Another instalment from the series “Curing Cancer with Photoshop“. This time, the hero is Eric Lam, breast cancer researcher and professor of molecular Oncology at Imperial College in London. Lam’s lab consists by his own count of “12-18 researchers”, Imperial had to stress that it “is made up from 15 females”. However, a number of problematic papers from Lam lab have male first authors.
This cancer fighting hub at Imperial is funded by British national funders Medical Research Council (MRC), Cancer Research UK and Biotechnology and Biological Sciences Research Council (BBSRC). Major research focus is on the Forkhead Box proteins, in particular FOXM1, a transcription factor implicated in cancer development. A pedestrian approach many cancer researchers take, where similar-looking papers on slightly varying topics of the same gene are produced, none of them too unorthodox or earth-shattering to attract scrutiny as to how those results came to be. Lam’s science is not controversial for that, and with 250 papers it is the combined quantity and mediocrity which let his research stand unquestioned. Yet if one does look closely, as the data integrity sleuth Clare Francis did and posted it on PubPeer, one finds duplicated western blot bands and irregularly spliced gels. Maybe it is a novel molecular pathway of FOX proteins, to be harnessed for a cancer cure?
What this article shows, is actually nothing scandalous. It is simply the boring banality of bad science in cancer research which PubPeer is full of. Lam is just one of many successful academic oncologists who knows what is needed to successfully publish papers in respectable society journals.
We can start with an early work from Lam lab, where Sunters et al J Biol Chem 2003 proved that FOXO3a and FOXO1a proteins are involved in apoptosis of cancer cells treated with a chemotherapeutic agent. Not really a surprising result, under the circumstances, but it still needed some tweaking. In a certain cancer cell line, this paclitaxel treatment had a bizarre effect of duplicating, or even triplicating gel bands inside the loading control. Apparently, an extracellular Photoshop-defence mechanism against undesired results gets activated, which then provides FOXO3a and FOXO1a phosphorylation in a tubulin-western blot band-dependent matter. Amazing science at the forefront of war on cancer at Imperial.
The Fernandez De Mattos et al, Mol Cancer Ther, 2008 paper from the Lam lab is also an interesting example in this regard. Who would have expected that the chemotherapeutic agent cisplatin will not only reduce the phosphorylation of FOXO3a protein, but also lead to a duplication of a p-FOXO3a gel band? Serendipitous discoveries like this is what drives cancer research!
Or this paper Francis et al, Int J Oncology, 2009, where Lam found out the key role of FOXM1 in breast cancer. The effect is that universal, that cancer cell lines produce identical western blots even when treated with utterly different pharmacological inhibitors, and in different publications. This is why two western blot panels from that paper are found in Mc Govern et al, Mol Cancer Ther, 2009, another Lam paper published by American Association for Cancer Research (AACR).
The effect spreads even further inside the McGovern et al 2009 paper. 4 loading controls were re-used, but not the fifth one because it apparently already was delegated to moonlight in the parallel Francis et al 2009 paper.
The following Lam study gives insights into his mentorship practice. Jimmy Kwok did his PhD udner Lam and is now head of London oncology division with the private healthcare provider HCA Healthcare UK. He is also a registered Ashtanga yoga instructor and helps cancer patients by selling them his books on yoga. There is lots for Kwok to mediate on himself though, for example on his paper Kwok et al Mol Cancer Research 2010 and his PhD thesis from same year.
Can yoga explain that western blot bands showing the expression of an oncogene look identical between control and chemotherapeutic agent-treated cells? Especially that these gels bands were squeezed vertically to look smaller when the therapy began?
The exactly same figure, with same duplication, is present in Kwok’s PhD thesis, supervised by Lam. The thesis also has other issues, like those amazing reproducible flow cytometry plots, despite different time points:
This was what Lam published together with another Imperial colleague, Simon Wagner (who is now with University of Leicester). The paper Batlle et al Mol Immunology 2009, published with Elsevier, proves that in the end, anything goes and nothing matters. Nobody cares, or looks, or actually peer reviews. Some brightness touch-up suffices to make results from different experiments interchangeable inside the same figure. Leukaemia solved, the easy Imperial way.
Other concerns in Lam’s papers are hidden irregular gel splicing, mismatching loading controls, which both make western blot results rather unreliable, or occasional duplications like in Barnouin et al J Biol Chem 2002.
