SGC7901 human gastric cancer cell line, BGC823 human gastric cancer cell line and BJ cells human normal fibroblast cell line were purchased from American Type Culture Collection (ATCC, Manassas, VA). All cell lines were maintained with 1640 complete medium (Gibco, Grand Island, NY) supplemented with 10% fetal calf serum (Gibco), at 37°C with 5% CO₂ atmosphere. Female nude mice (6-8 weeks) were purchased from Beijing HFK Bioscience (Beijing, China) and fed in SPF experimental animal rooms.

Primary CAFs were prepared by the outgrowth methods [12]. Primary gastric cancer tumor and adjacent normal samples were sterilely obtained after the surgery at the Zhejiang Provincial People’s Hospital (Table 1). Samples were sent to the laboratory within 2 hours. According to the clinical data, samples were divided into chemo-sensitive and chemoresistant groups. The samples were minced and cultured with F12 culture medium (HyClone, Los Angeles, CA) containing type II and IV collagenase (Sigma, San Francisco, CA) to digest at 37°C, 5% CO2 incubator for 2 hours, followed by filtration (BD Biosciences, Franklin Lakes, NJ). Then the cell precipitation was collected and seeded into 6-well plate in 2 mL F12 medium supplemented with 10% fetal bovine serum for 4 hours at 37°C. Next, the medium was replaced with fresh medium to remove the un-adherent cells and the remaining cells were collected. After 2-4 passages, we sorted the CD90 positive cells [13] with BD FACSAria III to collect fibroblasts and cultured for further study.

To obtain the cultured medium of CAFs (CAFs-CM), 2×10⁵ separated CAFs were seeded in the plates and cultured under conditioned circumstances. Twenty-four hours later, the medium was replaced by serum-free medium. Two days later, the supernatants were obtained and stored at –80°C for further use.

For the detection of CAF effect on chemotherapeutic drugs on BGC923 and SGC7901, we co-cultured CAFs (10⁵) and cancer cells (5×10⁴) for 24 hours, and we removed CAFs by labeling CD90. Then BGC923 and SGC7901 were seeded, followed by adding drugs as following: cisplatin (DDP; 2, 4, 6, 8, 10, 12, and 16 μg/mL; Sigma), doxorubicin (1, 5, 25, 50, 200, 500, and 800 μM; Sigma) or etoposide (0.01, 0.1, 1, 2, 20, 200, and 500 μM; Sigma) for 48 hours.

To examine the role of CAFs-CM (collected after 10⁵ CAFs culture for 24 hours) or rhIL-11 on cancer cells resistance to chemotherapy, we pre-treated the cancer cells with CAFs-CM or rhIL-11 (2.5, 5, 10, 20, and 50 ng/mL) for 24 hours, then employed 10 μg/mL DDP, 200 μM Etoposide and 20 μM doxorubicin for BGC823 cells and 8 μg/mL DDP, 200 μM Etoposide and 200 μM doxorubicin for SGC7901 cells at the presence or absence of IL-11 neutralizing antibody (25 μg/mL) or ruxolitinib (5 μM) for 48 hours. The apoptosis of cancer cells from tumor tissues were detect by the Annexin V and propidum iodide (PI) apoptosis detection kit (BD Biosciences) as guided.

Under some experiments, we used STAT3, IL-11R or Bcl2 knockdown BGC823 or SGC7901 cells to examine the effect of IL-11 on cell viability when treated with chemotherapeutic agents.

