Reproductive BioMedicine Online (2013) 27, 599– 610

www.sciencedirect.com www.rbmonline.com

SYMPOSIUM: FUTURES IN REPRODUCTION REVIEW

Preventing the transmission of mitochondrial DNA disorders: Selecting the good guys or kicking out the bad guys Hubert JM Smeets Unit Clinical Genomics, Department of Genetics and Cell Biology, School for Growth and Development and for Cardiovascular Research, Maastricht University Medical Centre, Maastricht, The Netherlands E-mail address: [email protected] Bert Smeets, PhD, is Professor in Clinical Genomics with a focus on mitochondrial disorders at Maastricht UMC, The Netherlands, combining research with genetic testing services. His research focuses on the genomics of mitochondrial disorders and involves identifying the genetic defect, studying the pathophysiology, characterizing new treatment options and preventing the transmission, the latter by preimplantation genetic diagnosis. He has published over 170 original research articles, reviews and book chapters (Hirsch index = 40). His group contains 60 people involved in clinical genomics research and services and exploits central genomics facilities for Maastricht UMC and adjoining universities from Belgium and Germany.

Abstract Mitochondrial disorders represent the most common group of inborn errors of metabolism. Clinical manifestations can be

extremely variable, ranging from single affected tissues to multisystemic syndromes. Maternally inherited mitochondrial DNA (mtDNA) mutations are a frequent cause, affecting about one in 5000 individuals. The expression of mtDNA mutations differs from nuclear gene defects. Mutations are either homoplasmic or heteroplasmic, and in the latter case disease becomes manifest when the mutation load exceeds a tissue-specific threshold. Mutation load can vary between tissues and in time, and often an exact correlation between mutation load and clinical manifestations is lacking. Because of the possible clinical severity, the lack of treatment and the high recurrence risk of affected offspring for female carriers, couples request prevention of transmission of mtDNA mutations. Previously, choices have been limited due to a segregational bottleneck, which makes the mtDNA mutation load in embryos highly variable and the consequences largely unpredictable. However, recently it was shown that preimplantation genetic diagnosis offers a fair chance of unaffected offspring to carriers of heteroplasmic mtDNA mutations. Technically and ethically challenging possibilities, such maternal spindle transfer and pronuclear transfer, are emerging and providing carriers additional prospects of giving birth to a healthy child. RBMOnline ª 2013, Reproductive Healthcare Ltd. Published by Elsevier Ltd. All rights reserved. KEYWORDS: chromosome–spindle transfer, mtDNA disease, nuclear genome transfer, preimplantation genetic diagnosis VIDEO LINK: http://sms.cam.ac.uk/media/1400848

Introduction Mitochondrial or oxidative phosphorylation disorders are complex diseases, caused by mutations in either nuclear

genes or in the mitochondrial DNA (mtDNA). Nuclear gene defects segregate as Mendelian diseases, whereas mtDNA defects are transmitted maternally. The latter occurs in about 15% of the cases, affecting about one in 5000

1472-6483/$ - see front matter ª 2013, Reproductive Healthcare Ltd. Published by Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.rbmo.2013.08.007

600 individuals (Rotig and Munnich, 2003). Carrier frequency for pathogenic mtDNA mutations in the population is higher – from 1:400 (Manwaring et al., 2007) to even >1:200 (Elliott et al., 2008) – but in general the mutation load remains below the level of clinical expression. Still, symptoms such as hearing loss can be present and undiagnosed individuals can turn out to be oligosymptomatic upon further investigation (Manwaring et al., 2007). Mitochondrial diseases can manifest with symptoms in many different organs and vary profoundly in severity and age of onset (reviewed in McFarland et al., 2010). Clinical manifestations may present in just a single affected tissue or organ, such as the loss of vision in Leber’s hereditary optic neuropathy (LHON), but a multisystemic or multiorgan involvement is more common. The clinical spectrum (Chinnery and Hudson, 2013) involves the brain (ataxia, dementia, migraine, myoclonus, neuronal loss and stroke), the peripheral nervous system (neuropathy), the heart (cardiomyopathy, conduction disorders, Wolfe-Parkinson-White syndrome), skeletal muscle (fatigue, myopathy, weakness), the liver (hepatopathy), the pancreas (diabetes), the eyes (optic neuropathy, ophthalmoplegia, retinopathy), the ears (sensori-neuronal hearing loss), the kidney (Fanconi syndrome, glomerulopathy), the colon (pseudo-obstruction), the blood (Pearson syndrome) and the gonads (ovarian failure). Well-known neurological syndromes, caused by mitochondrial dysfunction and partly due to mtDNA mutations, are Leigh syndrome (subacute necrotizing encephalomyelopathy), mitochondrial encephalopathy, lactic acidosis and stroke-like episodes (MELAS syndrome), neuropathy, ataxia and retinitis pigmentosa (NARP syndrome) and myoclonic epilepsy with ragged red fibres (MERRF syndrome). Fatally affected newborns represent the severe end of the spectrum. A frequent symptom in paediatric patients is developmental delay and failure to thrive. When at least two organ systems unexplained by other diseases are involved in a single person or in affected (maternal) relatives, then mitochondrial disorder must be considered. Clinicians should be aware that apparently unrelated symptoms might have a common genetic cause (McFarland et al., 2010). Given the potential for severe clinical disease in a child, the ability to prevent transmission of these inherited disorders using reproductive technology is highly desirable.

