Clinically translatable mitochondrial gene therapy in muscle using tandem mtZFN architecture

Abstract

Mutations in the mitochondrial genome (mtDNA) often lead to clinical pathologies. Mitochondrially-targeted zinc finger nucleases (mtZFNs) have been successful in reducing the levels of mutation-bearing mtDNA both in vivo and in vitro, resulting in a shift in the genetic makeup of affected mitochondria and subsequently to phenotypic rescue. Given the uneven distribution in the mtDNA mutation load across tissues in patients, and a great diversity in pathogenic mutations, it is of interest to develop mutation-specific, selective gene therapies that could be delivered to particular tissues. This study demonstrates the effectiveness of in vivo mitochondrial gene therapy using a novel mtZFN architecture on skeletal muscle using adeno-associated viral (AAV) platforms in a murine model harboring a pathogenic mtDNA mutation. We observed effective reduction in mutation load of cardiac and skeletal muscle, which was accompanied by molecular phenotypic rescue. The gene therapy treatment was shown to be safe when markers of immunity and inflammation were assessed. These results highlight the potential of curative approaches for mitochondrial diseases, paving the way for targeted and effective treatments.

Keywords: Adeno-Associated Viruses (AAV); Gene Therapy; Skeletal Muscle; Zinc Finger Nuclease (mtZFN); mtDNA Heteroplasmy Modification.

Figure 1. Schematic of the m.5024C>T mtZFNs and co-delivery strategy.

(A) The m.5024C>T mutation, occurring in mt-tRNA^Ala, results in a mismatch in the tRNA acceptor stem, as shown in the clover-leaf representation (top right). This in turn results in secondary structure disruption and lower levels of steady state mt-tRNA^Ala (Kauppila et al, 2016). To remodel the genotype, an appropriate pair of mtZFN heterodimers was optimized (bottom). Binding sites of both constructs are shown in blue or red, respectively. The mutation site base is capitalized. The 5’ and 3’ ends as well as the heavy (H) and light (L) mtDNA strands are also indicated. (B) The compact nature of mtZFNs enables viral co-encapsidation of both monomers in a single rAAV capsid. Both constructs are encoded in a single ORF by the insertion of a T2A peptide coding sequence (Liu et al, 2017). The residues of the T2A peptide remain appended to the proteins upon dissociation, adding a 17-amino acid chain to the C-terminus of the first protein and a residual proline on the N-terminus of the second one.

Figure 2. Effect of sequential arrangement of tandem mtZFN delivery in vitro.

(A) Western blot analysis of delivery of both heterodimers to HEK293 cells 48 h following transfection. Construct MTM25 was tagged with HA, whereas WTM1 was tagged with FLAG. Heterodimers were either delivered separately (+) or by expression from a single plasmid (T2A) in both sequential arrangements. The two bands are the precursor (p) and mature (m) forms of the peptides. The hammerhead ribozyme (HHR) was used to attenuate expression of the single constructs. All constructs were delivered in equimolar amounts. Coomassie brilliant blue (CBB) and Vinculin are loading controls. (B) Heteroplasmy shifting two weeks after electroporation, as assessed by pyrosequencing. Dual color red and blue denotes separate delivery of constructs on two plasmids. Purple color denotes expression of constructs from a single ORF. Mean ± SEM is displayed. Significance of treated samples was calculated with one-way ANOVA using Dunnett’s multiple comparison against ‘Vehicle’. Student t-test was used to calculate significance between both orientations of mtZFN in tandem. Each symbol is a biological replicate. Six data points were pooled from two independent biological experiments each having biological triplicates. Source data are available online for this figure.

Figure 3. Intravenous delivery of mtZFN using cardiotropic rAAV9.45.

(A) Schematic representation of experimental design. Tail vein administration of AAV9.45 encapsidated mtZFN, either separately or in the tandem configuration. (B) Heteroplasmy shift measured by pyrosequencing as the difference between initial ear biopsy and final muscle heteroplasmy values 65 days post-rAAV administration. Doses shown in increasing order were 5E11, 1E12, 5E12, and 1E13 vg/mouse for separate virions containing mtZFN monomers (red and blue) and 2.5E11, 5E11, 2.5E12, and 5E12 vg/mouse for the virions containing the tandem mtZFN configuration (purple). NB: Data points for 1E12, 5E12, and 1E13 vg/mouse doses used for separate mtZFN monomers are taken from Gammage et al (2018b) to allow direct comparison with the tandem configuration. Mean ± SEM are displayed, and statistics were performed using one-way ANOVA using Dunnett’s comparison to the control. N-numbers were 24 for control samples and 4 biological replicates for each experimental condition. Mean ± SEM are displayed, and statistics were performed using one-way ANOVA using Dunnett’s comparison to the control. (C) MtDNA copy number (mtCN) relative to untreated control samples measured by qPCR. Mean ± SEM are displayed, and statistics were performed using one-way ANOVA using Dunnett’s comparison to the control. N-numbers were 24 for control samples and 4 biological replicates for each experimental condition. Source data are available online for this figure.

Figure 4. Intramuscular administration of mtZFN.

