Overview
Manipulation of genomes is hardly new. Processes directly aimed at genetic modification were first developed in the 1970s and less directly we have been altering genomes for millennia through selective breeding. Recent technologies, most notably CRISPRCas9, nonetheless mark something new, the possibility of precisely specified and targeted changes to genome sequence. These advances have greatly enhanced the prospects for genomic enhancements of animals, and specifically of animals raised for food.
Although philosophers tend to be skeptical of ambitious claims made for future scientific achievements, for the purposes of this paper I shall assume that we are close to being able to make any arbitrarily selected change to living genomes and, importantly, with ever decreasing probability of off-target effects, changes to untargeted parts of the genome. This could be treated as simply an assumption for the sake of argument, but actually I think there are good reasons to treat the assumption as realistic. The progress achieved with CRISPR technology, and its spread through the biological research community, have been extraordinary. CRISPR-Cas9 may not be the technology at the end of this road, but the principles and techniques it embodies are likely to continue to lead to ever more precision in the so-called editing of genomes.
The second assumption underlying this talk will sound a lot less optimistic. Although I anticipate ever greater precision in the manipulation of genomes, I do not infer that this will lead to comparable ability to control phenotypes. Where phenotypic traits depend wholly or largely on a single gene of very large effect, it will indeed be possible to induce or prevent the effect. But a very small proportion of traits are of this kind. Where, as is typical, traits depend on many genes and many aspects of the environment in which the organism develops, the effects achievable with genomic editing will be quite limited. When a trait depends on many genes, it is almost certain that many of these will affect other traits, so that even if it were possible to alter many or all of them this would be likely to have many unwanted side effects. When the effects of a gene differ according to external aspects of the developmental environment genomic modification can only be effective in close coordination with control of the relevant environmental features. Hence, the possible range of modifiable traits is uncertain, but the range is a good deal less than is often supposed.
The magical genome
The genome is still frequently explained by appeal to metaphors such as ‘blueprint’, ‘recipe’, ‘programme’, etc. These metaphors suggest that ‘rewriting’ should allow almost indefinite changes in the phenotype, an idea that depicts the genome as having almost magical powers. If you want a pig with wings, should you not be able to draw the wings on your blueprint? Silly examples apart, the genome has no magical powers to determine the phenotype, which as a whole develops in response to a wide range of internal and environmental factors.
A more realistic, if disappointingly mundane view of the genome is as a particular kind of data store, providing the information necessary for the production of particular materials, primarily proteins. This account by no means excludes dramatic results from genomic modification. The cases most discussed in medical context concern the correction of debilitating genetic faults, but in the case of livestock such problems are much more cheaply, effectively, and on most views ethically, addressed by selective culling. There may, however, be developmental switches based, for example, on the availability or unavailability of a protein, or a protein with a different quality. This may change the course of development though generally in a way dependent on other parts of the genome and on other features of the environment. Examples that appear to be very largely free from unwanted side effects include double muscling in cattle breeds such as the Belgian Blue, which results from inactivation of the myostatin gene, and the polled gene discussed further below.
But for most traits, as the rapidly spreading technique of genome wide association studies (GWAS) hasdemonstrated, there are a large number of genes that have generally very small effects on the trait. While it might in theory be possible to boost the value of a variable trait to some extent by editing genes at a number of the relevant loci, the crucial problem with any such strategy is that these genes are likely to have multiple unknown effects on other traits.
Applications
A number of feasible and valuable modifications have been identified in various livestock species. One potentially important example is the editing out of PERVs, endogenous viruses in pigs that provided serious potential risks in using porcinederived organs for human xenotransplantation. Several significant examples have been described in cattle. Best known of these is the production of polled dairy cattle, replacing a gene found in Holsteins with one found in hornless Aberdeen Angus. This is a paradigmatic example of a developmental switch in a trait sufficiently peripheral to overall development to have, very probably, no unwanted side effects.
A rather different example is the enhancement of TB resistance. This has been achieved by Chinese scientists by adding an extra copy of a gene known to be associated with higher TB resistance, and appears to have further enhanced this resistance without any adverse side effects having yet been reported. It is not supposed that this change will eliminate bovine TB, but there is evidence that it will significantly mitigate the problem.
Ethical issues
There is no doubt that the use of this technology is likely to be controversial, and probably especially so in large animals for which people feel a good deal of empathy. In the remainder of this talk I shall consider some of the grounds that have been proposed for this ethical hostility, and very briefly assess how seriously these ethical considerations should be taken. A problem that should be highlighted at the outset, and which will be obvious enough from the experience of first generation GM crops, is that ultimately the decisive question will be less about abstract evaluation of perhaps subtle ethical arguments and more about public acceptance.
This last point highlights the importance of the so-called ‘Yuck Factor’, a term coined by bioethicist Arthur Caplan to refer generally to hostility to new technologies. Much more specifically, this has come to refer to a reaction to GM, especially transgenics. Since there is little reason in contemporary scientific practice to use actual material from other organisms, it can be argued that transgenics are no longer an issue, and this can be expected to mitigate future yuck reactions. Indeed, several regulators, including the US FDA, have decided that for this reason genomically edited organisms using technologies such as CRISPR should not count as genetically modified organisms (GMO).
This would, of course, be a very welcome decision for defenders of genomic editing, as it would potentially release these new technologies from the burdens of regulation that developed around GMOs. I must say, however, that I find this line of argument flimsy, and a dangerously insecure basis on which to rest the legitimation of a technological practice. Why should it be important whether a molecule inserted into a genome is derived whole from another organism or precisely the same molecule is synthesised from scratch? My own view is that to address the apparent widespread revulsion against genomic modification we need again to confront magical or mystical understandings of the genome, understandings that provide the background to frequent demands for purity and freedom from artificial contamination.
Another type of argument that is often brought against genomic modification generally, is the socalled slippery slope argument. At the bottom of the slope, very often is to be found a largely imaginary creature, the Designer Baby, though this is a long way down the slope from a cow with enhanced disease resistance. Perhaps we should think rather of the cow in The Hitchhiker’s Guide to the Galaxy that wants to be eaten. The general strategy is to suggest that even if the uses currently proposed for the technology look fairly harmless, if it is permitted it will soon be put to uses that are abominable.
As with the ‘Yuck’ arguments, one important response is to be more realistic about the kinds of applications that are likely to be possible. If one imagines that a genome is a plan on which just about anything can be written that takes the fancy of the planner, then the possible applications of genome editing are indeed alarming. But once we recognise that the genome is just one part of a highly integrated and coordinated system, then there will be very serious limits on the changes to that system possible simply by manipulating one part. Particular proposals should certainly be given very close scrutiny. But the idea that any arbitrary property might be written into any chosen organism is pure fantasy. To be convincing, therefore, slippery slow arguments need to be specific about the hazards that lie at the bottom of the slope and these, if necessary, should be amenable to specific regulation or legislation.
When we do turn to more relevant scrutiny of actual proposals for genome editing, a vital issue will of course be consideration of animal welfare. It is certainly possible to breed, and a fortiori to design, animals that have intrinsically higher suffering: flat-nosed dogs that have trouble breathing, and perhaps very high milk yield cows chronically subject to mastitis. A question that is pressing in assessing relevant aspects of animal welfare is what one takes as given. If the goals of husbandry are taken as given by the inescapable pressures of economic competition, for example, then modifying animals better to fit the conditions determined by these goals might be an effective way of improving animal welfare. Polled breeds of cattle, whether naturally or artificially made to be that way, avoid the pain of dehorning or disbudding, and moreover avoid the danger of causing harm to one another with their horns.
But perhaps animal welfare should come into the story earlier. If producing hornless cattle in whatever way is seen as a means to increase stocking density, which in turn is taken to reduce animal welfare, the process may be seen as a means to an ethically bad end. This thought leads me to the final point I want to address, that the question whether genome modification is acceptable, whether ethically or just to public opinion, may be difficult to extricate from broader discussions of the role of livestock production in the future of national or global food supply.
The bigger picture
Genome editing of plant crops has been promoted as an essential part of the effort to address future global food needs, and it is likely that the same argument will be deployed in support of genome editing of livestock. As readers of this publication will no doubt be aware, however, the role of cattle, especially in the production of meat, is controversial in respect of its impact on the environment. A 2014 study published in the US Proceedings of the National Academy of Sciences claimed that ending consumption of red meat would have a more strongly favourable impact on the environment than would abandoning the use of cars. The same study claimed that beef production, in particular, requires about an order of magnitude more resources than other modes of meat production. Claims of this kind are likely to feed resistance to applying major scientific investment in an already controversial technology to beef production.
The obverse of this argument is that if such changes can significantly decrease the environmental impact of beef production, it might provide a strong argument in favour of the technology. Of course, cattle are reared in different ways in a wide variety of environmental contexts and, as already noted, particular genomic modifications may have their desired effects only in specific environmental conditions. An important aspect of any environmental discussion in this area will be the range of alternative uses for land of particular kinds. Changes to the cattle stock that enable environmentally more favourable cattle rearing on land that is poorly suited to alternative uses are likely to be more warmly received than changes that enable higher intensity farming on land with alternative uses that have less damaging environmental impacts. And environmental impact here should be construed broadly: many people feel very strongly, for instance, about the importance of maintaining the kinds of landscape characteristic of much cattle rearing in this country.
While I have neither the expertise nor the space to assess these issues in any detail, I think it is safe to say that embedding the debate about genomic modification in a wider discussion of the social advantages and disadvantages of the forms of livestock production that they enable or enhance is likely to be an important feature of the ongoing discussion. Proposals that offer environmental enhancements are likely to receive a more friendly public reception than those seen to offer only improved profit margins.
Conclusions
The first step, I have argued, towards productive discussion of the issues raised by genome editing is to attempt to counter common misunderstandings of the nature of the genome. I have suggested that given a realistic understanding of the genome there is no clear general objection to genome editing of livestock. There may, however, be various specific concerns about animal welfare. Assessing these concerns will inevitably be sensitive to specification of the kinds of animal rearing that are in view, and these, in turn, feed into broad questions about the overall character of food production, and the wide array of positive and negative effects that different methods of animal rearing may have on the environment, broadly construed. If genome editing is defended on the basis of concern about future global food supplies, it is hard to avoid wider questions about the appropriate role of animals in the human food chain.