The Oncomine database (17,18) (www.oncomine.org) was used to search for the fold changes in hnRNP K expression levels in different types of tumor using the following screening conditions: i) Gene name, hnRNP K; ii) analysis type, cancer vs. normal; and iii) data type, mRNA. The associations between hnRNP K expression levels and different clinicopathological parameters of HNSCC were analyzed using the UALCAN online database (http://ualcan.path.uab.edu/index.html). In the UALCAN database, hnRNP K and HNSCC were used as the search criteria, and the different clinical pathological parameters, such as sample type, sex, age, ethnicity, tumor grade, individual cancer stages, nodal metastasis status and HPV status, were selected and the corresponding figures were downloaded. In the tumor-immune system interactions database (TISIDB; http://cis.hku.hk/TISIDB), HNRNPK was used as the gene symbol, with the search terms ‘clinical’ and ‘head and neck squamous cell carcinoma’ as the cancer type to generate survival curve.

FBS, DMEM and minimum essential medium (MEM) were purchased from Gibco; Thermo Fisher Scientific, Inc. The anti-hnRNP K primary antibody (cat. no. ab52600) was purchased from Abcam, while anti-β-Catenin (cat. no. 8480), anti-disheveled (Dvl)2 (cat. no. 3224), anti-c-Jun (cat. no. 9165), anti-Met (cat. no. 8198), anti-Cyclin-D1 (cat. no. 2978), anti-c-Myc (cat. no. 5605), anti-matrix metalloproteinase (MMP)7 (cat. no. 3801) and anti-β-actin (cat. no. 3700) primary antibodies were purchased from Cell Signaling Technology, Inc. Moreover, anti-rabbit IgG, HRP-linked antibody (cat. no. 7074) and anti-mouse IgG, HRP-linked antibody (cat. no. 7076) were purchased from Cell Signaling Technology, Inc. To confirm the outer β-catenin gene was expressed and the plasmid transfection was successful, flag expression level was analyzed using anti-flag (cat. no. M185-3L; Medical & Biological Laboratories).

CAL-27 (human tongue squamous cell carcinoma) and WI-38 (human embryo lung fibroblast) cell lines were cultured in DMEM, while the FaDu (human pharyngeal squamous cell carcinoma) cell line was cultured in MEM. All cells were purchased from National Infrastructure of Cell Line Resource (http://www.cellresource.cn/), and the cells were supplemented with 10% FBS and 100 U/ml penicillin and streptomycin, and maintained in a humidified atmosphere with 5% CO₂ at 37°C. The short hairpin RNA (shRNA/sh) sequence targeting hnRNP K (shhnRNP K; 5′-GTGCTGATATTGAAACAAT-3′) (GV248-hnRNP K) and the negative control (NC) lentivirus expressing green fluorescence protein (shNC; 5′-TTCTCCGAACGTGTCACGT-3′) (GV248-NC) were provided by Shanghai GeneChem Co., Ltd. The shhnRNP K and shNC (MOI=20) were transfected into CAL-27 cells (5×10⁵ cells per well) using polybrene according to the manufacturer's protocol. The duration of transfection was 12 h at 37°C followed by changing the fresh medium. Transfected cells were used for subsequent experiments after 72 h.

As the frozen shRNA after resuscitation was not sufficient to perform experiments, siRNA was used to verify the results of microarray results in order to save time. Small interfering RNA (siRNA/si) targeting hnRNP K (sihnRNP K; 5′-GAGCUUCGAUCAAAAUUGATT-3′) and a non-specific siNC siRNA (5′-UUCUCCGAACGUGUCACGUTT-3′) were purchased from Guangzhou RiboBio Co., Ltd. The target sequences in the two groups were identical to those previously described (12). The overexpression plasmid of β-Catenin (GV362-β-Catenin) and the plasmid-NC (empty vector, GV362-NC) were provided by Shanghai GeneChem Co., Ltd. The transfections were performed with Lipofectamine 2000, and cells with sihnRNP K and siNC (30 pmol) were harvested following 48 h of transfection at 37°C for further experimentation.

