Chemokine (CC motif) ligand 18 upregulates Slug expression to promote stem‐cell like features by activating the mammalian target of rapamycin pathway in oral squamous cell carcinoma

Abstract

Chemokine (CC motif) ligand 18 (CCL18) is involved in remodeling of the tumor microenvironment and plays critical roles in oncogenesis, invasiveness, and metastasis. We previously investigated the overexpression of CCL18 in primary oral squamous cell carcinoma (OSCC) tissues and its association with advanced clinical stage in OSCC patients. However, the underlying mechanisms of this CCL18‐derived activity remains unidentified. This study showed exogenous CCL18 increased cell migration and invasion and induced cell epithelial–mesenchymal transition (EMT), and that E‐cadherin, an epithelial marker, decreased and N‐cadherin, a mesenchymal marker, increased, compared to negative control in OSCC cells. Furthermore, we detected that CCL18 induced the acquisition of cancer stem(‐like) cell characteristics in oral cancer cells, but also found a significantly positive correlation between the expression of CCL18 and Bmi‐1 (P < 0.001) in OSCC surgical specimens by immunohistochemistry analysis. The expression of octamer‐binding transcription factor 4 and Bmi‐1 were significantly upregulated, and proportions of aldehyde dehydrogenase^high+ cells and CD133⁺ cells were markedly increased in CCL18‐treated cells compared to untreated cells. Sphere formation ability was observably enhanced when cells were continually exposed to high levels of CCL18. Moreover, CCL18 upregulated Slug expression by stimulating the mammalian target of rapamycin (mTOR) signaling pathway in OSCC cell lines. Inhibition of the mTOR pathway by INK128, or Slug knockdown by RNA interference, reversed CCL18‐induced EMT and the stemness response at both molecular and functional levels. In conclusion, our data suggested that CCL18 upregulated Slug expression to promote EMT and stem cell‐like features by activating the mTOR pathway in oral cancer. These findings provide new potential targets for the early diagnosis and treatment of OSCC.

Oral squamous cell carcinoma is the most common and aggressive epithelial tumor in the head and neck region, accounting for approximately 90% of oral malignancies and with a rising incidence in many countries. The 5‐year survival rate of OSCC patients has not significantly improved and remains <50%, despite extensive studies on pathogenesis, diagnosis, and therapy.1, 2, 3 Cervical lymph node metastasis and occasional distant organ metastasis are believed to be the leading causes of the high mortality of OSCC.4 So, it is necessary to elucidate mechanisms underlying OSCC tumorigenesis and development.

Chemokine (CC motif) ligand 18 is one of important molecular components in immunological and inflammatory processes, triggering biological activity in dendritic cells, monocytes/macrophages, fibroblasts, and cancer cells. It is involved in the remodeling of the tumor microenvironment and has crucial roles in oncogenesis, invasiveness, and metastasis.5, 6 Recent studies have shown that increased expression of CCL18 was observed in a variety of cancers and associated with clinicopathological features.5, 7, 8, 9, 10, 11, 12 We previously reported that CCL18 was overexpressed in primary OSCC tissues and associated with tumor TNM stage. Moreover, increased CCL18 acted in an autocrine manner to enhance cancer cells growth and invasion.13 However, the underlying molecular mechanisms by which CCL18 contributes to cancer cell invasion and metastasis in OSCC remain unclear.

Cancer stem(‐like) cells are a subgroup of cancer cells with a small percentage and have an ability to regenerate various cell types within tumors. Notably, CSCs are considered more resistant to radiotherapy and chemotherapy and responsible for tumor maintenance, metastasis, and recurrence.14, 15 Cancer stem(‐like) cells have been isolated from many cancer types, including head and neck squamous cell carcinoma,15 and accumulating evidence has shown a correlation between EMT and the acquisition of stem‐cell‐like traits in neoplastic cell populations.16 Epithelial–mesenchymal transition is a strictly regulated process during embryonic development and cancer progression through which an epithelial cell loses its cell polarity and cell–cell adhesion but acquires the capacity to migrate and metastasize.17, 18 Many groups have confirmed the hypothesis that the induction of EMT promotes cancer cell migration and invasion, and contributes to the therapeutic resistance and enrichment of cells with stem‐cell‐like features.15

In this study, for the first time, we evaluated the effects of exogenous CCL18 on EMT and stem‐cell like characteristics in OSCC cell lines, and then determined the relationship between CCL18 and CSC phenotypes in human OSCC specimens. Furthermore, we detected the potential mechanisms underlying the function of CCL18, and used Slug RNAi and mTOR inhibitor to reverse this function in vitro. Our results suggest that CCL18 promoted EMT and stem‐cell‐like traits of OSCC by activating the mTOR–Slug pathway, providing new targets for the early diagnosis and treatment of OSCC.

