In discussions of potential eugenics that have become inevitable in the CRISPR era a distinction keeps on being made between gene elimination and genetic enhancement. Typically, it seems to be implicitly assumed that these are the only options when it comes to human germline editing though, as we see in the recent moratorium proposal in Nature, there are other possibilities such as genetic “correction”–I will ignore these complexities here. In most discussions it is assumed that gene elimination is morally acceptable or at least less problematic than genetic enhancement.
Two arguments back this intuition. The first appeals to some general harm principle: we know that, in the case of eliminating harmful genes, we are far more likely to lessen total harm than increase it, something that may well be likely but is far more uncertain in the case of genetic enhancement. The second appeals to what has been a pretty well-worn historical consensus. Gene elimination is supposed to be our new, technologically more sophisticated, revitalized version of negative eugenics whereas genetic enhancement is a reincarnation of positive eugenics. Now, since the late 1930s, as positive eugenics began to come under increasingly critical scrutiny, negative eugenics has continued to have supporters. Though the distinction between gene elimination and genetic enhancement does not map cleanly into that between negative and positive eugenics, I will leave that argument for another occasion. I want to turn to an entirely different argument as to why gene elimination may well be less morally problematic than genetic enhancement.
A second implicit assumption about gene elimination versus genetic enhancement will be my focus here. This is the assumption that gene elimination would occur at the level of the population and that genetic enhancement at the level of the individual. (I will question this assumption below; let us work with it for the time being.)
Now, let us consider a prime–and, I hope, uncontroversial–case for gene elimination: the gene that causes Huntington’s disease. I will call it (this gene) HD for short. Strangely, there has been surprisingly little explicit attention to the conditions a gene must satisfy to be a valid target for elimination. What follows is an initial attempt to articulate these conditions (and comments are welcome) and to show that HD meets each of them.
We will begin with two harm conditions that try to ensure that the gene in question is genuinely harmful and that this judgment is not a matter of social prejudice:
- Deleterious effect condition: the gene in question is known to cause significant harm to an individual. Here, what “significant” means needs to be spelt out and one way of operationalizing it would be to say that the effect suffices to demand medical intervention as a routine choice. In a genetic context there are two complexities: the question of dominance and that of expressivity. For a dominant gene, a single copy suffices for the trait to be manifested; for a recessive gene two copies are necessary. The expressivity of a gene is the degree to which the trait is manifested (if at all). In the case of recessive genes, it makes sense (but this is not beyond question) to speak of expressivity only in the context in which the gene is homozygous (that is, the individual has two copies). I suggest that even recessive genes should be candidates for elimination so long as expressivity (which can be variable) is severe. In any case, HD is dominant and expressivity is almost always high though the problems may take years to get fully amplified.
- No bias condition: The determination that a gene leads to significant harm is clearly not made on the basis of culturally or otherwise relative societal values. That there is a problem here is indicated by the differences in medical responses across cultures to conditions such as mild hypertension and autism spectrum disorder. A proper discussion of this issue will take much more space than appropriate for this post and I will take it up elsewhere. Meanwhile we could use a rule that any serious disagreement about whether this condition is met should be taken to signify that it is not met. There is no such disagreement about HD.
Next we could impose a pragmatic limited options condition which could take a variety of forms. Here we codify so as to suggest we are driven to gene elimination because we have no other reasonable response.
- No management condition: Short of gene elimination there is no reasonable way to manage the trait that it causes. This is why gene elimination is attractive whereas it would seem beside the point to handle situations such as poor eyesight (assuming that we even have a genetic option). HD is exemplary with respect to this condition, as are genes for several other diseases such as cystic fibrosis and muscular dystrophy.
Next we would impose two standard technical conditions that are common in these discussions. They are both designed to ensure that our technology does indeed deliver what its proponents claim it can do and, given the tendency of enthusiasts to wax lyrical, it behooves us to display a healthy skepticism when stakes are as high as they are in human germline editing.
