Nature editors retract influential cancer paper with “unreliable” data 

Editors at Nature have retracted a 2015 paper on breast cancer metastases citing trouble with the data in the supplementary materials. 

The paper, “The hypoxic cancer secretome induces pre-metastatic bone lesions through lysyl oxidase,” was first published in May 2015 and has been cited 352 times, according to Clarivate’s Web of Science. 

This marks the second retraction for corresponding author Janine Erler, a professor in cancer biology at the University of Copenhagen. As previously reported by Retraction Watch, Nature in 2020 pulled a 2006 paper on which she was first author because of “image anomalies” and the absence of original data. Two other papers co-authored by Erler have been corrected and one additional paper has an expression of concern.

The now-retracted 2015 paper claimed to have found evidence that the enzyme lysyl oxidase (LOX) can be a driver of metastasis to the bone of some breast cancers. A 2016 report published in the Journal of Bone and Mineral Research disputed some of the claims made in the paper, and in March 2018, an anonymous commenter on PubPeer pointed out issues with the paper’s western blots. 

According to the April 2023 retraction notice: 

The Editors have retracted this paper because, in their view, data in Extended Data Fig. 1d and Extended Data Fig. 4a are unreliable. Issues with the original antibody used in Extended Data Fig. 1d and 4a have been identified. The western blots showing an increase in secreted LOX in bone-trophic and hypoxic MDA-MB-231 cells, and shRNA knockdown of LOX in 4T1 cells, present unreliable bands. R.L. agrees with the retraction. All other authors disagree with the retraction and believe the findings are still valid.

When asked to comment on the retraction, Erler forwarded a press release that she said represents the views of the 14 authors who do not agree with the decision to retract the paper. 

Here’s the statement: 

The University of Sheffield, Institute of Cancer Research, University of Copenhagen, and  Technical University of Denmark, have independently evaluated the evidence regarding the “issue” originally raised on PubPeer around two extended figure Western blots where the legends did not clearly state that the beta-actin blots were sample processing controls. 

Our collective assessment is that there is insufficient scientific justification for a retraction. Rather, a simple text correction, as originally proposed and agreed with Nature, should have been allowed to update the original figure legends. 

We believe a major conflict of interest has significantly influenced this case. All but one of the authors agreed to this previously approved correction and, similarly, all the authors, with the exception of the dissenting author, opposed to the retraction. 

We stand by the validity of the findings and support the 14 authors opposed to the retraction.

Anders Lund, director of the University of Copenhagen’s Biotech Research and Innovation Centre (BRIC) where Erler is a group leader, said the Danish Board on Research Misconduct, a government agency, concluded there was no research misconduct. 

“Prof. Erler was criticized for having mislabeled the supplementary figures,” he said of the board’s finding. “But there was no evidence of fraud in this conduct.”

In a call with Retraction Watch, Lund said the “major conflict of interest” cited in the statement refers to Erler’s ex-husband, Rune Linding, now a researcher at the Humboldt University of Berlin. Linding is the only author who agreed with the editor’s decision to retract the paper. 

“There were attempts earlier to settle this in a different way from retraction, corrections and so on,” said Lund, who is not one of the paper’s authors. “As far as I understand, Linding did not accept those suggestions.”

Linding told us: 

That is a false claim. The data issues were investigated by an international review panel and antibody experts who advised the Nature editors in their decision. The editors made their decision independently of course and I have seen no data or evidence suggesting that it is not the right decision.

Linding deferred to James Longden, a drug discovery biologist at e-therapeutics PLC, for further comment. Longden was a biologist at Linding’s lab at the time the research was being conducted, although he is not an author of the retracted paper. 

About a year ago, Erler failed to replicate the western blots after being asked to do so by Nature editors, according to Longden. 

“If the data could have been replicated, the paper would not be retracted,” Longden said in a call with Retraction Watch. “But there is data in that paper that cannot be repeated, and so therefore, the editors think it should be retracted and I agree with them.”

Co-corresponding author Alison Gartland, a professor of biology at the University of Sheffield, declined to comment on the record. 

A spokesperson for Nature told Retraction Watch:

I hope you will appreciate that, owing to our confidentiality policy, we cannot discuss the specific details regarding the editorial handling of individual cases.

Like Retraction Watch? You can make a tax-deductible contribution to support our work, follow us on Twitter, like us on Facebook, add us to your RSS reader, or subscribe to our daily digest. If you find a retraction that’s not in our database, you can let us know here. For comments or feedback, email us at [email protected].