Mitochondrial DNA The first description of a circular DNA located in the mitochondria dates from more than 40 years ago (Nass, 1966). The mtDNA has a number of unique characteristics that discriminates it from its nuclear counterpart. First, the mtDNA is a double-stranded circle (16,569 bp) with a structure and code different from the nuclear DNA. The mtDNA contains 37 genes, of which 13 genes encode OXPHOS subunits and 22 tRNA and two rRNA genes. Approximately 6% of the mtDNA is noncoding, located predominantly in the D-loop and involved in the replication and transcription of the mtDNA (Anderson et al., 1981). The mtDNA is compact. It contains no introns, several overlapping genes and incomplete termination codons. It mutates somatically during life as a result of reactive oxygen species, produced by the OXPHOS system, and through age-related damage.

HJM Smeets Next, the mtDNA is not a diploid but a multicopy genome. A cell contains hundreds of mitochondria and, dependent on the tissue involved, each cell carries between 500–10,000 mtDNA molecules, except for mature oocytes which have between 100,000 and 600,000 mtDNA molecules (Reynier et al., 2001). The higher the energy demands of a cell, the more mitochondria and mtDNA molecules it contains. In a cell, all mtDNA molecules can be identical, which is called homoplasmy. Alternatively, two (or more) types of mtDNA molecules that differ in sequence can coexist in the same cell, tissue or even in the same organelle, which is called heteroplasmy. Heteroplasmy levels may range between 0% and 100% and the majority of severe pathogenic mtDNA mutations is heteroplasmic. Clinical manifestations depend to a certain extent on the mutation load and no symptoms occur unless the mutant load (proportion of mutant mtDNA) exceeds a certain threshold of expression (Hellebrekers et al., 2012). This threshold varies between tissues and between different mutations. In some mtDNA disorders, onset and severity of symptoms are clearly related to mutation load (Black et al., 1996; White et al., 1999). Often, however, phenotype and mutation load correlate poorly (Chinnery et al., 1997; Thorburn and Dahl, 2001). Finally, the mtDNA is transmitted entirely through the maternal lineage. The mutation load inherited by the fetus from a heteroplasmic mother is affected by a segregational bottleneck, which is a restriction in the number of mtDNA molecules to be transmitted followed by a strong amplification of these molecules. During oogenesis, the number of mtDNA molecules is reduced and the resulting few mtDNA become the founders of all the 100,000 to 600,000 mtDNA molecules in the mature oocyte, resulting in considerable variation in mtDNA mutant load among individual oocytes and subsequently among offspring (Jacobs et al., 2006; Poulton et al., 2010). The ‘size’ of this bottleneck is still under debate, but seems to depend on the type of mtDNA mutation and may even be individual-dependent for certain mutations (Blok et al., 1997; Brown et al., 2001; Monnot et al., 2011). Also the mechanism by which mtDNA mutations segregate, either randomly or preferentially, is poorly understood and appears to differ among mutations and nuclear backgrounds. Apparently, homoplasmy based on uniparental inheritance seems to be the preferred and most healthy situation (Sharpley et al., 2012).

mtDNA defects and recurrence risks Disease causing mutations in the mtDNA can be due to large rearrangements, point mutations or a reduced copy number, in some cases leading to depletion of the mtDNA. One has to be aware that a mtDNA defect can be the primary cause of disease, but that it also can be the manifestation of nuclear gene defects (e.g. defects in genes involved in mtDNA maintenance causing multiple mtDNA deletions and/or mtDNA depletion), mitotoxic drugs (e.g. nucleoside reverse transcriptase inhibitors can induce mtDNA depletion) or ageing (e.g. multiple deletions). It is obvious that the recurrence risk – the likelihood that the mtDNA disease present in a patient will occur again in his or her offspring or sibling – is highly dependent on the

