HGSC tissues with different RFS were supplied by the Zhejiang Cancer Hospital Biospecimen Repository. The present study was approved by the Zhejiang Cancer Hospital Ethics Committee (2012-03-7) and informed consent was obtained from all participants prior to their inclusion. Histopathology and tumor grade were evaluated. Tumor stage was determined by a gynecologic oncologist based on FIGO classification criteria. All procedures were carried out in accordance with the approved guidelines of Zhejiang Cancer Research Institution.

Total RNA (without ribosomal RNA) of cisplatin-sensitive and -resistant A2780 cells was extracted and submitted for sequencing. The detailed procedures are described in previous reports (5). A gene was considered significantly different at fold change of >2 and P-value <0.001. Gene Ontology (GO) annotation was performed using the UniProt-GOA database (http://www.ebi.ac.uk/GOA). Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways were annotated using the online KEGG Automatic Annotation system. A similar approach was used for proteomics data analysis.

Total RNA was extracted from cultured cells with TRIzol reagent (Invitrogen; Thermo Fisher Scientific, Inc., Waltham, MA, USA) and1 µg used for cDNA synthesis with a PrimeScript RT reagent kit (Takara Biotechnology Co., Ltd., Dalian, China). Amplification and melting curve analyses were performed using a LightCycler 480 RT-qPCR system and LightCycler 480 SYBR-Green I Master Mix (Roche Applied Science, Mannheim, Germany). Amplification cycles were 95°C for 3 min, 35 cycles at 95°C for 30 sec, 55°C for 30 sec, 72°C for 30 sec, followed by 72°C for 3 min. β-Actin was used for normalization and the relative levels calculated using the 2^−ΔΔCt method. Tissue samples were obtained from patients with different recurrence-free survival (RFS) rates (RFS >12 months classified as the cisplatin-sensitive group and RFS ≤12 months as the cisplatin-resistant group), and cDNA hybrids of 32 samples were mixed as the control. Each PCR reaction was performed using one premixed reference template. All reactions were conducted in triplicate.

A2780 and SKOV3 cells were purchased from Cobioer (Nanjing, China) and verified by Shanghai Biowing Applied Biotechnology Co., Ltd. Expression plasmids containing CAMK2D (NM_172115) and SMARCA2 (NM_003070) were purchased from Shanghai GeneChem Co., Ltd., (Shanghai, China) and Lipofect2000™ obtained from Invitrogen; Thermo Fisher Scientific, Inc.

Cisplatin-sensitive A2780 and SKOV3 cells were seeded into six-well plates and allowed to reach 90–95% confluence at 37°C before transfection. Cells were transfected with expression plasmids containing CAMK2D or SMARCA2 and those containing empty vector used as a control. The ratio of plasmid to Lipofect2000 was 1:2. After 6 h transfection, the medium was replaced with complete culture medium. Cells were continuously cultured until harvesting for analysis after 48 h transfection.

Cisplatin-sensitive A2780 cells transfected with different plasmids were seeded in 96-well plates at a density of 10⁴ cells per well and cultured at 37°C in 5% CO₂. After 24 h culture, cells in each well were incubated with cisplatin at various concentrations for 24 h. The drug was removed by replacing with DMEM containing 20 µl MTS (Promega Corporation, Madison, WI, USA) for every 100 µl culture medium in each well. Cells were further incubated for 4 h at 37°C before the assay. Absorbance was measured using an automated microplate reader (SpectraMax M5) at 490 nm and all reactions were carried out in triplicate.

Cells were transfected with expression plasmids containing CAMK2D, SMARCA2 or empty vector, incubated for 48 h and treated with 10 µM cisplatin. At 24 h after treatment, the medium was removed and 1×10⁶ cells collected in 1.5 ml tubes. Next, cells were washed with PBS and centrifuged for 5 min at 200 g. The supernatant was discarded and 1 ml DNA staining solution added with 10 µl permeabilization solution. Cells were subjected to vortex oscillation for 5–10 sec followed by incubation at room temperature under protection from light, and subsequently prepared for flow cytometric analysis.

