Oncomine is an online data-mining platform which contains various microarray expression data. The data of Tomm34 gene expression in HNSCC was extracted from Oncomine database. The search conditions were as follows: 1) Gene: Tomm34; 2) Analysis Type: Cancer vs. Normal Analysis; 3) Cancer Type: Head and Neck Squamous Cell Carcinoma; 4) Analysis parameters were p-value < 1E-4, fold change >2, gene rank = 10%.

UALCAN is a comprehensive, user-friendly, and interactive web resource for analyzing cancer OMICS data [22]. Transcripts per million (TPM) values were used to measure the expression level of Tomm34, categorized into high expression group (TPM values are above the upper quartile) and low/medium expression group (TPM values are below the upper quartile) [22]. We compared the expression of Tomm34 in normal tissues and HNSCC, and analyzed the correlation of Tomm34 with clinicopathological variables such as tumor stage, tumor classification, race and weight and other indicators. Kaplan-Meier curves of Tomm34 expression and survival time were also evaluated.

80 archival formalin-fixed paraffin embedded (FFPE) primary OSCC biopsies were recruited from the tissue bank of Department of Oral Pathology, Hospital of Stomatology, Guanghua School of Stomatology, Sun Yat-sen University, Guangzhou, China for immunohistochemical (IHC) staining study. All of the specimens were histologically evaluated as squamous cell carcinoma which originated from oral cavity and oropharynx including tongue, buccal mucosa, gingiva, floor of mouth, hard palate, oropharynx and soft palate. Clinicopathological informations including gender, age, tumor size, lymph node metastasis, pathology differentiation, TNM status, disease-free survival (DFS) and overall survival (OS) were analyzed. The study protocol was approved by Clinical Research Ethics Committee of Hospital of Stomatology, Guanghua School of Stomatology, Sun Yat-sen University.

FFPE OSCC specimens were cut into 4 μm thick sections, de-paraffinized in xylene, and rehydrated in serial dilutions of ethanol. Endogenous peroxidase activities were subsequently blocked by a 10-min treatment with 3% H₂O₂ solution. Antigen retrieval was performed in EDTA (pH 9.0) for 2 min at 10.0 MPa. Following incubation in 5% BSA for 10 min to block endogenous biotin activity, sections were incubated with the primary antibody (Tomm34, 1:100, Proteintech Inc., United States; p16, 1:200, Abcam, United States) at 4°C overnight. The HRP-labeled secondary antibody was applied to the sections and incubated at 37°C for 40 min. For color development, sections were stained 3,3′-diamino-benzidine (DAB) substrate and hematoxylin. Human colon carcinoma tissue was utilized as the positive control, and samples which were incubated with PBS instead of primary antibody served as the negative controls.

The extent and intensity of staining were evaluated and scored by two independent pathologists. Twenty microscopic high-power fields of tumor tissues were randomly selected for assessment. A numeric intensity score of 0–4 was assigned to each case based on the percentage of positive neoplasm cells (0, negative, <10% positive cells staining; 1+, 10–25% positive cells with weak staining; 2+, 25–50% positive cells with moderate staining; 3+, 51–70% positive cells with strong staining and 4+, >70% positive cells with strong staining). Staining was dichotomized into low expression and high expression by score. Score 0–2 was considered as low expression of Tomm34, and score 3–4 was considered as high expression of Tomm34. Meanwhile, >70% positive cells with strong and diffuse nuclei staining (with or without cytoplasmic staining) (score 4) was considered as positive expression of p16 [9].

Data was analyzed by IBM SPSS statistical software V22 (IBM Corporation, United States). The χ2 test was utilized to assess the associations between the expression of Tomm34 or p16 and clinical characteristics of patients. Survival estimates were analyzed by Kaplan-Meier method with significance determined by the log rank test. A p value less than 0.05 (two-sided) was considered significant.

A total of 449 cases of different tumors of Tomm34 gene research results were collected in the Oncomine database. 30 research results showed statistical differences, where 22 of 30 cases reported overexpression of Tomm34 in tumor tissues and 8 cases reported lower expression in tumor tissues (Figure 1A).

