Gene Therapy for Mitochondrial DNA Disease

DS Kyriakouli, P Boesch, RW Taylor and RN Lightowlers Mitochondrial Research Group, Medical School, Newcastle University, Newcastle upon Tyne, UK

Defects of the mitochondrial genome cause a wide variety of Although far from the clinic, these approaches show promise. clinical disorders. Except for rare cases where surgery or Similarly, nuclear transfection with genes encoding restric- transplant is indicated, there is no effective treatment for tion endonucleases or sequence-specific zinc finger-binding patients. Genetic-based therapies are consequently being proteins destined for mitochondria has also proved success- considered. On account of the difficulties associated with ful in targeting mtDNA-borne pathogenic mutations. This is mitochondrial (mt) transfection, alternative approaches particularly important, as mutated mtDNA is often found in whereby mitochondrial genes can be engineered and cells that also contain normal copies of the genome, a introduced into the nucleus (allotopic expression) are being situation termed heteroplasmy. Shifting the levels of hetero- attempted with some success, at least in cultured cells. plasmy towards the normal mtDNA has become the goal of a Defects in the activities of multi-subunit complexes variety of invasive and non-invasive methods, which are also of the oxidative phosphorylation apparatus have been highlighted in this review. circumvented by the targeted expression of simple single Gene Therapy (2008) 15, 1017–1023; doi:10.1038/gt.2008.91; subunit enzymes from other species (xenotopic expression). published online 22 May 2008

Keywords: mtDNA; heteroplasmy; mitochondria; mtDNA disease

In brief Progress Prospects Strategies can be devised for gene therapy of Generation of numerous mice models heteroplasmic mitochondrial DNA disorders. for pathogenic mtDNA mutations that can routinely Allotopic and xenotopic expression show great be used to determine efﬁcacy of any therapeutic. promise as treatment concepts for mtDNA disease. Although such mice do exist, there are currently very The alternative oxidase may prove an attractive few. It is expected that within 5 years, there will be alternative for treating mtDNA disease. many models that can be used. The yeast soluble NADH oxidase can substitute for Determining the feasibility of using genes encoding defective human complex I. single enzyme alternatives to OXPHOS complexes as Allotopic expression and import of nucleic acids are possible interventions for mutations that affect being considered for rescuing pathogenic mt-tRNA protein-coding genes in mtDNA. mutations. Production of antigenomic drugs that will allow Rescue of mtDNA mutations may be possible targeting and removal of pathogenic mtDNA. Cur- through mitochondrial transfection. rent work is promising and it is likely that within 3–5 Manipulating the mitochondrial genome may lead to years a pharmaceutical could be developed for oral or rescue of an OXPHOS deﬁciency. intravenous delivery that will be able to modulate Germline therapy for mitochondrial DNA disease is levels of heteroplasmy. being considered.

Introduction matrix and encodes 13 polypeptides, all of which are The human mitochondrial genome (mtDNA) is found in integral components of the complexes that couple many copies in all nucleated cells (Figure 1). This small oxidative phosphorylation (OXPHOS). In addition to (16 569 bp) genome is housed in the mitochondrial these polypeptides, this remarkably compact molecule codes for the two ribosomal RNAs (16S and 12S mt- rRNA) and 22 tRNAs that are necessary and sufficient Correspondence: Professor RN Lightowlers, Mitochondrial for intramitochondrial protein synthesis. As all 13 Research Group, Newcastle University, The Medical School, mitochondrially encoded polypeptides are essential Framlington Place, Newcastle upon Tyne, Newcastle NE2 4HH, UK. E-mail: [email protected] members of the OXPHOS machinery that harnesses Received 13 February 2008; revised 11 April 2008; accepted 14 April cellular respiration to ATP production, it is not surpris- 2008; published online 22 May 2008 ing that mutations in this genome can cause a wide Gene therapy for mitochondrial DNA disease DS Kyriakouli et al 1018 the field to consider the prospect of gene therapy and related concepts. The approaches can roughly be divided into three groups: (1) rescue of a defect by expression of an engineered gene product from the nucleus (allotopic or xenotopic expression), (2) import of normal copies or relevant sections of mtDNA into mitochondria, and (3) manipulation of the mitochondrial genetic content. We briefly introduce these concepts below and indicate where significant progress has been made in the last 2 years.

