Chemoresistance to platinums, such as cisplatin, is of critical concern in the treatment of ovarian cancer. Recent evidence has linked epithelial–mesenchymal transition (EMT) as a contributing mechanism. The current study explored the connection between cellular responses to cisplatin and EMT in ovarian cancer. Expression microarrays were utilized to estimate the EMT status as a binary phenotype, and the transcriptional responses of 46 ovarian cancer cell lines to cisplatin were measured at dosages equivalent to 50% growth inhibition. Phenotypic responses to cisplatin were quantified with respect to cell number, proliferation rate and apoptosis, and then compared with the epithelial or mesenchymal status. Ovarian cancer cell lines with an epithelial status exhibited higher resistance to cisplatin treatment in the MTS assay than those with a mesenchymal status. Pathway analyses revealed the induction of G1/S- and S-phase genes (P=0.001) and the activation of multiple NF-κB (nuclear factor kappa-light-chain-enhancer of activated B cells) downstream genes (P=0.0016) by cisplatin selectively in epithelial-like cell lines. BrdU incorporation and Caspase-3/7 release assays confirmed impaired apoptosis in epithelial-like ovarian cancer cells. In clinical samples, we observed resistance to single platinum treatment and the selective activation of the NF-κB pathway by platinum in ovarian cancers with an epithelial status. Overall, our results suggest that, in epithelial-like ovarian cancer cells, NF-κB activation by cisplatin may lead to defective apoptosis, preferential proliferation arrest and a consequential decreased sensitivity to cisplatin.
Ovarian cancer is the most lethal gynaecological cancer, with 2013 statistics indicating 14 030 estimated deaths and 22 240 estimated new cases in the United States alone.1 In ovarian cancer therapeutics, platinum complexes have a central role as the first-line treatment option and are usually administered in combination with taxanes. Although ovarian cancer is a relatively chemosensitive disease, ~20–30% of patients demonstrate resistance to the platinum-based chemotherapy.2,3 Moreover, even after there has been efficient clearance of the tumour cells in response to standard therapy, many patients (70~90%) suffer from relapse within a window of months to years and the relapsed tumours typically acquire resistance to platinum.2, 3, 4 Thus, in order to improve patient outcomes, it is critical to overcome platinum resistance in ovarian cancer cells.
Cisplatin (cis-diamminedichloroplatinum [II]) is a platinum compound that forms DNA adducts, which are subsequently involved in the activation of various signal transduction pathways involved in DNA damage recognition, DNA repair, cell cycle arrest and apoptosis. Earlier studies have identified two broad mechanisms leading to platinum resistance in cancer cells: the first, a reduction in the formation of platinum–DNA adducts; the second, an impairment in the apoptotic response of cells to adduct products. The former includes decreased uptake, increased detoxification and export of platinum, and elevated DNA repair, whereas the latter can occur as a consequence of activated anti-apoptotic pathways after platinum treatment, exemplified by an increase in the activation of NF-κB (nuclear factor kappa-light-chain-enhancer of activated B cells), mitogen-activated protein kinase and Akt (otherwise referred to as Protein Kinase B).5, 6, 7, 8
Further complicating these mechanisms of platinum resistance, growing evidence has suggested an association between epithelial–mesenchymal transition (EMT) and cellular resistance to many cytotoxic reagents, including cisplatin.9, 10, 11, 12 EMT originally describes a transitional, biological process, during which polarized epithelial cells lose their epithelial architecture and are transformed into motile mesenchymal cells with an increased migratory and invasive capacity. This process is reported to be associated with cancer cell dissemination and invasion, evasion of apoptosis and drug resistance in several cancer types.9, 10, 11, 12 Recent studies have identified EMT not only as a cellular dynamic process (‘acquired’ EMT),13, 14, 15, 16 but also as an inherent nature of the cell (‘inherent’ EMT), which forms a spectrum ranging from epithelium to mesenchyme as a function of the expression level of EMT markers among clinical samples or cell lines, including ovarian cancer.17, 18, 19, 20 Multiple previous observations have implied a link between EMT (typically acquired EMT) and drug resistance in various cancers including ovarian,13, 14, 15,21, 22, 23, 24 breast25,26 and colorectal cancers.27 Upon single or multiple rounds of cisplatin or carboplatin treatment, ovarian cancer cells acquire not only resistance to platinum but also mesenchymal phenotypes.13, 14, 15,21 Recently, an expression profiling study of a panel of pancreatic cancer cell lines revealed that cell lines with epithelial characteristics are more sensitive to treatment with gemcitabine, 5-fluoro-uracil and cisplatin in comparison to cell lines with mesenchymal properties.18 Similarly, non-small cell lung carcinoma cell lines with mesenchymal properties show higher resistance to epidermal growth factor receptor and PI3K/Akt pathway inhibitors.17
In this study, we investigated the cellular responses to cisplatin, particularly in the context of EMT, using expression microarray analyses of a panel of ovarian cancer cell lines. The transcriptomic analyses revealed distinct responses between epithelial- and mesenchymal-like ovarian cell lines. Epithelial-like cell lines demonstrated preferential proliferation arrest and NF-κB activation by cisplatin, and these responses coincided with their cisplatin-resistant phenotype. In vivo ovarian cancers with epithelial status also exhibited a higher degree of resistance to and more NF-κB activation by single-agent platinum treatment than those with mesenchymal status.
Differential responses induced by cisplatin treatment in epithelial- and mesenchymal-like ovarian cancer
To investigate the cellular responses of ovarian cancer cells to cisplatin, we adopted the experimental strategy delineated in Figure 1a. Substantial variation in cisplatin sensitivity was observed across 46 ovarian cell lines (1.45–201.9 μM; 139.2-fold difference; Supplementary Table 1), which hampers a meaningful measurement of the biological effects of the drug on cells at a same concentration. Therefore, we performed transcriptomic analyses using cisplatin at the GI50 dosage (50% growth inhibitory concentration; that is, the cisplatin dosage required to cause a 50% reduction in the increase in viable cell number over 48 h as compared with untreated control cells). The use of GI50 measurements allowed us to conduct a fair comparison of cellular responses across all cell lines under biologically comparable conditions. In addition, the cisplatin-treated cells at the GI50 dosage are still viable; thus, the response is measurable, even in the presence of cisplatin.28 Each of the 46 ovarian cancer cell lines was treated with cisplatin at the GI50 dosage, with sham treatment serving as a negative control. According to a previous pharmacokinetics study, the peak cisplatin concentration in plasma was 12.99±4.68 μM after intravenous administration of a standard dosage of cisplatin (100 mg/m2).29 Therefore, the dosage applied in the current transcriptomic analyses was not a perfect reflection of the in vivo condition. RNA from sham- or cisplatin-treated cells was then extracted and subjected to microarray assays. The preferential activation of p53 downstream genes in TP53 wild-type cells as a response to cisplatin treatment (Figure 1b, Mann–Whitney U-test, P=0.0359; Supplementary Figures 1a–c and 2) demonstrated the validity of our experimental approach (see Supplementary Information for more details).
