Mitochondrial/Nuclear Transfer: A Literature Review of the Ethical, Legal and Social Issues

des articles de recherche originaux, des rapports gouvernementaux ou commandés par le gouvernement, des éditoriaux, des lettres aux éditeurs et des nouvelles de recherche. La revue présente et synthétise les arguments présents dans la littérature quant aux thèmes les plus fréquemment soulevés: terminologie; identité, relations et parentalité; dommage potentiel; autonomie reproductive; alternatives disponibles; consentement; impact sur des groupes d’intérêt spécifiques; ressources; « pente glissante »; création, utilisation et destruction des embryons humains; et bienfaisance. La revue conclut en identifiant les enjeux ELS spécifiques au M/NT et en appelant à une recherche de suivi longitudinale clinique et psychosociale afin d’alimenter le futur débat sur les enjeux ELS de preuves empiriques.


Introduction
In February 2015, the United Kingdom became the first country in the world to legalize a new in vitro fertilization (IVF) technology called mitochondrial/nuclear transfer (M/NT) [1][2][3][4]. M/NT aims to avoid the transmission of serious mitochondrial diseases from an affected mother to her progeny by using one of two techniques: maternal spindle transfer (MST) and pronuclear transfer (PNT). Both techniques result in offspring with genetic material from three different persons: the nuclear DNA (nDNA) of the two prospective parents and the mitochondrial DNA (mtDNA) of the egg donor [5][6][7][8][9]. These modifications of the germ-line are inheritable and, therefore, transmitted to the offspring's progeny [5,8,10,11]. When mtDNA carries mutations, it can result in serious, potentially fatal, and currently untreatable diseases such as Leigh's syndrome, affecting mostly the organs whose operation requires the most energy: the central nervous system, heart, liver, kidneys, etc. [12].
The UK's decision to legalize M/NT has provoked a heated debate regarding the ethical, legal and social issues (ELSI) related to the technique. UK regulations came into force on October 29, 2015 [1]. However, the first live birth of a boy following M/NT (MST) occurred in Mexico in 2016 [89][90][91][92] 1

Methodology Critical interpretive review
The ELSI debate on M/NT is taking place in research articles, commentaries, editorials, government-commissioned reports, letters to editors, and research news. We therefore chose to perform a critical interpretive review of all these relevant sources [18]. While considerable discussion also occurs in blog posts, we did not include them in the review due to great variability in their quality. Each step described below was performed independently by two researchers (RDL and SB).

Search
The first step -performed in July 2015 -consisted of a systematic search, with the key words and phrases listed in Table 1, of the following databases which were considered by the authors to be the most relevant for this review: PubMed, CINAHL (EBSCOHost), Science Direct, Embase and PsycInfo. A total of 1 082 potentially relevant titles were found. From these, 1 033 papers were excluded as they did not meet either of the inclusion criteria, which were: a focus on M/NT, discussion of ELSI (i.e., not having exclusively scientific content), availability in French or English, availability online, not being duplicate titles or conference abstracts. The database search thus generated 49 papers that were selected for the review. Key websites (Google Scholar and the Georgetown Library) and journals (Nature, Science, BMJ, Fertility and Sterility, Human Reproduction, Reproductive Biomedicine Online and The Lancet) likely to offer essential papers were also searched. Researchers in the field were consulted to help identify additional papers. Finally, relevant references from bibliographies of included papers were also included (using the same inclusion criteria as above). This generated 46 additional papers, for a total of 95 (see Appendix A).

Analysis
The analysis was performed by two independent researchers using NVivo10 [19] and adapting Burnard's [20] stage-by-stage process of content analysis. The coding of all papers was performed both inductively and deductively, separately by RDL and SB. The results were then compared one by one and discussed by RDL and SB. When necessary, the results were discussed by all three authors until consensus was achieved. The themes were generated by the researchers' codes and subsequently classified into the following categories: terminology; identity, relationships and parenthood; potential harm; reproductive autonomy; available alternatives; consent; impact on specific interest groups; resources; "slippery slope"; creation, use and destruction of human embryos; and beneficence.

