Crosstalk between tumor-associated macrophages and tumor cells promotes chemoresistance via CXCL5/PI3K/AKT/mTOR pathway in gastric cancer

Paraffin-embedded samples of primary lesions from 103 patients with gastric cancer who had undergone 5-FU based neoadjuvant chemotherapy prior to radical resection at Peking Union Medical College Hospital between 2015 and 2017 were used. Patients were divided into two groups according to the evaluation of pathological response based on the guidelines of College of American Pathologists (CAP) [20]. CAP 0, CAP 1 and CAP 2 were defined as pathological response whereas CAP 3 was defined as no pathological response. 67 patients were elected to pathological response group and 36 patients were elected to no pathological response group for further research. Clinical samples were gathered with written informed consent of patients according to a protocol reviewed and approved by the Institutional Review Board of Peking Union Medical College Hospital.

A total of 103 archived clinical samples were fixed in 10% formaldehyde solution, embedded in paraffin and serially severed into 4 μm sections. After deparaffinized in xylene and rehydrated in graded ethanol, microwave heating with sodium citrate retrieval buffer (pH 6.0) was performed for antigen retrieval. The endogenous peroxidase was inactivated by treatment with 3% H₂O₂ for 10 min. Tissue sections were incubated with blocking buffer followed by incubation with primary antibodies Anti-CD68 (1:100, Cell Signaling Technology, MA, USA), Anti-CD163 (1:100, Cell Signaling Technology, MA, USA), Anti-CD206 (1:100, Cell Signaling Technology, MA, USA) and Anti-CXCL5 (1:200, Abcam, Cambridge, UK) at 4 °C overnight. After washing with PBS, the secondary antibody horseradish peroxidase (HRP)-conjugated Anti-Rabbit IgG (1:100, Cell Signaling Technology, MA, USA) was added for 30 min’ incubation at room temperature. 3, 3ʹ-diaminobenzidine (DAB) regent was applied for visualizing staining subsequent to PBS washing, then all slices were re-dyed with hematoxylin, dehydrated and sealed for microscopic examination. At least three slices were taken from each tumor tissue and five independent fields were randomly selected from each slice for detection. CXCL5 immunoreactivity was scored by multiplying the staining percentage scores (~ 5% scores 0; 5% ~ 25% scores 1; 25% ~ 50% scores 2; 50% ~ 75% scores 3; 75% ~ 100% scores 4) and staining intensity scores (0, no staining; 1, weak; 2, moderate; 3, strong). A final score of 0–3 was defined as low expression, while others were defined as high expression. CD68, CD163 and CD206 immunoreactivity were analyzed by calculating the mean number of positive cells in 5 random 400-fold fields. Two independent pathologists observed and scored the slices without knowledge of the patients’ clinical information.

Human gastric cancer cell lines MKN45 and HGC27, and human mononuclear cells (THP-1) were acquired from the Cell Resource Center of Peking Union Medical College (Beijing, China). The cells were maintained in RPMI-1640 medium (Gibco, Carlsbad, CA, USA) incorporating 10% fetal bovine serum (FBS) (Gibco, Carlsbad, CA, USA) and 1% penicillin/streptomycin (Gibco, Carlsbad, CA, USA) in a humidified 37 °C incubator with 5% CO₂. 0.25% trypsin (NCM Biotech, Suzhou, Jiangsu, China) was administered in the logarithmic growth phase for cell digestion and passage.

Gastric cancer cell lines were regarded as 5-FU-sensitive (MKN45-S and HGC27-S) and the IC₅₀ of 5-FU was detected. 5-FU-resistent cell lines (MKN45-R and HGC27-R) were generated by repetitively exposing gastric cancer cells to increasing concentrations of 5-FU over a 10 month period and the acquired 5-FU resistance was confirmed by detecting the IC₅₀ of 5-FU and resistance index. THP-1 monocytes were differentiated into macrophages by 24 h incubation with 100 ng/ml phorbol 12-myristate 13-acetate (PMA) (Sigma-Aldrich, St.louis, MO, USA) for 24 h. Adherent cells were washed twice with culture medium followed by 24 h incubation to obtain the resting state of macrophages (M0). Method of detaching PMA-treated THP-1 cells from the culture dish was demonstrated in Additional file 1: Text S1.