Materials and Methods

Cell lines and tumor samples

The OSCC cell lines, HSC6 and CAL33, were kindly provided by J. Silvio Gutkind (NIH, Bethesda, MD, USA). SCC15 was purchased from ATCC (Manassas, VA, USA). The cells were incubated in a mixture of DMEM (Gibco, Grand Island, NY, USA) and FBS (Gibco) at a ratio of 9:1, 5% CO₂, 37°C.

A total of 45 OSCC patients who underwent surgical resection at the Department of Craniofacial Surgery, Guanghua School of Stomatology, Sun Yat‐sen University (Guangzhou, China) were enrolled in the study. Primary OSCC tissues were obtained from the most representative areas of each case, postoperatively. Medical records for all OSCC cases were reviewed for clinical information. We obtained informed consent from each patient. The study was approved by the Ethics Committee of Guanghua School of Somatology, Sun Yat‐sen University.

Information regarding reagents and antibodies is listed in Table S1.

Small interfering RNA

Slug siRNA and Si‐NC were designed by Ribobio (Guangzhou, China). Three different Slug siRNA duplexes were tested, and the sequences were as follows: SiSlug1,5ʹ‐GGACCACAGTGGCTCAGAA‐3ʹ; SiSlug2, 5ʹ‐GGAGCATACAGCCCCATCA‐3ʹ; and SiSlug3, 5ʹ‐CTTCAAGGACACATTAGAA‐3ʹ. Cells were seeded in 6‐well plates overnight and then transfected with siRNAs using the Lipofectamine RNAiMAX reagent and Opti‐MEM (Life Technologies, Carlsbad, CA, USA) according to the manufacturer's instruction.

Immunohistochemistry

Tissue sections from paraffin‐embedded OSCC were dewaxed in xylene and rehydrated in a graded alcohol series, soaked with 0.1% Triton X‐100, incubated in 3% H₂O₂ to eliminate endogenous peroxidase activity, heated in sodium citrate buffer (pH 6.0) for antigen retrieval, and then incubated in goat serum followed by incubation with CCL18 or Bmi‐1 primary antibody overnight at 4°C. After washing with PBST, the sections were incubated with secondary antibody for 30 min, and then visualized with DAB solution and counterstained with hematoxylin. The expression of CCL18 was quantified using a visual grading system based on the degree of staining. The positive cell percentages were classified as: 0, <5%; 1, 5–30%; 2, 30–70%; and 3, >70%. Staining intensity was graded as: 0, none; 1, weak; 2, moderate; and 3, strong. Five representative fields at 400× magnification were evaluated for each sample. A weighted staining value (S) was calculated by multiplying the positive cells percentage and the score of the staining intensity. Finally, all samples were assigned to three levels according to the S value: negative, S = 0; low expression, 0 < S < 6; high expression, 6 ≤ S ≤ 9. Bmi‐1 expression was defined as positive when typical nuclear staining was observed.

Western blot analysis

The cultured cells were lysed with RIPA buffer (Abcam, Cambridge, MA, UK), then centrifuged, and the precipitate discarded. The total protein concentrations of the supernatant were measured by the BCA protein assay kit (Cwbiotech, Beijing, China). Then the proteins were separated by 10% or 12% SDS‐PAGE and blotted onto PVDF membranes (Millipore, Billerica, MA, USA). The membranes were blocked in 5% non‐fat milk for 1 h at room temperature and then incubated with primary antibodies overnight at 4°C. Subsequently, the membranes were washed with TBST and incubated with HRP‐conjugated secondary antibody for 1 h. The immunoreacted bands were visualized with a highly sensitive chemiluminescence (ECL) detection system (Millipore). ImageJ (Bethesda, MD, USA) software was used to quantify the gray value (A) of the bands. Then we got the ratio (B) of the target protein's A to internal reference protein's A. Finally, the ratio (C) values of the test group's B to the control group's B from triplicate experiments were obtained for statistical analysis.

