Bitter legal fight leads to a retracted retraction

faseb june 2013Two years ago, the FASEB Journal retracted a paper that it had initially agreed to correct, after a dean at one of the author’s institutions said that a “well-recognized and top-class fact finding commission concluded that the publication contains gross flaws.” The retraction of the 2003 paper, as we noted at the time, punctuated a complicated case involving several investigations as well as legal maneuvering.

Now, the journal has retracted the retraction. Here’s the beginning of the notice:

Regarding the article titled, “Molecular analysis of Nogo expression in the hippocampus during development and following lesion and seizure,” by Susan Meier, Anja U. Bräuer, Bernd Heimrich, Martin E. Schwab, Robert Nitsch, and Nicolai E. Savaskan, published in the June 2003 issue of The FASEB Journal (FASEB J., 2003 Jun;17(9):1153—1155; doi:10.1096/fj.02-0453fje). The editors of The FASEB Journal received a letter date June 1, 2011, from Dr. Annette Grüters-Kieslich at Charite—Universitätsmedizin Berlin stating:

“In the year 2009 a series of reproaches in regard to scientific misconduct against Dr. Nicolai Savaskan reached the faculty of the Charite—Universitätsmedizin Berlin.”

“One of the manuscripts affected was published in the FASEB J in the year 2003: Meier S, Bräuer AU, Heimrich B, Schwab ME, Nitsch R, Savaskan NE. FASEB J. 2003 Jun;17(9):1153–5. A well-recognized and top-class fact finding commission concluded that the publication contains gross flaws. A key figure (Figure 14) and the conclusions drawn from it could not be underlined with the corresponding primary data. Therefore, the faculty has requested the senior author Dr. Nicolai Savaskan to retract the publication.”

In light of the “well-recognized and top-class fact finding” commission’s conclusions and the faculty’s recommendation to retract the article, the article was retracted and all versions were removed from the Web site.

Since receiving this communication, Annette Grüters-Kieslich at Charite—Universitätsmedizin Berlin has contacted the journal with an additional letter stating the following:

“We consulted you in writing 1 June 2011 concerning the publication of Meier S. et al, FASEB J 2003 17(9) 1153-5.”

“In a final evaluation of the investigations carried about by reason of the letter stating the underlying facts, we find that these investigations do not proof of intentional falsifications, manipulation of plagiarism in this work.” (sic)

“As a result of our investigations, we correspondingly suggest the recommendation of the Johann Gutenberg University of Mainz to publish a written erratum from the scientist, Dr. Savaskan and Prof. Nitsch, for the correction of the mistakes contained in this publication.” (sic)

In light of the new recommendation by Charite—Universitätsmedizin Berlin, the journal has reinstated the original article with the addition of the following erratum and addendum provided by the authors:

The notice then goes on, in some detail. Excerpts:

We have been informed about a data presentation error in one quantitative data set shown in Figure 14 B in our paper published in FASEB J (2003 Jun;17(9):1153–5. Epub 2003 Apr 8) [1]. There, the actin column for group 3 (lesion group 5DAL) is given larger, i.e. 887.7 a.u. than the original data file with 224.3 a.u. which contains the primary data actin quantification set. Nevertheless, the main conclusion of our paper given in Figure 14 A and C, i.e. the fact that an entorhinal cortex lesion induces upregulation of Nogo-A protein in the hippocampus, is not affected by this display error.

We decided to repeat the experiments given in the original Figure 14 which have been performed in the years 2001–2003.

After describing the methodology and new results, the authors conclude, somewhat emphatically:

From these reproduced and updated experimental results it can be concluded that the findings as reported in the paper in 2003 are valid and still stand. We apologize for any confusion the erroneous display may have caused. However, there is no doubt that the finding of a Nogo-A upregulation following entorhinal cortex lesion as reported in our FASEB J paper from 2003 is based on reproducible mRNA and protein data, and thus is a biological fact.

This has been a complicated case. Among the other developments in the case since mid-2011:

The publications

1.    U. Bräuer, N. E. Savaskan, M. Plaschke, S. Prehn, O. Ninnemann, and R. Nitsch. IG-molecule Kilon shows differential expression pattern from LAMP in the developing and adult rat hippocampus. Hippocampus 10 (6):632-644, 2000.

2.    U. Bräuer, N. E. Savaskan, M. Plaschke, O. Ninnemann, and R. Nitsch. Perforant path lesion induces up-regulation of stathmin messenger RNA, but not SCG10 messenger RNA, in the adult rat hippocampus. Anonymous. Anonymous.  Neuroscience 102(3):515-526, 2001.