Transwell chambers (6-well plates, 0.4-μm pore size; Corning, NY, USA) were used for co-culture. MKN45-S, HGC27-S, MKN45-R and HGC27-R cells were seeded onto the upper chambers, and M0 macrophages were placed in the lower chambers. After 48 h of co-culture, TAMs from 5-FU-sensitive TME (MS) and 5-FU-resistant TME (MR) were obtained and harvested for experimental analysis. To investigate the effect of TAMs with different phenotypes on gastric cancer cells, MS and MR were transferred to the upper chambers and gastric cancer cells were placed in the lower chambers for 48 h of co-culture.

A cell counting kit-8 (CCK-8; Dojindo, Kumamoto, Japan) assay was used to evaluate the inhibition of cell growth in response to varying concentrations of 5-FU (0, 5, 10, 20, 40, 80, 120, 160, 200 μg/ml). Briefly, cells were seeded onto 96-well plates at a density of 5 × 10³ cells per well in 100 μl of culture medium and incubated at 37 °C with 5% CO₂. After incubation for 24 h, varying concentrations of 5-FU diluted with the culture medium were added to each well and co-incubated for another 24 h. Then 10 μl of CCK-8 reagent was administered to each well for 2 h incubation at 37 °C. The optical density (OD) was detected by a microplate reader at 450 nm. Each experiment was repeated three times and each measurement was conducted three times.

Cells were seeded onto 6-well plates at a density of 500 cells per well for adhesion-dependent colony formation. 5-FU was added to the culture medium at a final concentration of 15 μg/ml and the culture medium that contained 5-FU was changed every 3–4 days. After 2 weeks, visible colonies were fixed with 4% paraformaldehyde for 15 min and stained with 0.1% crystal violet staining solution for 10 min. Then, the formed colony units were photographed and counted for analysis.

The concentrations of CXCL5 and CCL18 in culture supernatants from M0, MS, MR, MKN45-S, HGC27-S, MKN45-R and HGC27-R cells (1 × 10⁶ cells) were quantified by ELISA kits (Cell Signaling Technology, MA, USA) according to the manufacturer’s instructions. Absorbance was measured using a microplate reader. The concentration of the sample were estimated from the standard curve and the levels below the detection limit of the assay were perceived as zero.

Recombinant human CXCL5 (rhCXCL5) were purchased from R&D Systems (Minneapolis, MN, USA). 100 μg/ml stock solution of rhCXCL5 was achieved by dissolving 25 μg powder in 250 μl PBS, followed by adding 0.1% BSA in the final solution, and cells were treated with 10 ng/ml rhCXCL5 for 48 h.

THP-1 cells were seeded onto the upper chamber (6-well plates, 0.8 μm pore size; Corning, NY, USA) at a density of 2 × 10⁵ in 200 μl serum-free medium. M0, MS and MR cells were cultured with serum-free medium for 24 h, then the supernatants from above cells were collected and added to the corresponding lower chamber with or without CXCL5 neutralizing antibody (0.5 μg/ml; Abcam, Cambridge, UK). After incubation for 24 h at 37 °C with 5% CO₂, THP-1 cells that migrated to the lower chamber were measured by fixing and staining the inserts with 0.1% crystal violet staining solution and counting under a microscope (100-fold fields). Non-migratory cells were removed before the membrane was observed.

Total RNA was extracted from cells using TRIzol reagent (Invitrogen, Carlsbad, CA, USA) according to the manufacturer’s protocol. cDNA was synthesized from 1 μg of total RNA using 5 × PrimeScript RT reagent Kit (Takara, Dalian, China), and RT-qPCR was performed using TB Green Premix Ex Taq II (Takara, Dalian, China). Reactions were performed in triplicate and the relative mRNA expression was analyzed by the 2^−ΔΔCt method using GAPDH as an internal control. The forward and reverse primer sequences for the targeted genes are listed in Additional file 2: Table S1.