Allotopic and xenotopic expression show great promise as treatment concepts for mtDNA disease In 1988, Nagley and colleagues2 were able to show that the respiratory defect in yeast carrying mutations in the mitochondrial MTATP8 gene could be rescued if a normal copy of the gene was engineered and introduced into the nucleus, resulting in cytosolic translation of the gene product, ATPase 8. Phenotypic rescue required the addition of a well-deﬁned mitochondrial targeting and import sequence at its N-terminal. This ‘allotopically’ expressed mitochondrial gene product could be func- Figure 1 The human mitochondrial genome. This small (16 569 bp) genome is almost completely transcribed from both strands, initiating tionally integrated into the mitochondrial ATP synthase

from one of two promoters (IH1,IH2) on the Heavy (H-) strand or the (complex V) and the deﬁciency could be rescued. Several

single promoter (IL) on the Light (L-) strand. All of these promoters years later, this finding was extended to cultured human and elements involved in replication initiation are found in the cells. A different mitochondrial gene to encode a member displacement (D-) loop, the only major non-coding region. The of the ATP synthase, MTATP6, was also engineered for genome encodes 22 mt-tRNA (black diamonds), 2 mt-rRNA genes expression in the nucleus. Successful targeting of the (fuscia) and 13 protein-coding genes (olive, ND1-6 encoding members of NADH:ubiquinone oxido-reductase; blue, Cytb encoding ATPase 6 gene product to the mitochondrion was shown apocytochrome b of ubiquinol:cytochrome c oxido-reductase; orange, to partially suppress a respiratory-deficient phenotype COI-III encoding members of cytochrome c oxidase: aquamarine, caused by a pathogenic mutation in MTATP6. These ATPase 6, 8 encoding two members of Fo-F1 ATP synthase). experiments were met with optimism as a potential Important mutations that are referred to in this article are indicated. method for treating patients with mutations in mito- Single letter code is given for each mt-tRNA-encoding gene. chondrial protein-encoding genes such as those with neurogenic weakness, ataxia with retinitis pigmentosum variety of disorders such as mitochondrial encephalo- (NARP) which are caused by a specific point mutation pathy lactic acidosis with stroke-like episodes (MELAS), (m.8993T4G) in MTATP6. More recent and extensive myoclonic epilepsy with ragged red fibres (MERRF) and analysis by Bokori-Brown and Holt conflict with these leber’s hereditary optic neuropathy (LHON).1 In addi- data.3 Similar allotopic experiments were performed to tion, due to the ubiquity of mitochondria, the clinical rescue a cell line carrying the identical mutation in presentation of mitochondrial DNA disorders can MTATP6. The authors optimised the import of nuclear- be multi-system, but often causes profound chronic encoded ATPase 6 into the human cell lines but after progressive defects in muscle and nerve. careful analysis of assembled complexes were unable to Cells are multiploid with respect to mtDNA, with show integration of the imported protein into mature most somatic cells containing at least 103 copies of the functional ATP synthase. Restoration of complex V genome. Consequently, a pathogenic mutation may be activity in the original experiments was suggested to present in just one, many, or all, depending on when the have been due to random clonal variations in ATP mutation occurred and how the mutated molecule was synthesis as a consequence of aneuploidy in the replicated and segregated during development. For most transfected population. Despite these discrepancies, the disorders, the pathogenic mutation is recessive, requiring authors remained optimistic about the potential for a substantial percentage of the genome to carry the defect allotopic expression as a treatment method. before an effect on OXPHOS can be measured. This In a twist to this approach, Corral-Debrinski et al.4 ‘threshold’ varies between cell types and mutation sites have attempted to optimise the allotopic expression of but is normally between 60 and 95%. such highly hydrophobic mitochondrial proteins by suggesting that targeting of the transcript to the outer mitochondrial membrane may facilitate co-translational Strategies can be devised for gene translocation of the gene product, thus preventing the therapy of mitochondrial DNA disorders accumulation of cytosolic aggregates. Further evidence was generated by using primary fibroblasts from patients There is currently no effective treatment for the majority with mutations in MTND4 (a mitochondrial gene of these debilitating diseases. This has led researchers in encoding a member of complex I of the respiratory

