Retraction Watch

Tracking retractions as a window into the scientific process

Controversial CRISPR paper earns second editorial note

with 9 comments

Against the authors’ objections, Nature Methods has added an expression of concern to a 2017 paper that drew fire for suggesting a common gene editing technique could cause widespread collateral damage to the genome. The latest note — the second to be added in two months — alerts readers to an alternative interpretation of the findings.

When “Unexpected mutations after CRISPR–Cas9 editing in vivo” was published May 30, it immediately drew criticism from many of the top scientists working with CRISPR, including those associated with companies seeking to develop CRISPR-based therapies for humans. Share prices for the two largest companies pursuing CRISPR therapies, Editas Medicine and Intellia Therapeutics, dropped following publication of the article.

On June 14, the journal published a notice to alert readers to “technical criticisms” of the paper. Apparently, that wasn’t sufficient, because the journal is now providing more details on the nature of the criticisms, despite the objections of the paper’s authors:

The editors of Nature Methods are issuing an editorial expression of concern regarding this paper to alert our readers to concerns about interpretation of the data. Multiple groups have questioned the interpretation that single nucleotide changes seen in whole genome sequences of two CRISPR/Cas9-treated mice are due to the CRISPR treatment. Since the background genetic variation between the control mouse and the CRISPR-treated animals is not known, an alternative proposed interpretation is that the observed changes are due to normal genetic variation. We are in contact with the critics and with the authors to examine this matter further. We will update our readers once these investigations are complete. All the authors do not agree with the journal’s decision to issue an editorial expression of concern.

Study author Alexander Bassuk, of the University of Iowa, told us he and his co-authors first learned the journal would be publishing an EOC two weeks ago via email:

The journal did not really discuss their reasoning much with us.

Bassuk told Retraction Watch that the email came as he and his co-authors were preparing a response to critiques raised by Editas and Intellia, which he said the journal had requested about a month ago.

He added that the authors all disagreed with the expression of concern “because our studies point to unexpected mutations” caused by CRISPR/Cas9, though he welcomes “robust discussion” on this topic.

The “alternative proposed interpretation” mentioned in the EOC has been highlighted by two high-profile co-founders of Editas Medicine: George Church, of Harvard Medical School, and Keith Joung, of Massachusetts General Hospital. When we reported on the first editorial notice, Church told us:

If they haven’t ruled out that [the genetic difference] is due to the strain differences, it’s probably due to the strain differences, based on published strain characterizations.

On July 5, Joung and several co-authors posted a paper to the Biorxiv preprint server advancing the same idea:

Here we demonstrate that the simplest interpretation of Schaefer et al.’s data is that the two CRISPR-Cas9-treated mice are actually more closely related genetically to each other than to the control mouse. This strongly suggests that the so-called “unexpected mutations” simply represent SNPs and indels shared in common by these mice prior to nuclease treatment.

Bassuk said his team is “excited” to submit their response to criticisms of the paper, which is almost complete and will be sent to the journal in just a few days:  

We found that a deeper look at our already publicly available whole genome sequencing data, combined with Sanger sequencing, demonstrates many alleles (more than two in each case we examined), including multiple [single nucleotide variants], at CRISPR/Cas9 off-target mutation sites we originally reported. These multiple mutant alleles cannot be simply explained by parental inheritance, so this is very exciting news.

A Springer Nature spokesperson declined to confirm that it had asked Bassuk’s team for a reply to criticisms, saying it treats “individual correspondences as confidential.” As for why the journal felt another note was necessary:

We decided to issue an [Editorial EOC] at this time as, following initial consultations with authors, critics and referees, we believed it was important to provide further information to the community about concerns about the paper whilst we continued to investigate them.

Like Retraction Watch? Consider making a tax-deductible contribution to support our growth. You can also follow us on Twitter, like us on Facebook, add us to your RSS reader, sign up on our homepage for an email every time there’s a new post, or subscribe to our daily digest. Click here to review our Comments Policy. For a sneak peek at what we’re working on, click here.

Comments
  • Simon R July 26, 2017 at 1:44 pm

    This seems like nothing but a response from bigpharma flexing its muscle. Sadly, this would be the ideal case for performing the experiment again, and to support or refute the hypothesis; you know, as the field and nature of the scientific principle dictates. It’s why why have journal articles that allow anyone to replicate an experimental procedure.

    • Anonymous3 July 27, 2017 at 1:09 pm

      Simon – This is actually “small pharma” – very very very small pharma. Editas Medicine and Intellia Therapeutics both had ~$6 million in revenue last year and huge losses (you might not realize, “big” pharma gets 10s of billions in revenue, not just millions). Check the financials of any pharma company and you’ll see this.

      However, the issue shouldn’t be whether it is “big” or little pharma (or other non-scientific criteria), it should be about the quality of the science and whether it is credible and supported by the appropriate evidence.

      • Pointerout July 27, 2017 at 2:39 pm

        Yes, but how much VC do they have backing them?

      • Simon R July 27, 2017 at 5:49 pm

        Understandable. ‘Bigpharma’ was simply a catch-term for commercial interests in this case, not really a commentary on the revenue/market cap of any pharmaceutical company in particular.

  • Roland July 27, 2017 at 6:24 am

    One replicate in the control group. In my field that is an automatic rejection by the editor.

    • Deepak July 27, 2017 at 8:36 am

      In most cases yes. but in this case the control mouse must be from the same gene pool and hence single replicate is suffcient.

  • justtsumguy July 27, 2017 at 7:42 am

    This second editorial note doesn’t even say anything, just promises more information to come soon. Seems kind of pointless.

  • rfg July 27, 2017 at 9:24 am

    Also remarkable is how fast the expression of concern came.

    I guess that if a journal wants to move and move quickly they can.

    They journal can also act even if the authors and presumably their institution object.

    Could it be that the fact that the paper and complaints from the companies whose valuation compromised were both covered in the scientific and popular press?

  • Post a comment

    Threaded commenting powered by interconnect/it code.