All cell viability experiments were determined by MTT assay. Briefly, cells were seeded into 96-well culture plates at the density of 3,000 per well. Twelve hours later, cells were treated at the determined conditions. Cell growth was measured after addition of 10 μL MTT solution (0.5 mg/mL, Solarbio, Beijing, China). After 4 hours incubation at 37°C, the medium was replaced with 100 μL dimethylsulfoxide and vortexed for 10 minutes. Absorbance was measured at 570 nm by a microplate reader (Bio-Rad, Hercules, CA). Each experiment was performed for at least three times.

siRNA targeted human STAT3 (siRNA#1, GUGAAGUCAACAUGCCUGC; and siRNA#2, GCAGGCAUGUUGACUUCAC) and siRNA targeted human BCL-2 (siRNA#1, GCAGCAGCTGAACAACAT and siRNA#2, AATTAAAAAAGCAGCA) and control siRNA were purchased from RiboBio (Guanghzou, China). siRNA (50 μM) was transfected to gastric cancer cells with lipofectamine RNAiMax (Invitrogen, Carlsbad, CA) as guidance. Lentivirus vector with shRNA targeting IL-11 receptor (5'-GGACCATACCAAAGGAGAT-3') and negative control were purchased from Genechem Company (Shanghai, China).

To examine the percentage of CAFs in gastric cancer patients tissues, cells were incubated with CD90 (eBioscience, San Diego, CA) for 20 minutes at room temperature, after washed twice and then re-suspended in phosphate buffered saline (PBS). Flow cytometry was performed on the BD Canto II (BD Biosciences). 7-AAD was used to exclude the dead cells. IgG (eBioscience) was used as the negative control.

To detect the effect of IL-11 on apoptosis of cancer cells when treated with chemotherapeutic agents, SGC7901 cells (1×10⁶ per mouse) were subcutaneous injected into the nude mice. When the tumor volume reached 5 mm×5 mm, doxorubicin (5 mg/kg) monotherapy or combined with CAFs-CM, rhIL-11 (2.5 μg/kg), in the presence or absence of IL-11 neutralizing antibody (25 μg/mL) were used to treated the mice by tail intravenous injection every 2 days. PBS was used as control group. Mice were treated for 2 weeks. Then the mice were sacrificed to obtain the tumor tissues. After mincing and digesting the tumor tissues, the tissue suspension was filtered and washed by PBS, followed by staining with CD45 (eBioscience), Annexin V (eBioscience), and PI (eBioscience). Then flow cytometry was used to examine the apoptosis rate.

Two-microgram cDNA was performed as a template for the real-time polymerase chain reaction (PCR) analysis, with a SYBR Green Mix (Thermo Fisher, Waltham, MA) and Agilent Technologies Stratagene Mx3500P real-time PCR system. The ΔΔCt method was used for the relative quantification with actin as a reference. The primers used are listed as follows: human IL-11 forward primer 5′-CGAGCGGACCTACTGTCCTA-3′ and reverse primer 5′-GCCCAGTCAAGTGTCAGGTG-3′; human SDF-1 forward primer 5′-ATTCTCAACACTCCAAACTGTGC-3′ and reverse primer 5′-ACTTTAGCTTCGGGTCAATGC-3′; human HGF forward primer 5′-GCTATCGGGGTAAAGACCTACA-3′ and reverse primer 5′-CGTAGCGTACCTCTGGATTGC-3′; human VEGFα forward primer 5′-AGGGCAGAATCATCACGAAGT-3′ and reverse primer 5′-AGGGTCTCGATTGGATGGCA-3′; human FGF forward primer 5′-TGGCTATTTGGTGGGGATCAA-3′ and reverse primer 5′-GAAGAGGGCACTTCTCACTCC-3′; human PDGF forward primer 5′-GCAAGACCAGGACGGTCATTT-3′ and reverse primer 5′-GGCACTTGACACTGCTCGT-3′; human IL-11R forward primer 5’-CTGGGCTAGGGCATGAACTG-3’ and reverse primer 5′-CTGGGACTCCAAGTGCAAGA-3′; human SDF-1R forward primer 5′-ACTACACCGAGGAAATGGGCT-3′ and reverse primer 5′-CCCACAATGCCAGTTAAGAAGA-3′; human VEFGR forward primer 5′-AGGGCAGAATCATCACGAAGT-3′ and reverse primer 5′-AGGGTCTCGATTGGATGGCA-3′; human HGFR forward primer 5′-AGCAATGGGGAGTGTAAAGAGG-3′ and reverse primer 5′-CCCAGTCTTGTACTCAGCAAC-3′; human PDGFR forward primer 5′-TGGCAGTACCCCATGTCTGAA-3′ and reverse primer 5′-CCAAGACCGTCACAAAAAGGC-3′; human FGFR forward primer 5′-ACTACACCGAGGAAATGGGCT-3′ and reverse primer 5’-CCCACAATGCCAGTTAAGAAGA-3′; human GAPDH forward primer 5'-GGAGCGAGATCCCTCCAAAAT-3', reverse primer 5'-GGCTGTTGTCATACTTCTCATGG-3'. The expressions of IL-1 F9 (Hs00219742_m1) and IL-1 F9R (Hs002136-00_m1) were detected by using TaqMan primer sets (Applied Biosystems, Foster City, CA) [14].