Gene Therapy Gene therapy for mitochondrial DNA disease DS Kyriakouli et al 1019 chain) known to cause Leber’s hereditary optic neuro- termed as Ndi1. Back in 2001, Yagi and colleagues7 pathy, thereby avoiding the complexities of aneuploid showed that mitochondrial targeting of Ndi1 in human immortalised cell lines.5 Cis-acting elements in the cells defective in complex I activity corrected this defect. 30-untranslated regions from transcripts that had pre- In the intervening years, the authors have built on this viously been shown to localise to mitochondria (SOD2 observation and showed that in addition to the potential and COX10) were used in tandem with in-frame for Ndi1 activity to treat patients with mtDNA disease presequence coding regions to improve mitochondrial associated with mtDNA-borne MTND mutations, this import of the allotopically expressed proteins. Transfe- enzyme may prove useful for other complex I disorders. ctions resulted in the rescue of the deficiency as evidenced In particular, the authors have focussed on complex I by the restoration of OXPHOS and growth with galactose deficiency in Parkinson’s disease (PD). That the halluci- as a sole carbon source, a medium that requires efficient nogen MPTP (1-methyl-4-phenyl-1,2,3,6-tetrahydropyri- OXPHOS for cell proliferation. Although impressive, dine) causes a form of PD has been well described. This additional analysis by gel electrophoresis to show drug selectively inhibits complex I activity in humans, unequivocally that imported proteins are indeed inte- but has no effect on yeast Ndi1.7 Yagi and colleagues8 grated into fully assembled OXPHOS complexes is still have now demonstrated that viral-mediated delivery required. and expression of Ndi1 in the substantia nigra protected against neurodegeneration after feeding rats with MPTP. These data tentatively support the use of yeast Ndi1 The alternative oxidase may prove expression to compensate for any complex I deficiency in patients with PD. an attractive alternative for treating mtDNA disease Allotopic expression is an encouraging strategy, but does Allotopic expression and import require efficient mitochondrial import and correct of nucleic acids are being considered for integration of the allotropic protein into an assembled protein complex. Two entirely separate but equally rescuing pathogenic mt-tRNA mutations exciting prospects have received attention in the last 2 It has been a dramatic few years for reports of nucleic years. Human mitochondria rely on a single cyanide- acid import into mammalian mitochondria, any of which sensitive cytochrome c oxidase (COX) to reduce mole- may eventually prove beneficial for genetic therapies of cular oxygen. This multi-component complex is the mitochondrial disorders. In contrast to many eukaryotes, terminal member of the mitochondrial electron transfer mammals encode all tRNAs required for mitochondrial chain and contains three essential polypeptides encoded protein synthesis on the mitochondrial genome (mt- by mtDNA. Consequently, mtDNA mutations often tRNA). As many disorders of the mitochondrial genome cause defects in COX activity. Importantly, many other are caused by mutations in mt-tRNA encoding genes, species encode cyanide-insensitive alternative oxidases there has been great interest in rescuing these mutations (AOX). A very encouraging report recently showed that a by importing normal copies of the relevant tRNA from single subunit AOX from the sea squirt Ciona intestinalis the cytosol. could be synthesised in the cytosol of cultured human Tarassov et al.9 showed that a respiratory defect in cell cells and successfully targeted to the mitochondrial inner Lys 6 lines containing mutated mt-tRNA could be partially membrane. AOX caused no measurable cytotoxicity and rescued when the yeast homologue was expressed in the transfectants showed impressive levels of cyanide- nucleus. These results suggested that a cryptic tRNA insensitive oxidation of mitochondrial metabolites. The import system remains in human mitochondria. The authors also highlighted its important function as an exact tRNA import mechanism in human mitochondria anti-oxidant, preventing over-reduction of the ubiqui- is unclear, although the authors have begun to dissect the none pool, which is believed to be an important source of pathway in yeast mitochondria. Building on the theme of reactive oxygen species. To date, there has been no report tRNA import as a potential mechanism for treating of its use in rescuing a mitochondrial COX defect, but the mtDNA disorders, Mukherjee et al.10 have exploited the potential for its use in treating human COX deficiencies highly prolific tRNA import process in protists. Initially, is clear. a large tRNA import complex (RIC) was isolated from the inner mitochondrial membrane of Leishmania tropica and reconstituted into phospholipid vesicles. The com- The yeast-soluble NADH oxidase can plex comprised 11 major polypeptides, three of which substitute for defective complex I were mitochondrially encoded and five of the eight nuclear gene products were also members of the A similar and equally elegant approach has also been OXPHOS machinery. Remarkably, bathing cultured hu- used to by-pass defects in the initial part of the electron man cells in the purified RIC led to receptor-mediated transfer chain. Human complex I, NADH:ubiquinone import of the complex into the cytoplasm and integration oxidoreductase is a profoundly complicated assembly of into mitochondria via a mechanism that is currently approximately 45 proteins, 7 of which are encoded by the unclear.11 Compelling evidence of cytosolic tRNA import mitochondrial genome. Numerous mitochondrial defects comes from work with a cell line containing mtDNA have been linked with mutations in these mitochondrial with a large-scale deletion that removes various genes genes. In contrast, the yeast Saccharomyces cerevisiae does including MTTK, which encodes mt-tRNALys (Figure 1). not possess a similar membrane-bound complex, but Prior to incubation of cells with RIC, there was no instead, uses just a single subunit NADH oxidase, evidence of any mitochondrial protein synthesis, which

