Mitochondrial replacement by genome transfer in human oocytes: Efficacy, concerns, and legality

Abstract Background Pathogenic mitochondrial (mt)DNA mutations, which often cause life‐threatening disorders, are maternally inherited via the cytoplasm of oocytes. Mitochondrial replacement therapy (MRT) is expected to prevent second‐generation transmission of mtDNA mutations. However, MRT may affect the function of respiratory chain complexes comprised of both nuclear and mitochondrial proteins. Methods Based on the literature and current regulatory guidelines (especially in Japan), we analyzed and reviewed the recent developments in human models of MRT. Main findings MRT does not compromise pre‐implantation development or stem cell isolation. Mitochondrial function in stem cells after MRT is also normal. Although mtDNA carryover is usually less than 0.5%, even low levels of heteroplasmy can affect the stability of the mtDNA genotype, and directional or stochastic mtDNA drift occurs in a subset of stem cell lines (mtDNA genetic drift). MRT could prevent serious genetic disorders from being passed on to the offspring. However, it should be noted that this technique currently poses significant risks for use in embryos designed for implantation. Conclusion The maternal genome is fundamentally compatible with different mitochondrial genotypes, and vertical inheritance is not required for normal mitochondrial function. Unresolved questions regarding mtDNA genetic drift can be addressed by basic research using MRT.

among the OXPHOS protein complexes I, III, IV, and V, and are involved in OXPHOS activity. The displacement loop (D-loop) is a regulatory sequence that controls mtDNA replication and transcription.
In contrast, the nuclear genome encodes 74 polypeptides, including an OXPHOS complex and other mitochondrial proteins.
Mutations in mtDNA cause defects in the respiratory chain function and energy production within mitochondria, which can result in a wide range of clinical conditions and mitochondrial diseases.
Thus far, there is no fundamental treatment for mitochondrial diseases. Thus, mitochondrial replacement (MR) for preventing the inheritance of mutant mtDNA has been tested in animal models, and clinical research is ongoing in the United Kingdom (UK). Here, we reviewed recent developments in human models of MRT, the underlying biology, and the regulatory guidelines for current methods, particularly in Japan.

| mtDNA MUTATI ON S AND MITOCHONDRIAL DIS E A S E S
Many somatic cells contain approximately 1000 copies of mtDNA, whereas oocytes contain several hundred thousand copies of mtDNA. 1 Dysfunction in mitochondrial replication can result in single or multiple mtDNA mutations. The mitochondrial genome mutation rate is reported to be between two-and sixfold higher in non-vertebrates, and approximately 20 times higher in vertebrates, than that of nuclear genomes. 2 To date, the mechanisms that affect variations in mtDNA mutation rates are not well understood, although several hypotheses have been suggested. First, the different metabolic rates of various species can lead to higher ROS production during OXPHOS in mitochondria, resulting in oxidative damage and higher mtDNA mutation rates. Second, the number of genome replications per generation differs between mitochondrial and nuclear genomes. Furthermore, DNA polymerase γ, which is involved in the replication of the mitochondrial genome, has poor fidelity. 2 Mitochondrial disruption of energy production due to mtDNA mutations can affect different tissues and result in severe disease.
Mutations in mtDNA can cause defects in the respiratory chain function and energy production within mitochondria, resulting in a wide range of clinical conditions, such as liver dysfunction, bone marrow failure, pancreatic islet cell deficiency, diabetes mellitus, deafness, respiratory failure, stroke, heart disease, neurodegenerative disorders, and other diseases, collectively referred to as mitochondrial diseases. 3 Diseases resulting from mutant mtDNA show distinct patterns of inheritance owing to three features: maternal inheritance, mtDNA heteroplasmy (the proportion of mutant mtDNA relative to total mtDNA in a cell), and mtDNA replicative segregation.
The first feature refers to mutant mtDNA being subjected to maternal transmission through the oocyte cytoplasm and not from the father. However, some exceptions have been reported; for example, 17 individuals who inherited mtDNA from both parents were identified using sequencing data from the mitochondrial genomes of members from three unrelated families. 4 The second type, mtDNA heteroplasmy, is wherein phenotypic variability of mtDNA diseases is related to the high copy number of mtDNA in mammalian cells, which can therefore contain both mutant and wild-type mtDNA populations. 5 The third type, mtDNA replicative segregation, can yield various distinct genotypes in the mtDNA pool of each tissue. 6 Mitochondrial disease severity is dependent on the affected gene, percentage of mtDNA heteroplasmy and tissue distribution.
Two mitochondrial genetic bottlenecks have been reported to date. During oogenesis, primordial germ cells containing mutant mtDNA exhibit a dramatic reduction in mtDNA content, with discrete mitochondria containing approximately five mtDNA molecules. 7 Oocytes receive only a selected number of mtDNA molecules each, which are amplified accordingly to yield several hundred thousand mtDNA copies, as is observed in mature oocytes. 8 This selective reduction in mtDNA copies ensures that very few mtDNA copies can be clonally amplified, resulting in variability in the percentage of mutant mtDNA molecules in gametes and transmission to the next generation. This is referred to as the mitochondrial genetic bottleneck, and results in different loads of mutant mtDNA in different oocytes and variable transmission of pathogenic mtDNA from mother to offspring. mtDNA mutations are found in approximately 1 of 200 live births. 9 Human mitochondrial disease-causing mtDNA mutations were originally reported in 1988. 5,10 Since then, over 200 such mtDNA mutations have been discovered (as shown in mitomap [https://www.mitom ap.org/MITOMAP]); most of these mutations occur in a heteroplasmic context. Mitochondrial diseases affect approximately 1 out of 5000-10 000 adults. 11 However, owing to difficulties in diagnosis due to an extraordinarily broad spectrum of symptoms, the actual incidence is thought to be much higher.
Variability in the degree of mtDNA heteroplasmy arising from the mtDNA bottleneck can affect the phenotypes of mitochondrial diseases. There is a phenotypic threshold effect associated with the percentage of mutant mtDNA per cell, and its severity varies greatly depending on the specific mtDNA mutation and the organ affected. 3

