Bone mesenchymal stem cells are recruited via CXCL8‐CXCR2 and promote EMT through TGF‐β signal pathways in oral squamous carcinoma

Abstract Objectives Bone mesenchymal stem cells (BMSCs) play critical roles in tumour microenvironment. However, molecular mechanisms of how BMSCs to be recruited and effect subsequent tumour progression are poorly understood in oral squamous cell carcinoma (OSCC). Materials and Methods The distribution of CXCL8 was detected by immunohistochemical staining in OSCC tissues. The chemotaxis of conditioned media from different epithelial cells to BMSCs was examined by trans‐well assay. Real‐time quantitative PCR (qPCR) and ELISA were used to detect the expression of related cytokines and chemokine receptors. The migration of BMSCs was observed in BALB/c nude mice. The roles of BMSCs in proliferation, migration and invasion of OSCC were detected by CCK‐8, flow cytometry and trans‐well assay. Epithelial‐mesenchymal transition (EMT)–related markers were analysed by qPCR and Western blot in vitro, and growth was evaluated in BALB/c nude mice using subcutaneously implanted OSCC in nude mouse model in vivo. Results Using OSCC, we show CXCL8, secreted by OSCC, binds to exclusively CXCR2 in BMSCs to facilitate migration of BMSCs to OSCC. TGF‐β secreted by BMSCs subsequently induces EMT of OSCC to promote their proliferation, migration and infiltration. We also showed that the Ras/Raf/Erk axis plays a critical role in tumour progression. Conclusions Our results provide the molecular basis for BMSC recruitment into tumours, and how this process leads to tumour progression and leads us to develop a novel OSCC treatment target.

BMSCs become an especially important tool for regenerative medicine and a major cellular source for the biological therapies because of their self-renewal capacity, multipotency and immunomodulatory properties. It has been reported that BMSCs can be recruited into tumour microenvironment (TME) in response to a variety of biologically active molecules to repair tissues, balance cell homeostasis and participate in immune regulation. 3 BMSCs can be recruited by all kinds of chemokines in various cancers, for example CXCL16 in prostate cancer, 4 CXCL12 in breast cancer, 5 CXCL8/IL-8, CCL2/MCP-1 and CXCL1-2-3/GRO that play a role in the homing of BMSCs to liver tumour sites, 6 CCL5, 7 TGF-β and interleukin-17b, 8 CXCL10, placental growth factor 9 and platelet-derived growth factor receptor beta. 10 However, there is no report to uncover the chemotaxis mechanism of BMSCs in OSCC.
In cancers, however, BMSCs can secrete multiple bioactive factors to alter the key cellular functions of cells in TME, such as survival, apoptosis, maturation and differentiation. [11][12][13] On the one hand, BMSCs play a vital role in breast cancer development. 14 BMSCs promote the spread of breast cancer cells. 15 In osteosarcoma, exosomes secreted by BMSCs promote osteosarcoma cell proliferation. 16 On the other hand, in malignant gliomas, BMSCs can interact with TME, which leads to tumour shrinkage, and impair cell proliferation, vascularization and significantly prolonged animal survival. 17 BMSCs prevent the development of oncogenic lung cancer in the rat model. 18 In addition, it was also found that BMSCs have different effects on different biological behaviours of the same type of tumour. In thyroid cancer, BMSCs promote cancer cell proliferation and block cancer cell migration. 19 All these suggest that changes in TME affect the differentiation and proliferation of cancer cells. In this study, we tried to explore the chemotaxis of BMSCs and the role of BMSCs in OSCC migration, invasion and EMT by using in vitro and in vivo experiments.  (Table 1). Tumour tissues and relatively normal mucosal tissues were fixed by 10% formalin and embedded in paraffin.

| Patient samples and clinicopathological data
The study was approved by the ethics committee of Jilin University Stomatology Hospital.

