About these ads

Retraction Watch

Tracking retractions as a window into the scientific process

Harvard and the Brigham investigating leading heart group for “compromised” data

with 36 comments

circulationcoverCirculation has retracted a 2012 study by a group of Harvard heart specialists over concerns of corrupt data, and the university is investigating. The group was led by Piero Anversa, a leading cardiologist, and Joseph Loscalzo — who will be familiar to readers of Circulation as the editor in chief of that journal. (Anversa’s also on the editorial board).

The paper was titled “Cardiomyogenesis in the aging and failing human heart,”  and has been cited 30 times, according to Thomson Scientific’s Web of Knowledge. Here’s the abstract:

BACKGROUND:

Two opposite views of cardiac growth are currently held; one views the heart as a static organ characterized by a large number of cardiomyocytes that are present at birth and live as long as the organism, and the other views the heart a highly plastic organ in which the myocyte compartment is restored several times during the course of life.

METHODS AND RESULTS:

The average age of cardiomyocytes, vascular endothelial cells (ECs), and fibroblasts and their turnover rates were measured by retrospective (14)C birth dating of cells in 19 normal hearts 2 to 78 years of age and in 17 explanted failing hearts 22 to 70 years of age. We report that the human heart is characterized by a significant turnover of ventricular myocytes, ECs, and fibroblasts, physiologically and pathologically. Myocyte, EC, and fibroblast renewal is very high shortly after birth, decreases during postnatal maturation, remains relatively constant in the adult organ, and increases dramatically with age. From 20 to 78 years of age, the adult human heart entirely replaces its myocyte, EC, and fibroblast compartment ≈8, ≈6, and ≈8 times, respectively. Myocyte, EC, and fibroblast regeneration is further enhanced with chronic heart failure.

CONCLUSIONS:

The human heart is a highly dynamic organ that retains a remarkable degree of plasticity throughout life and in the presence of chronic heart failure. However, the ability to regenerate cardiomyocytes, vascular ECs, and fibroblasts cannot prevent the manifestations of myocardial aging or oppose the negative effects of ischemic and idiopathic dilated cardiomyopathy.

According to the retraction statement:

For the article by Jan Kajstura, Marcello Rota, Donato Cappetta, Barbara Ogórek, Christian Arranto, Yingnan Bai, João Ferreira Martins, Sergio Signore, Fumihiro Sanada, Alex Matsuda, James Kostyla, Maria-Virginia Caballero, Claudia Fiorini, David A. D’Alessandro, Robert E. Michler, Federica del Monte, Toru Hosoda, Mark A. Perrella, Annarosa Leri, Bruce A. Buchholz, Joseph Loscalzo, and Piero Anversa (Cardiomyogenesis in the aging and failing human heart. Circulation. 2012;126:18691881; DOI: 10.1161/CIRCULATIONAHA.112.118380), an ongoing institutional review by Harvard Medical School and Brigham and Women’s Hospital has determined that the data are sufficiently compromised that a retraction is warranted.

The American Heart Association, therefore, retracts the article.

We’ve emailed Anversa for comment and will add to this post as we learn more.

One thing to note: Harvard probably isn’t the only institution looking at this matter. Per the paper:

This work was supported by NIH grants and by grant NCRR RR13461. This work was performed in part under the auspices of the U.S. Department of Energy by Lawrence Livermore National Laboratory under Contract DE-AC52-07NA27344.

Please see an update on this post.

Like Retraction Watch? Consider supporting our growth. You can also follow us on Twitter, like us on Facebook, add us to your RSS reader, and sign up on our homepage for an email every time there’s a new post.

About these ads

36 Responses

Subscribe to comments with RSS.

    • Is the Brigham/Harvard list at PubPeer still in need of some length enhancement?

      Proc Natl Acad Sci U S A. 2008 Oct 7;105(40) Notch1 regulates the fate of cardiac progenitor cells.
      Boni A1, Urbanek K, Nascimbene A, Hosoda T, Zheng H, Delucchi F, Amano K, Gonzalez A, Vitale S, Ojaimi C, Rizzi R, Bolli R, Yutzey KE, Rota M, Kajstura J, Anversa P, Leri A.