Using the epithelial and mesenchymal expression signatures from a previous study (Supplementary Table 2),30 single-sample gene set enrichment analysis was performed on the expression data of each untreated cell line, which reflects the innate nature of cell line under standard culture conditions. The resulting enrichment scores (ESs) for epithelial and mesenchymal gene sets were then used to subdivide the 46 ovarian cell lines into 23 epithelial- and 23 mesenchymal-like cell lines (Figure 1c; Supplementary Table 1; see ‘Materials and Methods’). This allowed us to then investigate the pathways differentially regulated by cisplatin in epithelial- and mesenchymal-like ovarian cell lines. It is important to note that this EMT status is a distinct concept from molecular subtype, which was recently reported by three research groups.30, 31, 32 The relationship between EMT status and molecular subtyping in the cell line panel is shown in Supplementary Table 1 and Supplementary Figure 3.
Cellular responses to cisplatin were determined by subtracting the gene expression values of sham-treated cells from those of cisplatin-treated cells. Relying on the EMT status categorization, we detected differential responses between epithelial- and mesenchymal-like cells at the gene level, with 348 and 132 specific gene expression changes for epithelial- and mesenchymal-like cells, respectively (Mann–Whitney U-test, cut-off value of P<0.01; Supplementary Table 3). These findings suggested a differential pattern of pathway alteration by cisplatin.
To examine whether this differential pattern could be also observed at the pathway level, single-sample gene set enrichment analysis was again utilized to compute ESs to the expression data of sham- or cisplatin-treated cell lines using 6769 gene sets (MSigDB version 3.0; http://www.broadinstitute.org/gsea/downloads.jsp). Cisplatin-induced responses were then calculated by subtracting the ESs of sham-treated cells from that of cisplatin-treated cells. Consistent with our observations at the gene level, ovarian cancer cell lines with epithelial properties showed a significant upregulation of 267 pathways, whereas only 27 pathways demonstrated increased activities for mesenchymal-like cell lines (Figure 1d; Supplementary Table 4; Mann–Whitney U-test, cut-off values of P<0.01). In particular, we identified an upregulation of DNA checkpoint-related gene sets (including BIOCARTA ATM pathway and REACTOME activation of ATR in response to replication stress) and cell cycle-related gene sets (such as KEGG DNA replication, BIOCARTA cell cycle pathway and REACTOME G1/S transition) in epithelial-like cells; these results implicate the activation of cell cycle arrest by cisplatin. NF-κB pathway-related gene sets (exemplified by apoptosis via NF-κB pathway and tumour necrosis factor signalling via NF-κB) were also activated in epithelial-like cells. Of note, hypergeometric tests indicated the same significant enrichment of cell cycle- and NF-κB-related pathways in epithelial-like cells, with enrichment of 24 cell cycle-related gene sets from a total of 70 gene sets and 7 out of 14 NF-κB-related gene sets (Figure 1d; P=9.1 × 10−18 and P=1.8 × 10−6, respectively). Comparatively, ovarian cancer cell lines with a mesenchymal status showed a statistically significant upregulation of gene sets such as induction of apoptosis by extracellular signals and transition metal ion transport. Overall, these results demonstrate that cisplatin causes distinct responses between epithelial- and mesenchymal-like ovarian cancer cell lines at both the gene and pathway levels.
Cisplatin preferentially triggers G1/S- and S-phase gene expression and activates NF-κB pathway in epithelial-like ovarian cancer
Among the multiple DNA damage checkpoints in the cell cycle, previous studies have reported that cisplatin treatment induces cell cycle arrest at a checkpoint(s) in G2/M-, G1/S- or at other phases of the cell cycle, seemingly dependent upon the cell line analysed.33, 34, 35, 36, 37, 38, 39 Our observation that cell cycle-related genes are enriched in epithelial-like cells (Figure 1d) prompted us to identify the specific checkpoint where cells become arrested following cisplatin treatment. Using an expression signature curated from a previous study,40 the cisplatin-induced changes in ESs were statistically examined in epithelial- and mesenchymal-like ovarian cell lines. Mann–Whitney U-tests revealed that cisplatin treatment led to a significant enrichment in G1/S- and S-phase genes (P<0.001 for both phases) in epithelial-like ovarian cancer cells, with no selective enrichment for G2/M-phase or the other phase genes detected (Figure 2a). The selective induction of G1/S- and S-phase genes in epithelial-like cells was validated by quantitative reverse transcription–PCR with the same RNA used for the microarray assays (Supplementary Figure 1). This finding might suggest that epithelial-like cells are more susceptible to G1/S- and S-phase arrest than mesenchymal-like cell lines following cisplatin treatment. However, expression profiling of in vivo ovarian tumours (GSE15622)41 revealed that neither G1/S- nor S-phase genes were induced in tumours with an epithelial status, unlike that observed with the cell line model (Figure 2c). Moreover, the transcriptomic pattern for cell cycle phase genes does not wholly mirror the phenotypic pattern of cell cycle arrest, as observed using flow cytometric analyses of selected cell lines among the panel (Supplementary Figure 4 and Supplementary Information). These observations suggest a complex relationship between in vitro and in vivo cancer cells in terms of gene expression changes, as well as that between the transcriptome and the cellular phenotype.
We next confirmed the preferential activation of NF-κB-related pathways in epithelial-like ovarian cancer by comparing the averaged ES changes for NF-κB-related gene sets (n=14 extracted from MSigDB version 3.0; P=0.0016) between epithelial- and mesenchymal-like ovarian cancer cell lines (Figure 2b). This selective induction of NF-κB downstream genes in epithelial-like cells was validated by quantitative reverse transcription–PCR using the same RNA as that employed during the expression microarrays (Supplementary Figure 1). As an independent experiment to examine NF-κB activity in the cell, we performed enzyme-linked immunosorbent assays to measure the DNA binding activity of p65, a representative component of NF-κB protein dimers, and simultaneously measured NF-κB downstream genes by quantitative reverse transcription–PCR. The robust induction of the downstream genes (13.2–27.0-fold; Supplementary Figure 5) again confirmed NF-κB pathway activation by cisplatin treatment in epithelial-like ovarian cancer cells; however, the activation of p65 DNA binding capability (23.3–108.3%; Supplementary Figure 5) was marginal, and implied that one or more other NF-κB family proteins (c-Rel, Rel-B, p50 and p52) may be more responsible for this transactivation than p65. Importantly, using a publicly available data set (GSE15622),41 we observed similar NF-κB activation in response to carboplatin single treatment in in vivo epithelial-like ovarian tumours but neither to paclitaxel single nor to combined carboplatin plus paclitaxel (Mann–Whitney U-test, P=0.0303; Figure 2d), supporting the validity of the cell line model.
