August 8, 2018

expert reaction to critiques and the authors’ response with respect to an original report of the use of genome editing to correct a mutation in human embryos

On 2 August 2017, Nature published a paper by Shoukhrat Mitalipov and colleagues, Correction of a pathogenic gene mutation in human embryos, which reported the correction of a mutation implicated in a heritable heart condition in preimplantation human embryos using the CRISPR–Cas9 genome editing technique.

Nature has now published two Brief Communications Arising (BCAs) by Dieter Egli and colleagues and by Paul Thomas and colleagues. These BCAs are published alongside a Reply from the original authors, Shoukhrat Mitalipov and colleagues.

Dr Helen O’Neill, Programme Director, Reproductive Science and Women’s Health, University College London, said:

“While the original Ma et al paper was undoubtedly a landmark study in that it repaired a mutation in human embryos, it received criticism for the unanswered questions or, indeed, assumed interpretations that the results generated.

“One of these interpretations was that repair of the mutation (which was paternal in origin) had been mediated by the maternal genome, so despite providing a separate (exogenous) piece of DNA which they had assumed would be used by the embryos repair mechanism, instead, endogenous (maternal) DNA was used to correct it (a process called inter-homologue repair or IHR).

“At the earliest stages of fertilisation (before the embryo has even divided into two cells), both maternal and paternal genomes are present in close but separate proximity. One criticism of this assumption (Egli et al), therefore, was that this process would require the physical interaction of the maternal and paternal genomes and that at this stage of development, the two genomes are segregated in separate pronuclei so this would not be possible.

“This was addressed here by Ma et al and also in work carried out in MIT (Wilde et al) who challenged Elgi et al showing independent examples to support the idea that indeed IHC can occur (at least in mouse embryos and embryonic stem cells).

“The recent communication in response by Ma et al suggests that while it is true that they (parental genomes) are located in separate locations, that the inserted molecular scissors (Cas9) can remain for many hours in the embryo, perhaps indeed up to and after the parental genomes have joined.

“Further criticism of Ma et al suggested that other means of repair had not been ruled out and that indeed the depth of analysis was not sufficient to justify this assumption. For example, that Ma et al could not rule out the possibility that the so-called homozygous (“corrected”) embryos were actually compound heterozygotes (still containing a maternal and paternal genes) and that potential large deletions which had not been detected in their somewhat low-resolution screening, may, in fact, have encompassed the cut site on one allele.

“The recent rebuttal (Ma et al) contains many in-depth analyses of different outcomes and attempts to offer solutions, but many questions remain unanswered regarding both the repair mechanisms and cell-cycle timings.

“There is no doubt that full investigation of the whole genome sequence is necessary to account for all potential genetic rearrangements as it is difficult to predict outcomes due to the random nature of insertion and deletions (indels) produced when DNA is cut.

“These opposing arguments are excellent means of propelling proper validation of findings in research prior to publication but more importantly serve as benchmark setting for pre-clinical standard setting.

“While some aspects of the criticisms have been addressed, the unanswered elements are proof that unexplained systems remain, particularly in human embryo development which require further research.”

Prof Robin Lovell-Badge CBE FMedSci FRS, Group Leader, The Francis Crick Institute, said:

“We have known about the criticisms of the original Ma et al paper for some time as the communication from Egli et al had been posted on bioRxiv and those in that from Adikusama et al reflect results that are common in the field with mouse embryos, such as mosaicism and on-target deletions, which can be rather large, rather than repair of a mutation. These are generally accepted as issues that need to be taken into account in subsequent analysis for research – and they are precisely the issues that need to be solved if genome editing of human embryos is ever to be used clinically to avoid having a child with a genetic disease.

“Ma et al (in their BCA) provide some compelling data to indicate that they indeed had obtained homology directed repair (HDR) of a mutant allele coming from the father via the use of maternal DNA as a template. Moreover, their data suggests that this repair need not happen immediately after injection of Cas9 and guide RNA complexes (RNPs) either along with the sperm into oocytes (during intra-cytoplasmic sperm injection, ICSI) or after fertilisation (pronuclear or “S-phase” stages) and that it may not occur until at least after the first cell division (cleavage) to give two-cell embryos. If the repair can occur later, this may get around the criticism from Egli and colleagues that the maternal and paternal genomes are separated from each other until shortly before this first cleavage division.

“However, there are still too many unknowns and this just emphasises that we are not yet ready to use genome editing methods in the early embryo for “real”, i.e. to avoid having a child with a genetic disease.

“The low levels of mosaicism found when the genome editing components were introduced during ICSI compared with into zygotes still deserves an explanation, but may suggest that this reflects the earlier stage at which the DNA is cut by Cas9. However, if there is a time gap between the cut and its repair, this is indeed likely to lead to large deletions and to repair by NHEJ. Related to this, Ma et al also fail to provide an explanation of why the single maternal DNA copy was used for homology directed repair (HDR) rather than the many copies of exogenous DNA template in the experiments they had reported originally. Perhaps the exogenous DNA is eliminated rapidly by enzymes that digest broken DNA ends (exonucleases) before the HDR system can operate – but if so, these exonucleases would also be expected to produce large deletions in the cut genomic DNA. Perhaps this suggests that the repair mechanism using maternal DNA as a template reflects gene conversion or mismatch repair mechanisms that survey the paternal genome for defects. But if this is the case, why have we not previously noticed repair of paternally inherited deletions? The cut in the DNA mediated by Cas9 mediated might be an explanation.”

*’Large deletions induced by Cas9 cleavage’ by Thomas et al. published in Nature on Wednesday 8 August 2018. 

*’Inter-homologue repair in fertilized human eggs?’ by Egli et al. published in Nature on Wednesday 8 August 2018.

*’Ma et al. reply’ by Ma et al. published in Nature on Wednesday 8 August 2018.

Declared interests

Dr Helen O’Neill: “My research focuses on genome editing, but I have no affiliations with the authors.”

Prof Robin Lovell-Badge: “I am a senior group leader at the Francis Crick Institute. I use genome editing methods in my research in animals and cell lines, but not in human embryos. I serve on several committees with an interest in the topic of genome editing in human embryos, but have no financial interest in the topic or any other conflicts of interest.”