Inability to reproduce Dutch grad student’s data prompts two retractions from the cancer lit

We are watching an intriguing case out of the Netherlands, involving a young researcher whose dubious results have led to the retraction of a pair of papers.

The retracted articles, which appeared in 2008 in Cancer Research and the British Journal of Cancer, come from the lab of the prominent Dutch scientist Ed Roos, of the Netherlands Cancer Institute in Amsterdam. Both papers addressed the actions of certain chemokine receptors — molecules on cell surfaces that interact with blood proteins involved in the immune response — on the behavior of tumor cells.

The first author on each paper was Joost Meijer, at the time a graduate student in Roos’ shop.

The retraction notices contain essentially the same information, although in the case of the BJC article — “Effect of the chemokine receptor CXCR7 on proliferation of carcinoma cells in vitro and in vivo” — the letter is quite personal. Dated Jan. 4, 2011, it reads:

After a thorough investigation, we have recently concluded that key results contributed by the first author cannot be reproduced. As the senior and corresponding author, I have therefore decided to retract this paper.

The co-author has been informed, and agrees with this decision. I emphasise that there is no doubt whatsoever about her contribution.

I sincerely apologise for the inconvenience this may have caused to readers of this journal.

Dr Ed Roos

Division of Cell Biology

The Netherlands Cancer Institute

That paper, originally published in 2008, was cited 17 times, according to Thomson Scientific’s Web of Knowledge.

Here’s Cancer Research’s version, published Jan. 18 (we added a link to the original abstract):

We wish to retract the article titled “The Chemokine Receptor CXCR6 and Its Ligand CXCL16 Are Expressed in Carcinomas and Inhibit Proliferation“, which was published in the June 15, 2008 issue of Cancer Research (1). After a thorough institutional investigation, it was recently concluded that key results contributed by the first author cannot be reproduced. These results include the detection of CXCR6 and CXCL16 on the cell surface by FACS analysis, the knockdown of these proteins, and the effects of this knockdown on proliferation rates in vivo and in vitro. Five of the 6 authors agreed to this Retraction. Attempts on the part of the journal office to contact the first author, Joost Meijer, to determine whether or not he agreed with the Retraction were unsuccessful.

Janneke Ogink

Bas Kreike

Dimitry Nuyten

Karin E. de Visser

Ed Roos

The Netherlands Cancer Institute, Amsterdam, the Netherlands

That paper was cited 13 times since it was published in 2008. We should note — as you may have noticed if you clicked on the original abstract links — that the studies do not indicate they are retracted.

Setting aside the Case of the Missing Scientist for a minute, it turns out that the Netherlands Cancer Institute has been aware of the matter for months. How many isn’t clear, but we found the following statement on the center’s website, dated August 30:

The Netherlands Cancer Institute (NKI) has asked the editorial staff of two scientific journals to withdraw two of its publications. Both articles describe fundamental research on the mechanism of metastasis, carried out by one of NKI’s PhD students. After the first indications that the published results could not be reproduced an internal committee was installed to investigate the matter. The committee concluded that the PhD student conducted and documented his research in an inadequate way, below academic standards.

The NKI Board of Directors has informed the Dutch Cancer Society who has funded this research. The grant will be fully refunded. The PhD student is no longer working at the NKI. The NKI declares that no patients were involved in this research. The withdrawal of the articles has no consequences for existing or future medical treatments.

What makes the Cancer Research retraction a little more interesting is that Roos has been an associate editor of the journal since 2004, according to his website (update: an observant reader noted that the latest iteration of the journal’s ed board does not list Roos among the members). Somewhat surprisingly, Roos still lists both retracted articles among his group’s “key” publications.

We’ve reached out to Roos, but haven’t heard back. [Update, 9 a.m. Eastern, 1/26/110: See comments from Roos below.] Same with the two journals. We’ll update this post when we learn more. We’re particularly curious to find out when and how Roos learned that Meijer had been messing around in his lab. We’d also like to know when the two journals found out about the problematic papers — and why it took more than four months from the time the NKI issued its statement until the retractions appeared.

As for Meijer, according to his Linkedin page  he left the cancer institute in February 2009 for a post-doc position at the University of Amsterdam’s Academic Medical Centre. We confirmed with a scientist there that Meijer was working in the lab of Connie Bezzina, a cardiologist at AMC, until February 2010 but is no longer on the institution’s rolls (his Linkedin page states that he started in March 2009 and is still there). We couldn’t find any publications that listed the two as co-authors.

From Meijer’s biosketch on Linkedin (spelling unaltered):

Arritmogenic Right Ventricular Cardiomiopathy or ARVC is a hereditary hart disease and is seen 3 times as much in females and in males. There are many pathological chances in the right ventricle and one of them is the exchange of the myocardium by fatty fybrotic tissue. We are mainly interested in DSG2, a desmosomal protein, which was recently found to be mutated in about 20% of the ARVC patient. I’m trying to develop new mouse models to investigate the role DGS2 mutations in development of ARVC.