Recombinant human IL-4 and chemokine (C-C motif) ligand (CCL)-2, also termed macrophage chemotactic protein-1 (MCP-1) were purchased from PeproTech, Inc (Rocky Hill, NJ, USA). The monoclonal antibodies used in the present study were as follows: Rat anti-human IL-10-phycoerythrin (PE; cat. no. 554498), mouse anti-human TNFα-PE (cat. no. 340512), mouse anti-human cluster of differentiation (CD)206 (cat. no. 551135), rabbit anti-human CCL2 (BD Biosciences, San Jose, CA, USA) and mouse anti-human CD14-PE (cat. no. 325606; BioLegend, Inc, San Diego, CA, USA). AZD1480 was purchased from AstraZeneca Pharmaceuticals, (Waltham, MA, USA) and dissolved in dimethyl sulfoxide (Solarbio, Beijing, China) for in vitro investigations. Rabbit anti-human phosphorylated (p-) Janus kinase (JAK)2 (cat. no. 3776; 1:1,000 dilution), rabbit anti-human JAK2 (cat. no. 3230; 1:1,000 dilution), rabbit anti-human extracellular signal-regulated kinase (ERK; cat. no. 4695; 1:1,000 dilution), rabbit anti-human p-ERK (cat. no. 4377; 1:1,000 dilution), rabbit anti-human Akt (cat. no. 4685; 1:1,000 dilution), rabbit anti-human p-Akt (cat. no. 4058; 1:1,000 dilution) and rabbit anti-human p-signal transducer antibodies and activator of transcription (STAT)3 (Tyr705; cat. no. 9145; 1:1,000 dilution) were obtained from Cell Signaling Technology, Inc. (Danvers, MA, USA). Goat anti-mouse/rabbit immunoglobulin (Ig)G secondary antibodies were purchased from Sigma-Aldrich (St. Louis, MO, USA; anti-mouse cat. no. 91618-1G; anti-rabbit cat. no. 43413-1G; 1:1,000 dilution).

The present study was approved by the Ethics Committee of The Affiliated Hospital of Zhengzhou University (Zhengzou, China) in accordance with the Declaration of Helsinki. Written informed consent was obtained from all patients/patient’s family enrolled in the present study.

Primary ovarian cancer cells in ascites fluid were obtained during ascites diagnostic radical surgery, performed in six female patients (age, 25–60 years) with ovarian cancer at the Affiliated Hospital of Zhengzhou University (Zhengzhou, China), and were categorized as malignant Fédération Internationale de Gynécologie et d’Obstétrique stage III serous adenocarcinoma. The ascites fluid was stored in a refrigerator at 4°C for 1–2 h following collection. The supernatant was discarded and the cells (2×10⁵) were resuspended in RPMI-1640 medium (Invitrogen Life Technologies, Carlsbad, CA, USA) and centrifuged at 1,000 × g for 10 min (4°C) in order to isolate the primary ovarian cells. The resulting single tumor cells were cultured under stem cell conditions of serum-free Dulbecco’s modified Eagle’s medium/F12 supplemented (Invitrogen Life Technologies) with 20 ng/ml human recombinant epidermal growth factor (EGF; Invitrogen Life Technologies), 5 μg/ml insulin (Sigma-Aldrich), 0.4% bovine serum albumin (BSA; Sigma-Aldrich), 10 ng/ml basic fibroblast growth factor (bFGF; Invitrogen Life Technologies), 1% penicillin/streptomycin (Sigma-Aldrich) and 1% fungi-zone (Sigma-Aldrich). The samples were incubated at 37°C for 7 days and subsequently split and organized into cell clusters. The cells were analyzed within 7 days by flow cytometry, reverse transcription quantitative polymerase chain reaction (RT-qPCR) and western blotting.

