Michael Hertl, a leading dermatology researcher at Philipps-University Marburg, has been cleared of any wrongdoing in a case that spawned a retraction last year and two just-published retractions.
Felicitas Riedel, a legal officer for the university, tells Retraction Watch that the
…Committee for Scientific Misconduct of the Philipps-University Marburg closed the matter and submitted its results to the President of the University who in the meantime after examination consented with it.
The findings?
Prof. Dr. Michael Hertl is officially and completely exculpated. Based on an accurate fact-finding lasting for months the Committee adjudicated that Mr. Hertl neither committed any Scientific Misconduct himself nor overlooked such in his department by negligence. The latter means: if committed by a member of his group, Mr. Hertl could not foresee, recognize or prevent it. Please note that my permit to disclose personal data does not cover anyone but Mr. Hertl. Information on any other person involved cannot be given. Mr. Hertl was also completely exonerated within the framework of a disciplinary complaint from any accusation of breach of official duty.
The findings, of course, leave open the possibility that someone in Hertl’s lab may have committed misconduct.
Both new corrections appear in the Journal of Immunology. Here’s one:
Veldman, C., A. Höhne, D. Dieckmann, G. Schuler, and M. Hertl. 2004. Type I regulatory T cells specific for desmoglein 3 are more frequently detected in healthy individuals than in patients with pemphigus vulgaris. J. Immunol. 172: 6468–6475.
Fig. 2 of this article shows identical FACS images for the mIgG1 isotype control for CD25, CCR4, and TGF-β (all mouse IgG1-PE) as well as HLA-DR, CD45RO (both mouse IgG2a-PE), and CCR7 (rat IgG2a-PE), which is due to the use of inappropriate isotype controls by the first author of the study. Because we cannot rule out that observed differences in the expression patterns of HLA-DR, CD45RO, and CCR7 may be less significant when compared to the appropriate Ab isotype controls, these data have now been omitted in the revised Fig. 2, below. In addition, we have inserted an FACS image for CTLA-4 expression of the Th2 cells from the same original experiment. CTLA-4 expression was detected by a Cytochrome-labeled Ab (clone BNI3.1). The major observation of Fig. 2 that desmoglein 3-reactive Tr1 cells, but not Th2 cells, express glucocorticoid-induced TNFR and TGF-β is not affected by this revision. In Figs. 3, 4, and 5, proliferative T cell responses are expressed as mean cpm [3H]TdR ± SD. The authors apologize to the scientific community for any inconvenience these errors may have caused.
And here’s the other:
Veldman, C., A. Pahl, S. Beissert, W. Hansen, J. Buer, D. Dieckmann, G. Schuler, and M. Hertl. 2006. Inhibition of the transcription factor Foxp3 converts desmoglein 3-specific type 1 regulatory T cells into Th2-like cells. J. Immunol. 176: 3215–3222.
In Fig. 3, our group recently discovered that the FACS data for SCR2 at 0.4 μM were incorrectly labeled as 0.2 μM, and that the data for SCR2 at 0.4 μM were generated in an unrelated experiment and incorrectly inserted by the first author of the study. In addition, the first author used inadequate isotype controls for glucocorticoid-induced TNFR family-related receptor (GITR) (polyclonal IgG) that were identical to the isotype controls for CD4 (mIgG1). Even though we have shown in an independent experiment that desmoglein (Dsg)3-reactive Tr1 cells express GITR and that treatment with Foxp3 AS2 but not Foxp3 SCR2 leads to downregulation of GITR expression by Tr1 cells, we prefer to omit Fig. 3 based on the described inaccuracies. The overall message of the study, namely the observation that Tr1 cells treated with Foxp3 AS2 lose their suppressor function, gain a proliferative phenotype, and start to secrete the T cell growth factor, IL-2, is not affected by the errors of Fig. 3. Moreover, we have not directly linked expression of GITR or CTLA-4 to the suppressor function of the Tr1 cells. Neither have we found convincing evidence that the secretion of IL-10 and TGF-β by the Tr1 cells exclusively mediates their suppressor function since IL-10 and TGF-β secretion was also not altered by Foxp3 AS2 treatment. Because the ability of the Tr1 cells to produce IL-2 was directly linked to the loss of suppressor function, we speculate that the Dsg3-reactive Tr1 cells suppress T cell responses by the consumption of exogenous IL-2. The major finding of the study (i.e., that antisense-mediated loss of Foxp3 is directly associated with the loss of suppressor function) has been confirmed by a recent, independent study (Hansmann et al. 2012. J. Immunol. 188: 1275–1282) showing that downregulation of Foxp3 in Treg cells leads to Th2 differentiation. In Figs. 4 and 6, proliferative T cell responses are expressed as mean cpm [3H]TdR ± SD. The authors apologize to the scientific community for any inconvenience these errors may have caused.
The corrections match concerns raised by the Abnormal Science blog, as we noted last year, although the post in which the blog noted the issues seems to no longer be available. The papers have been cited 65 and 32 times, respectively, according to Thomson Scientific’s Web of Knowledge.
Mmmmm, “inappropriate isotype control”, that takes me back. Happy memories.
If you want to cheat with FACS, it is practically undetectable. Luckily, a lab in Regensburg has repeated the findings in Marburg, so there is nothing to see here.
J. Immunol. 172: 6468–6475.
“these data have now been omitted in the revised Fig. 2, below.”