We noted that the cell line panel used in this study has a mixed histology (serous origin, n=20; non-serous origin, n=26) as well as TP53 status (mutant, n=30; wild type, n=16) (Supplementary Table 1; Supplementary Information). Thus, we repeated analyses using only the cell lines derived from serous ovarian cancers or the cell lines with mutant TP53. Our analysis similarly revealed statistically significant enrichments of G1/S- and S-phase genes (Supplementary Figure 6A), and NF-κB-related pathways (Supplementary Figure 6B) in epithelial-like cell lines, indicating that the differential transcriptomic responses to cisplatin treatment were mainly dependent upon the epithelial and mesenchymal phenotypes and to a lesser extent on the histology or TP53 status (see Supplementary Information for more details).
Impaired apoptosis in epithelial-like ovarian cancer cells in response to cisplatin
To obtain mechanistic insight, we investigated the phenotypic responses of cells following cisplatin treatment at various concentrations using MTS [(3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium)], BrdU incorporation and Caspase-3/7 release assays for cell viability, proliferation and apoptosis, respectively (Figure 3a). MTS assays revealed that epithelial-like cell lines required a higher concentration of cisplatin to achieve 50% growth inhibition, indicating that epithelial-like cells are more resistant to cisplatin than mesenchymal-like cells (Mann–Whitney U-test, P=0.0213; Figure 3b). Statistically significant differences were not detected for the half-maximal inhibitory concentration (IC50) of BrdU incorporation or for the half-maximal response concentration (EC50) of Caspase-3/7 release among the 46 ovarian cancer cell lines (Mann–Whitney U-test; P=0.7285 (BrdU), P=0.1535 (Caspase-3/7)). However, it is important to note that the EC50 for Caspase-3/7 release was significantly higher than the IC50 for BrdU incorporation in epithelial-like cell lines in molar concentration (Wilcoxon signed-rank test; P=0.0156; Figure 3c). This observation implies a susceptibility of epithelial-like cells to undergo cell cycle arrest but not apoptosis at a given dosage of cisplatin and possibly some impairment in apoptosis. On the other hand, mesenchymal-like cells showed no difference in the EC50 of Caspase-3/7 release and IC50 of BrdU incorporation (Figure 3c), suggesting simultaneous stimulation of cell cycle arrest and apoptosis with cisplatin at a same dosage. We have verified that our observation was not due to the TP53 mutation status, as TP53 wild type was distributed evenly in epithelial- and mesenchymal-like cells (data not shown). There was no difference in GI50 (P=0.3385), IC50 of BrdU incorporation (P=0.9692) or EC50 of Caspase-3/7 release (P=0.6160) between p53 wild-type and mutant ovarian cell lines by Mann–Whitney U-test, clearly indicating that sensitivity to cisplatin in terms of cell number, proliferation and apoptosis was independent of the TP53 status. Moreover, epithelial-like cell lines of serous histology or harbouring mutant TP53, albeit not statistically significant, showed similar tendencies towards resistance to cisplatin (Supplementary Figure 6C) as well as towards a susceptibility to undergo cell cycle arrest (Supplementary Figure 6D). Taken together, these results suggest divergent mechanisms in the response of epithelial- and mesenchymal-like ovarian cancer cell lines to cisplatin.
As most of the previous studies have linked mesenchymal status with a drug-resistant phenotype,13, 14, 15,21, 22, 23, 24 we sought to assess the validity of our in vitro observation of epithelial-like ovarian cancer cells being more cisplatin-resistant by extending our study through the use of clinical ovarian tumour samples. We took advantage of the expression data derived from chemo-naive samples (GSE15622, GSE3149 and TCGA)32,41, 42, 43 that had accompanying information about the patients’ clinical responses to chemotherapy. Most of the patients in these data sets received chemotherapy with carboplatin, a platinum drug equivalent to cisplatin in terms of its mechanism of action. By following the same analytical procedure used for the cell line analysis, we first classified clinical ovarian cancer samples into epithelial- or mesenchymal-like tumours in each data set. On the basis of this assignment, we then performed statistical analyses to ascertain whether therapeutically resistant samples were enriched in tumours with an epithelial status. As the number of samples in each data set was limited (Supplementary Table 5), the assignments were combined before statistical evaluation. Supporting the validity of the in vitro cell line model, there is indeed a statistically significant association of epithelial status with resistance to cisplatin or carboplatin (Fisher’s exact test; P=0.0141; Figure 4). It is important to note that this association could be observed only on ovarian cancer patients who had undergone single platinum treatment. Similarly, an examination of the relationship between EMT status with overall survival revealed that patients with epithelial-like ovarian cancer may have poorer prognoses if the recipient of platinum single treatment, with marginal statistical significance by the log-rank test (P=0.0634; Supplementary Figure 7). Paclitaxel single treatment or combination therapy with cisplatin/carboplatin plus paclitaxel did not exhibit enrichment of resistant tumours with either an epithelial or a mesenchymal status, implying that the sensitivity of epithelial- or mesenchymal-like cells can be specific to a treatment regimen (Fisher’s exact test, P=0.5764 and 0.7004, respectively; Figure 4). In addition, the EMT status was not correlated with overall survival for patients with the combination chemotherapy (Supplementary Figure 7). Nevertheless, the sample size is still small, even after combining three independent data sets, and the result was discrepant from those described in two previous reports21,24 for which EMT was implied to be linked with chemotherapeutic resistance in ovarian cancer.
