Researchers ‘devastated’ after finding manipulated data in study of pediatric brain tumors

An international group of cancer researchers has lost an influential 2020 paper in Nature Neuroscience after finding problems with the data that triggered an institutional investigation.

The article, “Tumor necrosis factor overcomes immune evasion in p53-mutant medulloblastoma,” represented a potentially major advance in the treatment of pediatric brain tumors, according to Robert Wechsler-Reya, the director of the Tumor Initiation & Maintenance Program at the Sanford Burnham Prebys Medical Discovery Institute, in La Jolla, Calif., and the senior author of the paper, which has been cited 17 times, per Clarivate Analytics’ Web of Science:

The Nature Neuroscience paper was an important piece of work demonstrating that medulloblastoma cells can evade the immune system by shutting down expression of a key surface molecule called MHC-I, and showing that this evasion can be overcome by treating cells with a cytokine called tumor necrosis factor (TNF). It had significant implications for immunotherapy of medulloblastoma and other brain tumors, and formed the basis for a clinical trial that we were hoping to launch earlier this year.  

Those plans were shelved, the researcher said, after he learned of “anomalies” in the data in February: 

The machine used to collect many of the key data in the paper, called a flow cytometer, keeps detailed electronic records of how the data were collected, and examination of these records revealed that the photomultiplier tube (PMT) voltages had been selectively altered for some of the samples within certain experiments. (Ordinarily the voltage is kept the same throughout an experiment, so the samples can be compared to one another). In particular, the voltage seemed to have been manipulated to make it appear that MHC-I expression had increased in response to TNF when it had not. This type of alteration was found in a large percentage of the data in the paper. (We also examined the author’s data used in other publications, and found no evidence of problems in their data used for those papers.) As a consequence of these findings, we took three major actions:

First, we immediately contacted Sanford Burnham Prebys’ research integrity officer to initiate the formal process of research misconduct assessment/inquiry/investigation. This is an extended and confidential process, which is now wrapping up after a full research misconduct investigation. The purpose of this investigation is to determine whether there was intentional research misconduct, and if so, what sanctions there should be for the person involved. Once it is complete, the investigation findings will be shared with ORI, and they will determine if actions on their part are appropriate.

Second, the clinical trial that was being planned was put on hold (before final regulatory approval of the trial and before any patients had been treated), and eventually canceled. 

Finally, given the anomalies in the data, I asked several members of my lab to repeat the major experiments in the paper. This process took several months, and unfortunately, my lab members were unable to replicate many of the key findings. Based on these results, we decided that we could no longer stand by the conclusions of the paper, and that it would be appropriate to retract it. I contacted the journal in June, described what had happened, and initiated the process of retraction. Once they had approved the retraction in principle, I worked with them to craft the appropriate text, and this was published earlier this week.  

According to the retraction notice: 

The authors are retracting this paper owing to issues with the integrity of the data presented. Specifically, inconsistencies in the flow cytometry data in Figs. 2d, 4b, 5c,d, 5g,h, 6a,b and 7a–d, and Extended Data Figs. 2b, 3a–i, and 6c–j were found, with selected experimental samples collected using different settings from their corresponding controls. Thus, the data in these figures do not reflect the true fluorescence of the samples in question. In addition, we have been unable to replicate the key in vivo findings in the paper, including those presented in Fig. 7e–h and Extended Data Fig. 8d–k. These concerns undermine our confidence in the study and we therefore wish to retract the article in its entirety. We deeply regret this incident and apologize to the scientific community.

Wechsler-Reya added that the experience has been “devastating”:  

As you can imagine, this has been a very difficult and disheartening process for me and my lab. This was work that we were proud of and excited about, and it was devastating to learn that it was based on manipulated data. I am glad that we have had the opportunity to set the record straight, and fortunate that colleagues in the field have been supportive and understanding.  We are continuing to work on the mechanisms of immune evasion in pediatric brain tumors, and remain committed to finding more effective treatments for children with these devastating diseases.

Like Retraction Watch? You can make a one-time tax-deductible contribution by PayPal or by Square, or a monthly tax-deductible donation by Paypal to support our work, follow us on Twitter, like us on Facebook, add us to your RSS reader, or subscribe to our daily digest. If you find a retraction that’s not in our database, you can let us know here. For comments or feedback, email us at [email protected].