Cancer associated fibroblasts or BJ cells were seeded at the density of 5×10⁵ cells per well. Forty-eight hours later, the cells were washed twice with PBS and incubated with 2 mL serum-free Dulbecco's modified Eagle's medium for 24 hours. Then the supernatants were collected and stored at –80°C for further use. The levels of IL-11 in the CM were examined by using enzyme-linked immunosorbent assay kits (R&D Systems Inc., Minneapolis, MN), according to the manufacturer’s recommended protocol. Each experiment was performed in triplicate.

Whole cell lysates were prepared from BGC823 or SGC-7901 cells treated as previous designed and separated by sodium dodecyl sulfate polyacrylamide gel electrophoresis at 100 V for 1 hour. Next, the separated proteins were transferred to NC membranes (Millipore, Billerica, MA). The samples were blocked with 5% bovine serum albumin (BSA) in TBS containing 0.1% Tween-20 for 1 hour at room temperature and incubated overnight at 4°C with one of the following antibodies: JAK (1:200, Abcam, Cambridge, UK), phospho JAK (Ser473) (1:200, Abcam), STAT3 (1:300, Abcam), phospho STAT3 (Tyr705) (1:300, Abcam), BCL-2 (1:200, Abcam), GP130 (1:300, Abcam) and actin (1:500, Abcam). The samples were washed with TBST and incubated with HRP-conjugated secondary antibodies for 1 hour at room temperature. Proteins were visualized by ECL western blotting reagent (Thermo Fisher).

Tumor tissues were kept in 4% paraformaldehyde overnight, then processed, embedded in paraffin, and sectioned at 4 μm for further study. To detect the expression of JAK and STAT3 in gastric cancer tissues, antigen retrieval was done using citric acid and sodium citrate in a Microwave oven (Media, Guangdong, China). Then the sections or gastric cancer cells (fixed with 4% paraformaldehyde and permeabilized with 0.3% Triton X-100) were blocked with 5% BSA in PBS and incubated with p-JAK (1:200, Abcam) or p-STAT3 (1:500, Abcam) at 4°C overnight, followed by secondary antibodies (Thermo Fisher) for 1 hour at room temperature. Nuclei were stained with the DAPI solution (1 μg/mL). Confocal microscope (Olympus, Tokyo, Japan) was used to visualize the sections.

The sections of tumors were incubated with IL-11 (1:500, Abcam), IL-11 receptor (1:500, Abcam) or α-smooth muscle actin (α-SMA; 1:500, Abcam) at 4°C overnight, followed by signal amplification using an ABC HRP Kit (Thermo Fisher) and counter-staining with hematoxylin, dehydration with series of graded ethanol and cleaned with xylene. Microscope (Leica, Oskar-Barnack, Germany) was used to visualize the sections.

To test cell death and proliferation in patients’ samples, the sections of tumors were incubated with the mixture of terminal deoxynucleotidyl transferase, nucleotide, and reaction buffer as indicated by TUNEL Cell Death Detection Kit (Roche, Basel, Switzerland). Apoptosis percentage was calculated by percent of tumor cells with positive staining.

