The former president of the Joslin Diabetes Center has withdrawn a second article within a month of his first, and issued extensive corrections to another paper in the same journal, all due to figure errors.
In November, we reported that Carl Ronald Kahn — also affiliated with Harvard Medical School — had pulled a highly cited 2005 paper from The Journal of Clinical Investigation because of image duplication issues, which Kahn told us were introduced during figure assembly. This December, Kahn retracted a 2003 paper published in The Journal of Biological Chemistry (JBC)—again due to duplication issues that the authors believe “were inadvertently introduced during figure assembly.”
Here’s the retraction notice for “Bi-directional regulation of brown fat adipogenesis by the insulin receptor,” cited 46 times, according to Clarivate Analytics’ Web of Science, formerly part of Thomson Reuters:
This article has been withdrawn by the authors. The authors were recently made aware of duplicated images of PCR reaction products in panels B and D in Fig. 2. These duplications were inadvertently introduced during figure assembly. Review of the original data generated in the lab at that time (2000–2002), as well as subsequent studies, confirmed the conclusions of the manuscript. However, in the interest of maintaining accuracy in the published scientific literature and because the initial figures were not up to the standards of JBC, the authors wish to withdraw this article. The authors apologize for these errors.
Regarding the retraction, Kahn told us something very similar to the retraction notice – namely that the mistakes were introduced while assembling the figures, but the findings remain valid:
We were able to identify the original data two independent experiments generated in the lab at the time the work was performed (2000-2002). These were very similar to the published figure and confirmed the results and conclusions of the manuscript.
We reached out to first author Amelia Entingh-Pearsall, but have not heard back.
The paper has been discussed on PubPeer, with one user alleging issues in Figure 2.
One of the authors on the latest retraction, Cullen Taniguchi, now based at MD Anderson Cancer Center, was the first author on the JCI paper that was retracted late last year. Taniguchi told his contributions to the JBC paper were “minimal:”
I had performed a few pilot experiments for this project under the direct supervision of Dr. Pearsall (née Entingh) as a rotating summer student, and none of these data were used in the paper.
In October, Kahn and Entingh-Pearsall issued corrections to a 2004 paper, also published in JBC and again because of figure errors. Here’s the correction for “Differential roles of the insulin and insulin-like growth factor-I (IGF-I) receptors in response to insulin and IGF-I:”
There were several errors in this paper that the authors wish to correct. In the corrected Fig. 1, the panel showing the IGF1KO cells has been deleted because it appears to be an incorrect image. As Fig. 2 contains three example clones of IGFRKO cell differentiation, the presence of the IGFRKO panel in Fig. 1, which was not referred to under the “Results,” was redundant. We have also added dashed lines to indicate that the image was a composite of different tissue culture dishes. In the experiments shown in Fig. 6, A and B, the samples were loaded sequentially on the same gel, and the same basal sample served as 0 nM ligand for both insulin and IGF-1 stimulation. To create the double panel figures used in the paper, therefore, the basal sample lane was duplicated and spliced next to the IGF-1 stimulated samples. To clarify the confusion that this has created, a vertical white line has been added to the corrected Fig. 6 to all of the IGF-1 panels to indicate that the unstimulated lane to the left of the line has been spliced in and is the same unstimulated sample as appears in the insulin panels. The experiments in Fig. 7, B and C, were performed in the same fashion as those in Fig. 6. Therefore, the basal lanes shown in the IGF-1-stimulated panels are the same as the insulin-stimulated panels. For clarification, vertical white lines have been added to these figures to indicate the area of vertical splicing. None of these clarifications has any effect on the results of the study or their interpretation.
The 2004 study has been cited 77 times. A user also flagged figures 1 and 6 on PubPeer.
Kaoru Sakabe—data integrity manager at the American Society for Biochemistry and Molecular Biology (which publishes JBC)—told us:
A reader brought these articles to our attention. The details for the correction and withdrawal may be found in the accompanying notices.
We’ve identified two more corrections Kahn issued in 2016, also citing issues with images, such as labeling mistakes – one in JBC, and another in Cell Metabolism.
Like Retraction Watch? Consider making a tax-deductible contribution to support our growth. You can also follow us on Twitter, like us on Facebook, add us to your RSS reader, sign up on our homepage for an email every time there’s a new post, or subscribe to our daily digest. Click here to review our Comments Policy. For a sneak peek at what we’re working on, click here.
