Despite taking some serious hits, a 2006 letter in Nature isn’t going anywhere.
Years ago, a university committee determined that two figures in the letter had been falsified. The journal chose to correct the paper, rather than retract it — and then, the next year, published a correction of that correction due to “an error in the production process.” To round it out, in June of last year, Nature published a rebuttal from a separate research group, who had failed to replicate the letter’s results.
Still, the first author told us there are no plans to retract the paper, since the follow up experiments published in the corrections confirmed the paper’s conclusions.
“Enhanced bacterial clearance and sepsis resistance in caspase-12-deficient mice” has been cited 209 times since it was published in 2006, according to Clarivate Analytics’ Web of Science, formerly part of Thomson Reuters.
Early concerns about the paper led to a 2012 investigation by McGill University, where author Maya Saleh is based. The university committee found that two figures in the 2006 letter had been “intentionally contrived and falsified,” prompting a correction in 2013, which doesn’t mention the committee’s findings. Two years after that correction notice was itself corrected in 2014, Nature published a Brief Communications Arising from the paper, in which another research group discusses its unsuccessful efforts to replicate some of the Saleh et al results.
For those not familiar, a BCA is basically Nature’s way of flagging an important problem with a paper. The short papers are peer reviewed, challenge a central part of a paper’s conclusions, and provide new data as evidence.
Mohamed Lamkanfi, corresponding author on the BCA, told us via email that his lab has worked on a study for several years using the mouse line from Saleh et al, which is bred to be deficient in an enzyme involved in inflammation. They found two inconsistencies between the Saleh et al data and their own on how the deficiency affected other enzymes, so they created their own line of knockout mice. Again, they got different results than Saleh et al.
Lamkanfi told us:
As this dogma has dominated (and in our opinion obstructed) caspase-12 research for so long, we thought it relevant to dispute the claims in Nature.
Regarding the importance of getting this right, he explained,
Apart from pure scientific curiosity in understanding caspase-12 function, the biomedical relevance is in proclaimed links of caspase-12 variants with susceptibility to sepsis and other infections.
In response to questions about Lamkanfi’s work, Saleh responded with an email arguing that the context of the Lamkanfi experiments was different from her group’s experiments, and emphasized that science still has a ways to go before anyone fully understands caspase-12. Regarding whether the misconduct finding, which called four of her papers into question and resulted in two corrections, made her consider retracting the Nature paper, she told us it hadn’t, because “different experiments were conducted with the same results.” Those results were then published in the correction.
Nature appears to take a similar view. We asked a Nature spokesperson why the journal had chosen to publish a correction and a BCA, rather than retract, particularly in light of the committee’s finding of misconduct. She responded via email:
In general, the decision to retract a paper is judged according to whether the main conclusion of the paper no longer holds or is seriously undermined as a result of subsequent information coming to light of which the authors and editors were not aware at the time of publication. We refer you to the correction note for information.
Here’s that correction notice, which doesn’t mention anything about the committee’s finding:
When our Letter was under consideration at Nature, we originally showed co-immunoprecipitation between caspase-1 and wild-type caspase-12 or catalytically inactive caspase-12 (C299A) as part of Fig. 4. In response to reviewers’ comments, requesting co-immunoprecipitation with other caspases for specificity control, this original figure was removed from the manuscript and was later published as part of figure 6 of ref. 1. It was recently brought to our attention that the published Fig. 4c of our Letter is a composite image containing parts of the original figure (the immunoprecipitation lanes in Fig. 4c), and that the input lanes (‘Total Casp12’) are duplicated. Similarly, the anti-tubulin control right lanes of Fig. 4b are duplicates of the left lanes. We are unable to reconcile how these images were incorrectly assembled despite diligent efforts to do so. Figure 1 of this Corrigendum shows a correct Fig. 4c, representing a new and independent experiment, to replace Fig. 4c of our Letter. The interpretation of the data and the conclusions are unaffected; namely, that caspase-12 forms a complex with and co-immunoprecipitates with caspase-1 when co-expressed by transfection into human (HEK293T) cells. There is also some association of caspase-12 with caspase-5 (more so than previously described) but very little with caspase-9. The experiment is robust and has been repeated a total of four times (twice each by two workers, M.S. and Claudia Champagne). We have also re-probed the original western blot in Fig. 4b with anti-tubulin and provide a replacement panel for that loading control in Fig. 1 of this Corrigendum. Our conclusions remain unaltered and the original legend for Fig. 4 also remains correct.
The authors also added more detail to the legend of one of the figures. According to the journal, it updated the correction notice after details were omitted from the advance online version due to “an error in the production process.”
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.