To evaluate the anticancer effects of ruxolitinib combined with chemotherapeutic agents, SGC7901 cells (1×10⁶ per mouse) were subcutaneous injected into the nude mice. When the tumor volume reached 5 mm×5 mm, PBS, ruxolitinib (10 mg/kg), DDP (5 mg/kg), doxorubicin (5 mg/kg), etoposide (10 mg/kg), and ruxolitinib (10 mg/kg) combined with doxorubicin (5 mg/kg), DDP (5 mg/kg), etoposide (10 mg/kg) were used to treated the mice by tail intravenous injection every 2 days. Mice were treated for 3 weeks. To establish drug resistant animal models, SGC7901 cells (1×10⁶ per mouse) were subcutaneous injected into the nude mice. When the tumor volume reached 5 mm×5 mm, 50 ng IL-11 was injected into the mice by intratumor injection. After 4 days, PBS, ruxolitinib (10 mg/kg), DPP (5 mg/kg), doxorubicin (5 mg/kg), etoposide (10 mg/kg), and ruxolitinib (10 mg/kg) combined with doxorubicin (5 mg/kg), DPP (5 mg/kg), etoposide (10 mg/kg) were used to treated the mice by tail intravenous injection every 2 days. Mice were treated for 3 weeks. Tumor growth was recorded with the length (L) and width (W) of tumors by vernier calipers, and the tumor volume (V) was calculated by the formula V=(L×W²)/2. The data was monitored every 2 days and the survival of tumor bearing-mice was observed every day.

Results were presented as mean±standard error of mean and statistical significance was examined by an unpaired Student’s t test by the Graphpad 6.0 software (San Diego, CA). Survival analysis was performed by Kaplan-Meier method and evaluated using the log-rank test. p-value of < 0.05 was considered as statistically significant.

All studies were approved by the animal care and used committee of laboratory animal monitoring of People’s Hospital of Hangzhou Medical College. All the animal experiments were approved by the IACUC in People’s Hospital of Hangzhou Medical College.

The tumor stroma consists of various elements including CAFs, macrophages, lymphocytes and extracellular matrix, which are involved in the tumor initiation, metastasis, acquiring drug resistance and so forth [15,16]. Among these, CAFs is reported as an important component to mediate drug resistance [17,18]. To assess the expression landscape of CAFs in gastric cancer, we isolated fibroblasts from clinical gastric cancer patients’ samples resistant or sensitive to chemotherapy, and the CAFs are labeled as CD90⁺ by flow cytometry. We found that the percentage of CAFs in samples from chemo-resistant tissues was significantly higher compared with samples from chemo-sensitive tissues (Fig. 1A). The higher expression of CAFs in chemo-resistant tissues was also verified using α-SMA (Fig. 1B). These data indicated that the enriched CAFs may be involved in the process of gastric cancer cells chemo-resistance development. Herein, we isolated the CAFs from the tumor tissues and treated gastric cancer cell lines, BGC823 and SGC7901, with clinical chemotherapeutic agents (doxorubicin, DDP, and etoposide) after co-culture with or without those CAFs. We observed that the gastric cancer cells were more resistant to chemotherapeutic drugs in the CAFs co-culture conditions compared to monoculture (Fig. 1C-H). These data suggested that CAFs in gastric tumor sites result in the cancer cells drugs resistance development.