Gene Therapy Gene therapy for mitochondrial DNA disease DS Kyriakouli et al 1020 can normally be visualised after selective poisoning of This process of natural competence is not at all under- cytosolic protein synthesis. Following RIC addition, stood but has recently been used by Weissig and protein translation was restored for gene products whose colleagues to load isolated mitochondria that were genes were not removed by the deletion. This restoration subsequently shown to be engulfed by cells when co- is likely to be due to imported cytosolic tRNALys incubated.14 Uptake of isolated murine mitochondria had functioning in mitochondrial translation. initially been shown by the authors to restore respiration It is remarkable that the large RIC complex can be on fusion with human A549 lung carcinoma cells devoid transported across the plasma membrane and the outer of mtDNA. This internalisation, similar to the method mitochondrial membrane and integrated into the inner described over 25 years ago by Clark and Shay may membrane. Work with RIC has now been extended to prove to be a powerful technique for studying mitochon- show that complexes preloaded with antisense RNA drial gene expression in humans but is unlikely to be designed against various mt-mRNA can successfully immediately relevant to gene therapy for mtDNA transport these RNAs to the mitochondrial matrix of disease. intact cells in culture. This leads not only to selective Are there any other methods being investigated for impairment of protein synthesis from the mt-mRNA transfecting mitochondria in whole cells? Several years target, but also degradation of this mt-mRNA.12 This is a ago, there was substantial interest in the use of very striking result as it is consistent with an RNA- DQasomes. These nonviral mitochondriotropic liposome mediated targeting and degradation pathway similar to derivatives could be loaded with DNA and had been that noted in the cytosol. The latter is mediated by the shown to co-localise with mitochondria when added to RISC complex and dicer for which there is no evidence of cultured cells. Unfortunately, these earlier reports have mitochondrial localisation. Clearly, this work from the not been greatly extended in the past 2 years. Another laboratory of Adhya has quite profound consequences liposome-based carrier, MITO-Porter has also been for the field of mammalian mitochondrial transfection shown to co-localise to mitochondria in intact cells.15 and would benefit enormously from confirmatory MITO-Porter loaded with GFP was shown in isolated experiments from independent laboratories. mitochondria to be delivered to the intermembrane space and to co-localise with mitochondria in cultured cells. These data suggest that such liposomal preparations can fuse with the outer mitochondrial membrane, but will it Rescue of mtDNA mutations prove to be able to access the inner membrane that is may be possible through mitochondrial necessary for the cargo to access the mitochondrial transfection matrix? Another approach to mitochondrial transfection is protofection. To our knowledge, however, details of When a pathogenic nuclear mutation has been identified, this protein-mediated method for delivering complete one approach may be to re-introduce normal copies of mitochondrial genomes to mitochondria have not ap- the gene that might lead to rescue of the phenotype. peared as an original article in any peer-reviewed There are major obstacles in using this strategy to journal, so it is difficult to assess and to determine its overcome mitochondrial gene mutations. First, transfec- potential for mediating gene therapy. tion is difficult, as there are effectively three membrane barriers. Second, even assuming that exogenous DNA can be imported, gene expression will be transient. Stable integration may be mediated by homologous Manipulating the mitochondrial genome recombination. There is evidence that such recombina- may lead to rescue of an OXPHOS tion does occur in some somatic cells at low frequency, deficiency but it appears to be a rare event. Consequently, any imported DNA must contain the correct cis-acting Many patients that suffer from mtDNA disease are elements to promote DNA replication, resolution and heteroplasmic, harbouring two subpopulations, one maintenance. To date, these elements are not fully normal and one pathogenic. One approach to genetic characterised and even the detailed mechanism of therapy for these intractable disorders is to manipulate mtDNA replication itself is controversial. Finally, assum- the relative amounts of these two subpopulations. This is ing transfection and maintenance, there are still many particularly attractive, as most heteroplasmic pathogenic hundreds or thousands of mtDNA molecules that will mutations are recessive, such that levels do not have to outweigh the transfected molecule. How can it be be altered greatly before biochemical defects can be selected for and maintained in such a background? rescued. Which methods are currently being investigated What progress has been made towards overcoming for manipulating levels of heteroplasmy (Figure 2)? these barriers? It has been known for several years that Exercise training, either by endurance training to isolated plant mitochondria can spontaneously import a promote mitochondrial biogenesis or by resistance low level of naked linearised DNA. This observation has exercise protocols to promote muscle satellite cell now been extended to mammalian mitochondria, which activation has long been proposed as a means of treating also show a low efficiency of natural DNA import.13 patients with high levels of heteroplasmic mtDNA Unlike previous anecdotal reports of DNA uptake, data mutations in muscle. Endurance training improves showed that DNA could be imported and could act as a exercise capacity, inducing physiological adaptations, template for DNA synthesis. In addition, transcription which reverse the effects of muscle deconditioning and was shown to occur only when the template carried a lead to a resolution of the underlying mitochondrial transcriptional promoter element and the primary biochemical defect. Nevertheless, concerns have been transcript could be successfully processed and matured. raised as to whether endurance training has potentially

