CHCHD10 MUTATIONS SYNERGIZE WITH TDP-43 TO PROMOTE NEURONAL APOPTOSIS

CHCHD10 is a mitochondrial protein, in which multiple mutations were recently linked to familial ALS as well as sporadic ALS-FTLD(1-4). The ALS-linked CHCHD10 S59L mutation has been shown to reduce mitochondrial length and disorganize mitochondrial cristae morphology(1). Otherwise, the cellular function of CHCHD10 is completely unknown. While TDP-43 mutations are found in a small proportion of ALS and FTLD-TDP, TDP-43 pathology is also found in 59% of AD cases and also a large fraction of sporadic ALS and FTLD cases(5,6). Increasing evidence indicates that TDP-43 is highly neurotoxic, in large part, by inducing mitochondrial dysfunction(7,8). We utilized combinations of CHCHD10 knockdown and overexpression of wild type or mutant CHCHD10 (R15L & S59L) with/without TDP-43 in hippocampus-derived HT22 cells and primary neurons to assess mitochondrial health, apoptosis, as well as localization and interaction of CHCHD10 and TDP-43. In addition, we utilized a C. elegans model of har-1 (CHCHD10 ortholog) mutation or knockout for worm movement and mitochondrial health. In HT22 cells, wildtype and ALS/FTLD-linked R15L and S59L mutant Flag-CHCHD10 largely colocalized with mito-dsRed. However, significantly more R15L and S59L mutants were not colocalized with mitochondria compared to wild type CHCHD10. At the same time, R15L and S59L transfected cells showed more rounded and aggregated mitochondria, thereby reducing mitochondrial length. As such, recombinant His-tagged S59L variant showed reduced binding to isolated mitochondria, whereas recombinant R15L variant was consistently insoluble in E. Coli. While TDP-43 or CHCHD10 variants showed minimal toxicity per se, co-expression of TDP-43 with CHCHD10 mutations but not wildtype dramatically enhanced neurotoxicity, as evidenced by Annexin V/PI staining. Intriguingly, CHCHD10 mutations preferentially interacted with TDP-43 compared to wildtype. Both CHCHD10 knockdown and CHCHD10 mutations significantly reduced mitochondrial membrane potential and increased mitochondrial superoxide levels. In primary neurons, CHCHD10 mutations but not wildtype decreased the ratio of nuclear to cytoplasmic TDP-43 and reduced postsynaptic drebrin levels. In C. elegans, both the G73E mutation and knockout of har-1 significantly impaired worm movement and increased mitochondrial superoxide levels. We conclude that ALS/FTLD-linked CHCHD10 mutations and loss of CHCHD10 impair mitochondrial function principally via a pathway involving TDP-43 nuclear to mitochondrial localization.