A group of smoking researchers — no, not scientists who are on fire; scientists who study the effects of tobacco smoke — has retracted a 2009 article after deciding that they were no longer “satisfied with the quality of the data.”
The paper, “Cigarette Smoke–induced Oxidative/Nitrosative Stress Impairs VEGF- and Fluid Shear Stress–Mediated Signaling in Endothelial Cells,” came from the lab of Irfan Rahman, a lung disease expert at the University of Rochester. It appeared online in 2009 in Antioxidants & Redox Signaling, which will be familiar to readers watching the case of Dipak Das.
As the notice explains:
The corresponding author has sought retraction of this work from ARS. The author statement is copied below: We, the authors, wish to retract “Cigarette Smoke–induced Oxidative/Nitrosative Stress Impairs VEGF- and Fluid Shear Stress–Mediated Signaling in Endothelial Cells” by Edirisinghe et al (Antioxid Redox Signal, 12(12): 1355-1369. doi:10.1089/ars.2009.2874) because we are not satisfied with the quality of some of the data presented in the paper. Overall, however, the data are reproducible and the conclusions drawn were not affected. We apologize for any inconvenience this may have caused to the readers.
The paper has been cited 18 times, according to Thomson Scientific’s Web of Knowledge.
We tried to reach Rahman for more information but haven’t heard back from him yet. We’d like to know in particular why it took three years for the authors to figure out that they published an article with subpar data — and how those data could in turn be reproducible, exactly.
And anyway, how is it possible that the data were so poor as to require retraction, but good enough to be reproducible, and not damaging to the study conclusions? The short answer: Doesn’t seem likely.
Perhaps the “copied below” note in the second sentence of the notice reflects the same skepticism on the part of the editors. We’ve tried them for comment too.
Maybe the authors seek much higher standards for their own data than it is typically expected from mere mortals. If this causes some studies to go up in smoke, so be it.
I would imagine they are now using a new, somehow better method to determine the same things that are in the original paper but that this method simply does not give results consistent with what they’ve got back then. Reproducibility itself merely indicates consistency within a method, but does not say much about truth (along the lines of the ‘precision vs accuracy’ discussion). If the method itself turns out not to show what you believed it showed at some point in the past, you would just want to drop it (and the results) altogether, I take it?
I just hope the paper does not contain any Western blots. It is very easy to be dissatisfied with those…
If the data does “not..show what you believed it showed…in the past,” but the “conclusions drawn were not affected,” wouldn’t this have merited a new paper rather than a retraction? The emergence of an (alleged) new and improved method is in my opinion insufficient grounds for retraction. If it is sufficient grounds, then wouldn’t pretty much all published scientific papers eventually need to be retracted?
I suspect there is more going on here than a new method, or even a new-found elevation in the authors’ standards. Forgive the pun, but where there is smoke…
Each figure contains Western blots. There is an obvious duplication of a band in Fig 1A (Akt, lanes 1 and 3) as well as white rims around a lot of bands in Fig 7
The paper on the ARS website itself has been replaced by a single page notice indicating the retraction. However, the full MS is still available on UK PubMed Central.
http://europepmc.org/abstract/MED/19929443
As irre notes above, problems with the western blots…
Fig. 1A – pAKT blot spliced between lanes 4 & 5
Fig. 4A – pAKT blot spliced between 5&6, and 7&8.
Fig. 4A – pAKT blot – tops of lanes do not all line up evenly. Lanes 1 & 2 are slightly lower, so may have been spliced in.
Fig. 7A. p-eNOS blot contains 12.5 bands, but the eNOS blot below contains 12. Bands don’t line up.
My comment is awaiting moderation. But have a look of other Rahman and/or Edirisinghe publications, you will find a lot!
We edited a potentially libelous phrase out of your comment and have approved it.
Hi Ivan,
Another retraction and a correction from the same people as reported below by StrongDreams and fernando pessoa, a follow-post seems warranted!
Thanks Ivan! I got too excited when someone states that the conclusions drawn are still valid etc etc!
This my friend could be an opening into something much bigger (see Rakesh Kumar). We have to figure out whether it is just one bad apple. The first author Indika Edirisinghe has several papers with possible image issues [comment edited] where Irfan Rahman is not the PI. For example, check out Fig. 5A and you wonder who reviewed that paper and did not spot that (Reviewer 2: Stevie Wonder):
Edirisinghe I, Banaszewski K, Cappozzo J, McCarthy D, Burton-Freeman BM.