Preventing the transmission of mtDNA disorders nature of the underlying primary genetic defect. This is especially an issue for mtDNA rearrangements. Large single mtDNA deletions have a low recurrence risk of one in 25 in female patients and zero in males (Chinnery et al., 2004), but multiple mtDNA deletions due to autosomal recessive nuclear gene defects have a recurrence risk of 25% for other siblings, as both unaffected parents are carriers of a defect in the same gene. In case of autosomal dominantly segregating multiple DNA deletions, the recurrence risk is even 50%, as the transmission of only a single gene defect by a patient suffices to develop clinical symptoms in his or her offspring. This is in contrast to multiple mtDNA deletions, occurring as somatic mutations as part of the normal ageing process. These multiple mtDNA deletions do not explain the clinical phenotype and have no recurrence risk. Point mutations in the mtDNA can be pathogenic mutations, which can explain severe disease, but can also be risk factors that contribute to complex disease manifestations or neutral polymorphisms unrelated to disease. Because of the high mutation rate of the mtDNA, neutral polymorphisms are common (Voets et al., 2011). Until now, more than 150 confirmed pathogenic point mutations have been reported in the mtDNA, affecting protein coding genes or RNA genes (Hellebrekers et al., 2012). Most pathogenic point mutations are heteroplasmic, but homoplasmic disease causing point mutations in the mtDNA are known as well, generally causing less severe disease or with reduced penetrance, requiring additional, mostly unknown, genetic or environmental factors for clinical expression. Heteroplasmic point mutations are more often associated with severe disease. Many of these mutations present only in few or even single families. All mutations display clinical heterogeneity, but this is most extreme for the common m.3243A>G mutation, which can manifest with a variety of clinical manifestations, including the MELAS syndrome, deafness, cardiomyopathy, renal problems and/or diabetes. Symptoms may co-occur in a patient or differ between patients, even of the same family. Point mutations also emerge somatically, usually at low levels, although fixation of mutations can occur in dividing cells or cancer cells, and homoplasmy levels can be reached.

Preventing the transmission of mtDNA diseases A number of approaches exists to prevent the transmission of mtDNA disorders, all having their specific advantages and disadvantages and their technical and ethical constraints (Poulton et al., 2010). The remainder of this paper mainly concentrates on methods to prevent the transmission of point mutations in the mtDNA, which have been demonstrated to be prime causes of disease in affected family members and which are not secondary to nuclear or environmental factors.

Oocyte donation The use of donor oocytes with spermatozoa of the partner is a reliable method to prevent the transmission of mitochondrial disease caused by mtDNA mutations, with the drawback that the resulting child is only genetically related to the father and not to the mother. However, the woman

601 who carries and gives birth to a child will be recognized as its legal mother. The use of donor oocytes of close maternal relatives is not advisable in the case of familial disease since these may carry the same mtDNA mutations even though the mutation is undetectable in blood of the possible donor. A major problem can be the availability of sufficient oocytes for donation, which is obviously a key requirement for this treatment.

Prenatal diagnosis Conventional prenatal diagnosis (PND) during pregnancy on chorionic villi or amniotic fluid cells has its own complexity for heteroplasmic mtDNA mutations. Although the mutation load can be determined accurately in the DNA collected, it is the sampling and the interpretation which makes PND problematic. A number of criteria have been proposed to allow reliable PND in mtDNA disease (Poulton and Turnbull, 2000). First, the distribution of mutant mtDNA in all extra-embryonic and fetal tissues should be uniform. Second, the mutant load should not change over time. Third, a close correlation must exist between the mutant load and disease severity. Limited data are available to determine whether the mutation load determined adequately reflects the fetal mutation load due to differences between the tested chorionic villi or amniotic fluid cells and the actual fetus. These data, mainly on the NARP/Leigh syndrome mtDNA 8993T>C/G mutations (Dahl et al., 2000; Steffann et al., 2007; Thorburn and Dahl, 2001; White et al., 1999) and m.9176T>C (Jacobs et al., 2005), suggested that for these mutations the mutant load of extra-embryonic tissues was representative of the mutant load in the fetus. But differences have also been reported, questioning the reliability of chorionic villus sample analysis for mtDNA disorders (Marchington et al., 2006). Recent reports on the highly variable m.3243A>G mutation indicated that this mutation segregated relatively stable throughout the prenatal period, quite different from the postnatal segregation, in which the mutation load decreases with time (Bouchet et al., 2006; Monnot et al., 2011). Although still based on only few mutations and limited numbers of patients, these data suggest that during embryonic development mutation loads remain stable with time and across various tissues. If the mutation load in the fetus can be correctly determined, then the key remaining problem for PND for mtDNA disorders is that for the majority of mutations, either the phenotype cannot be accurately predicted from the mutation load or that insufficient data exist to judge this accurately. This would make it very difficult to interpret the clinical significance of the mutation load for the health of the future child and so for parents to decide on continuation or termination of a pregnancy. Exceptions exist and, for example, PND can be reliably performed for the m.8993T>G. For this mutation, prediction of the phenotype is possible for most mutation loads, although even in this case a grey zone of inconclusive results exists. If a safe margin is calculated, then a mutant load of G is a guarantee for healthy offspring, whereas a mutant load of 60% would give a 25% chance of a severe outcome (White et al., 1999).

