Human HNSC samples were gained from patients undergoing surgery at Qilu Hospital of Shandong University, China. The patients were randomly selected and no one received any chemotherapy before surgery. A total of 23 HNSC-HSCC tissues and paired adjacent tissues (15 pairs without lymph node metastasis and 8 pairs with lymph node metastasis) were gained. In lymph node metastasis group, 7/8 patients had alcohol habit and 6/8 patients with tobacco smoking history. In no lymph node metastasis group, 13/15 patients with alcoholism and 12/15 with smoking habit. Patients’ detailed clinical information were shown in Supplementary Material 1 . All patients signed informed consent and our research was approved by the Institutional Ethics Committee at Qilu Hospital.

The differential expression of CCDC60 mRNA among TCGA tumors was analyzed via the TIMER2.0 database (13). P-value cutoff < 0.05 and the statistical significance was examined by Wilcoxon test. CCDC60 expression in HNSC samples was detected by TCGA database (n = 546) via limma package in R ( Supplementary Material 2 ). The pROC R package was employed to plot the ROC curve of CCDC60 gene (14), and the predictive utility of CCDC60 for HNSC diagnosis was evaluated by the area under the curve (AUC). Values of AUC > 0.7 and P < 0.05 suggested that the gene had high predictive ability. The expression of CCDC60 in HNSC from different clinical characteristics was detected by the UALCAN database, including individual cancer stage, lymph node metastasis status, tumor grade, TP53 status, patient’s gender and age (15).

Apply TRIzol reagent (Sigma-Aldrich) to gain total RNAs from HNSC tissues, then reversely transcribed to cDNA with the reverse transcription kit (TaKaRa). qRT-PCR was conducted by SYBR qPCR Master Mix (Thermo Fisher). The 2^−ΔΔCt method was utilized to quantify the relative mRNA expression, and ß-actin was selected as the inner control. The sequences of CCDC60 were as follows: CCTCTTCCGCCAGCTCTGTG (sense) and CACCCGGGTCCTTTGGGTTC (antisense). The sequences of ß-actin were as follows: CACCATTGGCAATGAGCGGTTC (sense) and AGGTCTTTGCGGATGTCCACGT (antisense).

The genetic mutation of CCDC60 and alteration-related prognosis in HNSC samples was detected via cBioPortal (16). Kaplan-Meier (KM) plotter was applied to determine the relation between CCDC60 expression and clinical prognosis of HNSC patients with distinct clinicopathological characteristics, including overall survival (OS) and relapse-free survival (RFS) (17). The OS and disease-specific survival (DSS) maps of CCDC60 gene in HNSC samples were obtained from the DriverDBv3 database (18).

By the Spearman correlation test, we identified CCDC60 co-expression genes in HNSC samples from TCGA dataset (n = 520) via the LinkedOmics database (19). KEGG pathway and GO enrichment analysis of these genes in HNSC samples are available at Gene Set Enrichment Analysis (GSEA). We used the GeneMANIA database to construct the protein-protein interaction (PPI) networks of CCDC60 related co-expression genes in HNSC samples (20).

Human FaDu and Cal27 cell lines were preserved in our lab. The cells were cultured in DMEM/MEM medium (BasalMedia) containing 10% fetal bovine serum (FBS) in a humidified atmosphere containing 5% CO₂ at 37°C.

pENTER-CCDC60 overexpression plasmid was purchased from Vigenebio, and was transfected into Fadu and Cal27 Cells using Lipofectamine 3000 (Thermo Fisher), according to the manufacturer’s recommendations. 48 hours After transfection, cells were harvested or passaged for subsequent experiments.

Transiently transfected cells were maintained in culture medium for 14 days at 37°C. The colonies with more than 20 cells were counted. Fixed with 4% paraformaldehyde and stained with 0.1% crystal violet. Visible colonies were manually imaged and counted.

Approximately 5×10³ transfection cells were planted into 96-well plates and cultured for 24, 48, and 72 hours, respectively. After adding10 μL CCK-8 (Dojindo), the cells were incubated for 2 hours at 37°C in the dark. Finally, the value of optical density (OD) at 450 nm was measured by microplate reader.