Gene Therapy Gene therapy for mitochondrial DNA disease DS Kyriakouli et al 1021 muscle fibres, but low or undetectable levels in muscle satellite cells. The stimulation of muscle regeneration, and concomitant incorporation of satellite cell mtDNA into mature muscle (gene shifting), is therefore an attractive approach to manipulate mtDNA heteroplasmy levels in vivo. Indeed, anaesthetic injection, traumatic biopsy injury and resistance exercise (both eccentric and concentric) training protocols have all been shown to be of benefit in inducing muscle regeneration in individual patients. Ongoing studies in patients with single, large- scale mtDNA deletions suggest that 12 weeks of resistance training induces an increase in satellite cell number and muscle regeneration, leading to improve- ments in strength and oxidative capacity as reflected by a decrease in the percentage of muscle fibres deficient in mitochondrial COX activity.17 An elegant approach to removing pathogenic mtDNA is by targeting restriction endonucleases to mitochondria Figure 2 Manipulating mtDNA heteroplasmy. An approach to of heteroplasmic cells. First pioneered by Carlos Moraes treating patients who suffer from an mtDNA disorder where both in rodent and Masashi Tanaka in human cell lines, pathogenic and wild type mtDNA co-exist (heteroplasmy) is to specific mtDNA subpopulations in a heteroplasmic manipulate these levels with the aim of increasing the wild type at background can be degraded by endonucleases ex- the expense of the pathogenic mtDNA, potentially reversing the biochemical defect. In this figure, heteroplasmic mitochondria pressed from introduced genes engineered to generate containing both wild type (grey) and pathogenic (black) mtDNA enzymes with mitochondrial targeting sequences. Het- are schematically represented. Transverse muscle sections illustrate eroplasmy was shown to shift substantially towards the how high levels of mutated mtDNA in heteroplasmic patients can genome lacking the restriction site. This was potentially result in mosaic muscle fibre staining for cytochrome c oxidase of direct relevance to disease, as the heteroplasmic activity (left panel), a situation that can potentially be reversed by m.8993T4C mutation was shown to be markedly increasing the level of wild type mtDNA (right panel). Several different methods to modulate heteroplasmy are currently being reduced on targeting of the restriction site by endonu- explored. (A) Patients are being subjected to endurance or resistance clease SmaI. It could be envisaged that the engineered training, increasing muscle bulk and promoting growth of new gene encoding the targeted endonuclease could be muscle. (B) Targeting restriction endonucleases to mitochondria has delivered to patients, but the approach would appear already been successfully employed in tissue culture to remove to be limited to the very few human pathogenic mtDNA pathogenic mtDNA where the mutation generates a specific mutations that result in a restriction site unique to restriction site. This approach has also been used to modulate levels of heteroplasmy for polymorphic mtDNA in mice. (C) Cell the pathogenic mtDNA. Moraes and colleagues membrane crossing oligomers (CMCOs) are being optimised to have, however, shown that the generated restriction penetrate cells, target mitochondria and selectively bind to mutated site need not be unique.18 Heteroplasmic mice carrying mtDNA. These molecules are potentially capable of inhibiting two distinct genomes were subjected to adenoviral- mtDNA replication, although this has yet to be demonstrated. (D) mediated delivery of a gene encoding the targeted Zinc-finger binding proteins that show sequence binding specificity endonuclease ScaI. The two genomes each carried have been targeted to mitochondria. By fusing the ZFB chimaera to a DNA methylase, it was shown that mtDNA carrying a specific multiple ScaI sites, five in one genome and three in the mutation can be selectively methylated. It is postulated that ZFB second genome. Heteroplasmy was shown to shift in the proteins could be fused to DNases, leading to selective cleavage of liver and skeletal muscle, although high-level expression mutated mtDNA. did lead to mtDNA depletion. These data, however, are promising and suggest that if expression levels are controlled it may be possible to treat patients who have long-term deleterious effects, given that the mutated pathogenic mtDNA mutations that create additional mtDNA load in some patients was noted to restriction sites. increase following training. To assess the safety of A concept that received wide coverage several years endurance training, Taivassalo et al.16 studied the ago was the possibility of antigenomic therapy. This effects of 14 weeks of exercise in a group of patients required the design and use of a drug that could be with single, large-scale mtDNA deletions. Markers imported into mitochondria, bind selectively to the of physiological improvement (oxygen utilisation, mutated mtDNA and inhibit it from replicating, allowing skeletal muscle oxygen extraction, peak capacity the normal copy to propagate with time, and thus rescue for work and submaximal exercise tolerance) were the defective OXPHOS phenotype. But what could this evident, although genetic analysis did not reveal any antigenomic molecule be? At first, there was great change in the level of deleted mtDNA following exercise interest in the use of mitochondrially targeted peptide training. Further training for an additional 14 weeks nucleic acids (PNAs). These molecules still retained the maintained these benefits, whereas detraining resulted in ability to Watson:Crick base pair and showed remarkable a loss of the physiological adaptations. Follow-up studies binding selectivity even at physiological salt concentra- to assess whether long-term training might induce tions and temperatures. Initial experiments were en- changes in mtDNA mutation load are currently in couraging but eventually it was determined that the progress. targeted PNAs although co-localised to mitochondria in Some patients with sporadic mtDNA mutations whole cells did not get across the inner membrane, and harbour high levels of mutated mtDNA in mature so could not access mtDNA. Modified molecules termed