Reverse transcription–quantitative real‐time PCR

Total RNA was prepared with TRIzol (Invitrogen, Carlsbad, CA, USA) and reversed to synthesize cDNA according to the manufacturer's procedure (TaKaRa, Shiga, Japan). The real‐time PCR was carried out using the LightCycler 480 SYBR Green I Master system (Roche, Basel, Switzerland). The relative quantification of mRNA levels was evaluated by the comparative Cp (ΔΔCp) method with GAPDH as the internal control gene. The primers used were: for Slug, sense, 5′‐TATTTGGTTGGTCAGCACAGG‐3′ and antisense, 5′‐GACGCAATCAATGTTTACTCG‐3′; for OCT4, sense, 5′‐GGT ATTCAGCCAAACGACCA‐3′ and antisense, 5′‐CCTCTCACTCGGTTCTCGAT‐3′; for Bmi‐1, sense, 5′‐CCAGGGCTTTTCAAAAATGA‐3′ and antisense, 5′‐CCGATCCAATCTGTTCT GGT‐3′; and for GAPDH, sense, 5′‐GCACCGTCAAGGCTGACAAC‐3′ and antisense, 5′‐TGGTGAAGACGCCAGTGGA‐3′.

Immunofluorescence

The prepared cells were plated on confocal culture dishes and cultured normally overnight. Cells were then fixed with 4% formaldehyde for 20 min, permeabilized with 0.1% Triton X‐100 for 20 min, and blocked with goat serum for 30 min. Cells were treated with primary antibodies overnight at 4°C, followed by Dylight 594‐conjugated and Dylight 488‐conjugated secondary antibodies (1:200; Abcam) protected from light for 1 h at 37°C; subsequently, cell nuclei were stained with DAPI (Invitrogen) for 5 min. The confocal culture dishes were finally observed under a confocal laser scanning microscope (Carl Zeiss, Oberkochen, Germany) and representative fields of view at 200× magnification were randomly imaged for each group.

Transwell assay

Cell migration and invasion capacities were measured by a Transwell assay (Corning, Toledo, OH, USA). In contrast to the migration assay, the upper chamber of the insert was precoated with 0.1 mL (300 μg/mL) Matrigel matrix (Corning) for the invasion assay. In both assays, the prepared cells were seeded in the upper chamber with serum‐free medium, but the medium of the lower chamber was supplemented with 10% FBS as a chemoattractant. After incubation for 24 h, the cells were fixed with 4% formaldehyde. The cells not migrating or invading through the pores were removed with a cotton swab. Those that had migrated or invaded onto the lower surface of membrane were stained by crystal violet. Finally, five representative fields at 100 ×  magnification were randomly imaged and quantified for each well using a light microscope (Carl Zeiss).

Spheroid formation assay

Cells were seeded in low‐adhesion 6‐well plates (2000 cells/well) and cultured in DMEM/F12 (Gibco) without FBS. This tumor sphere medium was supplemented with N2 supplement, 20 ng/mL human recombinant basic fibroblast growth factor, and 20 ng/mL epidermal growth factor (Gibco) in the absence or presence of CCL18 (20 ng/mL) and/or INK128 (100 μM). After 10 days of incubation, the primary spheres larger than 100 μm were counted for each well. Then the primary spheres were dissociated into single cells and seeded in the same culture conditions. Ten days later, secondary spheres larger than 100 μm were similarly counted.

Flow cytometry

Cells were digested by 0.25% trypsin and 0.02% EDTA (Gibco). After centrifuged in media, the cells were washed and counted in PBS containing 0.5% BSA. They were then adjusted to a concentration of 1 × 10⁶ cells/mL and incubated within the allophycocyanin‐conjugated anti‐human CD133 for 45 min, and finally washed. The ALDH enzymatic activity was measured with the ALDEFLUOR kit (Stem Cell Technologies, Vancouver, BC, Canada) according to the manufacturer's protocol. The cells treated with ALDH inhibitor diethylaminobenzaldehyde (50 mmol/L) were used as a negative control. Flow cytometry analysis was carried out on CytoFLEX S (Beckman Coulter, Brea, CA, USA). Duplicates and dead cells were excluded by gating with forward scatter and side scatter.

