Last month, Ivan met David Vaux at the 3rd World Conference on Research Integrity in Montreal. David mentioned a retraction he published in Nature, and we thought it would be a great guest post on what it’s like to retract one of your own papers in an attempt to clean up the literature.
In September 1995 Nature asked me to review a manuscript by Bellgrau and co-workers, which subsequently appeared. I was very excited by this paper, as it showed that expression of CD95L on Sertoli cells in allogeneic mismatched testes tissue transplanted under the kidney capsule was able to induce apoptosis of invading cytotoxic T cells, thereby preventing rejection. As I wrote in a News and Views piece, the implications of these findings were enormous – grafts engineered to express CD95L would be able to prevent rejection without generalized immunosuppression.
In fact, I was so taken by these findings that we started generation of transgenic mice that expressed CD95L on their islet beta cells to see if it would allow islet cell grafts to avoid rejection and provide a cure for diabetes in mismatched recipients.
Little did we know that instead of providing an answer to transplant rejection, these experiments would teach us a great deal about editorial practices and the difficulty of correcting errors once they appear in the literature.
What we found was that unfortunately, these grafts were not protected, and indeed CD95L-expressing grafts seemed to provoke more, not less, of an inflammatory response. Puzzled by this, we decided to repeat the experiments by Bellgrau et al., but unlike them, we found that allogeneic mismatched testes grafts were rejected. A subsequent more thorough reading of the literature revealed that similar mismatched testes tissue grafts had been performed previously, both in the mouse and the rat, and their results were the same as ours, and opposite those of Bellgrau et al., i.e. the unmatched testes tissue was rejected.
Knowing that Nature had an explicit editorial policy to publish, in some form, work which refutes an important conclusion of any paper which appears in its pages, we submitted our findings describing the transgenic mice and our failure to replicate the work from Bellgrau et al. to Nature. We received two very positive reviews, but based on a third, very negative one, from Bellgrau et al., the editors decided not to publish our findings as a letter or as correspondence.
In 1996, we submitted our manuscript to Nature Medicine, but it was rejected without review, with the comment from the editor in chief, Adrian Ivinson, that he did “not think formal submission to Nature Medicine would be appropriate”. We then sent the manuscript to PNAS, where it has attracted 305 citations. Subsequently, another paper appeared describing transplants of beta cells from CD95L transgenic mice, and their findings were the same as ours, i.e. graft CD95L did not confer protection, but if anything, provoked inflammation. To our surprise, this paper appeared in Nature Medicine, accompanied by a News and Views by Lau and Stoeckert emphasizing the importance of the findings.
I was becoming increasingly frustrated by Nature’s refusal to abide by its own ethical policies to publish rebuttals, and Nature Medicine’s decisions apparently based on papers’ sources rather than their contents, when I had a flash of inspiration – I had published a News and Views extolling the virtues of Bellgrau et al.’s paper – now I could retract it!
I wrote to Phillip Campbell at Nature saying that I wished to retract my News and Views piece because I no longer had confidence in the findings on which it was based. My reasons for doubt were:
- We were unable to reproduce Bellgrau et al.’s findings;
- Three earlier groups who had published similar experiments had also come to the opposite conclusion;
- The failure of transgenic CD95L to protect allogeneic islet cells was contrary to the model they proposed.
I added “I regret having to take this course, but as Nature refuses to abide by its own ethical policy, namely to “publish refutations of any important conclusion that appears in its pages,” I am left with no other option.
Thankfully, Nature did agree to publish the retraction, but, perhaps unsurprisingly, they were unhappy with the wording. The retraction included just two sentences:
I wish to point out that I no longer stand by the views reported in my News and Views article “Immunology: Ways around rejection” (Nature 377, 576–577; 1995), which dealt with a paper in the same issue (“A role for CD95 ligand in preventing graft rejection” by D. Bellgrau et al. — Nature 377, 630–635; 1995). My colleagues and I have been unable to reproduce some of the results of Bellgrau et al., as reported by J. Allison et al. (Proc. Natl Acad. Sci. USA 94, 3943-3947; 1997).
This was accompanied by:
D. Bellgrau et al. consider that their results are reproducible and stand by them. They note, however, that the magnification in Figure1g of their paper should be 113 times, not 45 times as printed. Both groups believe that other published data support their views, and interested readers can contact them directly for further details. — Editor, Nature.
Note that they did not say that the results were reproducible, or that they had actually reproduced them, they just considered them to be reproducible. Indeed, no one, including Bellgrau et al., have subsequently reported reproducing these results. Furthermore, it turned out that Sertoli cells do not even express CD95L.
The retraction was published in 1998, and has attracted 16 citations of its own. However, of the 976 citations of the Bellgrau et al. paper, about 700 were subsequent to publication of the retraction, so it’s clear many remain unaware that its findings are questionable. Clearly, the processes that allow the scientific record to self-correct can be improved, not least by Nature.