Platinum resistance is a major obstacle in the development of effective ovarian cancer therapeutics. It is therefore becoming increasingly important to understand the underlying molecular mechanisms of this resistance. EMT has been reported to be critical not only in embryonic development but also in disease processes such as tumour progression.9, 10, 11, 12 Emerging evidence has emphasized a role for EMT in the development of drug resistance in cells. Indeed, highly invasive breast cancer cell lines with a more transitioned phenotype show a higher resistance to paclitaxel.25 Furthermore, the chronic or transient treatment of colorectal, breast and ovarian cancer cells with chemotherapeutic drugs, such as cisplatin, oxaliplatin, doxorubicin and paclitaxel in vitro, renders the cells drug resistant, which appears concomitant with the presence of a transitioned or an ‘EMTed’ phenotype.21,26,27,44,45 The overexpression of EMT-inducing transcription factors, such as SNAI1 and SNAI2, also confers resistance to apoptosis induced by DNA damage.44,46 Moreover, pancreatic cancer cell lines with mesenchymal properties exhibit a higher resistance to treatment with gemcitabine, 5-fluoro-uracil or cisplatin than cell lines with epithelial characteristics.18 Despite these previous findings13, 14, 15,21, 22, 23, 24,44,45,47 and contrary to our expectations, we found that cultured ovarian cancer cell lines with an epithelial status are more resistant to cisplatin than those with a mesenchymal status. This discrepancy may be derived from the fact that two distinct models of EMT were studied: ‘acquired’ and ‘inherent’ EMT.20 In the former, the cell acquires an EMTed phenotype through a dynamic process such as that caused by drug or growth factor treatment and the overexpression of EMT-inducing transcription factors.13, 14, 15, 16 By comparison, the latter describes an inherent nature of the cell, and forms a spectrum of EMT ranging from epithelium to mesenchyme as a function of the expression levels of several specific EMT markers.17, 18, 19, 20
A cell line pair—A2780 and its cisplatin-resistant derivative, A2780/CP70 (selected under long-term cisplatin treatment)—is a well-established model frequently used to study the molecular mechanism of cisplatin resistance in ovarian cancer.48,49 Other cell line models have similarly been employed.14,15 A2780/CP70, as well as other cisplatin-resistant derivatives in similar models, exhibits an EMTed phenotype, which manifests a mesenchymal cell morphology, increased motility and enhanced expression of EMT-inducing transcription factors.13, 14, 15,21 Although the acquisition of platinum resistance in the cell line models coincides with an EMTed phenotype as well as the activation of one or more pathways (as exemplified by epidermal growth factor receptor, mitogen-activated protein kinase, Jak-Stat and PI3K-Akt), the molecular mechanism(s) linking platinum resistance with EMT still remains largely elusive.13,14,22,23,50 Moreover, the EMT that occurs through cisplatin selection may simply be the result of an orchestrated cellular defence response in the resistant cells against platinum toxicity, as multiple types of stresses, such as oxidative stress (which can be also produced by platinum51,52), are capable of triggering a pathological EMT process.16 On the other hand, panels of in vitro cell lines have been used to represent the diverse heterogeneity of phenotypes observed in in vivo tumours in order to study therapeutic sensitivity to multiple drugs as well as the inherent status of the cells in terms of their EMT phenotype.17,18 In two previous studies, a mesenchymal status was correlated with resistance to chemotherapeutic drugs (gemcitabine, 5-fluoro-uracil and cisplatin) and to molecular target inhibitors (epidermal growth factor receptor and phosphoinositide 3-kinase inhibitors) in panels of pancreatic and non-small cell lung carcinoma cell lines, respectively.17,18 Importantly, epithelial-like non-small cell lung carcinoma cell lines showed a trend towards higher resistance for cisplatin, gemcitabine and vinorelbine, with marginal statistical significance.17 This is consistent with our observation that a correlation between EMT status and drug sensitivity depends on the treatment regimen. Furthermore, several lines of evidence have indicated that tumours with epithelial features are not necessarily sensitive to chemotherapy. An example is luminal-like breast cancers (characterized by epithelial features), which are more resistant than basal-subtype breast cancers (with mesenchymal phenotype) to anthracycline-based chemotherapy as well as to cisplatin single treatment.53, 54, 55 Conversely, testicular germ cell tumours are highly sensitive to cisplatin treatment, with a cure rate of >90%,56 even though these tumours are characterized by a high expression of the well-known mesenchymal markers, N-cadherin and vimentin.57,58
Unlike in the current study, recent clinical studies have implicated a link between EMT-related gene expression and relapse after platinum-based treatment24 as well as a link between EMT and innate resistance to platinum-based chemotherapy21 in ovarian cancer. Because these two studies seemingly relied on less stringent statistical evaluation and that the small number of samples in these studies and in our own is an obvious limitation, further examination with a higher number of samples is perhaps needed to obtain conclusive evidence.
To further scrutinize the relationship between epithelial status and cisplatin resistance, we focused on investigating the differential responses of epithelial- and mesenchymal-like cell lines to cisplatin treatment. One obvious distinction among the transcriptomic responses was enrichment of cell cycle-related gene sets in epithelial-like cells. In particular, cisplatin treatment prominently induced G1/S- and S-phase genes together with DNA checkpoint-related gene sets in vitro. Although the molecular mechanism(s) involved in the preferential induction of G1/S- and S-phase genes in epithelial-like ovarian cell lines remains to be elucidated, the findings presented here may indicate a shared molecular feature among cell lines with an epithelial status. In line with this, a distinct higher expression of cell cycle genes was indeed observed in luminal over basal breast cancer cell lines after doxorubicin or 5-fluoro-uracil treatment.28 Importantly, epithelial-like cells were susceptible to proliferation arrest but not to apoptosis at a given dosage of cisplatin. This is consistent with previous observations that proliferation arrest without cell death can be one mechanism of drug resistance and tumour relapse;59 the molecular mechanism of this impairment, however, is not clear in the current study. Taken together, the selective impairment of apoptosis can be a causative factor in the cisplatin resistance in epithelial-like ovarian cancer cells.
Consistent with previous reports, we observed an upregulation of NF-κB activity following cisplatin treatment.5,60 This finding may offer a clue as to the mechanism of selective impairment of apoptosis, and hence cisplatin resistance, in epithelial-like ovarian cancer cells. NF-κB is a complex and multi-faceted signalling pathway that can either contextually activate or suppress the apoptotic machinery in the cell.61 Emerging evidence, however, suggests that NF-κB may have an anti-apoptotic role against platinum agents in multiple cancer types, including ovarian cancer.62, 63, 64, 65 Indeed, inhibition of NF-κB has been shown to sensitize ovarian cancer cells to cisplatin as well as to other anti-tumoral agents.5,66, 67, 68 These observations may imply the potential therapeutic utility of the concomitant treatment of cisplatin and an NF-κB inhibitor to overcome innate cisplatin resistance in epithelial-like ovarian cancer cells.
Overall, our results show that epithelial- and mesenchymal-like ovarian cancer cells exhibit distinct responses following cisplatin administration. Our investigations have uncovered a selective apoptotic impairment in epithelial-like ovarian cancer cells, which is accompanied by NF-κB activation and increased resistance to cisplatin treatment. This distinction may suggest a need for differential therapeutic regimens in the treatment of ovarian cancers based on the EMT status of the cancer cell.