3.    U. Bräuer, N. E. Savaskan, M. H. Kole, M. Plaschke, L. M. Monteggia, E. J. Nestler, E. Simburger, R. A. Deisz, O. Ninnemann, and R. Nitsch. Molecular and functional analysis of hyperpolarization-activated pacemaker channels in the hippocampus after entorhinal cortex lesion. FASEB J. 15 (14):2689-2701, 2001.

4.    U. Bräuer, N. E. Savaskan, H. Kühn, S. Prehn, O. Ninnemann, and R. Nitsch. A new phospholipid phosphatase, PRG-1, is involved in axon growth and regenerative sprouting. Nat.Neurosci. 6 (6):572-578, 2003.

5.    U. Bräuer, N. E. Savaskan, M. Plaschke, O. Ninnemann, and R. Nitsch. Cholecystokinin expression after hippocampal deafferentiation: molecular evidence revealed by differential display-reverse transcription-polymerase chain reaction. Neuroscience 121 (1):111-121, 2003.

6.    U. Bräuer, R. Nitsch, and N. E. Savaskan. Identification of macrophage/microglia activation factor (MAF) associated with late endosomes/lysosomes in microglial cells. FEBS Lett. 563 (1-3):41-48, 2004.

were incriminated by Dr. Markus Kühbacher for scientific misconduct. Concerning publications 1, 2, 5 and 6 I have looked into the matter and there is no evidence for fabrication, falsification or plagiarism of the underlying data presented in the paper.

Professor J. Knop, Ombudsman of the Gutenberg University Mainz, looked into incriminations concerning publications 3 and 4. He declared that there is no evidence for fabrication, falsification or plagiarism of the underlying data presented in these papers.

have not provided evidence of deliberate falsification, manipulation or plagiarism in the investigated work of Dr. Savaskan.

We’ve asked for the complete reports from Charite, and have uploaded a number of files we’ve obtained that might shed light on the case for our German readers.

We’ve seen two retracted retractions before. One was due to a publisher error, but the other was a wrenching situation for the lead author.

Update, 10 a.m. Eastern, 6/5/13: Clarified “On February 20” sentence; this was a settlement agreement, not a court decision. Also added bullet point “In June 2012” with link to a court decision.

19 thoughts on “Bitter legal fight leads to a retracted retraction”

  1. How can anyone provide evidence of deliberate misconduct? I believe that the preponderance of evidence is what should be considered !

      1. Uh oh! That’s like a pretty dirty bomb. It shows how these things are handled in this “famous” Charité.

  2. We will have to see how this plays out once all the facts are known (if there are any German-reading reader of this site).

    But isn’t the phrase “A well-recognized and top-class fact finding commission” a big red flashing sign that the accompanying statements are bogus?

    Why pump up the prestige of this group rather then tell us the more mundane, but important, information as to whether this commission was really empowered by the university to adjudicate on this matter?

  3. The carpet of the Charité must be really dome-shaped by the huge number of cases which have been brushed under it.

    1. Err, doesn’t the court ruling say that Charité acted in violation of its own statutes regarding the investigation of scientific misconduct and the powers that the responsible committees may wield? My German’s never really been good…

        1. Why was the decision overruled by the court? As I know courts may overrule decisions for many reasons, some of which may not be related to the subject matter of the misconduct per se. In other words, the court’s decision is not an acquittal of the accused unless stated otherwise!

          1. There was no decision overruled, solely legitimate claim was not clear in the case of Nitsch

        2. There was no decision overruled, theree was solely decided that the legitimate claim in this case cannot be substantiated by Nitsch.

          1. Tombuzz is wrong. The decision was overruled by the higher administrative court:

            https://translate.google.de/?hl=de&tab=wT#de/en/Der%20Beschluss%20des%20Verwaltungsgerichts%20Berlin%20vom%201.%20November%202011%20wird%20mit%20Ausnahme%20der%20Streitwertfestsetzung%20ge%C3%A4ndert.%20Der%20Antrag%20des%20Antragstellers%20wird%20abgelehnt.

            There are many reasons for the fact that the higher court changed the decision of the first instance. One reason is translated as follows:

            https://translate.google.de/?hl=de&tab=wT#de/en/Die%20Bef%C3%BCrchtung%20des%20Antragstellers%20ausweislich%20seiner%20eidesstattlichen%20Versicherung%20vom%2029.%20Dezember%202011%2C%20der%20von%20ihm%20betreute%20Antrag%20gegen%C3%BCber%20der%20Deutschen%20Forschungsgesellschaft%2C%20%C3%BCber%20den%20im%20…%20entschieden%20werden%20solle%2C%20k%C3%B6nnte%20aufgrund%20des%20%E2%80%9Enach%20wie%20vor%20unausger%C3%A4umt%5Ben%5D%E2%80%9C%20Vorwurfs%20vermeintlichen%20wissenschaftlichen%20Fehlverhaltens%20abgelehnt%20werden%2C%20erscheint%20dem%20Senat%20nicht%20schl%C3%BCssig.