Kaoru Sakabe (data integrity manager at the American Society for Biochemistry and Molecular Biology (which publishes JBC)) might take a look at the figures in this CR Kahn paper
J Biol Chem. 2003 Nov 28;278(48):48453-66. Epub 2003 Sep 22.
Positive and negative roles of p85 alpha and p85 beta regulatory subunits of phosphoinositide 3-kinase in insulin signaling.
Ueki K1, Fruman DA, Yballe CM, Fasshauer M, Klein J, Asano T, Cantley LC, Kahn CR.
Author information
1Research Division, Joslin Diabetes Center and Harvard Medical School, Boston, Massachusetts 02215, USA.
Figure 3a.
http://i.imgur.com/WENMG4x.jpg
Figure 5c.
http://i.imgur.com/d5UVdkx.jpg
Figure 5a.
http://i.imgur.com/gOAdYZw.jpg
Pubpeer comments: https://pubpeer.com/publications/14504291
Mol Cell Biol. 2004 Jun;24(12):5434-46.
Suppressor of cytokine signaling 1 (SOCS-1) and SOCS-3 cause insulin resistance through inhibition of tyrosine phosphorylation of insulin receptor substrate proteins by discrete mechanisms.
Ueki K1, Kondo T, Kahn CR.
Author information
1Research Division, Joslin Diabetes Center, and Department of Medicine, Harvard Medical School, Boston, MA 02215, USA.
Figure 6a.
http://i.imgur.com/YlDLEmw.jpg
Figures 6a and 6b.
http://i.imgur.com/lFigzhZ.jpg
https://pubpeer.com/publications/15169905
Nat Cell Biol. 2005 Jun;7(6):601-11. Epub 2005 May 15.
Prediction of preadipocyte differentiation by gene expression reveals role of insulin receptor substrates and necdin.
Tseng YH1, Butte AJ, Kokkotou E, Yechoor VK, Taniguchi CM, Kriauciunas KM, Cypess AM, Niinobe M, Yoshikawa K, Patti ME, Kahn CR.
Author information
1Research Division, Joslin Diabetes Center, Children’s Hospital, Harvard Medical School, Boston, MA 02215, USA.
Figure 3b.
http://i.imgur.com/DUVh2ZC.jpg
Figure 5a.
http://i.imgur.com/XHuAZ2a.jpg
https://pubpeer.com/publications/15895078
Mol Cell Biol. 2005 Mar;25(5):1596-607.
Phosphoinositide 3-kinase catalytic subunit deletion and regulatory subunit deletion have opposite effects on insulin sensitivity in mice.
Brachmann SM1, Ueki K, Engelman JA, Kahn RC, Cantley LC.
Author information
1Beth Israel Hospital, NRB, Division of Signal Transduction, Department of Systems Biology, 10th Floor, 330, Brookline, MA 02215, USA.
Figures 3b and 5a.
http://i.imgur.com/uw1n89b.jpg
https://pubpeer.com/publications/15713620
So no independent investigation after all these issues?
I am beginning to wonder, given the very large number of retrations based on blot image errors, duplications,and manipulations, who does the assembly for publication, and who checks it in most cases. PI? Post doc? Lab tech? It makes me suspect it is someone who cannot actually read and discriminate western blots from one another.
That’s one theory.
Let us not forget the ultimate responsibility is of PI who is running the show, he/she has to be clear and confident about the data how it was collected, analyzed and presented to him for inclusion in the publication. No one else!
Do you think that people making mistakes with western, and other blots, is the equivalent to making typos? In any event blots are only crude measurements of what is going on in cells or organism. They are not in themselves and experimental intervention, or engineering.
It is equivalent to making typos while writing the numbers in your table of results, except that each typo accidentally makes your data look more conclusive than it really is.
Is it common, or a best practice for the level of involvement described below to be acknowledged by inclusion as one of the “authors” of a paper?
Taniguchi: “I had performed a few pilot experiments for this project under the direct supervision of Dr. Pearsall (née Entingh) as a rotating summer student, and none of these data were used in the paper.”
As I am only a casual reader of retractionwatch, the answer may be obvious to those in the industry, however, wouldn’t scientific literature be improved if authors identified their contributions up front? Seems like detailed attribution would credit those doing the difficult science, limit resume padding, and aid investigatory bodies if a paper later comes under criticism.
Well, depending on the amount of work involved, I would certainly consider including authors who made a contribution, even if that contribution is not necessarily used directly in the paper.
These days author contribution statements make this easy “X performed pilot experiments that were crucial to the successful development of technique Y” etc.
CM Taniguchi as first author on 2012 Nat Med paper.
Nat Med. 2013 Oct;19(10):1325-30. doi: 10.1038/nm.3294. Epub 2013 Sep 15.
Cross-talk between hypoxia and insulin signaling through Phd3 regulates hepatic glucose and lipid metabolism and ameliorates diabetes.
Taniguchi CM1, Finger EC, Krieg AJ, Wu C, Diep AN, LaGory EL, Wei K, McGinnis LM, Yuan J, Kuo CJ, Giaccia AJ.
Author information
1Division of Radiation and Cancer Biology, Department of Radiation Oncology, Center for Clinical Sciences Research, Stanford, California, USA.
https://pubpeer.com/publications/FB7695753BB37D69E5FC65189D8B5D
Figure 4. http://i.imgur.com/ZZH2uvl.jpg
CR Kahn and CM Taniguchi seond and third last authors on this 2009 paper.
Diabetologia. 2009 Jun;52(6):1197-207. doi: 10.1007/s00125-009-1336-5. Epub 2009 Apr 9.
Role of atypical protein kinase C in activation of sterol regulatory element binding protein-1c and nuclear factor kappa B (NFkappaB) in liver of rodents used as a model of diabetes, and relationships to hyperlipidaemia and insulin resistance.
Sajan MP1, Standaert ML, Rivas J, Miura A, Kanoh Y, Soto J, Taniguchi CM, Kahn CR, Farese RV.
Author information
1Research Service, James A Haley Veterans Hospital, Tampa, FL 33612, USA.
https://pubpeer.com/publications/EE929C5F5B16242E1C4C05520BFAE8
Figure 3b. http://i.imgur.com/ORuqDjK.jpg
Figure 7c. http://i.imgur.com/VxdVYWb.jpg
J Biol Chem. 2003 Aug 22;278(34):31964-71. Epub 2003 May 29.
Characterization of multiple signaling pathways of insulin in the regulation of vascular endothelial growth factor expression in vascular cells and angiogenesis.
Jiang ZY1, He Z, King BL, Kuroki T, Opland DM, Suzuma K, Suzuma I, Ueki K, Kulkarni RN, Kahn CR, King GL.
Author information
1Research Division, Joslin Diabetes Center, Harvard Medical School, Boston, Massachusetts 02215, USA.
https://pubpeer.com/publications/12775712
Figure 3C. http://i.imgur.com/kRoOuM5.jpg
Third retraction senior author.
Retraction of: J Biol Chem. 2003 Nov 28;278(48):48453-66. Epub 2003 Sep 22.
Positive and negative roles of p85 alpha and p85 beta regulatory subunits of phosphoinositide 3-kinase in insulin signaling.
Ueki K1, Fruman DA, Yballe CM, Fasshauer M, Klein J, Asano T, Cantley LC, Kahn CR.
Author information
1
Research Division, Joslin Diabetes Center and Harvard Medical School, Boston, Massachusetts 02215, USA.
2017 retraction notice.
http://www.jbc.org/content/292/13/5608
VOLUME 278 (2003) PAGES 48453–48466
This article has been withdrawn by the authors. In many of the experiments reported in this study, cells from mice of four genotypes were used (wild-type, p85α−/−, p85α+/−, and p85β−/−), but data from only three of the genotypes (wild-type, p85α−/−, and p85β−/−) were included in the final paper. As a result, there was splicing of the figures of several autoradiograms, which led to several duplicated or mislabeled lanes in the Western blots in Figs. 2B, 3C, and 5B. Although the experimental data generated in the lab from the same time period support the original conclusions of the study, and the studies by this lab and others have confirmed and extended the conclusions of the manuscript, in the interest of maintaining accuracy in the published scientific literature and because the initial figures were not up to the standards of JBC, the authors wish to withdraw this article. The authors apologize for these errors.
© 2017 by The American Society for Biochemistry and Molecular Biology, Inc.