Research integrity

The Legacy of Joan Massague

Cancer research giant Joan Massagué has no time to correct his old papers.

How come we haven’t cured cancer despite all the billions invested in cancer research? Despite all these Nature and Cell papers announcing the final breakthrough, with a simple cure finally found? Despite the finest minds of most venerated white men in elite US institutions working day and night to raise yet another multimillion research grant or biotech investment, to make cancer history?

Maybe our heroes are no heroes at all. Sure, there are some outright villains, like Carlo Croce, Pier Paolo Pandolfi, Michael Karin…. But the others? Whom can we trust?

For example, surely we can trust Joan Massagué, Spanish-born director of the Sloan Kettering Institute, part of Memorial Sloan Kettering Cancer Center (MSKCC) in New York? He is a titan of cancer research, credited with discovering the TGF-β signalling pathway and how to manipulate it to cure cancer.

But in science, it’s best never to have heroes. Also Massague’s science is tainted. In particular the older papers, back when he still was establishing his power.

In many cases, it’s stealth gel splicing. A practice which was retrospectively pardoned for old publications, not because that Franken-gel method was allegedly acceptable back then (as some claim), but because it is difficult to prove fraud without access to raw data, which is of course unavailable exactly because the papers are too old. And also because cancer research is so full of blatant fraud that nobody pays attention to spliced gels anymore. They should though.

But this trickery is why Massague published in all the big journals, reached the apex of the academic power hill, installed his loyal mentees in power positions all over the world, while your own manuscripts keep getting rejected for lacking novelty and impact, or worse, for contradicting the unassailable claims by elites like Massague.

Wei He, David C. Dorn , Hediye Erdjument-Bromage , Paul Tempst , Malcolm A.S. Moore , Joan Massagué Hematopoiesis controlled by distinct TIF1gamma and Smad4 branches of the TGFbeta pathway Cell (2006) doi: 10.1016/j.cell.2006.03.045

Why was it necessary to stitch a figure with just four lanes from who knows how many gels, why was it necessary to overexpose the images so no background and no traces of possible further gel splicing are visible? What did the raw data originally show, was the experiment even performed as described? How can this Franken-gel be used to claim that “The present findings may also have implications for pancreatic cancers and other tumors that suffer loss of Smad4 while retaining TGFβ receptors, Smad2/3, and TIF1γ“? Massague never replied to my emails.

Or this strange case:

In the Figure 6B, there is lane splicing, but also a strange object under the IgG bands of the Id2 blot, look closer:

If one scrutinises those old gels, they sometimes just don’t look natural.

Akiko Hata , Giorgio Lagna , Joan Massagué, Ali Hemmati-Brivanlou Smad6 inhibits BMP/Smad1 signaling by specifically competing with the Smad4 tumor suppressor Genes & Development (1998) doi: 10.1101/gad.12.2.186

Snarky PubPeer commenter: “I am sure the authors will get to work on this urgent matter of correcting an undisclosed splice in a 22 year old paper.
Something strange was done to the lower half of the aFlag blot. Is it a Franken-Gel? Or “just” digitally wiped clean?

Sometimes the seemingly carelessly innocent gel splicing leads to finding much more sinister things.

Joan Seoane , Celio Pouponnot , Peter Staller , Manuela Schader , Martin Eilers , Joan Massagué TGFbeta influences Myc, Miz-1 and Smad to control the CDK inhibitor p15INK4b Nature Cell Biology (2001) doi: 10.1038/35070086

What happened in the Anti-Flag image of the Figure 7b is a clear example of research misconduct. Look how bad the figure really is. I count at least 7 rectangles, a person who made this should be drummed out of science in shame. Or given a multimillion research grant and a huge lab with dozens of PhD students to train, I don’t know. Fact is that Massague made it to Sloan Kettering boss, and the first author Joan Seoane is now Director of the Translational Research Program at the Vall d’Hebron Institute of Oncology (VHIO) in Massague’s native Barcelona.

The penultimate author Martin Eilers, professor at the University of Würzburg, Germany, informed me:

The figure is from the Massague lab.”

A paper with such an outrageously fake figure should be retracted, but this is for Massague to decide. And this almighty god of US and Spanish cancer research decided to do exactly nothing at all.

Can we trust these two papers? It’s “just” irregular gel splicing, sure, but why was it necessary for a four- or five-lane gel? Was it because the results were not that impactful otherwise?

Qiaoran Xi , Wei He , Xiang H.-F. Zhang , Hong-Van Le , Joan Massagué Genome-wide impact of the BRG1 SWI/SNF chromatin remodeler on the transforming growth factor beta transcriptional program Journal of Biological Chemistry (2008) doi: 10.1074/jbc.m707479200

Irregular splicing renders the entire figure pointless. Because the readers and reviewers were mislead to believe they look at a properly controlled experiment, which it never was.

