Aetiology of Oral Cavity Squamous Cell Carcinoma An Exploration utilising Next Generation Sequencing

Abstract

Oral cavity Squamous Cell Carcinoma (OSCC) is a common form of head and neck cancerthrough the developed and developing world. Tobacco exposure is a significant risk factor inthe aetiology of OSCC, however 20% of these cancers occur in a non-smoking population.There have been multiple suggestions as to the causation of these cancers in the non-smokingpopulation, including the human papilloma virus (HPV) and chronic dental trauma. Theevidence on HPV has been divided, clouded by experimental design, differing methods ofHPV detection, small study populations and inclusion of data from oropharyngeal SCC in thestudy population. There has been limited evidence exploring the role of chronic dental traumain the aetiology of OSCC. This project sought to explore the aetiology of oral cavity cancers,by exploring the genetic and transcript differences in a population of OSCCs collected fromBrisbane Head and Neck Clinics between 2013 to 2015.Following participant recruitment and informed consent, a fresh tissue biopsy was taken fromeach individual, either at the time of diagnosis, or at the time of surgical treatment. DNA andRNA extraction was performed on the fresh tissue. Extracted DNA from 14 samples weretested using the HC2 High-Risk HPV test kit, to attempt to identify HPV in the tissue.However, none of the 14 samples tested positive with the HC2 kit. The OSCC samples weretested using PCR, with both a generic HPV primer (MY11/09) and a HPV16 specific primer.From these 14 samples that were taken, none showed the presence of HPV16, with only onesample having the potential presence of a low risk HPV present.As there was no conclusive evidence suggesting high-risk HPV plays a significant role in theaetiology of OSCC from the HC2 or PCR testing performed, this study progressed to DNAexome sequencing and RNA sequencing. The RNAseq data was first explored to understandthe RNA transcriptome. Following mapping of the RNAseq data to the human genome, thisstudy then mapped the RNAseq data to the HPV genome to attempt to identify HPV presencein the transcriptome. Only two samples were HPV16 positive (4.54%). These samplesshowed high expression of E6 and E7, indicating that HPV16 was a likely contributing factor,if not a causative factor in these two samples. Both of these samples were however fromcurrent smoking participants. A principal component analysis (PCA) attempted to explore whether the samples, whenarranged by the principal variance components, would cluster based upon the demographicscollected. This study utilised the PCA to determine whether the smoking status, tumour stageand tumour location of the participant would reveal clustering of samples based on the RNAtranscriptome. None of these demographic factors showed significant clustering, based on an80% confidence interval. A Weight Gene Co-expression Network Analysis (WGCNA)analysis was performed, to attempt to identify whether transcriptome differences could beexplained by the demographic details that were collected. Overall, more advanced diseaseshowed transcriptome changes consistent with the Warburg effect. Tumours from smokingparticipants showed a non-significant trend towards increased m-TOR signalling, TLR4signalling and EGF-EGFR signalling. This may potentially be of use in future therapeuticdevelopment.DNA exome analysis was then completed. Interpretation followed the procedure outlined bythe Broad Institute’s GATK Best Practice Workflow. As there was no matched normal tissuefrom the participants, data from the 1000 genomes project was used to help identify commongermline variants. Utilising the GenVisR package, commonly mutated genes in the data setwere visualised. Genes that were frequently mutated were genes commonly associated withSCCs, especially TP53. Other genes commonly mutated in the data set include HUWE1 andCDKN2A.There is significant future work that may be completed as a result of this research. If thisstudy was repeated, matched normal tissue would be collected, to help eliminate germlinevariants as a confounding factor of the sample analysis. Further demographic data may assistin explaining some the driving aetiology and may allow exploration of other factors, such assurvival. Finally, with increasing next-generation sequencing being performed, both inAustralia and around the world, collaborations between different head and neck centreswould allow increased participant numbers, enabling detection of subtle differences betweensub-groups, which may help with development of personalised therapeutics in the future.