CX3CL1 induces cell migration and invasion through ICAM‐1 expression in oral squamous cell carcinoma cells

Abstract Human oral squamous cell carcinoma (OSCC) has been associated with a relatively low survival rate over the years and is characterized by a poor prognosis. C‐X3‐C motif chemokine ligand 1 (CX3CL1) has been involved in advanced migratory cells. Overexpressed CX3CL1 promotes several cellular responses related to cancer metastasis, including cell movement, migration and invasion in tumour cells. However, CX3CL1 controls the migration ability, and its molecular mechanism in OSCC remains unknown. The present study confirmed that CX3CL1 increased cell movement, migration and invasion. The CX3CL1‐induced cell motility is upregulated through intercellular adhesion molecule‐1 (ICAM‐1) expression in OSCC cells. These effects were significantly suppressed when OSCC cells were pre‐treated with CX3CR1 monoclonal antibody (mAb) and small‐interfering RNA (siRNA). The CX3CL1‐CX3CR1 axis activates promoted PLCβ/PKCα/c‐Src phosphorylation. Furthermore, CX3CL1 enhanced activator protein‐1 (AP‐1) activity. The CX3CR1 mAb and PLCβ, PKCα, c‐Src inhibitors reduced CX3CL1‐induced c‐Jun phosphorylation, c‐Jun translocation into the nucleus and c‐Jun binding to the ICAM‐1 promoter. The present results reveal that CX3CL1 induces the migration of OSCC cells by promoting ICAM‐1 expression through the CX3CR1 and the PLCβ/PKCα/c‐Src signal pathway, suggesting that CX3CL1‐CX3CR1‐mediated signalling is correlated with tumour motility and appealed to be a precursor for prognosis in human OSCC.


| INTRODUC TI ON
Oral cancer is the most common type of head and neck squamous cell carcinoma (HNSCC), the most prevalent malignancy worldwide. 1 Human oral squamous cell carcinoma (OSCC) accounts for most oral cancers and takes various clinical forms with high mortality and morbidity rates, resulting from the cancer metastasis and locoregional recurrence. 2,3 Its survival rate is strongly correlated with the TNM (tumour size, nodal metastasis and distant metastasis) stage of the primary tumour. 4 Even thoughdiagnostic techniques and treatment modalities are advanced for OSCC, its 5-year relative survival rate is less than 50% for advanced-stage oral SCC because of large lesions and clinically detectable metastases to cervical lymph nodes at the time of diagnosis. 5 Thus, understanding and recognizing the molecular mechanism at the early stage is critical for prognosis of patients with OSCC.
C-X3-C is a chemokine family member, primarily affecting leukocyte migration and tumour growth and development. 6 C-X3-C motif ligand 1 (CX3CL1) is a multifunctional inflammatory chemokine that plays a central role in the pathogenesis of inflammationdriven malignancies. 7 CX3CL1 mediates immune cell adhesion in membrane-anchored form and acts as a chemotactic cytokine in the soluble form. 8 Studies also demonstrated that CX3CL1 and its unique chemokine receptor CX3CR1 played an essential role in cancer development. CX3CL1-CX3CR1 axis regulates numerous cellular functions relevant to cancer survival mechanisms, including proliferation, migration, invasion and apoptosis resistance. [9][10][11] However, contradictory data exist concerning the role in tumour progression in the specific cancer type for CX3CL1-CX3CR1. 7,[12][13][14] The CX3CL1-CX3CR1 axis was reported to predict better prognosis and fewer recurrences in hepatocellular carcinoma. 8 In contrast, activation of the CX3CL1-CX3CR1 axis promotes various neoplasia responses, including tumorigenesis and progression in ovarian carcinoma and gastric, pancreatic and lung cancer. [15][16][17][18] Nevertheless, the investigation of the stimulatory effects of CX3CL1 on the tumour cell motility and progression in OSCC cells at signalling mechanism and gene expression levels was limited.
Metastasis is associated with multiple steps in which tumour cells depart from the primary tumour site and migrate to other secondary locations through the bloodstream, lymphatic system or direct extension. 19 CX3CL1 promoted cell migration and lung metastasis in human osteosarcoma by upregulating intercellular adhesion molecule-1 (ICAM-1) expression. 20 The CX3CL1/ICAM-1 signalling network facilitated cell adhesion and transendothelial migration and initiated spinal metastasis in non-small-cell lung cancer. 21 ICAM-1, a transmembrane glycoprotein in the immunoglobulin superfamily, is constitutively expressed at a low basal level in vascular endothelial and epithelial cells, but its expression is upregulated for the stabilization of cell-cell interactions and endothelial transmigration during inflammation. [22][23][24] The expression of ICAM-1 reveals the correlation between the immune system and tumour growth, even though the pro-or anti-tumour immune roles of ICAM-1 are controversial. [25][26][27] ICAM-1 exhibits upregulated expression in malignant conditions, potentially participating in tumorigenesis and promoting the metastatic ability of melanoma, breast, gastric, pancreatic and lung cancers. [28][29][30] In contrast, some studies revealed that upregulation of ICAM-1 expression reflected low growth potential and good prognosis in breast, gastric and colorectal cancers. [31][32][33][34] In this study, we evaluated the role of CX3CL1 in the activation of ICAM-1 and their roles in cell movement, cell migration and invasion in OSCC. We utilized the OSCC cell lines SCC4 and SAS and found that CX3CL1 activates PLCβ, PKCα and c-Src. Using a combination of inhibitors or siRNAs for these kinases, we found a decrease in metastasis-associated events such as cell migration and invasion in OSCC. The data represented here contributes to the molecular characterization of the signalling events associated with CX3CL1 contributions to cell migration and invasion in OSCC.
Human recombinant protein CX3CL1 (300-31) was acquired from PeproTech and dissolved according to the manufacturer's manual.
All other chemicals and reagents were acquired from Sigma-Aldrich.