Cells were resuspended in 500 µl binding buffer after washing with PBS. Apoptosis analysis was performed using a FITC-AnnexinV/propidium iodide (PI) assay. Three groups were set as negative controls: i) Annexin-v-FITC was added to cells without PI; ii) PI was added to cells without Annexin-V-FITC and iii) neither Annexin-v-FITC nor PI was added to cells. Cells were incubated for 5 min under protection from the light after gentle vortex oscillation. The Annexin V-FITC/PI kit and PI were purchased from Hangzhou MultiSciences (Lianke) Biotech, Co., Ltd., (Hangzhou, China).

Cells were lysed using lysis buffer (4% SDC in 0.1 M Tris-HCl, pH 8.0) with SigmaFast protease inhibitor (Sigma-Aldrich; Merck KGaA, Darmstadt, Germany). Equal amounts of protein from the four cell sublines were reduced with 10 mM DTT and alkylated with 25 mM iodoacetamide. An in solution digestion was carried out with sequencing grade-modified trypsin/Lys-C (Promega Corporation) at 37°C overnight. The peptides were acidated with a final concentration of 0.5–1% trifluoracetic acid (TFA) and SDC removed via high-speed centrifugation. Tryptic peptides were desalted, centrifuged in a speedvac for drying and redissolved in 0.1% FA.

For LC-MS/MS analysis, peptides were separated by a 90 min gradient elution at a flow rate of 0.20 µl/min on a Thermo Scientific EASY-nLC 1000 HPLC system directly interfaced with a Thermo Scientific Q Exactive mass spectrometer. The analytical column was a Thermo Scientific Acclaim RPepMap RSLC (50 µm ID, 15 cm length, C18, 2 µm, 100 Å) and the precolumn was a Thermo Scientific AcclaimR PepMap100 column (100 µm ID, 2 cm length, C18, 5 µm, 100 Å). Mobile phases A and B consisted of 0.1% formic acid and acetonitrile with 0.1% formic acid, respectively. A Q Exactive mass spectrometer was operated in data-dependent acquisition mode using Xcalibur 2.2 SP1 software and there was a single full-scan mass spectrum in the orbitrap (350–2,000 m/z, 70,000 resolution) followed by 15 data-dependent MS/MS scans at 27% normalized collision energy (HCD). MS/MS spectra from each LC-MS/MS run were searched against ‘human. fasta’ from UniProt (release date March 19, 2014; 68406 entries) using Proteome Discoverer (PD) software (v.PD1.4; Thermo Fisher Scientific, Inc.). In terms of search criteria, full tryptic specificity was required, two missed cleavages allowed, carbamidomethylation (C) set as the fixed modification, oxidation (M) set as the dynamic modification, precursor mass tolerances set at 10 ppm, and the fragment mass tolerance set at 0.02 Da. The peptide false discovery rate (FDR) was calculated using Percolator provided by PD. Relative protein quantification was performed using MaxQuant software (v.1.4.0.8). Quantitation was only carried out for proteins with two or more unique peptide matches and PEP <0.001. Differentially expressed proteins were further confirmed by western blotting.

Proteins from cells were extracted in lysis buffer (Pierce; Thermo Fisher Scientific, Inc.) and a protease inhibitor cocktail (Sigma-Aldrich; Merck KGaA). Aliquots of protein samples (20 µl) were separated via SDS-PAGE. Samples were transferred onto PVDF membranes with a blotting system (ATTO) and membranes blocked with 5% non-fat milk for 2 h. Next, membranes were incubated overnight at 4°C with antibodies against SMARCA2, CAMK2D, cyclinE, Bcl-2, CDK4 and β-actin (1:1,000; ProteinTech Group, Inc., Chicago, IL, USA). After incubation with peroxidase-coupled anti-rabbit/mouse IgG (ProteinTech Group, Inc.) at 37°C for 2 h, bound proteins were visualized with ECL (Pierce; Thermo Fisher Scientific, Inc.) and detected using a BioImaging System (UVP, Inc., Upland, CA, USA). Relative protein levels were quantified using β-actin as a loading control.