Human dendritic cells were prepared, according to the method previously described by Patterson (21), with minor modifications (22). Briefly, peripheral blood monocytes (PBMCs) from seven female healthy donors (ages, 25–60 years) were isolated from the whole blood (50 ml samples obtained by venipuncture) by Ficoll-Hypaque gradient centrifugation (2,000 × g then 1,500 × g) (Pharmacia, St. Paul, MN, USA) of buffy coats and washed three times with phosphate-buffered saline (PBS; Invitrogen Life Technologies) prior to being resuspended in complete RPMI-1640 medium containing 10% fetal calf serum (HyClone, Logan, UT, USA), 2.5 mg/ml amphotericin B (Sigma-Aldrich), 100 mg/ml streptomycin sulfate (Sigma-Aldrich) and 100 U/ml penicillin. The cells (3×10⁶ cells/ml) were inoculated into a volume of 25 ml in 250 cm² plastic tissue culture flasks (Corning, Inc., Corning, NY, USA) for 45 min at 37°C in 5% CO₂ and were then plated at a density of 2×10⁵ cells/well into 6-well plates (Corning, Inc.) in duplicate. Rinsing with warm PBS removed the non-adherent cells, and the remaining adherent cells were harvested and cultured in complete RPMI-1640 medium. Human recombinant granulocyte macrophage colony-stimulating factor (800 U/ml; Immunex, Seattle, WA, USA) was added to the culture medium to isolate human macrophages.

For the co-culture investigations, 2×10⁵ OCSCs were seeded with an equal number of human PBMC-derived macrophages (2×10⁵) on day 7 of culture and then cultured for a further 3 days in RPMI-1640 medium. For the Transwell co-culture, cultivation was performed, as described previously (23). Briefly, 0.4 μm pore-size Corning Transwell inserts (VWR International, Inc., West Chester, PA, USA) were placed into 6-well plates with 2×10⁵ macrophages, initially seeded at the bottom of the chamber and 2×10⁵ OCSCs seeded onto the inserts, followed by culture for a further 3 days (24).

The macrophages (2×10 ⁵) were stained using anti-human CD14-PE, anti-mouse/human CD206-allophycocyanin (APC)/Cy7 and anti-human E-cadherin-fluorescein isothiocyanate (FITC; panisotype) for 20 min at 4°C to detect the macrophage polarization phenotype. To further detect intracellular cytokines, the macrophages were cultured in the presence of 10 μg/ml Brefeldin A (Sigma-Aldrich) to inhibit the secretion of cytokines (26). The cells (2×10⁵) were then incubated with specific antibodies for the two cytokines IL-10 and TNF-α, and were analyzed by flow cytometry.

The total RNA was isolated using TRIzol reagent (Invitrogen Life Technologies) and the optical absorbance ratio at 260/280 nm was measured using a ScanDrop100 nucleic acid analyzer (Analytik, Beijing, China) to determine the RNA content. Total RNA (5 μg) was reverse-transcribed into cDNA using M-MLV Reverse Transcriptase (Clontech Laboratories, Inc., Palo Alto, CA, USA). cDNAs were used as templates for qPCR. The qPCR mixture contained 5 ml SsoFast^™ EvaGreen Supermix (Bio-Rad Laboratories, Inc., Hercules, CA, USA), 1 ml cDNA (dilution, 1:50) and 2 ml each of the forward and reverse primers (1 mM) to a final volume of 10 ml. qPCR was performed using the cDNA as the template to amplify the mRNA of CCL-2 and cyclooxygenase (COX)-2, using specific primers synthesized by Sangon Biotech. Co., Ltd. (Shanghai, China). The reaction was performed as follows: 10 min at 95°C, followed by 40 cycles of 1 min at 95°C, 2 min at 63°C and 1 min at 72°C, prior to a final annealing step at 72°C for 10 min. The mRNA expression levels of CCL-2 and COX-2 were normalized against human GAPDH using the comparative Ct method (25). The sequences of the primers were as follows: CCL-2, sense 5′-GTTGACCCGTAAATCTGAAGC-3′ and antisense 5′-AAGGCATCACAGTCCGAGTCA-3′ and COX-2, sense 5′-CACATCCTCAATACCAGGTCC-3′ and antisense 5′-CAGAGATCCAGACTCGCATG-3′ (26,27).

