In what has to be one of the least-surprising results ever, researchers at the Sanger Institute in the UK find that CRISPR modification of dividing cells causes high rates of DNA rearrangements.
This study differed from others in that it 1) used cells that divide but are normal – eg., stem cells rather than tumor cells; 2) looked not just at the target site but thousands of nucleotides on either side of it; 3) created heterozygous cell lines and looked for gene conversion.
They found shockingly high levels of deletions, insertions, transpositions and conversions – anywhere from 10-60% of edited cells. Deletions, insertions and transpositions are the kind of DNA damage that frequently initiate transformation of normal cells to cancerous cells. Gene conversions may “uncover” recessive mutations that are harmless in single copies but cause disease when present in two copies. Everyone carries hundreds or thousands of these kinds of mutations.
Why is this result not surprising? Normal DNA repair processes in our cells are complex and tightly regulated: 
From DNA Repair Mechanisms and the Bypass of DNA Damage in Saccharomyces cerevisiae
Randomly screwing around with host DNA, rather than following a well-regulated process, is what viruses do. Viruses, of course, don’t care if they make sloppy changes that result in cancerous growth. And CRISPR operations look very much like a viral invasion. They set off a whole slew of cellular responses, including ones triggered by viral infections. Is it really a surprise that CRISPR’d cells behave aberrantly and make all kinds of mistakes?
Apparently it is to those desperate to cash in on the latest biotech gold rush. A whole set of clinical trials are already slated or underway (clinicaltrials.gov: NCT03081715, NCT03398967, NCT03166878, NCT02793856, NCT03044743, NCT03164135).
Not to mention the idiots who have taken up CRISPR as a “biohacking” technology.
When the first gene-editing technologies were proposed, the public was skeptical and alarmed. Scientists and the FDA proceeded with caution. Nonetheless, several people were sickened and killed by gene therapy. The obvious lesson is that biology is complex (duh) and we don’t really know how to extrapolate from the lab to the clinic. This is true of mature disciplines like small-molecule pharmacology and it is especially true of emerging disciplines like genome editing. Lots can go wrong in ways that we can’t even imagine yet.
Clinical trials of CRISPR should be halted immediately before sick and desperate patients are killed. There is no way that they can give truly informed consent – if the CRISPR community can be blindsided with the Sanger results, then we haven’t a clue of the risks. Surprises are inevitable, and they will be bad surprises. Experimenting on people like they were mice is not OK.