Total protein was extracted out of cells using ice-cold RIPA buffer (Thermo Scientific, Rockford, IL, USA) with Halt Protease and Phosphatase inhibitor Cocktail (Thermo Scientific, Rockford, IL, USA) for 15 min. Protein samples were sonicated followed by centrifugation at 12000 g for 15 min at 4 °C and the concentrations were detected by BCA Protein Assay Kit (Beyotime, Shanghai, China). Approximately 30 μg of denatured protein was fractionated by 10% sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) and transferred onto 0.45 μm polyvinylidene fluoride (PVDF) membranes (Millipore, Billerica, MA, USA). The PVDF membranes were blocked by TBST solution containing 5% skimmed milk for 1 h at room temperature and then incubated overnight at 4˚C with the primary antibodies against P-gp, Bcl-2, Bax, PTEN, PI3K, p-PI3K, AKT, p-AKT, mTOR, p-mTOR (1:1000, Cell Signaling Technology, MA, USA) and GAPDH (1:500, Cell Signaling Technology, MA, USA). Next, the membranes were incubated with corresponding secondary antibody HRP-conjugated Anti-Rabbit IgG (1:10000, Cell Signaling Technology, MA, USA) at room temperature for 1 h. SuperSignal West Pico PLUS Chemiluminescent Substrate (Thermo Scientific, Rockford, IL, USA) was used for visualizing the blots in a Kodak Image station (Tanon, China) (Additional file 3: Table S2).

M0, MS and MR cells were washed twice by PBS and filtered through a 100 μm mesh for flow cytometry. Then cells were counted, diluted to 1 × 10⁶ cells per 100 μl and subsequently stained with FITC-CD11b, PE-CD86, APC-CD163 and APC-CD206 antibodies (BioLegend, San Diego, CA, USA) followed by incubating in darkness at 4 °C for 15 min. Finally, the labeled cells were analyzed by BD Accuri C6 Plus flow cytometer (BD Biosciences, San Jose, CA, USA). The process was conducted in triplicate and data were analyzed by FlowJo software (Tree Star, Oregon, OR, USA) (Additional file 4: Figure S1).

Cell apoptosis was measured using Annexin-V-FITC Apoptosis Detection Kit (Dojindo, Kumamoto, Japan) according to the manufacturer’s protocol. In brief, cells were washed twice with PBS, after centrifugation, cells were suspended in 100 μl of 1 × binding buffer. Then 5 μl Annexin V-FITC and 5 μl propidium iodide (PI) were added to stain cells for 15 min in the dark. The stained cells were maintained on ice until apoptosis was measured using BD Accuri C6 Plus flow cytometer (BD Biosciences, San Jose, CA, USA). The process was conducted in triplicate and data were analyzed by FlowJo software (Tree Star, Oregon, OR, USA).

Data were expressed as mean ± standard deviation (SD) of at least three separate experiments. Student’s t-test or one-way analysis of variance (ANOVA) was used for difference analysis. Correlation of the expression level of CD163 or CD206 with CXCL5 was determined using Spearman rank-order correlation. Survival curves were analyzed using Kaplan–Meier and log-rank methods. All statistical analyses were calculated by SPSS 22.0 (SPSS Inc., Chicago, IL, USA) in conjunction with GraphPad Prism 8 (GraphPad Prism Software, Inc., San Diego, CA, USA), and P < 0.05 was considered statistically significant.

To investigate the biodistribution of macrophages in gastric cancer, we detected the expression of CD68, one of the markers of macrophages, in 103 clinical samples from patients receiving 5-FU based neoadjuvant chemotherapy by immunohistochemistry. CD68 positive macrophages were counted in 5 random 400-fold fields of each section. Results showed that the number of macrophages infiltrated in tissue from no pathological response group was significantly increased compared with that of pathological response group (Fig. 1, p < 0.001). The above finding indicated that infiltration of macrophages may play a critical role in the development and progression of chemoresistance in gastric cancer.