602 In conclusion, existing data indicate that although prenatal evaluation of mutation load is technically possible, it seems due to limitations in predicting the phenotype, only suitable in those cases where a high likelihood exists of offspring with either no mutation or a mutation load below the threshold of expression. This is the case for de novo mtDNA disease, when the unaffected mother of an affected child with in general high mutation levels does not elicit the presence of this mutation in any of her tissues available for testing (muscle, skin, blood, hair root). Although it remains always possible that she still has the mutation in other oocytes, this is not very likely and such a carrier has a good chance of producing healthy offspring without the mtDNA mutation, which was present in the previous child (Sallevelt et al., unpublished data). It is critical that such a carrier will be counselled correctly, as frequently wrong estimates for recurrence are given, based on the high mutation load in the affected child and not on the absence of the mutation in the mother. PND could be offered for reassurance in such a case. PND is also an option for carriers with a low mutation load in the tissues tested, in case these mutations demonstrate general skewing to the extremes (0% or 100% mutation load), as is the case for the m.8993T>G mutation. Under such circumstances, the majority of the fetuses will be healthy and not carry the pathogenic mutation. It is evident that in any case the decision to terminate a viable pregnancy based on PND can be extremely traumatic for many couples, even in the presence of a known abnormality.

Preimplantation genetic diagnosis Preimplantation genetic diagnosis (PGD) is an alternative to PND. Oocytes are fertilized in vitro and cells, usually from the 8-cell embryo, are biopsied and tested for the presence of a genetic defect. Unaffected embryos are transferred into the uterus. PGD avoids the dilemmatic choice of a pregnancy termination, which is a major advantage compared with PND in mtDNA disorders, which often results in multiple affected or ambiguous pregnancies. A technical advantage of PGD for mtDNA diseases is the high copy number of the mtDNA (up to 75,000 copies per blastomere) compared with the nuclear genes (two copies per blastomere). This makes the analysis less prone to artefacts such as amplification failure and allelic drop out, which can complicate PGD analysis of nuclear single-gene defects. Critical in offering PGD for mtDNA diseases is the representativeness of the mtDNA mutation load determined in blastomeres for the entire embryo and the possibility to characterize embryos with a mutation load below the threshold of expression, which can be considered to be healthy. PGD for mtDNA diseases by testing individual blastomeres has been reported for a number of mutations (m.3243A>G, m.8993T>G, m.8344A>G; Gigarel et al., 2011; Monnot et al., 2011; Sallevelt et al., 2013; Steffann et al., 2006; Thorburn et al., 2009). Although in principle, both blastomere and polar body biopsies are feasible for PGD, it has been shown that the mutation load determined in polar bodies does not reflect the mutation load in the embryo sufficiently, making polar biopsy analysis unsuitable for prevention of the transmission of mtDNA diseases