Terminology
The technology -as well as resulting offspring -is referred to using numerous terms ( Table 5). The expression "mitochondrial transfer" does not actually reflect what M/NT involves (i.e., the transfer of the spindle or the pronuclei -not of mitochondriafrom the prospective mother's egg to the donor's) [7]. Bredenoord et al. use the term "mtDNA modification" citing important previous use in the literature, while stating that "mtDNA replacement" is a more accurate expression [13]. Some terms arguably convey relative neutrality, such as "mitochondrial/nuclear transfer", while others are more value-laden, such as "mitochondrial therapy" [21]. In the present review, the phrase mitochondrial/nuclear transfer (M/NT) is used to designate both MST and PNT. [A] general assumption is that mtDNA does not really constitute our genetic make-up; it does not influence our phenotype, as it only governs cellular energy production. Modification of essential or defining characteristics is considered to be ethically more problematic, because it determines one's identity or personality. Modifying the nuclear DNA is therefore often regarded [as] more problematic than modifying the mtDNA [22, p.674].
A response to the above argument is that one's identity is about more than physical appearance and character traits. Most nuclear genes do not contribute to the physical appearance and character traits of a person, but rather are involved in fundamental processes, just like mitochondrial genes [23]. Hence, it is argued that the reasons for considering energy production to be an "unimportant physical function" are not clear [24, p.6] and that such premises could lead to the conclusion that it is acceptable to "tinker" with most genes [23]. mtDNA's role may be considerably underestimated and it may, in fact, have an impact on traits such as athleticism, fertility, ageing and intelligence, behaviour and health, variables that play a role in defining one's identity [9,22,25,26].
The qualities of DNA may thus be more relevant to the debate than its quantity [27]. Genes carried by mtDNA can have a pervasive effect on metabolism and physical development [28], potentially causing debilitating illness [Murdoch, in 29]. A person born free of mtDNA disease can thus have a very different life story, experience and even personality, because medical conditions can have a determining impact on one's self-perception and identity. If successful, M/NT can thus have a considerable influence on the resulting child's identity [6,13,14,30,[31][32][33]. While mtDNA may not directly impact character or physical traits, it may still be similar in some ways to nDNA modification, which is often perceived as more problematic than M/NT [22]. Study participants believed that because both mtDNA and nDNA can lead to serious disabilities, both kinds of DNA are identity-defining [16]. Like M/NT, "many other medical interventions, whether they involve genetic materials or not, are 'identity-altering' according to a variety of notions of identity" [31, p.55].
Some argue that M/NT is compatible with the child's "right to have one's future options kept open" [11, p.203] [see also 13,16,29], since it would eradicate the disease and therefore arguably broaden one's future options [11,13,16,29]. Others argue that M/NT would to a certain extent impede the child from making her/his own future choices [29].
The child's identity may be subject to confusion or ambiguity as s/he may be considered to have three genetic parents, leading to unknown social consequences [14,22,29,31,[35][36][37][38][39][40][41]. Some argue that having three genetic parents is not a major ethical concern [41], but others fear social stigmatization that may influence the child's conception of self [40]. Some suggest that the child should be informed, as in the case of other reproductive technologies involving a third-party contribution [37].
Ancestry tracing using mtDNA can provide genealogical information, which for some "suggests that the maternal input on the part of the donor is far from negligible" [42, p.19]. M/NT could make it impossible for the child to trace maternal genetic lineage [6], causing concerns regarding potential "ancestry confusion" [6,24,27,31,40,43]. However, it is argued that this confusion already exists due to egg, sperm and embryo donation [27].