Approximately 1.5×10⁵ transfection cells were planted into 6-well plates and incubated for 48 hours. The cells were fully digested, harvested and washed twice with ice-cold phosphate buffer saline (PBS). Next, the cells were resuspended in binding buffer, and stained with 5 µL of fluorescein isothiocyanate (FITC) annexin V and propidium iodide (PI) for 15 minutes at 37°C in the dark. The results of apoptosis were analyzed by CytEpert v2.0 (Beckman Coulter).

For cell cycle analysis, the cell cycle detection kit (Bestbio) was selected. After transfection, Fadu and Cal27 cells were digested by trypsin, washed twice with PBS, fixed with 75% ice-cold ethanol and stored at -20°C for 1 hours. The cells were resuspended and stained with RNaseA reagent (50 U/ml) and propidium iodide (PI, 50 µg/mL) at 4°C for 30 minutes. Finally, cell cycle was determined by a flow cytometer (Beckman Coulter).

After transfection, approximately 3×10⁵ cells were seeded into six-well plates. After 24 hours, scratch with a 10-ml micropipette tip in the center of the wells. Washed twice by PBS, the cells were incubated in 1% FBS medium. Scratch images were taken every 4 hours to compute the scratch healing area.

Transwell chambers were utilized to examine the invasion ability of Fadu and Cal27 cells. Approximately 1×10⁵ cells were planted into each well. The upper chambers were contained in 200 µl serum-free medium, while the lower chambers in the medium containing 20% FBS. After 24 hours incubation, fixed the cells and stained with 0.1% crystal violet (Solarbio). Finally, the cells were photographed under a microscope, and stained cells were counted.

Cell adhesion assay was conducted using the cell adhesion kit (BestBio). Added 100 μL coating buffer into 96 well plates and incubated at 4°C overnight. Wash each well with washing solution, then approximately 5×10⁴ transfection cells were planted and incubated for 1.5 hours at 37°C. Washed with medium and added 100 μL fresh medium. Finally, the cells were stained with 10 μL staining solution B and incubated for 2 hours at 37°C, and the OD value at 450 nm was measured using a microplate reader (BioTek, Winooski, VT, USA).

Applying TIMER2.0 database, we examined the connection between CCDC60 expression and the tumor-infiltrating lymphocytes (TILs) in HNSC, including CD4/CD8+ T cells, B cells, Tregs, Neutrophils, Macrophages, Monocytes, NK cells, Dendritic cells, and Mast cells. Next, the relation of CCDC60 expression with immune marker sets of immune cells was detected. These biomarkers of immune cells have been studied before (21–23). The analysis was based on the Spearman’s method and gene expression levels were displayed with log2 RSEM. Additionally, TISIDB database was developed to investigate the relationships of CCDC60 expression with immunostimulators, immunoinhibitors, chemokines, and receptors in HNSC (24).

R software (version 4.1.2) and GraphPad Prism7 were applied for data visualization and statistical analysis. The presence and strength of relationships between variables were determined by Spearman and Pearson correlation tests. The KM method was used to evaluate survival analyses. A ROC curve was plotted to examine the predictive performance of the CCDC60 risk score for HNSC diagnosis. The difference was statistically significant at p < 0.05.

We realized that the expression level of CCDC60 mRNA was significantly downregulated in most types of cancers via the TIMER2.0 database (P < 0.05, Figure 1A ). Compared with the adjacent normal tissues, CCDC60 expression was decreased in HNSC tissues from the TCGA database (n = 546), which was similar to the results of TIMER2.0 database (P < 0.001, Figure 1B ). Analysis of ROC curve suggested that CCDC60 gene had a good predictive performance for HNSC diagnosis in the TCGA database (AUC = 0.807, Figure 1C ). qRT-PCR results indicated that in 15 pairs of HNSC samples without lymph node metastasis, CCDC60 mRNA expression was lower than that in adjacent healthy tissues ( Figure 1D ). Similarly, in the 8 pairs of lymph node metastasis group, the consistent results were achieved ( Figure 1E ).