| Primary culture of BMSCs
Human BMSCs were isolated from the bone marrow of patients who undergone mandibular reconstruction with free fibular flap. Briefly, bone marrows were centrifuged by density gradient centrifugation with 1.073 g/mL Percoll (Sigma-Aldrich) at 1810 g for 30 minutes, and BMSC layer was collected and cultured in DMEM supplemented with 20% FBS, 100 U/mL penicillin and 100 μg/mL streptomycin.

| In vivo migration assay of BMSCs
All animals used in this study were approved by the Institutional Animal Care and Use Committee of Jilin University and maintained under specific pathogen-free conditions. All animal procedures were conducted according to the guidelines approved by the China Association of Laboratory Animal Care.
To further evaluate the migration of BMSCs in vivo, BMSCs were marked with 40μg/ml Au-PEI. Nine male BALB/c-nu/nu mice (4-6 weeks old) (Vital River Laboratory Animal Technology, Beijing, China) were used to create subcutaneous tumours by injecting CAL27 cells (2 × 10 6 cells/mouse) into the dorsal flank. After tumour diameters reached about 5 mm on 3 weeks, mice were randomly divided into three groups (n = 3), BMSCs (at 1 × 10 6 cells)-Au-PEI, BMSCs (at 1 × 10 6 cells)-Au-PEI-SB225002 and Au-PEI groups. Then, these three different combinations were injected into different mice through the tail vein. Mouse was anaesthetized by inhalation of a mixture of oxygen with 5% sevoflurane, photographed under IVIS (In Vivo Imaging Instruments, KODAK) using spectrum imaging system at Ex/Em 420 nm/600 nm after 24 hours post-injection. Then, heart, liver, spleen, lung, kidney and tumour were collected, rinsed in saline and photographed. Tumours were further frozen-sectioned, dying the nucleus with DAPI. Distribution of BMSCs-Au-PEI was assessed by direct visualization using fluorescence microscope.

| Effect of BMSCs on OSCC in vivo
Twenty-five male BALB/c nude mice (6 weeks old, 20-25 g) were randomly divided into five groups (n = 5), CAL27, BMSCs + CAL27, co-CAL27, co-CAL27 + U0126 and co-CAL27 + SB431542. CAL27 cells were treated with U0126 or SB431542 (10 nM/ml) for 24 h prior to co-culture with BMSC-CM, and then co-cultured with BMSC-CM + U0126 or SB431542 (10 nM/mL) for 5 days prior to in vivo injection. Different treated 2 × 10 6 cells of CAL27 were injected into the dorsal flank to establish subcutaneous tumours. The size of the tumour was measured every 3 days until 39 days after 7 days of injection. Tumour volume was calculated with the formula a × b 2 × 0.5, where a is the largest diameter, and b is the smallest diameter. After 40 days, tumour xenografts were collected, fixed, paraffin-embedded, sectioned and stained with H&E and IHC detection by Ki67, vimentin and snail. The same experiment was repeated once.
Therefore, a total of 50 male BALB/c nude mice were used. Some mice were lost during the experiments. At the last time point, 6 mice left for CAL27 group, 8 mice for BMSCs + CAL27, 6 for co-CAL27, 4 for co-CAL27 + U0126 and 4 for co-CAL27 + SB431542 from both experiments. Tumorigenesis occurred in all mice used in this experiment. Figure 6A shows the results obtained from both the experiments, and Figure 6B shows the results obtained from second in vivo experiment.

| Statistical analyses
All data were analysed using SPSS 19.0 statistical software. One-way ANOVA was used to compare means of two groups or more. All tests were two-sided. Data were presented as mean ± SD. All experiments in vitro were repeated three times. P values <.05 were considered to be significant.

| CXCL8 and CXCL8 receptor expression profiles in BMSCs in vitro
Data from immunochemistry staining demonstrated that expression of CXCL8 was the highest in OSCC, second is oral epithelial   Figure 1G). Interestingly, we recognized that migration of BMSCs was dose-dependent on CXCL8, migration of BMSCs increases with the increase in CXCL8 ( Figure 1H). These data indicate that CXCL8 plays critical role in the migration of BMSCs in vitro. In situ staining from Figure 1A also shows that the CXCL8 is an interesting cytokine ( Figure 1A). after BMSCs co-cultured with CAL27 ( Figure S1A and S1B).
Importantly, SB225002 inhibitor, which inhibits CXCL8 binding to CXCR2, could effectively and significantly block BMSC migration in trans-well assay ( Figure S1C and S1D). These further confirm that CXCL8 plays critical role through CXCL8-CXCR2 interaction in the migration of BMSCs.

