Crosstalk between cancer cells and macrophages promotes OSCC cell migration and invasion through a CXCL1/EGF positive feedback loop

The human oral tongue squamous carcinoma cell line Cal27, human epidermal keratinocyte cell line Hacat, and mouse acute monocytic leukemia cell line RAW264.7 were purchased from the China Center for Type Culture Collection. All of the cell lines were cultured in Dulbecco's modified Eagle's medium (VivaCell; Shanghai; China) with 10% fetal bovine serum (VivaCell; Shanghai; China), 100 U/ml penicillin, and 100 μg/ml streptomycin in a humidified 5% CO2 atmosphere at 37 °C.

Recombinant human CXCL1 and recombinant mouse EGF were purchased from R&D Systems Inc. (cat. no. 275-GR) and Abcam (cat. no. ab206643), respectively. AG1478 and SB225002 were purchased from Selleck Biotechnology Co., Ltd. (cat. no. S2728; cat. no. S7651). The antibodies used included rabbit anti-β-actin (cat. no. ab115777; Abcam), rabbit anti-E-cadherin (cat. no. ab40772 Abcam), rabbit anti-N-cadherin (cat. no. ab76011 Abcam), rabbit anti-CXCL1 (cat. no. ab206411 Abcam), rabbit anti-total EGFR (Zen Bio Science Co., Ltd), rabbit anti-phospho-EGFR (cat. no. R24173; Zen Bio Science Co., Ltd), rabbit anti-total P65 (cat. no. R25149; Zen Bio Science Co., Ltd), and rabbit anti-phospho-P65 (cat. no. 310013; Zen Bio Science Co., Ltd). DyLight 488 (A23220) and 800 (A23920) were purchased from Abbkine Scientific Co., Ltd. Anti-mouse PE CD86 (12-0862-81) and anti-mouse APC CD206 (17–2061-80) were purchased from Invitrogen, Thermo Fisher Scientific, Waltham, MA, USA. The anti-CXCL1 (YT2074) antibody for immunohistochemical staining was obtained from Immunoway, USA. The secondary antibody kit (GK600705) for IHC was purchased from Gene Tech, Shanghai, China.

Cancer cell lines were cultured in complete medium, and the medium was replaced by serum-free medium when the cell density reached approximately 90%. After 48 h, the conditioned medium (CM) was harvested, centrifuged at 2000×g for 10 min, filtered through 0.22 mm filters, and stored at 80 °C. Regarding TAM-conditioned medium, Raw264.7 cells were stimulated with tumor-conditioned medium for 48 h, and then serum-free medium replaced the conditioned medium. After 24 h, the conditioned medium (CM) was harvested, centrifuged at 2000×g for 10 min, filtered through 0.22 mm filters, and stored at 80 °C. Tumor- or TAM-conditioned medium was defined as TCM or TAM-CM.

Five healthy tissues and ten OSCC tissues were obtained from the Department of Oral Pathology of China Medical University. Human participants were not included in the study. Both healthy tissues and OSCC tissues were obtained using the paraffin-embedded method. Immunostaining was performed using the avidin–biotin-peroxidase method. Sections were stained using primary anti-CXCL1 (1:100) antibody. Then, a secondary antibody kit was used. Finally, sections were stained with Mayer’s hematoxylin, and sections were observed using a light microscope (Nikon Eclipse Ts2R, Japan).

For assessment of tumor cell migration, tumor cells (1 × 105) in 200 μl of serum-deprived medium were cultured in the upper chamber, and conditioned medium from macrophages stimulated with TCM or TCM and CXCL1 (100 ng/ml); TAM supernatant containing EGF (100 ng/ml) or EGF (100 ng/ml) and AG1478 (5 μM) with 10% FBS was added to the lower chamber. Moreover, cell invasion was assessed using a Transwell chamber (pore size 8 mm, Jet Biofil; Guangzhou; China) with Matrigel coating (BD Biosciences). For chemotactic migration and invasion, the non-migrated cells were removed from the surface of the upper chamber after 24 h, while after 48 h, the noninvasive cells were also scraped. The representative migrated or invaded cells were treated with 4% paraformaldehyde (product no. P0099; Beyotime Institute of Biotechnology) for 20 min and then with 0.5% crystal violet (product no. C0121; Beyotime Institute of Biotechnology) for 15 min. Cells were counted using a light microscope (Nikon Eclipse Ts2R, Japan) and analyzed by ImageJ software (National Institutes of Health, Bethesda, MD, USA).

