A group of researchers in Ireland has retracted their 2013 article on a possible new method for treating amyotrophic lateral sclerosis — ALS, also commonly called Lou Gehrig’s disease — after identifying errors in several images in the paper.
The article, “Acidotoxicity and acid-sensing ion channels contribute to motoneuron degeneration,” was published online in Cell Death & Differentiation (and appeared in the April 1 print issue, although we think that was a coincidence…).
According to the abstract:
Amyotrophic lateral sclerosis (ALS) is a fatal neurological condition with no cure. Mitochondrial dysfunction, Ca2|[plus]| overloading and local hypoxic|[sol]|ischemic environments have been implicated in the pathophysiology of ALS and are conditions that may initiate metabolic acidosis in the affected tissue. We tested the hypothesis that acidotoxicity and acid-sensing ion channels (ASICs) are involved in the pathophysiology of ALS. We found that motoneurons were selectively vulnerable to acidotoxicity in vitro, and that acidotoxicity was partially reduced in asic1a-deficient motoneuron cultures. Cross-breeding of SOD1G93A ALS mice with asic1a-deficient mice delayed the onset and progression of motor dysfunction in SOD1 mice. Interestingly, we also noted a strong increase in ASIC2 expression in motoneurons of SOD1 mice and sporadic ALS patients during disease progression. Pharmacological pan-inhibition of ASIC channels with the lipophilic amiloride derivative, 5-(N,N-dimethyl)-amiloride hydrochloride, potently protected cultured motoneurons against acidotoxicity, and, given post-symptom onset, significantly improved lifespan, motor performance and motoneuron survival in SOD1 mice. Together, our data provide strong evidence for the involvement of acidotoxicity and ASIC channels in motoneuron degeneration, and highlight the potential of ASIC inhibitors as a new treatment approach for ALS.
So much for that strong evidence. As the retraction notice states:
The authors wish to retract the above paper. The journal has recently been notified by a reader who expressed concerns about some of the figures in this paper. Dr. Áine Behan, co-investigator of this study, has acknowledged sole responsibility for errors affecting certain panels in Figures 2b and 5, in which data were processed with the result that the figure panels do not correctly reflect the original data. Dr. Behan and co-author Dr. Bridget Breen have acknowledged responsibility for clerical errors in Figures 8b and 2a, respectively. Dr. Behan and Dr. Breen regret these errors, of which the co-authors Marion Hogg, Ina Woods, Karen Coughlan, Mollie Mitchem and Jochen Prehn were completely unaware. We thank the reader for bringing this to our attention and sincerely apologize to the scientific community for any confusion or adverse consequences resulting from the publication of these data.
Of course, what “clerical errors” might mean is anyone’s guess. But the explanation for the flaws in figures 2b and 5 is, well, risible. How is saying that the “data were processed with the result that the figure panels do not correctly reflect the original data” different from saying “I falsified it?”
The paper has been cited three times, according to Thomson Scientific’s Web of Knowledge.
Behan’s LinkedIn page shows her as being with the Royal College of Surgeons in Ireland, and a co-founder of a gaming company called CorTechs. According to this website:
Cortechs designs products that integrate thought-controlled technology into games & brain training applications for the masses. Cortechs create a new opportunity to communicate using brain-sensing technology which has been limited and unaffordable before now.
“brain training applications for the masses”. That’s scary! Of even greater concern is the following statement: “Dr. Behan and Dr. Breen regret these errors, of which the co-authors Marion Hogg, Ina Woods, Karen Coughlan, Mollie Mitchem and Jochen Prehn were completely unaware”. Isn’t one of the rules of authorship (I think No. 3) by the ICMJE that ALL authors be aware of, approve of and take responsibility for the submission of a paper to a journal? Are Hogg, Woods, Coughlan, Mitchem and Prehn openly claiming that perhaps they are guest authors or are in fact not valid authors according to the ICMJE definitions? Or is it just that they want to scape-goat their responsibilities? I wonder how much pressure was placed on Dr. Behan to assume sole respnsibility?
https://pubpeer.com/publications/18665234
Are lanes 1 and 2 (GRP78) of Figure 1B identical to lanes 2 and 3 (CHOP) of Figure 2A?
Kögel and Prehn have a lot of papers with interesting figures, e.g. control lanes of C and E:
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2171851/figure/fig5/
Sorry, I forgot the links:
Are lanes 1 and 2 (GRP78) of Figure 1B identical to lanes 2 and 3 (CHOP) of Figure 2A?
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2154478/figure/F1/
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2154478/figure/F2/
Cell Death Differ. 2010 Mar;17(3):459-68. doi: 10.1038/cdd.2009.134. Epub 2009 Sep 25.
Contrasting patterns of Bim induction and neuroprotection in Bim-deficient mice between hippocampus and neocortex after status epilepticus.
Murphy BM, Engel T, Paucard A, Hatazaki S, Mouri G, Tanaka K, Tuffy LP, Jimenez-Mateos EM, Woods I, Dunleavy M, Bonner HP, Meller R, Simon RP, Strasser A, Prehn JH, Henshall DC.
