CROSS-DISEASE ANALYSIS OF MOLECULAR AND CELLULAR MECHANISMS IN KIF5A-ASSOCIATED NEURODEGENERATIVE DISORDERS

Kinesin family member 5A (KIF5A) is a neuron-specific kinesin involved in anterograde axonal transport. Heterozygous variants within the three KIF5A domains are linked to distinct neurodegenerative or neurodevelopmental disorders. Missense and nonsense mutations in KIF5A motor and stalk domains are associated with hereditary spastic paraplegia type 10 (SPG10) and axonal Charcot-Marie-Tooth disease (CMT), while frameshifts in its tail are linked to amyotrophic lateral sclerosis (ALS) and neonatal intractable myoclonus (NEIMY). To date, the molecular bases underpinning such clinical heterogeneity have been only partially elucidated, while the wide range of KIF5A mutations prompts the need to compare the molecular mechanisms of different variants. Therefore, in this thesis five key KIF5A mutants (the novel R17Q, R280C, R864*, N999Vfs*40, and C975Vfs*73) covering the whole spectrum of KIF5A-related pathologies were functionally characterised in comparison to wild type (WT) KIF5A in a homogeneous experimental setting. Overexpression in NSC-34 motor neuron-like cells resulted in an abnormal peripheral distribution for the CMT-linked R864* and the ALS-associated N999Vfs*40 KIF5A. Such aberrant localisation is due to an impairment in kinesin autoinhibition resulting from the loss or the modification of KIF5A tail domain, respectively. Moreover, expression of either R864* or N999Vfs*40 KIF5A altered mitochondrial distribution, suggesting the disruption of KIF5A transport competence. The N999Vfs*40 KIF5A mutant also formed sequestosome 1/p62 (SQSTM1/p62)-positive detergent-insoluble inclusions sequestering WT KIF5A, indicating a gain of toxic function. Cycloheximide chase assays in SH-SY5Y neuroblastoma cells demonstrated a shorter half-life for the SPG10-related R17Q and ALS-related N999Vfs*40 mutants compared to WT KIF5A and proteasomal blockage led to their accumulation into detergent-insoluble inclusions, demonstrating that these mutants are degraded by the ubiquitin-proteasome system and that their aggregation may depend on proteostasis impairment. Of note, R280C and N999Vfs*40 KIF5A both competed for degradation with proteasomal substrates. Interestingly, the aberrant biochemical behaviour of the ALS mutant N999Vfs*40 KIF5A was more severely recapitulated by the NEIMY mutant C975Vfs*73 KIF5A, which shares the last portion of its abnormal C-terminal tail yet causes a neurodevelopmental disorder, on the opposite end of KIF5A-linked phenotypic spectrum compared to ALS. Finally, integrated, inducible, isogenic lower motor neurons (i3LMNs) were transduced with lentiviral particles harbouring WT or mutant KIF5A to establish a model to evaluate the impact of KIF5A mutations on bidirectional axonal trafficking in a more physiologically relevant context than immortalised cells. Unfortunately, this strategy was ineffective most likely due to the autoregulation of mutant KIF5A protein levels in i3LMNs. Altogether, data collected in this thesis support the pathogenicity of newly identified KIF5A mutants, shed light on previously uncharacterised aberrant behaviours of recurrent variants, and demonstrate that both unique and shared mechanisms are implicated in the pathogenesis of KIF5A-related neurodegenerative and neurodevelopmental diseases.