Cell migration ability was assessed in vitro by wound healing. Cal27 cells (1 × 10^6/well) were seeded in 6-well plates treated with conditioned medium from macrophages stimulated with TCM or TCM and CXCL1 (100 ng/ml) or TAM-CM including EGF (100 ng/ml) or EGF and AG1478 (5 μM). After the tumor cells were grown to confluency, a scratch was created with a 200 µl pipette tip. Then, PBS was used to wash the plates several times. The plates were incubated for 24 h, and the tumor cells that migrated to heal the artificial wound were photographed with a light microscope (Nikon Eclipse Ts2R, Japan) and assessed with ImageJ software. The healing rate was calculated as (%) = (initial average scratch area − average scratch area at 24 h)/initial average scratch area × 100%.

RT‒PCR was used to measure the gene expression levels of CD206, CD86, Arg-1, EGF, CXCL1, cell transfection- and EMT-related markers in Raw264.7 cells and Cal27 cells. Total RNA was isolated by TRIzol Reagent (Takara Bio, Inc., Otsu, Japan), and extracted RNA was reversed transcribed by PrimeScript RT (Takara Bio, Inc., RR036A, Otsu, Japan). RT‒qPCR was performed using the SYBR Premix® Ex TaqTM Kit (Takara Bio, Inc, RR820A, Otsu, Japan) with the 7500 Fast System (Thermo Fisher Scientific). RT‒qPCR thermocycling conditions were as follows: 40 cycles of 95 °C for 30 s, 95 °C for 5 s, 60 °C for 34 s, 95 °C for 15 s, 60 °C for 60 s and 95 °C for 15 s. The internal control of the experiment was β-actin. Primers for β‑actin, CXCL1, EGF, E-cadherin, N-cadherin, Vimentin, CD206, CD86, and Arg-1 were obtained from Shanghai Sheng gong Biological Technology Co., Ltd. All primer sequences are shown in Table 1.

Cal27 cells were washed with PBS and then lysed using RIPA buffer (product no. P0013B; Beyotime Institute of Biotechnology) including 1 mM PMSF (Product no. ST507; Beyotime Institute of Biotechnology) to extract total proteins. Next, the protein concentration was determined by a BCA protein assay kit (product no. P0012s; Beyotime Institute of Biotechnology). Total proteins were separated using 10% or 12% SDS‒PAGE, and electrophoresis was run at 150 V. The separated proteins were transferred to a PVDF membrane at 220 mA. The PVDF membrane was incubated with 5% nonfat dry milk at room temperature for 1 h to block the immunoblots and probed at 4 °C overnight with specific primary antibodies, including rabbit anti-β-actin (1:2000), rabbit anti-CXCL1 (1:1000), rabbit anti-E-cadherin (1:1000), rabbit anti-N-cadherin (1:1000), rabbit anti-total EGFR (1:1000), rabbit anti-phospho-EGFR (1:1000), rabbit anti-total P65 (1:1000), and rabbit anti-phospho-P65 (1:500). Then, the membranes were incubated with the secondary anti-rabbit antibody DyLight 800 goat anti-rabbit IgG (1:2000) at room temperature for 1 h. The gray values of the bands were analyzed by Odyssey CLX (LI-COR, Lincoln, NE, USA) and ImageJ software (National Institutes of Health, Bethesda, MD, USA).