Materials and methods
Forty-six ovarian cancer cell lines (ovary1847, A2008, A2780, A2780cisR, C13, Caov-2, Caov-3, CH1, DOV13, DOV13A, DOV13B, FU-OV-1, Hey, HeyA8, HeyC2, IGROV-1, JHOS-2, JHOS-3, M41, OAW28, OAW42, OV56, OV90, OVCA420, OVCA429, OVCA432, OVCA433, OVCAR-10, OVCAR-2, OVCAR-3, OVCAR-5, OVCAR-8, OVK-18, PA-1, PEO1, RMG-I, RMG-II, SKOV-3, SKOV-4, SKOV-6, SKOV-8, TAYA, TOV-112D, TOV-21G, TYK-nu, UWB1.289) were gifts from Drs N Matsumura (Kyoto University, Kyoto, Japan) and S Murphy (Duke University, Durham, NC, USA), and have been described elsewhere.30,69,70 Cell lines were cultured in RPMI 1640 media (#23400021, Invitrogen, Carlsbad, CA, USA) with 10% fetal bovine serum (#S1810-500, Biowest, Nuaillé, France).
Measurement of phenotypic responses
Cisplatin (P4394) was purchased from Sigma Aldrich (St Louis, MO, USA), dissolved with Milli-Q water to a concentration of 5 mM, and stored as a stock solution at −20 °C before use. Forty-six ovarian cancer cell lines (23 epithelial-like and 23 mesenchymal-like) were tested for their sensitivity to cisplatin using CellTiter 96 AQueous Non-Radioactive Cell Proliferation Assay kit (#G5430, Promega, Fitchburg, WI, USA) based on MTS staining of cells, as described previously.30 Among these cell lines, 9 epithelial-like and 11 mesenchymal-like cells were further examined to quantify phenotypic changes in cell proliferation using the Cell Proliferation ELISA chemiluminescent BrdU kit (#11669915001, Roche Applied Science, Basel, Switzerland) and apoptosis using the Caspase-Glo(R) 3/7 Assay (#G8093, Promega) as per the manufacturers’ recommendations. Cells were seeded in 96-well plates at an optimal cell density (ranges 4000~12 000), as determined for each cell line to ensure that it reached 80% confluence by the end of the assay. Following an overnight incubation, cells were treated with five to six concentrations of cisplatin for 48 h, as required. The percentage of the cell population responding to cisplatin relative to the negative controls was measured. Dose–response curves were plotted using GraphPad Prism to derive the GI50, IC50 of BrdU incorporation and EC50 of Caspase-3 activation for each cell line in at least three independent experiments of quadruplicates.
Measurement of transcriptomic responses
To investigate transcriptomic responses to cisplatin, each of the 46 cell lines was treated with cisplatin at the GI50 dosage (1.45–201.9 μM). Sham treatment was used as a negative control. Cells were seeded in 100-mm plates at an optimal density, which was determined for each cell line to ensure that it reached 80% confluence by the end of the assay. Following an overnight incubation, cells were treated with cisplatin at the GI50 dosage for 48 h. RNA was then extracted from sham- or cisplatin-treated cell lines with RNeasy Mini Kit (Qiagen, Düsseldorf, Germany). Human Gene ST1.0 arrays (Affymetrix, Santa Clara, CA, USA) were used for the transcriptomic analyses. Expression data were deposited in the Gene Expression Omnibus (GEO) with the accession of GSE47856.
Data pre-processing of Affymetrix expression data
Robust Multichip Average (RMA) normalization was performed at the transcript level on the results from the Affymetrix Human Gene ST1.0 arrays using Affymetrix Power Tool 188.8.131.52 for all 46 sham- or cisplatin-treated ovarian cancer cell lines. The normalized data were subsequently standardized using ComBat71 to remove the batch effect. In this experiment, the cisplatin treatment assay was performed in triplicate on 20 cell lines, while single assays (without replicate) were performed on the remaining 26 cell lines. Taking advantage of the triplicate data, potentially fragile probes with strong variations (an s.d. of >0.2) within the triplicates were removed, decreasing the probe number from 33 297 to 21 329. To perform a fair comparison, the triplicate data were then log-averaged into one value so that one result for each cell line could be used in the following analyses.
Classification of cell lines into epithelial and mesenchymal status
The EMT status of ovarian cancer cell lines was determined by relying on the expression data derived from untreated cell lines, which reflect the innate nature of the cell line under normal culture conditions. Epithelial and mesenchymal expression signatures were generated by binary comparison between the transcriptional profiles of ‘epithelial’ (E-cadherinpositive/N-cadherinnegative detected by immunofluorescence) cell lines and those of ‘mensenchymal’ (E-cadherinlow or negative/N-cadherinpositive) cell lines with significant analysis of microarrays (false-discovery rate q=0) and receiver operating curve (ROC>0.85).30 This comparison yielded an epithelial gene set consisting of known epithelial cell markers, including DDR1, KRT8, KRT18, CDH1, CDH3, CLDN3, CLDN4 and EPCAM, and a mesenchymal gene set consisting of known mesenchymal cell markers, including ZEB1, CDH2, VIM and TWIST1 (Supplementary Table 2). Subsequently, single-sample gene set enrichment analysis was performed to assign an ES for epithelial or mesenchymal signature to each ovarian cancer cell line. The first principal component was then computed from the epithelial and mesenchymal ESs with Matlab (Natick, MA, USA). Using a median of the first principal component as a cut-off value, cell lines were finally subdivided into a binary category of ‘epithelial-like’ or ‘mesenchymal-like’. The cell line assignment is given in Supplementary Table 1.
Differential responses of epithelial- and mesenchymal-like cells following cisplatin treatment
To detect differential responses to cisplatin between epithelial- and mesenchymal-like cell lines, the transcriptomic responses to cisplatin were computed by subtracting the gene expression value of control (cisplatin untreated) cells from that of cisplatin-treated cells. Mann–Whitney U-test (P<0.01 as a cut-off value) was subsequently used to detect the differential transcriptomic responses between the expression changes by cisplatin treatment in epithelial-like cell lines with those in mesenchymal-like cell lines (Supplementary Table 3). Similarly, in order to examine whether this differential pattern could be also observed at the pathway level, single-sample gene set enrichment analysis72 was first applied to compute the ESs to the expression data following sham or cisplatin treatment (MSigDB version 3.0,73 6769 gene sets; http://www.broadinstitute.org/gsea/downloads.jsp). A cisplatin-induced change for each gene set was then computed by subtracting the ES of the sham-treated cell lines from that of the cisplatin-treated cell lines. Following this step, the changes in ES for the epithelial-like cells were compared with those in the mesenchymal-like cells to detect gene sets that showed differential responses to cisplatin (Mann–Whitney U-test, P<0.01; Supplementary Table 4).