Total protein was extracted from CAL-27, FaDu and WI-38 cells using cold RIPA lysis buffer (Sigma-Aldrich; Merck KGaA). Total protein was quantified using a bicinchoninic acid protein assay kit (Pierce; Thermo Fisher Scientific, Inc.), the protein samples were boiled with 2X SDS-PAGE protein loading buffer (Beyotime Institute of Biotechnology) for 15 min and then the cell lysates (20 µg per lane) were separated via 10% SDS-PAGE. The separated proteins were subsequently transferred onto nitrocellulose filter membranes (Pall Corporation) and blocked with 5% non-fat dry milk in Tris-buffered saline with 0.1% Tween-20 at room temperature for 1 h. The membranes were then incubated with the following primary antibodies overnight at 4°C: Anti-hnRNP K primary antibody (1:10,000, cat. no. 8480; Abcam), anti-β-Catenin (1:1,000, cat. no. 8480; Cell Signaling Technology), anti-Dvl2 (1:1,000, cat. no. 3224; Cell Signaling Technology), anti-c-Jun (1:1,000, cat. no. 9165; Cell Signaling Technology), anti-Met (1:1,000, cat. no. 8198; Cell Signaling Technology), anti-Cyclin-D1 (1:1,000, cat. no. 2978; Cell Signaling Technology), anti-c-Myc (1:1,000, cat. no. 5605; Cell Signaling Technology), anti-MMP7 (1:1,000, cat. no. 3801; Cell Signaling Technology), anti-Flag (1:1,000, M185-3L; Medical & Biological Laboratories) and anti-β-actin (1:1,000, cat. no. 3700; Cell Signaling Technology). Following the primary antibody incubation, the membranes were incubated with anti-rabbit IgG, HRP-linked antibody (1:5,000; cat. no. 7074) or anti-mouse IgG, HRP-linked antibody (1:5,000; cat. no. 7076) which was purchased from Cell Signaling Technology, Inc., for 1 h at room temperature. An enhanced chemiluminescence kit (cat. no. 32106; Thermo Fisher Scientific, Inc.) was used to visualize the membrane. β-actin was used as the loading control. Densitometric analysis was performed using ImageJ software (version 1.8.0; National Institutes of Health).

A total of 20 paired tumor and adjacent normal tissues (17 males and 3 females; median age, 63; age range, 44–84 years) were collected from patients with HNSCC at Beijing Tongren Hospital (Beijing, China) between March 2018 and March 2019. The patients had no history of radiotherapy or chemotherapy prior to surgery. Written, informed consent was obtained from all patients and the use of patient specimens for the experiments was approved by the Ethics Committee of Beijing Tongren Hospital, Capital Medical University (Beijing, China).

The patient tissue were fixed with 10% formalin at room temperature for 24 h. Briefly, 5-µm sections of paraffin-embedded tissue were deparaffinized in xylene and rehydrated in a descending series of ethanol at room temperature. The deparaffinized sections were incubated with 3% H₂O₂ at room temperature for 15 min to inhibit the endogenous peroxidase activity. The sections were then boiled in antigen retrieval solution for 8 min and blocked with 10% goat serum (cat. no. ZLI-9017; Beijing Zhongshan Golden Bridge Biotechnology Co., Ltd.) for 30 min at 37°C. Subsequently, the tissue sections were incubated with an anti-hnRNP K monoclonal antibody (1:250) overnight at 4°C, followed by incubation with anti-rabbit IgG, HRP-linked secondary antibody (1:5,000; cat. no. 7074) which was purchased from Cell Signaling Technology, Inc., for 1 h at room temperature. Fresh 3,3′-diaminobenzidine (DAB substrate solution:DAB concentrate, 20:1) was added to the sections for a color reaction at room temperature, and then, the samples were counterstained with hematoxylin for 10 sec at room temperature. Finally, the sections were dehydrated with an ascending series of ethanol and cleared with xylene at room temperature, sealed with a coverslip and visualized under a Zeiss Axio Imager Z2 Upright light microscope (magnification, ×20; Zeiss AG).