| Online database analysis
The CX3CL1 expression and clinical HNSCC patient data were downloaded from The Cancer Genome Atlas (TCGA) and analysed using an interactive web resource (UALCAN, http://ualcan.path.uab.edu/ index.html). 35

| Cell migration and invasion assay
Migration and invasion assay were performed using a 24-well transwell insert chamber (8.0 μm pore size; Corning Incorporated) without and with Matrigel-coated membrane, respectively. OSCC cells at a density of 5 × 10 4 cells per well were seeded onto the upper chamber with 5% CO 2 at 37°C. Approximately 500 μL of cell culture media containing 1% FBS was added to the lower chamber, which was supplemented with the recombinant protein CX3CL1 at three different concentrations (control, 10 and 30 ng/mL). After 24 h incubation, the cells migrating to the bottom chamber were fixed with 1% formaldehyde for 15 min and stained with 0.05% crystal violet for 30 min. The number of cells was captured using an inverted microscope analysed using ImageJ software. In addition, OSCC cells pre-treated with IgG or CX3CL1 monoclonal antibody (mAb) were also used for migration and invasion assay. Four independent experiments were conducted.

| Total RNA extraction and real-time polymerase chain reaction (qPCR) assay
Total RNA in OSCC cells was purified using TRI reagent (9424; Sigma-Aldrich) according to the manufacturer's instructions. The total RNA was reverse transcribed into complementary DNA (cDNA) using RT-PCR kit (PCRBIOSYSTEM), which was analysed using KAPA SYBR

| Total protein extraction and western blot assay
Total protein of OSCC cells was collected using lysis buffer containing protease inhibitor as previously described. 36

| Immunofluorescence assay
After SCC4 and SAS cells were pre-treated with specific inhibitors for 1 h, followed by recombinant protein of CX3CL1 for 2 h. Cells were fixed with 1% formaldehyde for 30 min, permeabilized with 0.05% Triton X-100 at room temperature for 5 min, and then blocked with buffer (Goal Bio; W-3400) at room temperature for 1 min. Cells were incubated with primary antibody specific for c-Jun (1:100) at 4°C overnight. Slides were washed three times with PBS and given appropriate FITC secondary antibody (1:100) at room temperature for 2 h. Then, cells were stained with DAPI (5 μg/mL) for 5 min and observed using an inverted microscope.

