Near-complete elimination of mutant mtDNA by iterative or dynamic dose-controlled treatment with mtZFNs

Abstract

Mitochondrial diseases are frequently associated with mutations in mitochondrial DNA (mtDNA). In most cases, mutant and wild-type mtDNAs coexist, resulting in heteroplasmy. The selective elimination of mutant mtDNA, and consequent enrichment of wild-type mtDNA, can rescue pathological phenotypes in heteroplasmic cells. Use of the mitochondrially targeted zinc finger-nuclease (mtZFN) results in degradation of mutant mtDNA through site-specific DNA cleavage. Here, we describe a substantial enhancement of our previous mtZFN-based approaches to targeting mtDNA, allowing near-complete directional shifts of mtDNA heteroplasmy, either by iterative treatment or through finely controlled expression of mtZFN, which limits off-target catalysis and undesired mtDNA copy number depletion. To demonstrate the utility of this improved approach, we generated an isogenic distribution of heteroplasmic cells with variable mtDNA mutant level from the same parental source without clonal selection. Analysis of these populations demonstrated an altered metabolic signature in cells harbouring decreased levels of mutant m.8993T>G mtDNA, associated with neuropathy, ataxia, and retinitis pigmentosa (NARP). We conclude that mtZFN-based approaches offer means for mtDNA heteroplasmy manipulation in basic research, and may provide a strategy for therapeutic intervention in selected mitochondrial diseases.

Figure 1.

Strategy for targeting the m.8993T>G NARP mutation using mtZFN and analysis of heteroplasmy and copy number upon short-term expression of mtZFNs. (A) Schematic of the general workflow for experiments that involve transient transfection of heteroplasmic cells with plasmids co-expressing mtZFN monomers and fluorescent marker proteins, FACS-based selection of cells expressing both mtZFN monomers and phenotypic evaluation of mtZFN-treated cells. The technical details and plasmids are as described in (27). (B) Detailed schematic of strategy for selective degradation of m.8993T>G mtDNA using mtZFN. Conventional dimeric, engineered mtZFN are directed to sequence adjacent to (COMPa, green) or including (NARPd, red) the mutated base position. Both monomers should bind the substrate only when the indicated nucleotide is mutated and not to the wild-type sequence. DNA double strand breaks should only be introduced into the mutant mtDNA molecule, leading to a shift in heteroplasmy. (C) Last cycle hot PCR restriction fragment length polymorphism (RFLP) analysis of mtDNA from FACS-enriched cells transiently expressing indicated mtZFN construct pairs. 143B and N100 are included as wild-type and 100% mutant digestion controls, respectively. (D) Quantification of RFLP heteroplasmy data from several biological replicates of FACS-enriched cells transiently transfected with mtZFN or control vectors. Data presented are from measurements made at 24 h, 18 days and 28 days post-transfection, as indicated. Statistical analyses were carried out using a two-tailed Student's t-test, P = 0.027 (24h); P = 0.00167 (18 days); P = 0.00146 (28 days); n = 3. ‘% untr. cells" indicates the baseline heteroplasmy of the N80 cell line used in these experiments. (E) Analysis of mtDNA copy number, performed by qPCR in quadruplicate, from samples tested in (C). Error bars = 1 S.D. (F) Wild-type mtDNA heteroplasmy shifts upon iterative expression and recovery cycles of mtZFN pairing NARPd(+)/COMPa(−). Inset; quantification of fold-changes in mtDNA heteroplasmy for mtZFNs and controls for each iteration of transfection/FACS/recovery. Measurements of heteroplasmy presented were taken 28 days post-transfection. ‘% untr. cels’ indicates the baseline heteroplasmy of the N80 cell line used in these experiments.

Figure 2.