ii) Relationships formed as a result of M/NT, donor's status and legal parenthood
M/NT may be considered controversial because, like other assisted reproductive technologies, it increases the number of parents a child may possibly have [44].
The third party genetic contribution may threaten some cultures' traditional concepts of family and parenthood [14,22,45]. However, three-(and more) parent families already exist: lesbian couples with a sperm donor, gay couples with a surrogate mother, reconstituted families, etc. [14,45]. M/NT is arguably closer to the "traditional nuclear family" than egg donation, since the prospective mother contributes her nDNA, carries, gives birth to and raises the child [30, p.14]. Nevertheless, M/NT challenges the definition of biological parenthood thus raising ethical and conceptual issues [13,46].
The availability of M/NT can be seen as a "double-edge sword" [47, p.3], making prospective parents responsible whether they choose to use or reject it. This may open the door to blame and possibly affect the parent-child relationship. Counselling of prospective families on how to manage the "three-parent family" may be needed, before and after conception, and at least until puberty -when children start asking questions about their origins [16]. Furthermore, the fact that the child was created to satisfy the parents' need of a genetically linked progeny may affect the relationship between the child and its parents by "commodifying" the child [24,30,42,48].
Some are concerned that the mtDNA donor may develop strong maternal feelings and may want to develop a relationship with the child. The prospective (nuclear) mother may feel that because of the donor's contribution, the offspring is not "fully hers" and confusion may result [49, p.1268]. However, current lack of data makes it challenging to anticipate the effect of M/NT on children's relationships with their parents and the donor, which will probably vary from one family to another, as it does in the case of reproductive egg donation [31].
When considering the third party's genetic contribution, questions about the privileges, rights, obligations and responsibilities of the donor are raised [16,40,43,49,50]. Some doubt whether children born following M/NT should know the donor's identity [16,43,51]. Paller-Rzepka presents some of the arguments put forward against the donor's recognition as a legal parent, highlighting the negative consequences of co-parentage on the child, as well as legislators' potential refusal to change existing legislation accordingly [50].
In the UK, when a child conceived with gamete donation after 2005 reaches adulthood, s/he can access the gamete donor's identity [52]. Moreover, the donor in these cases is not legally related to the resulting child and cannot obtain a parental order [29]. UK law allows the legal recognition of only one "mother" -the woman who carries and gives birth to the child [3,31]. Some argue that the same limitations should apply to M/NT [22,30,31]. According to the UK's M/NT regulations, mtDNA donors will be treated as tissue donors rather than as egg donors, with both children and the donor given access solely to limited nonidentifying information about each other [4,29,30]. It is unclear how mtDNA donors would be viewed in other jurisdictions [53].
Currently, there is a dearth of data regarding how mtDNA donors view the social significance of their donation [31]. Nevertheless, a voluntary system of contact or information exchange between the donor and the resulting child, arranged with mutual consent, is suggested [27,31,54,55].

Potential harm i) Harm to the future child
Issues regarding the limited available information about the safety and efficacy of M/NT as well as the potential physical or psychological harms to the resulting child, include: "mitochondrial disease, as a result of carryover of abnormal mitochondria and heteroplasmy"; "disorders due to nuclear-mitochondrial incompatibility"; "disorders related to aberrant epigenetic modifications"; "birth defects and other disorders associated with the specific mitochondrial manipulation technology procedure"; and "toxicities of reagents used in mitochondrial manipulation technologies" [56, p.20].
Because the goal of M/NT is a viable pregnancy, it "cannot be gradually phased into use" [31, p.65]. Furthermore, the germline effects of M/NT on the offspring will be irreversible [10,14,22,31,54]. The possible future discontinuation of M/NT as a treatment due to negative effects outweighing the benefits will be of little consolation to those children who are already negatively affected [31]. However, the risk of not offering M/NT to women who carry mutant mtDNA is that children may be born with diseases caused by mtDNA mutations [29,31].
The potential risks of M/NT raise the question of when it would be appropriate to offer it for the first time [10,16]. Introducing its use in humans through clinical trials is seen as essential, even if the design of such trials must accept some risks [5,6,16,39,55,58,59]. Many currently accepted reproductive technologies would not have been implemented without accepting some risk: e.g., IVF, prenatal genetic diagnosis (PGD) [5,22,14,27]. While the precautionary principle is suggested by some [10,22,29,50], others argue that prioritising precautions "stifles discovery or paralyzes scientific and technical progress…. To govern the introduction of new reproductive technology one could also adhere to a 'proof first' approach, placing the burdens on the regulator to demonstrate a high risk of serious harm" [22, p.673].
Finally, some raise concerns regarding the consequences that the prospective mother's health may have on the wellbeing of the child [31]. If the mother is affected by a serious mitochondrial disorder and consequently has a reduced life expectancy, she might have a limited ability to take care of her child.

