The technical risks of human gene editing

Abstract

A recent report from Dr He Jiankui concerning the birth of twin girls harbouring mutations engineered by CRISPR/Cas nucleases has been met with international condemnation. Beside the serious ethical concerns, there are known technical risks associated with CRISPR/Cas gene editing which further raise questions about how these events could have been allowed to occur. Numerous studies have reported unexpected genomic mutation and mosaicism following the use of CRISPR/Cas nucleases, and it is currently unclear how prevalent these disadvantageous events are and how robust and sensitive the strategies to detect these unwanted events may be. Although Dr Jiankui’s study appears to have involved certain checks to ascertain these risks, the decision to implant the manipulated embryos, given these unknowns, must nonetheless be considered reckless. Here I review the technical concerns surrounding genome editing and consider the available data from Dr Jiankui in this context. Although the data remains unpublished, preventing a thorough assessment of what was performed, it seems clear that the rationale behind the undertaking was seriously flawed; the procedures involved substantial technical risks which, when added to the serious ethical concerns, fully justify the widespread criticism that the events have received.

Introduction

The development of site-specific nucleases over the last decade now makes it possible to introduce precise changes into the DNA sequence of our cells (Carroll, 2017). In particular, RNA-guided CRISPR/Cas nucleases are very easy to design against specific genomic target sequences and high efficiencies of mutagenesis can be achieved (Sander and Joung, 2014). These qualities are making the therapeutic application of CRISPR/Cas nucleases to tackle genetic disease feasible, and there has been a diverse range of success stories published in preclinical models (Porteus, 2019). For example, CRISPR/Cas9 nucleases have been designed to ablate the mutation responsible for muscular dystrophy, restoring normal gene expression (Long, et al., 2014). In both small (Nelson, et al., 2016) and large (Amoasii, et al., 2018) animal models, viral delivery of these nucleases into the diseased muscle was shown to restore muscle condition and strength. Such studies typify the exciting new field of therapeutic gene editing and highlight its potential in tackling genetic disease.

The target cell for therapeutic gene editing needs to be carefully considered to maximise the therapeutic potential and its longevity. Clearly, editing a stem cell population would have clear advantages for a prolonged therapeutic effect. Indeed, CRISPR/Cas nucleases have been introduced into haematopoietic stem cells to correct the underlying mutations responsible for sickle-cell anaemia (Vakulskas, et al., 2018) and there is thus considerable interest in this approach for long-lasting treatments for genetic diseases of the blood.

The application of gene editing tools in somatic stem cell therapies has raised the possibility that they could be applied in the ultimate stem cell, the one-cell embryo, allowing the genetic correction to be permanent and thus present in all cells of the resulting individual. Furthermore, a successful genetic manipulation of the one-cell embryo would lead to the presence of the genetic change in the germ cells of the manipulated individual, and thus the inheritance of these genome edits would be achieved. Early work introducing CRISPR/Cas nucleases into the one-cell mouse embryo by microinjection demonstrated that genetic modification of the whole organism could be achieved efficiently and confirmed the inheritance of the genetic change in subsequent generations (Wang et al., 2013).

CRISPR/Cas nucleases in human embryos

Reports soon emerged that CRISPR/Cas could indeed be used to manipulate the human one-cell embryo. The first published studies used discarded tripronuclear zygotes and achieved mutagenesis efficiencies of up to 50%, with specific gene editing (i.e. the incorporation of information from a co-injected repair template) occurring at around 15% (Kang et al., 2016; Liang et al., 2015). This was quickly followed by a further report of efficient mutagenesis in healthy two-pronuclear embryos (Tang, et al., 2017). Two more substantial articles followed, both published in Nature (Ma et al., 2017; Fogarty et al., 2017). The first demonstrated efficient correction of a dominant pathogenic mutation at the MYBPC3 gene, encoding a cardiac myosin-binding protein, in heterozygous human embryos, and proposed a mechanism of inter-homologue repair using the wild-type allele as a repair template (Ma et al., 2017). The second study focussed on the use of the technology to explore the role of genes involved in preimplantation human development and achieved targeted mutation of the gene encoding the transcription factor OCT4 (POU5F1) in 71% of manipulated embryos (Fogarty et al., 2017). More recent studies have successfully achieved gene editing using exogenous repair templates to repair pathogenic mutations (Tang et al., 2018) or to introduce specific reporter sequences (Yao et al., 2018).