Statistical analysis

All statistical analyses were undertaken with spss 20.0 software (SPSS, Chicago, IL, USA). Data were analyzed using Student's t‐test or one‐way anova and were displayed as the means ± SEM of at least three independent experiments. The association between Bmi‐1‐positive and CCL18^high expression in immunohistochemistry experiments was analyzed using the χ²‐test. P‐values < 0.05 were considered statistically significant.

Results

Chemokine (CC motif) ligand 18 promoted EMT of OSCC cells

Previously, we have shown that CCL18 was overexpressed and positively correlated with advanced clinical stage in OSCC patients.13 Epithelial–mesenchymal transition was thought to play a key role in invasion and metastasis of many tumor types.17 To confirm whether CCL18 enhances OSCC invasion and metastasis through EMT, we selected two OSCC cell lines, HSC6 and CAL33, to undertake a Transwell assay. The results displayed that, after stimulation with 20 ng/mL CCL18, there were considerably more cells migrating or invading to the lower surface than the negative control (Fig. 1a,b). Moreover, we detected that the expression of the EMT‐associated molecule, E‐cadherin, an epithelial marker, decreased, whereas the expression of the mesenchymal marker, N‐cadherin, increased in HSC6 and CAL33 cells after 3 days of treatment with CCL18 (Fig. 1c). In addition, the immunofluorescence staining of the cell lines showed the same results with Western blot analysis (Fig. 1d). Hence, our results indicated that CCL18 induced EMT and then promoted migration and invasion of OSCC cells.

Figure 1.

Chemokine (CC motif) ligand 18 (CCL18) promoted epithelial–mesenchymal transition in oral squamous cell carcinoma. (a,b) HSC‐6 and CAL33 cells were cultured in the absence or presence of CCL18 (20 ng/mL) for 48 h, then their migration and invasion abilities were examined by Transwell assay. Representative pictures and the mean number of cells that migrated (a) or invaded (b) to the lower surface are shown. (c) Epithelial–mesenchymal transition ‐associated markers, E‐cadherin and N‐cadherin, were examined by Western blotting after cells were treated with or without CCL18 (20 ng/mL) for 2 or 3 days. α‐Tubulin was used as an internal control. (d) Immunofluorescence staining of E‐cadherin (red) and N‐cadherin (green) was indicated in HSC6 and CAL33 cells with or without CCL18 for 3 days. Nuclear DNA was stained with DAPI (blue). Data are presented as the means ± SEM of at least three independent experiments, *P < 0.05 compared to the non‐target control (NC).

Chemokine (CC motif) ligand 18 enhanced stem cell‐like characteristics in OSCC

A growing number of studies have suggested that EMT is associated with the gain of molecular and functional traits of stem‐like cells in normal and tumor cell populations.16 Here, we detected the mRNA and protein expressions of stemness‐related markers, OCT4 and Bmi‐1,19, 20 which were significantly upregulated in CCL18‐treated cells compared to untreated cells (Fig. 2a). Flow cytometry was carried out to analyze the percentages of ALDH^high+ cells and CD133⁺ cells, which are both used extensively to identify CSCs of oral cancer.19, 20 The results showed that the percentage of ALDH^high+ cells and CD133⁺ cells in HSC6 and CAL33 cells were significantly increased after treatment with CCL18 for 3 days (Fig. 2b). We also used a sphere formation assay to identify the stem‐like cells and found that CAL33 cells, but not HSC6 cells, efficiently formed tumor spheroids (data not shown). The sphere‐forming efficiency of CCL18‐treated CAL33 cells was markedly promoted, and evaluated and quantified as average primary and secondary sphere numbers (Fig. 2c). SCC15 cells, another OSCC line, were also found to have enhanced sphere‐forming capability in response to CCL18 (Fig. S1).

Figure 2.