ii) Harm to future generations
M/NT involves germ-line modifications that will be transmitted down the generations, thus complicating the risk/benefit analysis [61]. As with other reproductive technologies, the real impact of M/NT will not be known until numerous generations are born [31]. Thus, Baylis mentions the "right of subsequent generations to inherit an un-manipulated genome" [6, p.533-534].
As mitochondria are inherited maternally, male offspring do not transmit their potentially mutant mtDNA to progeny, which leads to a discussion of the ethical acceptability of selecting only male embryos for M/NT in order to avoid the transmission of health risks to subsequent generations [8,13,16,31,36,39,56,[62][63][64]. The UK Human Fertilisation and Embryology Authority (HFEA) ultimately rejected this idea, but recommends that any female born following M/NT be informed of the potential risks of transmitting mtDNA disease to her own children [8].

iii) Long-term follow-up
Because M/NT's effects may not manifest for many years, long-term and even intergenerational follow-up on the safety and efficacy of the technique is suggested, including "social research into how children born from mitochondria replacement feel about their origins" [39, p.5]. Some suggest making parental consent to long-term follow-up of resulting children a condition for participation in trials [16,31,54], but the HFEA recommends that permission to follow-up be obtained from the resulting children when they come of age [8]. Appleby argues that the procedure should only be offered to parents who plan on telling their child that s/he was born following the M/NT, since it would allow her/him to knowingly seek appropriate health care [36].
For some, long-term follow-up raises concerns regarding the burden put on children since "[b]eing made the enforced subjects of research over an indefinite period could be detrimental to their mental and emotional health" [30, p.7-8], while others believe that it would not be ethically problematic [16].
It is anticipated that it will be difficult to ensure long-term follow-up, since this has been a problem in the past for other assisted reproductive technologies [31]. A central register facilitating long-term follow-up that would be accessible to researchers is recommended [31,54], requiring government commitment to its maintenance [31,32,54].

iv) Harm to egg donors
Aside from ELSI related to regular egg donation, M/NT research requires an increase in availability of donated eggs, even as current shortage of donated eggs is acknowledged [22,30,31,32,59,65,66]. Moreover, eggs used for M/NT should be freshas opposed to cryopreserved -placing "time pressures on mitochondrial donors in terms of aligning their donation with someone's convenience other than their own" [30, p.11].

v) Harm to prospective parents
Risks of harm to prospective parents include failure to become pregnant [58,60], failure to give birth to a child [56], and health risks during childbirth due to the mitochondrial disorder affecting the mother-to-be [7]. Additionally, risks arising from hormonal stimulation and egg retrieval associated with IVF must be taken into account [6].

Reproductive autonomy
Reproductive autonomy is defined as "a principle concerning the non-interference in reproductive decision making" [31, p.70]. M/NT arguably widens the reproductive options available to women at risk of transmitting serious mtDNA diseases [16,29,31,37,[67][68][69][70][71]. However, "[b]ecause an experimental procedure is available, does that mean that every patient who wants the experimental procedure has a right to that procedure" [44, p.188] [see also : 48]? And "is the desire for a genetically related child a positive right and is there a corresponding social obligation to support its realization" [57]? Could M/NT "create needs people never knew they had" [Bonnicksen, in 57, p.19]? Conversely, potential pressure on women to undergo M/NT if it becomes available may result in diminished reproductive autonomy [30,39,56] and in the emergence of a parental duty to prevent the transmission of mtDNA diseases [30].

