The JAK2/STAT3 pathway inhibitor AG490 was purchased from APExBIO (Apexbio Technology LLC, Houston, TX, USA) and the β-TGF inhibitor SB431542 was obtained from Selleck Chemicals (Houston, TX, USA). Mouse anti-IL6 (cat. no. MAB206) was purchased from R&D Systems, Inc. (Minneapolis, MN, USA). The rabbit anti-α-SMA (cat. no. 55135-1-AP) and rabbit anti-Vimentin (cat. no. 10366-1-AP) antibodies were supplied from Proteintech Group, Inc. (Chicago, IL, USA). Other primary antibodies were provided by Abcam (Cambridge, MA, USA), including anti-E-cadherin antibody (cat. no. ab40772), anti-N-cadhein antibody (cat. no. ab76011), anti-Bax antibody (cat. no. ab32503), anti-Bcl-2 antibody (cat. no. ab32124), anti-caspase-3 antibody (cat. no. ab13847). Alexa Fluor 594-conjugated goat anti-rabbit IgG (H+L) (cat. no. SA00006-4), Alexa Fluor 488-conjugated AffiniPure goat anti-mouse IgG (H+L) (cat. no. SA00006-4) and HRP-conjugated AffiniPure goat anti-rabbit IgG (H+L) (cat. no. SA00001-2) or mouse IgG (H+L) (cat. no. SA00001-1) were purchased from Proteintech Group. The other reagents were obtained from Sigma-Aldrich (Merck KGaA, Darmstadt, Germany).

The human ovarian cancer cell line OVCAR3 was purchased from the cell bank of the Chinese Academy of Sciences (Shanghai, China). The cells were maintained in RPMI-1640 medium supplemented with 10% fetal bovine serum (FBS; Gibco; Thermo Fisher Scientific, Inc., Waltham, MA, USA), 100 U/ml penicillin and 100 mg/ml streptomycin at 37°C in a humidified atmosphere containing 5% CO₂. The cells in the logarithmic growth phase were used for further experiments.

Intraoperative cleaning was performed to cultivate CAFs. Briefly, tissues were washed with cleaning buffer (PBS with penicillin-streptomycin) and the epithelial and adipose tissues were removed. The remaining connective tissues were cut into 1×1×1 mm pieces. The fragments were cultured in DMEM supplemented with 5% FBS and 1% antibiotic-antimycotic solution and incubated at 37°C in a humidified atmosphere containing 5% CO₂. The culture medium was refreshed every 2 or 3 days. After the cellular fusion of the cells, the cell passage ratio was 1:2. The third-generation cells were used for verification.

Normal ovarian specimens removed during myomectomy were selected to cultivate normal fibroblasts (NFs) following the above-mentioned steps. Fresh ovarian cancer tissue was used for the primary culture of ovarian cancer cells. Since fibroblasts are easier to adhere than endothelial cells, most fibroblasts were removed by incubating the cells at different durations.

To confirm CAFs and NFs, the isolated cells grown on slides were fixed with 4% paraformaldehyde for 20 min and then washed with PBS. The cells were then blocked with 10% normal goat serum for 1 h. Subsequently, the slides were incubated with primary antibodies against α-SMA (1:100) or Vimentin (1:100) at 37°C for 1 h, followed by incubation with HRP-conjugated secondary antibody at 37°C for 1 h. Finally, the cells were stained with a DAB substrate (Abcam).

Paraffin-embedded sections of ovarian cancer tissues were obtained in a routine manner. Slides were deparaffinized in xylene and serially rehydrated with alcohol and water, and then epitope retrieval was performed by heating the sections in a microwave oven. The sections were then blocked in 10% normal goat serum for 1 h and incubated with primary antibodies against α-SMA (1:100) or IL-6 (1:100) at room temperature for 1 h. They were then incubated in Alexa Fluor 594-conjugated goat anti-rabbit IgG (H+L) and Alexa Fluor 488-conjugated AffiniPure goat anti-mouse IgG(H+L) secondary antibodies for 30 min. Subsequently, the sections were washed with PBS, counterstained with DAPI (Abcam) at room temperature for 10 min and mounted using fluorescent mounting medium to protect the specimens.