Data sets derived from publicly available databases
To examine the reproducibility of the findings from the in vitro model in the in vivo setting, we took advantage of expression data from clinical samples (GSE15622: http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE15622; GSE3149: http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE3149 and TCGA (provisional version, downloaded in March 2013): http://cancergenome.nih.gov/).32,41, 42, 43 We used samples that had accompanying information regarding the patients’ clinical responses to chemotherapy. Whereas clinical samples in GSE3149 and TCGA were obtained from chemo-naive patients, GSE15622 expression data were derived from biopsied ovarian cancer samples before and after treatment. Most of the patients in these data sets received chemotherapy with carboplatin. The clinical responses in GSE3149 and TCGA studies32,42,43 were assessed per Response Evaluation Criteria in Solid Tumor (RECIST),74 whereas those in GSE15622 were defined on the basis of the change in serum CA-125 level.41 As clinical responses defined according to the serum CA-125 level were previously demonstrated to be comparable to that determined by RECIST,75 data from GSE1562, GSE3149 and TCGA were able to be combined for further statistical evaluation. Note that overall survival information is only available for data from GSE3149 and TCGA.
EC50, half maximal response concentration; EMT, epithelial–mesenchymal transition; ES, enrichment score; GI50, 50% growth inhibition concentration; IC50, half maximal inhibitory concentration; NF-κB, nuclear factor kappa-light-chain-enhancer of activated B cells.
Gene Expression Omnibus
Siegel R, Naishadham D, Jemal A . Cancer statistics, 2013. CA Cancer J Clin 2013; 63: 11–30.
Agarwal R, Kaye SB . Ovarian cancer: strategies for overcoming resistance to chemotherapy. Nat Rev Cancer 2003; 3: 502–516.
Coleman RL, Monk BJ, Sood AK, Herzog TJ . Latest research and treatment of advanced-stage epithelial ovarian cancer. Nat Rev Clin Oncol 2013; 10: 211–224.
Leitao MM Jr., Chi DS . Surgical management of recurrent ovarian cancer. Semin Oncol 2009; 36: 106–111.
Mabuchi S, Ohmichi M, Nishio Y, Hayasaka T, Kimura A, Ohta T et al. Inhibition of NFkappaB increases the efficacy of cisplatin in in vitro and in vivo ovarian cancer models. J Biol Chem 2004; 279: 23477–23485.
Shen DW, Pouliot LM, Hall MD, Gottesman MM . Cisplatin resistance: a cellular self-defense mechanism resulting from multiple epigenetic and genetic changes. Pharmacol Rev 2012; 64: 706–721.
Kelland L . The resurgence of platinum-based cancer chemotherapy. Nat Rev Cancer 2007; 7: 573–584.
Weir NM, Selvendiran K, Kutala VK, Tong L, Vishwanath S, Rajaram M et al. Curcumin induces G2/M arrest and apoptosis in cisplatin-resistant human ovarian cancer cells by modulating Akt and p38 MAPK. Cancer Biol Ther 2007; 6: 178–184.
Davidson B, Trope CG, Reich R . Epithelial-mesenchymal transition in ovarian carcinoma. Front Oncol 2012; 2: 33.
Thiery JP, Acloque H, Huang RY, Nieto MA . Epithelial-mesenchymal transitions in development and disease. Cell 2009; 139: 871–890.
Helleman J, Smid M, Jansen MP, van der Burg ME, Berns EM . Pathway analysis of gene lists associated with platinum-based chemotherapy resistance in ovarian cancer: the big picture. Gynecol Oncol 2010; 117: 170–176.
Tomaskovic-Crook E, Thompson EW, Thiery JP . Epithelial to mesenchymal transition and breast cancer. Breast Cancer Res 2009; 11: 213.
Yue P, Zhang X, Paladino D, Sengupta B, Ahmad S, Holloway RW et al. Hyperactive EGF receptor, Jaks and Stat3 signaling promote enhanced colony-forming ability, motility and migration of cisplatin-resistant ovarian cancer cells. Oncogene 2012; 31: 2309–2322.
Rosano L, Cianfrocca R, Spinella F, Di Castro V, Nicotra MR, Lucidi A et al. Acquisition of chemoresistance and EMT phenotype is linked with activation of the endothelin A receptor pathway in ovarian carcinoma cells. Clin Cancer Res 2011; 17: 2350–2360.
Wintzell M, Lofstedt L, Johansson J, Pedersen AB, Fuxe J, Shoshan M . Repeated cisplatin treatment can lead to a multiresistant tumor cell population with stem cell features and sensitivity to 3-bromopyruvate. Cancer Biol Ther 2012; 13: 1454–1462.
Kalluri R, Weinberg RA . The basics of epithelial-mesenchymal transition. J Clin Invest 2009; 119: 1420–1428.
Byers LA, Diao L, Wang J, Saintigny P, Girard L, Peyton M et al. An epithelial-mesenchymal transition gene signature predicts resistance to EGFR and PI3K inhibitors and identifies Axl as a therapeutic target for overcoming EGFR inhibitor resistance. Clin Cancer Res 2013; 19: 279–290.
Arumugam T, Ramachandran V, Fournier KF, Wang H, Marquis L, Abbruzzese JL et al. Epithelial to mesenchymal transition contributes to drug resistance in pancreatic cancer. Cancer Res 2009; 69: 5820–5828.
Yang D, Sun Y, Hu L, Zheng H, Ji P, Pecot CV et al. Integrated analyses identify a master microRNA regulatory network for the mesenchymal subtype in serous ovarian cancer. Cancer Cell 2013; 23: 186–199.
Huang RY, Chung VY, Thiery JP . Targeting pathways contributing to epithelial-mesenchymal transition (EMT) in epithelial ovarian cancer. Curr Drug Targets 2012; 13: 1649–1653.
Haslehurst AM, Koti M, Dharsee M, Nuin P, Evans K, Geraci J et al. EMT transcription factors snail and slug directly contribute to cisplatin resistance in ovarian cancer. BMC Cancer 2012; 12: 91.
Latifi A, Abubaker K, Castrechini N, Ward AC, Liongue C, Dobill F et al. Cisplatin treatment of primary and metastatic epithelial ovarian carcinomas generates residual cells with mesenchymal stem cell-like profile. J Cell Biochem 2011; 112: 2850–2864.
Gupta N, Xu Z, El-Sehemy A, Steed H, Fu Y . Notch3 induces epithelial-mesenchymal transition and attenuates carboplatin-induced apoptosis in ovarian cancer cells. Gynecol Oncol 2013; 130: 200–206.
Marchini S, Fruscio R, Clivio L, Beltrame L, Porcu L, Nerini IF et al. Resistance to platinum-based chemotherapy is associated with epithelial to mesenchymal transition in epithelial ovarian cancer. Eur J Cancer 2013; 49: 520–530.