Chemokine (CC motif) ligand 18 (CCL18) enhanced stem cell‐like characteristics in oral squamous cell carcinoma (OSCC) cells. (a) HSC6 and CAL33 cells were treated with or without CCL18 (20 ng/mL) for 3 days. Then the protein and mRNA expressions of OCT4 and Bmi‐1, were examined by Western blotting and reverse transcription–quantitative real‐time PCR. α‐Tubulin was used as an internal control. (b) Percentages of aldehyde dehydrogenase (ALDH)^high+ cells and CD133⁺ cells are presented in each of the indicated OSCC cell lines with or without 3‐day CCL18 treatment. (c) CAL33 cells were cultured in low‐adherence conditions to form spheres in 6‐well plates with or without continuous stimulation of 20 ng/mL CCL18. Images of primary and secondary spheres are shown (left), and average numbers of primary and secondary spheres (>100 μm) were calculated for each well (right). Data (a–c) are presented as the means ± SEM of at least three independent experiments. *P < 0.05 compared to the non‐target control (NC). (D) Two representative examples of OSCC tissues stained with CCL18 and Bmi‐1 antibodies using consecutive tissue sections. The association between CCL18 and Bmi‐1 expression in 45 human OSCC specimens was shown.

It was then examined whether human OSCCs with high expression of CCL18 also showed features of stem‐like cells. A total of 45 OSCC specimens were collected and used for immunohistochemical analysis. High positivity for CCL18 was present in 27/45 OSCC cases and Bmi‐1 stained positive in 29/45 OSCC cases. The clinicopathological association of Bmi‐1 expression in OSCC is shown in Table 1. Bmi‐1 was positively expressed in OSCC tissues and was related to advanced clinical stage. Notably, high expression of CCL18 was significantly associated with Bmi‐1 positivity (P < 0.001; Fig. 2d). Representative immunostaining of consecutive tissue slides from two cases with CCL18 and Bmi‐1 antibodies were shown in Figure 2(d). Taken together, these data provided evidence that CCL18 could contribute to enhance CSC features in OSCC.

Table 1.

Clinicopathological association of Bmi‐1 expression in oral squamous cell carcinoma (OSCC) specimens

Characteristics	No. of cases	Bmi1		P‐value
		Positive	Negative
Age, years
<50	16	10	6	0.840
≥50	29	19	10
Gender
Male	32	20	12	0.669
Female	13	9	4
T – primary tumor
T1 + T2	26	14	12	0.082
T3 + T4	19	15	4
N – regional lymph node
Non‐metastasis	38	17	11	0.502
Metastasis	17	12	5
Histological grade
Well	18	11	7	0.703
Moderately + poorly	27	18	9
Clinical stage
I+II	22	11	11	0.048
III+IV	23	18	5

Forty‐five primary OSCC tissues were obtained from representative areas of each OSCC patient, postoperatively. Medical records for all OSCC cases were reviewed for clinical information. Bmi‐1 expression was defined as positive when a typical nuclear staining was observed. The association between clinicopathology features and Bmi‐1‐positive expression was analyzed using the χ²‐test. Bold text indicates significance.

Chemokine (CC motif) ligand 18‐induced EMT and stemness initiated by Slug overexpression

It is recognized that Snail and Slug (Snail2) are EMT‐associated transcriptional factors.21 To identify whether they were related to overexpressed CCL18, we examined the expressions of Snail and Slug. The results indicated that only Slug protein was upregulated in response to CCL18, but not Snail (Fig. 3a). Analyzed by qRT‐PCR, the mRNA level of Slug was also upregulated by CCL18 (Fig. 3a). To further confirm the role of Slug in CCL18‐induced EMT and stemness of OSCC cells, we used Slug siRNA to knock down endogenous Slug expression. Three independent siRNA sequences (SiSlug1, SiSlug2, and SiSlug3) were designed to target Slug; finally, SiSlug1 (SiSlug) and Si‐NC were selected for subsequent experiments. The knockdown efficiency of the three independent siRNA sequences was presented in the Figure S2.

Figure 3.