Macrophages (2×10⁵) in Transwell plates were lysed using 200 μl lysis buffer (Invitrogen Life Technologies) containing 20 mmol/l HEPES, 25 mmol/l MgCl, 5 mmol/l KCl, 0.5% (v/v) complete protease inhibitor and Triton X-100. The debris was removed by centrifugation at 12,000 × g at 4°C for 10 min. Western blotting was performed using the total protein from cell lysate homogenates. Equal concentrations of the protein (80 μg) were separated using 8% precast SDS-PAGE gels (Invitrogen Life Technologies) and electrophoretically transferred onto a polyvinylidene fluoride membrane (Bio-Rad Laboratories, Inc.). The membrane was subsequently probed with a 1:150 dilution of polyclonal primary YM-1 or PPARγ antibodies (BD Transduction Laboratories, San Jose, CA, USA) or a 1:1,000 dilution of mouse anti-human prostaglandin E2 (PGE2; Cell Signaling Technology, Inc.), Akt, JAK or ERK antibodies, at 37°C for 2 h. Membranes were then washed in PBST five times and then incubated with goat anti-mouse/rabbit IgG (1:5,000; Sigma-Aldrich) secondary antibodies. Bands were visualized using an enhanced chemiluminescence western blot detection system (GE Healthcare Life Sciences, Chalfont, UK) was performed. The intensity of the bands were measured using ImageJ software, version 1.48 (National Institutes of Health, Bethesda, MA, USA). For western blot, GAPDH served as a loading control.

The data were analyzed using SPSS 17.0 statistical software (SPSS, Inc., Chicago, IL, USA) and the data are expressed as the mean ± standard deviation. Methods of least significant difference were applied appropriately to evaluate the differences between various groups. P<0.05 was considered to indicate a statistically significant difference.

The majority of macrophages in the tumor microenvironment exhibit an M2-like phenotype, which is closely associated with tumor development and progression (28). Previous studies have demonstrated that there is a molecular interaction between stem-like cells and macrophages (29,30), however, whether OCSCs affect macrophage polarization, and the specific underlying molecular mechanisms remain to be elucidated. In the present study, following co-culture of OCSCs at equal cell densities for 72 h, under direct cell-cell contact, the expression of CD206, a specific marker of the M2 phenotype, was significantly increased compared with the control (11.7 vs. 40.6%; P<0.05; Fig. 1A). A previous investigation confirmed that macrophages of the M2 phenotype increased the expression of anti-inflammatory cytokines, including IL-10, and inhibited the expression of inflammatory cytokines, including TNFα (30). The present study observed that co-culture with OCSCs under direct cell-cell contact led to a significant increase in the expression of IL-10 (41.3%) and a decreased expression of TNFα (5.8%) in the macrophages compared with the controls (12.5 and 39.5%, respectively; Fig. 1B and C). This result indicated that the macrophages with the M2 phenotype and with anti-inflammatory activity were induced by OCSCs and confirmed that OCSCs had the potential to convert macrophages into the M2 phenotype. To further examine whether the OCSC-induced upregulation of CD206 was dependent on soluble factors and/or direct cell-cell contact, the OCSCs and macrophages were co-cultured in a Transwell system. The results demonstrated that co-culturing of macrophages with OCSCs in Transwells increased the CD206+ macrophage population to a similar extent as under conditions of direct cell-cell contact (40.6 vs. 42.3%; P<0.05; Fig. 1D and E). Additionally, no significant change was detected in the macrophages cultered alone, indicating that the soluble factors secreted by OCSCs, rather than direct cell-cell contact, contributed to macrophage polarization. Changes in macrophages depend predominantly on the specific cytokines expressed, including YM1, YM2 and resistin-like molecule α1, PPAR-γ is the cytokine responsible for increasing the expression of the M2 gene (31). The expression levels of YM-1 and PPAR-γ are used as specific markers of alternatively activated macrophages, therefore, their expression levels were assessed by western blotting. The results revealed a significant increase in the expression levels of YM-1 and PPAR-γ when the macrophages were co-cultured with OCSCs in Transwells, compared with macrophages cultured alone (Fig. 1F and G). This further confirmed that anti-inflammatory M2 macrophages were induced following co-culture with OCSCs.