CAL-27 cells were seeded into 96-well plates at a density of 1,000 cells/well with three replicates per condition. The plates were incubated at 37°C in a humidified incubator with 5% CO₂ for 24, 48, 72, 96 or 120 h, and the cell number was checked by cell counter every 24 h.

CAL-27 cells were seeded into 96-well plates at a density of 800 cells/well with three replicates per condition. The plates were incubated at 37°C in a humidified incubator with 5% CO₂ for 24, 48, 72, 96 or 120 h and the cell viability was analyzed every 24 h. Briefly, every 24 h, 100 µl serum-free DMEM containing 10 µl CCK-8 reagent (Dojindo Molecular Technologies, Inc.) was added to each well and incubated for 2 h according to the manufacturer's protocol. The absorbance was measured for 450 nm using microplate reader at room temperature. Data were analyzed from ≥3 independent experiments.

CAL-27 cells were seeded into 6-well culture plates (Corning Inc.) at a density of 1,000 cells/well, with each condition set up in triplicate wells. The cells were cultured for 10 days at 37°C constant temperature CO₂ incubator to induce colony formation (>50 cells per colony). Following the incubation, the cells were fixed with 4% paraformaldehyde at room temperature for 30 min, and then stained with 1% crystal violet at room temperature for 30 min. The colonies were counted and images were captured using a Zeiss Axio Imager Z2 light microscope.

The EdU incorporation assay was performed with transfected CAL-27 cells using the Cell Light EdU Apollo^® 567 In Vitro Imaging kit (Guangzhou RiboBio Co., Ltd.). For each group, 1×10⁴ cells/well were seeded into 24-well plates and incubated overnight at 37°C constant temperature CO₂ incubator. The EdU assay was performed according to the manufacturer's protocol; however, the cell nuclei were stained with DAPI instead of Hoechst 33342 for 30 min at room temperature, and the fixative used was 4% paraformaldehyde for 30 min at room temperature. Five randomly selected fields of view of EdU-positive cells were visualized using Leica DMi8 fluorescence microscope (magnification, ×20). Image analysis was performed using ImageJ software (version 1.8.0; National Institutes of Health). Data were analyzed from ≥3 independent experiments.

A total of 1×10⁶ CAL-27 cells/well were plated into 6-well plates and cultured with complete medium in an incubator at 37°C overnight until 100% confluence. Subsequently, a linear scratch wound was generated in the cell monolayer using a 200-µl pipette tip. After washing with PBS 3 times, the cells were cultured with serum-free DMEM at 37°C with 5% CO₂. At 0, 12 and 24 h after the scratch was made, images were captured using a Zeiss Axio Imager Z2 light microscope (magnification, ×10). The wound area (%) was analyzed using ImageJ software (National Institute of Health).

CAL-27 cells were transfected with shhnRNP K and shNC as described above. A total of 5×10⁴ cells/well were resuspended in 200 µl DMEM without FBS and seeded into the upper compartments of Boyden chambers (Falcon; Corning Inc.) The lower chamber was filled with 700 µl DMEM containing 10% FBS. Following 12 h of incubation at 37°C, the non-migratory cells remaining in the upper chamber were removed with cotton swabs and the migratory cells in the lower chamber were washed with PBS three times, fixed with 4% paraformaldehyde at room temperature for 30 min and stained with crystal purple for 15 min at room temperature. The stained cells were visualized in five randomly selected fields using a Zeiss Axio Imager Z2 light microscope (magnification, ×100).

Animal experiments were approved by the Institutional Animal Care and Use Committee of the Institute of Laboratory Animal Sciences, Peking Union Medical College (approval no. GR-16002; Beijing, China). The nude mice were housed in pathogen-free environment with 12-h light/dark cycle, controlled humidity (50%) and temperature (26°C) and had free access to food and water.