Current thinking suggests that CAFs are capable of remodeling the surrounding cells through secreting soluble factors [5,19]. In line with this concept, we added the conditioned medium from CAFs to the gastric cancer cells followed by chemotherapeutic drugs treatment respectively. And we observed enhanced drug resistance in BGC823 cells treated with the CAFs-CM compared with control monoculture (Fig. 2A). Similar results were obtained when tested with SGC7901 (Fig. 2B), indicating that CAFs mediated chemoresistance of gastric cancer cells through soluble factors. To further investigate the exact cytokines of CAFs-induced resistance, we screened the expressions of major cytokines (IL-11, stromal cell-derived factor-1, hepatocyte growth factor, fibroblast growth factor, platelet-derived growth factor, vascular endothelial growth factor, and IL-1 F9) by CAFs and normal fibroblasts. And we observed that the expression of IL-11 was significantly higher in CAFs than that from normal samples (nearly 90-fold), while no significant different expression of the other cytokines was observed (Fig. 2C and D). Correspondingly, the IL-11 receptor had higher expression in SGC7901 cells co-cultured with CAFs-CM compared with monoculture (Fig. 2E). Based on these data, we hypothesized that IL-11 may work as a key player to mediate drug resistance in gastric cancer cells. So, we applied gradient recombinant human IL-11 to pre-treat BGC823 and SGC7901 cells followed by chemo-therapeutic drugs treatment. We found that rhIL-11 significantly enhanced the drug resistance in a dose dependent manner (S1A and S1B Fig.). To further verify this, we treated BGC823 cells with CAFs-CM or recombinant human IL-11 in the presence or absence of IL-11 neutralizing antibody. In line with our hypothesis, BGC823 cells treated with IL-11 or CAFs-CM showed enhanced drug resistance while the addition of IL-11 neutralizing antibody in CAFs-CM reversed the acquired drug resistance in gastric cancer cells (Fig. 2F). And testing SGC7901 cells achieved parallel results (Fig. 2G). While knocking down IL-11 receptor reversed the drug resistance (S1C and S1D Fig.). To further demonstrate the acquired drug resistance of gastric cancer cells induced by the IL-11, we injected CAFs-CM, recombinant human IL-11, CAFs-CM combined with IL-11 neutralizing antibody into the gastric tumor-bearing mice. Then we treated the mice with doxorubicin by tail vein injection. Interestingly, we observed that co-injecting with CAFs-CM or rhIL-11 remarkably suppressed the doxorubicin efficacy to inhibit tumor volume, while addition of IL-11 neutralizing antibody in CAFs-CM reversed the drug resistance (Fig. 2H). In consistent, combining doxorubicin with IL-11 neutralizing antibody prolonged the survival time of mice (Fig. 2I), which is in line with our results in vitro. Together, these data suggested that IL-11 secreted by CAFs is involved in the gastric cancer cells to acquire chemo-resistance. Moreover, we observed enhanced IL-11 and IL-11 receptor expression in tumor tissues from chemo-resistant gastric patients compared to the samples from chemo-sensitive gastric patients (Fig. 2J and S1E Fig.). And the expression of IL-11 and IL-11R expression has no influence to the cell survival or apoptosis in tumor tissues (S1F and S1G Fig.). Above data reminded us that accumulated CAFs participate in the drug resistance of gastric cancer cells through the secretion of IL-11.

In tumor microenvironment, cytokines are often characterized as functional redundancy and tissue-specific activities [20,21]. Current studies have shown that over-activated IL-11 contributed to the activation of the gp130 through binding to its receptor in solid tumors [6,11], which afterwards activated JAK/STAT3 pathway that was involved in regulating angiogenesis, metastasis, resistance to apoptosis, cell proliferation, and immune evasion [22]. Herein, we supposed that IL-11 might participate in the drug resistance development through the activation of JAK/STAT3 pathways. So we detected the p-JAK and p-STAT3 expression in BGC823 and SGC7901 cells treated with CAFs-CM, rhIL-11, and CAFs-CM combined with IL-11 neutralizing antibody. Increased expression of p-JAK and p-STAT3 was observed in cells treated with CAFs-CM or rhIL-11 while the addition of IL-11 neutralizing antibody in CAFs-CM reversed the phenomenon in both BGC823 and SGC7901 cells (Fig. 3A and B), indicating that IL-11 might activate the JAK/STAT3 pathway in gastric cancer cells. To further expound the signaling pathway induced by IL-11, we detect the expression of GP130, JAK, and STAT3 by western blotting and found the activation of GP130/JAK/STAT3 signaling pathway under the condition of CAFs-CM or rhIL-11 treatment, while adding IL-11 neutralizing antibody reversed the phenomenon in both BGC823 and SGC7901 cells (Fig. 3C and D).