For the dual immunofluorescence staining of CAFs and NFs grown on chamber slides, the cells were fixed with 4% paraformaldehyde for 20 min and then washed with PBS. Subsequently, the cells were blocked with 10% normal goat serum for 30 min and the slides were sequentially reacted with primary and secondary antibody as above described. Following counterstaining with DAPI, the slides were sealed with fluorescent mounting medium and then examined using a fluorescence microscope.

The expression of IL-6 secreted by cancer cells or fibroblasts was evaluated using ELISA. The cells were seeded into six-well plates at a density of 2×10⁵/ml and cultured for 48 h. Subsequently, culture supernatants were collected and centrifuged at 12,000 × g for 10 min and the IL-6 level was determined using an IL-6 ELISA kit (human Quantikine^®; R&D Systems) according to the manufacturer's instructions.

Total RNA was isolated using an Ultrapure RNA kit (CW Biotech, Beijing, China) according to the manufacturer's instructions. Then 1 µg of purified RNA was reverse transcribed into first-strand cDNA using the HiFiScript cDNA Synthesis kit (CW Biotech) and the IL-6 expression at the mRNA level was determined using UltraSYBR Mixture real-time PCR (CW Biotech). β-actin was selected as the housekeeping gene and the specific primers were synthesized by Genewiz (Beijing, China) with the sequences as follows: IL-6 forward, 5′-GTCCAGTTGCCTTCTCCC-3′ and reverse, 5′-GCCTCTTTGCTGCTTTCA-3′; β-actin forward, 5′-CCCGAGCCGTGTTTCCT-3′ and reverse, 5′-GTCCCAGTTGGTGACGATGC-3′. The relative expression of IL6 was calculated using the 2^−ΔΔCt method (16) and each experiment was performed in triplicate.

After treatment for 48 h, the cells were harvested and lyzed in RIPA buffer and the protein concentration was determined using the BCA kit (CW Biotech). Equal amounts of total protein were separated by SDS-polyacrylamide gel electrophoresis (SDS-PAGE) and then electro-transferred onto a polyvinylidene fluoride (PVDF) microporous membrane (Millipore; Merck KGaA, Darmstadt, Germany). The membranes were incubated with primary antibodies at 4°C overnight, followed by incubation with secondary antibodies at room temperature for 1 h. The blots were visualised by ECL chemiluminescence (Millipore; Merck KGaA).

The invasion assays were performed using chambers with polycarbonate filters (8-µm pore size; BD Biosciences, Franklin Lakes, NJ, USA). First, 200 µl cell suspension containing 3×10⁵ cells was seeded into the upper chamber of the insert pre-coated with Matrigel (1:8 dilution; BD Biosciences). The cells were maintained in medium without serum and medium supplemented with 10% FBS was used as a chemoattractant in the lower chamber. After 24 h of incubation, the cells that did not invade through the membrane were wiped off using a cotton swab. Subsequently, the membranes were fixed with 4% paraformaldehyde for 30 min and stained with 0.5% crystal violet. The penetrating cells were quantified on five randomly selected fields.

Following the treatment of the OVCAR3 cells with co-culture supernatant or the transfection with IL-6 shRNA for 24 h, equal numbers of cancer cells (6×10² cells) were incubated in 96-well plates with 100 µl medium for 24, 48 and 72 h. The number of cells was estimated using a CCK-8 assay. Briefly, 10 µl CCK-8 was added to each well and the cells were incubated for 1 h. The absorbance was determined at a wavelength of 450 nm.

The cells were transfected with IL-6 shRNA or treated with co-culture supernatant for 48 h, followed by treatment with paclitaxel (30 nM) for 24 h. Subsequently, the cells were stained with PI and FITC-labelled Annexin-V and analyzed using flow cytometry (BD Biosciences). The cells that were negative for PI, but positive for FITC were considered early apoptotic cells, whereas the cells that were positive for both PI and FITC were considered late apoptotic or necrotic cells. Flow cytometry data were analyzed using the FlowJo software (BD Biosciences).