The bigger problem is that in cancer research, all these issues in Lam’s papers are nothing special. None of them. Those problematic western blots Lam published are boringly common. Just in UK, there are many other cases of photoshopped cancer research discussed on PubPeer, my site alone documents ICR London, Cardiff University , or even another lab at Imperial. Even more, Lam’s own superior Iain McNeish, Head of Division of Cancer of Department of Surgery and Cancer at Imperial published something rather inappropriate in his own youth, also at Imperial, in McNeish et al, Cancer Gene Ther, 2001. This is just one example of duplicated gel bands in that paper of his, more is on PubPeer.
Even more can be found on PubPeer in the works in of McNeish et al 2001 last author, Nick Lemoine, formerly professor of molecular pathology at Imperial and now Chair of the Bart’s Cancer Institute in London. That was also mostly posted by pseudonymous sleuth Clare Francis, who also informed the universities involved. Not that anyone cares, just like with ICR, also on Bart’s case it will the institute director investigating himself. Lemoine is also Chair of trustees at MRC, in case anyone is concerned about MRC’s attitude to research integrity, they can complain to Professor Lemoine.
It is basically a culture of mutual back-scratching and one hand washes the other in cancer research. Lam might be not the biggest shark in the pond, but one quickly sees that those above him and at the helm of British cancer research are, let’s put it this way, conflicted themselves. Yet this kind of bad science quickly leads to clinical trials, with millions of pounds, dollars or euros wasted, where thousands of cancer patients are given false hope and subjected to potentially dangerous therapies based on a flimsy, unreliable or plain fake lab data. Those same people unable to present a convincing western blot without the help of Photoshop, keep asking the public to donate to cancer research.
Maybe this is why there is so little progress relative to the enormous financial effort worldwide. What if the really radical approach to war on cancer would be not throwing good money after bad in Moonshot programmes, not inventing new crazy theories of cancer, not stripping patients of basic protection so new methods can be quickly tested on them on a hunch, but introducing some basic research integrity for a change? It would cost very little, in fact it would liberate enormous sums which more honest scientists could use for better science.
Update 25.11.2018. Same day and just a couple of hours after this article appeared, Lam’s Head of Department Iain McNeish went to PubPeer to share the results of an investigation on his own problematic paper with Nick Lemoine, in this comment. In brief, the gel band duplications were confirmed and original data was not available. This was exactly the reason why Imperial and Bart’s Cancer Institute decided not to retract or even to correct the paper. And anyway, some of the duplicated bands were, as McNeish said, “only control lanes”. McNeish announced to be now in charge of monitoring and punishing exactly same transgressions as his own paper shows.
In Britain, goats are appointed gardeners, by other goats.
Quote by McNeish:
“None of these issues in my opinion change the conclusions made in the papers. I see no merit in retracting these important publications or publishing an erratum at this late stage, although agree that there have been errors made in producing the figures.
It should also be noted that the original data was no longer available for review. […] The investigation stated that the conclusions made from the published studies were valid and important, even accounting for these reporting errors.
I recommend that measures are put in place to avoid such issues surrounding research data quality and reporting in future. […]
As first author, I accept the findings of this investigation. The Division of Cancer at Imperial College has now established a data integrity committee that will investigates suspected data irregularities and also to review the current practices relating to research data storage and reporting.”
If you are interested to support my work, you can leave here a small tip of $5. Or several of small tips, just increase the amount as you like (2x=€10; 5x=€25). Your generous patronage of my journalism will be most appreciated!
Haematologica. 2010 Jul;95(7):1081-9. doi: 10.3324/haematol.2009.017178 . Epub 2010 Feb 23.
Bone marrow mesenchymal stromal cells non-selectively protect chronic myeloid leukemia cells from imatinib-induced apoptosis via the CXCR4/CXCL12 axis.
Vianello F1, Villanova F, Tisato V, Lymperi S, Ho KK, Gomes AR, Marin D, Bonnet D, Apperley J, Lam EW, Dazzi F.
Department of Haematology, Kennedy Institute of Rheumatology, Imperial College, London, UK.
In the case of Eric Lam, I think you are being unfair to Imperial. They clearly have a very robust scientific fraud investigation procedure in place:
Click to access Annual-Research-Misconduct-Report-for-2015-16.pdf
They just need to apply it to Prof Lam.
Cough… I saw how UCL investigated Martin Birchall. Pity I can’t publish their preliminary report.