Equal numbers of shhnRNP K-and shNC-transfected cells (5×10⁶ cells) which were resuspended with PBS were subcutaneously injected into the axilla of the left forelimb of 12 male BALB/c nude mice (six mice/group; age: 6-weeks-old; weight: 21 g; HFK Bioscience, http://www.hfkbio.com/). The tumor volume (mm³) was determined every 2 days from the 5th day after injection to the study endpoint using the following formula: Tumor volume = (a × b²)/2, where ‘a’ was the longest tumor diameter and ‘b’ was the shortest tumor diameter. The maximum tumor volume of the study was 603.7 mm³. All animals were sacrificed 4 weeks later by exposure to CO₂ (25 l/min) at a chamber displacement rate of 30% and subsequent cervical dislocation, and the tumors were excised. Death was verified by complete cessation of breathing, the lack of a heartbeat and dilated pupils.

Total RNA was extracted from cultured shhnRNP K and shNC cells using TRIzol^® reagent (Invitrogen; Thermo Fisher Scientific, Inc.), according to the manufacturer's protocol. Gene expression profiles were analyzed by Shanghai OE Biotechnology Corporation (https://www.oebiotech.com/). The hybridized signals were assessed with Agilent SurePrint G3 Human gene expression v3 microarrays (Agilent Technologies, Inc.) for the downregulated genes. A fold change threshold of ≤-2 and P≤0.05 were considered to indicate a statistical significance. The differentially expressed genes were then subjected to functional term enrichment analysis using the Gene Ontology (GO) database (19) and signaling pathway enrichment analysis using the Kyoto Encyclopedia of Genes and Genomes (KEGG) database (20) using GeneSpring GX software (version 14.9; Agilent Technologies, Inc.).

RT-qPCR was performed as previously described (12). The reverse transcription protocol was: 42°C for 60 min and 70°C for 5 min. The protocol of PCR was as follows: 95°C, 2 min; 95°C, 10 sec; 60°C, 30 sec for 40 cycles in total. The primers used for the qPCR are provided in Table I and the expression levels were quantified using the 2^−ΔΔCq method (21).

Statistical analyses were performed using GraphPad Prism software (version 5.0; GraphPad Software, Inc.) and data are presented as the mean ± SD from three independent experiments. Group comparisons between two groups were performed using an unpaired two-tailed Student's t-test. Multiple group comparisons were performed using a one-way ANOVA followed by a Bonferroni's post hoc test. The overall survival analysis for low and high expression levels of hnRNP K was performed using the Kaplan Meier method and a log-rank test was used to determine the P-value. P<0.05 was considered to indicate a statistically significant difference.

To obtain an initial insight into hnRNP K expression patterns, data mining using the Oncomine and UALCAN databases was performed and the results are presented in Figs. 1A and S1. Compared with the normal tissues, the mRNA expression levels of hnRNP K were upregulated in brain and CNS cancer, cervical cancer, melanoma, myeloma and head and neck cancer (Fig. 1A), and the expression levels of hnRNP K in primary tumor of HNSCC was more upregulated than normal head and neck tissues. Furthermore, hnRNP K expression levels were positively associated with the grade, human papillomavirus status and lymph node metastasis of patients with HNSCC (Fig. S1E, G and H), while no significant difference were recorded between the expression levels of hnRNP K and the sex, age, ethnicity or tumor stage of the patients with HNSCC (Fig. S1B-D and F). In addition, data mining in the TISIDB revealed that high expression levels of hnRNP K were associated with the significantly poorer survival of patients with HNSCC compared with patients with low expression levels (Fig. 1B). Furthermore, the expression levels of hnRNP K were determined in two HNSCC cell lines (CAL-27 and FaDu) and human embryonic lung cells (WI-38) using RT-qPCR and western blotting analysis. The results indicated that hnRNP K expression levels were upregulated in CAL-27 and FaDu cells compared with the WI-38 cells (Fig. 1C and D). Next, 20 pairs of paraffin-embedded cancerous and adjacent normal tissue samples from patients with HNSCC were analyzed using immunohistochemistry. It was revealed that hnRNP K was mainly expressed in the nuclei of tumor tissue compared with the adjacent normal tissue, the expression levels of hnRNP K in the cancer specimens were markedly increased (Fig. 1E). Collectively, these data indicated that hnRNP K may have an important role in the development of HNSCC.