| PRE VENTI ON OF THE TR AN S MISS I ON OF AB NORMAL mtDNA US ING PRE-IMPL ANTATI ON G ENE TIC TE S TING FOR MITOCHONDRIAL DISORDER S
Currently, no curative treatments for mitochondrial diseases have been developed, and treatment remains limited to symptomatic management. Increasing identification of mutations and advanced genomic techniques have improved the diagnosis of mitochondrial disorders. In clinical practice, pre-implantation genetic testing (PGT), an in vitro fertilization (IVF)-based technique developed three decades ago, 12 is one option to produce embryos without mutant mtDNA or with few heteroplasmic mtDNA mutations. PGT-M is the approach used for the diagnosis of mitochondrial disease, and PGT is conventionally used for monogenic disorders. In the current review, the term PGT-MIT would be used for convenience to describe PGT, which is limited to the diagnosis of mitochondrial diseases. PGT is based on the genetic analysis of one or several cells obtained from embryos at the cleavage stage 13 and/or the blastocyst stage. [14][15][16][17] This approach enables the selection of mutation-free embryos or embryos with low mutation load, before their transfer to the uterus.
Since the first successful case in which PGT-MIT reduced the risk of neurogenic ataxia retinitis pigmentosa in a patient with the m.8993T > G mutation, 18 a number of children have been born with the help of PGT-MIT, by enabling the selection of embryos with low mutation levels. [19][20][21] Although current conventional approaches are useful for reducing or eliminating the risk of mtDNA, pre-implantation genetic diagnosis has several limitations. For example, pre-implantation genetic diagnosis is not available for women who harbor homoplastic mtDNA mutations. Additionally, animal studies using experimentally constructed primate embryos containing heteroplasmic mtDNA showed marked variation in the levels of heteroplasmic mtDNA between blastomeres, 22  shown not to improve the overall pregnancy outcomes. 23 However, initial studies have shown the presence of DNA in polar bodies, 24 blastocoel fluid, and spent culture medium, 25 which could facilitate the development of non-invasive methods for PGT.