J Biol Chem. 2004 Mar 26;279(13):12005-8. Epub 2004 Jan 23.
AMP-activated protein kinase plays a role in the control of food intake.
Andersson U1, Filipsson K, Abbott CR, Woods A, Smith K, Bloom SR, Carling D, Small CJ.
Medical Research Council Clinical Sciences Centre, Cellular Stress Group and Endocrine Unit, Imperial College London, Hammersmith Campus, Du Cane Road, London W12 ONN, United Kingdom.
Triplicated data figure 3-25, page 194, PhD Thesis Jimmy Kwok Imperial College, London 2010.
If you click on figure 3-25 above to enlarge you can see that the signal inside the yellow rectangles is not uniform, but does have features. A line slanting downwards towards the right in the upper part of the rectangles, and a distinct spot in the lower left of the rectangles.
Figures 3-27 and 3-34, PhD Thesis Jimmy Kwok Imperial College, London 2010.
Minor point: figures 3-14 and 3-15, PhD Thesis Jimmy Kwok Imperial College, London 2010, seem to contain the same information.
Figure 3.1B, PhD Thesis Julie Millour, Imperial College London.
Click to access Millour-J-2013-PhD-Thesis.pdf
Figure 3.4, PhD Thesis H R Upekha Karunarathna, Imperial College, London 2017. PhD supervisor prof Eric Lam.
More similar than you might expect.
Figure 5.2, PhD Thesis H R Upekha Karunarathna, Imperial College, London 2017.
PhD supervisor prof Eric Lam.
Figure 3-2 (lowest panel), PhD Thesis H R Upekha Karunarathna, Imperial College 2017.
PhD supervisor prof Eric Lam.
Figure 4.8, PhD Thesis H R Upekha Karunarathna, Imperial College 2017.
PhD supervisor prof Eric Lam.
Oncogenesis. 2016 Aug 15;5(8):e252. doi: 10.1038/oncsis.2016.57.
RNF168 cooperates with RNF8 to mediate FOXM1 ubiquitination and degradation in breast cancer epirubicin treatment.
Kongsema M1, Zona S1, Karunarathna U1, Cabrera E2, Man EP3, Yao S1, Shibakawa A1, Khoo US3, Medema RH4, Freire R2, Lam EW1.
Department of Surgery and Cancer, Imperial College London, London, UK.
Unidad de Investigación, Hospital Universitario de Canarias, Instituto de Tecnologías Biomédicas, Ofra s/n, La Laguna, Tenerife, Spain.
Department of Pathology, Li Ka Shing Faculty of Medicine, The University of Hong Kong, Hong Kong SAR, China.
Division of Cell Biology, The Netherlands Cancer Institute, Amsterdam, The Netherlands.
Cancer Discov. 2012 Feb;2(2):156-71. doi: 10.1158/2159-8290.CD-11-0256.
The OPCML tumor suppressor functions as a cell surface repressor-adaptor, negatively regulating receptor tyrosine kinases in epithelial ovarian cancer.
McKie AB1, Vaughan S, Zanini E, Okon IS, Louis L, de Sousa C, Greene MI, Wang Q, Agarwal R, Shaposhnikov D, Wong JL, Gungor H, Janczar S, El-Bahrawy M, Lam EW, Chayen NE, Gabra H.
Ovarian Cancer Action Research Centre, Division of Cancer, Imperial College London Hammersmith Campus, London, United Kingdom.
Putting it together.
Several examples image manipulation in two PhD theses supervised by prof Eric W Lam:-
1. PhD Thesis Jimmy Kwok, 2010,
2. PhD Thesis H R Upekha Karunarathna, 2017,
3. Single instance on image triplcation, PhD Thesis Julie Millour, 2012/13.
PhD Thesis Jimmy Kwok, 2010.
Original document can be found here: https://spiral.imperial.ac.uk/handle/10044/1/5940
Figure 3-18C. Same data vertically stretched to represent different treatments (with and without drug) and different time points.
Figure 3-23. Same FACS profiles, or slightly stretched FACS proflies, to represent different time points.
Figure 3-25. Image triplication.
If you click on https://imgur.com/Ibj1qiK to enlarge you can see that the signal inside the yellow rectangles is not uniform, but does have features.
A line slanting downwards towards the right in the upper part of the rectangles, and a distinct spot in the lower left of the rectangles.
