Generation of mitochondrial DNA-edited mice using a TALE-based adenine base editor

초록

Background: Mitochondrial diseases, caused by mitochondrial DNA (mtDNA) mutations, are severe rare disorders that remain under investigation. Generating animal models is technically challenging because CRISPR-Cas guide RNAs cannot enter mitochondria. To overcome this, protein-only base editors using transcription activator-like effectors (TALEs) fused to deaminases were developed. Aim: This study aims to utilize a TALE-based adenine base editor (TALED) to generate mouse models harboring pathogenic mtDNA mutations, providing an in vivo platform for studying mitochondrial diseases. Methods: TALED mRNAs were microinjected into mouse zygotes (MT-ATP6, MT-Rnr1 target were tested). Upon confirming editing efficiency in cultured blastocysts via next-generation sequencing (NGS), microinjected 2-cell embryos were transferred into surrogate mothers. Live pups were genotyped to confirm successful mtDNA editing. Results: Since the original TALED harboring TadA8e deaminase caused embryonic lethality upon microinjection, we utilized a variant with reduced toxicity, TALED-V28R. For the MT-ATP6 target, the average heteroplasmic rate was 42% in edited blastocysts (n = 17). Following embryo transfer, the resulting pup (n = 6) exhibited an average heteroplasmic rate of 62.4%, confirming the successful generation of highly mtDNA mutated models. Conversely, targeting MT-Rnr1 initially yielded a low heteroplasmy of 7.7% (n = 25) in blastocysts. By substituting the TALE N-terminal and DddA components with engineered versions, we increased the average blastocyst heteroplasmy to 42.2% (n = 25), which subsequently yielded mutant pup (n = 6) with an average heteroplasmy of 37.5%. Further phenotypic analyses confirmed functional in vivo mitochondrial defects, as evidenced by decreased heart rates (bpm) in MT-ATP6 mice and decreased ATP levels in MT-Rnr1 muscle biopsies. Conclusions: MtDNA-mutant mouse models serve as invaluable platforms for investigating mitochondrial diseases. Although the specific models generated herein do not perfectly replicate human patient mutations, our demonstration may pave the way for generating precise disease models to elucidate pathological mechanisms and evaluate future therapeutic interventions.