Mitochondrial targeted meganuclease as a platform to eliminate mutant mtDNA in vivo

Abstract

Diseases caused by heteroplasmic mitochondrial DNA mutations have no effective treatment or cure. In recent years, DNA editing enzymes were tested as tools to eliminate mutant mtDNA in heteroplasmic cells and tissues. Mitochondrial-targeted restriction endonucleases, ZFNs, and TALENs have been successful in shifting mtDNA heteroplasmy, but they all have drawbacks as gene therapy reagents, including: large size, heterodimeric nature, inability to distinguish single base changes, or low flexibility and effectiveness. Here we report the adaptation of a gene editing platform based on the I-CreI meganuclease known as ARCUS^®. These mitochondrial-targeted meganucleases (mitoARCUS) have a relatively small size, are monomeric, and can recognize sequences differing by as little as one base pair. We show the development of a mitoARCUS specific for the mouse m.5024C>T mutation in the mt-tRNA^Ala gene and its delivery to mice intravenously using AAV9 as a vector. Liver and skeletal muscle show robust elimination of mutant mtDNA with concomitant restoration of mt-tRNA^Ala levels. We conclude that mitoARCUS is a potential powerful tool for the elimination of mutant mtDNA.

Fig. 1. mitoARCUS construct and mitochondrial expression.

a For an MIT nuclease (against the mouse mtDNA tRNA-Ala mutation [MIT 11–12]), a pair of engineered CHO lines were produced to carry either the wild-type (WT) or mutant mtDNA target site in the nuclear DNA. The target site was positioned between direct repeats of a GFP gene such that cleavage of the target site promotes homologous recombination events between repeated regions to yield a functional GFP. In addition, there is a target site for a positive control nuclease (“CHO 23-24”) incorporated next to the MIT target site (left panel). Each of the cell lines was transfected with mRNA encoding MIT 11–12 or CHO 23-24 (control) and cells were assayed by flow cytometry 2, 5, 7, 9, and 12 days post transfection for the percentage of GFP+ cells (the average GFP fluorescence for the different time points is shown in the right panel). As the different time points are not biological replicates, no statistic was applied. b mitoARCUS gene construct for ex vivo expression includes CMV promoter, mitochondrial localization sequence (MLS) of Cox8 or Cox8/Su9, Flag tag for immunological detection, Meganuclease (ARCUS) sequence, and PolyA tail. c Immunofluorescence done on HeLa cells 24 h after transfection with mitoARCUS. MitoTracker stains mitochondria red, Flag stains mitoARCUS green, and merged image shows colocalization (yellow) of mitoARCUS to mitochondria. Images taken at ×40 magnification. This experiment was repeated twice with identical results. (d) Western blot depicting mitoARCUS expression (FLAG) in HEK293T cells 24 h after transfection with either CF or CSF construct. Lanes CF + GFP and CSF + GFP depict protein expression in cells transfected with mitoARCUS constructs in which we added a GFP sequence. Lane Unt represents untransfected cells. Lane GFP represents cells transfected with GFP only. Tubulin (Tub) expression was used as a loading control. This experiment was performed once.

Fig. 2. mitoARCUS effect on heteroplasmic cells carrying the tRNAAla mutation (m.5024C>T).

a Example of FACS cell sorting gating. Cells were sorted by the presence of GFP co-expression: “Black” cells (bottom gate) and “Green” cells (top gate). b RFLP-HOT PCR analysis of two independent transfections and cell-sorting experiments of heteroplasmic cells carrying 50% m.5024C>T mutation. Mutant levels in the Green cell populations (Gr) were compared to Untransfected cells (U). This experiment was done once. c Quantification of heteroplasmy shift from the two cell-sorting experiments in cells carrying 50% mutation described in b. Results were compared to Untransfected cell heteroplasmy. d RFLP “last-cycle hot” PCR analysis of heteroplasmic cells carrying high heteroplasmic mutant load (90%) transfected with mitoARCUS over time. This experiment was repeated three times with similar results. e Quantification of Fig. 2d. Values are normalized to Untransfected cells. Black cells are named Blk (n = 4). f Total mtDNA levels were checked in highly mutant cells transfected with mitoARCUS and compared to untransfected cells 24 h after transfection, and followed for 3 weeks after transfection (n = 3). p values are related to untransfected cells (100%). g Oxygen consumption rate (OCR) was deduced in cells carrying high levels of heteroplasmic mutant mtDNA that were transfected with mitoARCUS and grown for 3 weeks [n = 3 (CTRL), n = 7 (Unt), n = 3 (Blk), n = 7 (Gr)]. Data are mean ± SEM. Statistical analysis was performed using two-tailed Student’s t-test.

Fig. 3. AAV9-mitoARCUS effect in treated juvenile mice.

a Representative western blottings (W.B.) of homogenates (top panels) with Flag antibody for AAV9-mitoARCUS samples and GFP antibody for AAV9-GFP samples. RFLP “last-cycle hot” PCR analysis (RFLP, bottom panels) of DNA samples from the same injected animals at 6, 12, and 24 weeks PI. Similar analyses were performed for each animal. b Quantification of heteroplasmy shift shown as a percent change in heteroplasmy across all tissues at 6, 12, and 24 weeks PI normalized to brain tissue. Heteroplasmy levels of the heart (H), tibialis anterior (TA), quadriceps (Q), gastrocnemius (G), kidney (K), liver (L), and spleen (Sp) were compared to that of the brain (B) (negative for expression of mitoARCUS). c Quantification by RT-PCR of total mtDNA levels in the skeletal muscle, liver, and brain at 6 and 24 weeks PI using ND1 and ND5 mitochondrial primer/probes normalized to 18S (nuclear DNA). Data are mean ± SEM of n = 4 (with exception of 6 weeks GFP (n = 5) and 24 weeks mitoARCUS (n = 3)). Statistical analysis was performed using two-tailed Student’s t-test.

Fig. 4. AAV9-mitoARCUS effect in treated adult mice.

a Representative western blottings (W.B.) of homogenates (top panels) with Flag antibody for AAV9-mitoARCUS samples and GFP antibody for AAV-GFP samples. RFLP “last-cycle hot” PCR analysis (RFLP, bottom panels) of DNA samples from the same injected animals, at 6, 12, and 24 weeks PI. Similar analyses were performed for each animal. b Quantification of heteroplasmy shift shown as percent change in heteroplasmy across all tissues at 6, 12, and 24 weeks PI normalized to brain tissue. Heteroplasmy of the heart (H), tibialis anterior (TA), quadriceps (Q), gastrocnemius (G), kidney (K), liver (L), and spleen (Sp) were compared to that of the brain (negative for expression of mitoARCUS). c Quantification by qPCR of total mtDNA levels were measured in the heart, skeletal muscle, liver, and brain (B) at 6 and 24 weeks PI using ND1 and ND5 mitochondrial primer/probes normalized to 18S (nuclear DNA). Data are mean ± SEM of n = 4. Statistical analysis was performed using two-tailed Student’s t-test.

Fig. 5. mitoARCUS-induced increase in mt-tRNA^Ala in the liver.

a Northern blot analysis of juvenile mouse liver 24 weeks PI probed for mt-tRNA^Ala and total RNA loading (28S and 18S). b Quantification of mt-tRNA^Ala (from panel a) normalized to 28S rRNA (n = 3). Riboruler high-range molecular marker was used, but the lowest marker was 200 nt, whereas the tRNA are ~75 nt. c Quantification of mt-tRNA^Ala by qPCR compared to levels of mt-tRNA^Val in juvenile mouse liver 24 weeks PI. RNA samples from AAV9-mitoARCUS-treated animals compared to AAV9-GFP controls and WT liver samples (n = 3 with the exception of GFP n = 4). d Quantification of mt-tRNA^Ala compared to levels of mt-tRNA^Val in adult mouse liver. RNA samples from AAV9-mitoARCUS-treated animals compared to AAV9-GFP controls and WT liver samples, at 24 weeks PI (n = 4 with the exception of WT n = 3). Data are mean ± SEM. Statistical analysis was performed using two-tailed Student’s t-test.