Zebrafish models of uveal and cutaneous melanoma for preclinical studies

Uveal melanoma (UM) is the most common primary cancer of the eye and its prognosis is strongly influenced by the occurrence of metastases, which are both rapidly developing and mostly fatal. The most frequent driver mutations occur in a small number of genes including GNAQ, GNA11, BAP1, CYSLTR2 and SF3B1. Due to a lack of suitable animal models, the mechanism through which mutations in these genes cause or cooperate in UM initiation and progression is still largely unknown. We aimed at generating transgenic strains expressing the human mutant proteins in zebrafish uveal melanocytes, using the kita promoter. We used the binary Gal4/UAS system to express the mutant genes mentioned above. Moreover, we performed xenotransplantation experiments with uveal melanoma human and zebrafish cell lines in optically-clear, immunocompromised, zebrafish larvae. Transplanted fish developed melanoma near the site of transplantation in two weeks and showed metastatic growth within one month of age. This approach could be used for short-term assays in larvae, and be further developed for long-term uveal melanoma studies. In parallel, we performed a chemical screen using a transgenic model previously generated in our laboratory, where oncogenic RAS is expressed under the kita promoter. As adults transgenic kita:RAS develop cutaneous melanoma with high frequency and uveal melanoma with a much lower percentage. Larvae showed an increased number of melanocytes already at 3 days post fertilization (dpf) as the earliest evidence of abnormal melanocyte growth. Using this model we performed a chemical screen based on automated detection of a reduction of melanocytes number caused by any of the 1280 FDA or EMA approved drugs of the Prestwick library. The analysis showed that 55 molecules were able to reduce by 60% or more the number of melanocytes per embryo. We identified clotrimazole, as the best candidate. The molecule is an azole derivative acting on the energetic metabolism of melanoma cells. We further tested two compounds for each of the 5 pharmacological classes, and a farnesyltransferase inhibitor (lonafarnib), that inhibits an essential post-translational modification of HRAS and suppresses the hyperpigmentation phenotype. Combinations of clotrimazole and lonafarnib showed the most promising results in zebrafish embryos, allowing a dose reduction of both drugs. We performed validation of these observations in the metastatic human melanoma cell line A375M, and in normal human epithelial melanocytes (NHEM) as control cells, in order to investigate the mechanism of action of clotrimazole in blocking the proliferation of transformed melanocytes. Viability assay and analysis of energy metabolism in clotrimazole treated cells show that this drug specifically affects melanoma cells in vitro and transformed melanocytes in vivo, having no effects on NHEM or wild type larvae. Similar effects were observed with another hit of the antifungal class, miconazole. Furthermore, we show that the effects of clotrimazole are mediated by the inhibition of hexokinase activity and suggest further testing of clotrimazole in combinatorial treatments. In conclusion, this thesis investigated different possibilities of modeling the rare cancer uveal melanoma in zebrafish, using both transgenic and transplantation approaches, and developed a pipeline for a high-throughput, semi-automated chemical screen in a zebrafish melanoma that identified clotrimazole and miconazole as targeting a metabolic vulnerability in melanoma cells.