| Effects of CXCL8-CXCR2 on BMSC migration in vivo
To evaluate whether CXCL8-CXCR2 plays the same role as in vitro in BMSC migration towards OSCC in vivo, CAL27 cell lines were used to establish xenograft animal model of OSCC. In order to track migration of BMSCs, BMSCs were labelled by Au-PEI. Figure 2A and 2B shows that Au-PEI synthesized for this study was a 2-3 nm with red fluorescence, which could efficiently label BMSCs ( Figure 2C). After 24-h delivery, mice were detected by fluorescence molecular tomography. Figure 2D clearly shows that there were few small red fluorescence spots in Au-PEI group, was very stronger and big red fluorescence spot, which was located at the xenograft tumour area in the Au-PEI-BMSC group, and was significant and dramatical decrease in red fluorescence in Au-PEI-BMSCs + SB225002 compared with the Au-PEI-BMSC group, there were significant differences between Au-PEI-BMSC group and Au-PEI group or Au-PEI-BMSCs + SB225002. In order to exclude possible interference of autofluorescence, tumour, liver, heart, spleen, lung and kidney were isolated and collected to observe them in the dissociated situation. Figure 2E shows that red fluorescence signals were mainly in the tumour and very few in the liver and lung in the Au-PEI-BMSC group, while red fluorescence signals were significantly decreased and mainly in the lung and very few in the tumour in the Au-PEI-BMSC + SB225002 group. Compared with these two groups, red fluorescence signals were observed in the lung and some tumours were observed in the Au-PEI group ( Figure 2E). As shown ( Figure 2F), BMSCs migrated to tumour stroma in OSCC rather than tumour. These in vivo experimental data, especially data from SB225002 inhibitor, demonstrate that CXCL8/ CXCR2 also play an important role in the migration of BMSCs.

| BMSC promoted migration of OSCC
To identify whether BMSCs promoted CAL27 and FaDu cell migration, we performed a trans-well invasion assay, and we found that OSCC cells exhibited an enhanced vertical migratory capacity after treatment with BMSC-CM. Moreover, this migratory capacity increased along with co-culture time ( Figure 4A). As shown in Figure 4B, to further confirm this result, we performed a scratch assay to assess the plane migratory capacity of cells, CAL27 and FaDu cells treated with BMSC-CM migrated nearly 100% of the scratch within 24 hours, while control groups failed to close the scratch after 24 hours, which produced similar results with transwell invasion assay.

| BMSCs promoted OSCC cells to express EMT relative molecules through TGF-β1/Ras/Raf/Erk signalling pathway
We and FaDu cells after inhibiting TGF-β1 and Erk ( Figure S2C and Figure S2D). These results suggested that the proliferation ability was inhibited, migration ability was decreased in OSCC cells when  Figure 6A shows that there were significant differences in tumour volume in BMSC-treated mice compared with control mice; meanwhile, mix culture and CM culture have not difference. Tumours in mice with added inhibitor cells were significantly smaller than those in the control group ( Figure 6A and 6B). Furthermore, we observed that the Ki67 of BMSCs + CAL27 and co-CAL27 groups was strongly positive compared with the control group, but inhibitor groups were lower than the control group. The results suggest

F I G U R E 5 BMSCs promoted OSCC cells that express EMT relative molecular changes by TGF-β1/Ras
BMSCs promote the formation of xenograft tumours. In addition, we found that expressions of vimentin and snail were increased in BMSCs + CAL27 and co-CAL27 groups compared with the matched CAL27 group. Consistently, the increased vimentin and snail expressions in inhibitor groups were decreased. These results indicate BMSCs also promote the extent of tumour invasion and EMT ( Figure 6C).

