Genomics-based Assessments of Stemness and DNA Damage Response in Oral Cancer

Abstract

Gingivobuccal oral squamous cell carcinoma (OSCC-GB) is a prominent clinical subtype of head and neck squamous cell carcinoma in India, predominantly affecting habitual users of smokeless tobacco. The genetic basis of OSCC-GB, especially the role of cancer stem cells (CSCs), is not well understood. CSCs are a tumor subpopulation with self-renewal, tumorogenicity, and therapy-resistant properties. Our study elucidates the role of CSCs in OSCC-GB primary tumors and a novel cell line using a custom somatic mutation, bulk and single-cell RNA-sequencing (scRNA-seq) differential gene expression, molecular dynamics (MD), and machine learning (ML) approaches. Somatic mutation analysis was conducted in primary OSCC-GB tumors based on CD44 CSCs marker from tobacco-chewing patients. The sourced data consists of primary tumors, sorted into CD44+Lin- and CD44-Lin- matched groups using fluorescence-activated cell sorting, followed by RNA-seq. We found DNA damage response (DDR)-related known mutational signatures, notably 1 bp T/(A) nucleotide insertions and C>T mutations in the OSCC-GB subgroups, indicating disease etiology linked to smokeless tobacco-related carcinogens. The differential somatic mutation, functional impact predictions, and survival analyses highlight DDR-related genes, particularly CREBBP, in CD44+Lin- OSCC-GB. Nonsense-mediated mRNA decay (NMD)-associated frameshift insertion in the CREBBP histone acetyltransferase domain disrupts acetyl-CoA ligand binding site, impairing acetyltransferase activity and perturbing TP53 activity. The loss-of-function (LoF) finding was rationalized using a 1-microsecond all-atom MD study. Since TP53 represses CD44 expression, reduced TP53 can increase CD44 expression and might drive stemness and tumor progression risk under DDR conditions in CD44+Lin- OSCC-GB. Subsequently, an RNA-seq data-based combined analysis of somatic mutation and gene expression was performed using orosphere (Oro) CSCs experimental model from a novel Indian-origin OSCC-GB cell line ‘IIOC019’ derived from an oral cancer patient with habitual smokeless tobacco use. Our previously reported mutational signatures, C>T mutations and 1bp T/(A) nucleotide insertions, were confirmed in IIOC019. Our study identified a key contributor gene, SON, in IIOC019 Oro, whose LoF variant could be linked to enhanced CSCs-like traits associated with nuclear speckles dysregulation under DDR conditions. This enhanced stemness was linked to the loss of TP53-mediated repression and splicing regulation of CD44. This LoF finding was supported by bulk and scRNA-seq differential gene expression studies. Finally, we leveraged the generalizability of the ML prediction model for ‘Tumor Status’ for a comparative somatic mutation analysis between ‘With Tumor’ and ‘Tumor Free’ OSCC-GB patients. Our results showed that support vector machines classified the ‘Tumor Status’ classes at a mean accuracy of 89% based on clinical features. Our previously reported mutational signatures C>T were identified in OSCC-GB subgroups. The study identified MAPKAP1 gene as a significant player in ‘With Tumor’ OSCC-GB. LoF-associated MAPKAP1 missense mutation in SIN1 domain might alter its phosphorylation activity via mTORC2 signaling pathway. This LoF finding was linked to CREBBP-TP53-CD44-mediated stemness and supported by mutational structural analysis. In summary, our findings reveal that that the cross-talk among CREBBP, SON and MAPKAP1 genes might further promote stemness and disease progression via TP53 under DDR conditions, potentially leading to increased mortality rates among Indian OSCC-GB patients.