The primary sphere K3 cell line was used. This cell was isolated from a surgical specimen from a HNSCC patient, and the CSC properties of the cell line have been validated using a number of functional assays, testing self-renewal capability, stem cell marker expression, chemoresistance and in vivo tumorigenicity, as previously reported (5). Cells were cultured in serum-free Dulbecco's modified Eagle's medium (DMEM)-Ham's F-12 (F12) medium supplemented with human recombinant basic fibroblast growth factor (bFGF; 10 ng/ml; R&D Systems, Minneapolis, MN, USA), N2 supplement (1 ml per 500 ml medium; Gibco; Thermo Fisher Scientific, Inc., Waltham, MA, USA), and epidermal growth factor (EGF; 10 ng/ml; R&D Systems). The primary β-catenin antibody was purchased from Santa Cruz Biotechnology, Inc. (Dallas, TX, USA), and the secondary anti-mouse IgG antibody was obtained from Jackson ImmunoResearch Laboratories (West Grove, PA, USA).

To assess self-renewal in vitro, the cells were dissociated into a single-cell suspension, seeded in a 24-well plate at a density of 500 cells per well, and cultured in serum-free medium, with EGF and bFGF added every other day. Untreated cells (cells not treated with EGF or bFGF) were used as a control. Tumorspheres were allowed to grow for 7 days, and spheres with a diameter >30 µm were counted.

RT-qPCR and western blotting experiments were performed as previously described (5). RT-qPCR analysis of Wnt signaling target genes (β-catenin, c-myc and cyclin D) was subsequently performed on an iCycler IQ real-time detection system (Bio-Rad Laboratories, Hercules, CA, USA), using IQ Supermix with SYBR-Green (Bio-Rad Laboratories). The sequences of the human-specific primers used were as follows: Oct4 forward, 5′-GCAATTTGCCAAGCTCCTGAA-3′ and reverse, 5′-GCAGATGGTCGTTTGGCTGA-3′; Sox2 forward, 5′-CCTCCGGGACATGATCAG-3′ and reverse, 5′-TTCTCCCCCCTCCAGTTC-3′; Twist forward, 5′-CAGTCTTACGAGGAGCTGCA-3′ and reverse, 5′-TCTTGCTCAGCTTGTCCGAG-3′; Snail forward, 5′-CACCTCCAGACCCACTCAGAT-3′ and reverse, 5′-CCTGAGTGGGGTGGGAGCTTC-3′; Slug forward, 5′-CTCCATGCCTGTCATACCAC-3′ and reverse, 5′-GGAAAGAGGAGAGAGGCCAT-3′; β-catenin forward, 5′-GGCAGTGCGTTTAGCTGGTG-3′ and reverse, 5′-AGCTTGGGGTCCACCACTAG-3′; c-myc forward, 5′-GTATGTGGACGGCTTCTCGC-3′ and reverse, 5′-GCTGCTGTCGTTGAGAGGGT-3′; and cyclin D forward, 5′-GGAGCTGCTGCAAATGGAGC-3′ and reverse, 5′-CAGAGGGCAACGAAGGTCTG-3′.

ALDH activity was analyzed using an Aldefluor assay kit (Stem Cell Technologies, Vancouver, BC, Canada). Briefly, cells were suspended in Aldefluor assay buffer containing the ALDH substrate (1 µmol/l per 1×106 cells). A sample of cells was stained with the specific ALDH inhibitor diethylaminobenzaldehyde (DEAB) as a negative control. Fluorescent ALDH+ cells were detected in the green fluorescence channel, and samples treated with the specific ALDH inhibitor DEAB were used as the control to set the gates defining the ALDH+ region. Flow cytometric sorting was conducted using a FACSAria flow cytometer (Becton Dickinson, San Jose, CA, USA).

Cells were dissociated into a single-cell suspension and then plated in a 24-well plate at a density of 1×105 cells per well under serum-free culture conditions. The cells were treated with cisplatin at the indicated concentrations and then cultured at 37°C in a humidified 5% CO2 atmosphere. Subsequently, 24 h later, 50 µl of MTT solution (5 mg/ml in PBS) was added to each well, and the plate was incubated at room temperature for 2 h. Absorbance was measured on a SpectraMax 190 device (Molecular Devices, LLC, Sunnyvale, CA, USA) at a wavelength of 570 nm.

