Molecular patterns and mechanisms of tumorigenesis in HPV-associated and HPV-independent sinonasal squamous cell carcinoma

Abstract

Mechanisms of tumorigenesis in sinonasal squamous cell carcinoma (SNSCC) remain poorly understood due to its rarity. A subset of SNSCC is associated with human papillomavirus (HPV), but it is unclear whether HPV drives tumorigenesis or acts as a neutral bystander. Here, we show that HPV-associated SNSCC shares mutational patterns found in HPV-associated cervical and head and neck squamous cell carcinoma, including lack of TP53 mutations, hotspot mutations in PI3K and FGFR3, enrichment of APOBEC mutagenesis, viral integration at known hotspots, and frequent epigenetic regulator alterations. We identify HPV-associated SNSCC-specific recurrent mutations in KMT2C, UBXN11, AP3S1, MT-ND4, and MT-ND5, with KMT2D and FGFR3 mutations correlating with reduced overall survival. We establish an HPV-associated SNSCC cell line, showing that combinatorial small-molecule inhibition of YAP/TAZ and PI3K synergistically suppresses clonogenicity. Combining YAP/TAZ blockade with vertical PI3K inhibition may benefit HPV-associated SNSCC, whereas targeting MYC and horizontal inhibition of RAS/PI3K may suit HPV-independent SNSCC.

Fig. 1. High-throughput sequencing of HPV-independent SNSCC reveals mutational patterns in COSMIC genes.

A Whole-exome sequencing was performed in HPV-independent SNSCC with matched normal DNA (n = 7) and somatic variants were assessed. Mutations in COSMIC genes are represented. B Whole-exome sequencing was performed in HPV-independent SNSCC without matched normal DNA (n = 12) and somatic variants were assessed using a panel of normal genomes. Mutations in COSMIC genes are represented. Source data are provided as a Source data file.

Fig. 2. High-throughput sequencing of HPV-associated SNSCC reveals mutational patterns in COSMIC genes.

A Whole-genome sequencing was performed in HPV-associated SNSCC with matched normal DNA (n = 12) and somatic variants were assessed. Mutations in COSMIC genes are represented. B Whole-genome sequencing was performed in HPV-associated SNSCC without matched normal DNA (n = 25) and somatic variants assessed using a panel of normal genomes. Mutations in COSMIC genes are represented. Source data are provided as a Source data file.

Fig. 3. APOBEC pathway signature 13 is enriched in HPV-associated SNSCC.

A APOBEC signature 13 in the HPV-independent SNSCC compared to HPV-associated SNSCC cohort with matched normal DNA. Statistical significance was determined using an unpaired two-tailed t-test and a Mann-Whitney U test (P = 0.0171). The center line of the boxplot represents the median, the box spans the interquartile range (IQR, 25th to 75th percentile), and the whiskers extend to the minimum and maximum values within 1.5 × IQR. Outliers beyond this range are shown as individual points. Median values: HPV-independent = 0.001932 (n = 7), HPV-associated = 0.3492 (n = 12). B APOBEC signature 13 in the HPV-associated SNSCC cohort. C Stacked bar chart showing the relative contribution (%) of mutational signatures in individual HPV-associated and HPV-independent SNSCC samples. No significant differences were observed for signatures #1, 5, 7 and 29. Source data are provided as a Source data file.

Fig. 4. Clinical associations with overall survival and mutation status.

A HPV-independent SNSCC with TP53 mutations demonstrate an association with significantly decreased overall survival (P = 0.0093, Log-rank test: χ² = 6.757, df = 1; TP53 normal (n = 13), TP53 mutant (n = 6)). B HPV-associated SNSCC with KMT2D mutations demonstrate an association with significantly decreased overall survival (P = 0.0377; Log-rank test: χ² = 4.316, df = 1; KMT2D normal (n = 26), KMT2D mutant (n = 11)). C HPV-associated SNSCC with FGFR3 mutations demonstrate an association with decreased overall survival (P = 0.0516; Log-rank test: χ² = 3.789, df=1; FGFR3 normal (n = 31), FGFR3 mutant (n = 6)). D HPV-associated CESC from TCGA with KMT2D mutations demonstrate an association with decreased overall survival (P = 0.0105; Log-rank test: χ² = 6.54, df = 1; KMT2D normal (n = 123), KMT2D mutant (n = 17)). E HPV-associated HNSCC from TCGA with KMT2D mutations and overall survival (P = 0.4117; Log-rank test: χ² = 0.6738, df = 1; KMT2D normal (n = 67), KMT2D mutant (n = 11)). Survival curves were compared using a two-tailed Log-rank (Mantel–Cox) test. Source data are provided as a Source data file.

Fig. 5. Viral integration site analysis for HPV-associated SNSCC.

Viral integration sites for HPV-associated SNSCC are displayed for each chromosomal location. At each integration site the top row indicates the histology type while the bottom row indicates the HPV serotype. Source data are provided as a Source data file.

Fig. 6. Pathway alterations in sinonasal squamous cell carcinoma.

Incorporation of mutation and copy number alteration analysis reveals targetable pathways in HPV-associated SNSCC (left) and HPV-independent SNSCC (right). Source data are provided as a Source data file. Created in BioRender. Team, S. (2025) https://BioRender.com/r25r431.

Fig. 7. Combinatorial blockade of PI3K and YAP/TAZ synergistically inhibits HPV-associated SNSCC clonogenicity in vitro.

A Cell viability analysis of NCI-197 cells 72 h after treatment with alpelisib by impedance assay. Data are (average, n = 3) representative of three independent experiments done in triplicates. B Cell viability analysis of NCI-197 cells 72 h after alpelisib treatment by ATP measurement. Data are mean ±SEM of three independent experiments (n = 3) done in triplicates. C Dose response of alpelisib and (D) Dose response of verteporfin assessed by colony formation assay of NCI-197 cells. Data represents mean ± SEM of three independent experiments (n = 3) done in triplicates. P-values were calculated by ordinary one-way ANOVA with Tukey’s multiple comparison test. Significant p-values are in comparison with control group. E Combination effects of alpelisib and verteporfin assessed by colony formation assay of NCI-197 cells. Data represents mean ± SEM (n = 3 per group) of three independent experiments done in triplicates. P-values were calculated by ordinary one-way ANOVA with Tukey’s multiple comparison test. F Loewe synergy scores and p-values (bootstrapping of a dose-response matrix) calculated using SyngeryFinder software for alpelisib and verteporfin dose responses (100, 250 nM). Source data are provided as a Source data file.