In this research investigation, the original counts and corresponding clinical information of RNA sequencing data (Level 3) for 528 HNSCC samples and 44 normal samples were obtained on July 1^st, 2021 from TCGA dataset (https://portal.gdc.cancer.gov/). Also, the expression profiling data of 22 pairs of HNSCC were obtained from the Gene Expression Omnibus (GEO) database (https://www.ncbi.nlm.nih.gov/geo/). The data were analyzed in SPSS (Version 25, IBM Corp., USA), and the results were processed using Graphpad Prism (Version 8, GraphPad Software, USA). In addition, the UALCAN database (http://ualcan.path.uab.edu/) was referenced to investigate the relationships between the expression levels of NTF2 and various clinical features.

In this study, the prognostic values were determined by the Kaplan-Meier curves, and univariate and multivariate Cox regression analysis was performed. The R packages “RMS” and “RMDA” packages were utilized to perform the nomogram, calibration, and decision curve analysis (DCA) based on the results of the multivariate Cox proportional risk analysis. Nomogram and DCA were used to evaluate and compare the predictive models containing the clinical outcomes. All of the above-mentioned analysis methods and R packages were performed using R software version 4.0.3, and P < 0.05 was considered to be statistically significant.

Differentially expressed genes were identified and analyzed using the “limma” R package and wilcoxon tests based on the expression levels of the NTF2 (26, 27). A false discovery rate (FDR) < 0.05 and a |log2(fold change)| >1 were set as the thresholds. Then, Gene Ontology (GO) and the Kyoto Encyclopedia of Genes and Genomes (KEGG) were utilized when using the “cluster profiler” R package. Finally, the “ggplot2” R package was adopted to visualize the experimental results.

The Tumor Immune Estimation Resource (TIMER) database (https://cistrome.shinyapps.io/timer/) was referenced in this study to analyze six subsets of tumor-infiltrating immune cells (28). The immune checkpoint-related genes (SIGLEC15, TIGIT, CD274, HAVCR2, PDCD1, CTLA4, LAG3, and PDCD1LG2) were extracted from TCGA database to explore the tumor immunology. Then, Tumor Immune Dysfunction and Exclusion (TIDE) was used to predict the potential immune checkpoint blocking (ICB) responses (29). The R software packages “ggplot2”, “pheatmap”, and “ggpubr” were used in this research for graph visualization. In the aforementioned analysis processes, a P value of less than 0.05 was considered to be statistically significant.

During this study’s experimental processes, human nasopharyngeal carcinoma cell line 5-8F cells and human pharyngeal squamous carcinoma cell line Fadu cells were cultured in RPMI1640 (iCell, China) and MEM medium (Gibco, USA), respectively. All of the media contained 10% heat-inactivated fetal bovine serum (FBS) and 1% penicillin/streptomycin. Also, all of the cells were incubated at 37°C under the condition of 5% CO₂.

The NTF2 was knocked down in 5-8F and Fadu cell lines via siRNA transfection (Hanbio, China). Then, the 5-8F and Fadu cells were collected for subsequent studies after transfection for 48 hours.

The total RNA was extracted according to the instructions of the Trizol Reagent (Invitrogen, USA), and real-time quantitative PCR (RT-qPCR) was performed using SYBR Green (Vazyme, China). The relative expression levels of the genes were analyzed using the ΔΔCT method and normalized to GAPDH.

The specific primers and siRNA sequences were listed in Table 1.

In order to assess the protein expression levels of the NTF2, a RIPA buffer and protease inhibitors (Gene-Protein Link, China) were used to disrupt the cells. Then, 20 μg of total protein was separated on 15% SDS-PAGE gel. The antibodies specific for NTF2 (66063-1, Proteintech, China) (1:1000) and GAPDH (T004, Affinity, China) (1:2000) were utilized for probing the proteins. Then, specific proteins were visualized using the method provided by the AI600 Imaging System (GE, USA).

