Sweden is a tolerant country, which is a very good thing. Unfortunately, sometimes this Swedish tolerance seems ill-advised. Dishonest scientists caught faking data are happily given another chance and fat funding, like the case of the diabetes researcher Pontus Boström shows.
This scientist was found to have fabricated data during his PhD studies with late Sven‐Olof Olofsson at the University of Gothenburg, and went afterwards to publish a seminal paper in Nature with the biggest godfather of the diabetes research field, Bruce Spiegelman. Also this high-impact study turned out to be irreproducible by other researchers and a likely artefact of erroneous antibody use. Yet due to his impressive publishing record and unwavering support of the mighty Spiegelman, Boström was invited to head a group leader position at several Swedish universities, while he settled on the best offer by the Department of Cell and Molecular Biology at the prestigious Karolinska Institutet (KI) in Stockholm, supported by the elite EU funder ERC. All despite his previous convictions of research misconduct in Gothenburg and ensuing retractions of two meeting abstracts, which were at all times perfectly known to all parties involved.
UPDATE 7.07.2016: just as my article went online, KI chose to finally reply to my email. Tomas Ahlbeck, KI Press secretary declared:
“About Pontus Boström:
The research group was discontinued during the autumn 2015. Mr Boström is since 31 Dec 2015 not an employee at KI. His postdocs are not at KI anymore, they continue their work at other Universities abroad”.
The ERC press team informs me that Boström’s grant funding was stopped:
“Dr Pontus Boström’s grant agreement (Novel_myokine) was terminated on 31 December 2015”.
Boström was obviously sacked by KI and lost his ERC funding, please take it into account while reading this article.
The 2012 Gothenburg investigation found that Boström manipulated excel sheets for a manuscript, which publication was subsequently prevented. Now, new evidence on PubPeer suggests serious image integrity problems in his publications, including his and Spiegelman’s papers in Nature and Cell. Here are two examples of apparently duplicated western blot bands and image re-use in papers from Olafsson’s Gothenburg lab, which Boström signed as first author in Diabetes, 2010 and Nature Cell Biology 2007, respectively.
Boström’s case has its entertainment value, because it is similar to that of Suchitra Sumitran-Holgersson, but in reverse. Basically, KI and University of Gothenburg traded cheaters. Sumitran-Holgersson was found guilty of deliberate data manipulation while working at the KI and had to retract a paper, yet she was soon acquitted under most bizarre circumstances (see my detailed report here). As result, she was allowed to retain all her funding and her professorship at the Sahlgrenska University Hospital of the University of Gothenburg. Meanwhile, Sumitran-Holgersson and her colleagues are being investigated for faking ethics approvals in order to perform questionable experiments in regenerative medicine on human patients, at least one of whom has died (details in my follow-up report here).
I obtained key original documents of the Gothenburg investigation. First, I present a note submitted by Boström’s former Gothenburg colleague Jan Borén, chair of the Department of Molecular and Clinical Medicine at the Sahlgrenska University Hospital, to their Vice-Chancellor (original here):
“Professor Sven‐Olof Olofsson, from the Institute of Medicine, died unexpectedly on December 28, 2011, leaving his research group without a principal investigator. To support the research group, the undersigned participated in their group meetings and project meetings. This was an obvious course of action because Sven‐Olof and the undersigned had discussed the projects on a frequent basis.
The research group decided to complete a project that focuses on syntaxin‐5 and its role in fatty acid‐induced insulin resistance, a project that they had worked on for several years. The project had been inactive for about 6 months but there was a manuscript, Decreased syntaxin‐5 in skeletal muscle in type 2 diabetes impairs insulin action on AKT by inhibiting its sorting to the plasma membrane, which had been submitted to the journal Diabetes, and then rejected on 19 July, 2011 (Appendix 1). The research group, together with their collaborator Kurt Højlund at Aarhus University, Denmark, decided to revise the work and submit it to another journal (Appendix 1). We therefore designed different control experiments, but these did not give the expected results. When this situation occurs, there are in principle three different explanations:
− The control experiments are incorrectly designed or executed
− The interpretation of the original data is incorrect
− The original data are incorrect
We went through these possibilities and excluded the first two. The undersigned therefore asked Linda Andersson [who took the lead of the lab after Olofsson’s death, -LS] in the research group to carefully review the original data. Linda checked the original data and reported to the undersigned on April 24 that the values she found were not consistent with those found in the manuscript (Appendix 2). She then checked additional files and found the same phenomenon, which she reported on April 26 (Appendix 3). The original data were calculated by Pontus Boström who was a PhD student in the research group and defended his thesis on 21 December, 2007. He then continued to work in the research group in parallel with his medical training, until he moved to the US in 2009 to undertake postdoctoral training at the Dana‐Farber Cancer Institute, Harvard Medical School [with Bruce Spiegelman, funded for 2 years by the Swedish Heart-Lung Foundation -LS]. Pontus Boström has been working at the Karolinska Institute in Stockholm since spring 2012.