| Activator protein 1 (AP-1) luciferase assay
Human oral squamous cell carcinoma cells were seeded in a 12well plate for 24 h and transfected with an AP-1 luciferase plasmid (Promega) for 24 h according to the manufacturer's instructions.
Cells were treated with specific inhibitors and CX3CL1 recombinant protein for 24 h. The microplates were read for luciferase by 2030 Multilabel Reader (VICTOR™X2; PerkinElmer).

| Chromatin immunoprecipitation (ChIP) assay
After the pre-treatment of inhibitors for 1 h and followed by

| Statistical analysis
All results were analysed by using Sigmastat program. All the quantified results were expressed at the mean ± standard deviation (SD) and analysed with one-way ANOVA followed with Fisher's least significant different (LSD) post hoc test. All results were considered p < 0.05 as a significant difference.

| Clinicopathological characteristics of CX3CL1
To confirm the clinical significance of levels of CX3CL1 expression in OSCC, analysis of GEO microarray 3524 OSCC tissue samples revealed higher levels of CX3CL1 mRNA expression compared with levels in adjacent non-tumour samples (p < 0.05; Figure 1A). As shown in Figure 1B, we found higher levels of CX3CL1 mRNA expression in TCGA sample of HNSC tissue compared with the oral mucosal tissue samples, and levels of CX3CL1 expression were significantly associated with clinical disease staging (p < 0.05). Results of IHC staining for levels of CX3CL1 in patients were with a highergrade OSCC than in those with a lower-grade OSCC; the level of CX3CL1 expression was reflected by the tumour stage ( Figure 1C,D).
To assess the prognostic value of CX3CL1 expression in patients with OSCC, we investigated the associations among the CX3CL1 expression, overall survival (OS). We observed higher levels of CX3CL1 expression in patients with OSCC, which was also correlated with poor OS (p = 0.016) ( Figure 1E). The results suggest that CX3CL1 is overexpressed in OSCC and is correlated with clinical stage and poor prognosis.

| CX3CL1 upregulates OSCC cell migration and invasion
CX3CL1 corresponds with node metastasis and early tumour stage in OSCC patients. 37 Cancer metastasis involves sequential steps, such as cell movement, migration and invasion. In this study, we investigated the mechanism through which CX3CL1 regulates cell movement, cell migration, cell invasion and cell proliferation in OSCC. The wound-healing assay and the Transwell migration assay revealed that CX3CL1 induced cell motility in the SCC4, SCC25 and SAS cell lines (Figure 2A,B). Furthermore, treatment with different concentrations of CX3CL1 enhanced cell invasiveness ( Figure 2C). While simulation with CX3CL1 did not affect the proliferative capacity of oral cancer cells ( Figure S1). However, these activities were inhibited when the cells were treated with CX3CL1 neutralizing antibody ( Figure 2D,E). These findings confirmed that CX3CL1 promotes cell motility and metastatic activity of OSCC cells.

| CX3CL1 upregulates cell migration via intercellular adhesion molecule-1 (ICAM-1) expression
Cell adhesion molecules play a significant role in cell migration and cell-matrix interactions, and tumour dissemination. 38,39 The expression of adhesion molecules enables the interaction between tumour cells and the surrounding stroma, resulting in tumour metastasis. 40 Therefore, the function of ICAM-1 and VCAM-1 in OSCC cells was explored. Records from the TCGA dataset analysis revealed higher levels of ICAM-1 and VCAM-1 in tumour specimens compared with the normal specimens ( Figure 3A,B).