Available alternatives
Alternatives to M/NT include adoption, IVF with egg donation, pre-implantation genetic diagnosis (PGD) and pre-natal genetic diagnosis (PND). Given available alternatives for preventing the birth of an affected child and for parenting a healthy child, the question of the reasonableness of using the M/NT arises [31]. However, these alternatives are not without limitations. Adoption and IVF with egg donation do not address the parents' desire to have a genetically related child [8,[14][15][16]22,29,32,37,59].
Finding an egg donor can be challenging, especially when considering that the prospective mother's relatives may also be carriers [22]. PGD and PND are not a suitable technique for women with homoplasmy or a very high level of heteroplasmy [8,66,72]. On the other hand, PND can detect mutation in the fetus [22], but this is not necessarily decisive [10].
The drawbacks of these alternatives may be outweighed by being considered less ethically challenging than M/NT [10,28,73]. However, "PGD and other forms of reproductive technology are more prone to promote some kind of eugenics (by their active selection) than M/NT. […] [PGD] involves embryos that have already been created, rather than dealing with the potential of embryos" [14, p.4].

Consent i) Consent of the future child
The most commonly raised concern related to consent of the prospective child is similarly pertinent to already accepted reproductive technologies: while the prospective mother is the one undergoing the intervention, the one who bears most of the potential risks is the offspring, and who obviously could not consent at the time of the decision [5,[9][10][11]22,31,39,60,[74][75][76].

ii) Consent of prospective parents
While everyone can agree that participants in M/NT trials must be fully informed of the potential risks and benefits of M/NT, "[g]iven the uncertainty -particularly in the early days of human trials -as to the safety and/or efficacy of the new techniques, is it possible that a true, informed consent could actually be given" [29, p.84]? Indeed, while "new genetic technology widens reproductive options for couples, it makes them dependent on experts to make an informed decision, and concepts in mitochondrial inheritance are particularly difficult" [10, p.5].
Possible pressure to undergo M/NT can jeopardize free and informed consent [30]. Couples "may forgo rational considerations of risks, benefits, and long-term consequences" and "are likely to fall victim of the therapeutic misconception in which they 'deny the possibility that there may be major disadvantages to participating in clinical research that stem from the nature of the research process itself'" [Appelbaum et al., in 43, p.188]. The fact that researchers and clinicians will be working very closely might further obfuscate the frontiers between research and clinical practice [61]. There is a lot to learn about the way mitochondrial disease is "experienced, measured and communicated", including "how [the] mutation ratio combines with experience to provide estimations of risk and projections of the future, how normative decisions are made about acceptable levels of mitochondrial mutation and how the boundaries between normal and pathological are negotiated by both patients, families and health professionals" [47, p.9].
Informed consent will vary from one family to another and the acceptable level of risk will remain a personal matter [31]. The "difference across patients in the severity of expected offspring symptoms in the event that [M/NT] is not taken will shape the decision of choosing the treatment versus waiting for the outcomes of further research" [64, p.1346].

Impact on specific interest groups i) Persons suffering from diseases caused by mtDNA mutations
Some raise concerns regarding stigmatization of persons currently afflicted by mitochondrial disease [29,30,31,37], including a possible "knock-off effect on attitudes towards disabled people more generally" [39, p.18]. However, these arguments apply equally to currently socially accepted reproductive technologies aiming to avoid the transmission of genetic diseases [37,39]. Additionally, there is concern about the stigmatization of children born following M/NT as well as of parents who choose not to undergo the procedure [39].

ii) Scientists and researchers
The implementation of M/NT, or lack thereof, affects the relationship science has with society [10,22]. A specific area of research affected by M/NT is "[h]istorical and anthropological research on human population migration patterns and demographic history us[ing] mtDNA analysis and provid[ing] useful evidence of the geographical origins of humans, likely population sizes, and migration patterns" [6, p.533]. However, this issue was already addressed when egg and embryo donation were debated [27,30].