Here another paper by Massague, with another German collaborator, Christoph Niehrs, institute director at University of Mainz and German Cancer Research Centre (DKFZ).

Darya Onichtchouk , Ye-Guang Chen , Roland Dosch , Volker Gawantka , Hajo Delius , Joan Massagué, Christof Niehrs Silencing of TGF-beta signalling by the pseudoreceptor BAMBI Nature (1999) doi: 10.1038/46794 

Niehrs wrote back to me:

thank you for the valuable hint. I contacted the co-author responsible for the data from the then collaborating laboratory of Joan Massagué and we will place an erratum.

Niehrs and his co-author Ye Guang Chen then commented on PubPeer:

We note that the shown lanes with the arrows were indeed from separate gels and immunoblots that were spliced together later. The reason was that there were too many samples in that experiment to be loaded on one single gel, so they had to be split. Importantly though, all samples were from the very same experiment and are comparable. Nevertheless, we should have indicated this merging clearly and apologize for this omission. The omission does not affect the conclusion nor any statement of the publication.

I am not sure if this makes sense. The first two lanes from the Total HA blot and the first 4 lanes of the Total Flag blot clearly have a different background from the rest. They can never be from the same gel, i.e. experimental analysis. Most obviously these spliced-on lanes stem from some other gels, which renders the entire figure and all of its conclusions void. At best.

Will there be at least a correction now? Presumably Massague said no. Let’s have a look at even older papers.

I Reynisdóttir , K Polyak , A Iavarone , J Massagué Kip/Cip and Ink4 Cdk inhibitors cooperate to induce cell cycle arrest in response to TGF-beta Genes & Development (1995) doi: 10.1101/gad.9.15.1831

No Photoshop there, it’s the old artisan craft of paper-scissors-glue data fudgery which only the old generation of scientists learned to perform. The penultimate author Antonio Iavarone is professor at the Columbia University and has a PubPeer record of his own.

In another case of a very old Massague paper, the PubPeer criticism seemed not valid, like the alleged lane duplication, highlighted with blue boxes here:

Joaquı́n Arribas , Fernando López-Casillas , Joan Massagué Role of the juxtamembrane domains of the transforming growth factor-alpha precursor and the beta-amyloid precursor protein in regulated ectodomain shedding Journal of Biological Chemistry (1997) doi: 10.1074/jbc.272.27.17160

But a closer look revealed that the two “-PMA” “Media” blots completely lack the first lane, they are just cut-off there. The first author, Joaquin Arribas, now director of preclinical research at the aforementioned VHIO in Barcelona, did not reply when asked to explain.

Update 8.02.2022: Arribas now replied with an explanation:

“The lanes you refer to were not loaded. I did this experiment – 25 year ago! – as follows: cells were pulsed with radioactive amino acids for a short period of time. Then, labelling media was removed, cells were washed and fresh media was added. At the indicated times cells and conditioned media were harvested to analyse cell lysates and secreted material independently. Of course, there is no sample corresponding to media that has been conditioned 0 minutes by cells (as there is no samples corresponding to lysates or media of cells treated for 0 minutes with PMA). I did not add anything digitally, the blots just start there. Perhaps painting borders to openly show that there is no media conditioned for 0 minutes would have made clearer the presentation of the data.

But all those were small potatoes of Massague’s legacy. The great oncologist was also mentor to Stacy Blain, now associate professor at SUNY Downstate and biomedical entrepreneur, who went on to have a PubPeer record of her own. Here some examples of Blain’s and Massague’s joint contribution to cancer research:

Beverley J. Warner , Stacy W. Blain, Joan Seoane , Joan Massagué Myc Downregulation by Transforming Growth Factor β Required for Activation of the p15 Ink4b G 1 Arrest Pathway Molecular and Cellular Biology (1999) doi: 10.1128/mcb.19.9.5913

These papers are ancient, no journal will want to bother about those. But science like this formed the cancer research of today.

Stacy W. Blain, Ermelinda Montalvo , Joan Massagué Differential interaction of the cyclin-dependent kinase (Cdk) inhibitor p27Kip1 with cyclin A-Cdk2 and cyclin D2-Cdk4 Journal of Biological Chemistry (1997) doi: 10.1074/jbc.272.41.25863 

Fig 2C

For a paper with just two authors, it’s not difficult to find the culprits.