Source
Department of Physiology and Medical Physics, Royal College of Surgeons in Ireland, Dublin, Ireland.
PMID: 19779495
Please compare Actin panel in figure 1D Cell Death Differ. 2010 Mar;17(3):459-68
http://www.ncbi.nlm.nih.gov/core/lw/2.0/html/tileshop_pmc/tileshop_pmc_inline.html?title=Click%20on%20image%20to%20zoom&p=PMC3&id=2950266_nihms235005f1.jpg
with Actin panel in figure 2A FASEB J. 2010 Mar;24(3):853-61.
http://www.ncbi.nlm.nih.gov/core/lw/2.0/html/tileshop_pmc/tileshop_pmc_inline.html?title=Click%20on%20image%20to%20zoom&p=PMC3&id=3231945_z380031076720002.jpg
For reference:
FASEB J. 2010 Mar;24(3):853-61. doi: 10.1096/fj.09-145870.
PMID: 19890018
I’m fairly experienced in digital image analysis, but I fail to see a problem in the above linked cases. The bands clearly appear to be different, no?
Sorry, I referred to the first post that linked the pubpeer entry and the figure from the JCB journal.
I didn’t check the pubpeer entry. I just noticed there was an entry on this group there. As for the JCB paper the control lanes are clearly spliced. Another obvious (so many dots on the blots) example is in Figure 5 of the following paper, where the last two lanes of B have been reused in F:
Skorupa A, King MA, Aparicio IM, Dussmann H, Coughlan K, Breen B, Kieran D, Concannon CG, Marin P, Prehn JH.
J Neurosci. 2012 Apr 11;32(15):5024-38. doi: 10.1523/JNEUROSCI.6366-11.2012.
PMID: 22496549
The original paper can still be accessed here:
http://www.nature.com/cdd/journal/v20/n4/full/cdd2012158a.html. You may judge for yourself the “clerical errors” referred to in the retraction notice.
Having spent a good deal of time on this blog the past two days, I would like to contribute once more. Firstly, I wish to say that retractionwatch makes for some very interesting reading, and I say that from the position of an editor that has already seen a fair share of fabrication and manipulation cases.
Regarding the paper that this blog post is about, I think that the retraction of the paper is the right decision. The quoted figure panels obviously indicate that the data have been manipulated/fabricated.
I had a closer look at the other studies as well.
I read the PNAS paper, and the bands look indeed somewhat similar. Whether they are absolutely identical I cannot tell from the resolution that I have available. There may or may not be something to it. Importantly, the samples would be expected to originate from different materials.
I have had a look at the other linked studies as well, and overall I have to say that I disagree with the commenters’ characterization of the images.
The actin signals used in the FASEB and Cell Death Differ papers indeed appear identical. Seeing the experimental design, I suppose that both the Bim and Puma signals in the two papers were captured from the same membrane. If that is the case, it would be correct to show the same loading control. It obviously would be more desirable that the authors had shown a separate experiment. However, I would in no way put this into the context of a blog post on data fabrication or falsification.
Similarly, it seems that in the next paper (J Neurosci) linked by user Junk Science indeed the same signals have been shown in two subpanels of the linked figure. However, while I am not a neuroscientist, it seems to me from the figure caption that the authors show the data twice in the context of two separate comparisons and that the treatment conditions referred to for the lanes in questions indeed appear to be identical for both subpanels. Again, I would in no way put this into the context of data fabrication or falsification. This double display/repeat comparison may in fact support the storytelling, the flow of the full paper and its accessibility to the reader (but again, I am not from that field). As an editor and author I have seen cases where manuscript reviewers requested rearrangements of data and display items resulting in figures as shown here.
Back to the JCB paper and the spliced lanes pointed out by user Junk Science. Obviously, the lanes are spliced. But why would that be a problem if the data are from the same gel or membrane. In fact, in the figure legend the authors clearly state that “White lines indicate that intervening lanes have been spliced out.” Again, I would in no way put this into the context of data fabrication or falsification.
On top of that a link to the pubpeer site (which I was not familiar with before, thanks for the heads up) admittedly was blindly posted by Junk Science without investigating whether the comment posted there has any merit. Again, based on the information available to me I would in no way put the linked study into the context of data fabrication or falsification. Likewise, I find it unacceptable and irresponsible to post a comment that some of the authors have “a lot of papers with interesting figures”.