In the present study, 5-8F cells were inoculated in 6-well plates containing MEM with 10% FBS at a density of 4 × 10⁵ cells/well. Subsequently, when 90% of the cells had become fused after 16 to 24 hours of culture, the confluence cells were lined with a 200 µL pipette. Then, the injured monolayer cells were washed with phosphate-buffered saline (PBS) for the purpose of removing the cell debris. In the next experimental step, the MEM with 10% FBS was replaced with a serum-free medium. Images were obtained at 0, 6, and 24 hours, and each experiment was independently performed at least three times. The scratch areas were evaluated using Image J, and the cell migration rates were calculated using the following formula:

Paraffin-embedded tissue was selected from 66 HNSCC patients who had undergone surgery in the Affiliated Hospital of Guilin Medical University between April 2012 and October 2019. The patients enrolled in this study were diagnosed with primary squamous cell carcinoma of the larynx, with no other malignancies in the mouth, oropharynx, or pharynx, and no previous history of radiotherapy or chemotherapy. In addition, both clinical and pathological data were collected, including the patients’ ages, differentiation grades, lymph node metastasis, and survival periods. The follow-up and postoperative management data of the patients were collected by telephone or from outpatient medical records. The tumor stages were classified according to the tumor node metastasis (TNM) staging system (2017) of the Union for International Cancer Control (UICC). The samples obtained from the 66 HNSCC patients and ten adjacent normal tissue were each divided into 4 µm thick sections for this study’s immunohistochemical (IHC) analysis process. A primary monoclonal antibody (66063-1, Proteintech, China) was used to detect the expression levels of the NTF2. Secondary antibodies to mouse IgG were obtained from the IHC kit (#CW2069, Beijing Cowin Bioscience Co., Ltd., China). Mouse IgG was used as a negative control to exclude false positive results. The staining intensity was assessed by histology score (H-score) and semi-quantitative analysis was performed by two independent pathologists who had not been informed of the sources of the clinical samples (30, 31). The following formula was applied:

The studies involving human participants were reviewed and approved by the Ethics Committee of the Affiliated Hospital of Guilin Medical University. All of the patients/participants provided their written informed consent to participate in this study.

Two-tailed student t-tests were performed in SPSS to evaluate the statistical significance in this study. The significance differences of P < 0.05, 0.01, and 0.001 were symbolized as *, **, and ***, respectively.

The mRNA expressions of NTF2 in human cancer cells were analyzed using the UALCAN database. It was found that when compared with the corresponding normal tissue, higher expressions of NTF2 were observed in the majority of the cancer types, including HNSCC (P < 0.001); bladder urothelial carcinoma (BLCA); breast invasive carcinoma (BRCA); cholangiocarcinoma (CHOL); colon adenocarcinoma (COAD); esophageal carcinoma (ESCA), and so on (Figures 1A, B). In addition, significant increases in the NTF2 expressions in HNSCC cases were observed in 44 cases of tumor tissue with paired adjacent normal tissue (P < 0.001) (Figure 1C). Furthermore, similar results were observed in 22 pairs of HNSCC samples from the GSE6631 cohort in the GEO database (P < 0.001) (Figure 1D). Therefore, the findings suggested that NTF2 may play a vital regulatory role in the development and progression of HNSCC.

This study investigated the NTF2 expression levels in various HNSCC subgroups using the UALCAN database. It was observed that the NTF2 expression levels were higher in the HPV negative group than in the positive group (P < 0.001; Figure 2D). In addition, the NTF2 expression levels were significantly up-regulated in both the men and women tumor groups, respectively (P < 0.001) (Figure 2A). The same results were observed for the different age groups (P < 0.001; Figure 2B); pathological grade groups (P < 0.001; Figure 2C); HPV infection groups (P < 0.001; Figure 2D); tumor stage groups (P < 0.001; Figure 2E); and lymph node metastasis groups (P < 0.001; Figure 2F) among the HNSCC cases. However, there were no significant differences observed among the clinical subgroups.

Then, the prognostic values of the NTF2 were determined. Following a median expression level, the patients were divided into the following two groups: High expression group (n = 264) and low-expression group (n = 264), as detailed in Figure 3A. The patients in the high expression group were observed to have remarkably higher mortality rates than those in the low-expression group (Figure 3B). In addition, the Kaplan-Meier survival curves also showed that the survival rates of the high expression patients were significantly lower than those of the low-expression patients (P = 0.000714; Figure 3C).