Measurement of EGF and CXCL1 levels in the supernatant of the TAMs and Cal27 cells was conducted using mouse EGF ELISA kits (BOSTER Biological Technology Co. Ltd., Wuhan, China; EK0326) and human CXCL1 ELISA kits (R&D Systems Inc., Minneapolis, MN, USA; cat. no. #DY275), respectively. For the standard solution or sample, 100 μl/well was added to the 96-well containing specific antibody incubated at room temperature for 2 h. After washing three times with 300 μl diluted washing buffer, each well was incubated with the diluted HRP conjugate at room temperature for 1 h. Then, the diluted washing buffer was used to wash each well three times. Each well was incubated with 100 µl chromogenic substrate in the dark for 30 min. Finally, 90 μl stop solution was added to every well, and the absorbance of each well was measured by a microplate reader at 450 nm. By constructing a standard curve, the results of the sample solution were obtained.

The Cal27 cells were treated with 0.5% Triton solution for 10 min. PBS was used to wash the Cal27 cells three times. Then, the Cal27 cells were incubated with primary antibodies against one of the EMT markers (E-cadherin) at 4 °C overnight. The cells were washed three times with PBS again. The cells were incubated with the corresponding secondary anti-rabbit antibodies with DyLight 488 (1:200) at room temperature for 1 h. Finally, the cells were counterstained with DAPI at room temperature for 10 min. Moreover, tissue sections were incubated using primary anti-CXCL1 (1:100) antibody. Then, a secondary antibody kit was used. The staining was observed and photographed with a fluorescence microscope (Nikon Eclipse Ts2R, Japan).

The rate of cell proliferation was determined using a Cell Counting Kit-8 assay (product no. C0038; Beyotime Institute of Biotechnology) according to the manufacturer’s instructions. Cal27 cells were plated in 96-well plates at a ratio of 3 × 10^3/well. Then, the cells were incubated in TAM-CM with EGF (100 ng/ml) or EGF (100 ng/ml) and AG1478 (5 μM) for 24 and 48 h. Ten microliters of CCK-8 solution was added to every well, and the cells were cultured at 37 °C for 1 h. The absorbance of each well was measured at 450 nm (Tecan Mechelen, Belgium).

Cal27 cells were cultured in 6-well plates at a ratio of 1 × 10^3/well and stimulated with TAM-CM containing EGF (100 ng/ml) or EGF (100 ng/ml) and AG1478 (5 μM) for 2 weeks. The incubation medium was replaced every two days. Subsequently, the cells were treated with 4% paraformaldehyde and then with 0.5% crystal violet for 30 min, and colonies were counted with a light microscope.

Cal27 cells (1 × 10^5/well) were seeded in 6-well plates and transfected with double-stranded small interfering RNA (siRNA) (Shanghai Gene Pharma Co., Ltd., Shanghai, China). Sense and antisense strands for siRNAs were as follows: si-CXCL1-1 (5′-ACUCAAGAAUGGGCGGAAATT-3′) (5′-UUUCCGCCCAUUCUUGAGUTT-3′); si-CXCL1-2 (5′-CCAAGAACAUCAAAGUGUTT-3′) (5′-ACACUUUGGAUGUUCUUGGTT-3′); and si-CXCL1-3 (5′-GAUGCUGAACAGUGACAAATT-3′) (5′-UUUGUCACUGUUCAGCAUCTT-3′). Cal27 cells transfected for 24 h were used for subsequent experiments.

After RAW264.7 cells were cultured with TCM or TCM and CXCL1 for 48 h, they were stained with anti-mouse PE CD86 (Thermo Fisher Scientific, Waltham, MA, USA) and anti-mouse APC CD206 (Thermo Fisher Scientific, Waltham, MA, USA) on ice for 20 min according to the dilution ratio recommended by the manufacturers. Data analysis was performed using FlowJo software.

Prism 8.0 for Windows (GraphPad Software, Inc., La Jolla, CA, USA) was used for data analysis. Measurement data are presented as the mean ± standard deviation (SD), and statistical significance between groups was determined by Student’s t test or analysis of variance (ANOVA). Western blotting, wound healing assay and Transwell assay results were assessed by ImageJ software. Differences were considered statistically significant at *P < 0.05, **P < 0.01, and ***P < 0.001.

