Fig 2: The ALDH2 levels affected 5-fluorouracil (5-FU) response in HNSC-derived cells. (A) Cancer Cell Line Encyclopedia (CCLE) and Genomics of Drug Sensitivity in Cancer (GDSC) databases demonstrate the ALDH2 levels and the IC50 values for 5-FU in 13 head and neck cancer cell lines. (B) Quantitative RT-PCR measured the endogenous ALDH2 levels in 7 HNSC-derived cell lines. (C) Western blot analysis of HNSC cell lines of DOK-, CAL27- cells (high endogenous ALDH2 level), TW2.6 cell (medium endogenous ALDH2 level), and SAS cell (low endogenous ALDH2 level). (D) MTT assay showed the cell viability of SAS, TW2.6, and DOK cells after treatment with different concentrations of 5-FU. (E) Quantitative RT-PCR (left panel) determined the ALDH2 levels in DOK/shLuc (control) and DOK/shALDH2 cells. The response of DOK cells to 5-FU treatment (right panel) was evaluated across a final concentration of 50 µM for 72 hours.

Fig 3: The gene and protein interacting networks with ALDH2. (A) STRING (left panel) shows the interaction between ALDH2 and ALDH6A1, which was experimentally determined. GeneMANIA (right panel) analysis using the networks including 'physical interactions', 'co-expression', and 'shared protein domains' highlight the genes, including ALDH6A1, with functions enriched in 'oxidorectase activity', 'small molecule binding', and 'catalytic activity'. (B) The correlation between ALDH2 and ALDH6A1 mRNA levels (TIMER 2.0) (C) The ALDH6A1 levels in HNSC tumor and normal tissues (UALCAN). (D) The survival curves of female patients with low ALDH6A1 and male patients with high expression of ALDH6A1 (UALCAN).

Fig 4: The ALDH2 mRNA levels in HNSC patients. (A) TIMER 2.0 displays the differential expression levels of ALDH2 in tumor and adjacent non-tumor tissues across all cancer types in TCGA. The asterisk labels statistical significance. (B) The UALCAN platform indicates that the ALDH2 level in tumors (n = 520) is significantly lower than in normal tissues (n = 44) in the TCGA-HNSC cohort. (C) The ALDH2 levels in tumor tissues across different stages. (D) The ALDH2 levels in HPV+ (n = 41), HPV- (n = 80) tumors, and normal tissues. (E, G) TIMER 2.0 demonstrates the frequencies of HRAS and AJUBA mutations in HPV- and HPV+ HNSCs. (F, H) ALDH2 mRNA levels in HRAS- and AJUBA-mutated HNSCs compared to their wild types. Statistical significance: *P < 0.05, **P < 0.01, ***P <0.001.

Fig 5: The relationships between the ALDH2 level and immune subtype, molecular subtype, and immunostimulatory molecules. The TISIDB database displayed the levels of ALDH2 in the TCGA-HNSC cohort (n = 522) in different (A) immune subtypes and (B) molecular subtypes. (C) A heatmap shows the association between the ALDH2 levels and representative immunostimulators in various cancer types, where an arrowhead indicates the significance of HNSC. (D-F) The top three correlated transcripts were KLRK1, CD27, and TNFRSF14.