Genome editing in mitochondria corrects a pathogenic mtDNA mutation in vivo

Abstract

Mutations of the mitochondrial genome (mtDNA) underlie a substantial portion of mitochondrial disease burden. These disorders are currently incurable and effectively untreatable, with heterogeneous penetrance, presentation and prognosis. To address the lack of effective treatment for these disorders, we exploited a recently developed mouse model that recapitulates common molecular features of heteroplasmic mtDNA disease in cardiac tissue: the m.5024C>T tRNA^Ala mouse. Through application of a programmable nuclease therapy approach, using systemically administered, mitochondrially targeted zinc-finger nucleases (mtZFN) delivered by adeno-associated virus, we induced specific elimination of mutant mtDNA across the heart, coupled to a reversion of molecular and biochemical phenotypes. These findings constitute proof of principle that mtDNA heteroplasmy correction using programmable nucleases could provide a therapeutic route for heteroplasmic mitochondrial diseases of diverse genetic origin.

Figure 1. Strategy to eliminate m.5024C>T and in vivo mtDNA heteroplasmy modification.

A Illustration of mtZFN strategy. A wild-type specific monomer (WTM1), bind upstream of m.5024 in wild-type and mutant genomes; a mutant specific monomer (MTM25) binds preferentially to the mutated site. Dimerization of obligatory heterodimeric FokI domains produces DNA double-stand breaks resulting in specific depletion of mutant mtDNA. B Pyrosequencing of m.5024C>T heteroplasmy from MEFs transfected with controls or MTM25/WTM1 at differing concentrations facilitated by tetracycline-sensitive HHR . Change (Δ) in m.5024C>T heteroplasmy is plotted. utZFN is a mtZFN that does not have a target site in mouse mtDNA . n = 5 (mtZFN, low expression), 8 (mtZFN, high expression), 4 (all other conditions) biologically independent cell cultures (Table S2). Error bars indicate SD. Statistical analysis performed: two-tailed Student’s t-test. Vehicle/mtZFN low expression p = 0.000021, vehicle/mtZFN high expression p = 0.000083. Measure of center is the mean. C Scheme of in vivo experiments. MTM25 and WTM1 are encoded in separate AAV genomes, encapsidated in AAV9.45 then simultaneously administered by tail-vein (TV) injection. Animals are sacrificed at 65 days post-injection. D Western blot of total heart protein from animals injected with 5*10¹² vg MTM25 and/or WTM1. Both proteins include the HA tag and are differentiated by molecular weight. This blot was performed twice with similar results. Raw data are available for this panel (Fig. S9). E Pyrosequencing of m.5024C>T heteroplasmy from ear [E] and heart [H] total DNA. Change (Δ) in m.5024C>T is plotted. n = 20 (vehicle), 3 (WTM1 only), 4 (all other conditions) animals (Table S2). Error bars indicate SEM. Statistical analysis performed: two-tailed Student’s t-test. Vehicle/intermediate dose p < 0.00001, Vehicle/high dose p < 0.00001. Measure of center is the mean. F Assessment of mtDNA copy number by qPCR. n = 8 (vehicle), 4 (all other conditions) animals (Table S2). Error bars indicate SEM. Statistical analysis performed: two-tailed Student’s t-test p = 0.007931. Measure of center is the mean.

Figure 2. Reduction of m.5024C>T mtDNA heteroplasmy results in phenotype rescue.

A Illustration of mt-tRNA^ALA bearing the m.5024C>T mutation. Given the nature and position of this mutation, transcribed tRNA molecules containing the mutation mispair are unlikely to fold correctly or be aminoacylated, resulting in reduced steady-state levels of mt-tRNA^ALA at high levels of m.5024C>T heteroplasmy . B Quantification of high-resolution northern blot data from total heart RNA extracts. mt-tRNA^ALA and mt-tRNA^CYS abundance was normalized to 5S rRNA. n = 8 (vehicle), 4 (all other conditions) animals (Table S2). Error bars indicate SEM. Statistical analysis performed: two-tailed Student’s t-test. Vehicle/intermediate dose p < 0.00001, vehicle/high dose p = 0.00011. Measure of center is the mean. C Principal component analysis (PCA) plot of metabolomic data for intermediate dose AAV-treated mice and age/initial heteroplasmy-matched controls acquired by LC-MS (Table S2). n = 3 (vehicle), 4 (AAV) animals. D Total metabolite levels of pyruvate from samples measured in C. n = 3 (vehicle), 4 (AAV) animals. Error bars indicate SEM. Statistical analysis performed: one-tailed Student’s t-test. p = 0.046403. Measure of center is the mean. E Total metabolite levels of lactate from samples measured in C. n = 3 (vehicle), 4 (AAV) animals. Error bars indicate SEM. Statistical analysis performed: one-tailed Student’s t-test. p = 0.03505. Measure of center is the mean. E Total metabolite levels of aspartate from samples measured in C. Error bars indicate SEM. n = 3 (vehicle), 4 (AAV) animals. Measure of center is the mean