Genetic instabilities drive melanoma development, disrupting immune editing and leading to resistance against revolutionary therapies like immune checkpoint therapies (ICT), as seen in uveal (UM) and cutaneous (CM) subtypes. This thesis adopts a multidisciplinary approach integrating analytical tools in patient samples to reveal key drivers of uveal and cutaneous melanoma malignancy. It further reverses to the bench side to validate an important driver by creating a novel preclinical model that provides new foundations for future biological and therapeutic studies.

In CM, integration of multiple transcriptomic datasets from patients who had received ICT provided new insights into ICT resistance, often linked to cold tumors with deficient antigen presentation. Impaired antitumor immune responses were frequently associated with reduced β2-microglobulin (β2M) levels, which correlated with poor ICT responses in CM. Integrative transcriptomic analyses identified novel β2M-associated biomarkers, with CD1D playing a critical role in natural killer T (NKT) cell functions. Epigenetic profiling showed that methylation regulated β2M and CD1D expression, influencing ICT outcomes. Modulation of these pathways was suggested to enhance ICT efficacy in CM.

In UM, BAP1 loss consistently drove aggressive tumor behavior across populations. Chromosomal analyses in Southeast Asian UM patients revealed distinct genetic features, including less frequent monosomy 3 and more frequent chromosome 1q gains, both linked to poor progression-free survival. These findings underscored the critical and consistent role of BAP1 loss while highlighting the importance of region-specific molecular profiling. To investigate BAP1-dependent tumor-immune changes, CRISPR-engineered BAP1⁻/⁻ mouse melanocyte models were developed, demonstrating lipid metabolic reprogramming and immunosuppressive traits characteristic of high-risk UM. This model offers a robust platform for testing novel immunotherapies targeting BAP1-driven pathways.

This research established a comprehensive preclinical framework to connect clinical observations with translational applications in melanoma. By elucidating key genetic and immune drivers, it provided a robust platform for developing and refining novel therapeutic strategies, particularly immunotherapies.