Cell lines

The human epithelial ovarian cancer cell lines SKOV3, A2780, OVCAR3, OVCAR4, IGROV1, ES2 and COC1 were obtained from the Biological Cell Institute of Chinese Peking Union Medical College (Chinese Academy of Medical Science) and cultured in HG-DMEM (Gibco Invitrogen, Carlsbad, CA) supplemented with 10% FBS (HyClone, South Logan, UT) at 37 °C and 5% CO2.

SP cell sorting

When ovarian cancer cells had reached a logarithmic growth phase, they were analysed by Moflo (Beckman Coulter, Fullerton, CA). Cells were digested with 0.25% trypsin (Sigma–Aldrich), washed twice and resuspended in HG-DMEM with 2% FBS at a concentration of 1 × 106 cells/mL. Hoechst 33342 was added at a final concentration of 5 μg/mL, incubated for 90 min in the dark, and then washed twice. The cells were kept at 4 °C, and 1 μg/mL propidium iodide (Sigma–Aldrich) was added before sorting. Verapamil (50 μg/mL, a calcium ion tunnel antagonist) was added for 30 min at 37 °C before adding Hoechst 33342 to control cells, which blocked the fluorescent efflux of SP cells in ovarian cancer cells. SP cells were sorted and cultured in vitro. When SP cells reached a logarithmic growth phase, second SP and non-SP (NSP) sorting was performed to improve the ratio of SP cells.

Stable isotopic labelling by amino acids in cell culture (SILAC)

The SILAC membrane protein identification and quantitation kit from Invitrogen (Carlsbad, CA) was used for cell labelling and fractionation according to the manufacturer’s instructions. Briefly, SKOV3 and A2780 human epithelial ovarian cancer cells were seeded and cultured in modified DMEM supplemented with 10% dialyzed FBS and either 0.1 mg/ml light L-lysine or heavy [U-13C6] L-lysine. After six doubling times, SP and non-SP cells were harvested and sorted on a Moflo (Beckman Coulter, Fullerton, CA). A total of 106 SP cells labelled with [U-13C6] L-lysine and non-SP cells labelled with light L-lysine were mixed at a 1:1 ratio and lysed in 1.6 ml membrane protein lysis buffer with 0.08% benzonase nuclease on ice for 30 min. The lysate was mixed with 0.4 ml of 1.25 M sucrose solution and centrifuged at 500 g for 10 minutes at 4 °C to remove the nuclear fraction. The supernatant was centrifuged at 100,000 g for 1 hour at 4 °C to obtain the membrane pellet, which was resuspended in 20 μl Nu AGE LDS sample buffer and resolved by SDS–PAGE (dodecyl sulfate, sodium salt-polyacrylamide gel electrophoresis). The entire gel lane was cut into 45 fractions and subjected to in-gel trypsin digestion as described in previous work [15]. Trypsinized peptides labelled with light or heavy L-lysine were analysed using nanoelectrospray LC–MS/MS (liquid chromatograph-mass spectrometer/mass spectrometer) with a Q-TOF API-US mass spectrometer (Waters Corporation, Milford, MA). After candidate peptides were determined with 95% confidence, corresponding proteins were identified using Mascot software with minimum confidence of 80% (Matrix Science, Boston, MA) and validated using immunoblotting with specific antibodies.

Sphere forming and re-plating assay

Single suspension cells were cultured in ultralow-attachment 6-well plates (Corning, Corning, NY, USA) at 500 ~ 2000 cells/ml sphere culture medium for 2–3 weeks. Once spheres reached approximately 150 μm in diameter, they were digested with Accutase (Invitrogen, Carlsbad, CA, USA) into single cells for replating culture, flow cytometric analysis or sorting. The sphere culture medium was composed of DMEM/F12 (Hank’s medium) (Invitrogen, Carlsbad, CA, USA) 0.4% bovine serum albumin (Sigma, St. Louis, MO, USA), 4 mM L-glutamine, 1 mM sodium pyruvate, 0.1 mM MEM nonessential amino acids, 20 ng/ml recombinant human epidermal growth factor (Invitrogen Carlsbad, CA, USA), 20 ng/ml basic fibroblast growth factor (Invitrogen, Carlsbad, CA, USA), 5 μg/ml insulin, and and 10 ng/ml leukaemia inhibitory factor (LIF, Peprotech, Rocky Hill, NJ).

Immunofluorescence analysis

To perform immunofluorescence analysis, cells were cultured onto glass slides. After fixation with ice-cold 4% paraformaldehyde for 10 minutes and permeabilization with Triton-X-100 for 30 minutes, cells on the slides were blocked with 3% BSA (bovine serum albumin) for 1 hour and incubated with antibody at 4 °C overnight. Following 5 minutes of washing with PBS 3 times, the slides were incubated with a fluorescein-conjugated IgG antibody (Santa Cruz Biotechnology, CA) in the dark at room temperature for 1 hour. They were also counterstained with propidium iodide for 10 minutes. Confocal images were acquired on a Radiance 2100 confocal laser-scanning microscope (Bio–Rad, Hercules, CA, USA).

Immunoblotting

Immunoblotting was performed as previously described [16]. The CDC50A antibody used in the study was from Santa Cruz Biotechnology (Santa Cruz, CA, USA). All blots were cut prior to hybridisation with antibodies during blotting.

Reverse transcription-polymerase chain reaction (RT–PCR)

Total RNA was extracted from cells using the RNA Mini Kit (Qiagen), reverse-transcribed into cDNA, and amplified for 30 cycles in 25 μl reactions with 10 pmol primers The PCR products were electrophoresed in 1% agarose gels. Amplification of β-action was used as a control. Primers for stem cell-associated genes are listed in Supplementary Table 1.