A previous study demonstrated that CXCL1 was highly expressed in OSCC tissues [20]. In our study, CXCL1 expression was determined in healthy tissues and OSCC tissues by immunohistochemical staining and immunofluorescence, as shown in Fig. 1A–D. The results demonstrated that CXCL1 was overexpressed in OSCC tissues. High levels of CXCL1 in OSCC have been related to the malignant progression of OSCC. The role of CXCL1 in OSCC and the underlying mechanism need to be studied.

In addition to OSCC tissues, CXCL1 is also highly expressed in many OSCC cell lines. Our PCR and ELISA results consistently indicated that the mRNA level and secretion level of CXCL1 were higher in Cal27 cells than in Hacat cells (Fig. 2A, B). However, the roles of tumor cell-secreted CXCL1 in the progression of OSCC are not fully clear. It has been reported that the CXCL1-mediated interaction of cancer cells with tumor-associated macrophages promotes tumor progression in bladder cancer. To verify that CXCL1 promotes OSCC cell migration and invasion through crosstalk with macrophages, Raw264.7 cells were cultured with tumor conditional medium (TCM) or TCM and CXCL1. Then, the conditioned medium of macrophages stimulated with TCM or TCM and CXCL1 was used to stimulate Cal27 cells. The wound-healing and Transwell assays demonstrated that the supernatant of TCM-induced macrophages promoted Cal27 cell migration and invasion. Moreover, the supernatant of macrophages treated with TCM and CXCL1 further induced migration and invasion (Fig. 2C, D). These results illustrated that CXCL1 may induce the migration and invasion of OSCC cells through crosstalk with macrophages.

According to the results above, CXCL1 could induce OSCC cell migration and invasion via macrophages. Thus, we further determined whether CXCL1 affects the polarization of macrophages to promote the aggressive behavior of OSCC. The results demonstrated that recombinant CXCL1 increased the mRNA levels of Arg-1 and CD206 while decreasing the level of CD86 in macrophages (Fig. 3A–C). We also found that compared to the control group, the group treated with recombinant CXCL1 exhibited upregulated levels of CD206 (M2 surface marker) and reduced levels of CD86 (M1 surface marker) (Fig. 3D–G). It has been reported that colorectal cancer cell-derived CXCL1 promotes the secretion of epithelial growth factor (EGF) in an autocrine manner by binding to C-X-C motif chemokine receptor 2 (CXCR2) in colorectal cancer cells [28]. Moreover, CXCR2 was also expressed on the macrophage surface and found to mediate tumor invasion and metastasis [25], so we hypothesized that OSCC cell-derived CXCL1 may promote macrophages to secrete EGF via CXCR2. Compared with TCM, TCM combined with recombinant CXCL1 further induced the secretion of EGF in macrophages (Fig. 3H, I). However, knockdown of CXCL1 in Cal27 cells led to the opposite effect (Fig. 3J, K). We used an inhibitor of CXCR2 (SB225002) to block the CXCL1/CXCR2 pathway [29]. The gene and protein levels of EGF were reversed (Fig. 3L, M). These results demonstrated that CXCL1 promoted M2 macrophage polarization and induced the secretion of EGF in macrophages.

In the tumor environment, autocrine EGF has multiple functions, such as promoting the invasion, metastasis and EMT of cancer cells [6, 30]. In previous research, the functions of EGF mainly depended on its receptor EGFR [31]. To further explore whether macrophage-derived EGF exerts a cancer-promoting effect via EGFR, recombinant-incubated Cal27 cells were treated with or without an inhibitor of EGFR (AG1478). Then, colony formation and CCK-8 assays were used to evaluate cell proliferation and viability. The colony formation results revealed that recombinant EGF significantly increased the proliferation ability of Cal27 cells, while AG1478 treatment led to a decrease in cell viability (Fig. 4A). There was no difference in the proliferation ability of Cal27 cells at 24 h. However, after stimulation with recombinant EGF for 48 h, the proliferation ability of Cal27 cells was increased, and these effects could be abrogated by AG1478 treatment (Fig. 4B). The 24 h wound-healing assay was employed to determine the migration ability of the Cal27 cells. The increased healing rate of Cal27 cells induced by recombinant EGF was decreased by AG1478 (Fig. 4C, D). Moreover, the results of the Transwell assay showed that the migration and invasion abilities of Cal27 cells were prominently increased following 24 h of EGF treatment. We also performed knockdown experiments by using an inhibitor of EGFR, AG1478. The results showed that AG1478 could abrogate the increased cell migration and invasion (Fig. 4E, F). In summary, these data demonstrated that EGF promoted the proliferation, migration and invasion of OSCC cells.