            The higher administrative court didn’t believe what Prof. Dr. Robert Nitsch argued in his affidavit.

  4. Ugh, this is a contorted case. AFAI understand the PDF linked in the original post above, there are two commissions, two universities and two scientists involved (Nitsch and Savaskan) in this case.

    The court only asserts the suspending effect of the petition of Nitsch. The rest seems to be a morass of hair-splitting between gross negligence and purposeful misconduct, between retraction by the authors (plural, allowable) and retraction by the journal (not allowable) etc.

    Nitsch claims that such a retraction would harm his scientific freedom, in particular, his chance to score in the next round of founding decisions. The court seems to deny the idea of the opposing party that this is none of Nitsch’s business as the accusations are leveled against Savaskan. Hope that helps to get a hunch.

  5. Don’t we all know how commissions can be installed to be used for political aims?
    Should read this press release and you get an idea that the dean of the Charité, Annette Grüters-Kieslich fought against Nitsch:

    http://www.zeit.de/studium/hochschule/2013-06/charite-faelschung-urteil-savaskan

    What is most worrying is the fact that such a big organisation as the Charité violates its own statutes and regulations. Not so once apon a time Germany has had a time where organisations defined the rules on their own.
    Annette Grüters-Kieslich follows here good old tradition of Germans? C’mon, we believed this behaviour has been trained of existence over the last decades.

  6. I just discovered a VERY STRANGE “detail”: the erratum / addendum was published in 2013. In fact, the pdf was created on May 21, 2013, according to the document properties. However, the date printed on the erratum / addendum is August 2011 ?! Even more confusing is the fact that according to the Journal’s website it allegedly has been published in 2011 (last article in the issue 8 of volume 25). Is anyone able to clarify this mess? Is this a mistake? Misinformation? A scandal?

  7. Just came across this. Looking at figure 14 in the original paper:
    http://www.fasebj.org/content/suppl/2013/05/30/17.9.1153.DC1/meier.pdf

    and comparing to figure 2 in the correction:
    http://www.fasebj.org/content/25/8/2853.full.pdf+html

    We can see that the relative expression levels are completely different. For the controls, the original data shows 5-6 times lower expression than the new data. For the lesioned animals, the original data shows relative expression less than half of what is seen in the new data. However, using the correct actin data for the original lesioned group means that the value reported in the paper should be multiplied by ~4x. Thus the “correct” original data gives a value of ~1.5x of the new data.

    To recap, in this direct replication we should expect similar results. Similar results does not mean “calculate p-value greater or less than 0.05 both times”, it means the results should be similar. The point of the normalization to actin levels is to allow such comparisons (if this does not work why do it?).

    Yet 8 years later the relative expression in the control animals has risen to 5x of the original, while for the lesioned animals it has dropped by about 1/3. So how reliable are these methods and results anyway?

    Another point. Looking at figure 14 in the original paper and substituting the correct actin levels (~200 a.u.) would lead us to conclude that it is not so much the nogo expression that has changed due to the lesion, but the supposed housekeeping gene (actin).

    Now they did possibly use a different anti-actin antibody (they report it was from sigma in the first paper, abcam in the second) as well as different gel and blotting materials so this is not actually a direct replication. We also have no idea what actually occurred for the hippocampus extraction and protein purification as the only reference is to Chen 2000 which is a nature paper (that of course contains no details).

    Another thingt. Unfortunately both papers use “dynamite plots”
    http://biostat.mc.vanderbilt.edu/wiki/pub/Main/TatsukiRcode/Poster3.pdf

    This type of chart prevents us from viewing the distribution of results, however by looking at the image in the figure 2 of the correction it is apparent that the expression of nogo-A in the controls is varying wildly and that there is a floor effect occurring (many data points very near zero). Why is this?

    I’m not even sure the presence or not of fraud matters since:
    1) We aren’t sure what they actually did in terms of protein extraction and how this may have been affected by the lesion surgery. Are we sure that they are actually measuring only the amount of protein present and not the amount their procedure can successfully extract due to shorter time between sacrifice and extraction? The control actin lanes look messier on average, is this degradation?
    2) Huge variability in the housekeeping gene data for the original study.
    3) Undetectable levels of nogo-A protein in a number of the control animals but high levels for others in the new study is unexplained.

    a final quote:
    “…there is no doubt that the finding of a Nogo-A upregulation following entorhinal cortex
    lesion as reported in our FASEB J paper from 2003 is based on reproducible mRNA and protein data, and thus is a biological fact.”

    The observed difference could be the result of any aspect of the surgical procedure such as the ketamine. No study that simply compares the averages of two groups should make such strong claims.

Leave a Reply to Dan Zabetakis Cancel reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.