An alternative strategy for achieving site-specific change within the genome has been reported: base editing, which relies upon the fusion of enzymatic domains to the CRISPR/Cas machinery, capable of chemically converting one nucleotide base to another (Rees and Liu 2018). Base editing has also been successfully applied within the human one-cell embryo (Li et al., 2017; Liang et al., 2018), and pathogenic mutations have been successfully corrected using this technology (Liang et al., 2017; Zeng et al., 2018).

These studies demonstrate that it is now feasible to achieve mutations and edits in human embryos at manageable frequencies and suggest that the tools for therapeutic germline editing are now available. There are numerous ethical concerns surrounding this technology which have been widely discussed and reviewed elsewhere (van Dijke et al., 2018), but a major requirement before germline editing can be considered is to assess the safety of the manipulations. Investigations of CRISPR/Cas9 mutagenesis both in cell culture experiments and in embryos have highlighted a number of disadvantageous consequences, where research is needed to assess and mitigate the risk and optimise methods permitting the reliable detection of such events. There thus remains an appreciable amount of technology development and optimisation to be done before therapeutic editing can be considered.

Off-target mutagenesis

Soon after the demonstration that CRISPR/Cas9 could be used for targeted manipulation of the mammalian genome, reports emerged that its use carries a risk of unintended mutagenesis at closely matched genomic sequences (Fu et al., 2013). This so-called off-target mutagenesis is also more pronounced than initially expected as the commonly used Cas9 enzyme can tolerate certain mismatches within its targeting sequence (Fig. 1A). Many of the studies addressing off-target mutagenesis have been performed in cell culture experiments where the CRISPR/Cas enzymes are transfected into millions of cells, the genomic DNA of which is then deep-sequenced to ascertain levels of accuracy. These types of experiment may overestimate the risk of off-target mutagenesis occurring when the CRISPR/Cas nucleases are applied in a single cell, i.e. the one-cell embryo. Indeed, one carefully controlled study in mouse used whole genome sequencing on a trio (sequencing both parental and offspring DNA) to address off-target mutagenesis resulting from a one-cell embryo microinjection experiment and the authors were unable to detect any events in the founder mice analysed (Iyer et al., 2018). In contrast, a larger study investigated founder rodent lines generated with multiple CRISPR/Cas9 enzymes addressing a number of different target sequences and found that almost 30% of the mutant lines harboured putative off-target mutations (Anderson et al., 2018). Interestingly, base editors designed to convert cytidine to thymidine residues were also found to have substantial off-target effects when applied within the mouse one-cell embryo (Zuo et al., 2019). Some of the human studies have also analysed the resulting embryos for off-target effects. Candidate off-target sites, localised using bioinformatic approaches, have been analysed by either Sanger or next generation sequencing. One of the studies confirmed an off-target mutation in two of the resulting embryos (Liang et al., 2015). All of the other studies found no evidence for significant levels of off-target mutation (Fogarty et al., 2017; Kang et al., 2016; Ma et al., 2017; Tang et al., 2017). In contrast with the results from the cytidine to thymidine base editors in mouse (Zuo et al., 2019), only low or entirely absent levels of off-target mutagenesis were detected when these reagents were applied in human embryos (Li et al., 2017; Liang et al., 2017; Zeng et al., 2018).

Figure 1.