The undersigned therefore contacted Pontus Boström (Appendices 4‐8) and asked him to explain why:
- The raw data have been altered without an explanation of why this has happened.
This is notable since the altered values give significant differences between the groups which were not different in the first analysis
- Western gels or qPCR files for the new data are missing
- In the altered values, only the digits with the highest value differ (illustrated in the table below). […]
When the undersigned failed to clarify these ambiguities, the Vice‐chancellor of Gothenburg University, Pam Fredman, and the Head of the Institute of Medicine, Hans Carlsten, were informed on August 16, in accordance with the Handling procedure at Gothenburg University for cases of suspected research misconduct. The Vice‐Chancellor, Pam Fredman, then contacted the university lawyer Kristina Ullgren, who on August 30 asked the undersigned to submit this summary.
The undersigned has also informed Sven‐Olof Olofsson’s collaborators and research leaders at other universities. The research group has also checked the raw data for published work, but has not found similar ambiguities”.
I contacted Borén about the background of the Boström misconduct investigation and received this comment:
“These issues are difficult but I think it’s critical that we handle them transparent and don’t try to hide scientific misconduct. We must learn from mistakes. I learned this from my former PhD supervisor, late Professor Sven-Olof Olofsson. Science integrity and honesty was critical for him, and I’m glad that the investigation fully cleared him from any suspicions of scientific misconduct.”
In early 2014, Boström was investigated and found guilty of “conscious preparation and presentation of falsified results”, as evidenced by the full report of the investigative committee and the Vice-Chancellor’s decision which I make available exclusively (both in Swedish). There is also an English-language report of the external investigator Ugo Moens, professor for medical biology at the University of Tromsø in Norway. Moens analysed Boström’s excel files and found out that “it is beyond doubt that primary data have been changed when composing the xls.files”. He also determined that Boström manipulated data to create a significant difference (100% vs 74,5%) between two experimental groups which were actually same (100% vs 98,3%). The rigged data was then used by Boström to present a phony result in a manuscript subsequently submitted to two journals, as mentioned by Borén above.
These were the translated “Conclusions from the Committee on Academic Misconduct” of the 1.5 year long Boström investigation, official number Dnr D 2012/26 (highlights mine):
“The Committee on Academic Misconduct has taken into account that several people were involved in the research project in question during the time period that the inquiry relates to. The investigation has received answers to questions regarding the extent to which there are modifications to the raw data and to what extent these modified data formed the basis of scientific manuscripts. Against this background, the Committee on Academic Misconduct then assessed to what extent there had been a deliberate alteration of data with the aim of obtaining a favorable outcome to the research project.
The Committee on Academic Misconduct could not come to any conclusion other than that alteration of data has occurred in this case. There is nothing to suggest anything other than that the modified data, through having been changed in a certain way, shows a finding that makes the modified data more scientifically interesting. This finding would not have been shown if the data had not been modified in the way that happened.
The modified data can be traced to Pontus Almer Boström, because they were first found in Pontus Almer Boström’s folder on the Wallenberg Laboratory’s file server and in the form of the diagrams attached to two emails that he sent to Sven-Olof Olofsson.
The claim that the modified data originates from any other person has not been shown to be credible.
The Committee on Academic Misconduct finds that the data have been modified in a systematic manner, which contradicts Pontus Almer Boström’s statement that he did not knowingly corrupt or modify any data.
The Committee on Academic Misconduct finds that Pontus Almer Boström is guilty of misdemeanor with reference to point 7a) “conscious fabrication, corruption or suppression of basic material” and point 7b) “conscious preparation and presentation of falsified results” and therefore finds reason to believe that scientific misconduct has occurred.
The Committee on Academic Misconduct’s opinion is unanimous”.
Boström had to retract two conference abstracts: Andersson et al, Atherosclerosis 2011 and Højlund et al, Diabetologia 2011. He had to fully admit having faked his data, yet towards RetractionWatch, Bostöm attacked his critics by speaking of “an old conflict”, “a very biased investigation” and “science politics”. He was then supported by his former mentor Spiegelman, who declared to be “confident” in Boström’s work in his lab. But who else is confident in that 2012 Nature paper?
Together with Boström, Spiegelman has discovered a miracle cure for obesity: the elusive hormone Irisin, a soluble version of a cell membrane-bound receptor FNDC5. Irisin is allegedly a produced by exercising muscle and turns white adipose tissue into brown one, which then literally burns energy. Basically, a fantasy come true for the pharma industry: pop an irisin pill and you lose weight and straighten your muscles while watching TV, and you heat your room on top. Everything was set to tap the huge and bloated obesity market in US: irisin was patented by Boström and Spiegelman, then licenced to a pharma company and scheduled to go from 2014 on for clinical treatment of obesity and type 2 diabetes. If you have not heard of this miracle cure yet, it is because very soon doubts arose about what these two scientists claimed to have discovered.