Between January 1st, 2009 and June 30th, 2014, 269 patients underwent initial treatment at the Department of Gynecologic Oncology of Shandong Cancer Hospital, Affiliated to Shandong University. Patients with FIGO stage IIA-IV ovarian cancer underwent cytoreductive surgery (the scope of surgery included uterus, double accessory, omentum, appendix with or without pelvic or abdominal metastases). Complete clinical data of the patients and pathological paraffin blocks were obtained. Follow-up was performed using the following three methods: Medical records, telephone follow-up and outpatient review. The deadline of the follow-up was June 30th, 2016.

Two observers determined the results based on the expression of interstitial IL-6. For the reading of the pathological sections, a positive result was determined when the cytoplasm and nucleus of the fibroblasts were stained yellow in the stroma of cancer tissues. The proportion of positive cells was assessed by five visual fields, without considering the staining intensity. The expression of IL-6 was divided into two grades: High expression group, in which at least 50% of stromal fibroblasts was positive; low expression group, in which the percentage of positive stromal fibroblasts was <50%.

All statistical analyses were performed using the SPSS version 17.0 software (SPSS, Inc., Chicago, IL, USA). The relationship between the expression of interstitial IL-6 and other clinical features and treatment responses was examined by linear regression and multiple linear regression analysis.

Purified CAFs and NFs were identified by immunofluorescent staining using two antibodies as follows: i) Anti-Vimentin antibody, which is often used for the identification of fibroblasts and ii) anti-α-SMA antibody, which is a marker of activated fibroblasts. The expression of Vimentin was detected in both CAFs and NFs. However, α-SMA was almost not expressed in NFs, while it was overexpressed in CAFs (Fig. 1A). This finding indicated that fibroblasts in the tumor stroma were activated, which are called ‘cancer associated fibroblasts’.

Subsequently, we detected the expression of IL-6 in CAFs and NFs. IL-6 was highly expressed in CAFs expressing the activity marker α-SMA. Furthermore, α-SMA and IL-6 were co-localized in CAFs. In contrast, the expression of IL-6 was weak in NFs with sparse α-SMA expression (Fig. 1B). Similarly, IHC staining of ovarian epithelial carcinoma revealed that IL-6 was also mainly expressed in interstitial cells of ovarian cancer and co-localized with α-SMA in interstitial CAFs (Fig. 1C). This finding indicated that CAFs were the main source of IL-6 secretion in ovarian cancer.

Since IL-6 is a secretory protein, we further assessed the expression of IL-6 in the culture supernatants of CAFs, NFs, primary cancer cells and OVCAR3 cells using an ELISA kit. The expression of IL-6 in CAFs was significantly higher than that in NFs, primary cancer cells and OVCAR3 (Fig. 1D). Detected by RT-PCR, the expression of IL-6 mRNA in CAFs was significantly higher than that in NFs, primary cancer cells and OVCAR3 (Fig. 1E). This result also revealed that the secretion of IL-6 in ovarian cancer was mainly from interstitial CAFs.

OVCAR3 cells were incubated with the culture supernatants collected from CAFs and NFs. The CCK-8 proliferation assay revealed that the culture supernatant collected from CAFs significantly enhanced the proliferation of OVCAR3 cells (Fig. 2A). Similarly, the invasive potential of OVCAR3 cells treated with culture supernatant from CAFs was significantly higher than that of cells treated with culture supernatant from NFs (Fig. 2B). However, with the addition of IL-6 mAb, the effect of CAF supernatant on the proliferation and invasion of OVCAR3 cells was significantly suppressed (Fig. 2A and B). This finding indicated that CAF-derived IL-6 was an important factor in promoting the proliferation and invasion of the OVCAR3 cells.

EMT is an important step through which ovarian cancer cells obtain malignant biological behavior (12,17). Subsequently, we examined the EMT markers of OVCAR3 cells after the treatment of the culture supernatant of CAFs. The results revealed that the culture supernatant of CAFs significantly upregulated interstitial markers N-cadherin and Vimentin and decreased the expression of epithelium marker E-cadherin in OVCAR3 cells, indicating that the EMT of the OVCAR3 cells was significantly enhanced. However, the effect of CAF supernatant on the expression of EMT markers in OVCAR3 cells was significantly attenuated by the addition of IL-6 mAb. There were no significant changes in the EMT markers of OVCAR3 cells after treatment with NF supernatant (Fig. 2C), indicating that CAF-derived IL-6 promoted the EMT of OVCAR3 cells.