| MR THER APY (MRT )
The severity of clinical phenotypes in most maternally inherited mitochondrial diseases varies according to the percentage of mutant mtDNA (mtDNA heteroplasmy). Although PGT is a useful option for reducing the risk of abnormal mtDNA transmission, the exact threshold for the mtDNA heteroplasmy ratio is still unclear. 26 Even if embryos containing a mutation ratio below the threshold give rise to a successful pregnancy, the stochastic replicative segregation of heteroplasmic mtDNA during cell division can result in genetic mosaicism within each organ. Therefore, to reduce the risk of mtDNA disease transmission, approaches to replace abnormal mitochondria with normal mitochondria from healthy donors using the nuclear transfer (NT) technique have been proposed.
The nuclear genome can be transferred by NT techniques, such as maternal spindle transfer (MST; Figure 1) and germinal vesicle transfer from unfertilized oocytes or first polar body transfer (PB1T), second polar body transfer (PB2T), and pronuclear transfer (PNT) from zygotes (Table 1). MST is the procedure for transferring the karyoplast (which contains nuclear DNA with a small amount of cytoplasm, surrounded by the plasma membrane) into an enucleated donor oocyte before fertilization. Because these reconstructed oocytes or embryos are obtained from healthy donor oocytes or zygotes, such NT procedures are thought to reduce the risk of transmission of mutant mtDNA to the next generation. Several studies have reported the live birth of healthy offspring from mouse embryos at the pronucleus stage 27 and from rhesus macaque oocytes at the metaphase stage of meiosis II (MII). 28 In an early human study, Craven et al reported a successful rate of blastocyst development after the transfer of pronuclei obtained from abnormal human embryos (unipronuclear or tripronuclear). 29 Their findings showed a decrease of approximately 50% in the blastocyst formation rate after genome exchange (8.3%, 3/36) compared with that of unmanipulated embryos (n = 76), partly owing to the absence of either a maternal or paternal genome. There was variation in mtDNA carryover, that is, the transfer of the mtDNA genotype from the nucleus donor to the embryos receiving pronuclei (8.1 ± 7.6%, n = 8). With optimization of the NT technique, the average mtDNA heteroplasmy ratio at pre-implantation stages decreased by less than 2% (1.68 ± 1.81%, n = 9).
The Nuffield Council on Bioethics published a report in 2012 titled "Novel techniques for the prevention of mitochondrial DNA disorders: an ethical review." In this report, the authors detailed their approval for research on the prevention of mtDNA disorders. 30 Paull et al performed genome exchange at the MII stage (MST) 1 and used artificial activation of unfertilized oocytes to exclude the F I G U R E 1 Spindle transfer for mitochondria replacement. Spindles are enucleated from both oocytes, and the carrier's spindle (orange) is fused with the enucleated healthy donor oocyte (blue). After nuclear exchange, the resulting oocyte, which consists of nuclear DNA from the carrier and cytoplasm from the donor oocyte, is subjected to in vitro fertilization (IVF) with the partner's sperm [Colour figure can be viewed at wileyonlinelibrary.com] influence of paternal factors on nuclear DNA-mtDNA incompatibility. The blastocyst development ratio in genome-exchanged oocytes (37%, 7/18) was comparable to those in unmanipulated IVF embryos (52%, 54/103) and parthenotes (33%, 7/21). In addition, the successful ratio of stem cell derivation from genome-exchanged oocytes (15%, 3/18) was also comparable to that from parthenotes

| MINIMIZING mtDNA C ARRYOVER
To avoid mtDNA heteroplasmy in MRT, it is critical to ensure inheritance of a single maternal mtDNA lineage. The human egg contains more than 100 000 copies of mtDNA; however, the sperm contains only approximately 100 copies. 33 Although a few reports have shown that paternal mtDNA can be passed on to the offspring, 4 sperm mitochondria are ubiquitinated inside the oocyte cytoplasm and later subjected to proteolysis during pre-implantation development. 34 Unlike after natural fertilization, when ubiquitination and proteolysis can eliminate a small amount of paternal mtDNA heteroplasmy, technical improvements are needed to avoid mtDNA heteroplasmy after MRT.
Hyslop et al modified the PNT protocol so that pre-implantation embryo development could be improved by minimizing mtDNA carryover. 35 In the original protocol, sucrose was added to the manipulation medium to facilitate enucleation and fusion. However, it was later removed because of the osmotic effect, which may increase mtDNA carryover. Before karyoplast fusion, the enucleated karyoplast was vitrified. Excess cytoplasm from karyoplasts was removed before fusion, and karyoplasts were then fused with enucleated fresh cytoplasm; 79% of blastocysts showed less than 2% mtDNA heteroplasmy, and none showed greater than 5% mtDNA heteroplasmy.

| MITO CHONDRIAL G ENE TIC DRIF T (mtDNA G ENOT YPE RE VER S ION)
The mechanism of mtDNA replicative segregation is still not clear.
Even with minimal mtDNA carryover, there is concern that mito-  mtDNA mutation load was transferred in utero. Subsequently, the neonate was born with an mtDNA mutation load between 2.36% and 9.23% in each tissue (amnion, 6.77%; buccal epithelium, 3.52%; foreskin, 9.23%; hair follicles, 5.59%; and urine precipitate, 2.36%). The baby was reported to be healthy at 7 months of age. 44 Another baby was born in Ukraine using the MRT technique. 45 In this case, genome transfer was not used for disease prevention, but instead for rejuvenation of aging oocytes to treat embryo arrest.

| E THI C AL CON CERN S AND LEG AL CONTROVER S IE S REG ARDING DELIVERY AF TER G ENOME TR AN S FER FOR THE PRE VENTION OF MITOCHONDRIAL DISE A SE
In

| REG UL ATING THE HANDLING OF NINE T YPE S OF EMB RYOS PRODUCED BY CLONING TECHNI QUE S A S " S PECIFIED EMB RYOS" IN JAPAN
In Although several news and review articles have indicated that Japan had previously released draft guidelines with a more permissive stance on human embryo genome editing, which did not outlaw germline editing for reproduction, 48,49 anything that is intended for reproduction, including human cloning or hybrid individuals, is