All experiments were performed in triplicate and data are presented as the mean ± standard deviation. A nonparametric Student's t-test was used for the analysis of paired groups. All calculations were done with SPSS v.13.0 software (SPSS, Inc., Chicago, IL, USA). P<0.05 was considered to indicate a statistically significant difference.

Cisplatin resistance is one of the main problems encountered by cancer patients during cancer therapy. While numerous studies to date have identified genes associated with cisplatin resistance, the underlying mechanism remains unclear due to its complexity. To clarify the underlying pathways involved in the mechanism of cisplatin resistance development in ovarian cancer and identify contributory genes, we performed deep RNA sequencing, as described in previously, to compare the mRNA expression profiles between cisplatin-sensitive and -resistant A2780 cells (5). All RNA-seq data were analyzed using Bioinformatics technology (available upon request). In the present study, differentially expressed mRNAs between A2780 and A2780-DR cells were obtained from analyses using the DAVID bioinformatics platform (david.abcc.ncifcrf.gov/). GO and KEGG pathway analyses were included to further establish the biological relevance of differentially expressed mRNAs. The mRNAs showing significant alterations between A2780 and A2780-DR cells were classified into several groups of biological processes, including protein-DNA complex assembly (22; 1.038%), positive regulation of apoptosis (71; 3.35%), nucleosome assembly (21; 0.991%), L-serine metabolic process (6; 0.283%) and regulation of the cell cycle (58; 2.738%; Fig. 1A). Proteins encoded by the identified mRNAs were classified into several KEGG pathways, including adherens junction (15; 0.708%), SNARE interactions in vesicular transport (9; 0.425%), glioma (13; 0/614%), ubiquitin-mediated proteolysis (24; 1.13%) and glycine, serine and threonine metabolism (8; 0.378%; Fig. 1B).

To determine whether these genes were variably expressed in cisplatin-sensitive (A2780) and -resistant (A2780-DR) cells, six mRNAs were selected for RT-qPCR validation. As shown in Table I, some of these mRNAs were reported previously and others identified in the present study. Levels were quantified with a LightCycler 480 Real-Time PCR system and amplified with the aid of specifically designed primers (Table II). β-Actin was used for normalization and the relative levels calculated using the ΔΔCq method. The results showed higher expression levels of CAMK2D ARHGAP29 and SMARCA2 in cisplatin-resistant cells (A2780-DR), compared with sensitive cells (A2780), while the ABCB1 and NKAIN3 level was not (Fig. 2A). On the other hand, the expression levels of those genes in SKVO3 cells before and after cisplatin treatment didn't show the same pattern as in A2780 cells. SMARCA2 mRNA expression, but not that of CAMK2D, was significantly higher in SKVO3 cells after cisplatin treatment (Fig. 2B), indicating distinct roles of SMARCA2 and CAMK2D in different tumors. Based on this finding, we conducted further detailed studies on these two genes in association with cisplatin resistance development.

CAMK2D is strongly associated with epithelial ovarian cancer (EOC) risk and SMARCA2 is frequently implicated as a tumor suppressor. To determine whether these two genes are putatively involved in cisplatin resistance, their mRNA levels in 32 selected patients with different RFS rates after hospital treatment were determined. Based on the RT-qPCR results, patients with shorter RFS (RFS <12 months, no. 01-019) had significantly higher SMARCA2 and CAMK2D levels, compared to those with longer RFS (RFS >12 months, no. 1-13; Fig. 2C). Our data indicate a higher risk of ovarian cancer recurrence in EOC patients with elevated CAMK2D expression.

To further verify whether SMARCA2 or CAMK2D influence cisplatin resistance, plasmids expressing either gene were transfected into A2780 cells and cisplatin sensitivity evaluated. After transfection, CAMK2D and SMARCA2 levels were evaluated via RT-qPCR. The viability of three cell lines (A2780, A2780-CAMK2D and A2780-SMARCA2) was measured after cisplatin treatment. The survival rate of A2780 cells was decreased to <20% after treatment with 20 µM cisplatin while that of A2780-CAMK2D cells was 38 and 30% for A2780-SMARCA2 cells (Fig. 3A and B).