Resources
An important debate is centred on whether the costs associated with developing and introducing M/NT technology are justified by their supposed benefits [6,31]. Some argue costs are not justified because potential beneficiaries are "a very small minority for whom there are other reproductive options" [6, p.534]. However, others say that categorizing potential beneficiaries as a "very small minority" tends to diminish their pain [27]. The annual number of women who can benefit from M/NT in the UK had been estimated between 10 [37,74,77] and approximately 150 [69,71].
An argument for the costs of the technology not being justified is that it does nothing for those already affected by the targeted diseases, and that the money invested in M/NT research might be better spent on looking for treatment options [16,22,28]. However, "families may use their own resources to pursue M/NT…minimizing public expenditure while increasing scientific knowledge and experience" [37, p.345]. While it is true that the targeted conditions affect a relatively small portion of the population, the financial cost of mtDNA diseases on children and their families is significant [27,37]. The usual cost-benefit analysis calculations may not be directly applicable to the case of M/NT, since it prevents disease "at the very beginning of life", whereas many therapies used as examples in the debate take place "at the end of life", thus making the net gain of M/NT "considerable" [22, p.676], especially if everyone can access it regardless of their financial situation [16,23,31].
An aspect of the uncertainty involved in implementing a new technology is the way in which it can be applied for other purposes.

Slippery slope
A slippery slope is defined as "[i]ntroducing or accepting a technology or application A that in itself is not morally problematic, would be problematic if doing so makes it impossible to avoid the subsequent introduction or acceptance of another technology or application B that is morally unacceptable" [22, p.675].
The third slippery slope argument is as follows: "If PNT or MST were approved for treatment in a jurisdiction, and the techniques became accessible and acceptable to prospective parents, clinicians or patients might then ask to use the techniques for purposes other than the avoidance of the transmission of serious disease, or which have no therapeutic intention" [31, p.81].
Three examples of such "misuse" are given: a) using M/NT as an assisted reproductive technique for older, perimenopausal women seeking to have children [17,22,31,66,70]; b) using M/NT to establish a genetic link between the prospective child and more than two prospective parents when the prospective mother does not carry mtDNA mutations [6,27,30,31,44,73,82]; and c) using M/NT "to allow a woman with a major genetic problem in her nuclear genes to create a genetic link with her child through the use of her mitochondria, without passing on nuclear DNA" [31, p.82].
Fourth, there may be a legal slippery slope whereby "the lifting of the UK ban may facilitate lifting of the ban and initiation of mitochondrial replacement in other countries" [74, p.154]. This argument may be refuted by the ultimate independence of different jurisdictions' legal systems [81,84].
Some claim that the slippery slope effect is an important enough ethical concern to impede the adoption of M/NT: "once one kind of germ-line therapy is accepted, other kinds will almost certainly follow" [42, p.20] [see also : 72]. However, the slipperiness of the slopes may be doubted. A "clear legal distinction between modification to the different genomes [exists], thereby forming a practical barrier to the threat of 'slippery slope' arguments" [31, p.65] [see also : 27,37,39]. Moreover, "this slippery slope is purely philosophical -not methodological, and it is certainly one that is easy to prevent by regulations" [27, p.518]. While technological innovations often carry the potential of abuse with them, this does not necessarily mean that they should not be implemented when the use does not constitute abuse [27,51].
Some question the wrongness of all those things at the bottom of the slopes: eugenics, reproductive cloning, etc. [22,39]. They argue that these "horrors worthy of a dystopian fiction novel" [61, p.3] are perceived as such because of an argument of last resort, often used to vilify possible applications of genetics in assisted reproduction: "unnaturalness" [55, p.1965]. Moreover, it is pointed out that this kind of debate is not new and that similar fears arise when other new technologies first become available [31,51,55].