Based on the analysis of cancer stage, lymph node metastasis status, tumor grade, patients’ gender, age and TP53 status, CCDC60 expression was significantly downregulated in different subgroup of HNSC patients (P < 0.05, Figures 1F–K ). Interestingly, in the subgroup of lymph node metastasis, CCDC60 expression was significantly lower in the N2 group than N0 group (P < 0.01, Figure 1G ). In the tumor grade subgroup, we noticed the increasing expression of CCDC60 with increasing tumor grade in HNSC patients ( Figure 1H ). CCDC60 expression in male was significantly higher than female, and its expression was obviously different in various age subgroups (P < 0.05, Figures 1l , J ). Additionally, CCDC60 expression was decreased in the TP53 mutant group compared with the nonmutant (P < 0.01, Figure 1K ). We concluded that CCDC60 was significantly decreased in HNSC tissues and could be a promising indicator for HNSC identification and diagnosis.

Analysis of genetic alteration showed that the alteration frequency of CCDC60 was 1.80% in 523 HNSC samples, including amplification (0.6%) and mutation (1.2%, Figure 2A ). As shown in Figure 2B , the diploid was the most common copy number variations in HNSC patients. Between the altered and unaltered groups, difference in OS probability was not significant, suggesting that overall alteration of CCDC60 was not the cause of worse outcome led by low CCDC60 expression (P = 0.115, Figure 2C ). Compared with the low group, high CCDC60 expression was significantly connected with longer OS (HR = 0.645, P = 0.0357) and DSS (HR = 0.541, P = 0.0281) of HNSC patients ( Figures 2D, E ). KM plotter demonstrated that CCDC60 expression was related to better OS (HR = 0.600, P = 0.002); however not with RFS (HR = 1.610, P = 0.210) ( Figures 2F, G ).

Based on the clinicopathological factors, we evaluated the relation between CCDC60 expression and clinical prognosis in HNSC. As shown in Table 1 , we noticed that increased CCDC60 mRNA expression was linked with OS in male (HR = 0.560, P = 0.0024), white race (HR = 0.570, P = 0.0011) and tumor grade 3 (HR = 0.430, P = 0.0058). Specifically, upregulated CCDC60 expression was related to longer OS and RFS in stage 2 (OS, HR = 2.430, P = 0.0355; RFS, HR = 0.190, P = 0.0393), stage 3 (OS, HR = 0.460, P = 0.0416) and stage 4 (OS, HR = 0.480, P = 0.001) of HNSC patients. The above analysis indicated that CCDC60 expression was associated with a favorable outcome in HNSC, highlighting the function of CCDC60 in predicting the prognosis of HNSC, and the prognostic value of CCDC60 for HNSC patients was determined by its distinct clinical characteristics.

We have identified the CCDC60 related co-expression genes using the LinkedOmics database [FDR(BH) < 0.001, Supplementary Material 3 ], among which 5520 genes were positively linked with CCDC60 and 1688 genes were negatively correlated ( Figure 3A ). The PPI network elaborated the potential connections between CCDC60 related co-expression genes ( Figure 3B ). Figures 3C , D indicated the top 50 genes positively or negatively correlated with CCDC60 in HNSC. GO analysis suggested that these genes were mainly involved in the peptide cross-linking, peptidase complex, protease binding and translation factor activity, RNA binding ( Figures 3E–G ). KEGG pathway indicated enrichment in the NOD-like receptor signaling pathway, associated with chemical carcinogenesis ( Figure 3H ). The above analysis revealed the potential function of CCDC60 in HNSC, which CCDC60 might act as a tumor suppressor that inhibited the HNSC progression through sphingolipid signaling pathway.