HJM Smeets (Gigarel et al., 2011). The mutation load determined in blastomeres accurately reflects the mutation load in the embryo (Figure 1), although one should keep in mind that data on human preimplantation embryonic mtDNA mutant levels are limited and biased towards specific mutations and that conclusions are often too easily generalized. Studies have shown similar mtDNA heteroplasmy levels in the blastomeres of a blastocyst, and minimal variation in time and between tissues throughout embryo-fetal development (Monnot et al., 2011; Sallevelt et al., 2013; Steffann et al., 2007). This indeed suggests that mutation levels in a blastomere reliably predict the mutant load of the future child. Data from embryos, in which the mutation load has been determined in every cell of the embryo, show that indeed the majority of single cells have a comparable mutation load (Figure 1). However, individual outliers exist and therefore it is advisable to always perform PGD on two blastomeres of a single embryo and use the highest mutation load in case of discrepancies (Sallevelt et al., 2013). Verification with PND could be offered to minimize uncertainty, although PND could due to the limitations mentioned above create uncertainty as well. Still, it would help improve understanding of the developmental aspects of mtDNA disease. The issue of genotype–phenotype correlation and predicting the phenotype from the mutation load is less important in PGD than in PND. In PND the decision to either continue or terminate a pregnancy depends on the accuracy of this prediction. In PGD the key factor is the threshold of expression of a mutation. For common mutations this can be determined on the available data, showing considerable variation in the mutation thresholds (Poulton and Bredenoord, 2010). For the rare or private mutations, this is often unknown. As this lack of data would preclude a general usage of PGD for mtDNA mutations, a study was performed to test if a minimal mutation level may exist, which remains below the pathogenic thresholds of all individual mtDNA mutations, assuming that a minimal level of wild-type mtDNA will be sufficient to prevent the occurrence of clinical symptoms and to investigate the existence of a ‘safe’ cut-off point below which the chance of being affected is acceptably low, was determined irrespective of the mtDNA mutation. PGD could then be offered to select most likely unaffected embryos with mutation loads in blastomeres under this threshold. To calculate the chance of being unaffected for mtDNA mutation carriers, a systematic review was performed on different pathogenic mtDNA mutations (Hellebrekers et al., 2012), using only data on mutation levels in muscle, which were more stable and expected to reflect mtDNA mutation load in embryonic tissues better than the often rapidly changing mtDNA mutation load in dividing cells. Data on muscle mutant level and clinical status of 159 different pathogenic mtDNA mutations, retrieved from 327 pedigrees, were combined, covering 385 affected individuals and 19 unaffected mutation carriers. The work included checking for a selection bias and for differences dependent on the nature of the mutations, resulting in the inclusion of only the definite pathogenic, familial mutations. As only few unaffected individuals underwent a muscle biopsy, an assumption had to be made for calculations on the general proportion of affected and unaffected

Preventing the transmission of mtDNA disorders

603

Figure 1 Preimplantation genetic diagnosis (PGD) for a carrier of the m.3243A>G mutation. (A) The pedigree of the family with an affected mother and an unaffected daughter with the mutation load in different tissues. (B) The analysis of five embryos during a PGD cycle; despite the high-mutation load in the carrer, one of the embryos is practically mutation-free (1% mutation load). (C) The non-transferred embryos were dissected and individual cells were analysed; in general, individual cells reflect the entire embryos, although single outliers were observed (data not shown). Bars = 100 lm.

individuals in families, which obviously differs from the proportion of biopsies from affected and unaffected relatives. The a-priori prevalence of being affected was estimated by averaging the proportion of affected siblings from the probands with a familial mtDNA mutation. The calculation implied a general a-priori probability of being affected of 0.477 (95% confidence interval 0.415–0.540). Using this a-priori probability and the pooled data on muscle mutant levels from all affected and unaffected individuals from

the pedigrees of familial mtDNA mutations, the risk of being affected at a certain mutation level was calculated with 95% confidence intervals (Figure 2). This graph can be used to estimate a mutant level threshold at which the risk of being affected is acceptably low and embryos would be eligible for transfer in PGD. For example, a 95% or higher chance of being unaffected was found at a muscle mutant level of 18% or less (95% confidence limit). At mutation level 18%, the predicted probability of being affected is 0.00744.

604

HJM Smeets

Figure 2 Estimated probability of being affected at a certain mtDNA mutation level in muscle (logistic regression model), assuming a general probability of being affected of 0.477. Muscle mutant level from all affected individuals (n = 195) and unaffected carriers (n = 19) from the pedigrees with a familial mtDNA mutation were pooled. (Reproduced by permission of Oxford University Press from Hellebrekers et al., 2012).