Creation, use and destruction of human embryos
The "development of [M/NT] may necessitate the creation of embryos especially for research" [22, p.672] [see also : 10]. While research on human embryos is unacceptable to some [22], others argue that it is ethically acceptable when it aims to alleviate human suffering, and more specifically, that of children at risk of inheriting mitochondrial diseases [37,82]. Adopting a similar perspective and starting from the premise that few alternatives are available to affected families, the Nuffield Council on Bioethics (NCB) Working Group "agreed that, with the appropriate oversight, research that may destroy or alter eggs or embryos (and which may develop treatments which require the same), is justifiable in seeking to prevent serious genetic illnesses being transmitted" [31, p.85].

Beneficence
Physicians have a "duty to act to benefit their patients -both prospective parents and children," and parents have a duty to ensure the children they bring into this world do not suffer needlessly [37, p.345]. Authors point out that M/NT "falls within the good medical practice of preventing serious illness" [29, p.88] and "could offer significant health and social benefits to individuals and families, who could potentially live their lives free from what can be very severe and debilitating disorders" [9, p.74].
While no one debates the validity of the beneficence argument, the question remains "whether the benefit of averting severe disease overrides societal objections to changing the human germ line" [70, p.827]. The lack of debate regarding what constitutes an "acceptable risk-benefit balance" is lamented. "[W]e may be facing a whole new set of ethical questions related to intergenerational risk/benefit analysis" [61, p.4], that may require novel approaches. Braude suggests that prospective parents are "best placed to balance the risks" and benefits of the technology [83, p.1].
How much weight do the benefits carry? Some authors argue that we are nothing less than morally obliged to reduce suffering in future people at risk of disease [30,82]. Others bring to our ethical attention "the potential to remove an entire class of adverse and truly devastating genetic mutations from affected families and possibly from the human population as a whole" [27, p.518] [see also : 26,37,63]. It may be our duty as parents and society to prevent future children from inheriting genetic disorders [30,60]. Yet, given the recognition that it is possible to prevent future suffering without recourse to M/NT, the risk/benefit question becomes: "Does fulfilling some prospective parents' desire to be genetically related to their child override societal objections to changing the human germ line?" [6,9,11,14,31,37,57,83] Indeed, the NCB's Working Group argues that "[i]n light of the health and social benefits to individuals and families living free from mitochondrial disorders, and where potential parents express a preference to have genetically-related children…if the PNT and MST techniques are proven to be acceptably safe and effective, on balance it would be ethical for families wishing to use them to do so" [31, p.88].
This view is based on the often-implicit recognition of prospective parents' desire to be genetically related to their child as "being natural or even instinctive, whilst also being influenced by contemporary cultural and social norms" [31, p.68]. Others, while acknowledging the importance of such a desire, question whether it constitutes a right [16,24,48], especially when juxtaposed with ethical concerns. Then again, such questioning puts in doubt the acceptability of other socially sanctioned reproductive technologies whose main benefit is ensuring genetic relatedness between parents and children [14,53].
While it may seem that the scientific and bioethics communities are largely undecided on the matter, the literature suggests that there is, in fact, almost a consensus, with 15 articles [26,29,31,34,37,39,41,51,53,61,63,83,[85][86][87] explicitly stating that "ethical concerns are outweighed by the arguments in favour of permitting mitochondria replacement" [39, p.4], and only 1 explicitly stating the opposite [24]. The remaining reviewed articles do not make explicit their position on where the ethical balance lies.

Limitations
A potential limitation of the present review is that a significant portion of the debate took place in the blogosphere as well as other types of popular media. While some blogs have a higher degree of credibility than others, it was decided that such texts should be excluded from a review of the scientific literature on the topic of M/NT. Nonetheless, the large number of commentaries, editorials, and letters to the editor that were included in the present review compensates for the omission of blogs and popular media. Moreover, aside from the usual caveats of qualitative analysis -albeit mitigated by the fact that two researchers coded the sources independently -several of the identified ELSI overlapped with one another (for example, the issue of financial "resources" invested in research on N/MT is often discussed in connection with the impact of M/NT on "persons suffering from diseases caused by mtDNA mutations", since the resources invested in N/MT are not intended to cure these people).