Disadvantageous outcomes of CRISPR/Cas mutagenesis within the 1-cell embryo. (A) Off-target mutagenesis. In addition to the correct mutagenesis event, similar sequences elsewhere in the genome are also mutated leading to unpredictable effects. (B) Mosaicism. Prolonged activity of the nuclease within the developing embryo can lead to different mutations in different parts of the resulting individual. (C) Large deletions. The left panel shows the intended mutagenesis event where a target gene is inactivated. The right panel shows a potential consequence of a large deletion event, where a neighbouring gene is also inactivated. (D) On-site damage. The top panel shows the intended mutagenesis event, with a CRISPR/Cas nuclease specifically recognising only the mutant allele, leading to its inactivation. The bottom panel shows the biallelic mutation that could occur if the CRISPR/Cas nuclease is not able to discriminate the mutant sequence from the normal sequence.

The risk of off-target mutagenesis is thus clearly dependent upon the target sequence and can be reduced by designing CRISPR/Cas nucleases against truly unique genomic sequences; a number of online algorithms are available to facilitate this improved design (Haeussler et al., 2016; Hodgkins et al., 2015). These bioinformatic assessments of off-target risk can be somewhat flawed however, as many of the available tools do not take into consideration human genetic sequence variation. A true off-target profile and thus risk assessment for a selected CRISPR/Cas nuclease can only really be accomplished by establishing a personalised genome. Indeed, studies have suggested that naturally occurring human SNPs can alter the off-target landscape of site-specific nucleases substantially (Lessard et al., 2017).

It has also been shown that the concentration and persistence of the nuclease can increase the chance of off-target cleavage (Kim et al., 2014; Zuris et al., 2015), and subsequently, a number of approaches aimed at limiting the activity of the nuclease have been shown to reduce the level of off-target mutation (Chen et al., 2016; Shen et al., 2018a). Structural investigations and molecular evolution of the Cas9 nuclease have enabled the design of variant sequences which show increased levels of accuracy (Kleinstiver et al., 2016; Slaymaker et al., 2016) and reduced risk of off-target mutagenesis in rodent models (Anderson et al., 2018). Furthermore, orthologues of Cas nucleases from alternative bacterial species have been shown to have increased levels of accuracy (Kim et al., 2016; Teng et al., 2018). A recent study also improved the accuracy of CRISPR/Cas effectors by altering the structure of its cofactor guide-RNA (Kocak et al., 2019), highlighting a different approach to addressing this problem.

Taken together, there is a measurable risk of off-target mutagenesis when applying CRISPR/Cas nucleases in cells and embryos, but there has been significant technology refinement and bioinformatics tool development to reduce these risks substantially. Nonetheless, more research is needed in this area to assess the risk, improve the accuracy of the enzymes and explore methods for detecting these off-target events.

Mosaicism

The microinjection of CRISPR/Cas nucleases into mouse zygotes soon revealed that the CRISPR/Cas nucleases frequently retain activity after the first embryo cleavage event. Consequently, there is an appreciable risk that individual cells of the two-cell or even four-cell embryo harbour different combinations of wild-type and mutant alleles. The resulting organism is thus frequently a genetic mosaic, with different combinations of mutations in different parts of the animal (Fig. 1B). This phenomenon is easily demonstrated by microinjection experiments in mouse embryos which use CRISPR/Cas nucleases designed against coat colour genes. Addressing the tyrosinase gene, the loss of function of which leads to the albino phenotype, clear somatic mosaicism was evident as the majority of founder offspring show a speckled, patchy coat colour, rather than complete albinism (Yen et al., 2014). Clear evidence of mosaicism was also found in the human studies (Fogarty et al., 2017; Ma et al., 2017; Tang et al., 2017; Yao et al., 2018).

Mosaicism may be tackled by altering the timing of the nuclease activity within the one-cell embryo, effectively restricting its activity to the one-cell stage. Reducing the half-life of the nuclease through the use of a destabilised version of Cas9 was shown to reduce mosaicism whilst editing non-human primate embryos (Tu et al., 2017). Another improvement was achieved by delivering the nuclease to in vitro fertilised embryos by electroporation, permitting the delivery of the CRISPR/Cas machinery at a very early developmental stage, even before pronuclei have formed (Hashimoto et al., 2016). One study in human embryos was able to effectively eliminate mosaicism by introducing the CRISPR/Cas reagents at the same time as performing the fertilisation by intracytoplasmic sperm injection (Ma et al., 2017).