Two 5-FU-resistant gastric cell lines, MKN45-R and HGC27-R, were established by continuous exposure of the parental cells to increasing concentrations of 5-FU. To confirm that they were resistant to 5-FU, increasing concentrations of 5-FU (0, 5, 10, 20, 40, 80, 120, 160, 200 μg/ml) were added and cell viability was then detected by CCK-8 assay. The results demonstrated that 5-FU decreased cell viability in a dose-dependent manner (Fig. 2A, B). The IC₅₀ values of MKN45-R and HGC27-R cells increased to 176.31 ± 8.43 μg/ml and 219.15 ± 10.25 μg/ml respectively, remarkably higher compared with their parental cells MKN45-S (34.97 ± 2.92 μg/ml) (p < 0.001) and HGC27-S (41.52 ± 3.75 μg/ml) (p < 0.001). In addition, western blot analysis indicated that the expression levels of chemoresistance-associated protein P-gp (HGC27: p < 0.001; MKN45: p < 0.001) and anti-apoptotic protein Bcl-2 (HGC27: p < 0.001; MKN45: p < 0.001) were significantly increased, and pro-apoptotic protein Bax (HGC27: p = 0.004; MKN45: p = 0.006) was significantly decreased in MKN45-R and HGC27-R cells (Fig. 2C, D). Uncropped western blot images were demonstrated in Additional file 5: Figure S2. As a result, the above data denoted that 5-FU-resistant gastric cell lines were effectively established.

Human monocyte cell line THP-1 was differentiated into macrophage (M0) via treatment with 100 ng/ml PMA for 24 h. M0 macrophages were subsequently co-cultured with MKN45-R, HGC27-R cells and their parental cells respectively in a non-contact Transwell system for 48 h, and then TAMs from 5-FU-sensitive TME (MS) and 5-FU-resistant TME (MR) were obtained (Fig. 3A). RT-qPCR was performed to explore the effect of co-culturing gastric cells on M0 cells. As a result, the levels of M1-related gene expression of CD86, TNF-α and IL-12 were down-regulated in MS and MR cells relative to M0 cells. Furthermore, the expression levels were significantly lower in MR cells than in MS cells (Fig. 3B, C). In contrast, the levels of M2-related genes expression of CD163, CD206, IL-10, Arg-1 and VEGF-A were up-regulated after co-culturing with gastric cells, and the expression levels were remarkably higher in MR cells than in MS cells (Fig. 3B, C). Flow cytometry analysis was used to characterize the surface makers of M0, MS and MR cells. It showed that the percentage of CD86⁺CD11b⁺ cells (M1 macrophages) was decreased in MS and MR cells, and the percentage was significantly lower in MR cells than in MS cells (HGC27: p = 0.032; MKN45: p = 0.028). Besides, the percentages of CD163⁺CD11b⁺ (HGC27: p = 0.040; MKN45: p = 0.029) and CD206⁺CD11b⁺ cells (HGC27: p = 0.016; MKN45: p = 0.039) (M2 macrophages) were increased in MS and MR cells, and the percentage was higher in MR cells than in MS cells (Fig. 3D). Taken together, the above results indicated that gastric cancer cells can modulate macrophages to M2-like polarization, and chemoresistant cells seem to be more effective in inducing macrophages polarization to M2 phenotype.