Tomm34 gene ranked in 1,045 position among the total expressed genes of HNSCC(p = 1.38 E −10) (Figure 1B). Moreover, Tomm34 expression was significantly upregulated in the HNSCC tissues paired normal tissues (Figure 1C).

For further study, the expression of Tomm34 in HNSCC was investigated in subgroups via UALCAN analysis. Tomm34 expression varied in HNSCC with different grades (Figure 1D). The Kaplan-Meier curves showed that higher level of Tomm34 correlated with poorer overall survival (OS) (Figure 1E).

Tomm34 even staining was seen in the cytoplasm of tumor cells, nuclear expression was rarely seen (Figures 2A,B). Weak positive expression was seen in keratinocytes and some tumor stroma. In the cohort of 80 OSCC, 48 cases fulfilled the Tomm34 positive judgment criteria, and the overall positive rate was 60%.

The relationship between Tomm34 staining and clinicopathological factors were shown in Table 1. The expression of Tomm34 was closely related with the TNM clinical stage of OSCC (p < 0.001), tumor size (p < 0.001), lymph node metastasis (p = 0.001), drinking history (p = 0.019), but not with pathology differentiation, age, gender, smoking history and site (p > 0.05).

To investigate the effects of Tomm34 on patient survival in more detail, we next assessed the relationship between the Tomm34 expression level and patient overall survival by Kaplan-Meier survival analysis. The overall survival was extremely down-regulated in the Tomm34-positive expression group compared with the Tomm34-negative expression group. The Kaplan-Meier curves indicated that higher level of Tomm34 correlated with poorer overall survival (OS, p = 0.114) and disease-free survival (DFS, p = 0.044) (Figures 2C,D).

To separate the HPV-related cases in OSCC, p16 immunohistochemistry was utilized in the cohort. The expression of p16 was observed in 33.75% of the primary OSCC tumors (27/80), distinguished by strong staining in >70% (scored 4) of tumor cells. p16 was homogeneously detected in the nuclei and cytoplasm of tumor cells (Figure 3A).

The relationship between the p16 expression and clinicopathological factors was shown in Table 2. The expression of p16 was strongly associated with earlier TNM classification by WHO (p = 0.046) and lymph node metastasis (p = 0.003), but not with tumor site, age, gender, tumor size, pathology differentiation, cigarette and drinking history.

The relationship between the p16 expression level and patient overall survival was further assessed by Kaplan-Meier survival analysis. The overall survival rate was markedly down-regulated in the p16-negative group compared with the p16-positive group. The Kaplan-Meier curves indicated that patients with p16-positive staining exhibited better OS (p = 0.015) and DFS (p = 0.008) (Figures 3B,C).

HPV-related status of this OSCC cohort has been described by p16 staining. To further analyze the relationship of Tomm34 and HPV status, the expression of Tomm34 was reassessed in HPV-positive OSCC (n = 27) and HPV-negative OSCC (n = 53), respectively. As shown in Table 3, the higher expression of Tomm34 was consistently correlated with larger tumor size (p = 0.01 in HPV-positive OSCC and p = 0.004 in HPV-negative OSCC) and higher TNM classification (p = 0.008 in HPV-positive OSCC and p < 0.001 in HPV-negative OSCC). However, drinking history was only related with Tomm34-postive expression group in HPV-positive OSCC (p = 0.044), while positive node category was only related with Tomm34-postive expression in HPV-negative OSCC (p = 0.001).

Furthermore, survival outcomes of four subgroups based on the expression of Tomm34 and p16 were examined (Figures 4A,B). Detailed pairwise comparisons of groups were shown in Table 4 (OS) and Table 5 (DFS). In HPV-positive OSCC subgroups, the survival curves of Tomm34 (+) p16 (+) group and Tomm34 (−) p16 (+) group exhibited rather little difference (OS, p = 0.824; DFS, p = 0.782). While in HPV-negative OSCC, statistically decreased OS (p = 0.046) and DFS (p = 0.020) rates were observed between Tomm34 (+) p16 (−) group and Tomm34 (−) p16 (−) group. In addition, Tomm34 (+) p16 (−) group also showed dramatically decreased survival rates compared with Tomm34 (+) p16 (+) group (OS, p = 0.014; DFS, p = 0.008) and Tomm34 (−) p16 (+) group (OS, p = 0.050; DFS, p = 0.015).