Safety matters — can we be sure that CRISPR-Cas9 is not producing unwanted genetic alterations?

Recent debate over the safety of CRISPR/Cas9 genome editing following a study that suggested it can cause hundreds of unexpected mutations (1) left me puzzled. The research (see BioNews 903), published in Nature Methods and carried out in three living mice, could indeed be criticized for the lack of stringent controls and technical errors. In addition, a number of sloppy mistakes suggested a misinterpretation of the data and therefore incorrect conclusions (2). It is hard to believe the assertion that CRISPR-Cas9 editing caused so many mutations. And it is also hard to believe how such sloppiness passed the scrutiny of one of the most highly-ranked scientific journals.

In defence of the authors, they did the right thing – they used whole-genome sequencing to assess possible effects of CRISPR-Cas9-mediated genome editing in their experimental system. What they found - from their point of view - was quite alarming, and by publishing the study they wanted to make the data available to the public and warn the scientific community.

The puzzling part to me is the reaction of the companies Intellia Therapeutics and Editas Medicine and their calls for the paper to be retracted on the grounds of flawed design and interpretation (see BioNews 905). Intellia is working on permanently editing disease-associated genes in the human body with a single treatment course, whereas Editas Medicine is dedicated to treating patients with genetically defined diseases. The base technology used by both companies is CRISPR-Cas9.

Why did they get so upset? The notion that CRISPR-Cas9-mediated genome editing may not be flawless brought down the share values of the companies. If they are so cocksure that the technology is flawless, the companies must have proof of that from their own pipelines. Instead just launching an attack to put minds of the investors at peace, they could make available a few examples of whole-genome sequencing data sets before and after CRISPR-Cas9 gene editing from their own work. This would show (I assume) that the technology is indeed in their hands very precise and, therefore, safe.

Failure to do that, makes me question whether they do whole-genome sequencing before and after CRISPR-Cas9 gene editing. I cannot help but ask the question, what is their quality control?

How they can even think about developing clinical products without demonstrating that there are no unwanted genome alterations following genome editing? Regulatory bodies, concerned about patient safety, would never approve clinical studies without such proofs. Thus, naturally, the companies' shares have ended up plunging.

Such an approach is standard in other fields of emerging therapies. If you, for example, work in cellular therapy with pluripotent stem cells, either human embryonic stem cells (hESCs) or human induced pluripotent stem cells (hiPSCs), you would have to demonstrate the efficiency of your differentiation protocol by showing that not a single cell remained in a pluripotent state afterwards (which can lead to teratoma formation) - even  though the probability of this happening is very low. The field has been struggling for years with the challenge of validating a hESC/hiPSC-derived cell dose for the absence of pluripotent stem cells.

And whole-genome sequencing is a part of the quality control in hiPSC-based clinical trials due to the fear that the reprogramming of somatic cells into hiPSC might cause mutations - even when the method used to reprogram the cells in the first place does not involve integrating DNA into the genome. Hence the possibility of introducing mutations due to reprogramming is almost zero. (3)

With CRISPR-Cas9, we are dealing with a genome editing technology – and a remote possibility of unwanted genome alterations cannot be neglected. The whole genome sequencing before and after the editing should be a paradigm of quality control for any serious research study, not only clinical work.

And cost should not be an issue. According to the US National Human Genome Research Institute, the price of generating a high-quality 'draft' whole human genome sequence had fallen below US$1,500 (4) by the end of 2015, whereas the cost of a whole-exome sequence was generally below US$1,000. With such low costs, using this method for quality control should be a no-brainer for any CRISPR-Cas9 gene editing research study, not only for clinical applications.

How about CRISPR-Cas9-mediated genome editing in human embryos? The UK's HFEA (Human Fertilisation and Embryology Authority) last year granted a research licence for this (see BioNews 837) and, therefore, the same rules should apply as for any other research study. The complexity of the system and number of cells available would determine quality control. How investigators who have the licence granted will tackle the issue, remains to be seen.