Epithelial-mesenchymal transition (EMT) is an important underlying mechanism of primary tumorigenesis and metastasis [31]. During EMT, epithelial cells transform from epithelial cells highly expressing epithelial markers (E-cadherin) to mesenchymal cells overexpressing mesenchymal markers (N-cadherin and Vimentin), which then promote the migration and invasion of cancer cells [32]. Subsequently, we further investigated whether EGF enhanced the migration and invasion of Cal27 cells by inducing EMT. TAM-CM supplemented with EGF or EGF and AG1478 was used to incubate Cal27 cells for 48 h. The RT‒PCR results showed that EGF increased the mRNA levels of Vimentin and N-cadherin and decreased the mRNA level of E-cadherin in Cal27 cells. However, AG1478 had exactly the opposite effect (Fig. 5A). Western blot analysis and immunofluorescence staining showed that the EGF-induced suppression of E-cadherin was abrogated by AG1478. Simultaneously, the EGF-induced high protein level of N-cadherin was decreased by AG1478 treatment (Fig. 5B–D). As shown in Fig. 5E, the morphology of Cal27 cells became elongated and spindle-shaped when cells were stimulated by EGF, while the Cal27 cells became cobblestone-shaped when they were treated with AG1478. The above results revealed that EGF could greatly promote the EMT of OSCC cells.

Former studies have revealed that a number of signaling pathways activated by EGF could result in increased cell migration and invasion [5, 27]. EGF mainly functions by binding to EGFR. Therefore, an inhibitor of EGFR (AG1478) was used to suppress EGF/EGFR signaling [33]. Western blot analysis demonstrated that EGFR expression was slightly upregulated and that the p-EGFR signaling pathway was significantly activated in Cal27 cells stimulated with recombinant EGF (100 ng/ml) for 30, 60 and 120 min (Fig. 6A–C). Furthermore, Western blot analysis was used to detect the protein expression of NF-κB pathway members in Cal27 cells treated with EGF (100 ng/ml) or EGF (100 ng/ml) combined with AG1478 (5 μM). The analysis showed that after treatment with EGF (100 ng/ml), phosphorylation of EGFR and p65 in Cal27 cells was increased. However, AG1478 inhibited the activation of the NF-κB/p-65 pathway (Fig. 6D–F). A recent study identified the NF-κB/p-65 pathway as the mediator of OSCC cell migration and invasion [34‐36]. Therefore, these results demonstrated that EGF may promote the invasion and migration of OSCC cells through the EGFR/NF-κB signaling.

Previous studies have revealed that EGF increases the production of CXCL1 in hepatocellular carcinoma cells [37]. Therefore, it was extremely necessary to determine whether CXCL1 and EGF could mediate crosstalk between OSCC cells and macrophages and further induce the aggressiveness of OSCC cells in the tumor microenvironment. As shown in Fig. 6G–I, the mRNA expression of CXCL1 was upregulated in Cal27 cells after stimulation with recombinant EGF. Moreover, the protein expression of EGF in Cal27 cells treated with EGF was also upregulated. However, when we used an inhibitor of EGFR (AG1478) to inhibit the EGF/EGFR signaling pathway, the level of CXCL1 was also suppressed [30]. These results indicated that CXCL1 could promote OSCC cell invasion and metastasis through a CXCL1/EGF-mediated positive feedback loop in macrophages and OSCC cells.