| CX3CL1 induces ICAM-1 expression through CX3CR1 to promote tumour cell motility in OSCC
The CX3CR1 is a specific receptor for CX3CL1 and is involved in CX3CL1-mediated cancer progress. 41 To investigate associations between CX3CR1 expression in OSCC, we first analysed levels of CX3CR1 mRNA expression in normal and tumour tissue. Data from the GSE13601 microarray showed that, compared with the non-tumour tissues, the mRNA expression of CX3CR1 was significantly downregulated in the tumour tissues (p < 0.01; Figure 4A). In addition, data from the GSE78060 microarray showed that CX3CR1 expression was positively correlated with

| CX3CL1 induces cell motility by PLCβ, PKCα and c-Src activation
Studies indicated that the activation of PLCβ, PKCα or c-Src has a critical role in tumour growth, cell migration, invasiveness and metastasis in breast cancer, colon carcinoma, lung squamous cell carcinoma and hepatocellular carcinoma cells. [42][43][44][45] Here, we investigated whether CX3CL1 activated these signalling pathways in OSCC. The Moreover, incubation with CX3CL1 obviously induced phosphorylation of PLCβ, PKCα and c-Src ( Figure 5D,E). Finally, we certified our finding by using siRNA to suppress these pathways activation, and the data showed that transfection with PLCβ, PKCα and c-Src siRNA apparently reversed cell migration and ICAM-1 expression after CX3CL1 incubation ( Figure 5E-G). Similarly, the CX3CL1induced phosphorylation of PLCβ, PKCα and c-Src was reversed by pre-treatment with CX3CR1 neutralized antibody ( Figure S2). These results suggest that the PLCβ/PKCα/c-Src signalling pathways contribute to CX3CL1-promoted cell migration in OSCC.

| AP-1 activation and c-Jun phosphorylation are involved in the CX3CL1-induced tumour cell motility and ICAM-1 expression in OSCC cells
The nuclear transcription factor AP-1, associated with cell proliferation and tumorigenesis, has been identified as a transcription regulator of ICAM-1 expression. 46  We also checked whether CX3CR1, PLCβ, PKCα and c-Src were the upstream regulator of AP-1, with pre-treatment of CX3CR1 neutralized antibody, PLCβ, PKCα or c-Src inhibitors, the CX3CL1induced c-Jun nuclear translocation, representing for AP-1 activation, was reversed by pre-treatment with CX3CR1 neutralized antibody, PLCβ, PKCα and c-Src inhibitors ( Figure 7A), as well as c-Jun phosphorylation ( Figure 7B). Finally, to further examine the transcriptional activation of AP-1 is responsible for CX3CL1 effects, the luciferase reporter assay was performed. The AP-1 activity was significantly increased after CX3CL1 treatment ( Figure 7C). However, in the presence with pathway inhibitors of CX3CR1, PLCβ, PKCα and c-Src, the AP-1 activity was reversed after CX3CL1 treatment ( Figure 7D). In addition, transcriptional activation of AP-1 was further investigated whether it participates in CX3CL1-promoted ICAM-1 expression. The chromatin immunoprecipitation (ChIP) assay was examined for evaluation and it was suggested that these pathway inhibitors could abolish the binding ability of CX3CL1-induced c-Jun to the AP-1 binding element on the ICAM-1 promoter ( Figure 7E). These results demonstrated that the CX3CL1-driven CX3CR1, PLCβ, PKCα and c-Src pathway is involved in the regulation of AP-1 expression and nuclear translocation, and subsequently affected the expression of AP-1dependent ICAM-1.

| DISCUSS ION
Human oral squamous cell carcinoma is one of the common malignant tumours occurring in the oral cavity with a low 5-year survival rate, especially in the advanced stage. Its low survival rate is associated with regional metastasis involving penetration of cervical lymph nodes and distant metastasis resulting from the invasion-metastasis cascade involving activation of various kinases and molecules. 47,48 Understanding the pathogenesis of cancer metastasis at the molecular level and kinase pathway could offer approaches for therapeutic targeting. 49,50 The present study demonstrated that the CX3CL1-

| CON CLUS IONS
In the present study, high levels of CX3CL1 expression were evi- writing -original draft (equal); writing -review and editing (equal).

ACK N OWLED G EM ENT
This study was supported by grants from the Ministry of Science and Technology (NSTC-111-2320-B-038-038) and Taipei Medical University (TMU108-AE1-B47).

CO N FLI C T O F I NTER E S T S TATEM ENT
The authors declare that they have no conflict of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
The datasets generated for this study can be accessed upon request to the corresponding author.