CAMK2D and SMARCA2 were additionally overexpressed in another ovarian cancer cell line, SKVO3, to validate their effects. The survival rate of cisplatin-treated cells increased from 20 to 25% in SKVO3-CAMK2D and 38% in SKVO3-SMARCA2 (Fig. 3C and D). Data from cell viability analyses in both A2780 and SKVO3 cell lines indicate that overexpression of either CAMK2D or SMARCA2 increases cell resistance to cisplatin.

Conversely, knockdown of CAMK2D and SMARCA2 was performed in A2780-DR cells via siRNA transfection. Upon suppression of these genes, the survival rate of A2780-DR under cisplatin treatment was decreased, compared to that of control A2780-DR cells (Fig. 3E).

Previous GO and KEGG pathway analyses suggest that CAMK2D and SMARCA2 are associated with regulation of the cell cycle and apoptosis. To establish the participation of these two genes in the above processes, A2780 cells were transfected with CAMK2D or SMARCA2, harvested after cisplatin treatment and subjected to cell cycle and apoptosis analyses. Apoptosis analyses showed that overexpression of CAMK2D and SMARCA2 lead to a significant reduction in apoptosis of A2780 cells treated with cisplatin. Overall, >20% A2780 cells underwent apoptosis after treatment, while only ~10% apoptotic cells were detected in cells expressing CAMK2D or SMARCA2 (Fig. 4A). Similar results were obtained with the SKVO3 cell line, but the extent of apoptosis was lower than that in A2780/DDP (Fig. 4B). Our data suggest that CAMK2D and SMARCA2 increase cancer cell resistance to cisplatin mainly via reducing apoptosis. Cell cycle analyses disclosed that after cisplatin treatment, the percentages of G1 and G2/M cells were decreased while cells in the S phase were simultaneously increased, indicative of drug-induced defects in G2/M progression. Upon CAMK2D and SMARCA2 overexpression in A2780, the G2/M cell population was decreased ~4% after cisplatin treatment while that in control A2780 cells was decreased ~8%, indicating a reduction in sensitivity to cisplatin-induced inhibition of cell proliferation (Fig. 5A). We additionally assessed the levels of CDK4, cyclin E and Bcl-2 related to the cell cycle and apoptosis, respectively, in the same sample group. Expression of these three proteins was higher in the A2780-CAMK2D and A2780-SMARCA2 cell lines than control A2780 cells (Fig. 5B).

To further elucidate the specific mechanisms by which CAMK2D or SMARCA2 contributes to cisplatin resistance in ovarian cancer, quantitative proteomic analysis was conducted in A2780, A2780/DDP, A2780-SMACRA2 and A2780-CAMK2D cells. Protein expression changes after cisplatin treatment or SMARCA2 and CAMK2D overexpression were measured and change fold values of >1.5 or <0.67 selected for further analysis (available upon request). After cisplatin treatment, about 305 proteins (group I) involved in several processes of metabolism and disease development were highly expressed, 315 proteins (group II) displayed changes in expression after transfection with SMARCA2, and 250 (group III) were altered after transfection with CAMK2D. In total, ~212 proteins were altered in groups I and II, 172 proteins in groups I and III, and 186 proteins in groups II and III (Fig. 6A). Based on KEGG pathway analysis, the majority of these proteins contribute to spliceosome assembly, RNA degradation, oxidative phosphorylation, pyrimidine metabolism and the development of two neurodegenerative diseases (Huntington's and Parkinson's) (Fig. 6B). To further clarify the relationship between these proteins and CAMK2D or SMARCA2 overexpression, we selected PDI and RAN and verified their expression patterns in A2780-CAMK2D, A2780-SMARCA2 and A2780 with or without DDP treatment (Fig. 6C). The signals for these two proteins in A2780-CAMK2D, A2780-SMARCA2 and A2780/DDP were significantly higher than those in A2780 cells. Further studies are warranted to ascertain how these proteins are involved in the development of cisplatin resistance of ovarian cancer and the underlying mechanisms.