What turned a cancer researcher into a literature watchdog?

Sometime in the middle of 2015, Jennifer Byrne, professor of molecular oncology at the University of Sydney, began her journey from cancer researcher to a scientific literature sleuth, seeking out potentially problematic papers.

The first step was when she noticed several papers that contained a mistake in a DNA construct which, she believed, meant the papers were not testing the gene in question, associated with multiple cancer types.  She started a writing campaign to the journal editors and researchers, with mixed success. But less than two years later, two of the five papers she flagged have already been retracted.

When asked why she spent time away from bench research to examine this issue, Byrne told us: 

We wanted to make a reasonable comprehensive description of a phenomenon that if correct, is important. Many of those papers may be fine, but if there is any hint of a systematic issue in the literature that is not as described, that is really serious.

The five papers focused on one cancer gene, TPD52L2, which Byrne cloned back in 1998. It’s associated with breast cancer and acute lymphoblastic leukemia.

Working with her colleague Cyril Labbé at the University of Grenoble Alpes in France, who developed SciDetect software that detects fake scientific papers, Byrne found five papers that all appeared to use a faulty primer. She and Labbé recently published their results in the journal Scientometrics.

All of the flagged papers focused on the effect of TPD52L2 gene knockdowns in various cancer cell lines. The papers were published between 2014 and 2015, and have no common authors. However, most used the same faulty DNA sequence to create cell lines that do not even knock down the intended gene. The papers also contained similar figures.

Byrne and Labbé’s first step: Contact the editors of the journals which published the five papers: Cancer Biotherapy and Radiopharmaceuticals, Cell Biology International, the International Journal of Clinical and Experimental Medicine, and Cellular and Molecular Biology.

She expressed two main concerns: the DNA constructs did not target the TPD52L2 gene, and five independent groups used the same faulty DNA constructs.

Her initial analysis may be only the tip of the iceberg — looking through PubMed and Google Scholar, she found 30 more papers using the same faulty TPD52L2 primers.

One editor at Cellular and Molecular Biology, the first to respond, offered an opportunity for Byrne to write a letter to the editor. She declined, saying she wanted the journal to launch an investigation.  hen she emailed follow up questions, the editor never replied.

But ultimately, her efforts resulted in two retractions.

Here is the first notice, for “Lentivirus-mediated TPD52L2 depletion inhibits the proliferation of liver cancer cells in vitro,” published February 2015 by the International Journal of Clinical and Experimental Medicine:

Several similar articles were published within the similar time frame of this article as disclosed by a knowledgeable informant and subsequently confirmed by the Editorial Office. Although the authors claimed the data were original and the actual experiments were sourced out to a biotechnology company, they failed to response to editor’s further inquiries and concerns in a timely manner. Therefore, the entire article has been retracted in accordance with this journal’s policy and Editorial decision.

The 2015 paper has been cited once, according to Clarivate Analytics’ Web of Science, formerly part of Thomson Reuters.

Here’s another retraction, issued by Cell Biology International for “Knockdown of tumor protein D52-like 2 induces cell growth inhibition and apoptosis in oral squamous cell carcinoma:”

The above article, published online on 13 October 2014 in Wiley Online Library (http://onlinelibrary.wiley.com/doi/10.1002/cbin.10388/abstract), has been retracted by agreement between the authors, the journal Editor, Sergio Schenkman, and John Wiley & Sons Ltd. The retraction has been agreed because the authors discovered after publication that one of the cell lines described in the article had been unintentionally misidentified. The experiments described in the article as being conducted on Human Oral Squamous Cell Carcinoma cell line KB were in fact conducted on a Human Oral Epidermal-like Cancer cell line. The authors and publisher apologise for any inconvenience. References He Y, Chen F, Cai Y and Chen S (2015) Knockdown of tumor protein D52-like 2 induces cell growth inhibition and apoptosis in oral squamous cell carcinoma. Cell Biology International 39: 264-271. doi: 10.1002/cbin.10388.

The 2014 paper has been cited once.

The other three papers flagged by Byrne and Labbé are:

• Lentivirus-mediated knockdown of tumor protein D52-like 2 inhibits glioma cell proliferation, published 2014 by Cellular and Molecular Biology, cited twice.
• Tumor protein D52-like 2 accelerates gastric cancer cell proliferation in vitro, published 2015 by Cancer Biotherapy and Radiopharmaceuticals, not cited.
• Tumor protein D52-like 2 contributes to proliferation of breast cancer cells, published 2015 by Cancer Biotherapy and Radiopharmaceuticals, cited once.

We contacted the authors and editors of the journals that published the five papers. Like Byrne, we haven’t heard from any of the authors. But some editors responded.

Donald J. Buchsbaum, co-editor of Cancer Biotherapy and Radiopharmaceuticals, which published two of the papers, told us:

We are continuing with our investigation of the allegation, and at this time, we cannot determine how long that may be.

Sergio Schenkman, editor-in-chief of Cell Biology International, which has already retracted a paper,  told us that the authors agreed that the cell lines were not correct, and were willing to retract their paper.

In their Scientometrics paper, Byrne and Labbé make a fairly serious allegation: Given that the authors of one retracted paper admit they outsourced the experiments to a company, Byrne and  Labbé suspect the other researchers may be purchasing data – in this case with similar technical mistakes — to produce the similar papers.

The researchers analyzed the text of the papers and found that some manuscripts share more than 50 percent of their words, but those words were not in the same order, which makes these potential paper factories even more difficult to find.

The issue is particularly salient for Byrne who works in a cancer hospital where she sees patients receive treatment based on research. Although insisting that people are “innocent until proven guilty,” she told us:

People do cancer research to improve cancer sufferers’ lives. It’s not funny, the consequences (of fraudulent research) can be dire. This is the start of the pipeline that translates to better cures for patients.  If the start of the pipeline is basically a sham, then what’s the point in the end?

Like Retraction Watch? Consider making a tax-deductible contribution to support our growth. You can also follow us on Twitter, like us on Facebook, add us to your RSS reader, sign up on our homepage for an email every time there’s a new post, or subscribe to our daily digest. Click here to review our Comments Policy. For a sneak peek at what we’re working on, click here.