      Maximise the image for figure 3

      http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2563067/figure/F3/

      Observe the prolific lane splicing and control band reuse in the individual ChIP gels in parts I and J.

      Scrutineer

      April 10, 2014 at 3:26 am

      • Nice catch!

        I’d even go so far as to say that in panel I, the right lane (Ig) in the Nkx2.5 panel bears an uncanny resemblance to the Ig lane in the adjacent Hes1 panel. In addition the marker ladder is identical between all 3 panels in panel I. Same thing in Panel J – ladders are identical between the 3 images, and the Ig lanes are identical too (the latter appear to originate from the one on the right which is not spliced).

        So, this is not simple splicing to skip unwanted lanes, but duplication of information between panels meant to represent different experiments. As per JCB guidelines, any image manipulation that “changes the scientific content of an image” is to be avoided.

        Paul Brookes

        April 10, 2014 at 8:52 am

        • One other curiosity and I am sure it is no more than harmless boasting; but the given size label is 200 kb. Feeling a tad inadequate, I had to google a bunch of images of ChIP analysis to reassure myself that 200 bp was probably meant, rather than 200,000 bp. A 200 kb PCR band? That would be awesome!

          Scrutineer

          April 10, 2014 at 3:36 pm

      • Proc Natl Acad Sci U S A. 2008 Oct 7;105(40) at Pubpeer:

        https://pubpeer.com/publications/627750E871B794D17ECE68B6DBCAAB#fb7567

        david hardman

        April 11, 2014 at 6:59 pm

        • Perhaps this blot-heavy article would also add value to the recently enhanced PubPeer list?

          Numb regulates glioma stem cell fate and growth by altering epidermal growth factor receptor and Skp1-Cullin-F-box ubiquitin ligase activity.
          Jiang X, Xing H, Kim TM, Jung Y, Huang W, Yang HW, Song S, Park PJ, Carroll RS, Johnson MD.
          Stem Cells. 2012 Jul;30(7):1313-26. doi: 10.1002/stem.1120.
          PMID: 22553175

          Figure 5 E Left side input 3 lanes convert into right side input lanes with horizontal inversion and vertical stretching.

          http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3963835/figure/F5/

          Figure 6 Actin controls in E and F are the same but the experiments are different.

          http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3963835/figure/F6/

          Scrutineer

          April 12, 2014 at 5:44 pm

  1. Wow, and I always thought that Anversa was one of the “untouchables”. I heard his lab had been raided by Harvard, but I didn’t get any details and it has been months since that rumor was going around. I guess this is the end result. A lot of people for a very long time have seriously doubted the quality of his data.

    John

    April 9, 2014 at 3:51 am

  2. This paper was a big deal in the field. Along with work from Richard Lee (also at the Brigham), it forwarded the concept that the heart is not a terminally differentiated organ, and that cardiomyocytes undergo constant renewal in the adult. The whole thing is predicated on there being a spike of 14C in the atmosphere during nuclear testing era (hence the involvement of Lawrence Livermore lab), with the decay since then allowing estimation of a cell’s “actual” age versus the chronological age of the donor.

    It’s intriguing if you look at an earlier paper from the same authors (Kajstura & Anversa) with an almost identical title (PMID 21088285). In Figure 6 of that paper, cells are labeled for phospho-histone H3 (blue) and propidium iodide (green). In the co-labeling you can see that blue and green co-localize in the nucleus but they don’t overlap a whole lot. There’s distinct green vs. blue staining in the nucleus undergoing mitosis. Now look at the similar Figure 7A of the retracted paper. They swapped the colors, so now nuclei are labeled blue and the phospho-H3 (mitosing cells) is labeled green. The problem is, the green (inset) is an absolute carbon copy of the blue. It’s almost as if they just took the blue image and re-colored it green. If this was a real mitosing cell, surely the green/blue (phosphoH3/PI) should stain like in their earlier 2010 paper? Looky… I even done gone made a imgur for y’all… http://imgur.com/5EuzeCA

    So yeah, I’d agree with their assessment that the data are “unreliable”.

    Also, am I the only one who thinks it’s kinda weird that two groups working at the same institute (Anversa and Lee) would publish rather similar findings 3 months apart in different journals with no overlapping authors whatsoever? (Lee published his findings in PMID 23222518). Did these guys really not know what was going on down the hall? The other question therefore, is whether the data in Lee’s paper are now also bought into doubt by this retraction?

    Paul Brookes

    April 9, 2014 at 4:39 pm

    • To your questions in the last paragraph, as I was personally involved in some of this nonsense.
      1. Did these guys really not know what was going on down the hall?
      Yes and no. Everyone in the field knew that the Anversa work didn’t pass the smell test. Lee and Anversa groups had no interaction at the Brigham and were at odds with each other over the findings of their respective papers.
      2. Is Lee’s paper now brought into doubt?
      Not anymore than it was before. Honestly, if you read both papers, Lee’s group did much more rigorous analysis.

      N

      April 9, 2014 at 5:15 pm

      • Yes, many people have disassociated themselves from Anversa’s group for this very reason. Lee is an excellent scientist. I could see the two groups having an invisible wall between them.

        John

        April 14, 2014 at 6:37 am

    • Analysing figure A on the left. Observe the red background, bright red spots at 5 o’clock in the boxed blue/green image and adjacent to the green arrowed nucleus above it – they are identical. There is no overlap, as if there were we would see a non-blue and non-green colouring, depending on the intensity of each pixel – kinda turqoise – which we don’t see.

      Similarly, on the right panel, boxed green nucleus and arrowed blue nucleus – contrast and compare the red worm-like structure beneath both images. They are identical.

      Therefore the background of each image is identical, suggesting that if the authors were requested to send the orignal datasets – it would be impossible to do so, as there is clearly only one datset, coloured differently.

      Stewart

      April 9, 2014 at 5:25 pm

    • Hi,
      could you please provide me the full reference of the “figure 6A” you have mentioned above? I saw your uploaded image on imgur and I’ve downloaded the paper “Myocyte Turnover in the Aging Human Heart, Circ Res. 2010;107:1374-1386″ but I could not find the mentioned figure in that paper…

      Cheers!

      Moshka Baskitt

      April 9, 2014 at 9:45 pm

      • Ach! My Bad. Sorry, I mixed up the PMIDs (you’ll see why in a minute). The Correct PMID is 20522802.

        There are 2 similar papers, both in the same volume of the journal from Anversa/Kajstura.

        One is PMID 20522802, Circ Res. 2010 Jul 23;107(2):305-15. “Cardiomyogenesis in the adult human heart.” This is the one with the Figure 6A that should be contrasted with the retracted paper.

        Then there’s PMID 21088285 Circ Res. 2010 Nov 26;107(11):1374-86. “Myocyte turnover in the aging human heart.” It appears to use many of the same techniques. 7 authors are the same between both, and 6 of those appear on the 2012 retracted paper.

        It would be helpful if the journal could provide some further information on exactly why “the data are sufficiently compromised”. If it was due to a specific mistake in just this one paper then the others are fine. If it was due to an underlying problem with the method itself then it might suggest these other papers from the same group using the same methods should also be questioned.

        Paul Brookes

        April 10, 2014 at 9:20 am

        • Thanks a lot for clarifying it!

          In this case I will consider the PMID switch a genuine mistake instead of misconduct, intentional alteration or fraud… ;)

          Cheers!

          Moshka Baskitt

          April 11, 2014 at 1:35 am

    • Even Anversa and Lee are listed in the same institute (BWH), their lab are not in the building. In fact, they are not even in the same city (one is Boston and one is Cambridge).

      Ruei-Zeng

      May 10, 2014 at 10:57 pm

  3. Piero Anversa does research cardiology, but he is NOT a cardiologist but a pathologist.

    Watcher

    April 10, 2014 at 8:57 am

    • Does that matter? He is doing “research” in the area of cardiology, and pathological cardiology in particular. You don’t have to listen to hearts.

      fernando pessoa

      April 10, 2014 at 4:05 pm

      • It doesn’t except that a cardiologist would know that cardiac myocytes don’t regenerate or divide

        Watcher

        April 10, 2014 at 4:18 pm

  4. Free Radic Biol Med. 2010 Oct 15;49(7):1238-53. doi: 10.1016/j.freeradbiomed.2010.07.020. Epub 2010 Aug 1.
    Cardiac overexpression of insulin-like growth factor 1 attenuates chronic alcohol intake-induced myocardial contractile dysfunction but not hypertrophy: Roles of Akt, mTOR, GSK3beta, and PTEN.
    Zhang B, Turdi S, Li Q, Lopez FL, Eason AR, Anversa P, Ren J.

    https://pubpeer.com/publications/314063891818854B9524964F4595B1#fb7655

    david hardman

    April 14, 2014 at 10:02 am

    • Given the rate that Harvard has been appearing in Retraction Watch over the last few days, do you think they are at last doing some “spring cleaning”?

      Anyway, regarding this newly interesting 2010 paper, there might be a bit more to figures 9 and 10 than is so far discussed on the PubPeer link?

      For in the legend, it say that A to C and B to D are intimately related, the lower item being the phospho-version of the upper. That is perhaps why the loading controls are the same since the usual procedure would be to reprobe the blot to get the phospho-version; just slightly shifted on the gel so an almost identical band curvature. A cursory glance at these figures might convince you that, sadly, the P-proteins are on different gels to the non-P-proteins. And normally that would mean that the loading controls should be too. I expect you will want to see for yourself:

      http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2947341/figure/F9/

      http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2947341/figure/F10/

      Scrutineer

      April 16, 2014 at 3:25 pm

    • I believe in this case, although the involvement of Anversa is a coincidence, but in fact closer scrutiny is required of the senior author (Jun Ren). The following were reported to ORI a while ago:

      PMID 01907315 Fig. 4B blot splicing.
      PMID 21535395 Fig. 3B = PMID20110268 Fig. 2E blots flipped.
      PMID 20477759 Fig. 8 re-used of GAPDH control blot, same thing in Fig. 8D vs. 10C, again in 10A vs. 10D.
      PMID 19942091 Fig. 5A/B loading control re-use, shifted by 1 lane.
      PMID 20445824 Fig. 1A vs. 3A, loading control re-use.

      Birds of a feather, etc.

      vhedwig

      April 16, 2014 at 4:50 pm

      • For the 2010 FRBM paper in question, we did run the same sample for pan and phosphorylated proteins on separate gels while running the loading control on another gel. The loading volume was identical for all 3 gels. This has been a “combersome practice” that got passed around by graduate students here that I did not enforce to change. It can be seen in a number of our publications. Thanks for the insightful comments – we should modify this procedure. Jun Ren, University of Wyoming

        Jun Ren

        April 22, 2014 at 12:40 am

  5. Another group of Harvard University published a research paper in Nature in November 2013(Nature 503: 481, 2013). They claimed a cell fate switching in a bacterium and their “discovery” was heralded as “Systems biology: How bacteria choose a lifestyle” (Nature 503: 476, 2013).
    However, their so-called “discovery” was based on what I think is incorrect composition and compilation of data. This is because their so-called “cell fate switching” during their so-called “lifespan” of so-called “single bacterium” might well be an illusion resulting from combining life segments from multiple bacteria into a so-called “single bacterium”. This suspicion was reinforced in one part with my attention to authors’ description on their unique algorithm to “automatically connect the lineage to the dead cell’s closet relative. In all cases, this is the sister cell of the former mother”.

    In order to confirm this suspicion and better understand their research I sent my initial inquiry to all four authors including two corresponding authors on December 20th, 2013. I also asked them to send me some key data underlying their published figures. However my inquiry was not replied by any of them. My follow-ups on December 24th, 2013 and January 3rd, 2014 were also ignored. I should point out that my January 3rd email was copied to the Editor-in-Chief of Nature in a hope that the journal will ask the authors to answer scientific inquiry and provide key data for validation.
    Because I couldn’t get any response from the authors I submitted my Communications Arising directly to Nature on Jan. 20th 2014. But Nature rejected my submission on Jan. 23rd, 2014 without any peer review. My protest to Nature’s rejection was not replied. Then a modified version of the Nature-rejected manuscript was submitted to Science on Jan. 27th, 2014 as a response to Science’ Editorial calling for “demonstrating excellence in transparency and instilling confidence in results”. But this manuscript entitled “Transparent Discussion on a Bacterial Cell Fate Switching Study” was rejected by Science, too.

    Following Science’s suggestion “Although your analysis is interesting, we feel that the scope and focus of your paper make it more appropriate for a more specialized journal”, the Science-rejected manuscript was published on Feb. 3rd, 2014 in Logical Biology, a journal I founded in 2000. The complete version of this paper can be accessed free of charge at http://im1.biz/albums/userpics/10001/LB2014V13N1A2_CellFate.pdf. Any criticism on this publication will be published in Logical Biology if the author(s) wish to do that.

    Because I still need the critical data underlying the published work in Nature from this Harvard University research group I emailed Harvard University Faculty of Arts and Sciences and received a response from Dr. Gearoid Griffin, the Research Integrity Officer for the Faculty of Arts and Sciences at Harvard University. Per his suggestion I filed “A Complaint against Potential Research Misconduct by Harvard University Faculties and Students”.
    In this complaint I expressed the following two wishes:
    1. By presenting my PUBLISHED criticism to their PUBLICATION through some authoritative means, it may lead to a transparent discussion on some controversies in science.
    2. By issuing a request from some authoritative parties for obtaining custody of all research records and evidence, I may have opportunities to perform more analyses on their research and thus resolve some lingering concerns over their discoveries.

    My complaints submitted to the Harvard University Faculty of Arts and Sciences was also copied to the Office of Research Integrity (ORI) and later sent to Nature.

    So far, the Harvard University has not made any real progress on the case. The most recent response received from Dr. Griffin on April 2nd, 2014 states that “because the allegations involve a faculty member from the medical school, I am not in a position to move forward until I have discussed the matter with them. I have referred the matter to the medical school and I am currently waiting for a response”.

    So if the Harvard University Medical School does not pay attention to this matter then the Harvard University Faculty of Arts and Sciences will unable move this case forward. Then will the whole case become dead?

    Shi V. Liu

    April 17, 2014 at 1:02 pm

    • You publish your “paper” in your journal that you founded… that’s the least of your problems from your little rant.

      George Comfy

      October 1, 2014 at 1:00 am

  6. Piero Anversa’s wikipedia entry.

    http://it.wikipedia.org/wiki/Piero_Anversa

    david hardman

    April 9, 2014 at 6:17 am

  7. From Wikipedia: “Piero Anversa…considerato uno dei”

    I don’t speak Italian, but I did Latin for a year at school, so when I read this, I thought it meant “Piero Anversa… is considered a god”.

  8. That is an entertaining translation, but in that case “dei” translates in English as “of”.

    Z

    April 11, 2014 at 9:14 am

  9. Nope, it does not mean that. My husband speaks Italian (and French, and Spanish, and German, and Russian… :) . He says that “dei” means “gods”, but more frequently it means “of”. In this case, you chopped the sentence after “of” (dei). The sentence would have continued with something like “considered one of the top blah blah” . Nothing to do with god in this context…

    Wendy Grossman

    April 11, 2014 at 6:37 pm


We welcome comments. Please read our comments policy at http://retractionwatch.wordpress.com/the-retraction-watch-faq/ and leave your comment below.

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

Follow

Get every new post delivered to your Inbox.

Join 35,939 other followers

%d bloggers like this: