCGA database (24, 25). However, chromosome 8p21, where this gene is located, was indicated to contain a tumor suppressor gene that was implied by frequent LOH (nearly 50%). Hence, because TOPK is transactivated without any activating genetic alterations in most of the cases, we assume that this chromosomal loss is biologically not so significant.

In this study, we first confirmed that TOPK was highly and frequently transactivated in ovarian epithelial carcinomas but less frequently in borderline malignancy tumors. In addition, high TOPK expression was significantly associated with poor prognosis of ovarian cancer patients, particularly those at early stages, indicating that TOPK may be a strong indicator for the aggressive phenotype in the early-stage ovarian cancer. Interestingly, we have not observed any difference in the prognosis between the TOPK-high and TOPK-low groups when we analyzed advanced-stage ovarian cancer patients, probably because the number of genetic alterations generally increases during cancer progression (26, 27) and then multiple progressive factors other than TOPK may diminish the clinical significance of TOPK in advanced EOC cases. In fact, some genes, which were expressed specifically and highly in metastatic ovarian tumors, were suggested as independent predictors of poor survival in advanced ovarian cancer patients (28).

TOPK is classified into the MAPK pathway and enhances cell migration by modulating a PI3K/PTEN/AKT-dependent signaling pathway (29). In addition, we previously reported that TOPK is highly activated during cell mitosis and indispensable in cancer cell proliferation (7, 17). The results in our previous studies (15) are concordant with immunohistochemical analysis in this study that showed a significant correlation between Ki-67 and TOPK expression. Our immunoblot results clearly revealed high TOPK expression in most of ovarian cancer cell lines that are sensitive to TOPK inhibitors. Generally, IC₅₀ values of OTS514 and OTS964 were correlated in most of the cell lines that we examined. However, some of them, such as OVCAR3 and A2870, showed relatively less sensitivity to OTS514 but showed enhanced sensitivity to OTS964. One possible mechanism is that OTS964 is an N,N-dimethylated derivative of OTS514, and this dimethylation modification might reduce efflux of the compound in some cancer cells and therefore could increase its intracellular concentration, as exampled by other compounds (30, 31).

One of important downstream targets of TOPK is FOXM1 that is an oncogenic transcription factor, overexpressed in 87% of serous ovarian cancer, and associated with cell proliferation and poor prognosis in ovarian cancer (19, 32, 33). Our previous studies showed that inhibition of TOPK kinase activity reduced stability of TOPK protein itself through the downregulation of autophosphorylation and resulted in reduction of the FOXM1 transcriptional level (11, 18). Similarly, in ES-2 and SKOV3 ovarian cancer cell lines, we found that TOPK inhibitors reduced the TOPK protein level and suppressed transcription of FOXM1, which probably led to the growth inhibition of ovarian cancer cells.

We also demonstrated the in vivo efficacy of a TOPK inhibitor in a peritoneal dissemination mouse model that recapitulates advanced ovarian cancers. We chose an ES-2 cell line that has the most aggressive characteristics among the intraperitoneal xenograft models and showed 16 days of median survival after tumor cell transplantation (34). Even in this aggressive xenograft model, we could show significantly elongated survival in mice that were orally administered with 25 mg/kg or 50 mg/kg of OTS514. In addition, we sacrificed 6 mice that survived until day 32, among 40 mice treated with OTS514 and found no evidence of tumor in their abdomen, indicating complete regression of tumors as we previously observed in lung cancer xenograft models (15). For adverse events, because we did not perform blood test in this study, we are unable to address a possibility of leukocytopenia-related death in OTS514-treated mice, particularly 2 mice that showed significant weight loss, as our previous xenograft models showed (15). Finally, our ex vivo results using the patient-derived ovarian cancer cells strongly suggested that TOPK inhibitors would be very effective to suppress the growth of ovarian cancer cells regardless of the sites of tumors or the histological subtypes.

In conclusion, high TOPK expression is strongly associated with poor prognosis in ovarian cancer patients and our in vitro, in vivo, and ex vivo results suggest the therapeutic potential of TOPK inhibitors that can be immediately translated into clinical evaluation.

J. Park is a consultant/advisory board member for OncoTherapy Science. Y. Nakamura is an employee of, has ownership interest (including patents), is a consultant/advisory board member for, and reports receiving commercial research grants from OncoTherapy Science. No potential conflicts of interest were disclosed by the other authors.

Conception and design: Y. Ikeda, J.-H. Park, Y. Nakamura, K. Hasegawa

Development of methodology: T. Miyamoto, Y. Imai

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): T. Miyamoto, N. Takamatsu, T. Kato, Y. Imai, K. Hasegawa

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): N. Takamatsu, A. Iwasa, S. Okabe, Y. Nakamura, K. Hasegawa

Writing, review, and/or revision of the manuscript: Y. Ikeda, J.-H. Park, T. Miyamoto, T. Kato, K. Fujiwara, Y. Nakamura, K. Hasegawa

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): T. Miyamoto, S. Okabe, Y. Imai, K. Fujiwara, K. Hasegawa

Study supervision: Y. Nakamura, K. Hasegawa

We thank Mrs. Yuko Ijima, Mrs. Akiko Miyara, and Dr. Akira Kurosaki for their excellent technical assistance and contribution to sample collection, and Mrs. Akiko Taira and Mrs. Kazue Toyota for their contribution to the in vivo and in vitro study.

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.