Retraction Watch

Tracking retractions as a window into the scientific process

Mystery box sinks immunology paper

with 27 comments

Screen Shot 2014-06-30 at 12.30.32 PMBiochemical Journal has pulled a 2006 paper for an undisclosed “background subtraction box” in an image – which, if you take a not-particularly-close look at the figure to the right, means somebody added a black rectangle over the control lane.

Here’s the notice: for “Phosphorylation of Ser158 regulates inflammatory redox-dependent hepatocyte nuclear factor-4a transcriptional activity”:

It has been brought to our attention that Figure 1(A) contains a background subtraction box in the lane labelled ‘Control + IgG’. As this band represented an important negative control, this has serious potential to affect the conclusions drawn in the paper, and therefore we are retracting this paper. We thank the Editorial Board of the Biochemical Journal for pointing out this error.

Lead author Paul Kuo was at Duke at the time; he is now chair of surgery at Loyola University.

Here’s the abstract for “Phosphorylation of Ser158 regulates inflammatory redox-dependent hepatocyte nuclear factor-4α transcriptional activity:”

In IL-1β (interleukin 1β)-stimulated rat hepatocytes exposed to superoxide, we have previously identified an IRX (inflammatory redox)-sensitive DR1 [direct repeat of RG(G/T)TCA with one base spacing] cis-acting activator element (nt –1327 to –1315) in the iNOS (inducible nitric oxide synthase) promoter: AGGTCAGGGGACA. The corresponding transcription factor was identified to be HNF4α (hepatocyte nuclear factor-4α). HNF4α DNA binding activity and transactivation potential are tightly regulated by its state of phosphorylation. However, the functional consequences of IRX-mediated post-translational phosphorylation of HNF4α have not been well characterized. In the setting of IL-1β + H2 O2 , HNF4α functional activity is as- sociated with a unique serine/threonine phosphorylation pattern. This indicates that an IRX-sensitive serine/threonine kinase path- way targets HNF4α to augment hepatocyte iNOS transcription. In the present study, following identification of phosphorylated residues in HNF4α, serial mutations were performed to render the target residues phosphorylation-resistant. Electrophoretic mobi- lity-shift assays and transient transfection studies utilizing the iNOS promoter showed that the S158A mutation ablates IRX- mediated HNF4α DNA binding and transactivation. Gain-of- function mutation with the S158D phosphomimetic HNF4α vec- tor supports a critical role for Ser158 phosphorylation. In vitro phosphorylation and kinase inhibitor studies implicate p38 kinase activity. Our results indicate that p38 kinase-mediated Ser158 phos- phorylation is essential for augmentation of the DNA binding and transactivation potential of HNF4α in the presence of IL- 1β + H2 O2 . This pathway results in enhanced iNOS expression in hepatocytes exposed to pro-inflammatory cytokines and oxidative stress.

The paper has been cited 22 times, according to Google Scholar. We’ve reached out to Kuo and the editor, and will update with any new information we get.

Update, 5/19/16, 4:11 p.m. Eastern: We have posted an update to this story.

Update, 5/20/16, 3:45 p.m Eastern: Based on new developments in this story, we have changed the headline to better reflect the narrative.

  • Neuroskeptic (@Neuro_Skeptic) July 8, 2014 at 11:31 am

    Let me guess what the authors will say…

    “Although we strongly regret this error, it was unintentional, and also a lab technician (who is not an author) did it. We are pleased to report that we have conducted further experiments showing that the conclusions of the paper are still valid.”

    • Dan Zabetakis July 8, 2014 at 12:49 pm

      Alas, you are probably right.

      Anyway, we should probably call this a “foreground subtraction box”.

    • BB July 8, 2014 at 4:18 pm

      More like shoddy MSpainting… The laziness of these fraudsters is sometimes truly amazing, they did not even bother using the clone stamp tool or stg. similar, I bet they simply draw a black rectangle in MS Paint. My eyes are bleeding…

      Some of the other “papers” commented on pubpeer are even worse:

      Look at this:

      Were all the reviewers and the editor on acid? No one managed to spot that the most important band lane was simply copypasted within a pitch dark rectangle?

      • David Hardman July 9, 2014 at 2:19 am

        Ann Surg. 2012 Feb;255(2):319-25. doi: 10.1097/SLA.0b013e31823e3a1c.
        Osteopontin regulates epithelial mesenchymal transition-associated growth of hepatocellular cancer in a mouse xenograft model.
        Bhattacharya SD, Mi Z, Kim VM, Guo H, Talbot LJ, Kuo PC.
        Author information

        Department of Surgery, Duke University Medical Center, Durham, NC, USA.

        Supported by NIH R01 GM065113 (PCK), T32 GM069331 (LJT), and T32 CA093245 (SDB)

        Figure 1. Please compare signal lane marked as 6 with Non-Metastatic Tenascin-c panel with darker version lane marked as 4 Metastatic SMA panel.

        Please compare lanes marked as 7 and 8 Non-Metastatic Tenascin-c panel with darker version lanes marked as 5 and 6 Non-Metastatic SMA panel.

        Metatstatic SMA panel. Vertical, straight changes background between lanes 2,3 and 4.

        Non-Metastatic SMA panel. Vertical, straight changes background between lanes marked as 6,7 and 8.

        • David Hardman July 9, 2014 at 2:42 am

          Surgery. 2008 Aug;144(2):182-8. doi: 10.1016/j.surg.2008.03.007. Epub 2008 May 16.
          Osteopontin mediates Stat1 degradation to inhibit iNOS transcription in a cecal ligation and puncture model of sepsis.
          Guo H1, Wai PY, Mi Z, Gao C, Zhang J, Kuo PC.
          Author information

          1Department of Surgery, Duke University Medical Center, Durham, NC, USA.

          Figure 1.

          Figure 1. LIVER panel.

          Please compare right CLP 12 h OPN (-/-) P-Stat1 band with left CLP 12 h WT P-Stat1 band.

          • David Hardman July 9, 2014 at 3:19 am

            Cancer Res. 2007 May 1;67(9):4088-97.
            Thrombin-cleaved COOH(-) terminal osteopontin peptide binds with cyclophilin C to CD147 in murine breast cancer cells.
            Mi Z, Oliver T, Guo H, Gao C, Kuo PC.
            Author information

            Department of Surgery, Duke University Medical Center, Durham, North Carolina 27710, USA.

            Figure 5.
            Figure 5C. Lower cell lysate beta-actin panel. Please compare bands lanes 1 and 8.
            Bands lanes 1 and 8 cell lysate Pro-MMP2 panel are different.

            Figure 5C. Medium Pro-MMP2/Active-MMP2 panel.
            Please compare Pro-MMP2 bands lanes 1 and 8. Active-MMP2 bands are different.

        • BB July 10, 2014 at 5:30 am

          If you take a close look at the representative in vivo images, you can see that the scale bars display different upper and lower thresholds, though it impossible to read the actual numbers.

      • JATdS July 9, 2014 at 2:31 am

        If you magnify that image and look at the fringe of the black box, you will see a fairly faint “white halo” around most of the box, which indicates that the authors manipulated the lane to conceal a positive band and to make it appear as a negative one. In other words, the authors had a positive result, but made it look negative, i.e., manipulation. Automatically, that means all results are valid. The authors hould now be forced (by the publisher?) to contact the 22 journals that cited their paper and request an erratum. I think this should be the new norm in science publishing related to retractions. If however, the publisher is responsible for the errors that lead to a retraction, then it should be the publisher that should then contact the journals or books that have cited the retracted paper.

        • JATdS July 9, 2014 at 2:31 am

          Erratum: “that all results are valid” should read “that all results are invalid”.

  • David Hardman July 9, 2014 at 3:47 am

    J Immunol. 2007 Feb 1;178(3):1870-81.
    Osteopontin induces ubiquitin-dependent degradation of STAT1 in RAW264.7 murine macrophages.
    Gao C1, Guo H, Mi Z, Grusby MJ, Kuo PC.
    Author information

    1Department of Surgery, Duke University Medical Center, Durham, NC 27710, USA.

    This work was supported by National Institutes of Health Grants R01-GM65113 and R01-AI44629 (to P.C.K.) and R01-AI506296 (to M.J.G.).

    Figure 2.

    Figure 2b. iNOS panel. Please compare band LPS lane with band LPS+siRNA MM lane.

    Figure 9.

    Figure 9B.

    Please compare Control band siRNA MM SLIM panel with LPS+MG132 band WT SLIM panel.

    Please compare LPS+MG132 band siRNA MM OPN panel with LPS band WT SLIM panel.

    Please compare Control WT SLIM band with LPS+MG132 band siRNA MM SLIM panel and with LPS+MG132 band WT OPN panel.

  • Iain Moal July 9, 2014 at 3:54 am

    The citation count is reported from Google Scholar, while in previous articles it was obtained from Thomson Reuters Web of Science. Does anybody know why that is?

    • Ivan Oransky July 9, 2014 at 6:03 am

      Thanks for the query. We’ve been having some technical difficulties with our usual Web of Knowledge access. We’ve asked the relevant people to look into it.

  • David Hardman July 9, 2014 at 5:13 am

    Am J Physiol Cell Physiol. 2003 Apr;284(4):C1090-9. Epub 2002 Dec 4.
    Serine/threonine phosphorylation regulates HNF-4alpha-dependent redox-mediated iNOS expression in hepatocytes.
    Guo H1, Wei J, Inoue Y, Gonzalez FJ, Kuo PC.
    Author information

    1Duke University Medical Center, Durham, North Carolina 27710, USA.

    This work was supported by National Institutes of Health Grants AI-44629 and GM-65113 (to P. C. Kuo) and the American College of Surgeons Clowes Development Award (to P. C. Kuo).

    Figure 7.

    Please compare signal right panel with horizontal mirror image signal right 2 lanes left panel.

    The treatments are different.

    Figure 6. Vertical, straight, grey streaks/ vertical change background between lanes 3 and 4. There are only 4 lanes.

  • David Hardman July 9, 2014 at 5:30 am

    Carcinogenesis. 2003 Dec;24(12):1871-8. Epub 2003 Aug 29.
    Osteopontin-dependent CD44v6 expression and cell adhesion in HepG2 cells.
    Gao C1, Guo H, Downey L, Marroquin C, Wei J, Kuo PC.
    Author information

    1Department of Surgery, Duke University Medical Center, Durham, NC 27710, USA.

    Pubpeer comment:

  • BB July 9, 2014 at 6:19 am

    J Biol Chem. 2002 Feb 15;277(7):5054-60. Epub 2001 Dec 6.
    Hepatocyte nuclear factor-4alpha mediates redox sensitivity of inducible nitric-oxide synthase gene transcription.
    Guo H1, Cai CQ, Kuo PC.

    Full text:

    Please compare lane 2 and 3 of Fig1. with lane 1 and 2 of Fig3.

    What is your opinion? This paper has not been commented on pubpeer!

    • Neuroskeptic (@Neuro_Skeptic) July 9, 2014 at 7:03 am

      They’re identical with contrast changed. I think Fig 1 is the original. In Fig 3, the two bands look out of place and there’s a vertical background change to the right of them (at the bottom).

      • BB July 10, 2014 at 6:47 am

        Thanks, that was my impression as well. However, now that I wanted to take a second look at these figures I can no longer see to the images “403 Forbidden” though the text is accesible. Is it only me?

        • David Hardman July 10, 2014 at 7:25 am

          In reply to BB July 10, 2014 at 6:47 am

          Same problem. It might be a temporary technical issue so worth checking again in a few days.

          • Neuroskeptic (@Neuro_Skeptic) July 10, 2014 at 9:06 am

            I can’t access the images, and the PDF is also down. Can this be a coincidence? Maybe.

          • Neuroskeptic (@Neuro_Skeptic) July 10, 2014 at 9:13 am

            Ah, it’s affecting other papers as well. Not just this one.

          • BB July 10, 2014 at 4:06 pm

            It seems that it was only temporary. An additional observation: please compare band lane 2 of Fig3 with lane 6 of Fig4 (especially after horizontal flipping and enhancing the contrast in Fig3) and lane 1 of Fig3 with lane 7 of Fig4.

  • fernando pessoa May 22, 2015 at 12:59 pm
  • fernandopessoa April 26, 2016 at 4:17 am

    2015 correction.

    J Am Coll Surg. 2013 Jul;217(1):17-26; discussion 26. doi: 10.1016/j.jamcollsurg.2013.02.025. Epub 2013 Apr 23.
    Osteopontin up-regulates critical epithelial-mesenchymal transition transcription factors to induce an aggressive breast cancer phenotype.
    Li NY1, Weber CE, Mi Z, Wai PY, Cuevas BD, Kuo PC.
    Author information
    1Department of Surgery, Loyola University, Maywood, IL 60153, USA.
    Erratum in
    J Am Coll Surg. 2015 Aug;221(2):642.

    2015 correction notice.

  • Joey Joe Joe Shabadoo April 27, 2016 at 7:45 pm

    There are now 28 papers being discussed on pubpeer with Paul Kuo as an author, most as last author. Example:

  • Post a comment

    Threaded commenting powered by interconnect/it code.