Cells were seeded in 24-well plates at a density of 1×105 cells per well. Cells were transiently transfected using Polyexpress^® (Excellgen, Rockville, MD, USA). For the luciferase assay, cells were transfected with 0.5 µg of the lymphoid enhancer factor-reporter pTOPFLASH/pFOPFLASH along with 50 ng of pRL-SV40. TGFβ1 treatment was initiated at 24 h post-transfection and performed for 48 h. Luciferase activity was monitored using the Dual-Glo luciferase assay system (Promega Corporation, Madison, WI, USA). The transfection efficiency was normalized to the cotransfected Renilla luciferase activity according to the manufacturer's protocol.

Experimental data were statistically assessed using either 2-tailed Student's t-test or analysis of variance with Scheffe's post-hoc analysis. P<0.05 was considered to indicate a statistically significant difference.

The capacity for self-renewal is one of the putative traits of CSCs (5). Thus, it was evaluated whether TGFβ1 treatment enhances the self-renewal capacity of HNSCC CSCs by performing a sphere-forming assay. TGFβ1 (2 ng/ml) treatment significantly increased the sphere formation ability of HNSCC CSCs with a diameter of >30 µm compared to the control treatment group (Fig. 1).

Oct4 and Sox2 are critical regulators of pluripotency in the mammalian embryo, and deregulated expression of these genes can be found in HNSCC CSCs (13). Thus, whether TGFβ1 treatment increases the transcriptional expression of Oct4 and Sox2 was examined. The results showed that mRNA expression levels of Oct4 and Sox2 were significantly increased in the TGFβ1 treatment group compared to the control treatment group (Fig. 2A). ALDH is considered to be a marker of HNSCC CSCs (14). Therefore, ALDH activity was examined in TGFβ1- and control-treated HNSCC CSCs. Fig. 2B shows that the proportion of ALDH+ cells was significantly increased in the TGFβ1 treatment group compared with the control treatment group (7.24 vs. 19.28%).

HNSCC CSCs possess a chemoresistance ability against anticancer drugs, and the ABCG2 transporter is responsible for this resistance phenotype (5). Therefore, chemoresistance levels were compared using MTT assay subsequent to cisplatin administration at various concentrations with the TGFβ1 and control treatment groups. The results showed that TGFβ1-treated HNSCC CSCs had significantly increased cisplatin resistance compared with control-treated HNSCC CSCs (Fig. 3A). Furthermore, an increased ABCG2 expression level was identified in TGFβ1-treated HNSCC CSCs compared to control-treated HNSCC CSCs, indicating that TGFβ1 may enhance the chemoresistance of HNSCC CSCs through increased expression of ABCG2 (Fig. 3B).

EMT is a key process in tumor invasion and metastasis (15). Previously, it was reported that the EMT process can generate cancer cells with a stem cell phenotype (15). Thus, whether TGFβ1 can increase central regulators of the EMT process, such as Twist, Snail and Slug, was investigated. As shown in Fig. 4, mRNA expression levels of all three EMT regulators were increased in TGFβ1-treated HNSCC CSCs compared with control-treated HNSCC CSCs.

It was previously suggested that Wnt/β-catenin contributes to the stemness of HNSCC cells (16). Thus, whether TGFβ1 acts as an upstream stimulator of Wnt/β-catenin signaling was investigated. Administration of TGFβ1 increased the expression of β-catenin, which is an effector of the Wnt pathway (Fig. 5A). Also, TGFβ1 treatment increased the activity of a Wnt/β-catenin-dependent reporter as a functionally more relevant indicator (Fig. 5B). Furthermore, the transcript levels of Wnt/β-catenin signaling target genes, such as the c-myc and cyclin D1 genes, were increased in TGFβ1-treated HNSCC CSCs (Fig. 5C). Overall, these data suggest that TGFβ1-induced stemness may be associated with Wnt pathway upregulation in HNSCC CSCs.