Quantitative real-time polymerase chain reaction (qRT-PCR)

The extraction of total RNA from cells and synthesis of cDNA have been described on RT-PCR. Specific quantitative real–time PCR experiments were performed using PowerUp™ SYBR™ Green Master Mix (Applied Biosystems™).

Fluorescence-activated cell sorting (FACS) and analysis

Single cells were counted and diluted in HBSS+ buffer (1 × HBSS, 2% foetal bovine serum, 10 mM HEPES, pH 7.2, 1% penicillin-streptomycin, Invitrogen, Carlsbad, CA, USA) to obtain up to 107 cells per mL. Each sample was stained at 4 °C with antibody. Primary cells isolated from clinical tumours were also stained with lineage (Lin) markers, including CD235a, CD45, CD31 and CD140a. After washing, the cells were resuspended and sorted on a Moflo Beckman Coulter (Fullerton, CA, USA) or analysed on an LSRII Fortessa (BD Biosciences). Isotype-matched primary and secondary antibodies were used as controls.

shRNA construction and lentivirus package

Four shRNAs targeting CDC50A were designed and synthesized by GenePharma (Shanghai, China) and cloned into pGPU6/GFP/Neo vectors. After the four shRNA plasmids plus negative control shRNA plasmids were transiently transfected into 293FT cells for 48 hours, the expression of CDC50A was evaluated using Western blotting. Compared with the control shRNA-expressing construct, shCDC50A-974 was the most effective in reducing CDC50A expression (Supplementary Fig. 1). The target sequence of the negative control shRNA was 5′ TTCTCCGAACGTGTCACGT 3′. To generate cells stably expressing the shRNA, shRNA TMEM30A-homo-974 was cloned into the pLKO.1-GFP lentivirus vector, which was then cotransfected with the packaging plasmids pCMV-dR8.91 and pCMV-VSV-G into 293FT cells with PEI transfection reagent. Viruses were harvested 48 hours post transfection and filtered with 0.45 μm syringe filters (Millipore, Milford, MA, USA).

Construction of CDC50A expression vector

Primers for CDC50A transcriptional variant 1 (GENE ID: 55754, HUGO Gene Nomenclature Committee) were designed and synthetised (Qiagen, up 5′ -GCGGAATTCGCCACCATGGCGATGAACTATAAC – 3′; down 5′ -GCCGCGGCCGCTTACTTATCGTCGTCATCCTTGTAATCTCCTCCTCCAATGGTAATGTCAGCTG - 3′, 1086 bp). CDC50A gene was amplified (35 cycles, 25 μl reactions with 10 pmol primers) and verified by sequencing. To generate cells stably up-regulated expressing CDC50A, the PCR amplified product was cloned into the pLVX-IRES-GFP virus vector, which was then cotransfected with the packaging plasmids pCMV-dR8.91 and pCMV-VSV-G into HEK-293 T cells. Viruses were harvested 48 hours post transfection and filtered with 0.45 μm syringe filters (Millipore, Milford, MA, USA).

Mice

All of the procedures involving animals in this study were approved by the Animal Ethics Committee of PUMCH in accordance with institutional and Chinese government guidelines for animal experiments. Nod;Scid (NOD. CB17-Prkdcscid) mice were purchased from the Institution of Laboratory Animal Sciences, Chinese Academy of Medical Sciences, Beijing, China. NSG (NOD. Cg-PrkdCscidIl2rgtm1Wjl/SzJ) mice were obtained from Jackson Laboratory, Bar Harbor, ME, USA. Mice were housed under specific pathogen-free conditions with sterile acidified water and irradiated food. Female mice of 4 to 6 weeks old were used in the study. At least 3 mice were used in each group. Mice were killed by cervical dislocation.

Tumour xenograft

Cells were diluted and mixed with Matrigel (BD Biosciences, San Jose, CA) at a 1:1 ratio. The mixtures were implanted subcutaneously into the scapular region of mice [17]. Tumour development was monitored by palpation and visual inspection twice a week. Mice were killed and xenograft tumours were harvested for subsequent sorting when tumours were confirmed by palpation or visual. If there were no tumours after 5 months of cell implantation, mice were killed and further determine whether the tumours were formed by anatomy. Tumour tissue was confirmed by immunohistochemistry.

Tumour tissue

Fresh tumour tissues or ascites were collected from 23 patients with epithelial ovarian cancer from June 2014 to May 2015. Among them, 16 patients were diagnosed with ovarian high-grade serous carcinoma during primary debulking surgery, and their tumour recurrence outcome and PFI were followed. This work conformed to the guidelines explained in the Declaration of Helsinki and was approved (Approval No. S-072) by the Ethics Committee of Peking Union Medical College Hospital (PUMCH, Beijing, China). Informed consent was obtained from all patients and/or their legal guardians. The pathologic diagnosis for all patients was performed by two experienced gynaecologic pathologists. Tumour specimens were analysed with FACS, and CDC50A+Lin− cells were counted. PFI was defined as the time interval between the dates of completion of standard platinum-based chemotherapy and that of the first confirmed sign of disease recurrence or the last follow-up.

Statistical analysis

Data analyses were performed with SPSS 19.0 software. T tests and one-way analysis of variance (ANOVA) were used to compare differences between two study groups. Survival curve was drawn with univariate Cox regression analysis. Hazard ratio (HR) was calculated after adjusting for the optimal debulking surgery. All the tests were two sided, and p values less than 0.05 were considered statistically significant.