Identifying novel genetic drivers of melanoma initiation, metastasis, and drug resistance using zebrafish

Abstract

Sun-exposed cutaneous melanoma has the highest genomic mutational burden of all tumor types. Progression towards metastasis is responsible for most melanoma-related mortality, yet no recurrent mutations specifically driving metastasis have been identified. Instead, a progressive increase in copy number variation (CNV) and recurrent transcriptional states correlating with invasion and/or drug resistance have been described in metastatic melanoma. However, our knowledge on genes that underlie non-focal CNVs and invasion/resistance transcriptional states remains limited. Unlike sun-exposed melanoma, non-sun exposed melanoma genomes lack UV mutational signatures, and their pathogenesis is less well understood. In particular, acral melanoma shows a strong correlation with APOBEC3 mutational signatures at sites of structural variation, but the role of APOBEC3 mutagenesis in tumor initiation has not been studied. In this thesis, I set out to functionally identify novel drivers of cutaneous melanoma progression and to investigate the role of APOBEC3 DNA-hypermutators in tumor initiation by leveraging zebrafish mosaic transgenic melanoma models. First, I used a functional genomics approach to identify candidate CNV progression drivers. I identified HDGF and SETDB1 on chromosome 1q as additive functional drivers of melanoma progression and prognosis, thus pinpointing an evolutionary advantage that may underlie one of the most significant arm-level amplifications observed in metastatic melanoma. Next, using a discovery-driven approach for epigenetic/transcriptional regulators accelerating melanoma development we identified SATB2 as a novel driver of metastatic melanoma. Transcriptional rewiring by SATB2 to a neural crest mesenchyme-like program drives melanoma invasion and resistance to FDA-approved BRAF inhibitor Vemurafenib. Lastly, I tested whether any of the 7 APOBEC3 DNA-hypermutator family members can initiate melanoma. Melanocyte-specific expression of APOBEC3 genes alone did not result in tumor formation, but APOBEC3A and APOBEC3B cooperated with BRAFV600E to initiate malignant melanoma, likely through WNT pathway activation. The studies described in this thesis further our molecular understanding of cutaneous melanoma progression and show APOBEC3A/B hypermutators in cooperation with an oncogene to be sufficient to initiate tumors.