Complementary in vitro models of neurodegeneration: metabolic imbalance, protein aggregation and context- dependent neuroprotection.

Neurodegenerative diseases are a class of pleiotropic disorders that arise from the interplay of multiple stressors, including metabolic imbalance, chronic inflammation and proteotoxic stress. This thesis aimed to explore two converging axes of neurodegeneration in vitro: hyperglycaemia-induced metabolic stress and protein aggregation-driven proteotoxicity, using complementary retinal and glial-like human cell models. ARPE-19 retinal pigment epithelial cells were exposed to sustained high glucose to model diabetic retinal neurodegeneration and to assess the modulatory effects of β-hydroxybutyrate (BHB). Chronic hyperglycaemia induced a state of sub-lethal dysfunction characterised by reduced viability, impaired migration and clonogenicity, cell-cycle alterations and a proinflammatory cytokine profile. BHB partially counteracted these changes, supporting survival and cell-cycle progression, improving functional parameters and attenuating selected inflammatory mediators, consistent with a role as a metabolic and immunomodulatory factor. In parallel, U251 glioma cells expressing mutant huntingtin (cytoplasmic inclusions) or the NKX6-2 L163V variant associated with spastic ataxia type 8 (nuclear inclusions) were used to model proteotoxic stress. Several compounds were tested to explore their effects on pathological protein aggregation: Pt(IV) tris-chelate complexes, platinum-based nucleoside derivatives and natural compounds (flavonoids, a cannabinoid and curcumin-derived molecules). They did not abolish aggregate formation, but many induced clear, compartment- specific remodelling of aggregate morphology and, in selected cases, reduced the prevalence of nuclear inclusions. Overall, the findings highlight that ARPE-19 and U251 cells provide complementary insights into metabolic and proteotoxic mechanisms of neurodegeneration, while emphasising the need to align candidate interventions with the dominant pathogenic drivers in each context. Future work should extend cross-testing of compounds across models, deepen mechanistic analysis and move towards multi-compound, multi-axis strategies better suited to the pleiotropic nature of neurodegenerative disorders.