J. Immunol. 176: 3215–3222.
“our group recently discovered” By what method was that?
“we prefer to omit Fig. 3 based on the described inaccuracies.”
Not standing by the original data twice.
“The major finding of the study (i.e., that antisense-mediated loss of Foxp3 is directly associated with the loss of suppressor function) has been confirmed by a recent, independent study (Hansmann et al. 2012. J. Immunol. 188: 1275–1282)” may be true, but you should prove your own experiments.
Both corrections are in the same issue of the same journal.
the papers that they lean on as confirmation are here
http://www.ncbi.nlm.nih.gov/pubmed/22210907
http://www.ncbi.nlm.nih.gov/pubmed/19283780
Confirmation is the wrong word, consistent with is more accurate – since they don’t use the antisense approach, rather in vitro expansion. I don’t have institutional access to journals, but if in vitro expansion involves the use of any growth factors all kinds of confounding factors are introduced.
Your final paragraph is interesting and worthy of follow up.
Other articles covering Hertl’s coauthors Schuler and Veldman are still up on the abnormal science blog.
http://abnormalscienceblog.wordpress.com/2011/09/11/sms-alert-in-erlangen-controls-and-other-trivia/
http://abnormalscienceblog.wordpress.com/2011/09/13/sms-alert-in-erlangen-the-power-of-gates-in-flow-cytometry/
http://abnormalscienceblog.wordpress.com/2011/09/19/sms-alert-in-erlangen-the-mystery-of-a-cancer-study/
The curiously opaque comment from Ms. Riedel can only be interpreted as being legalese dictated word for word by the University’s lawyers.
It won’t take a rocket scientist to conclude that 1. Veldman will take the rap and that 2. abnormal science has had to pull an article or face a defamation lawsuit.
I guess we will see, in the end, what Veldman’s fate will be? He is the first author of the two studies and is now fingered with fabricating these figures. Reasons for the pulled article on the blog? My guess is we will never know…
I seem to remember when this affair was first published on the Abnormal Science Blog, that there was a mention that Prof. Hertl had threatened to sue the author of the blog, so this might well have happened.
C. Veldman I found back as “Referent für Eu- und internationale Forschungsförderung” at the University of Marburg in Germany, if that’s still accurate, so he seems to have left research.
Honestly, I find it always hard to believe that PIs who end up with retractions due to faked data, never seem to have observed anything fishy before the request for retraction.
Here is an official Marburg university webpage where C Veldman talks about taking up his new position and what he offers.
http://www.uni-marburg.de/informationen/mitarbeiter/nachrichten-archiv/EU-Referenten
“I would like to inform you of the fact that I, Dr. Christian Veldman, on 16.2.2009, have taken up my activity for all interests of EU-research support within the Philipps university of Marburg. I would like to serve to simplify the application for EU projects and to raise the success rate.”
“Referent für Eu- und internationale Forschungsförderung” means
Adviser for Eu support and international research support
That does not seem appropriate. Booted upstairs to management.
This statement from the Uni’s legal grunt is surely disquieting?
“..if committed by a member of his group, Mr. Hertl could not foresee, recognize or prevent it.”
Now, one should make allowances for the German-English translation; but not being able to “recognize” fraud while being the ambitious, proud, go-getting head of a cutting edge research laboratory is not something for which one should be exonerated (scientifically speaking). We are talking about a medical doctor publishing frequently on distressing skin diseases. One who can’t recognize fraud in his own research work when it is shoved past his nose?
Ideally, an honest researcher who has been stung by an in-group fraudster would find this a painful experience. In order to avoid future episodes, discussions with, education of, training of, the team members would be the order of the day so that it doesn’t happen again. If it hurts, you’re not going to let a culture of sloppiness pervade the work environment.
But the legalese might well be correct in Dr. Hertl’s case. I was annoyed enough by it to go figure-sleuthing: Hertl may indeed be unable to RECOGNIZE compromised data in his publications. Readers of Retraction Watch might be interested in this newer 2011 paper* involving Hertl’s lab (PMID: 21281804 ):
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3069870/
Figure 3 panel Q is a slightly cropped and overlapping slice from the same image as in panel O. The two exactly matching bright “dots” rule out that it is a different slice from an unfortunate monkey’s eosophagus**. This is not good at all because these are supposed to be two different antibodies.
Figure 4 exhibits reuse of a cell-based assay. Fig. 4A right-most dish is reused for Fig. 4C top right. This falls under the burgeoning category of “data inflation” whereby the same illustration is claimed to make a new point. This is not prima facie fraud since the same experiment is NOT being relabelled. But, since researchers reading scientific papers assume that each part of a figure is a NEW experiment, it is a devious way of implying that more work has been done than has been done. Well whatever happened to replication as part of the scientific method? Don’t these people even have a spare replicate to use for the second part of the figure?
Some advice for the exculpated good doctor Hertl: The internet never forgets. You already have four corrections and one retraction to your name (2004, 2006, 2007, 2009-2009). Your papers are going to be trawled over by figure sleuths for the foreseeable future. If you don’t want to have to keep removing your appearance on anti-fraud blogs, then don’t let your colleagues fabricate figures. RTBM – that’s “read the bloody manuscript” – before you send it in for publication.
*Credit due to the publishers for the open access. There is no excuse – go and see for yourself.
**Which kind of monkey!? There is more than one sort as it happens. Monkey is a remarkably un-scientific term here.
EU meaning European Union.