pENTER-CCDC60 overexpressed plasmid was successfully transfected ( Figure S1 ). In colony formation assay, we realized that compared with the control group, both Fadu and Cal27 cells overexpressing CCDC60 formed fewer clones (P < 0.05, Figure 4A ). Consistently, the proliferation abilities of both cell types were significantly attenuated after pENTER-CCDC60 plasmid transfection at 24, 48 and 72 hours in the CCK-8 assay (P < 0.05, Figure 4B ). Since dysfunction of chromosome formation results in mitosis failure and apoptosis, to identify the mechanism by which CCDC60 affects the tumorigenesis of HNSC, we detected the influences of CCDC60 on apoptosis and cell cycle progression. By the flow cytometry, we found that there was no significant difference in the rate of apoptosis between these groups (P > 0.05, Figure 4C ). CCDC60 overexpression increased the proportion of cell cycle arrest in the G0-G1 phases, while the number of cells in the S-phase was obviously decreased ( Figure 4D ). These results showed that CCDC60 overexpression could inhibit the proliferation, result in G0-G1 cell cycle transition arrest in HNSC.

The findings of wound-healing and transwell assays showed that compared to that of control, overexpression of CCDC60 significantly inhibited the migration of Fadu and Cal27 cells (P < 0.05, Figure 5A ), and the number of invading pENTER-CCDC60 transfected cells was markedly decreased (P < 0.01, Figure 5B ). In cell adhesion assay, we observed the number of adhesion cells treated with pENTER-CCDC60 plasmid was higher than that in the control group (P < 0.05, Figure 5C ). The meaningful results suggested that CCDC60 could markedly reduce the migration and invasive activities, while promote the capacity to cell adhesion in HNSC.

Tumor-infiltrating lymphocytes (TILs), as an integral part of the tumor microenvironment, was correlated with the clinical outcomes and therapy responses in cancers. The above findings supported a prognostic role of CCDC60 in HNSC, but its function within the TME remained unknown. As definitely presented in Figure 6 , CCDC60 expression was significantly positive linked with the infiltration levels of B cells (Rho = 0.419, P = 2.37e-22), CD4+ T cells (Rho = 0.234, P = 1.49e-07) and Tregs (Rho = 0.203, P = 5.66e-06). Furthermore, CCDC60 expression was negatively related to the infiltration levels of CD8+ T cells (Rho = -0.363, P = 8.59e-17), Macrophages (Rho = -0.169, P = 1.68e-04), Monocytes (Rho = -0.205, P = 4.45e-06), DCs (Rho = -0.283, P = 1.62e-10) and NK cells (Rho = -0.233, P = 1.79e-07). These analyses suggested that CCDC60 had a close connection with the immune infiltrating cells in HNSC, and further studies of their interactions and potential immune-related mechanisms are needed.

Subsequently, our study explored the association based on the immune marker sets of TILs. Our analysis revealed that CCDC60 was closely correlated with the majority of gene markers of immune cells in HNSC ( Table 2 ). We noticed that the expression levels of marker set of M1 macrophages (NOS2 and IRF5), TAMs (CCL2) had strong associations with CCDC60 expression in HNSC ( Figures 7A , C ). CCDC60 expression was significantly linked to T cell exhaustion markers (PDCD1, PDCDLG2 and CTLA4) in HNSC ( Figure 7B ), suggesting that CCDC60 may have an impact on the immune escape in HNSC. For Treg cells, CCDC60 showed significant correlation with FOXP3, CCR8, STAT5B and TGFB1 in HNSC ( Figure 7D ). High CCDC60 expression was significantly connected with the infiltration level of DCs in HNSC, DC markers such as HLA-DPB1, HLA-DRA, HLA-DPA1, CD1C and ITGAX were also significantly correlated with CCDC60 expression ( Figure 7E ). The above findings strongly suggested that CCDC60 played a key function in the regulation of immune infiltration in HNSC.