F I G U R E 6
BMSCs promoted tumorigenesis and EMT in murine xenograft model of CAL27 cells. A, The initial in vivo growth analysis of xenografts. There was a significant difference in tumour size between these groups up to day 39 after xenograft implantation in nude mice. B, Representative photographs of xenografts nude mice and xenografts at day 42 after tumour implantation. C, The expression of Ki67, vimentin and snail in xenografts tumours. Representative images of Ki67, vimentin and snail immunohistochemistry at 10× and 40× magnification. *, P < .05; **, P < .01; and ***, P < .001  [29][30][31][32][33] Interestingly, we also recognized that CXCL8 is positive in human OSCC tissues in situ, and two OSCC cell lines, CAL27

| D ISCUSS I ON
and FaDu, can secrete CXCL8 into CCM in vitro ( Figure 1). Also, CXCL8 involves in BMSC migration and induces CXCR2 receptor expression in BMSCs (Figure 1 and Figure S1). Previous studies have reported that CXCL8 can bind to two receptors, CXCR1 and CXCR2.
CXCR1 and CXCR2 express on many kinds of cells, leucocytes,  (Figure 2), which indicates that the recruitment of BMSCs depends on interaction between CXCL8 and CXCR2 ( Figures S1 and   S2). These results suggest that CXCL8-CXCR2 indeed play important roles in tumorigenesis and development of OSCC.
EMT is a critical process of tumorigenesis of epithelial tumours.
Upon EMT, cancer cells acquire migratory and invasive properties. 36 Our data clearly showed that BMSCs promote the EMT of OSCC cells leading to cancer cell proliferation, decrease apoptosis and necrosis of OSCC cells and increase migration of OSCC (Figures 3, 4 and 5).
Further, we have found that TGF-β/Ras/Raf/Erk signalling pathway has emerged in the OSCC cells after co-culture ( Figure 5 and is an important one, which is the first Raf isoform identified as a potential cellular oncogene. 39 TGF-β1 is the most central mediator in the proliferation and EMT of epithelial cells. ERK, a member of the MAPK family, is the major participant in the regulation of cell growth, differentiation and EMT. 40 In our results, TGF-β1 and p-Erk are increased in co-culture condition, which indicates that both are activated ( Figure 4). Importantly, both inhibition assays for TGF-β1 and p-Erk result in significantly decreasing the proliferation and migration of OSCC cells in vitro ( Figure S2), and significantly decreasing tumour size compared with the control group in vivo ( Figure 6). BMSC is a critical component of the host-response network in tumours and provides a potential novel anti-cancer target for the treatment of OSCC. Data from the in vivo experiments ( Figure 6) may also suggest that the time we used for co-culture with BMSCs or condition medium was enough to reprogramme CAL27 cells or turn on some important signals in the CAL27 cells resulting in CAL27 cell progressive growth. This is our new hypothesis needed to further study. In this study, we still do not understand effects of BMSCs on normal epithelial cell, such as HaCaT from our current study although we demonstrate that BMSCs can promote tumour. We will perform more in vitro and in vivo experiments to reveal the truth in the further study.

ACK N OWLED G EM ENTS
This study was supported by grants from National Key R&D

CO N FLI C T O F I NTE R E S T
The authors declare that they have no competing interests.

AUTH O R CO NTR I B UTI O N S
ML contributed to conception and design, data acquisition, analysis and interpretation, drafted and critically revised the manuscript; ZY contributed to data acquisition, analysis and interpretation, and critically revised the manuscript; BW contributed to design and data acquisition, and critically revised the manuscript; LX contributed to design and data acquisition, and critically revised the manuscript; LX, ZD, CY and ZS contributed to data acquisition and critically revised the manuscript; LQ contributed to design and critically revised the manuscript; SH and L.QL. contributed to conception, design, data acquisition, analysis and interpretation, and drafted and critically revised the manuscript. All authors gave final approval and agreed to be accountable for all aspects of the work.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available within the article and its Appendix or on reasonable request from the corresponding author.