Cheng GZ, Chan J, Wang Q, Zhang W, Sun CD, Wang LH . Twist transcriptionally up-regulates AKT2 in breast cancer cells leading to increased migration, invasion, and resistance to paclitaxel. Cancer Res 2007; 67: 1979–1987.
Li QQ, Xu JD, Wang WJ, Cao XX, Chen Q, Tang F et al. Twist1-mediated adriamycin-induced epithelial-mesenchymal transition relates to multidrug resistance and invasive potential in breast cancer cells. Clin Cancer Res 2009; 15: 2657–2665.
Yang AD, Fan F, Camp ER, van Buren G, Liu W, Somcio R et al. Chronic oxaliplatin resistance induces epithelial-to-mesenchymal transition in colorectal cancer cell lines. Clin Cancer Res 2006; 12: 4147–4153.
Troester MA, Hoadley KA, Sorlie T, Herbert BS, Borresen-Dale AL, Lonning PE et al. Cell-type-specific responses to chemotherapeutics in breast cancer. Cancer Res 2004; 64: 4218–4226.
Ostrow S, Egorin M, Aisner J, Bachur N, Wiernik PH . High-dose cis-diamminedichloro-platinum therapy in patients with advanced breast cancer: pharmacokinetics, toxicity, and therapeutic efficacy. Cancer Clin Trials 1980; 3: 23–27.
Tan TZ, Miow QH, Huang RY, Wong MK, Ye J, Lau JA et al. Functional genomics identifies five distinct molecular subtypes with clinical relevance and pathways for growth control in epithelial ovarian cancer. EMBO Mol Med 2013; 5: 983–998.
Tothill RW, Tinker AV, George J, Brown R, Fox SB, Lade S et al. Novel molecular subtypes of serous and endometrioid ovarian cancer linked to clinical outcome. Clin Cancer Res 2008; 14: 5198–5208.
Cancer Genome Atlas Research N. Integrated genomic analyses of ovarian carcinoma. Nature 2011; 474: 609–615.
Yip HT, Chopra R, Chakrabarti R, Veena MS, Ramamurthy B, Srivatsan ES et al. Cisplatin-induced growth arrest of head and neck cancer cells correlates with increased expression of p16 and p53. Arch Otolaryngol Head Neck Surg 2006; 132: 317–326.
Qin LF, Ng IO . Induction of apoptosis by cisplatin and its effect on cell cycle-related proteins and cell cycle changes in hepatoma cells. Cancer Lett 2002; 175: 27–38.
Gatti L, Supino R, Perego P, Pavesi R, Caserini C, Carenini N et al. Apoptosis and growth arrest induced by platinum compounds in U2-OS cells reflect a specific DNA damage recognition associated with a different p53-mediated response. Cell Death Differ 2002; 9: 1352–1359.
Sorenson CM, Eastman A . Mechanism of cis-diamminedichloroplatinum(II)-induced cytotoxicity: role of G2 arrest and DNA double-strand breaks. Cancer Res 1988; 48: 4484–4488.
Demarcq C, Bunch RT, Creswell D, Eastman A . The role of cell cycle progression in cisplatin-induced apoptosis in Chinese hamster ovary cells. Cell Growth Differ 1994; 5: 983–993.
Mueller S, Schittenhelm M, Honecker F, Malenke E, Lauber K, Wesselborg S et al. Cell-cycle progression and response of germ cell tumors to cisplatin in vitro. Int J Oncol 2006; 29: 471–479.
Lin-Chao S, Chao CC . Reduced Inhibition of DNA Synthesis and G(2) Arrest during the Cell Cycle of Resistant HeLa Cells in Response to cis-Diamminedichloroplatinum. J Biomed Sci 1994; 1: 131–138.
Mizuno H, Nakanishi Y, Ishii N, Sarai A, Kitada K . A signature-based method for indexing cell cycle phase distribution from microarray profiles. BMC Genomics 2009; 10: 137.
Ahmed AA, Mills AD, Ibrahim AE, Temple J, Blenkiron C, Vias M et al. The extracellular matrix protein TGFBI induces microtubule stabilization and sensitizes ovarian cancers to paclitaxel. Cancer Cell 2007; 12: 514–527.
Bild AH, Yao G, Chang JT, Wang Q, Potti A, Chasse D et al. Oncogenic pathway signatures in human cancers as a guide to targeted therapies. Nature 2006; 439: 353–357.
Verhaak RG, Tamayo P, Yang JY, Hubbard D, Zhang H, Creighton CJ et al. Prognostically relevant gene signatures of high-grade serous ovarian carcinoma. J Clin Invest 2013; 123: 517–525.
Kurrey NK, Amith K, Bapat SA . Snail and Slug are major determinants of ovarian cancer invasiveness at the transcription level. Gynecol Oncol 2005; 97: 155–165.
Kajiyama H, Shibata K, Terauchi M, Yamashita M, Ino K, Nawa A et al. Chemoresistance to paclitaxel induces epithelial-mesenchymal transition and enhances metastatic potential for epithelial ovarian carcinoma cells. Int J Oncol 2007; 31: 277–283.
Kajita M, McClinic KN, Wade PA . Aberrant expression of the transcription factors snail and slug alters the response to genotoxic stress. Mol Cell Biol 2004; 24: 7559–7566.
Du F, Wu X, Liu Y, Wang T, Qi X, Mao Y et al. Acquisition of paclitaxel resistance via PI3Kdependent epithelialmesenchymal transition in A2780 human ovarian cancer cells. Oncol Rep 2013; 30: 1113–1118.
Parker RJ, Eastman A, Bostick-Bruton F, Reed E . Acquired cisplatin resistance in human ovarian cancer cells is associated with enhanced repair of cisplatin-DNA lesions and reduced drug accumulation. J Clin Invest 1991; 87: 772–777.
Perez RP . Cellular and molecular determinants of cisplatin resistance. Eur J Cancer 1998; 34: 1535–1542.
Singh A, Settleman J . EMT, cancer stem cells and drug resistance: an emerging axis of evil in the war on cancer. Oncogene 2010; 29: 4741–4751.
Simons AL, Ahmad IM, Mattson DM, Dornfeld KJ, Spitz DR . 2-Deoxy-D-glucose combined with cisplatin enhances cytotoxicity via metabolic oxidative stress in human head and neck cancer cells. Cancer Res 2007; 67: 3364–3370.
Buttke TM, Sandstrom PA . Oxidative stress as a mediator of apoptosis. Immunol Today 1994; 15: 7–10.