Our findings showed that CCDC60 expression was closely connected with immunoinhibitors (P < 0.05, Figure 8A ), CCDC60 was positively related to ADORA2A (Rho = 0.170. P = 1.00e-04), LGALS9 (Rho = 0.200, P = 4.39e-06), PDCD1 (Rho = 0.122, P = 5.18e-03), BTLA (Rho = 0.173, P = 7.58e-05), VTCN1 (Rho = 0.311, P = 4.72e-13), CD96 (Rho = 0.150, P = 5.98e-04) and CD244 (Rho = 0.102, P =2.03e-02), while negatively correlated with TGFB1 (Rho = -0.298, P = 4.71e-12), TGFBR1 (Rho = -0.207, P = 1.87e-06), PDCD1LG2 (Rho = -0.233, P = 8.16e-08), IL-10 (Rho = -0.095, P = 3.08e-02) and CD274 (Rho = -0.104, P = 1.79e-02). Additionally, the expression of CCDC60 was significantly connected with immunostimulators (P < 0.001, Figure 8B ). CCDC60 was positively correlated with CD27 (Rho = 0.227, P = 1.70e-07), CD40LG (Rho = 0.211, P = 1.27e-06), LTA (Rho = 0.149, P = 6.40e-04), KLRK1 (Rho = 0.178, P = 4.39e-05), TMIGD2 (Rho = 0.156, P = 3.47e-04), TNFRSF13B (Rho = 0.276, P = 1.70e-10), TNFRSF13C (Rho = 0.322, P = 6.30e-14), TNFRSF14 (Rho = 0.206, P = 2.09e-06), TNFRSF17 (Rho = 0.298, P = 4.40e-12) and TNFRSF18 (Rho = 0.191, P = 1.11e-05), while CD276 (Rho = -0.251, P = 6.62e-09), PVR (Rho = -0.298, P = 2.00e-11) and NT5E (Rho = -0.336, P = 4.33e-15) were negatively correlated. The above results reminded us that CCDC60 had a regulatory effect on the immune interactions and mediated tumor immune escape in HNSC, which could improve the response to immunotherapy for HNSC patients.

Subsequently, we investigated the relationship between CCDC60 and chemokines in HNSC. We demonstrated that CCDC60 expression was positively linked with CCL14 (Rho = 0.196, P = 6.98e-06), CCL19 (Rho = 0.269, P = 5.13e-10), CCL20 (Rho = 0.158, P = 2.99e-04), CCL28 (Rho = 0.277, P = 1.33e-10), CX3CL1 (Rho = 0.275, P = 2.06e-10), CXCL17 (Rho = 0.377, P < 2.20e-16) and XCL2 (Rho = 0.172, P = 7.94e-05), and negatively with CCL3 (Rho = -0.158, P = 3.07e-04), CCL7 (Rho = -0.170, P = 9.49e-05), CCL11 (Rho = -0.157, P = 0.32e-03), CCL13 (Rho = -0.199, P = 5.08e-06), CCL27 (Rho = -0.152, P = 5.04e-04), CXCL5 (Rho = -0.183, P = 2.81e-05) and CXCL11 (Rho = -0.158, P = 3.08e-04) (P < 0.001, Figure 9A ). We also found that the expression of CCDC60 was strongly related to chemokine receptors (P < 0.05, Figure 9B ). Except CXCR1 (Rho = -0.132, P = 2.44e-03), the expression of other chemokine receptors was positively correlated with CCDC60, including CCR2 (Rho = 0.115, P = 8.35e-03), CCR4 (Rho = 0.094, P = 3.17e-02), CCR6 (Rho = 0.192, P = 1.04e-05), CCR7 (Rho = 0.168, P = 1.15e-04), CCR10 (Rho = 0.107, P = 1.42e-02), CX3CR1 (Rho = 0.180, P = 3.56e-05), CXCR3 (Rho = 0.109, P = 1.29e-02), CXCR4 (Rho = 0.132, P = 2.53e-03), CXCR5 (Rho = 0.238, P = 3.84e-08), CXCR6 (Rho = 0.104, P = 1.73e-02). This study revealed the close relationships between CCDC60 expression and chemokines as well as chemokine receptors and further proved that CCDC60 may be a promising immunomodulatory factor in HNSC.