Stacy W. Blain, Joan Massagué Different sensitivity of the transforming growth factor-beta cell cycle arrest pathway to c-Myc and MDM-2 Journal of Biological Chemistry (2000) doi: 10.1074/jbc.m006496200

You have to look closer, the last p27 band was digitally inserted to cover up the result Drs Blain and Massague must have disapproved of.

Those were historical records, such easy to spot forgeries rarely happen in elite US labs nowadays. Not because they became honest, but because the methods have changed. The technology of the western blot, where every manipulation is evident if you bother to look, is becoming outdated or even embarrassing. But still, when western blot is used in newer papers, things like these happen:

Qing Chen , Adrienne Boire , Xin Jin , Manuel Valiente , Ekrem Emrah Er , Alejandro Lopez-Soto , Leni S. Jacob , Ruzeen Patwa , Hardik Shah , Ke Xu , Justin R. Cross , Joan Massagué Carcinoma–astrocyte gap junctions promote brain metastasis by cGAMP transfer Nature (2016) doi: 10.1038/nature18268 

The TBK panel is same in both figures, just at different brightness. Massague replied on PubPeer right away in October 2016 and announced a correction:

You are right that the TBK1 strip is the same in both figures, 4a and Ex9a. These are repeats of the same experiment done in two different cell lines. The equal loading of all the tracks is still clear based on the IRF3 and Actin data. Therefore, the validity of the results fortunately is not compromised by the error.

The Nature correction from 2017 declared that there was even more:

“In this Article, Extended Data Figs 7d and 9a, presenting results obtained with the H2030-BrM3 cell line, contain errors that were introduced during the assembly of these figures in parallel with the corresponding panels in Figs 3b and 4a, which show the results of the same experiment but using a different cell line (MDA231-BrM2). The p65 loading control from Fig. 3b and the TBK1 loading control from Fig. 4a were duplicated in Extended Data Figs 7d and 9a, respectively. The STAT1 strip in Extended Data Fig. 7d was horizontally reversed. T[…] These errors do not alter the results or conclusions of the Article.”

Of course none of this can ever alter the results or conclusions of the Article, because articles like this are just conveyor-belt bullshit without any meaning for actual cancer therapies. Sure, the study claimed to have discovered a cure for brain cancer:

The orally bioavailable modulators of gap junctions meclofenamate and tonabersat break this paracrine loop, and we provide proof-of-principle that these drugs could be used to treat established brain metastasis.

A clinical trial with these drugs with brain tumour patients was registered by MSKCC in 2015 and never started. Probably because the results and conclusions of Massague’s Nature paper are that reliable.

Original photo: Gerstner Sloan Kettering

If these preclinical discoveries do reach the clinical phase, they reliably fail in trials, but not without first having earned their professorial inventors hundreds of thousands or even millions of dollars. In this regard, Massague and his colleagues declared in that Nature paper “no competing financial interests“. In reality he holds several patents for treatment of brain tumours with meclofenamate and tonabersat. A bit reminiscent of how MSKCC president and another Spaniard, Jose Baselga, was dismissed in 2019 for hidden multi-million dollar-heavy conflicts of interests (actually, hiding financial COIs is standard practice in biomedical research). Money can’t buy everything though: Baselga died in 2021 from a neurodegenerative disease aged just 61, and then it became impolite to even mention his earlier sacking.

Here an even newer paper from the Massague lab, it’s all diagrams now, no more treacherous western blots.

Karuna Ganesh , Harihar Basnet , Yasemin Kaygusuz , Ashley M. Laughney , Lan He , Roshan Sharma , Kevin P. O’Rourke , Vincent P. Reuter , Yun-Han Huang , Mesruh Turkekul , Ekrem Emrah Er , Ignas Masilionis , Katia Manova-Todorova , Martin R. Weiser , Leonard B. Saltz , Julio Garcia-Aguilar , Richard Koche , Scott W. Lowe , Dana Pe’er , Jinru Shia , Joan Massagué L1CAM defines the regenerative origin of metastasis-initiating cells in colorectal cancer Nature Cancer (2020) doi: 10.1038/s43018-019-0006-x

I wonder if in “Fig. 1 […] panel g, the data points for sample “190P-Primary tumor” and “41Li-Liver metastasis” were inadvertently duplicated?

Massague replied on PubPeer right away:

We regret to have missed this accidental duplication in the course of putting together the figure. Remarkably, our error was there from the outset, yet the authors, the referees, and the editors all missed it through two rounds of review! We are thankful that the reader caught this error at last.

The reviewers and editors missed it because papers by someone like Massague are peer-reviewed differently than papers from pedestrian scientists. A correction was issued a few months later:

Correction September 2020: “In the version of this article initially published, the values for CRC190P (159, 165, 160, 40, 22) were incorrect in the Source Data file for Fig. 1g (a duplication of the values for CRC41Li), which resulted in an error in the plot in Fig. 1g. Those values are now correct (103, 118, 98, 20, 36) in the Source Data, and the correct Fig. 1g is now provided.