Large deletions and rearrangements

The mutagenesis occurring following the application of CRISPR/Cas nucleases relies upon the innate DNA repair machinery of the target cells. Most frequently, the induced double-strand break (DSB) is repaired by non-homologous end joining, which can lead to the introduction of small deletions and insertions. The mutations are frequently small in size, the most common being a single-nucleotide insertion or deletion (Chakrabarti et al., 2018; Taheri-Ghahfarokhi et al., 2018). Nonetheless, larger deletions do occur and there is evidence to suggest that, on occasion, the repair event can result in large kilobase-scale deletions. In one in vivo study, introducing CRISPR/Cas nucleases as a virus to correct a mutation in the Otc gene, an appreciable rate (6.5%) of disruptive large deletions was found (Yang et al., 2016). More recently, various studies have observed large deletions occurring at significant levels following the use of CRISPR/Cas9 in vitro (Kosicki et al., 2018) and when applied within the one-cell embryo (Parikh et al., 2015; Shin et al., 2017). These unexpectedly large deletions have the capacity to delete whole genes or cause misregulation of nearby expressed sequences (Fig. 1C).

In addition to large deletions, a mouse study has revealed that complex rearrangements can occur following the application of nucleases in the one-cell embryo (Boroviak et al., 2017). These events seem particularly prevalent when using multiple CRISPR/Cas nucleases which cleave in cis. The repair of the resulting two or more DSBs can result in the deletion of the intervening sequence (which is often the aim of the experiment), but also the inversion of the sequences can occur, as well as many surprising duplications and insertion events.

Interestingly, the prevalence of these deletions and rearrangements may have been underestimated since these complex events are often invisible to the molecular assays used to genotype the resulting mutations. Simple PCR-based genotyping strategies can be compromised by the deletion encompassing the primer binding sites. Short-read next generation sequencing technologies are not well equipped to detect and assess genomic inversions and duplications. Genomic technologies based on long-reads or more traditional assessment of target locus integrity by Southern blotting or fluorescent in situ hybridisation analysis may help detection, but these methods are difficult to apply and may not very applicable for genotyping embryo biopsy material.

The difficulty in detecting large deletions has fuelled controversy surrounding the inter-homologue repair mechanism proposed when CRISPR/Cas9 nucleases are used to selectively ablate a pathogenic mutation present heterozygously (Ma et al., 2017). It has been suggested that the inability to detect a rearranged or damaged mutant allele could lead to the misinterpretation that the allele has been repaired from the intact wild-type allele (Adikusuma et al., 2018; Egli et al., 2018), although follow-up analysis of the original study provided evidence arguing against this explanation (Ma et al., 2018). At the moment, we know too little about the dominant DNA repair machinery active within the early preimplantation embryos and these discussions highlight the requirement for further research to fully establish what repair events are likely, how they can be harnessed for therapeutic effect and how disadvantageous large deletions and rearrangements of the target locus can be detected.

On-site damage and biallelic modification

The high efficiency of CRISPR/Cas nucleases frequently leads to the mutagenesis of both autosomal copies of a target gene. Where a complete loss of function is therapeutic, this is, of course, advantageous; however, there are frequent situations where only one copy of a gene needs to be addressed, in particular when trying to correct or ablate dominant heterozygous mutations. Although CRISPR/Cas nucleases can be designed against the mutated copy of a gene, the tolerance of Cas9 for small mismatches may make it challenging to design nucleases that can discriminate between a mutant copy and a normal copy of a gene (Fig. 1D). Interestingly, in the human study which successfully applied CRISPR/Cas nucleases to correct a dominant heterozygous mutation in the MYBPC3 gene (Ma et al., 2017), the mutation chosen for this proof-of-concept study was a 4-bp deletion. This relatively large mutation allowed the nuclease to be designed specifically against this mutant allele, thus eliminating the risk of mutating the wild-type allele. The majority of disease-associated mutations, however, are single-nucleotide changes where discrimination may be challenging, and mutagenesis of the normal copy of the gene or even reprocessing and subsequent mutagenesis of the correctly repaired mutation would be expected to occur at appreciable frequencies. Development of enzymes with a higher level of discrimination may help the selective correction of the mutant alleles. The developed Cas enzymes and orthologues with reported higher accuracy may be very useful in this context (Kim et al., 2016; Kleinstiver et al., 2016; Slaymaker et al., 2016; Teng et al., 2018).

Another approach to help improve the predictability of gene editing outcomes is emerging from the analysis of large numbers of mutagenesis events. It has become clear that, for a specific target sequence, certain mutational outcomes can be quite common. Part of the explanation for this lies in regions of microhomology that lie upstream and downstream of the target site (Bae et al., 2014), but recent studies have identified other key attributes of the underlying primary sequence that can be used to predict the predominant pattern of repair (Allen et al., 2018; Chakrabarti et al., 2018; Shen et al., 2018b; Taheri-Ghahfarokhi et al., 2018). These studies demonstrate the power of using machine learning tools and large data set analysis to help make CRISPR/Cas mutagenesis more predictive. With the help of these new tools, it could be possible to achieve the desired repair of a mutation simply by design considerations alone. Indeed, one of the recent papers in this area has already confirmed the feasibility of this approach by using the predictive outcome of DSB processing to repair a pathogenic mutation (Shen et al., 2018b).

The reported birth of gene-edited twins

Despite technical improvements addressing the shortcomings, there remains uncertainty about the prevalence, extent and detection of genomic damage and mosaicism. Considerable research and technical development are needed to quantify and address the issues before therapeutic gene editing can be considered. Given these unresolved safety concerns, it was alarming to hear the reports emerging late last year from the Southern University of Science and Technology, Shenzhen, China, which suggested that human embryos had been manipulated by CRISPR/Cas9, reimplanted into the mother and carried to term. The principal scientist involved in this study, Dr He Jiankui, reported his results at the Second International Summit on Human Genome Editing in Hong Kong (National Academies of Sciences, 2019) and described the birth of twin girls, Lula and Nana, who both carried CRISPR/Cas9-engineered mutations. Dr Jiankui attempted to introduce loss-of-function mutations into the gene encoding the CCR5 chemokine receptor, a co-receptor for certain subtypes of HIV virus.

A naturally occurring CCR5 variant involving a deletion of 32 bp introduces a premature STOP codon into the gene, resulting in the expression of a truncated protein which is not able to act as an HIV co-receptor. Subsequently, individuals homozygous for this so called ∆32 mutation are resistant to infection by certain HIV subtypes. There has been wide interest in this mutation, since an HIV-infected patient has been effectively cured of viral infection by an allogeneic stem-cell transplantation with haematopoietic stem cells from a donor homozygous for this ∆32 CCR5 variant (Hutter et al., 2009). The aim of Dr Jiankui’s study was to engineer loss-of-function mutations within the CCR5 gene in human embryos generated by IVF from parents where the father was infected with HIV. In doing so, the goal was to protect the resulting embryos from HIV infection.

Flawed scientific rationale and experimental design

The scientific rationale behind the study is questionable for a number of reasons. Firstly, there are established protocols involving semen washing which can be used to reduce the risk of infection when using HIV-infected semen in assisted reproductive therapy (Zafer et al., 2016). There appears no need to invoke genome editing for this purpose. Secondly, CCR5 is a co-receptor for one subtype of HIV; a different chemokine receptor, CXCR4, can also act as a co-receptor for different classes of HIV. In patients with CCR5 mutations, CXCR4-tropic HIV subtypes can, albeit inefficiently, enter the cells through this alternative receptor which would continue to be expressed (Agrawal et al., 2004). Potentially the edited offspring would thus still be susceptible to HIV infection, despite engineered mutations in their CCR5 receptor gene.