To investigate the function of hnRNP K in HNSCC cells, shRNA was used to establish a stable hnRNP K-knockdown CAL-27 cell line. Both hnRNP K mRNA and protein expression levels in the shhnRNP K-transfected cells were downregulated compared with the shNC-transfected cells (Fig. 2A and B). The results of the CCK-8 and absolute cell count assays indicated that the knockdown of hnRNP K significantly inhibited CAL-27 cell viability and proliferation compared with the shNC-transfected cells following 120 h of incubation (Fig. 2C and D). Furthermore, the results of the colony formation assay revealed that the hnRNP K-knockdown CAL-27 cell line formed significantly fewer and smaller clones compared with the clones formed in the shNC group (Fig. 2E and F). The EdU incorporation assay also demonstrated that hnRNP K silencing resulted in a markedly reduced percentage of CAL-27 cells compared with the shNC group (Fig. 2G). In addition, the hnRNP K knockdown significantly inhibited the migratory ability of CAL-27 cells compared with the shNC group (Fig. 2H-K). Taken together, the present results indicated that hnRNP K may promote the cell viability, proliferation and migration of CAL-27 cells.

To verify the oncogenic role of hnRNP K in CAL-27 cells in vivo, shhnRNP K- or shNC-transfected CAL-27 cells were inoculated into BALB/c nude mice. Tumors derived from the shhnRNP K cells exhibited slower growth compared with the tumors derived from the shNC group (Fig. 3A-C). The differences in the tumor volume and weight between the two groups were significant, which was shown as the tumor weight and volume significantly decreased in the shhnRNP K group compared with the shNC group (Fig. 3A-C), while there were no significant differences recorded in the body weight of the mice between the two groups (Fig. 3D). These results indicated that hnRNP K may promote the tumor formation by CAL-27 in vivo.

To determine the molecular mechanisms underlying the phenotypic changes in hnRNP K-knockdown HNSCC cells, mRNA microarray analysis was performed. GO functional term enrichment analysis of the significantly differentially expressed genes of shhnRNP K knockdown cells compared with the NC was used to identify the role of differentially expressed hnRNP K. The results identified that ‘Cell adhesion’ was the most highly enriched biological process of the downregulated genes, which suggested that hnRNP K may be an important modulator of HNSCC metastasis, while receptor agonist activity was the most highly enriched molecular function and extracellular region was the most highly enriched cellular component (Fig. 4A). Furthermore, the downregulated genes were also relevant to the ‘Response to virus’ and ‘Defense response to virus’, suggesting that hnRNP K may be involved in the immune response (Fig.4A).

KEGG signaling pathway enrichment analysis of the significantly differentially expressed genes was used to elucidate the pathways and molecular interactions of hnRNP K in HNSCC. The top 30 pathways associated with the downregulated mRNAs are listed in Fig. 4B. The ‘Wnt signaling pathway’ was the top pathway enriched by the downregulated genes, which implied that the Wnt signaling pathway may be a target for hnRNP K (Fig. 4B). Next, CAL-27 cells were transfected with sihnRNP K and siNC, and RT-qPCR (Fig. 4C) and western blotting analysis (Fig. 4D and E) revealed that hnRNP K, β-catenin, Dvl2, c-Jun, Met, Cyclin-D1, c-Myc and MMP7 expression levels were significantly downregulated in the sihnRNP K-transfected cells compared with the siNC-transfected cells. To determine if the β-Catenin pathway was involved in hnRNP K regulating phenotype of HNSCC, β-Catenin was overexpressed with plasmid (Fig. 4F). Following the simultaneous transfection with si-hnRNP K and a β-Catenin overexpression plasmid, the expression levels of β-Catenin, c-Myc and c-Jun were significantly restored compared with the cells transfected with the si-hnRNP K only (Fig. 4G and H).

These results indicated that the Wnt/β-Catenin signaling pathway may be involved in the hnRNP K-driven inhibition of proliferation and migration in HNSCC.