In the present investigation, both univariate and multivariate Cox proportional risk analyses were performed. The results revealed that the NTF2 expression levels, patient ages, and TNM stages were independent prognostic factors (Figures 3D, E), which were included to establish an accurate prediction model. This study’s nomogram provided a graphical representation of the aforementioned factors, and the prognostic risks for an individual patient could be calculated by the points associated with each risk factor, as detailed in Figure 3F. In addition, as shown in Figure 3G, the calibration plots showed excellent agreement between the actual probabilities and the estimated probabilities at 1, 3, and 5 years.

An immunohistochemical staining method was used to detect the expression levels of NTF2 in the tumor samples from 66 HNSCC patients and 10 normal tissue samples. The results revealed that the NTF2 was highly expressed in the HNSCC tissue when compared with the normal tissue (Figures 4A, B). Meanwhile, there was no significant correlation observed between the NTF2 expression levels and the patient ages, genders, pathological grades, tumor stages, lymph node metastasis, or smoking habits (Table 2). The median follow-up timeframe for all of the examined patient cases was 36.2 months (ranging from 1.0 to 99.9 months). At the final follow-up times, it was determined that in 49 cases (74.2%), the patients had survived, and in 17 cases (25.8%) the patients had died. The Kaplan-Meier analysis results showed that the high expression levels of NTF2 were closely related to significant reductions in overall survival (P = 0.0066) in the HNSCC case samples (Figure 4C).

In the present study, the co-expressed genes related to NTF2 were identified by mining data from TCGA database. This study’s volcano map and heat map with positive and negative correlations with NTF2 in HNSCC were shown in Figures 5A, B, respectively. A total of 119 genes associated with NTF2 (P < 0.05) were used in the GO and KEGG enrichment analyses in order to explore relevant biological functions and pathways. The top 30 critical terms for the enrichment analysis of the biological processes (BP), cellular components (CC), and molecular functions (MF) were detailed in Figures 5C–E. The first 10 KEGG pathways of the related genes were shown in Figure 5F.

The associations between the NTF2 expression levels and the infiltrating immune cells were analyzed. The results showed that NTF2 expression levels were negatively correlated with the B cells (PSpearman = -0.35; P < 0.001); CD4⁺ T cells (PSpearman = -0.25; P < 0.001); CD8⁺ T cells (PSpearman = -0.17; P < 0.001); neutrophils (PSpearman = -0.17; P < 0.001); macrophages (PSpearman = -0.11; P = 0.0017); and dendritic cells (PSpearman = -0.16; P < 0.001) (Figure 6A). The immune checkpoint-related genes (SIGLEC15, TIGIT, CD274, HAVCR2, PDCD1, CTLA4, LAG3, and PDCD1LG2) were extracted and analyzed. It was found that the CD274, CTLA4, LAG3, PDCD1, and TIGIT genes displayed negative correlations with the NTF2 expressions (Figure 6B). The potential ICB responses indicated that the NTF2 high expression group had a poor efficacy for immune checkpoint blockade treatments (Figure 6C).

In order to investigate the role of NTF2 in HNSCC cells, two siRNAs targeting NTF2 (si1, si2) were transfected into 5-8F and Fadu cells respectively. It was found that when compared with the control cells treated with empty vector, the NTF2 was significantly silenced at the mRNA (Figure 7A) and protein levels in the knockdown group (Figure 7B). This study then tested the effects of cell proliferation in-vitro. Using the CCK-8 assay, it was found that the knockdown of the NTF2 could inhibit HNSCC cell proliferation after 48 hours (P < 0.001) and 72 hours (P < 0.001) of culturing (Figures 7C, D). In addition, the wound-healing assay was used to assess the capacity of cancer cell migration. The results revealed that the cells treated with NTF2 siRNAs showed lower migration rates than the control cells after 6 hours (P < 0.001) and 24 hours (P < 0.001) of culturing (Figures 7E, F).