Gene Therapy Gene therapy for mitochondrial DNA disease DS Kyriakouli et al 1022 as ‘cell membrane crossing oligomers’ (CMCO) have Future prospects for genetic therapy of mtDNA now been designed, with greater polarity than traditional disease PNAs. It has now proved possible to target and import Considering the recent output, it is clear that scientists are CMCO:PNA hybrid molecules into mitochondria in now claiming a variety of methods for transfecting whole cells. Although these new molecules have great mitochondria. Although impressive, it is difficult to see promise as potential drugs for treating heteroplasmic how this work will be of direct relevance to treatment patients, many important experiments need to be unless the transfected material can be autonomously performed to determine whether they can inhibit maintained. Allotopic expression and the use of simple mtDNA replication in whole cells and be delivered to single-subunit enzymes that can potentially replace the defective mitochondria in animals. activity of OXPHOS complexes are both areas of great The possibility of selective targeting of mtDNA potential. The drawback is that they will require the 19 has also been explored by Minczuk et al. These delivery of foreign genes to affected cells, which is a similar authors have attempted to use the short zinc-finger problem to many conventional gene therapy approaches. peptides that show sequence-specific DNA binding Manipulating heteroplasmy remains a particularly attrac- capabilities. The authors generated chimaeric tive idea, especially if the intervention is non-invasive or is ZFPs linked to a DNA methylase carrying an additional mediated by a pharmaceutical agent. One area that has mitochondrial presequence. Following import, chi- been resurrected recently and shows great promise is the maeras were shown to bind mtDNA and to leave a use of CMCO-based molecules that can enter cells and signature methylation pattern at various sites on the mitochondria naturally and initial experiments suggest mitochondrial genome, a genome that is not normally these molecules can influence heteroplasmy and mitochon- methylated. In addition, by using the NARP m.8993T4G drial gene expression. Before any of these therapeutic cell line, popular for many of these studies, the approaches can be used in clinical trials, it will be necessary authors showed methylation patterns that differed to show their efficacy in animal models. Unfortunately, it from the wild-type mtDNA control. By linking these has proven difficult to establish mice that are heteroplasmic domains to a nuclease it may be possible to target a for pathogenic mutations. To move forward, it is important mutated subpopulation in a heteroplasmic patient. This that more models become available. is an exciting and impressive achievement, but its long-term usage will be limited by the requirement for transfection and successful expression in defective cells and tissues. References