Figures 3-27 and 3-34. Cytoplasmic becomes nuclear, and nuclear becomes cytoplasmic.
Figures 3-14 and 3-15 contain the same information, the legends are the same, just number of words per line different.
PhD Thesis H R Upekha Karunarathna, 2017.
Original document can be found here: https://spiral.imperial.ac.uk/handle/10044/1/52459
Figure 3.4. Same FACS profiles for different treatments and different time points.
Highly problematic (two stage image manipulation).
Figure 5.2. 180 degree rotation and vertical stretching. Same data used to represent different things.
Figure 3.2C. Same image shifted down and across the page used to represent control samples from different cell lines.
Figure 4.8. Same FACS profile for three different treatments and two different time points.
Problematic, but only one instance in this thesis.
Image triplication PhD Thesis Julie Millour, 2012/13.
Original document can be found here: https://spiral.imperial.ac.uk/handle/10044/1/17849
Figure 3.1, see:
“Publication of a thesis consists of an overview that contains a “bar code” visualizing the amount of plagiarism, the exact bibliographic information of the theses (and if online, a link) and then a page overview that links to the fragments of plagiarism found on the page. Each fragment is classified as to the kind of plagiarism and must be verified by a second person before it is acceptable. Interesting parts are linked to separately and there is a final report published when the group feels the work is done.”
Problem is that text can be any old story, plagiarised or not. More effective would be analysis of priomary data in PhD theses.
This would help put an end to the marketisation of education by the corporate state (the universities).
Page 161, PhD Thesis Jimmy Kwok, Imperial College, London 2010. PhD supervisor prof Eric Lam.
Upper 2 panels, figure 3.2C, page 130, PhD Thesis H R Upekha Karunarathna, Imperial College 2017.
PhD supervisor prof Eric Lam.
Straight edges in some parts of some lanes, but not in other lanes.
Bands becoming wider above straight edge, lane 2 right panel.
Resorting to “image compression” does not answer these observations.
Blood. 2009 Mar 12;113(11):2498-507. doi: 10.1182/blood-2008-06-161281 . Epub 2008 Dec 3.
Regulation of multiple myeloma survival and progression by CD1d.
Spanoudakis E1, Hu M, Naresh K, Terpos E, Melo V, Reid A, Kotsianidis I, Abdalla S, Rahemtulla A, Karadimitris A.
Department of Haematology, Imperial College Healthcare National Health Service (NHS) Trust, Hammersmith and St Mary’s Hospital, Imperial College London, London, UK.
Cancer Res. 2001 Nov 1;61(21):8005-13.
Direct relation between BCR-ABL tyrosine kinase activity and cyclin D2 expression in lymphoblasts.
Deininger MW1, Vieira SA, Parada Y, Banerji L, Lam EW, Peters G, Mahon FX, Köhler T, Goldman JM, Melo JV.
Department of Haematology, Faculty of Medicine, Imperial College of Science, Technology and Medicine, Hammersmith Hospital, London W12 0NN, United Kingdom.
I used to think that Imperial College was better than UCL. I no longer believe that.
Are the publisshing practices noted in this article, and the comments, one of the reasons they are coining it in?
“Imperial College London secured the largest amount of open-call funding in 2017-18, according to a Times Higher Education analysis of data provided by six research councils – £91.5 million, up from £72.7 million last year, a 25.8 per cent increase over a period in which the overall amount of funding distributed dropped by 8.6 per cent.”
“UCL remained in second place, securing £89.2 million, up from £76.4 million.”.
Problematic data figure 3.2A, page 96, PhD Thesis Vidhya Varghese, Imperial College 2016.
PhD Supervisor Dr Laura Kenny, PhD co-supervisor prof Eric Lam.
Original document can be found here: https://spiral.imperial.ac.uk/handle/10044/1/53384
Figure 3.2A. See:
Blood. 2001 May 1;97(9):2604-10.
The influence of INK4 proteins on growth and self-renewal kinetics of hematopoietic progenitor cells.
Lewis JL1, Chinswangwatanakul W, Zheng B, Marley SB, Nguyen DX, Cross NC, Banerji L, Glassford J, Thomas NS, Goldman JM, Lam EW, Gordon MY.
LRF Centre for Adult Leukaemia, Department of Haematology, Imperial College School of Medicine, Hammersmith Campus, London, United Kingdom.