With respect to the experimental design, the genome engineering strategy adopted did not involve the incorporation of the naturally occurring ∆32 mutation, despite the fact that a previous study in human embryos showed successfully that the ∆32 mutation could be incorporated at the CCR5 gene with a repair template (Kang et al., 2016). Instead, a random mutagenesis approach was adopted, albeit at the same position within the gene, which would be expected to lead to the incorporation of de novo mutations within the gene. The biological consequences of these novel mutations are impossible to predict and could lead to a global CCR5 knockout by affecting mRNA or protein stability.

Importantly, it has been suggested that stable expression of the CCR5 ∆32 variant may be important for HIV resistance (Agrawal et al., 2007); thus, mutations that cause a global knockout might not lead to the immunity which was the primary goal of the study. More worrying, results from Ccr5 knockout mouse models are revealing that there may be other consequences of CCR5 loss of function, besides HIV entry. Ccr5 loss of function led to an increased severity following infection of influenza virus (Falcon et al., 2015) and West Nile virus (Durrant et al., 2015) and have implicated CCR5 function in neuronal plasticity (Zhou et al., 2016) and recovery after brain injury (Joy et al., 2019). Another recent study explored human UK Biobank data to assess the impact of the CCR5 ∆32 variant on longevity and reported an estimated 21% increase in mortality for individuals homozygous for this mutation (Wei & Nielsen, 2019). Mutations at CCR5, especially those with uncharacterised consequences on protein stability and expression, might thus be expected to have unpredictable and disadvantageous consequences.

Genotyping data from Dr Jiankui’s study

The results of Dr Jiankui’s study remain unpublished, and thus, the primary data has not been peer-reviewed, making it difficult to assess the study. However, from his presentation at the Second International Summit on Human Genome Editing, we can learn some of the genotyping approaches he adopted in an attempt to mitigate the known problems of CRISPR/Cas nucleases outlined above.

Following CRISPR/Cas9 injection, the embryos were cultured to the blastocyst stage in vitro and the CCR5 genotype was then assessed by preimplantation genetic diagnosis (PGD) using whole genome sequencing on trophectoderm biopsies. Two embryos were selected for transplantation, the first of which (Lulu) harboured +1-bp and −4-bp frameshift alleles, predicted to encode truncation mutations that are similar but different from the natural occurring ∆32 mutation. As discussed above, the biological activity of these mutations is completely unclear.

The second embryo (Nana) harboured a 15-bp deletion on one allele, whilst the other allele remained unedited. This 15-bp deletion results in an in-frame deletion that would effectively remove five amino acids from the mature CCR5 peptide chain. Dr Jiankui hypothesised that this allele might destabilise the CCR5 structure near the HIV-binding site, but no experimental validation of this assumption was presented. The in-frame deletion would almost certainly result in a mature protein of completely unknown function; one could envisage a dominant effect which might compromise the function of organ systems in which CCR5 is known to play a role.

The fact that an embryo with an in-frame deletion and with a remaining intact copy of the CCR5 gene was selected for implantation is highly questionable. The unedited CCR5 allele would still be expressed, and thus, a normal version of this chemokine receptor would remain on the surface of T-cells, thus rendering the cells infectable by HIV, at odds with the primary goal of the study.

Dr Jiankui highlighted in his presentation that the parents made the decision concerning whether this embryo should be implanted, although it is entirely unclear how the parents were advised or counselled. Indeed, in the question-and-answer session that followed his presentation, it became clear that the scientists directly involved in the study may have performed the genetic counselling of the parents, rather than correctly qualified and trained genetic counsellors.