J Agric Food Chem. 2011 Aug 24;59(16):8616-24. doi: 10.1021/jf201116y. Epub 2011 Jul 28.
PMID: 21761876
There are also examples where he is not the first author, but basically all Westerns in this paper have been manipulated:
Yao H, Edirisinghe I, Yang SR, Rajendrasozhan S, Kode A, Caito S, Adenuga D, Rahman I.
Am J Pathol. 2008 May;172(5):1222-37. doi: 10.2353/ajpath.2008.070765. Epub 2008 Apr 10.
PMID: 18403597
One example where he is not author at all. Fig 5, you cannot anchor the bands at all:
Extracellular superoxide dismutase protects against pulmonary emphysema by attenuating oxidative fragmentation of ECM.
Yao H, Arunachalam G, Hwang JW, Chung S, Sundar IK, Kinnula VL, Crapo JD, Rahman I.
Proc Natl Acad Sci U S A. 2010 Aug 31;107(35):15571-6. doi: 10.1073/pnas.1007625107. Epub 2010 Aug 16.
PMID: 20713693
Start digging people, I have to work.
Another correction by Edirisinghe related to image manipulation.
http://www.ncbi.nlm.nih.gov/pubmed/23327621
Thanks StrongDreams, it’s the one I reported above.
http://www.ncbi.nlm.nih.gov/pubmed/19238689
How are the Texas curcuminators doing?
Thanks, hopefully, more is coming.
That looks like a review article. I guess it either relied on or re-used figures from the good Dr. E. (but I can’t access it to be sure). I imagine Dr. R. is quite distressed by now. Shows what happens when you let your students show you finished paper figures and don’t demand to see the originals. I really hate it that I now have to ask every one of my students to show me original films and printouts and not just lab meeting summaries, but I guess that’s how things are these days.
Actually, that review was retracted for “substantial overlap” and not figure problems, according to the retraction notice. Dr. Rahman publishes an awful lot of review articles.
And in some cases you see the finished figures and don’t spot the similarities between Control and H2O2 (4 hr Hoehst dye) in Figure 1A. PMID: 15456740 FASEB J. 2004 Dec;18(15):1897-9. Epub 2004 Sep 28.
Oxidative stress and cigarette smoke alter chromatin remodeling but differentially regulate NF-kappaB activation and proinflammatory cytokine release in alveolar epithelial cells.
Moodie FM, Marwick JA, Anderson CS, Szulakowski P, Biswas SK, Bauter MR, Kilty I, Rahman I.
“I imagine Dr. R. (Rahman) is quite distressed by now. Shows what happens when you let your students show you finished paper figures and don’t demand to see the originals.”
Sorry for being concrete, but how do you know that? Usually as the work is going along whoever is leading the group, and others, will see the primary data. There will usually be a session where people are asked if some to be presented data comes from particular primary data. People will ask if this control is the control for particular results.
“Usually as the work is going along whoever is leading the group, and others, will see the primary data.”
IF the PI ever saw the blue film. When I started in science 25 years ago, we had weekly lab meetings and brought our x-ray films and blot membranes and put them right on an overhead projector for the whole room. These days, every lab meeting I attend (the work-in-progress meetings in my lab and the labs I collaborate with) is done with Powerpoint. The PI gets weekly updates of the work in his or her lab but may never actually see an original film unless he or she insists.
I prefer to think that most of these episodes are the result of students who are in a rush to get a pretty looking version of something they already believe to be true, and the PI has only failed to check the original films. I don’t want to think that PIs might actually be instructing their labs to fake up better looking blots (although that seems to have happened in at least some of the cases noted on the blog).
Here’s an old one, this might really be an accident?
http://respiratory-research.com/content/9/1/6
Of course it is, just like this one where they happened to flip the two uppermost far-right IKKa and p-s176/180-IKKa blots in 6C:
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2396248/figure/fig6/
Wow. Do you use some kind of software too or just have really good eyes?
The one on top has some additional blurring. Probably they applied some intensity threshold on the bottom one
I am impressed! Bravo Junk Science!