Dosage of mtZFN substantially alters the efficiency of heteroplasmy shifting and mtDNA copy number depletion/repletion profiles. (A) Representative dot plot indicating a typical FACS gating strategy used to separate ‘high’ and ‘low’ mtZFN-expressing transfectants. (B) Western blot analysis of total cellular proteins from FACS-enriched samples at 24 h post-transfection by SDS-PAGE, probing for mtZFN expression with antibodies to the HA (NARPd (Nd)) or FLAG (COMPa (Ca)) epitopes. β-actin and a section of Coomassie stained gel (CBB) are shown as loading controls. p, precursor isoform; m, mature isoform. (C) Last cycle hot PCR RFLP of mtDNA from cells transiently expressing ‘high’ or ‘low’ quantities of mtZFN or control vectors. (D) Quantification of collated heteroplasmy data from several biological replicates of transient mtZFN or control vector expression. Data presented are from measurements made at 24 h, 18 days and 28 days post-transfection, as indicated. Statistical analyses were carried out using a two-tailed Student's t-test, P = 0.044 (‘low’ 24 h); P = 0.032 (‘low’ 18 days); P = 0.011 (‘low’ 28 days); n = 3. ‘% untr. cells’ indicates the baseline heteroplasmy of the N80 cell line used in these experiments. (E) Analysis of mtDNA copy number, performed by qPCR in quadruplicate, from samples tested in D. Statistical analyses were carried out using a two-tailed Student's t-test, P = 0.036 (mtZFN, 24 h). Error bars = 1 S.D.

Figure 3.

Dynamic control of mtZFN expression using a tetracycline-sensitive 3′ hammerhead ribozyme (HHR) reveals dosage-dependent effects on heteroplasmy and mtDNA copy number depletion/repletion profiles. (A) Schematic of mtZFN transgene with the hammerhead ribozyme (HHR) incorporated. When transcribed, mRNA encoding mtZFN is constitutively degraded following HHR cleavage, resulting in substantially lower quantities of translated protein. However, in the presence of tetracycline the HHR element is stabilized, inhibiting catalysis and producing larger quantities of translated protein. CMV, cytomegalovirus promoter, MTS, mitochondrial targeting sequence, F, FLAG epitope tag, NES, nuclear export signal, BGH pA, bovine growth hormone polyadenylation signal. (B) Western blot analysis of total cell lysates from FACS-enriched samples at 24 h post-transfection, probing for expression of COMPa(−) mtZFN or COMPa(-)HHR mtZFN (FLAG) in the indicated concentrations of tetracycline (tet) or doxycycline (dox) in the medium. β-actin and a section of Coomassie stained gel (CBB) are shown as loading controls. p, precursor isoform; m, mature isoform. (C) Quantification of collated heteroplasmy data from several biological replicates of transient mtZFN expression. Measurements made at 28 days post-transfection. Filled black triangle indicates spectrum of tetracycline concentrations (0, 25, 250 μM). n = 3, statistical analyses were undertaken using a two-tailed Student's t-test, P = 0.0015 (vector/mtZFN), P = 0.0002 (vector/mtZFN-HHR, 0 μM), P = 0.009 (mtZFN-HHR, 0 μM/mtZFN-HHR, 250 μM). ‘% untr. cells’ indicates the baseline heteroplasmy of the N80 cell line used in these experiments. (D) Analysis of mtDNA copy number, performed by qPCR in quadruplicate, from samples tested in C. Statistical analyses were undertaken using a two-tailed Student's t-test, P = 0.047. Error bars = 1 SD.

Figure 4.

Treatment with mtZFN rescues the mitochondrial defect associated with m.8993T>G and is accompanied by rewiring of cellular metabolism. (A) Oxygen consumption rates (OCR) in empty vector-transfected cybrid cells harbouring 80% of m.8993T>G (N80) and cells of the same origin transfected with NARPd(+)-HHR/COMPa(−)-HHR with either 25 μM (N10) or 250 μM (N45) tetracycline present in the culture medium. These cell lines were generated concurrently within a single experiment. Statistical analysis was undertaken using a two-tailed Student's t-test (P = 0.0001865, n = 7). (B) ATP-linked mitochondrial respiration in N80, N45 and N10 cells, calculated as the ratio of OCR in the presence or absence of oligomycin. (C) Energy charge state analysis of N80, N45 and N10 cells, calculated using values for adenosine phosphate species detection by LC-MS by the established method (35). (D) Principal Component Analysis (PCA) of intracellular metabolites from N80, N45 and N10 cells as measured by LC-MS-based metabolomics. Score plot of principal component 1 and 2, explaining 55.7% and 16.1% of total variance, respectively, is shown. (E) Heatmap representation of intracellular levels of citrate and aconitate in N80, N45 and N10 cells, as measured by LC–MS. Of note, citrate abundance was found to be significantly different between N80 and N45, as well as between N80 and N10, while aconitate was significantly upregulated in N10 compared to N80 cells (Supplementary Figure S8).