Based on pre-test odds of 0.477/0.523, this results in a negative diagnostic likelihood ratio of 0.0082. It is evident that PGD of heteroplasmic mtDNA mutations should be based on adequate (pre-test) case-by-case counselling, considering the uncertainties linked to this risk-reduction strategy (Bredenoord et al., 2008a; Poulton and Bredenoord, 2010). The current estimations can be supplied as guidance to all families with a mtDNA disorder, irrespective of the point mutation, on a case-by-case basis and may depend on the data generated by performing an actual PGD cycle. The exact threshold should be discussed and determined during counselling of each individual couple, since it also depends on other factors such as disease manifestation and clinical severity in the family, perception of risk, the availability of embryos below the threshold and general fertility issues associated with IVF (Hellebrekers et al., 2012). Limited experience has demonstrated that the carriers tested so far are all able to produce oocytes with mutation loads far below the 18% threshold of expression (Gigarel et al., 2011; Monnot et al., 2011; Sallevelt et al., 2013; Steffann et al., 2006; Thorburn et al., 2009). This will most likely also depend on the mutation load of the carrier herself, the nature of the mutation and the distribution of the mutation load among individual oocytes. High mutation loads in the carrier will not be a problem, in cases where the mutation displays a random distribution in oocytes, such as the m.3243A>G mutation (Brown et al., 2001; Monnot et al., 2011), or a skewed distribution, often going to the extremes (0% and 100%), such as the m.8993T>G/C mutation (Steffann et al., 2006; White et al., 1999). These carriers will very likely produce oocytes below the threshold of expression. The frequency of these oocytes will define the numbers of oocytes or PGD cycles required. A narrow distribution of the mutation load in oocytes and a level above the threshold of expression in the carrier could be a contraindication for PGD. It is clear that the choice for a lower threshold than 18% or even 0%

might reduce the chance of identifying an embryo matching such criteria considerably. Anyhow, each PGD cycle will provide additional information on heteroplasmy levels in a woman’s oocytes and the chance of selecting presumably healthy embryos below the requested threshold of expression. How far will PGD for mtDNA disorders help in preventing the transmission of mtDNA disease? PGD of heteroplasmic mtDNA mutations is technically safe and the results are reliable if carried out on two blastomeres. Carriers of all heteroplasmic mtDNA mutations have a fair chance of having healthy offspring by applying PGD, irrespective of the mutation. A cut-off mutation percentage at which the risk of being affected is acceptably low as a risk reduction strategy. Based on limited PGD cycles for specific mutations, it is expected that most mtDNA mutation carriers will have oocytes below this threshold. As most implanted embryos will carry a low level of mutated mtDNA it is clear that the future female generation may face the same choice or dilemma as their mother. Selection of male embryos by including sex analysis during PGD would definitely eliminate mtDNA disease for future generations without further testing. For female embryos this is not the case, although it is likely that because of low mutation levels of the implanted embryos, it will be possible to select completely mutation-free embryos for the next generation, also clearing the female lineage of mtDNA disease (Samuels et al., 2013). Ethical issues apply to PGD in general and more specific to PGD for mtDNA disorders (Bredenoord et al., 2008a; Poulton and Bredenoord, 2010). Although PGD for mtDNA disorders cannot fully guarantee an unaffected child, it is reasonable to assume that in most cases a healthy embryo will be selected for transfer and PGD will eliminate the chance of having a severely affected child. PGD does not only contribute to positive health outcomes, but also enhances the reproductive autonomy of the couples involved. Still, a risk-reducing strategy raises issues of

Preventing the transmission of mtDNA disorders

605

Figure 3 Nuclear genome transfer in mtDNA disease. (A) Chromosome–spindle transfer: the chromosome–spindle complex is removed from the oocyte of a carrier with mutated (red) mitochondria and fused to an oocyte with healthy (green) mitochondria from which the chromosome–spindle complex is removed; the reconstructed oocyte is then fertilized. (B) Pronuclear transfer: after fertilization, the pronuclei (dark blue) are removed from a zygote with mutated (red) mitochondria and fused to a zygote with healthy (green) mitochondria from which the pronuclei have been removed (light blue).

parental and medical responsibilities, but none of these supply convincing arguments to regard this as unacceptable. It is obvious that a responsible use of PGD for mtDNA disorders will raise additional points for discussion: to what extent can parents ask or be asked to take more or less risks; how to balance embryo quality and mutation load, because the best embryo to achieve a pregnancy does not need to have the lowest mutation load; the number of cycles that can be justified to find the best possible embryo. The additional issue of selecting male fetuses in order to eliminate the risk for future generations is in principle morally acceptable within the context of PGD for mtDNA disorders (Bredenoord et al., 2010).

