Integrated analysis of oral tongue squamous cell carcinoma identifies key variants and pathways linked to risk habits, HPV, clinical parameters and tumor recurrence

Abstract

Oral tongue squamous cell carcinomas (OTSCC) are a homogeneous group of tumors characterized by aggressive behavior, early spread to lymph nodes and a higher rate of regional failure. Additionally, the incidence of OTSCC among younger population (<50yrs) is on the rise; many of whom lack the typical associated risk factors of alcohol and/or tobacco exposure. We present data on single nucleotide variations (SNVs), indels, regions with loss of heterozygosity (LOH), and copy number variations (CNVs) from fifty-paired oral tongue primary tumors and link the significant somatic variants with clinical parameters, epidemiological factors including human papilloma virus (HPV) infection and tumor recurrence. Apart from the frequent somatic variants harbored in TP53, CASP8, RASA1, NOTCH and CDKN2A genes, significant amplifications and/or deletions were detected in chromosomes 6-9, and 11 in the tumors. Variants in CASP8 and CDKN2A were mutually exclusive. CDKN2A, PIK3CA, RASA1 and DMD variants were exclusively linked to smoking, chewing, HPV infection and tumor stage. We also performed a whole-genome gene expression study that identified matrix metalloproteases to be highly expressed in tumors and linked pathways involving arachidonic acid and NF-k-B to habits and distant metastasis, respectively. Functional knockdown studies in cell lines demonstrated the role of CASP8 in a HPV-negative OTSCC cell line. Finally, we identified a 38-gene minimal signature that predicts tumor recurrence using an ensemble machine-learning method. Taken together, this study links molecular signatures to various clinical and epidemiological factors in a homogeneous tumor population with a relatively high HPV prevalence.

SMARCD1 is a transcriptional target of specific non‐hotspot mutant p53 forms

“…An oncogene is ideally expected to exhibit increased expression in tumor compared to normal tissues. We identified significant elevation of SMARCD1 transcript in tumor compared to normal SCCOT samples (Figure a and Table S1c) corroborated in a previously published tumor versus matched normal SCCOT microarray gene expression dataset (Krishnan et al, ; Figure b). SMARCD1 expression was consistently higher in tumor over normal samples (Figure c) validated in several individual HNSCC subtypes as well (Figure S7) as well as other cancer types (Figure d) as determined from the Firebrowse database (firebrowse.org).…”

A minimal set of internal control genes for gene expression studies in head and neck squamous cell carcinoma

“…The gene expression profiling using Illumina HumanHT-12 v4 expression BeadChips (Illumina, San Diego, CA), raw data collection and data processing using the R package lumi 7 and individual batch-correction using ComBat 8 is described previously 9 10 . The raw signal intensities from arrays were exported from GenomeStudio for pre-processing and analyzed using R. Gene-wise expression intensities for tumor and matched control samples from GenomeStudio were transformed and normalized using VST (Variance Stabilizing Transformation) and LOESS methods, respectively, using lumi 7 and top genes with least across-sample variance in expression were selected.…”

“…In the discovery and validation stage, HNSCC tumor: adjacent normal paired samples ( n =63 and n =14 respectively), were used. Previously published whole-genome microarray data 9 10 , RNA-seq data from TCGA 11 and data from individual studies in the literature were used in the discovery stage. For the validation set, 10 HNSCC tumor: normal pairs were from a completely independent cohort (which were not a part of discovery cohort) and 4 were used from the in house discovery cohort.…”

Genomic sequencing and precision medicine in head and neck cancers

“…Aside from the TCGA dataset, which is composed primarily of white patients, large sequencing analyses have been performed on Asian patient groups from India and Singapore [44–46]. Comparing the alteration frequencies in these populations has indicated that there may be significant differences in the mutation rates in CDKN2A , NOTCH1 , and PIK3CA [47].…”