COX-2 is an inducible enzyme involved in prostaglandin biosynthesis and often overex-pressed in several epithelial malignancies, including ovarian and breast cancer (32,33). The overexpression of COX-2 increases the motility and invasion of cancer cells in several types of tumor, including ovarian cancer (34,35). The CC chemokine, CCL2, is overexpressed in ovarian cancer tumor microenvironments and is important for the progression of ovarian cancer. COX-2 and CCL2 have been implicated in inducing M2-type macrophage polarization (36,37), therefore, the present study evaluated the expression levels of CCL2 and COX-2 by western blot analysis of the samples obtained from Transwell co-cultures of OCSCs and macrophages. The results demonstrated that the expression levels of CCL2 and COX-2 were increased compared with the culture of either alone (Fig. 2A). Quantitative intensity analysis revealed that the expression of CCL-2 was increased ~3-fold and the expression of COX-2 was increased ~5-fold (Fig. 2A; P<0.05). RT-qPCR analysis revealed that the mRNA expression levels of CCL2 and COX-2 were increased in the OCSCs following Transwell co-culture (Fig. 2B). Inhibition of COX-2 significantly decreased the expression levels of YM-1 and PPAR-γ, however, no significant difference was detected following the inhibition of CCL-2 using a CCL-2 antibody (Fig. 2C). These findings suggested that COX-2, but not CCL-2, is important for altering the phenotype of tumor-associated macrophages from M1 to M2 in OCSC-induced macrophages.

Previous studies have suggested that the frequently upregulated expression of COX-2 in tumor cells causes increased PGE2 metabolism, which can facilitate tumor cell transformation, proliferation, angiogenesis and infiltration or migration (38,39). Whether the OCSC-mediated conversion of the macrophage M2 phenotype, via increased expression of the COX-2/PGE pathway, is important remains to be elucidated. In the present study, the expression of PGE2 was further evaluated and the results demonstrated that OCSCs co-cultured with macrophages led to a significant increase in the expression of PGE2 (Fig 3A). The addition of the COX-2 inhibitor, NS398, to the OCSC Transwells prevented the increased expression of PGE2 (Fig. 3A). Following treatment with the anti-PGE2 antibody, OCSCs exhibited decreased expression levels of YM-1 and PPAR-γ (Fig. 3B), indicating that the COX-2/PGE2 pathway is important in altering macrophage polarization.

In mammals, the COX-2/PGE2 signaling pathway is propagated by the activation of several other proteins and pathways, including the transduction and activation of JAK/STAT and the ERK and phosphatidylinositol 3-kinase (PI3K)-Akt pathways. To investigate the intracellular signaling mechanisms responsible for the altered polarization effect of OCSCs on macrophages, the present study examined the changes in the phosphorylation levels of JAK, ERK and Akt in the macrophages following OCSC stimulation. As shown in Fig. 4A, the phosphorylation of Akt was increased following co-culture with OCSCs. Similarly, co-culturing with OCSCs significantly increased the phosphorylation of ERK and JAK (Fig. 4A).

Furthermore, treatment with the anti-PGE2 antibody abrogated the increased expression of p-JAK, however, no significant effect was observed on the expression levels of p-ERK or Akt (Fig. 4B). In addition, inhibition of JAK with AZD1480 caused a significant decrease in the production of YM-1 and PPAR-γ (Fig. 4C and D), suggesting that the JAK signaling pathway exerts important effects in the COX-2/PGE2-induced polarization of M2 macrophages.