As phosphoraltion of JAK/STAT3 reflects the active form, these data suggested that JAK/STAT3 signaling may be a key player in the IL-11 induced drug resistance. In support of this idea, JAK inhibitor (ruxolitinib) or shSTAT3 was used to detect the acquired drug resistance maintain induced by the IL-11. And we found that the inhibition of JAK or knockdown of STAT3 reversed the drug resistance induced by IL-11 co-culture in both BGC823 and SGC7901 cells when treated with DDP, doxorubicin, and etoposide (Fig. 3E-H). Collectively, these results reminded us that the JAK/STAT3 signaling pathway activation may work as an important player in IL-11/IL-11R axis induced drug resistance in gastric cancer cells.

A diverse range of molecular mechanisms have been implicated in drug resistance, including cancer cells dormancy, genes damage repair, overexpression of multi-drugs resistant proteins and so on. Anti-apoptotic family members have also been identified as a central regulator in cancer drug resistance development [23,24]. Among anti-apoptosis proteins family, Bcl2 plays an important role in cell anti-apoptosis activities and is overexpressed in various malignant tumors. To determine whether Bcl2 participate in the drugs resistance development in gastric cancer cells induced by IL-11, we treated BGC823 and SGC7901 cells with CAFs-CM or rhIL-11 and we found the expression of Bcl2 was significantly higher in the presence of CAF-CM and rhIL-11 in both mRNA and protein levels. And addition of IL-11 neutralizing antibody in CAFs-CM results in the down-regulation of Bcl2 respectively (Fig. 4A and B), suggesting that Bcl2 may be involved in the drug resistance induced by IL-11 in gastric cancer. For this, we used the siRNA to knocked down the expression of Bcl2 and detect the drug resistance in gastric cancer cells treated with CAFs-CM or rhIL-11. CAFs-CM or rhIL-11 pre-treatment significantly enhanced the resistance of BGC823 and SGC7901 cells to DDP, doxorubicin and etoposide, while silencing Bcl2 effectively reversed the drug resistance (Fig. 4C and D). Together, these data suggested that Bcl2 may work as the downstream of IL-11/JAK/STAT3 pathway to participate in the process of acquiring drug resistance in gastric cancer.

Next, we investigated the effect of IL-11 on JAK/STAT3/Bcl2 pathway induced drug resistance in vivo. Here, we generated the mice models bearing SGC7901 cells model and treated with doxorubicin, ruxolitinib (JAK inhibitor) or the combination. We found that doxorubicin or ruxolitinib alone could slightly inhibit the tumor growth, while the combination remarkably slowed the tumor growth (Fig. 5A, left). In consistent, combining doxorubicin with ruxolitinib significantly prolonged survival time compared to control group, while monotherapy had slight effect (Fig. 5A, right). Similarly, we found that DDP/etoposide or ruxolitinib alone could slightly inhibit the tumor growth, while the combination remarkably slowed the tumor growth (Fig. 5B and C, left). In consistent, combining DDP/etoposide with ruxolitinib significantly prolonged survival time compared to control group, while monotherapy had slight effect (Fig. 5B and C, right). These results suggested that blocking IL-11/JAK/STAT3 pathway may be helpful to enhance the chemotherapeutic effect for gastric cancer.

Further, we generated the xenograft mouse bearing SGC7901 cells model. Then IL-11 was injected into the mice by intratumor injection. After four days, mice were treated with doxorubicin, ruxolitinib, or the combination. The IL-11 treated mice showed significantly resistant to the doxorubicin, while combining doxorubicin with ruxolitinib successfully slowed the tumor growth (Fig. 5D, left). In consistent, combining doxorubicin with ruxolitinib significantly prolonged survival time, while monotherapy had limited effect compared to control group (Fig. 5D, right). Similar results were achieved when treated NOD-SCID mice bearing SGC-7901 cells model with DDP and Etoposide combined with or without ruxolitinib (Fig. 5E and F). These data suggested the potential therapeutic value of ruxolitinib combining with chemotherapy drugs in gastric cancer treatment.