1 McFarland R, Taylor RW, Turnbull DM. Mitochondrial disease— its impact, aetiology, and pathology. Curr Top Dev Biol 2007; 77: 113–155. Germline therapy for mitochondrial DNA 2 Nagley P, Farrell LB, Gearing DP, Nero D, Meltzer S, Devenish RJ. disease is being considered Assembly of functional proton-translocating ATPase complex in yeast mitochondria with cytoplasmically synthesized subunit 8, a Patients with mtDNA disease often wish to have polypeptide normally encoded within the organelle. Proc Natl children. However, the knowledge that mtDNA disease Acad Sci USA 1988; 85: 2091–2095. is a genetic disorder and can be transmitted down the 3 Bokori-Brown M, Holt IJ. Expression of algal nuclear ATP maternal line can preclude this decision. Would it be synthase subunit 6 in human cells results in protein targeting to possible to generate a zygote from the parents’ gametes mitochondria but no assembly into ATP synthase. Rejuvenation using standard in vitro fertilisation techniques and then Res 2006; 9: 455–469. replace the defective mitochondrial genetic complement? 4 Kaltimbacher V, Bonnet C, Lecoeuvre G, Forster V, Sahel JA, A single-cell zygote containing normal copies of mtDNA Corral-Debrinski M. mRNA localization to the mitochondrial would need to be enucleated to produce a cytoplast. An surface allows the efﬁcient translocation inside the organelle of a egg from the patient would then be fertilised in vitro with nuclear recoded ATP6 protein. RNA 2006; 12: 1408–1417. the sperm from the partner. Finally, an embryo would 5 Bonnet C, Kaltimbacher V, Ellouze S, Augustin S, Benit P, Forster V have to be reconstructed from pronuclei carefully et al. Allotopic mRNA localization to the mitochondrial surface removed from the patients fertilised oocyte and fused rescues respiratory chain defects in ﬁbroblasts harboring to the cytoplast.20 Studies in mice have indeed suggested mitochondrial DNA mutations affecting complex I or V that it may be possible to prevent transmission of subunits. Rejuvenation Res 2007; 10: 127–144. mutated mtDNA by pronuclear transfer. Following 6 Hakkaart GA, Dassa EP, Jacobs HT, Rustin P. Allotopic expression reconstruction of such a murine embryo, fused zygotes of a mitochondrial alternative oxidase confers cyanide resistance to were transferred to the oviducts of pseudopregnant human cell respiration. EMBO Rep 2006; 7: 341–345. females and offspring were shown to be normal.21 Could 7 Seo BB, Nakamaru-Ogiso E, Flotte TR, Matsuno-Yagi A, Yagi T. In vivo complementation of complex I by the yeast Ndi1 enzyme. such an approach be possible in humans? Following the Possible application for treatment of Parkinson disease. J Biol decision in 2005 by the UK Human Fertilisation and Chem 2006; 281: 14250–14255. Embryology Authority to approve research into the 8 Marella M, Seo BB, Nakamaru-Ogiso E, Greenamyre JT, feasibility of this approach in humans, experiments have Matsuno-Yagi A, Yagi T. Protection by the NDI1 gene against been initiated using abnormally fertilised embryos. neurodegeneration in a rotenone rat model of Parkinson’s Although results of these studies have yet to be disease. PLoS ONE 2008; 3: e1433. published, this approach may bring hope to patients 9 Tarassov I, Kamenski P, Kolesnikova O, Karicheva O, Martin RP, with mtDNA disorders who wish to have children and Krasheninnikov IA et al. Import of nuclear DNA-encoded RNAs for whom oocyte donation is either not possible or not into mitochondria and mitochondrial translation. Cell Cycle 2007; desired. 6: 2473–2477.