Chemokine (CC motif) ligand 18 (CCL18)‐induced epithelial–mesenchymal transition and stemness were initiated by Slug overexpression in oral squamous cell carcinoma cells. (a) Cells were treated with or without CCL18 (20 ng/mL) for 2 or 3 days. Protein expressions of Slug and Snail in HSC6 and CAL33 cells were examined by Western blot. α‐Tubulin was used as an internal control. The mRNA level of Slug was detected by reverse transcription–quantitative real‐time PCR (qRT‐PCR). (b–f) HSC6 and CAL33 cells were treated with CCL18, CCL18+non‐terget control siRNA (Si‐NC), CCL18+SiSlug for 3 days. (b) Expression levels of Slug, E‐cadherin, N‐cadherin, OCT4, and Bmi‐1 were detected by Western blot. α‐Tubulin was used as an internal control. mRNA levels of OCT4 and Bmi‐1 were detected by qRT‐PCR. (c,d) Migration and invasion abilities were detected by Transwell assays. (e) Immunofluorescence staining of E‐cadherin (red) and N‐cadherin (green) was indicated. Nuclear DNA was stained with DAPI (blue). (f) Percentages of aldehyde dehydrogenase (ALDH)^high+ cells and CD133⁺ cells were detected by flow cytometry. Data are presented as the means ± SEM of at least three independent experiments. *P < 0.05 compared to non‐target control; **P < 0.05 compared to CCL18; ***P < 0.05 compared to CCL18+Si‐NC.

In HSC6 and CAL33 cells treated with SiSlug, the protein expressions of Slug, N‐cadherin, OCT4, and Bmi‐1 were decreased and E‐cadherin was increased, despite imposing a continuous 3‐day stimulation of CCL18. By contrast, the CCL18‐stimulated cells treated with Si‐NC showed the same expression levels of these proteins with CCL18‐stimulated cells (Fig. 3b). Moreover, similar results were presented by qRT‐PCR for mRNA levels of OCT4 and Bmi‐1 (Fig. 3b) and by immunofluorescence staining for E‐cadherin and N‐cadherin (Fig. 3c). Subsequently, we found that the migration and invasion ability was weakened (Fig. 3d,e), and the percentages of ALDH^high+ and CD133⁺ cells were reduced (Fig. 3f) in Slug‐knockdown cells, compared to control groups. These results indicated that knockdown of Slug protected OSCC cells from EMT and stem‐like cell transformation.

Chemokine (CC motif) ligand 18 upregulated Slug expression by activating mTOR signaling

A proteomics analysis reported that CCL18‐mediated invasion of ovarian cancer was strongly correlated with the mTORC2 pathway.22 Here, we explored the changes in mTOR signaling in HSC6 and CAL33 cells treated with CCL18. Western blot analyses showed that p‐Akt (Thr308), p‐Akt (Ser473), and p‐mTOR were upregulated, whereas the total Akt and mTOR were unaffected (Fig. 4a). Next, INK128, a potent and selective mTOR ATP binding site competitor (Fig. S3), was used to block mTOR activation in cells before being treated with CCL18. We observed the mRNA and protein expressions of Slug were downregulated (Fig. 4b). These observations suggested that CCL18 activated mTOR signaling and then upregulated Slug expression in OSCC cells.

Figure 4.

Chemokine (CC motif) ligand 18 (CCL18) upregulated Slug expression through protein kinase B (Akt)/mammalian target of rapamycin (mTOR) signaling in oral squamous cell carcinoma cells. (a) HSC6 and CAL33 cells were treated with CCL18 for 15, 30, or 45 min, and then harvested for Western blot to detect phosphorylated (p‐) and total (T‐) AKT/mTOR proteins. α‐Tubulin was used as an internal control. (b) Protein and mRNA expressions of Slug were examined by Western blot and reverse transcription–quantitative real‐time PCR after HSC6 and CAL33 cells were treated with CCL18 and/or INK128 (an inhibitor of mTOR, 100 nM). α‐Tubulin was used as an internal control. Data are presented as the means ± SEM of at least three independent experiments. *P < 0.05 compared to NC; ^** P < 0.05 compared to CCL18.

Inhibition of mTOR pathway reversed CCL18‐induced EMT and stemness

To further investigate the role of the mTOR pathway in CCL18‐induced EMT and stemness in OSCC cells, we treated HSC6 and CAL33 cells with CCL18 in the presence of INK128, and detected that the expression of N‐cadherin, OCT4, and Bmi‐1 was decreased, but the expression of E‐cadherin was increased, compared to that in the absence of INK128 (Fig. 5a). Immunofluorescence staining of E‐cadherin and N‐cadherin displayed the same results as Western blotting (Fig. 5b). Simultaneously, the proportions of ALDH^high+ cells and CD133⁺ cells were downregulated (Fig. 5c), the migration and invasion ability (Fig. 5d,e) was weakened, and the forming spheres of CAL33 cells (Fig. 5f) and SCC15 cells (Fig. S1) were decreased, when the cells underwent CCL18 stimulation in the presence of INK128. All results showed that the mTOR inhibitor efficiently blocked CCL18‐induced EMT and enrichment of stem‐like cells.

