Gyngell and colleagues consider that the recent Nuffield Council report does not go far enough: heritable genome editing (HGE) is not just justifiable in a few rare cases; instead, there is a moral imperative to undertake it. We agree that there is a moral argument for this, but in the real world it is mitigated by the fact that it is not usually possible to ensure a better life. We suggest that a moral imperative for HGE can currently only be concluded if one first buys into an overly deterministic view of a genome sequence, and the role of variation within in it, in the aetiology of the disease: most diseases cannot simply be attributed to specific genetic variants that we could edit away. Multiple, poorly understood genetic and environmental factors interact to influence the expression of diseases with a genetic component, even well understood ‘monogenic’ disorders. Population-level genome analyses are now demonstrating that many genetic ’mutations' are much less predictive than previously thought 1. Furthermore, HGE might introduce new risks just as it reduces old ones; or remove protections not yet clearly delineated.
Statistics from Altmetric.com
How can we know that edits would result in a ‘better’ life?
In order to use HGE responsibly, we would need to be confident that it would improve the welfare of the resultant person. There are currently no examples where we know at a molecular level exactly what HGE will do, but can be completely confident that there are no off-target or unintended effects. Genome editing technologies are often inefficient and sometimes imprecise,1 2 and there is no ‘normal’ genome to aim for. The potential harms of HGE are currently so unknown that it is far from clear that using HGE would enhance the welfare of the resulting child. For some ‘Mendelian’ conditions, the lifetime chance of manifesting the associated disease is lower than the chance of never doing so3: HGE of such variants could not be classed as a moral imperative . In some families with, for example, a dominantly inherited cancer risk, people who test negative for the ‘disease-causing’ variant in their family may still experience the associated cancer and HGE might not confer the hoped-for immunity from the disease. Multiple, poorly understood factors influence how genetic conditions are expressed within families, and population studies indicate that the predictive value of many genetic "mutations" varies significantly depending on the context in which they were ascertained4. Until we better understand what determines the penetrance of genomic variants, HGE is premature.
Value of ‘genetic relatedness’
Gyngell et al accept that many people have a strong preference for genetically related offspring. But what does this mean? Approximately 99.9% of the genetic code is the same in everyone5; the degree to which we share variation in the remaining 0.1% is therefore at stake. This consists of some 4– 5 million variants per person; 100,000 of these will be rare6 and so our knowledge of their clinical significance is usually limited and is often purely hypothetical. The impact of any given variant will depend on its context within a genomic background, and the influence of environmental factors, some of which will be stochastic.7 8 Attempting to predict the effect of each variant separately in a pointillist fashion oversimplifies this issue,9 and means that once we stray beyond considering HGE for well-understood, highly penetrant single gene disorders, we quickly cease to know what we might be doing. Furthermore, the more extensively we undertake HGE in a given embryo, the more we reduce the ‘genetic relatedness’ that forms a key part of the given rationale for this technology. Efforts to correct ‘normal’ properties (such as a 1 in 3 lifetime risk of developing cancer) or influence complex traits like intelligence will ultimately undermine the genetic relatedness that HGE is sold as preserving.
Polygenic risk distracts from the case for HGE
Polygenic risk is generated by a combination of variants scattered throughout the genome, most of which are not co-inherited. A child of a parent with a high polygenic risk of a particular condition will inherit only some of these and also inherit a complement of protective and susceptibility factors from the other parent, creating a completely new polygenic risk profile for him/her. Most importantly, many variants used in gauging polygenic risk are markers of disease risk, rather than agents of pathogenesis themselves, presumably lying physically close to the ‘real’ contributors to the disease. Applying HGE to these markers would be entirely futile as well as potentially harmful.10 Other variants are double-edged, for example, a variant in HBB might cause sickle cell trait while also conferring relative protection against malaria. Research to date has largely focused on finding susceptibility factors for diseases rather than protective factors, so many variants may have ‘positive’ effects on health of which we are unaware.
Although the authors acknowledge that HGE for polygenic risk is ‘a long way away from being plausible’, we argue that presenting HGE as a potential mitigator for polygenic risk unduly distracts from the most compelling arguments for HGE, which relate to the very unusual situations where a parent would inevitably pass on a severe, fully penetrant disorder every time he/she had a genetically related child. However, we think that these situations are rarer than Gyngell et al suggest, and we are concerned that arguing on a more widespread moral imperative for HGE might skew a public discourse—that already verges on an inappropriate genetic determinism—in the wrong direction. We would like to encourage more debate about HGE in these extremely rare cases, without diluting the arguments by including HGE for polygenic, multifactorial or poorly penetrant conditions.
We agree that the ready availability of technically perfect HGE might create responsibilities to use it in certain extremely unusual cases. We argue that polygenic risk is an inappropriate substrate to edit against, and that even for monogenic disorders, the relationship between the presence of a genomic variant and the expression of a disease is often more complex than the popular discourse around genomics might suggest. HGE may have an important role when considering the tiny number of people who do not have other routes to having genetically related children unaffected by severe, highly penetrant genetic conditions. Here, we need to look at the status we give to ‘genetic relatedness’, and the potential risk of off-target or unknown effects when using HGE that might undermine the aspiration to result in a ‘better’ life.
Contributors Both authors wrote and revised this commentary.
Funding AML’s work is supported by funding from a Wellcome Trust collaborative award 208053/Z/17/Z.
Competing interests None declared.
Provenance and peer review Not commissioned; internally peer reviewed.
Correction notice This article has been amended since it was first published online. This article has been changed from a Response to a Commentary article.
Patient consent for publication Not required.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.