Cell culture and conditioned medium preparation
Human epithelial ovarian carcinoma cell lines A2780 and A1847 were gifts from Dr. Stuart Aaronson (Mount Sinai School of Medicine, NY, USA). The TOV-112D, SKOV3, IOSE and ES-2 cell lines were purchased from ATCC (Manassas, VA, USA). The SKOV-I6 and A2780-I4 cell lines were selected from the parental cell lines SKOV3 and A2780 by 6 or 4 times of in vitro invasion assays, respectively . The OVS1 cell line was established from a serous type of human ovarian tumor obtained from Taipei Veterans General Hospital (TVGH), Taiwan [33,34,35]. IOSE cells were cultured in medium 199:MCDB 105 (1:1) medium contains 10% fetal bovine serum (FBS) (Sigma-Aldrich, St Louis, MO, USA). Cells except IOSE, were grown in Dulbecco's Modified Eagle’s Medium (DMEM), supplemented with 10% FBS and penicillin/streptomycin (Invitrogen, Waltham, MA, USA). Human monocytic leukaemia THP-1 cells were grown in RPMI 1640 medium, supplemented with 10% FBS and penicillin/streptomycin (Invitrogen). Normal human adipose derived stromal cells (hADSC) were a gift from Dr. Chang Cheng-Chi, Graduate Institute of Oral Biology, National Taiwan University, Taipei, Taiwan. hADSCs were grown in DMEM/F12 (1:1) medium (Invitrogen) and supplemented with 10% FBS and 1 ng/mL basic fibroblast growth factor (R&D Systems, Minneapolis, MN, USA). All the cells were maintained in the exponential growth phase at 37° C in 5% CO2. For collection of conditioned medium, 1 × 106 cells were seeded in a 10-cm culture dish with DMEM and 10% FBS overnight. Next day, the culture medium was substituted with RPMI 1640 medium containing 0.5% FBS for an additional 24 h. The conditioned medium was collected and filtered through a low protein binding 0.45-μm polyvinylidene fluoride (PVDF) filter (Millipore, Burlington, MA, USA) and stored in a – 80 °C freezer before use. Reagents used in this study are listed in Additional file 1: Table S1.
Exon array analysis
Briefly, total RNA of cell lines were extracted, and analyzed by Affymetrix Human Exon 1.0 ST Array (Affymetrix, Santa Clara, CA, USA). The raw data was normalized and analyzed by GeneSpring software (Agilent Technologies, Santa Clara, CA, USA), followed by Gene Ontology annotation analysis, and classified into GO molecular function domains. Genes that expression changes greater than 1.5 fold were selected for heatmap analysis. The heatmap was plotted in R with pheatmap package (Lucent Technologies, Murray Hill, NJ, USA). Color scale bar indicates log2 fold change after column normalization.
Public domain data analysis
The sources of gene expression profiling and clinical pathological characteristics of Tothil cohort (GSE9891) were downloaded from Oncomine (www.oncomine.org). The median of POSTN expression levels were used as cut off points for the overall survival analysis.
Gene expression manipulation
For overexpression experiments, POSTN cDNA was subcloned into the pcDNA4 mammalian expression vector (Invitrogen) and was transfected into low POSTN expressing cells. POSTN stably expressing cells were maintained in medium containing zeocin. For knockdown experiments, the high POSTN expressing cells were infected with pLKO.1-POSTN or control pLKO.1-Luc lentiviral vector (National RNAi core facility, Academia Sinica, Taipei, Taiwan) and were maintained in the complete medium containing puromycin. siRNA and shRNA clones used in this study are listed in Additional file 1: Table S2.
Western blot analysis
Total cell lysates were prepared in RIPA lysis buffer supplemented with protease inhibitor and phosphatase inhibitor cocktails. Proteins were separated by SDS-PAGE, transferred onto the PVDF membrane and non-specific binding was blocked 1 h in 5% skim milk. Proteins were detected by applying respective specific primary antibody for 1 h, followed by incubation with horseradish peroxidase-conjugated secondary antibody for 1 h at room temperature. The chemiluminescence system was used to visualize the signals. Antibodies against specific proteins in this study are listed in Additional file 1: Table S3.
