Mitochondrial genome editing reveals the impact of complex I heteroplasmic variants on human neurogenesis

Mitochondrial DNA (mtDNA) variants, whether maternally inherited or acquired somatically during aging, can be particularly detrimental to cells that rely on mitochondrial metabolism such as neurons. The functional consequences of mtDNA variants are thought to depend on their heteroplasmy, defined as the proportion of mtDNA molecules carrying that variant. However, investigating the specific effects of heteroplasmic mtDNA variants on human neuronal function has been challenging. Studies using patient-derived cells cannot conclusively attribute observed phenotypes to a single mtDNA variant, while conventional genome editing approaches are not applicable to mtDNA. Here, we use DddA-derived cytosine base editor (DdCBE) to introduce pathogenic heteroplasmic mtDNA variants into Neurogenin 2 (NGN2)-inducible human induced pluripotent stem cells (iPSCs). We focus on two variants in the genes MT-ND1 and MT-ND5 belonging to complex I (CI), the largest component of the mitochondrial respiratory chain and a frequent contributor to human disease, including the incurable neurodevelopmental disorder Leigh syndrome. mtDNA-edited iPSC lines exhibit low off-target editing and maintain similar heteroplasmy following subclonal isolation and differentiation into NGN2-induced neurons (iNs). Although the CI variants do not substantially affect mitochondrial energy metabolism in undifferentiated iPSCs, they reduce mitochondrial membrane potential in iNs where they drive a transcriptional reprogramming indicative of impaired neuronal commitment. These findings suggest that heteroplasmic CI variants are sufficient to disrupt neural fate specification even at relatively low heteroplasmy. More broadly, mtDNA editing in human iPSCs provides a powerful platform to investigate how mitochondrial fitness and heteroplasmy influence development and disease.