- Accuracy condition: Exactly the targeted gene–all of its relevant sequence and nothing else–would be removed (and, if so desired, replaced by an an innocuous alternative). We are not yet in a situation where we can confidently say that this condition can be fully met for any human gene (including HD). But this is where the promise of CRISPR-associated technologies lie.
- Safety condition: Endorsing safety is easy and paying lip service to it is even easier. But, in the context of human germline editing we are not close to achieving consensus on what it means and how it should be judged. Of course, that means that, no matter the extent to which HD satisfies the other conditions, we must still proceed with caution and ensure that this condition is met.
Finally, from my perspective, the most important conditions are two veridicality conditions. These are conditions that the relevant gene must satisfy (that is, they are not about the state of our technology). :
- High penetrance condition: The penetrance of a gene is the conditional probability that the corresponding trait is manifested given the presence of the gene. (We will assume that we interpret “presence of the gene” as requiring both copies in the case of recessive genes.) If a gene has low penetrance, the risk of intervention in the germline (and there always is some risk of unexpected consequences) is too high to endorse gene elimination. If a gene has high penetrance that risk may prove acceptable? But where should the cutoff be? This will indicate the amount of risk that a society is willing to tolerate which is a question about social norms. I have not seen any serious discussion of this issue and it is high time that it begins. (The case of HD is easy because it has penetrance close to 1 [or 100 %].)
- Gene specificity condition: I have explained this condition in an earlier post: modification of the intended gene results in an effect on the intended phenotype and only that intended phenotype. Suffice it here to for me to underline how important this condition it. We do not have experimental proof that HD satisfies this condition but the circumstantial evidence is good enough for me: we know how repeats in the mutation increases the severity of the disease and, as far as I can tell, has no other phenotypic effect. Stuart Newman deserves particular credit for recognizing the importance of specificity (without using the term).
In my view, once these conditions are satisfied, a gene can be an appropriate target for elimination and HD is a good example so long as the accuracy and safety conditions can be met. If any moral scruples are left, it seems to me that they would have to be of the type that would delegitimize any human germline editing (and, consequently, any eugenic measure of this sort).
What sort of reason could back such a perspective? A popular answer seems to be that we should not be “playing God”; that we should not consciously interfere into the future genetic profile of humanity. But any such claim has very strange consequences. Suppose you decide, personally, after finding out that you are a carrier of HD that you would not have any children. (I know of such individuals; this is not an artificial philosophers’ example.) You will have violated the “playing God” criterion. I wonder whether, perhaps, people’s discomfort with editing out genes such as HD ultimately arise from a distrust of what seems like very invasive technology penetrating into our bodies?
Now contrast this situation with genetic enhancement. The typical situation envisioned is one in which parents choose to enhance some trait, e.g., fair skin (very likely in India where I now am), intelligence (very likely in a wide variety of places), or some physical capacity related to athletic success (also very likely in a wide variety of places). Trouble arises that in many societies of the neoliberal North, in particular, the United States, such parental choices are regarded as well-nigh sacrosanct, For instance, parents can choose not to vaccinate their children for religious or even for nebulous “personal” or “philosophical” reasons.
In such circumstances, while there would presumably be agreement that the technical and veridicality conditions should be met, there would presumably be no agreement that parents must meet the no bias condition. And, thus, we become open to charges that we could not only be influenced by social prejudices but would even pander to them and possibly reify them. The Indian preference for fair skin (and it in under tangible genetic control) is a good example: born of the horrors of the caste system devised some thousands of years ago, it celebrates and panders to the racism popularized by Europeans for the last two centuries.
But, before I leave this topic altogether, two caveats are necessary. First, what if a parent chooses to remove a gene such as HD from offspring with no concern for its fate in the population as a whole? Wouldn’t such a decision be as subject to societal prejudice as in the case of genetic enhancement? Yes, but it would not qualify as gene elimination. Perhaps a better term would be genetic “correction” (used in Lander and colleagues’ moratorium proposal). Second, what if genetic enhancement is carried out for the whole population? What if it satisfies all the conditions listed above (which suitably altered harm and limited options conditions)? I have to say that this fantasy intrigues me but I will leave a discussion to another occasion.