Clusterization in head and neck squamous carcinomas based on lncRNA expression: molecular and clinical correlates

Long non-coding RNAs (lncRNAs) have emerged as key players in a remarkably variety of biological processes and pathologic conditions, including cancer. Next-generation sequencing technologies and bioinformatics procedures predict the existence of tens of thousands of lncRNAs, from which we know the functions of only a handful of them, and very little is known in cancer types such as head and neck squamous cell carcinomas (HNSCCs).

Here, we use RNAseq expression data from The Cancer Genome Atlas (TCGA) and various statistic and software tools in order to get insight about the lncRNome in HNSCC. Based on lncRNA expression across 426 samples, we discover five distinct tumor clusters that we compare with reported clusters based on various genomic/genetic features. Results demonstrate significant associations between lncRNA-based clustering and DNA methylation, TP53 mutation, and human papillomavirus infection. Using "guilt-by-association" procedures, we infer the possible biological functions of representative lncRNAs of each cluster. Furthermore, we found that lncRNA clustering is correlated with some important clinical and pathologic features, including patient survival after treatment, tumor grade, or sub-anatomical location.

We present a landscape of lncRNAs in HNSCC and provide associations with important genotypic and phenotypic features that may help to understand the disease.