CRISPR editing human embryos - did it work or not?

Controversy surrounding last year's report of a disease gene being edited out of human embryos continues with a fresh round of evidence.

A group led by Dr Shoukhrat Mitalipov of the Oregon Health and Science University in Portland published a study in Nature last year describing how they used CRISPR/Cas9 genome editing to repair a genetic variant in human embryos responsible for a heart condition (see BioNews 911 and 912).

But their results were challenged soon after, with other researchers suggesting that the findings of the Nature paper did not prove a 'repair' of a faulty gene, but rather the absence of a disease-causing mutation (see BioNews 916).

The original study used sperm harbouring the mutated gene to fertilise healthy eggs. The subsequent embryos were injected with CRISPR/Cas9 reagents, designed to edit out the disease gene.

Three-quarters of the resulting embryos did not have the faulty gene – without editing, around half would be expected to be disease-free.

Crucially, however, the disease gene appeared to have been fixed in an unexpected way. The paper reported a discovery regarding the cells' intrinsic repair mechanism: instead of the paternal gene being repaired with healthy template DNA delivered to the cell, the team said the findings suggested the faulty paternal allele was repaired using the copy from the mother's genome, contributed by the egg. This would represent a previously unknown DNA repair mechanism. 

It is this conclusion that has been the source of intense debate – which Dr Mitalipov says he welcomes. He and his team have now published new data in Nature backing their original findings. Key criticism also comes in the form of two letters published together in the recent issue of Nature.

The first, from scientists including Dr Maria Jasin at the Memorial Sloan Kettering Cancer Centre in New York City and Dr Dieter Egli at the Columbia University, was the critique that originally sparked debate when it was published last year on preprint server bioRxiv.

It challenges the idea of the mother's gene being used as a template on the grounds that the maternal and paternal genomes are physically separated in the early embryo. In their view this should isolate them from one another, preventing the maternal genome being available as a repair template.

'The genomes are separated by what are, in cellular terms, intergalactic distances,' Dr Tony Perry at the University of Bath, who was not involved in the critique, told Nature News.

A second letter, from researchers including Professor Paul Thomas at the University of Adelaide, Australia, questioned the methods used by Dr Mitalipov's group to detect the purported correction.

'We think there are other explanations,' they said in the letter. The signatories' own work suggests that large deletions of DNA can occur in CRISPR genome editing, and they think this might have caused Dr Mitalipov's team to misinterpret the results.

But Dr Mitalipov and his colleagues remain confident, writing in their own paper in Nature: 'The additional retesting of the samples confirmed our original conclusion that this repair occurred as we described.'

They also refuted the criticisms with a reanalysis of the DNA of the edited embryos using different techniques, which they say proves their original conclusion, citing research from the Massachusetts Institute of Technology that suggests the same repair mechanism occurs in mice.

'Independent replication of our work provides additional evidence that our discovery may lead to the prevention of inherited disease,' Dr Mitalipov said.