Figure 5.

Inhibition of the mammalian target of rapamycin (mTOR) pathway reversed the chemokine (CC motif) ligand 18 (CCL18)‐induced epithelial–mesenchymal transition and stemness in oral squamous cell carcinoma cells. (a–e) HSC6 and CAL33 cells were treated with CCL18 (20 ng/mL) and/or INK128 (100 nM) for 3 days. (A) Expression levels of E‐cadherin, N‐cadherin, OCT4, and Bmi‐1 were detected by Western blot. α‐Tubulin was used as an internal control. (b,c) Migration and invasion abilities were detected by Transwell assays. (d). Immunofluorescence staining of E‐cadherin (red) and N‐cadherin (green) is indicated. Nuclear DNA was stained with DAPI (blue). (e) Percentages of aldehyde dehydrogenase (ALDH)^high+ cells and CD133⁺cells were detected by flow cytometry. (f) CAL33 cells were cultured in low‐adherence conditions to form spheres in 6‐well plates exposed in CCL18 (20 ng/mL) in the absence or presence of INK128. Average numbers of primary and secondary spheres (>100 μm) were calculated for each well (right). All the data are presented as the means ± SEM of at least three independent experiments. *P < 0.05 compared to NC; **P < 0.05 compared to CCL18.

Discussion

It is well established that chemokines in the tumor microenvironment play important roles in cancer progression. Our previous study reported that chemokine ligand CCL18, predominantly produced by cancer epithelial cells, was abundantly expressed in primary OSCC tissues and associated with an advanced clinical stage.13 However, the mechanism of CCL18‐stimulated oncogenesis and development in OSCC remained elusive. In this study, we showed that CCL18 accelerated the abilities of migration, invasion, and sphere formation, and upregulated expression of EMT and CSC‐associated markers in OSCC cell lines. Furthermore, we found a positive correlation between CCL18 and CSC biomarkers in human OSCC tissues. Moreover, we found the mTOR–Slug‐dependent pathway was involved in these CCL18‐induced activities, and silencing Slug or inhibiting mTOR signaling reversed CCL18‐modulated EMT and stemness.

Recently, growing evidence has highlighted that CCL18 was overexpressed and related to tumorigenesis and metastasis in various cancers.23, 24 Several groups have found that elevated CCL18 induced EMT in cancer cells.11, 25, 26, 27 Here, our results showed that the high level of CCL18 accelerated the migration, invasion, and E‐ to N‐cadherin switch of OSCC cells, which implied CCL18 was a vital factor involved in the induction of EMT in OSCC.

Accumulating evidence suggests that tumor cells that have undergone EMT display many similarities to CSCs, which are rare subtypes of cancer cells with unique abilities to self‐renew and differentiate into progenitor cells, and are responsible for drug resistance, metastasis, and recurrence.15, 16 Currently, OCT4, Bmi‐1, ALDH, and CD133 are the common markers used to identify CSCs in OSCC.19, 20 Our data showed that the mRNA and protein expressions of OCT4 and Bmi‐1 were upregulated, as well as the percentages of ALDH^high+ cells and CD133⁺ cells after treatment with CCL18 in vitro. In addition, CSCs are characterized by their tumorsphere‐forming ability in vitro, and tumorsphere formation assays are widely used to identify, isolate, enrich, maintain, or expand potential CSC subpopulations from various types of cancers.28 Interestingly, we found that CAL33 and SCC15 cells, but not HSC6 cells, efficiently formed tumorspheres, which was also strengthened by the high level of CCL18. The reasons might be the heterogeneity of normal stem cells where these CSCs originated, the presence of several different activated signaling pathways, or the different expression patterns of various CSC markers in each OSCC subtype. Other groups also discovered that only a few cancer cell lines could establish stable spheres, such as one‐third of thyroid cancer cell lines29 and one‐quarter of hepatocellular carcinoma cell lines.30 Furthermore, we found high levels of CCL18 and Bmi‐1 positivity were significantly associated in human OSCC specimens. These results provided evidence that CCL18 stimulation enhanced the stem cell‐like characteristics of OSCC cells.