Migration and invasion assays
The in vitro migration and invasion assays were performed using Transwells (Corning, Corning, NY, USA). For the cancer cell migration assay, 2.5 × 104 cells were seeded in uncoated Transwells, with 8.0 μm pores (Corning). For the invasion assay, 1 × 105 cells were seeded in a Matrigel-coated chamber with 8.0 μm pores (BD Bioscience, Franklin Lakes, NJ, USA). The migrated or invaded cells in polycarbonate membranes were stained and counted using previously described methods [36, 37]. For THP-1 monocytic cell migration assay, 2.5 × 105 cells were seeded onto the inserts with a porous PET (Corning) membrane (pore size, 8.0 μm) in 350 μL of RPMI 1640 supplemented with 0.5% FBS. The inserts were placed in the wells, which contained cancer cell-derived conditioned medium, and incubated for 6 h at 37 °C. The migration of the THP-1 cells was visualized under a microscope and quantified by counting three randomized 100 × fields of the migrated cells on the underside of insert.
Quantitative RT-PCR analysis
The qRT-PCR was performed as described previously . Total RNA from the cultured cells was extracted using TRIzol reagent (Invitrogen) following the protocols recommended by the manufacturer. First-strand cDNA was generated by the SuperScript® III First-Strand Synthesis System (Invitrogen) using oligo-dT primer. The KAPA SYBR FAST Universal qPCR Kit (KAPA Biosystems, Wilmington, MA, USA) was used for gene detection in a CFX96 real-time PCR detection system (Bio-Rad, Hercules, CA, USA). The 2−∆CT or 2−∆∆CT method was used to calculate the relative expression of specific genes. The expression level of the β-actin (ACTB) housekeeping gene was used as an internal control. The primer sequences are listed in Additional file 1: Table S4. The annealing temperature for all the primer pairs was 60 °C.
The expression level of secreted TGF-β1 or TGF-β2 was detected by using human TGF-beta1 ELISA kit (#ELH-TGFb1) or human TGF-beta2 ELISA kit (#ELH-TGFb2-1) (RayBiotech, Peachtree Corners, GA, USA) according to user manual. Briefly, standard or sample were added to each pre-coated well and incubate 2.5 h at room temperature with gentle shaking. Add biotinylated antibody, streptavidin solution and finally add TMB One-Step substrate reagent for color visualization. The protein levels of TGF-β1 or TGF-β2 were detected under 450 nm by ELISA reader.
Ovarian tumor RNA samples were obtained from the Department of Obstetrics and Gynecology, China Medical University Hospital, Taichung, Taiwan, with the approval of the institutional review board (IRB#CMUH107-REC1-095). Informed written consent was obtained from all the patients who participated in the study. Tumor samples were collected during debulking surgery. The identities of the patients from whom the pathological specimens were obtained remained anonymous. In addition, a commercial ovarian cancer cDNA was purchased from OriGene (Rockville, MD, USA) and used in this study. Detail information was summarized in Additional file 1: Tables S5 and S6.
In vivo xenograft studies
Animal experiments were conducted as previously described . Three xenograft models were used in this study: subcutaneous (s.c.), intraperitoneal (i.p.) and orthotopic metastasis models. In the subcutaneous and intraperitoneal models, SKOV3-derived cells (1 × 106) were harvested and resuspended in 100 μL of PBS. Tumor xenografts were established by injecting SKOV3-derived cells into the dorsal flank (s.c.) or abdominal cavity at right lower quadrant (i.p.) of mice. In the orthotopic metastasis model, SKOV3 or its derived highly invasive subline SKOV-I6 cells (1 × 106) were resuspended in 20 μL of PBS containing 50% Matrigel (BD Biosciences) and intra-bursally injected into ovary capsules of the mice. Female NOD/SCID mice (aged 6–8 weeks, National Laboratory Animal Center, Taipei, Taiwan) were randomly assigned to each experimental group. For the s.c. tumor growth model, the dimensions of the xenografts were measured by callipers every week and tumor volume was calculated using the formula V = (π × length × width2)/6. For the i.p. model, tumor growth and abdominal metastases were monitored by in vivo imaging system (PerkinElmer, Waltham, MA, USA). The mice were sacrificed 4 weeks (s.c. and i.p. models) or three weeks (orthotopic metastasis model) after inoculation of the tumor cells. All the xenografts were fixed in 10% neutral buffered formalin, embedded in paraffin and cut consecutively into 4-µm sections for further pathological examination after hematoxylin and eosin staining and for subsequent immunohistochemistry analysis.
