Molecular Analysis of Recurrent Translocations in Mucoepidermoid Carcinoma

Abstract

Head and neck cancers include a diverse group of malignancies, and pathogenesis is driven by different recurring somatic mutations. In head and neck squamous cell carcinoma (HNSCC), these mutations include single nucleotide variants of several different genes, as well as HPV viral integration. By contrast, many salivary gland tumors are characterized by genomic translocations, resulting in frequent gene fusions. For example, mucoepidermoid carcinomas (MEC) have prevalent CRTC1-MAML2 fusions, while hyalinizing clear cell carcinomas (HCCC) have prevalent EWSR1-ATF1 fusions. Exploring the molecular phenotypes caused by these driver mutations and others will better explain the mechanisms of tumorigenesis and growth in salivary gland tumors, and is therefore necessary to identify potential targets for future patient treatments. In my thesis, I investigate the hypothesis that driver mutations, such as CRTC1-MAML2 and EWSR1-ATF1, alter transcription regulation in salivary gland tumors, varying based on tumor type, fusion status, and grade. In this thesis, I begin by using molecular techniques to differentiate two salivary gland tumors, MEC and HCCC, which are difficult to differentiate by standard histopathology approaches. Using RNA sequencing (RNAseq), I identify a 354 gene signature that differentiates both malignancies. These genes are significantly enriched for an ATF1 binding motif, consistent with the EWSR1-ATF1 fusion found in HCCC. These differentially expressed genes include IGF1R, SGK1, and SGK3, which are elevated in HCCC tumors. This, and other differentially expressed genes in this signature, describe examples of differing molecular pathology between MEC and HCCC. I then seek to further understand the genetic underpinning of MEC. Within MEC tumors, the most common somatic translocation forms the CRTC1-MAML2 fusion. I map the CRTC1-MAML2 breakpoint in four MEC-derived cell lines, via long-read sequencing. I also identify a series of genomic translocations leading to this fusion and uncover a TERT promoter rearrangement in NCI-H292. Subsequent TERT break apart FISH reveals TERT copy number increase and translocation events in all four cell lines tested. These experiments reveal complex genomic rearrangement leading to CRTC1-MAML2 formation and a novel TERT driver mutation. Thus, I discover and validate TERT as a novel MEC driver. While the CRTC1-MAML2 fusion is the most common MEC driver mutation, patients with CRTC1-MAML2, or less commonly CRTC3-MAML2, positive tumors have a better prognosis. Therefore, using RNAseq on 48 MEC tumors, I identify gene expression patterns associated with tumor CRTC1/3-MAML2 fusion status and grade. Gene expression signatures associated with fusion status are enriched for gene sets involving cellular respiration, including oxidative phosphorylation and the electron transport chain. Moreover, changes to T and B cell infiltration are associated with MAML2 fusion status and grade, respectively. Therefore, I perform spatial RNA sequencing to measure the effect of CRTC1-MAML2 activity throughout the MEC tumor microenvironment. I identify spatial overlap between CRTC1-MAML2 associated gene expression and many other transcripts, including VEGFA and CTNNB1. These data suggest that CRTC1-MAML2-associated gene expression affect a variety of biological processes throughout the tumor microenvironment. Overall, these data describe a pattern of gene regulation dependent on tumor type, fusion status, and grade. These gene expression changes, coupled with novel driver mutations, such as TERT translocation, affect multiple cancer phenotypes throughout the tumor microenvironment. These biological processes play a role in the molecular etiology of HCCC and MEC tumors, uncovering several pathways which are opportunities to advance targeted therapies, which may improve the survival of MEC and HCCC patients.