Gastric cancer cells were co-cultured with macrophages of different phenotypes in a non-contact Transwell system for 48 h, then macrophages were discarded and gastric cancer cells were collected for further analysis (Fig. 4A). CCK-8 assay was adopted to investigate whether the resistance to 5-FU was affected by macrophages. As indicated in Fig. 4B and C, co-culturing with MS and MR macrophages enhanced the IC₅₀ values of HGC27-S cells from 42.58 ± 6.93 μg/ml to 66.74 ± 8.53 and 126.65 ± 10.23 μg/ml respectively (p = 0.003; p = 0.006). Similarly, co-culturing with MS and MR macrophages enhanced the IC₅₀ values of MKN45-S cells from 36.37 ± 6.35 μg/ml to 67.42 ± 7.82 and 103.54 ± 13.29 μg/ml respectively (p = 0.006; p = 0.002). In addition, cell colony formation assay also demonstrated that co-culturing MR macrophages significantly promoted 5-FU-resistance of gastric cancer cells, whereas the effect of MS macrophages was remarkably smaller than that of MR macrophages (Fig. 4D, E; HGC27: p = 0.003; MKN45: p = 0.005). Chemotherapy-induced apoptosis was also determined through apoptotic flow cytometry assay. In brief, after co-culturing with MS and MR macrophages, gastric cancer cells were then treated with 5-FU at the concentration of 15 μg/ml for 24 h. The result indicated that the rate of apoptosis was significantly lower in gastric cancer cells co-cultured with MR macrophages as compared to that cultured alone or co-cultured with MS macrophages (Fig. 4F, G; HGC27: p = 0.002, p = 0.004; MKN45: p = 0.002, p = 0.003). Western blot analysis further indicated that the expression levels of chemoresistance-associated protein P-gp and anti-apoptotic protein Bcl-2 were significantly increased, and pro-apoptotic protein Bax was significantly decreased in gastric cancer cells co-cultured with MR macrophages (Fig. 4H, I; HGC27: p = 0.007, p = 0.005, p < 0.001; MKN45: p < 0.001, p < 0.001, p < 0.001). Uncropped western blot images were demonstrated in Additional file 6: Figure S3. All the findings above indicated that 5-FU-resistant gastric cancer cells can effectively induce macrophages polarization to M2 phenotype, which in turn promotes 5-FU-resistance in gastric cancer cells.

It was speculated that the different ability of MS and MR macrophages to enhanced 5-FU-resistance of gastric cancer cells was due to the different level of secreted cytokines. After reviewing the relevant literature, eleven cytokines (CCL1, CCL2, CCL3, CCL4, CCL5, CCL17, CCL18, CCL22, CXCL1, CXCL2 and CXCL5) were selected for analyzing through RT-qPCR. As indicated in Fig. 5A and B, of the eleven cytokines, only CCL18 and CXCL5 whose transcription levels were consistent with the trend of previous results (Fig. 4B–I) and the transcription level in MR macrophages was significantly higher than that in MS macrophages. The levels of CCL18 and CXCL5 protein in culture supernatants were then further detected by ELISA. The results denoted that the levels of CCL18 and CXCL5 in 5-FU-sensitive and 5-FU-resistant gastric cancer cells were below the detection limit of the assay and recorded as zero (Fig. 5C-F). In addition, the concentrations of CCL18 in culture supernatants from MS and MR macrophages were very low and the difference was not statistically significant (Fig. 5C, D; HGC27: p = 0.149; MKN45: p = 0.372), which could not explain the difference between MS and MR macrophages in enhancing 5-FU-resistnace of gastric cancer cells. By contrast to CCL18, CXCL5 levels in MS and MR macrophages were remarkably high and the difference was statistically significant (Fig. 5E, F; HGC27: p < 0.001; MKN45: p < 0.001). Based on the results, we focused on CXCL5 in the further studies to verify its role in 5-FU-resistance of gastric cancer cells. 5-FU-sensitive gastric cancer cells were cultured alone or co-cultured with MR macrophages with the presence of rhCXCL5 (10 ng/ml) or CXCL5 neutralizing antibody (0.5 μg/ml) for 48 h followed by being treated with different concentrations of 5-FU (0, 5, 10, 20, 40, 80, 120, 160, 200 μg/ml) for 24 h, then the cell survival rates were assessed. As indicated by CCK-8 assay, co-culturing MR macrophages or addition of rhCXCL5 could induce 5-FU-resistance, whereas CXCL5 neutralizing antibody decreased MR macrophages-mediated resistance to 5-FU in gastric cancer cells (Fig. 5G, H). Taken together, the above results denoted that CXCL5 derived from M2-polarized macrophages was one of the major chemokines that associated with 5-FU-resistance in gastric cancer cells.

