The ovarian cancer cell lines A2780 and SKOV3 were obtained from the American Type Culture Collection (Rockville, MD, USA). The cells were cultured in RPMI 1640 medium (Thermo Fisher Scientific, Waltham, MA, USA) supplemented with 10% fetal bovine serum (FBS, Thermo Scientific) and 1% penicillin/streptomycin (Invitrogen Life Technologies, Carlsbad, CA, USA) and maintained at 37 °C in a humidified atmosphere under 5% CO2 conditions.
Spheroid-forming CSCs were generated from A2780 and SKOV3 cells following a method as previously described . Briefly, to isolate spheroid-forming cells, 80–90% confluence of A2780 and SKOV3 monolayers were detached using trypsin/EDTA solution (Thermo Scientific) and seeded in CSC culture medium containing Neurobasal™ medium (Thermo Scientific) with EGF (R&D Systems, Minneapolis, MN, USA), basic fibroblast growth factor (R&D Systems, Minneapolis, MN, USA), penicillin/streptomycin (Invitrogen Life Technologies, Carlsbad, CA, USA), Glutamax supplement (Thermo Scientific), B-27 supplement (Thermo Scientific), and HEPES (Sigma-Aldrich, St Louis, MO, USA) on ultra-low attachment 100-mm2 culture plates (Corning Inc., Corning, NY, USA).
Gene expression omnibus data set analysis
We analyzed the Gene Expression Omnibus (GEO) dataset GSE53759 on ovarian cancer cell lines, including ovarian carcinoma IGROV-1 cell monolayer and spheroid-derived cells (n =3). To see differentially expressed genes between ovarian cancer and normal samples, GSE29450, GSE36668, GSE14001, and GSE69428 were analyzed. GSE124766 was used to see the level of PIK3R3 in high-grade serous ovarian cancer (HGSOC) tumor organoids and tumor tissues. We then used GEO2R(http://www.ncbi.nlm.nih.gov/geo/geo2r/) to identify differentially expressed mRNAs in the abovementioned cell types of the dataset.
Real-Time Quantitative Reverse Transcription PCR
Total RNA from monolayer cells (A2780 and SKOV3) and spheroid cells (A2780-SP and SKOV3-SP) was extracted using TRIzol® (Thermo Scientific) according to the manufacturer’s protocol. RNA (1 µg) was reverse-transcribed to cDNA using an RT First Strand kit (Cat. No. 330401, Qiagen, Valencia, CA, USA) and mixed with Power SYBR GREEN PCR Master Mix (Thermo Scientific) for real-time quantitative reverse transcription PCR (RT-qPCR) on an Applied Biosystems™ StepOne™ Real-Time thermal cycler (Thermo Scientific). The following primers were used: PIK3R3 (forward primer 5′-ATG TAC AAT ACG GTG TGG AGT ATG-3′ and reverse primer 5′-GCT GGA TCC ATT TCA AT-3′); GAPDH (forward primer 5′-GAG AGA CCC TCA CTG CTG-3′ and reverse primer 5′-GAT GGT ACA TGA CAA GGT GC-3′); OCT4 (forward primer 5′- TTTTGGTACCCCAGGCTATG ′ and reverse primer 5′- GCAGGCACCTCAGTTTGAAT-3′). All primers were purchased from Bionics (Seoul, Korea). For quantification purposes, GAPDH was used to normalize the mRNA level. The experiment was performed in triplicate.
Proteins from monolayer cells (A2680 and SKOV3) and spheroid cells (A2780-SP, SKOV3-SP) were collected and extracted using RIPA buffer (Sigma-Aldrich). The Bio-Rad Protein Assay (Bio-Rad, Hercules, CA, USA) was used to determine protein concentration. The protein extract was separated using 10% sodium dodecyl sulfate–polyacrylamide gel electrophoresis and transferred onto a polyvinylidene difluoride membrane (Bio-Rad, Hercules, CA, USA). For blocking, the membrane was incubated with 8% (w/v) nonfat dry milk in PBS–Tween 20 (PBST; 0.05%, Sigma-Aldrich) at 4 °C overnight. After washing with PBST, the membranes were incubated with primary antibodies against PIK3R3 (1:1000, Abclonal, Woburn, MA, USA), OCT4 (1:1000, Abclonal), and GAPDH (1: 5000, Sigma-Aldrich) in PBST at 4 °C overnight. After washing with PBST, the membranes were incubated with HRP-conjugated secondary antibodies (Bio-Rad, Hercules, CA, USA) at room temperature for 1 h, and the protein bands were visualized via enhanced chemiluminescence (GE Healthcare Biosciences, Piscataway, NJ, USA).
