Global gene expression profile of proliferative verrucous leukoplakia and its underlying biological disease mechanisms

Abstract

Background: Proliferative verrucous leukoplakia (PVL) is a rare and enigmatic oral potentially malignant disorder which almost invariably results in oral squamous cell carcinoma (OSCC). The aims of this project were to use transcriptome profiling to characterise PVL gene expression patterns for biomarker identification and gain insight into the molecular aetiopathogenesis of PVL.

Methods: Forty-three oral cavity mucosal biopsies from 32 patients with oral lesions clinically compatible with either PVL or non-PVL conventional oral leukoplakia (OLK) underwent transcriptome profiling by RNA sequencing. Data was analysed by hierarchical clustering, differential gene expression, functional enrichment and network analysis, sparse partial least squares discriminant analysis sPLS-DA, and immune cell phenotypic estimation.

Results: We found 464 genes significantly differentially expressed at least 2-fold between PVL and non-PVL OLK (193 up and 271 down). HOX genes, including HOXA1 and HOXB7, keratin-associated proteins (KRTAPs) and olfactory receptor G proteins (OR) were significantly upregulated in PVL. Other upregulated genes in PVL included FOS, WNT16 and IFNA1. Pathway analysis showed that there was a significant downregulation of connective tissue signalling in PVL. Classifying multivariate models based upon 22 genes discriminated PVL from non-PVL OLK. Bioinformatic profiling showed that immune cell profiles in PVL and OLK were similar except that fibroblast markers were reduced in PVL.

Conclusion: These results demonstrate that PVL and conventional OLK are molecularly distinct with upregulation of many cancer-associated genes. They provide insight into the pathogenesis of PVL and show that biomarker based molecular diagnostics is feasible to discriminate and inform diagnosis and management.

Keywords: Biomarkers; Oral leukoplakia; Proliferative verrucous leukoplakia; RNA sequencing; Transcriptome.