Figure 2. J Inflamm (Lond). 2010; 7: 11.
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2831890/figure/F2/
The 4 bands in the p-AKT panel, figure 4A, J Inflamm (Lond). 2010; 7: 11. look like the 4 left-most bands in the eNOS panel, Figure 4A, Antioxid Redox Signal. 2010 Jun 15;12(12):1355-69.
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=2864667_fig-4.jpg
Those two panels are of different things.
In reply to strongDreams February 3, 2013 at 5:57 pm
Not all changes are for the better. “New, worse washing detergent!”. They don’t tell you that at the supermarket.
Are the powerpoint presentations so the PIs don’t have to pay attention, but can be on the mobile phone, or on aeroplanes at the same time?
J Inflamm (Lond)
. 2010 Feb 13;7(1):11. doi: 10.1186/1476-9255-7-11.
Cigarette smoke regulates VEGFR2-mediated survival signaling in rat lungs
John A Marwick 1, Indika Edirisinghe, Gnanapragasam Arunachalam, Christopher S Stevenson, William Macnee, Paul A Kirkham, Irfan Rahman
Affiliation
1National Heart and Lung Institute, Imperial College London, UK. [email protected].
PMID: 20205917
Images very similar for different time points.
https://pubpeer.com/publications/01BAE656F97EED0947C378F78C5539
FASEB J. 2004 Dec;18(15):1897-9. doi: 10.1096/fj.04-1506fje. Epub 2004 Sep 28.
Problematic data figures 4A and 6A. Much more similar than expected.
See: https://imgur.com/ELCAJVX
Oxidative stress and cigarette smoke alter chromatin remodeling but differentially regulate NF-kappaB activation and proinflammatory cytokine release in alveolar epithelial cells
Fiona M Moodie 1, John A Marwick, Charlotte S Anderson, Patryk Szulakowski, Saibal K Biswas, Mark R Bauter, Iain Kilty, Irfan Rahman
Affiliation
1
ELEGI and Colt Research Laboratories, MRC Centre for Inflammation Research, University of Edinburgh Medical School, Edinburgh, UK.
https://www.urmc.rochester.edu/people/23899711-irfan-rahman
Under publications:
12/2004
Moodie FM, Marwick JA, Anderson CS, Szulakowski P, Biswas SK, Bauter MR, Kilty I, Rahman I. “Oxidative stress and cigarette smoke alter chromatin remodeling but differentially regulate NF-kappaB activation and proinflammatory cytokine release in alveolar epithelial cells.” FASEB journal : official publication of the Federation of American Societies for Experimental Biology.. 2004 Dec 0; 18(15):1897-9. Epub 2004 Sep 28.
FASEB J. 2008 Jul;22(7):2297-310. doi: 10.1096/fj.07-099481. Epub 2008 Feb 8.
Problematic data figure 5A. Much more similar after vertical stretch than expected.
See: https://imgur.com/lgw3xQZ
VEGFR-2 inhibition augments cigarette smoke-induced oxidative stress and inflammatory responses leading to endothelial dysfunction
Indika Edirisinghe 1, Se-Ran Yang, Hongwei Yao, Saravanan Rajendrasozhan, Samuel Caito, David Adenuga, Chelsea Wong, Arshad Rahman, Richard P Phipps, Zheng-Gen Jin, Irfan Rahman
Affiliation
1
Department of Environmental Medicine, Lung Biology and Disease Program, University of Rochester Medical Center, 601 Elmwood Ave., Rochester, NY 14642, USA.
Biochem Biophys Res Commun. 2010 Feb 26;393(1):66-72. doi: 10.1016/j.bbrc.2010.01.080. Epub 2010 Jan 25.
Problematic data figure 1A. Much more similar than expected.
See: https://imgur.com/7dsNdxK
SIRT1 regulates oxidant- and cigarette smoke-induced eNOS acetylation in endothelial cells: Role of resveratrolGnanapragasam Arunachalam 1, Hongwei Yao, Isaac K Sundar, Samuel Caito, Irfan RahmanAffiliations collapseAffiliation1Department of Environmental Medicine, Lung Biology and Disease Program, University of Rochester Medical Center, Rochester, NY 14642, USA.PMID: 20102704 PMCID: PMC2830376 DOI: 10.1016/j.bbrc.2010.01.080