Slug (Snail2) is a member of the Snail family of zinc‐finger transcription factors, which could bind specifically to a subset of E‐box motifs in target promoters, such as the E‐cadherin promoter, and are key mediators of EMT and CSC enrichment in many tumors.31, 32 In this study, we found that CCL18 upregulated Slug mRNA and protein expressions, and Slug knockdown by RNAi blocked EMT and stemness of HSC6 and CAL33 cells, which proved that Slug was required to induce theses phenotypes in OSCC cells.

Mammalian target of rapamycin is a Ser/Thr kinase, and its activity is misaligned in several human diseases, including cancer.33 There is no doubt that the classical pathway of PI3K/Akt/mTOR signaling, one of the most important intracellular pathways, is frequently activated in diverse cancer types. Disorders of the PI3K/Akt/mTOR signaling pathway mediated through molecular aberrations contribute to tumor development and therapeutic resistance, as well as CSC biology.34, 35 In this study, we detected that elevated CCL18 upregulated Slug expression to modulate EMT and the gain of stem cell‐like properties by promoting the phosphorylation of Akt and mTOR, and these effects could be blocked by INK128 (a potent and selective TORC1/2 dual inhibitor) in vitro. Hence, these results highlighted that the CCL18/mTOR/Slug pathway is involved in the regulation of EMT and stemness in OSCC.

The underlying mechanisms of cancer often contain a series of complex aberrations that stimulate critical cellular signaling pathways in oncogenesis. Therefore, CCL18‐related signaling may also relate to more than Akt/mTOR/Slug pathway in cancer. Zhang et al.36 reported that CCL18 binding to Nir1 activated two signaling pathway in breast cancer: (i) the Akt/LIMK/cofilin pathway, which regulates actin polymerization and rearrangement of the cell cytoskeleton; and (ii) the Akt/GSK3β/Snail pathway, which induces EMT. Lin et al.37 suggested that CCL18‐induced endothelial–mesenchymal transformation and pro‐angiogenesis through ERK and Akt/GSK3β/Snail signaling in breast cancer. These reports and our study commonly implied that Akt may be a key node in CCL18 facilitating cancer development. Interestingly, in our study, the CCL18‐related transcription factor that induced EMT in OSCC was Slug, not Snail, as in breast cancer, which reflected the heterogeneity between different tumors. In addition, ERK1/2/NF‐κB, Pyk2/Src, or ELMO1/Dock180 signaling were stimulated by CCL18 and then regulated migration and invasion in some carcinomas.23, 24, 25, 38, 39, 40 Above all, CCL18 might be involved in multiple intracellular signaling participating in cancer development, and further research is required to explore the function and mechanisms of CCL18 in different cancer types.

In conclusion, increased CCL18 activated mTOR signaling to upregulate Slug expression level, and subsequently modulated EMT and stem cell‐like characteristics in OSCC. Thus, our findings not only provide new insight into the role of CCL18 in the development of cancer but also offer new potential therapeutic targets for early diagnosis and treatment of OSCCs.

Disclosure Statement

The authors have no conflict of interest.

Abbreviations

Akt

protein kinase B

ALDH

aldehyde dehydrogenase

CCL18

chemokine (CC motif) ligand 18

CSC

cancer stem(‐like) cell

EMT

epithelial–mesenchymal transition

GSKβ

glycogen synthase kinase 3β

mTOR

mammalian target of rapamycin

OCT4

octamer‐binding transcription factor 4

OSCC

oral squamous cell carcinoma

PI3K

phosphatidylinositol 3‐kinase

qRT‐PCR

reverse transcription–quantitative real‐time PCR

Si‐NC

non‐target control siRNA

Supporting information

Fig. S1. Spheroid formation assays for SCC15.

Fig. S2. Transfection efficiency of the three independent Slug siRNA sequences.

Fig. S3. INK128 inhibited mammalian target of rapamycin (mTOR) activity.

Table S1. Reagents and primary antibodies information.

Cancer Sci 108 (2017) 1584–1593