Nuclear genome transfer Nuclear genome transfer (Figure 3) involves the transfer of the nuclear genome from an oocyte or zygote with mutated mtDNA in the cytoplasm (donor) to an enucleated acceptor oocyte or zygote of a healthy donor (acceptor) with presumably normal, mutation-free mtDNA (Craven et al., 2011). Consequently, the offspring would not carry the mtDNA

mutation present in the mother and would not suffer from the familial mtDNA disease. (Immature) spindle–chromosome transfer is carried out at the level of the oocyte (Paull et al., 2013; Tachibana et al., 2009, 2013) and pronuclear transfer at the level of the zygote (Craven et al., 2010). These methods can be applied to carriers of all pathogenic mtDNA mutations, heteroplasmic and homoplasmic alike, but similar to oocyte donation require the availability of sufficient donor oocytes or zygotes, which could be a major hurdle for clinical application. The feasibility of spindle–chromosome transfer has been demonstrated in primate offspring (Macaca mulatta) and human oocytes, which were subsequently fertilized. Because a little bit of donor cytoplasm is always included in the transfer, acceptor cells were tested for the presence of donor mtDNA. No donor mtDNA was detectable in skin and blood of the monkeys, spindle transfer embryos and embryonic stem cells, derived from blastocysts, with an initial sensitivity of 3% (Tachibana et al., 2009) and later of 1% (Tachibana et al., 2013). The feasibility of spindle–chromosome transfer was also demonstrated by the generation of sufficient mutation-free blastocysts suitable for transfer. For humans, retrieval of 12 MII oocytes would generate at least two

606 blastocysts for transfer, which seems appropriate for clinical practice. Finally, the issue of synchronous retrieval of both donor and acceptor oocytes was difficult to manage in initial protocols using fresh oocytes (Tachibana et al., 2009). However, it was shown that transfer of vitrified spindle–chromosome complexes, derived from monkey oocytes, into fresh cytoplasts, yielded similar results compared with fresh controls (Tachibana et al., 2013). In another study (Paull et al., 2013) vitrification or cooling was performed on human spindle–chromosome complexes, leading to a reversible destabilization of spindle microtubules and partial depolymerization of the complex. Transfer of such an incompletely assembled, immature spindle–chromosome complex was followed by parthenogenetic development. The mtDNA, transferred with the nuclear genome, was initially detected at levels below 1%, which decreased in parthenogenetic blastocysts and derived stem-cell lines to undetectable levels, and remained undetectable after passage for more than a year, clonal expansion, differentiation into neurons and cardiomyocytes and cellular reprogramming. Based on these data it seems feasible to apply (immature) spindle–chromosome transfer between oocytes to produce offspring, in which the mtDNA present in the mother is reduced to undetectable amounts. Comparable results were obtained with another approach, involving pronuclear transfer between abnormally fertilized oocytes (Craven et al., 2010). Minimizing the size of the karyoplast has been successful in reducing the amount of mtDNA carry over to undetectable or very low levels (G as a model system. Hum. Mutat. 32, 116–125. Nass, M.M., 1966. The circularity of mitochondrial DNA. Proc. Natl. Acad. Sci. U. S. A. 56, 1215–1222. Paull, D., Emmanuele, V., Weiss, K.A., Treff, N., Stewart, L., Hua, H., Zimmer, M., Kahler, D.J., Goland, R.S., Noggle, S.A., Prosser, R., Hirano, M., Sauer, M.V., Egli, D., 2013. Nuclear genome transfer in human oocytes eliminates mitochondrial DNA variants. Nature 493, 632–637. Poulton, J., 2012. Nuclear tranfers to prevent maternal transmission of mitochondrial DNA disease. BMJ 345, e6651–e6652. Poulton, J., Bredenoord, A.L., 2010. 174th ENMC international workshop: applying pre-implantation genetic diagnosis to mtDNA diseases: implications of scientific advances 19–21 March 2010, Naarden, The Netherlands. Neuromuscul. Disord. 20, 559–563. Poulton, J., Turnbull, D.M., 2000. 74th ENMC international workshop: mitochondrial diseases 19–20 november 1999, Naarden, the Netherlands. Neuromuscul. Disord. 10, 460–462. Poulton, J., Chiaratti, M.R., Meirelles, F.V., Kennedy, S., Wells, D., Holt, I.J., 2010. Transmission of mitochondrial DNA diseases and ways to prevent them. PLoS Genet. 6, e1001066. Reynier, P., May-Panloup, P., Chretien, M.F., Morgan, C.J., Jean, M., Savagner, F., Barriere, P., Malthiery, Y., 2001. Mitochondrial DNA content affects the fertilizability of human oocytes. Mol. Hum. Reprod. 7, 425–429. Rotig, A., Munnich, A., 2003. Genetic features of mitochondrial respiratory chain disorders. J. Am. Soc. Nephrol. 14, 2995–3007. Sallevelt, S.C., Dreesen, J.C., Dru ¨sedau, M., Spierts, S., Coonen, E., van Tienen, F.H., van Golde, R.J., de Coo, I.F., Geraedts, J.P., de Die-Smulders, C.E., Smeets, H.J., 2013. Preimplantation genetic diagnosis in mitochondrial DNA disorders: challenge and success. J. Med. Genet. 50, 125–132. Samuels, D.C., Wonnapinij, P., Chinnery, P.F., 2013. Preventing the transmission of pathogenic mitochondrial DNA mutations: can we achieve long-term benefits from germ-line gene transfer? Hum. Reprod. [Epub ahead of print]. Sharpley, M.S., Marciniak, C., Eckel-Mahan, K., McManus, M., Crimi, M., Waymire, K., Shi Lin, C., Masubuchi, S., Friend, N., Koike, M., Chalkia, D., MacGregor, G., Sassone-Corsi, P., Wallace, D.C., 2012. Heteroplasmy of mouse mtDNA is genetically unstable and results in altered behavior and cognition. Cell 151, 333–343. Steffann, J., Frydman, N., Gigarel, N., Burlet, P., Ray, P.F., Fanchin, R., Feyereisen, E., Kerbrat, V., Tachdjian, G., Bonnefont, J.P., Frydman, R., Munnich, A., 2006. Analysis of mtDNA variant segregation during early human embryonic development: a tool for successful NARP preimplantation diagnosis. J. Med. Genet. 43, 244–247. Steffann, J., Gigarel, N., Corcos, J., Bonniere, M., Encha-Razavi, F., Sinico, M., Prevot, S., Dumez, Y., Yamgnane, A., Frydman, R., Munnich, A., Bonnefont, J.P., 2007. Stability of the m.8993T->G mtDNA mutation load during human embryofetal development has implications for the feasibility of prenatal diagnosis in NARP syndrome. J. Med. Genet. 44, 664–669. Tachibana, M., Sparman, M., Sritanaudomchai, H., Ma, H., Clepper, L., Woodward, J., Li, Y., Ramsey, C., Kolotushkina, O., Mitalipov, S., 2009. Mitochondrial gene replacement in primate offspring and embryonic stem cells. Nature 461, 367–372. Tachibana, M., Amato, P., Sparman, M., Woodward, J., Sanchis, D.M., Ma, H., Gutierrez, N.M., Tippner-Hedges, R., Kang, E., Lee, H.S., Ramsey, C., Masterson, K., Battaglia, D., Lee, D., Wu,