Gene Therapy Gene therapy for mitochondrial DNA disease DS Kyriakouli et al 1023 10 Mukherjee S, Basu S, Home P, Dhar G, Adhya S. Necessary 16 Taivassalo T, Gardner JL, Taylor RW, Schaefer AM, Newman J, and sufficient factors for the import of transfer RNA Barron MJ et al. Endurance training and detraining in mitochon- into the kinetoplast mitochondrion. EMBO Rep 2007; 8: drial myopathies due to single large-scale mtDNA deletions. 589–595. Brain 2006; 129 (Part 12): 3391–3401. 11 Mahata B, Mukherjee S, Mishra S, Bandyopadhyay A, Adhya S. 17 Taivassalo T, Gardner JL, Taylor RW, Schaefer AM, Haller RG, Functional delivery of a cytosolic tRNA into mutant mitochon- Turnbull DM. Resistance exercise training in mitochondrial dria of human cells. Science 2006; 314: 471–474. myopathy due to single, large-scale mtDNA deletions: implica- 12 Mukherjee S, Mahata B, Mahato B, Adhya S. Targeted mRNA tions for therapy. Neuromuscul Disord 2007; 17: 829. degradation by complex-mediated delivery of antisense RNAs 18 Bacman SR, Williams SL, Hernandez D, Moraes CT. Modulating to intracellular human mitochondria. Hum Mol Genet 2008; 17: mtDNA heteroplasmy by mitochondria-targeted restriction 1292–1298. endonucleases in a ‘differential multiple cleavage-site’ model. 13 Koulintchenko M, Temperley RJ, Mason PA, Dietrich A, Light- Gene Ther 2007; 14: 1309–1318. owlers RN. Natural competence of mammalian mitochondria 19 Minczuk M, Papworth MA, Kolasinska P, Murphy MP, Klug A. allows the molecular investigation of mitochondrial gene Sequence-specific modification of mitochondrial DNA using a expression. Hum Mol Genet 2006; 15: 143–154. chimeric zinc finger methylase. Proc Natl Acad Sci USA 2006; 103: 14 Katrangi E, D’Souza G, Boddapati SV, Kulawiec M, Singh KK, 19689–19694. Bigger B et al. Xenogenic transfer of isolated murine mitochon- 20 Brown DT, Herbert M, Lamb VK, Chinnery PF, Taylor RW, dria into human rho(0) cells can improve respiratory function. Lightowlers RN et al. Transmission of mitochondrial DNA Rejuvenation Res 2007; 10: 561–570. disorders: possibilities for the future. Lancet 2006; 368: 87–89. 15 Yamada Y, Akita H, Kamiya H, Kogure K, Yamamoto T, 21 Sato A, Kono T, Nakada K, Ishikawa K, Inoue S, Yonekawa H Shinohara Y et al. MITO-Porter: a liposome-based carrier system et al. Gene therapy for progeny of mito-mice carrying pathogenic for delivery of macromolecules into mitochondria via membrane mtDNA by nuclear transplantation. Proc Natl Acad Sci USA 2005; fusion. Biochim Biophys Acta 2008; 1778: 423–432. 102: 16765–16770.

Gene Therapy