1.1 DFNA5 expression is upregulated in human HNSCC
The analysis of microarray data from HNSCC and adjacent normal tissues in the Oncomine database showed that DFNA5 expression was upregulated in HNSCC (Figure 1A). This pattern was confirmed by the analysis of gene expression data deposited in TIMER (Figure 1B) and the transcriptome sequencing data from TCGA analyzed with GEPIA (Figure 1C, D). In addition to bulk analysis of HNSCC data, we also used Oncomine to compare DFNA5 expression in healthy and cancer tissues from datasets corresponding to different HNSCC subtypes. These results showed that DFNA5 upregulation is a general characteristic of HNSCC (Figure 1E–I). Further subgroup analysis of various clinicopathological features of HNSCC patients from the TCGA database revealed that in subgroup analyses based on gender, age, race, disease stage, tumor grade, human papillomavirus infection, and nodal metastasis status, the transcription level of DFNA5 in HNSCC patients was consistently significantly higher than that in healthy subjects (Figure 2).
Furthermore, we used HNSCC cell lines for in vitro validation, demonstrating a significant increase in the level of DFNA5 compared to that of control cells (Figure 2I). Collectively, these results suggest that the expression of DFNA5 may be a potential diagnostic marker of HNSCC.
1.2 High DFNA5 expression predicts poor prognosis in HNSCC patients
We speculated that the expression of DFNA5 is related to the prognosis of HNSCC patients. To test this hypothesis, we evaluated the prognostic value of DFNA5 using the Kaplan-Meier plotter, TISIDB, and UALCAN (Figure 3A–C, respectively). All three analyses showed that high DFNA5 expression levels are associated with a poor prognosis in HNSCC patients. Moreover, this association was consistent regardless of tumor grade (Figure 3D), race (Figure 3E), or gender (Figure 3F).
To further understand this correlation and assess the potential underlying mechanism, we utilized the Kaplan-Meier plotter to explore the relationship between the DFNA5 expression level and specific clinical characteristics of HNSCC patients. As shown in Table 1, high expression of DFNA5 was significantly correlated with worse overall survival (OS) only in male patients (P = 0.041). Moreover, racial background did not have a statistically significant effect. With respect to tumor stage, DFNA5 overexpression was associated with poorer OS in HNSCC patients at all stages except for stage 3. More importantly, we found that high expression of DFNA5 corresponded with a worse OS in patients whose tumors were infiltrated by immune cells such as basophils, B-cells, T-cells, and macrophages (Table 1). These results suggest that DFNA5 may be a reliable biomarker for HNSCC prognosis, at least in male patients.
1.3 DFNA5 regulates cell adhesion in HNSCC
Lists of the top 200 genes co-expressed with DFNA5 were created by analyzing TGCA datasets with cBioPortal (Figure 4A) and UALCAN (Figure 4B). Comparison of the two lists revealed 125 common co-expressed genes (Figure 4C). DAVID was then used to perform functional annotation of these 125 genes. GO analysis indicated that these genes were mainly involved in biological processes of locomotion, cell adhesion, and cell migration (Figure 4D), which was consistent with the enrichment in the respective cellular components and the proposed molecular functions (Figure 4E-F). In addition, KEGG pathway analysis showed enrichment in pathways of focal adhesion and the interaction between the actin cytoskeleton and the extracellular matrix receptor (Figure 4G). Collectively, these data suggest an essential role of DFNA5 in regulating cell adhesion in HNSCC.
1.4 Identification and analysis of DFNA5-related hub genes
CytoHubba, a Cytoscape plug-in, was used to identify potential hub genes for DFNA5 function based on the density of maximum neighborhood component. The top 10 genes in the network included ITGB1, ITGA3, ITGB4, ITGA6, PXN, ITGA5, LAMC2, LAMA3, PLEC, and LAMB3 (Figure 5A). GEPIA was then used to analyze the expression level of each hub gene and its correlation with OS. All genes except for PLEC had higher expression levels in HNSCCs than in normal tissues. Moreover, high expression levels of ITGB1, LAMA3, PLEC, and LAMB3 correlated with poor OS in HNSCC patients (Figure 5B–K). These results further support that DFNA5 mainly affects tumor progression in HNSCC by regulating cell adhesion through processes such as integrin-related molecular pathways.
1.5 High DFNA5 expression results in decreased lymphocyte infiltration in HNSCC
To further explore the function of DFNA5 in HNSCC and its prognostic potential, we focused on the infiltration of immune cells in the HNSCC tumor microenvironment. Specifically, we used TIMER to investigate whether the expression of DFNA5 is related to the level of immune cell infiltration. Figure 6A shows that the DFNA5 expression level was significantly negatively correlated with the infiltration of lymphocytes in HNSCC, such as B cells (cor = –0.223, P = 8.57e-07) and CD8 T cells (cor = –0.223, P = 2.99e-07). Furthermore, we analyzed the correlation between DFNA5 expression levels and various immune cell markers in HNSCC, including subsets of T cells, B cells, M1 and M2 macrophages, neutrophils, natural killer cells, and dendritic cells. The results showed that the DFNA5 expression level was significantly negatively correlated with B cell and CD8 T cell markers (Figure 6B). We conclude that a high level of DFNA5 expression significantly correlates with decreased tumor local lymphocyte infiltration, which is an important negative prognostic factor in HNSCC patients.