In the present study, we treated OVCAR3 cells with culture supernatants of CAFs and NFs. The phosphorylation levels of JAK2 and STAT3 proteins in OVCAR3 cells treated with culture supernatant from CAFs were significantly higher than those of cells treated with culture supernatant from NFs. Furthermore, after the addition of IL-6 mAb, the phosphorylation levels of JAK2 and STAT3 proteins were decreased (Fig. 3A).

AG490, the JAK2/STAT3 signaling pathway specific inhibitor, was applied to OVCAR3 cells treated with the culture supernatant from CAFs. Subsequently, the expression of interstitial markers N-cadherin and Vimentin were decreased and the expression of epithelium marker E-cadherin was increased in OVCAR3 cells. This result indicated that CAF-derived IL-6 mediated the EMT in OVCAR3 cells via the JAK2/STAT3 pathway (Fig. 3B).

OVCAR3 cells were treated with the culture supernatants collected from CAFs and NFs. After 24 h, the cells were treated with paclitaxel to induce apoptosis, which was confirmed by flow cytometry. We found that the number of apoptotic cells was decreased in OVCAR3 cells treated with CAF supernatant. After the addition of IL-6 mAb, paclitaxel resistance was reduced and paclitaxel-induced apoptosis was promoted (Fig. 4A). Furthermore, we assessed the expression of apoptotic-related protein in OVCAR3 cells by western blot analysis. Our results indicated that the expression of pro-apoptotic protein was decreased and the expression of apoptosis-suppressing protein was enhanced in cells treated with CAF supernatant compared with the cells treated with NF supernatant. As displayed in Fig. 4B the addition of IL-6 mAb neutralized the expression of these genes, indicating that CAF-derived IL-6 was involved in paclitaxel resistance of OVCAR3 cells.

SB431542, an inhibitor of TGF-β, was applied to the culture system. Furthermore, EMT in OVCAR3 cells treated with CAF supernatant was inhibited, resulting in the downregulation of the interstitial markers N-cadherin and Vimentin as well as the upregulation of epithelium marker E-cadherin (Fig. 3B). Furthermore, the number of apoptotic cells treated with paclitaxel was increased after the addition of SB431542 and AG490 (Fig. 4C). These findings indicated that CAFs could induce the EMT of cancer cells via the IL-6/JAK2/STAT3 pathway, resulting in increased apoptosis resistance, decreased expression of the pro-apoptotic protein Bax and caspase-3-p17, and increased expression of the apoptosis-suppressing protein Bcl-2 (Fig. 4B). Finally, it may lead to pactitaxel resistance.

Between January 2009 and June 2013, 255 patients underwent cytoreductive surgery and chemotherapy with TP at the Department of Gynecologic Oncology in Shandong Cancer Hospital Affiliated to Shandong University. In Table I the baseline patient characteristics are displayed. A total of 14 patients were excluded due to incomplete clinical and pathological information. We divided the patients into a chemotherapy-sensitive group (n=160) and chemotherapy-resistant group (n=95) according to whether or not the tumor relapsed within 6 months after the last chemotherapy. The patients were divided into a low expression group (n=176, Fig. 5A) and a high expression group (n=79, Fig. 5B) according to the staining of IL-6 protein in interstitial cells. The results revealed that the sensitivity rate of chemotherapy in patients with low expression of IL-6 was 69.3% and the sensitivity of chemotherapy in patients with high expression of IL-6 was 48.1%. The Chi-square test revealed that there was a significant difference between the two groups (P<0.05). Univariate and multivariate analyses revealed that age, CA125, interstitial IL-6 expression and cytoreduction satisfaction were closely related to the sensitivity of the TP regimen in ovarian cancer (P<0.05, Table II).