Decoding cancer plasticity: long non-coding RNAs and microenvironmental signals driving phenotypic transitions in melanoma and prostate cancer

Cellular plasticity is a fundamental mechanism driving cancer progression, therapeutic resistance, and lineage reprogramming. Beyond the accumulation of genetic alterations, tumour cells can undergo dynamic phenotypic transitions that enable adaptation to environmental pressures and anticancer therapies. This thesis investigates these processes in two tumour types, melanoma and prostate cancer. The first part of the study focuses on phenotype switching in melanoma. Transcriptomic analyses identified two distinct cellular states, a proliferative and an invasive state, and revealed the long non-coding RNAs (lncRNAs) LINC00504 and LINC00520 as key components of the regulatory network controlling this transition. Both lncRNAs were identified as direct MITF targets and were associated with poor patient survival. Functional studies demonstrated that they participate in a competitive endogenous RNA (ceRNA) network with specific miRNAs to sustain the proliferative program, whereas their silencing promotes a switch toward a more invasive phenotype. Despite defining a proliferative cellular state, their role in sustaining uncontrolled proliferation, both at the primary and metastatic sites, may ultimately drive disease progression and poor clinical prognosis. The second part of the thesis investigates lineage plasticity in prostate cancer using the ProMPt (Prostate cancer Modelling Platform) organoid biobank. We identified Epidermal Growth Factor (EGF) as a major driver of culture-induced resistance that masks androgen receptor (AR) dependency under standard conditions. To overcome this limitation, we developed an organoids-fibroblasts co-culture system that restores sensitivity to androgen deprivation while maintaining long-term organoid growth. Using this platform in combination with CRISPR-mediated genome editing, we demonstrated that distinct genetic backgrounds direct different adaptive trajectories under therapeutic and microenvironmental stress. In particular, APC loss favoured a trajectory toward a neuroendocrine-like state, whereas RB1 inactivation promoted a double-negative, lineage-plastic phenotype. Overall, this work highlights how genetic alterations, non-coding regulatory networks, and microenvironmental signals cooperate to drive tumour evolution and therapeutic resistance. Furthermore, it establishes optimized experimental models that provide physiologically relevant platforms for studying cancer plasticity and identifying novel therapeutic vulnerabilities.