For the FACS analysis, the differentiated M1 and M2 macrophages were collected and stained with anti-CD68 (eBioscience, Waltham, MA, USA), anti-CD206 (eBioscience) or anti-CD80 (eBioscience) antibody following the protocols recommended by the manufacturer. Antibodies against specific proteins in this study are listed in Additional file 1: Table S4.
Immunohistochemistry and immunofluorescence
Briefly, paraffin-embedded ovarian cancer tissue sections (4 μm) on poly-l-lysine-coated slides were deparaffinized and rinsed with 10 mM Tris–HCl (pH 7.4) and 150 mM sodium chloride. Peroxidase was quenched with methanol and 3% hydrogen peroxide. The slides were then placed in 10 mM citrate buffer (pH 6.0) at 100 °C for 20 min in a pressurized heating chamber. After incubation with POSTN (1:100), F4/80 (1:100), CD206 (1:500) or α-SMA (1:500) primary antibodies individually for 1 h at room temperature, the slides were thoroughly washed three times with PBS. Bound antibodies were detected using the EnVision Detection Systems Peroxidase/DAB, Rabbit/Mouse kit (Dako, Glostrup, Denmark). The slides were then counterstained with haematoxylin. Finally, the slides were photographed under a microscope (BX50, OLYMPUS, Tokyo, Japan). Negative controls were obtained by performing all the steps but omitting the primary antibodies. For immunofluorescence staining, cells were fixed with 4% paraformaldehyde/PBS for 30 min followed by permeabilizing cells with 0.5% Triton X-100/PBS for 15 min at room temperature. Nonspecific binding sites were blocked using 1% BSA/PBS for 1 h. Primary specific antibody was applied for 1 h, followed by fluorophore-conjugated secondary antibody incubation for additional 1 h. Cells were washed with 0.1% Tween 20/PBS (PBST) and were mounted with anti-fading agent. Images were captured under Leica TCS SP5 (Leica Microsystems, Wetzlar, Germany) and were processed using LASX software (Leica Microsystems). Antibodies against specific proteins in this study are listed in Additional file 1: Table S3.
Proximity ligation assay (PLA)
The PLA was applied for in situ detection of endogenous protein–protein interaction between POSTN and integrin β3 or POSTN and integrin β5. This assay was carried out as described in user manual (Duolink® PLA DUO92008, Sigma-Aldrich). In brief, cells were seeded on glass coverslips overnight then fixed by ice-cold 100% methanol for 15 min at − 20 °C. Aspirated fixative, rinsed three times in PBS then added blocking solution to samples for 30 min at 37 °C. Primary antibodies anti-POSTN antibody (SC-46655, Santa Cruz Biotechnology, Dallas, TX, USA), anti-integrin β3 antibody (#13166, Cell Signaling Technology, Danvers, MA, USA) and anti-Integrin β5 antibody (#3629; Cell Signaling Technology) were diluted 1:100 in buffer and added to samples for overnight at 4 °C. Washed in PBST then followed by applying PLA plus and minus probes to samples for 1 h at 37 °C. Washed in PBST twice. Ligation was performed by ligase application for 30 min at 37 °C, followed by amplification with polymerase for 2 h at 37 °C. Washed samples with 2 × SSC twice then 0.2 × SSC once. Mounted samples with mounting media with DAPI. Signal was detected by confocal microscopy.
In vitro co-culture model
Co-culture of THP-1 with ovarian cancer cells was done in RPMI 1640 supplemented with 10% FBS in 6-well cell culture inserts with a permeable PET membrane (pore size, 0.4 μm). The tumor cells were seeded at a density of 2 × 105 cells per insert in 1.5 mL of medium. The THP-1 cells were seeded at a density of 2 × 105 cells in the lower compartment in 3 mL of medium. After co-culturing for 5 days, the THP-1 cells were harvested for FACS analysis, RNA extraction and qRT-PCR analysis.
Kaplan–Meier analysis using the p value of the log-rank test was applied to determine the power of POSTN as a marker for overall patient survival. All data are the mean ± SD, unless otherwise specified. The means ± SD represent data from three independent experiments. One-way ANOVA was used to compare the means among three or more independent groups. Student’s t-test was used to compare the means between two groups to be compared. A p value of less than 0.05 was considered statistically significant.