PIK3R3 knockdown was performed with an siRNA using the following target sequences: sense (5′-GGA CUU GCU UUA UGG GAA A dTdT-3′) and antisense (3′-dTdT CCU GAA CGA AAU ACC CUU U-5′) (Bionic). The jetPRIME transfection reagent (Polyplus, New York, NY, USA) was used according to manufacturer's instructions. The number of tumor spheroids was determined after 3 and 5 days of transfection, and images depicting CSC morphology were acquired by using EVOS 7500 machine (Thermo Scientific). The experiments were repeated three times.
Spheroid formation assay
Briefly, A2780-SP cells (2 × 103) were seeded in 6-well ultra-low attachment plates (Corning) with or without 10 μM PF-04691502 (Sigma). After 7 days, the number of tumor spheroids was determined, and images were acquired by using EVOS 7500 machine (Thermo Scientific). The experiments were repeated three times.
Gene correlation expression analysis
The starBase database (http:://starbase.sysu.edu.cn/, containing 32 The Cancer Genome Atlas [TCGA]-associated multidimensional datasets, including those for ovarian cancer) was used to study the correlation between the expression levels of PIK3R3 and SOX2 and other genes in TCGA ovarian cancer (OC) cohort. The results were analyzed statistically using the Pearson’s correlation coefficient.
Comparison of protein expression levels
Human protein atlas (HPA; https://www.proteinatlas.org) was used to compare the expression levels of PIK3R3 in normal ovaries and ovarian carcinoma .
Comparison of gene expression levels
TNM plotter (https://www.tnmplot.com) enables real-time comparison of gene expression changes among normal, tumor, and metastatic tissues. This web tool was used to determine PIK3R3 expression using TCGA datasets, and a direct comparison among normal, tumor, and metastatic tissues was conducted using the Mann–Whitney U test .
Receiver operating characteristic plotter for drug sensitivity
To determine the effect of PIK3R3 expression on potential anticancer drug treatment, we used available data from receiver operating characteristic (ROC) plotter (http://www.rocplot.org/), a transcriptome-level open-access database for biomarker validation and independent drug treatment response prediction. Samples with relapse-free survival at 6 months and pathological response to PIK3R3 (202743_at) were used, and the chemotherapy drug Avastin® was used for targeted therapy.
Correlation analysis between PIK3R3 and somatic mutations
muTarget (https://www.mutarget.com/), a tool based on TCGA, provides correlations between mutations and expression of the gene of interest in cancer. The ‘Target’ analysis module was used to find mutations that alter the expression of target genes in ovarian cancer. The candidate gene PIK3R3 was used as an input to determine the correlation between PIK3R3 and somatic mutations in ovarian cancer.
protein–protein interaction data
To identify the interacting partners for PIK3R3 in ovarian tissues, we used the interactome-atlas tool (www.interactome-atlas.org) (Fig. 6A). Filter for tissue expression was selected with ovary.
Kyoto encyclopedia of genes and genomes anlysis
To determine the function of PIK3R3 interacting genes, we performed Kyoto encyclopedia of genes and genomes (KEGG) pathway enrichment analysis . For analysis of KEGG pathway enrichment, we used the web-based DAVID v6.8 (https://david.ncifcrf.gov/tools.jsp).
All data are represented as the mean ± standard deviation of the mean (SD) and all the analyses were carried out at least three times. For the statistical comparisons, the Student's t-test was performed using GraphPad Prism 9 (GraphPad Software, La Jolla, CA, USA).