(A) Schematic representation of experimental design. Left legs were injected with rAAV buffering solution containing HA-eGFP mock construct as a positive marker for injection. Right legs were injected with rAAVs encoding mtZFNs, either in separate viral particles or encapsulated in tandem, diluted with control solution for left leg. Treatments were delivered in equimolar amounts, with four biological replicates (mice) per condition. Statistics were performed using paired t-test. (B) Heteroplasmy shift measured by pyrosequencing as the difference between initial ear biopsy and final muscle heteroplasmy values 65 days post rAAV administration. Mean ± SEM are displayed and statistics for both experiments were performed by paired t-test. N-number was four for both experimental conditions. (C) Relative mtCN in both left and right legs assessed by qPCR. Mean ± SEM are displayed and statistics for both experiments were performed by paired t-test. p = 0.826 for WTM1 + MTM25 and p = 0.1014 for MTM25-T2A-WTM1. (D) Pooled and averaged ratios from digitally quantified and normalized northern blot. Statistics were performed using paired t-test. p = 0.049 for WTM1 + MTM25 and p = 0.0298 for MTM25-T2A-WTM1. N-number was four for both experimental conditions. Source data are available online for this figure.

Figure 5. Intravenous delivery of mtZFNs using myotropic rAAV.

(A) Schematic representation of experimental design. Tail vein administration of AAVMYO encapsidated mtZFNs, either separately or in tandem. (B) Heteroplasmy shift measured by pyrosequencing as the difference between initial ear biopsy and final muscle heteroplasmy values 65 days post-rAAV administration. Doses shown in increasing order are 4E10 (4 biological replicates), 2E11 (4 biological replicates) and 1E12 (8 biological replicates) vg/mouse for both virions containing mtZFN monomers (red and blue) and 2E10 (4 biological replicates), 1E11 (4 biological replicates) and 5E11 (8 biological replicates) vg/mouse for the virions containing the tandem mtZFN configuration, with 4 control animals. Mean ± SEM are displayed, and statistics were performed using one-way ANOVA using Dunnett’s comparison to the control (C) mtCN relative to untreated control samples were measured by qPCR. Mean ± SEM are displayed, and statistics were performed using one-way ANOVA using Dunnett’s comparison to the control. Doses shown in increasing order are 4E10 (4 biological replicates), 2E11 (4 biological replicates), and 1E12 (8 biological replicates) vg/mouse for both virions containing mtZFN monomers (red and blue) and 2E10 (4 biological replicates), 1E11 (4 biological replicates), and 5E11 (8 biological replicates) vg/mouse for the virions containing the tandem mtZFN configuration, with 4 control animals. Source data are available online for this figure.

Figure 6. Innate immune responses to muscle gene therapy.

RT-qPCR measurements of transcripts related to pro-inflammatory cytokines and the type I interferon response obtained for both separate and tandem administrations. The following transcripts were analyzed: (A) Il1b, (B) Tnfa, (C) Ifnb1, (D) Cxcl10, (E) Ddx58, (F) Isg20. The doses were 4E10 (4 biological replicates), 2E11 (4 biological replicates), and 1E12 (8 biological replicates), vg/mouse for virions containing mtZFN monomers (red and blue), and 2E10 (4 biological replicates), 1E11 (4 biological replicates), and 5E11 (8 biological replicates) vg/mouse for virions containing the tandem mtZFN configuration, with 4 control animals. Samples were collected 65 days post-rAAV administration. Samples were performed in technical quadruplicate and fold change calculated compared to age and heteroplasmy matched control mice. Bar plot shows mean ± SEM. Statistical analysis was conducted using one-way ANOVA, with results compared to the control group utilizing Dunnett’s test. Source data are available online for this figure.

Figure 7. Dose titration of AAVMYO encoding tandem mtZFN.

(A) Schematic representation of experimental design. (B) Heteroplasmy shift measured by pyrosequencing as the difference between initial ear biopsy and final muscle heteroplasmy values 65 days post rAAV administration. Doses shown in increasing order 2E10 (4 biological replicates), 1E11 (4 biological replicates), 2.5E11 (8 biological replicates), 5E11 (8 biological replicates), and 2.5E12 (4 biological replicates) vg/mouse of each virion containing the tandem mtZFN configuration, with 4 control animals. Mean ± SEM are displayed, and statistics were performed by one-way ANOVA with Dunnett comparison to control. (C) Pooled and averaged ratios from digitally quantified and normalized northern blot. Statistics were performed using t-test between control and mice treated with 5E11 (8 biological replicates) of the tandem virus. 3 C57BL/6J wild-type mice and four control mice were analyzed. (D) mtCN relative to untreated control samples measured by qPCR. Mean ± SEM are displayed, and statistics were performed using one-way ANOVA using Dunnett’s comparison to the control. Doses shown in increasing order 2E10 (4 biological replicates), 1E11 (4 biological replicates), 2.5E11 (8 biological replicates), 5E11 (8 biological replicates), and 2.5E12 (4 biological replicates) vg/mouse of each virion containing the tandem mtZFN configuration, with 4 control animals. Source data are available online for this figure.