There is accumulating evidence indicating that resistance to apoptosis is responsible for chemoresistance. 5-FU exerts the anti-tumor efficiency mainly though mediating apoptosis of tumor cells. Therefore, we determined whether M2-polarized macrophages mediated 5-FU-resistance through regulation of apoptosis in gastric cancer cells. 5-FU-sensitive gastric cancer cells were cultured alone or co-cultured with MR macrophages with the presence of rhCXCL5 (10 ng/ml) or CXCL5 neutralizing antibody (0.5 μg/ml) for 48 h followed by being treated with 15 μg/ml 5-FU for 24 h, then the rates of apoptosis were analyzed. It was revealed that co-culturing MR macrophages or rhCXCL5 could both reduce apoptotic proportion, however, applying CXCL5 neutralizing antibody decreased MR macrophages-mediated resistance to apoptosis in gastric cancer cells (Fig. 6A, B). Western blot analysis further denoted that co-culturing MR macrophages or addition of rhCXCL5 increased the expression levels of anti-apoptotic protein Bcl-2, whereas pro-apoptotic protein Bax was significantly decreased in gastric cancer cells. In addition, the ability of MR macrophages to affect the expression of Bcl-2 and Bax was remarkably weakened by CXCL5 neutralizing antibody (Fig. 6C, D). Previous studies have revealed that the PI3K/AKT/mTOR signaling pathway plays a critical role in tumor progression through regulating cell proliferation, apoptosis and chemoresistance. Based on this, the potential effect of M2-polarized macrophages on the activation of the PI3K/AKT/mTOR pathway was further investigated. As indicated in Fig. 6E and F, co-culturing MR macrophages and rhCXCL5 could both significantly increased the expression of phosphorylated PI3K, AKT and mTOR, whereas treatment with CXCL5 neutralizing antibody remarkably weakened the ability of co-culturing MR macrophages to activate this signaling pathway. Uncropped western blot images were demonstrated in Additional file 7: Figure S4. Taken together, the above results indicated that CXCL5 derived from M2-polarized macrophages inhibited apoptosis and activated the PI3K/AKT/mTOR pathway to increase 5-FU-resistance in gastric cancer cells.

To identify how macrophages affected the recruitment of monocytes in gastric cancer, chemotaxis assay was performed. It has already been determined that the levels of CXCL5 were different in M0, MS and MR macrophages, we speculated that the effect of macrophages on recruitment of monocytes coincided with the secretion level of CXCL5. Data from chemotaxis assay indicated that M0, MS and MR macrophages could attract THP-1 cells, the chemotaxis-inducing effect of MR macrophages was significantly stronger than that of M0 and MS macrophages (p < 0.001; p < 0.001), and the CXCL5 neutralizing antibody could weakened the MR macrophages-induced chemotaxis in gastric cancer (Fig. 7A; p < 0.001). To further verified the role of M2-polarized macrophages-derived CXCL5 in gastric cancer, the expressions of CXCL5, CD163 and CD206 in clinical samples from 103 patients were determined by immunohistochemistry. The result demonstrated that the high expression of CXCL5 correlated with the high infiltration of CD163 positive macrophages (r = 0.595, p < 0.001) and CD206 positive macrophages (r = 0.603, p < 0.001) (Fig. 7B-D). Based on the data above, we speculated that gastric cancer cells, especially chemoresistant cells could induce macrophages polarization to M2 phenotype, which lead to the increased secretion of CXCL5. High level of CXCL5 could in turn recruit monocytes to form more M2-polarized macrophages and further promote the development of chemoresistant microenvironment in gastric cancer.

To investigate the clinical relevance of CXCL5 in prognosis of patients with gastric cancer, we detected the expression level of CXCL5 in clinical samples from 103 patients by immunohistochemistry. The representative images of low (n = 46) and high (n = 57) expression of CXCL5 were showed in Fig. 8A. Furthermore, the survival analysis indicated that the overall survival (OS) rates of patients with higher expression of CXCL5 were significantly lower than that of patients with lower expression of CXCL5 (Fig. 8B; p < 0.001). Taken together, the above results demonstrated that CXCL5 can be used as an effective biomarker for predicting chemoresistant and prognosis in gastric cancer.