Discussion
The present critical interpretive review of the ELSI of M/NT highlights the issues most prevalent in the literature. Most of them -namely, identity of the child, relationship with the donor, parenthood, unknown risks to the offspring, risks to egg donors, the existence of available alternatives, consent of the yet unborn child, the impact of the technology on persons suffering of the same disability, the slippery slope argument, and the use of human embryos -have already been raised in relation to other assisted reproduction technologies that have since been accepted by society, such as IVF, PGD and intracytoplasmic sperm injection (ICSI).
The ELSI particular to M/NT are those stemming from the fact that M/NT involves germ-line modification, especially potential harms to future generations. Additionally, M/NT crossing the line into germ-line modification shapes the issues of intergenerational follow-up and the slippery slopes to nuclear DNA modification and misuse of the technique. These novel issues specific to M/NT call for a new framework for risk-benefit analysis, new research ethics guidelines, a register for longterm follow-up, a new system of contact or information exchange between the child and the donor, and an assessment of how M/NT will affect egg donors.
Arguments in the debate on the identity of the child go beyond those in the context of gamete donation, since they address the child's genetic link to three persons. Some of the arguments both for and against the implementation of M/NT are based on how the child's identity is affected by M/NT, and this is influenced by views on the nature of the genetic contribution of mtDNA [16]. These views in turn affect both the terminology chosen to refer to M/NT (e.g. "three-parent babies") and views on the relationship between the donor and child [39].
Given the diversity of potential ethical arguments against implementing M/NT in humans, it is possible to lose sight of the ethical argument in favour: beneficence. Some of the articles reviewed ignore this argument, possibly because the duty to benefit seems so obvious that it goes without saying and does not require discussion.
As M/NT was applied clinically to humans for the first time in 2015, its risks have yet to be studied. This contributes to much of the debate taking place in the media rather than in scientific articles. Commentaries, editorials, letters to the editor, and research news items made up the majority of the papers included in this review, which may lead to an overrepresentation of certain issues. Furthermore, of the 30 research articles reviewed, only two were based on empirical qualitative studies [16,47], neither of which analyzes specifically the views of M/NT's potential users (although Dimond has interviewed people suffering of mitochondrial diseases, they were not necessarily potential M/NT users). Even in the HFEA's public consultation, the focus group meeting included only 7 patients -some patients did, however, attend the "Open consultation meetings" that took place in two different cities [39]. Potential users are the target audience of M/NT and arguably the main stakeholder; they must therefore take significant part in the debate.
It is important to heed the calls of researchers, especially for follow-up research [5,10,24,54]. This should include longitudinal clinical research on all parties involved, as well as psychosocial longitudinal research to equip the ELSI debate with reliable evidence regarding questions such as: are there psychosocial harms and identity issues for the child and for the future generations, and if so, what are they? And what is the impact on egg donors, persons currently suffering of mtDNA diseases, and scientists?
Conceived the study (VR, RDL); designed the study (RDL, VR); carried out the search and coding (RDL, SB); analyzed the data (RDL, SB); wrote the manuscript (RDL, SB, VR); prepared, read, and approved the final draft for publication (RDL, SB, VR). This review was made possible thanks to the Bourse d'excellence of the Programmes de Bioéthique, Département de médecine sociale et préventive of the École de Santé publique de l'Université de Montréal. The authors would like to thank the reviewers for their very helpful comments that greatly improved the paper.
Stanislav Birko is an editor of the journal. He was not involved in the evaluation nor the review of this manuscript.
Reviewer evaluations are given serious consideration by the editors and authors in the preparation of manuscripts for publication. Nonetheless, being named as a reviewer does not necessarily denote approval of a manuscript; the editors of Canadian Journal of Bioethics take full responsibility for final acceptance and publication of an article.