610 D., Jensen, J., Patton, P., Gokhale, S., Stouffer, R., Mitalipov, S., . Towards germline gene therapy of inherited mitochondrial diseases. Nature 493, 627–631. Thorburn, D.R., Dahl, H.H., 2001. Mitochondrial disorders: genetics, counselling, prenatal diagnosis and reproductive options. Am. J. Med. Genet. 106, 102–114. Thorburn, D.R., Wilton, L., Stock-Myer, S., 2009. Healthy baby girl born following pre-implantation genetic diagnosis for mitochondrial DNA m.8993T>G mutation. Mol. Genet. Metab. 98, 5–6. Voets, A.M., van den Bosch, B.J., Stassen, A.P., Hendrickx, A.T., Hellebrekers, D.M., Van Laer, L., Van Eyken, E., Van Camp, G., Pyle, A., Baudouin, S.V., Chinnery, P.F., Smeets, H.J.M., 2011. Large scale mtDNA sequencing reveals sequence and functional

HJM Smeets conservation as major determinants of homoplasmic mtDNA variant distribution. Mitochondrion 11, 964–972. White, S.L., Collins, V.R., Wolfe, R., Cleary, M.A., Shanske, S., DiMauro, S., Dahl, H.H., Thorburn, D.R., 1999. Genetic counselling and prenatal diagnosis for the mitochondrial DNA mutations at nucleotide 8993. Am. J. Hum. Genet. 65, 474–482. Declaration: The author reports no financial or commercial conflicts of interest. Received 17 February 2013; refereed 26 July 2013; accepted 1 August 2013.

Preventing the transmission of mitochondrial DNA disorders: selecting the good guys or kicking out the bad guys.

Mitochondrial disorders represent the most common group of inborn errors of metabolism. Clinical manifestations can be extremely variable, ranging fro...
1MB Sizes 0 Downloads 0 Views