Known problems of CRISPR/Cas mutagenesis were addressed

Concerning off-target site mutagenesis, theoretical off-target sites were bioinformatically assessed using the parental genomes to allow the consequence of SNP variation to be considered. As mentioned above, naturally occurring genome variation has the potential to alter the off-target landscape substantially (Lessard, et al., 2017) and it is interesting that this study did indeed ascertain a personalised genome for this purpose. These theoretical off-targets were combined with experimentally reported off-target sites from published in vitro experiments which used the same target site to establish a panel of risk sites within the genome. PGD using whole genome sequencing on biopsied cells from the embryos revealed the presence of a single intergenic off-target site within Lulu’s genome. A decision was thus made to knowingly implant an embryo harbouring a CRISPR/Cas-induced mutation at an off-target site; although localised to an intergenic region of the genome, the functional consequences of this off-target mutation are unclear. Intriguingly, this off-target was no longer detected in fetal DNA analysis obtained from maternal blood during gestation and in the cord and placental samples obtained at birth, indicating that the initial PGD result might have been an artefact of whole genome amplification or reflect a low-level mosaicism in the trophectoderm cells biopsied.

Concerning mosaicism, the CRISPR/Cas9 reagents were applied during the ICSI fertilisation procedure using the same approach as adopted previously (Ma et al., 2017). Similarly, despite what appears to be mosaic sequence traces in the PGD Sanger sequencing, the results of the whole genome sequences revealed equal proportions of two alleles in each of the implanted embryos, suggesting that the embryos were not genetic mosaics.

Concerning large deletion analysis, the presence of large deletions was investigated by searching for chimeric sequencing reads arising from two regions of the genome in cis. Interestingly, in one edited embryo, not selected for implantation, a 6-kb deletion at the CCR5 target site was indeed found, confirming the prevalence of this kind of repair outcome. The embryos chosen for implantation revealed no evidence of large deletions.

Despite these relatively thorough sequencing experiments which, to a degree, seek to mitigate the known problems of CRISPR/Cas mutagenesis outlined above, it is unclear how thorough and complete the analysis was. Without an in-depth assessment of the primary data, it is impossible to know whether the investigations were sufficient to completely rule out non-specific mutagenesis events, large deletions or genomic rearrangements.

Conclusions

Given the above technical and scientific issues, combined with the grave ethical concerns (Krimsky, 2019), it is not surprising that Dr Jiankui’s study has been widely condemned as being a reckless and premature use of the technology. As a direct result, experts in the field have called for a moratorium on germline editing (Lander et al., 2019), which has been widely supported by the scientific community. It is important to recognise however that this moratorium concerns the implantation of edited embryos. Indeed, many scientists, clinicians, patient groups and ethicists support that research is needed to understand and address the risks involved. There is thus an understanding that this research may necessitate the use of human embryos, and the argument has been made that intentionally refraining from engaging in life-saving research is not morally defensible (Savulescu et al., 2015). However, at this point in time, at the very beginning of this emerging field with many unknowns, the implantation of edited embryos cannot be justified.

Research is needed to fully understand the repair outcomes occurring following the action of CRISPR/Cas9 nucleases within the early embryo, especially those involving repair templates. Further improvements in the accuracy of nucleases by either mutagenesis or molecular evolution, or by mining the bacterial and archeal kingdoms for alternative more accurate enzymes, would be advantageous. Improving methodologies for the detection of off-target mutation and large deletion and rearrangement events is also an area where continued research would be beneficial. Furthermore, a more thorough examination of the effects of base editing technology would also be of considerable interest.

PGD provides one alternative strategy for combatting genetic disease. However, there are concerns that, with reproductive age increasing in the Western world and given that PGD is known to impact reproductive success (Steffann et al., 2018), the number of viable embryos from which healthy individuals can be selected may frequently be too low for PGD selection to provide an effective solution. There are, of course, also situations where PGD cannot provide a solution, for example, where one patient is homozygous for a mutation. It is thus not too far-fetched to imagine a growing necessity to consider human germline editing in the future. It is thus clear that, in parallel with research addressing the safety concerns, the debate exploring ethical aspects of human germline editing must continue.

Funding

Wellcome Trust Core Award Grant (203141/Z/16/Z).

Conflict of interest

The author has nothing to declare.