The scientific reliability of their Nature paper was soon vigorously debated on PubPeer, but also in research literature. Some immediately questioned the role of FNDC5 in muscle function and exercise (Timmons et al, 2012). Harold Erickson, cell biology professor at Duke University, published in 2013 his detailed criticisms, among which was the lack of validation for the novel irisin antibody which Spiegelman and Boström based their discovery on. Later on, Erickson teamed up with Steffen Maak, professor at Leibniz Institute for Farm Animal Biology, to show by the simple means of molecular weight analysis by western blot that the irisin antibody was unspecific and the ground-breaking Nature-result most likely an artefact (Albrecht et al 2015). Boström and Spiegelman apparently never bothered about proper antibody specificity controls: their experiments delivered exciting results, and no one questioned what exactly they were detecting.
Despite all counter-evidence against his high-impact discoveries, Spiegelman’s enormous (and as I was told, rather intimidating and oppressive) influence in the diabetes field hardly took a dent. Soon science journalists proclaimed “Irisin Redeemed”, all because the irisin discovery was reproduced, by antibody-independent method of mass spectrometry [update: as Maak comments below, also this paper relied on the faulty antibody]. The unbiased reproducer was actually Spiegelman himself (together with his original irisin-co-discoverer and Harvard department colleague Steven Gygi), but this hardly makes any difference to some scientists, as I learn again and again. The fact that no obesity or diabetes cures (or for that matter, no sufficiently trustworthy follow-up research) came out of irisin, seems hardly relevant. Certainly one can ignore an occasional blot duplication, which a PubPeer user reported for the Boström et al Nature paper now:
Science moves one irreproducible break-through discovery at a time. Prior to their irisin scoop, Boström and Spiegelman found a way to easily train an athlete’s heart by suppressing the function of the protein C/EBPβ. This was published also not somewhere, but in Cell (Boström et al 2010). There were some strange band duplications as well, apparently some of Boström’s western blots from his Diabetes paper with Olafsson found their way in his Cell paper with Spiegelman:
It seems, as long as patients don’t start dying, scientific misconduct is not only tolerated by academic institutions in Sweden (and elsewhere), but actively covered-up, while whistle-blowers are being threatened. It is all for a good cause after all: securing public research funding.
KI has a fine history of such behaviour, the Paolo Macchiarini scandal had all: dead and abused patients, fake medical records, harassed whistle-blowers, misconduct investigation disregarded and even trashed patient samples (for more, read my original report and this letter by a disappointed investigator Bengt Gerdin to the KI ex-rector Anders Hamsten). All because Swedish government was throwing big money for stem cell research around, and Macchiarini was a genius at catching it.
In Bostöm’s case, nobody died, thus he retains his position at KI. Apparently, KI hopes to cash in on his irisin patents with Spiegelman. Despite (or most likely, because of) his certified data fakery, Boström is also a fine money magnet: he received in an ERC starting grant of €2 Million, valid till end of 2017, to study (despite misconduct and irreproducibility) yes, what else: irisin. Though that probably non-existing hormone does not help at all against obesity or diabetes, it creates lots of cash out of thin air: Boström also pulled a “Collaborative Project Award” with the Mayo Clinic in US.
Then there was an up to SEK 4 Million Ingvar Carlsson Award from the Swedish Foundation for Strategic Research. Only last month, Boström was given another SEK 4 Million (~€420,000) from the Hållsten Research Foundation (where he is a fellow since 2014) to employ some research assistants. One wonders what exactly KI expects these junior scientists to learn under Boström’s tutelage.
Update 9.07.2016. A first hand source (identity known to me) gave me these details:
“Since the misconduct was only proven for a manuscript that hadn’t been published, Pontus kept his PhD title and his position at KI (under some conditions). During 2015 [his lab members, -LS] were told that Pontus decided to end the group and would leave KI by the end of the year. Not sure if this indeed was his own decision or whether KI wanted to have him go [Boström’s ERC funding was terminated on 31.12.2015, -LS]. Pontus still has an affiliation with KI for 2016 for economical and administrative matters.
Pontus left Academia. He’s an MD, has been working one day a week at a GP practice during his time at KI (http://old.slso.sll.se/SLPOtemplates/SLPOPage1____2683.aspx) and I assume that he now works there full-time”.
Update 19.07.2016 Boström was funded by ERC for almost 2 years, between the misconduct verdict by University of Gothenburg in early 2014 and its termination in December 2015. ERC has now answered my inquiry about how much he and KI received:
“the total ERC Starting Grant funding initially planned for the project ‘Irisin – a novel myokine protective against metabolic disease’ was €1,999,433.00 of which Dr Boström has received € 666,858.46″.