There may be more in Massague’s papers if one bothers to search. But what’s the point? None of that affects his conclusions.

PS: because a fish always rots from its head, at MSKCC even plagiarism is acceptable.

Contact form


I thank all my donors for supporting my journalism. You can be one of them!
Make a one-time donation:

I thank all my donors for supporting my journalism. You can be one of them!
Make a monthly donation:

Choose an amount


Or enter a custom amount

Your contribution is appreciated.

Your contribution is appreciated.

DonateDonate monthly

15 comments on “The Legacy of Joan Massague

  1. Like many others, I would say 80%+, of the scientists in renowned institutions across the world they not only not have time to correct old and new papers but also actively engage in relationships with governments and high impact journals. And therefore are unable to answer to very simply questions like if HIV-1 or Sars-Cov-2 were isolated for example.


  2. ” A bit reminiscent of how MSKCC president and another Spaniard, Jose Baselga, was dismissed in 2019 for hidden multi-million dollar-heavy conflicts of interests (actually, hiding financial COIs is standard practice in biomedical research).”

    Problematic data to boot:-


  3. ” A bit reminiscent of how MSKCC president and another Spaniard, Jose Baselga, was dismissed in 2019 for hidden multi-million dollar-heavy conflicts of interests (actually, hiding financial COIs is standard practice in biomedical research).”

    Not good at arithmetic.


  4. “For example, surely we can trust Joan Massagué, Spanish-born director of the Sloan Kettering Institute, part of Memorial Sloan Kettering Cancer Center (MSKCC) in New York?”

    Cancer Res. 2005 Mar 15;65(6):2186-92.

    p73β-Mediated Apoptosis Requires p57kip2 Induction and IEX-1 Inhibition

    Susana Gonzalez, Manuel M. Perez-Perez, Eva Hernando, Manuel Serrano and Carlos Cordon-Cardo
    DOI: 10.1158/0008-5472.CAN-04-3047 Published March 2005

    Susana Gonzalez1,2, Manuel M. Perez-Perez1, Eva Hernando1, Manuel Serrano2, and Carlos Cordon-Cardo1

    1Department of Pathology, Memorial Sloan-Kettering Cancer Center, New York, New York and

    2Molecular Oncology Program, Spanish National Cancer Centre, Madrid, Spain

    Requests for reprints:
    Carlos Cordon-Cardo, Department of Pathology, Memorial Sloan-Kettering Cancer Center, New York, NY. Phone: 212-639-7746; Fax: 212-794-3186; E-mail:

    What a coincidence!


    When all the elm trees in the elm grove have Dutch elm disease, yet this one doesn’t?

    Carlos Cordon-Cardo has notched up more than a few:


  5. Joan Massage/David M Sabatini
    Nobel Price candidate/Nobel Price candidate
    Are there more parallelisms related to female scientists?


  6. alfricabos

    One of my all time favorite fake cancer research paper:


  7. Regarding the final paper mentioned, Ganesh et al. (2020), doi: 10.1038/s43018-019-0006-x: such mistakes when putting figures together do in fact happen, and I can belief this to be an accident. I’ve made similar mistakes myself. Fortunately, so far, I always caught them before the final publication of a paper. But it has happened that I found mistakes like this during proof correction, after more than one round of reviews.

    I think such mistakes are easier to make when creating diagrams like this compared to something like western blots or photographs. In the creation of a diagram figure, there is always some assembly to be done. You’re selecting which columns in your file have the labels, which ones have the values, which ones the error values, etc. Occasionally you might select column 237 again instead of 239…

    Photographic figures, including western blots, should normally be presented close to how they were created, with next to no work done to the figure. So, in theory, there should be less chances of errors there.


  8. Surely we can also trust big scientist Eric Lander.
    According to Science, White House science adviser Eric Lander resigns after bullying investigation.
    Another staffer not quoted by name told Politico that Lander has a “Jekyll and Hyde personality,” pleasant in public meetings but harsh behind closed doors.
    Is this bullying investigation limited to White House? Eric Lander is the president and founder director of the Broad Institute (MIT, Harvard).
    Broad Institute, Whitehead Institute (Sabatini), … what’s happening at MIT?


  9. alfricabos

    Exactly right about the culture clash. Few people in academia, especially low paid research staff on visa, would dare complaining about the boss being demeaning (or worst). Business as usual, even if things are starting to change, very slowly.


Leave a Reply

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

You are commenting using your 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 )

Connecting to %s

%d bloggers like this: