Editing human embryos

The paper published last week by Junjiu Huang and colleagues (1) on gene editing in human embryos (see BioNews 799) is, I expect, the first of several that we will see this year. There has been much excitement among scientists about the power of new gene-editing methods - particularly the CRISPR/Cas9 system, which is relatively simple to use and generally very efficient. The method has already been used in a wide variety of animal models and in human cell lines. The possibility of using such methods to genetically modify human embryos has been on the cards since these methods were first described, and recently some senior scientists and commentators have called for a moratorium to halt any such attempts.

I disagree with such a moratorium, which is in any case unlikely to be effective. I am fully supportive of research being carried out on early human embryos in vitro, especially on embryos that are not required for reproduction and would otherwise be discarded. If the techniques work, there are many interesting questions that could be asked about the role of specific genes in early human embryo development, especially as there is accumulating evidence that the stages of development of human of embryos may differ from that of other mammals. For example, we have known for a long time that human trophoblast cells types, which make up much of the placenta, are different from those in mice. And a recent study from Azim Surani's group revealed how Sox17 is critical for germ-cell development from human embryonic stem cells, whereas the same gene is not required in mice (2).

The paper from China is the first to ask if the CRISPR/Cas9 methods work in human embryos, and the answer provided is very equivocal. Yes, they do, but inefficiently and with several problems. In particular, there are 'off target' effects, where genes distinct from that being targeted (which in this case is the gene encoding beta-globin, a red blood cell protein), have ended up being mutatede. This has occurred at a much higher frequency than has been found in other cases where the techniques have been applied to human cells in culture or to mouse embryos. Moreover, even within the beta-globin gene, there was a high frequency of incorrect editing, resulting in yet more errors in the DNA sequence.

These problems lead the authors to suggest that a lot more work will be required to alter the techniques for use in human embryos. However, there are a few issues with the design of the experiments. First, the authors used abnormally fertilised embryos, presumably because they did not want to be accused of using embryos that could undergo development to term if implanted. However, it is possible that inappropriate DNA repair mechanisms had been activated in such abnormal embryos and that these caused the additional errors. Second, it would have made sense to test out the techniques and reagents (notably the guide RNA) using human embryonic stem cells, before actually using them on human embryos. They also chose a gene target that might itself be problematic, given that it is part of a closely linked family of globin genes with highly related sequences, making it hard to target one without affecting the others.

There are already variations on the CRISPR/Cas9 method that might solve some of the problems encountered. And ways to improve the efficiency and fidelity of the editing process are fast appearing in print, so the methods are getting close to the point where we can no longer use the simple excuse that they should not be applied in humans because they may be unsafe. This makes it important that we consider the reasons why gene editing might be applied to humans, whether these are justified, and in what circumstances.

The Nature News (3) report on this research suggests that previous attempts by the authors to publish their work had failed at least in part due to ethical issues. Had this work been done in the UK, under the excellent regulatory system provided by the Human Fertilisation and Embryology Authority (HFEA), any ethical concerns would have been solved before the work could have started. Indeed, with an HFEA licence, research of this sort could be conducted in the UK, and, with justification, it would be possible to use normally fertilised embryos. The 14-day limit on research in vitro would apply and it would be illegal in the UK to implant any such manipulated embryo into a woman for further development. 

In China, there are guidelines issued by the central government, rather than laws governing research and clinical applications using human embryos. These guidelines tend to be followed because it is not a good idea to fall on the wrong side of the Chinese Government.

In the USA, it is not possible to do such research with federal funding, and although some states have rules that would forbid it irrespective of funding source, it would be legal with non-federal funds in many states. Indeed, in addition to research in vitro, it would in theory be possible in some states to actually implant the embryos. There is no equivalent to the HFEA in the USA, and  no federal rules governing assisted conception. Clinics are supposed to follow guidelines issued by professional organisations, such as the American Society for Reproductive Medicine, but many do not, or they are selective about which parts of the guidelines to follow. There may be local ethical review committees that clinics need to consult, but these are variable in quality and in the decisions they make. 

Perhaps the Food and Drug Administration could impose a US-wide ban on clinical application of genome editing as they did on methods of mitochondrial replacement on the grounds that the gene editing methods are a form of gene therapy, which is in their remit. But this would contrast with other US regulators, who have already indicated that crop varieties generated using these techniques may not constitute genetically modified organisms, because the changes can be so subtle (4).

It would be impossible to do these types of experiment in some European countries, notably Germany and Italy. But other countries, such as Spain and Sweden, have rather liberal laws, and they may well permit research in vitro. In Germany it is only possible to carry out a procedure on an early human embryo that would not cause it harm. For example, preimplantation genetic diagnosis (PGD) was until 2011 illegal in Germany because embryos that are found to have a genetic defect have to be discarded and therefore 'harmed'. It might be interesting to ask, if the gene editing would be of benefit to the embryo would it be legal to transfer it into a woman to obtain a child? I suspect the answer would come back 'no' as this might constitute an affront to 'human dignity', but it has never been clear - at least to me - how 'human dignity' can apply to a pre-implantation embryo. If the gene editing had corrected a gene defect, in what way would the embryo have been harmed?

Much of the fuss about the possibility of germline gene editing is misplaced, in my view, because there are very few instances where it would be necessary to correct a gene defect, which was the ultimate aim of the work reported by Huang and colleagues. For recessive single-gene defects - such as beta-thalassaemia - alternative techniques, notably PGD, can be used to choose embryos for implantation that would be free from disease.

The situation and the arguments for or against become less clear when dealing with dominant mutations, sex-linked diseases, or where multiple gene defects may be present within an embryo. The arguments become even more contentious when dealing with 'enhancement'. However, while we work towards using the methods to make disease-resistant crops and animals, should we deny this possibility for humans?

There is a need for caution, but also for reasoned and well-informed debate. Only then can we have appropriate and proportionate regulations to govern the use of these powerful and important techniques.

This article was originally published under the title 'No need for a moratorium on genetically modified humans'. This was not the author's title, and did not entirely reflect his views. BioNews apologises for any misunderstanding this may have caused.