Bioethics

The other argument against allowing genome editing for clinical uses is concern for the safety of future offspring (and of this offspring’s offspring). At this stage, safety is indeed an issue and the efficiency of genome editing on embryos remains low, with mosaic embryos (i.e. embryos that have abnormal numbers of chromosomes in certain cells resulting in genetically different cells coexisting in the same organism) being the main known drawback of these technologies (Vassena et al. 2016). Despite this, some studies have proven the feasibility of gene editing in animals (Heo et al. 2014; Shao et al. 2014; Yoshimi et al. 2014; Zou et al. 2015), even though the efficiency of genetically modifying zygotes with Cas9 ranges between 0.5 and 40% (Araki and Ishii 2014). In addition, a recent study demonstrated the feasibility of preventing the onset of a genetic disorder such as cataract development (Wu et al. 2013) and the injection of Cas9 into primate zygotes led to the birth of genetically modified offspring (Liu et al. 2014; Niu et al. 2014).

There are two sets of arguments for introducing CRISPR and other gene editing technologies into the clinic, provided that safety concerns are properly addressed. In this section I first outline the first group of arguments, which concerns the benefits of genome editing for future children (and their children too) and for prospective parents (Gyngell et al. 2016; Reyes and Lanner 2017). In the following section, I present additional reasons why genome editing could be a morally preferable alternative to PGD: genome editing would not be subjected to some of the critiques moved against PGD.

The moral reasons that ground the case for PGD (the welfare of future children and the reproductive autonomy of prospective parents; Pennings et al. 2007; Buchanan et al. 2001; Harris 1992) can be extended to defend the clinical use of genome editing in reproduction. It is widely accepted that reproductive autonomy and respect for parental discretion in reproduction are values worth defending 4 (Buchanan et al. 2001; Harris 1992; Robertson 1996). With respect to reproductive autonomy, genome editing would be comparatively better than PGD: it would offer an alternative to this technology for those aforementioned cases where PGD is not effective or for prospective parents who wish to increase their chances of having mutation‐free embryos. In this sense, genome editing could be said to enhance reproductive autonomy. With respect to the welfare of the child, the case in favour of genome editing seems prima facie stronger than the case in favour of PGD. Unlike the latter technology, whereby embryos implanted can be carriers of the parents’ mutated gene, genome editing would allow modification of the genetic makeup of embryos who would consequently develop into mutation‐free offspring. In other words, genome editing would prevent the occurrence of genetic diseases in future generations, while PGD can sometimes only prevent the occurrence of genetic diseases in the child that develops from the implanted embryo (Gyngell et al. 2016).

There are, however, other arguments in favour of preferring genome editing to PGD. PGD is a contested practice as its scopes are not therapeutic (i.e. PGD does not treat embryos) but rather selective (i.e. PGD selects the embryos that should be transferred in utero. Asch and Barlevy 2012; Parens and Asch 2003). PGD as a means to select embryos that have a decreased risk of developing into a child with a genetic condition is seen as ethically troubling for two reasons: firstly, because it goes against the traditional ends of medicine and ‘selects out’ rather than ‘cures’ persons affected by genetic conditions (MacKellar and Bechtel 2014), and secondly, because decisions on which embryos should be selected are said to embody value judgements regarding people living with certain disabilities (Knoppers et al. 2006; Parens and Ash 2003), a critique of screening technologies that became known as the ‘expressivist argument’ or ‘expressivist objection’ (Buchanan 1996; Shakespeare 2006).

PGD (at the moment) and CRISPR (potentially in the future) are two technologies that enable similar ends: in both cases, these technologies increase the chances of giving parents genetically related offspring unaffected by specific genetic conditions. Despite the similarity of the outcomes (i.e. healthy child), the means used are rather different. PGD is a form of genetic testing that allows screening for abnormalities in early embryos and to subsequently implant only those with a decreased risk of developing a certain condition. Instead, CRISPR and other gene editing technologies are tools for gene therapy that allow the modification of embryos or of gamete cells in order to avoid the occurrence of certain conditions in the future child (and in future generations).

Following this distinction of means, there is a sense that while PGD entails the selection of embryos, CRISPR is more akin to therapy . At this point, however, it is important to note that CRISPR and other genome editing technologies can be considered both therapeutic and non‐strictly‐therapeutic (or, following Wrigley et al. “pre‐emptively therapeutic”; Wrigley et al. 2015, p. 636). I am not trying to violate Aristotle’s principle of non‐contradiction on the impossibility that contradictory assertions can be both true at the same time here. What I mean is rather that whether these technologies are therapeutic depends on what sort of factual and moral considerations are taken into account. If the focus is on the prospective parents, then CRISPR can be considered therapeutic in some instances because it could be a solution (or a treatment?) for those couples who would not otherwise be able to conceive children that are related to them and that are free from the risk of developing (or have a decreased risk to develop) the condition that affects them.

If the focus is on the future children, we have two possible interpretations: following the view that equates embryos with persons, CRISPR is therapeutic because it treats the embryos (i.e. it treats persons), whereas PGD is selective because it selects in/out the embryos (i.e. it selects out persons). If, however, we are more inclined to think of embryos as beings with the potential to develop into persons (i.e. potentiality view, arguably a more widely shared position), then CRISPR is not straightforwardly therapeutic, because there is no person to be treated at the moment that we use the technology. 5 Despite this remark, I argue that there is a sense whereby genome editing can still be considered therapeutic, or, as mentioned above, pre‐emptively therapeutic. In order to assess whether CRISPR can be considered pre‐emptively therapeutic, it is necessary to determine whether embryo X (i.e. the embryo that exists prior to the application of CRISPR) is identical to newborn X + about 9 months(i.e. the child that is born after the application of CRISPR on embryo X). This assessment matters for the ethical debate on PGD and genome editing because if these two entities (embryo X and newborn X + about 9 months) are identical, then PGD would be more problematic than CRISPR as the first would be a selective technology, whereas the second would be a therapeutic technology. A brief explanation of the question of identity is needed before proceeding with the discussion on PGD and CRISPR and the ethics thereof. Currently, ethicists and philosophers involved in the debate on reproductive genetic technologies seem to be divided on whether genome editing technologies applied to embryos are identity‐affecting technologies or not, as this largely depends on the circumstances taken into account. 6 When I say “identity‐affecting” I refer to the idea of numerical identity and to the metaphysical problem of determining how we can rightly refer to one and the same person in any different set of circumstances, despite the changes that the person undergoes over time. Thus, for instance, there is numerical identity between a person X and a person Y only if person X and Y are the same person. To put it simply, I am numerically identical to the person that is writing this paper at the moment. The challenge of any account of numerical identity is then to explain what determines the entity that we in fact are despite the changes that we undergo over time. In this sense, if I grow taller or if I lose an eye due to an accident, I am still numerically identical to the entity I was before having that accident or when I was shorter. This is the case because changes such as losing an eye or growing taller are largely considered contingent to numerical identity, namely they do not change the entity that I am.