Sarrio D, Rodriguez-Pinilla SM, Hardisson D, Cano A, Moreno-Bueno G, Palacios J . Epithelial-mesenchymal transition in breast cancer relates to the basal-like phenotype. Cancer Res 2008; 68: 989–997.
Carey LA, Dees EC, Sawyer L, Gatti L, Moore DT, Collichio F et al. The triple negative paradox: primary tumor chemosensitivity of breast cancer subtypes. Clin Cancer Res 2007; 13: 2329–2334.
Silver DP, Richardson AL, Eklund AC, Wang ZC, Szallasi Z, Li Q et al. Efficacy of neoadjuvant Cisplatin in triple-negative breast cancer. J Clin Oncol 2010; 28: 1145–1153.
Masters JR, Koberle B . Curing metastatic cancer: lessons from testicular germ-cell tumours. Nat Rev Cancer 2003; 3: 517–525.
Bremmer F, Hemmerlein B, Strauss A, Burfeind P, Thelen P, Radzun HJ et al. N-cadherin expression in malignant germ cell tumours of the testis. BMC Clin Pathol 2012; 12: 19.
Miettinen M, Virtanen I, Talerman A . Intermediate filament proteins in human testis and testicular germ-cell tumors. Am J Pathol 1985; 120: 402–410.
Shah MA, Schwartz GK . Cell cycle-mediated drug resistance: an emerging concept in cancer therapy. Clin Cancer Res 2001; 7: 2168–2181.
Godwin P, Baird AM, Heavey S, Barr MP, O'Byrne KJ, Gately K . Targeting nuclear factor-kappa B to overcome resistance to chemotherapy. Front Oncol 2013; 3: 120.
Oeckinghaus A, Hayden MS, Ghosh S . Crosstalk in NF-kappaB signaling pathways. Nat Immunol 2011; 12: 695–708.
Xiao G, Fu J . NF-kappaB and cancer: a paradigm of Yin-Yang. Am J Cancer Res 2011; 1: 192–221.
Samanta AK, Huang HJ, Bast RC Jr., Liao WS . Overexpression of MEKK3 confers resistance to apoptosis through activation of NFkappaB. J Biol Chem 2004; 279: 7576–7583.
Liu GH, Wang SR, Wang B, Kong BH . Inhibition of nuclear factor-kappaB by an antioxidant enhances paclitaxel sensitivity in ovarian carcinoma cell line. Int J Gynecol Cancer 2006; 16: 1777–1782.
Bentires-Alj M, Barbu V, Fillet M, Chariot A, Relic B, Jacobs N et al. NF-kappaB transcription factor induces drug resistance through MDR1 expression in cancer cells. Oncogene 2003; 22: 90–97.
Solomon LA, Ali S, Banerjee S, Munkarah AR, Morris RT, Sarkar FH . Sensitization of ovarian cancer cells to cisplatin by genistein: the role of NF-kappaB. J Ovarian Res 2008; 1: 9.
Parajuli B, Lee HG, Kwon SH, Cha SD, Shin SJ, Lee GH et al. Salinomycin inhibits Akt/NF-kappaB and induces apoptosis in cisplatin resistant ovarian cancer cells. Cancer Epidemiol 2013; 37: 512–517.
Escarcega RO, Fuentes-Alexandro S, Garcia-Carrasco M, Gatica A, Zamora A . The transcription factor nuclear factor-kappa B and cancer. Clin Oncol (R Coll Radiol) 2007; 19: 154–161.
Matsumura N, Huang Z, Mori S, Baba T, Fujii S, Konishi I et al. Epigenetic suppression of the TGF-beta pathway revealed by transcriptome profiling in ovarian cancer. Genome Res 2011; 21: 74–82.
Kondoh E, Mori S, Yamaguchi K, Baba T, Matsumura N, Cory Barnett J et al. Targeting slow-proliferating ovarian cancer cells. Int J Cancer 2010; 126: 2448–2456.
Johnson WE, Li C, Rabinovic A . Adjusting batch effects in microarray expression data using empirical Bayes methods. Biostatistics 2007; 8: 118–127.
Verhaak RG, Hoadley KA, Purdom E, Wang V, Qi Y, Wilkerson MD et al. Integrated genomic analysis identifies clinically relevant subtypes of glioblastoma characterized by abnormalities in PDGFRA, IDH1, EGFR, and NF1. Cancer Cell 2010; 17: 98–110.
Subramanian A, Tamayo P, Mootha VK, Mukherjee S, Ebert BL, Gillette MA et al. Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles. Proc Natl Acad Sci USA 2005; 102: 15545–15550.
Kang J, D'Andrea AD, Kozono D . A DNA repair pathway-focused score for prediction of outcomes in ovarian cancer treated with platinum-based chemotherapy. J Natl Cancer Inst 2012; 104: 670–681.
Rustin GJ, Timmers P, Nelstrop A, Shreeves G, Bentzen SM, Baron B et al. Comparison of CA-125 and standard definitions of progression of ovarian cancer in the intergroup trial of cisplatin and paclitaxel versus cisplatin and cyclophosphamide. J Clin Oncol 2006; 24: 45–51.
Nakamura Y . Isolation of p53-target genes and their functional analysis. Cancer Sci 2004; 95: 7–11.
We thank Drs K Yamaguchi, N Matsumura and S Murphy for kindly providing us with a panel of ovarian cancer cell lines. We thank Dr R Jackson for her careful editing of the English language. This work was supported in part through a grant from the Cancer Science Institute of Singapore, the Institute of Molecular and Cell Biology at A*STAR, Singapore, the Vehicle Racing Commemorative Foundation in Japan and the Princess Takamatsu Cancer Research Fund.
The authors declare no conflict of interest.
Supplementary Information accompanies this paper on the Oncogene website
About this article
Cite this article
Miow, Q., Tan, T., Ye, J. et al. Epithelial–mesenchymal status renders differential responses to cisplatin in ovarian cancer. Oncogene 34, 1899–1907 (2015). https://doi.org/10.1038/onc.2014.136
Overcoming Limitations of Cisplatin Therapy by Additional Treatment With the HSP90 Inhibitor Onalespib
Frontiers in Oncology (2020)
International Journal of Molecular Sciences (2020)
Morin exerts anti-metastatic, anti-proliferative and anti-adhesive effect in ovarian cancer cells: an in vitro studies
Molecular Biology Reports (2020)
MicroRNA-302 represses epithelial-mesenchymal transition and cisplatin resistance by regulating ATAD2 in ovarian carcinoma
Experimental Cell Research (2020)
Epithelial-to-mesenchymal transition is a prognostic marker for patient outcome in advanced stage HNSCC patients treated with chemoradiotherapy
Radiotherapy and Oncology (2020)