Taken together, I consider the respective comments above to be highly questionable and inappropriate and based on sloppy analysis. In the context of the original blog post on a rightfully retracted paper, these comments are a huge disappointment (at least for a novice on this site). If commenters expect quality and integrity from science and scientists, they will need to provide that as well. I find these comments unacceptable, from the perspective of an experienced scientist, author and editor, and also from the perspective of a formerly young postgraduate student and postdoctoral researcher. I am all for documenting fraud and misconduct where it occurs, but this needs to be based on a thorough in-depth analysis which includes reading the paper, at the very least the associated figure legends. Every such contribution needs to be based on solid ground. Let us not forget about the enthusiastic researchers and their hard work that is being put into scientific papers, whether they are junior or senior. To have a career in science is difficult enough, and false allegations, directly expressed or indirectly hinted at, can destroy hopes, careers and reputations. Posting on here can be done anonymously, quick and easy, but it goes hand-in-hand with being responsible for such mostly invisible but dramatic damage that can be done through “false positives”. I would like to see a more sensible approach on this site, and I hope that the previous posters and readers can see some value in the point that I just made.
Since I am more at home in the chemistry domain, a discipline that had its own rattles and rolls recently, I would like to link to the following blog post which puts it very well: http://www.rheothing.com/2013/09/fraud-in-literature-blogs-and-witchhunts.html
Miley Cyrus
If you do not see anything wrong with images at the ‘resolution available’ to you – perhaps you could state what the resolution is as that would show what you mean.
If you are able to see individual pixels and still not see any fabrication – that is somewhat unfathomable as the image manipulation is as clear as the sky is blue on a sunny day.
What I suggest is you put the image into something like powerpoint and enlarge the image, say, 800%. That resolution will show pretty much anything. Consider changing the contrast too.
Miley Cyprus November 21, 2013 at 5:28 pm
“The actin signals used in the FASEB and Cell Death Differ papers indeed appear identical.”
That is image and data resue in another paper. You are not supposed to do that.
Format: AbstractSend to
J Neurosci. 2008 Dec 24;28(52):14056-61. doi: 10.1523/JNEUROSCI.3399-08.2008.
Control of motoneuron survival by angiogenin.
Kieran D1, Sebastia J, Greenway MJ, King MA, Connaughton D, Concannon CG, Fenner B, Hardiman O, Prehn JH.
Author information
1
Department of Physiology and Medical Physics and Royal College of Surgeons in Ireland Neuroscience Research Centre, RCSI, Dublin 2, Ireland.
Much more similar than you would expect.
https://pubpeer.com/publications/1FA2EBA4A326D65A49C257E31A0D69#1 https://pubpeer.com/publications/1FA2EBA4A326D65A49C257E31A0D69#2 https://pubpeer.com/publications/1FA2EBA4A326D65A49C257E31A0D69#6
Proc Natl Acad Sci U S A. 2007 Dec 18;104(51):20606-11. Epub 2007 Dec 11.
Deletion of the BH3-only protein puma protects motoneurons from ER stress-induced apoptosis and delays motoneuron loss in ALS mice.
Kieran D1, Woods I, Villunger A, Strasser A, Prehn JH.
Author information
1
Department of Physiology and Medical Physics and Royal College of Surgeons in Ireland Neuroscience Research Centre, Royal College of Surgeons in Ireland, 123 St. Stephen’s Green, Dublin 2, Ireland.
Much more similar than you would expect.
https://pubpeer.com/publications/7F91083EB58E4E946D2B3245576F68#1
Dis Model Mech. 2019 Aug 13;12(8). pii: dmm040238. doi: 10.1242/dmm.040238.
Vascular regression precedes motor neuron loss in the FUS (1-359) ALS mouse model.
Crivello M1, Hogg MC1,2, Jirström E1,2, Halang L1, Woods I1, Rayner M1, Coughlan KS1, Lewandowski SA3,4, Prehn JHM1,2.
Author information
1
Department of Physiology and Medical Physics, Centre for the Study of Neurological Disorders, Royal College of Surgeons in Ireland, 123 St. Stephen’s Green, Dublin 2, Ireland.
2
FutureNeuro Research Centre, Royal College of Surgeons in Ireland, St. Stephen’s Green, Dublin 2, Ireland.
3
Tissue Biology Laboratory, Department of Medical Biochemistry and Biophysics, Center for Molecular Medicine, Karolinska Institute, Scheeles v. 2, 17177 Stockholm, Sweden.
4
Department of Physiology and Medical Physics, Centre for the Study of Neurological Disorders, Royal College of Surgeons in Ireland, 123 St. Stephen’s Green, Dublin 2, Ireland [email protected]
Much more similar than you would expect.
https://pubpeer.com/publications/6C40F304AB24C8A548551673DDF8DF#1
Cell Death Differ. 2009 Sep;16(9):1238-47. doi: 10.1038/cdd.2009.52. Epub 2009 May 15.
Angiogenin protects motoneurons against hypoxic injury.
Sebastià J1, Kieran D, Breen B, King MA, Netteland DF, Joyce D, Fitzpatrick SF, Taylor CT, Prehn JH.
Author information
1
Department of Physiology and Medical Physics and RCSI Neuroscience Research Centre, Royal College of Surgeons in Ireland, Dublin, Dublin 2, Ireland.
Much more similar than you would expect (different treatments, not overlay).
https://pubpeer.com/publications/7840919D4CFF0760DA9F03D100DEBC#1